.

The Art of C++

Herbert Schildt

McGraw-Hill/Osborne
2100 Powell Street, 10th Floor
Emeryville, California 94608
U.S.A.

To arrange bulk purchase discounts for sales promotions, premiums, or fund-raisers, please contact
McGraw-Hill/Osborne at the above address. For information on translations or book distributors outside the U.S.A.,
please see the International Contact Information page immediately following the index of this book.

The Art of C++

Copyright © 2004 by The McGraw-Hill Companies. All rights reserved. Printed in the United States of America. Except

籍
as permitted under the Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form
or by any means, or stored in a database or retrieval system, without the prior written permission of publisher, with the
exception that the program listings may be entered, stored, and executed in a computer system, but they may not be

书
reproduced for publication.

1234567890 CUS CUS 01987654

典 书
ISBN 0-07-225512-9

Publisher

经 读
Brandon A. Nordin

+
Vice President & Associate Publisher

不
Scott Rogers

C+
Editorial Director

典
Wendy Rinaldi

Project Editor

经
Jenn Tust

Acquisitions Coordinator

无
Athena Honore

Technical Editor
Fahad Gilani

Copy Editor
Sally Engelfried

Proofreader
Marian Selig

Indexer
Sheryl Schildt

Composition
Carie Abrew, Lucie Ericksen

Illustrator
Melinda Lytle, Kathleen Edwards, Gregg Scott

Series Design
Roberta Steele
Cover Design
Jeff Weeks

This book was composed with Corel VENTURA™ Publisher.

Information has been obtained by McGraw-Hill/Osborne from sources believed to be reliable. However, because of the
possibility of human or mechanical error by our sources, McGraw-Hill/Osborne, or others, McGraw-Hill/Osborne does
not guarantee the accuracy, adequacy, or completeness of any information and is not responsible for any errors or
omissions or the results obtained from the use of such information.

About the Author

Herbert Schildt is a leading authority on the C, C++, Java, and C# languages, and is a master Windows programmer.
His programming books have sold more than 3 million copies worldwide and have been translated into all major
foreign languages. He is the author of numerous C++ bestsellers, including C++: The Complete Reference, C++ From
the Ground Up, C++: A Beginner’s Guide, and STL Programming From the Ground Up. His other bestsellers include
C: The Complete Reference, Java 2: The Complete Reference, and C#: The Complete Reference. Schildt holds a
master’s degree in computer science from the University of Illinois. He can be reached at his consulting office at (217)
586-4683.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Preface
Since the early days of FORTRAN, the on-going development of computer languages has followed what can best be
described as an evolutionary path, with the efforts of the past shaping developments in the future. In this process,
weak features fell away and misguided efforts terminated in dead-ends. Over the years, these evolutionary forces
distilled into undiluted form, the pure essence of what a programming language should be. The result was C++, and no
other language holds a more important place in the history programming.

There are many reasons for the success of C++. Its syntax is terse, yet elegant, its object- model is streamlined and
conceptually clean, and its libraries are carefully crafted and interlocking. These features by themselves, however, are
not what earned C++ a place in history. Instead, it is these things coupled with the power that C++ puts into the hands
of the programmer. No language before or since has given the programmer more direct control over the computer.
With C++, the programmer is master of the machine—and this is the way that all programmers want it.

No boundaries, no limits, no constraints. That’s C++.

What’s Inside

书 籍
This book is different from most other books on C++. Whereas other books teach the basics of the language, this book
shows how to apply it to a wide range of interesting, useful, and, at times, mysterious programming tasks. In the

典 书
process, it displays the power and elegance of the C++ language. Thus, it is through the art of C++ that the artistry of
C++’s design is displayed.

经 读
In general, the book contains two types of applications. The first type I call “pure code” because they focus on
expanding the C++ programming environment, itself. The garbage collector in Chapter 2, the thread control panel in

+
Chapter 3, and the custom STL container in Chapter 8 are examples. The second type shows how C++ can be applied

不
to a variety of computing tasks. For example, Chapter 5 develops a restartable Internet downloader, Chapter 6 shows

+
how to create financial applications, and Chapter 8 applies C++ to Artificial Intelligence.

C 典
The book ends with a unique and interesting piece of code: the Mini C++ interpreter, which interprets a small subset of
C++. Mini C++ gives insight into how the keywords and syntax of C++ work together to create the grammar of the
language. Moreover, it lets you get “inside the language,” showing some of the reasons behind C++’s design. Although

经
Mini C++ is fun to use as-is, it can also serve as a starting point for your own language development, or be adapted to
work as an interpreter for any other language.

无
Each chapter develops code that you can use as-is, without changes. For example, the garbage collector in Chapter 2
is applicable to many programming tasks. However, the real benefits result when you use the applications as starting
points for your own development. For example, the Internet file downloader in Chapter 8 could be enhanced to start a
download at a specific time, or to watch a download site, waiting to download until an updated file is posted. In general,
think of the various programs and subsystems as launching pads for your own projects.
 Knowledge of C++ Is Assumed
This book assumes that you have a solid grounding in the fundamentals of the C++ language. You should be able to
create, compile, and run C++ programs. You should be able to use pointers, templates, and exceptions, understand
copy constructors, and be familiar with the most common parts of the standard library. Thus, this book assumes that
you have the skills that one would acquire in a course on C++.

If you need to refresh or enhance your basic knowledge, I recommend the following books.

C++ From the Ground Up

C++: A Beginner’s Guide

籍
C++: The Complete Reference

All are published by McGraw-Hill/Osborne.

典 书 书
+ 经 不 读
C+ 典
无 经
 Don’t Forget: Code on the Web
Remember, the source code for all of the examples and projects in this book is available free-of-charge on the Web at
www.osborne.com.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 More From Herb Schildt
The Art of C++ is just one in the Herb Schildt series of programming books. Here are some others that you will find of
interest.

To learn more about C++, we recommend the following.

C++: The Complete Reference

C++: A Beginner’s Guide

Teach Yourself C++

籍
C++ From the Ground Up

书
STL Programming From the Ground Up

书
To learn about Java programming, you will find these books especially helpful.

经 典 读
Java 2: The Complete Reference

+ 不
The Art of Java (Co-authored with James Holmes)

C+
Java 2: A Beginner’s Guide

典
Java 2: Programmer’s Reference

无 经
To learn about C# we suggest the following books:

C#: A Beginner’s Guide

C#: The Complete Reference

If you want to learn about the C language, the foundation of all modern programming, then the following titles will be of
interest.

C: The Complete Reference

Teach Yourself C
 Chapter 1: The Power of C++

Overview
C++ is all about power: the power to control the machine at its lowest level, the power to produce highly optimized
code, the power to interact directly with the operating system. This power runs deep and wide. With C++ you have
detailed control over objects, including their creation, destruction, and inheritance; access to pointers; and support for
low-level I/O. You can add new features by defining classes and overloading operators. You can construct your own
libraries and hand-optimize code. You can even “break the rules” when necessary. C++ is not a language for the timid.
It is for the programmer who demands and deserves the world’s most powerful programming language.

Of course, C++ is not just raw power. It is power channeled, focused, and directed. Its careful design, rich libraries,
and subtle syntax yield a programming environment that is both flexible and agile. Although C++ excels in the creation
of high-performance system code, it is suitable for any type of programming task. For example, its string handling is

籍
second to none, its math and numeric capabilities make it excellent for scientific programming, and its ability to
produce fast object code makes it perfect for any CPU-intensive task.

书
The purpose of this book is to demonstrate the power, scope, and agility of C++. It does so by applying it to a varied
cross section of applications. Some of the applications demonstrate the capabilities of the language itself. These are
called “pure code” examples because they show the expressiveness of the C++ syntax and the elegance of its design.

典 书
The garbage collector in Chapter 2 and the C++ interpreter in Chapter 9 are examples. Other applications illustrate the
ease with which C++ can be applied to common programming tasks. For example, the download manager in Chapter

经 读
5 illustrates the ability of C++ to create high-performance networked code. Chapter 6 applies C++ to various financial
computations. Collectively, the applications show the wide ranging versatility of C++.

+ 不
Before moving on to the applications, however, it is good to reflect for a moment on the things that make C++ such a

+
great language. To this end, this chapter takes some time to point out several of the features that define the “power of
C++.”

C 典
无 经
 A Terse yet Rich Syntax
If there is one fundamental defining characteristic of C++, it is the terseness of its syntax. C++ defines only 63
keywords. In a seeming contradiction, much of the power of C++ is derived from not building into the language more
features than are necessary. Instead, C++ defines a rich but compact syntax that provides the control statements,
operators, data types, and object-oriented features necessary for any modern language—but not more! Thus, the C++
syntax is clean, consistent, and uncluttered.

This lean philosophy has two important benefits. First, the keywords and syntax of C++ apply to all environments in
which C++ can be used. That is, the core features of C++ are universally available to all applications, independent of
the execution environment. Features that are sensitive to the execution environment, such as multithreading, are left to
the operating system, which is best able to handle them efficiently. Thus, C++ does not attempt a “one size fits all”
solution, which can lead to degraded runtime performance.

Second, a streamlined, logically consistent syntax allows complex constructs to be expressed with clarity. This is an
important advantage in a world in which programs have grown to truly enormous sizes. While it is, of course, true that

籍
a poor programmer will write poor C++ code, a good programmer can create amazingly clear and concise code. This
ability to transparently represent complicated logic is one of the reasons that the C++ syntax has become the near

书
universal language of programming.

经 典 读 书
+ + 不
C 典
无 经
 Powerful Libraries
Of course, the modern programming environment requires many features beyond those supported by the C++
keywords and syntax. C++ provides access to those features in a truly elegant way: through its standard library. C++
defines what is arguably the most well-designed library available in any mainstream language. Its function library,
which is a carryover from C, contains a wide-ranging array of non-object-oriented functions, such as char * string
handling, character manipulation, and conversion functions, which are used by programmers everywhere. The C++
class library provides object-oriented support for I/O, strings, and the STL, among others.

Because of its reliance on library routines rather than keywords, new functionality can be added to C++ by simply
expanding its library rather than by inventing new keywords. This makes it possible for C++ to adapt to changes in
programming environments without requiring changes to the core language. Thus, C++ combines two seemingly
conflicting traits: stability and flexibility.

Even in its function and class libraries, C++ takes a lean, “less is more” approach, which avoids the “one-size-fits-all”
trap. The libraries offer only those features that can be reasonably implemented for a wide range of programming

籍
environments. For features unique to a specific environment, C++ enables access to the operating system. Thus, the
C++ programmer has access to all features of the execution platform. This approach lets you write highly efficient

书
code that best utilizes the attributes and capabilities of the execution environment.

经 典 读 书
+ + 不
C 典
无 经
 The STL
There is one part of the standard class library that is so important that it deserves separate attention: the Standard
Template Library (STL). The creation of the STL was a pivotal event that changed the way that programmers think
about and use a language’s library. Its impact was so profound that it has influenced the design of subsequent
languages. For example, the Collections frameworks of both Java and C# are directly patterned after the STL.

At its core, the STL is a sophisticated set of template classes and functions that implement many popular and
commonly used data structures, which the STL refers to as containers. For example, it includes containers that
support vectors, lists, queues, and stacks. Because the STL is constructed from template classes and functions, its
containers can be applied to nearly any type of data. Thus, the STL provides off-the-shelf solutions to a wide variety of
programming problems.

Although its practical importance to the C++ programmer is not to be underestimated, there is a larger reason that the
STL is important. It was in the vanguard of the software component revolution. From the beginning, programmers have
sought ways to reuse their code. Because development and debugging is a costly process, code reuse is highly

籍
desirable. In the early days, code reuse was achieved by cutting and pasting source code from one program to
another. (Of course, this approach is still used today.) Next, programmers created reusable function libraries, such as

书
those provided by C++. This was soon followed by standardized class libraries.

Building on the foundation of a class library, the STL took the concept one important step further: it modularized the

典 书
library into generic components that are applicable to a wide range of data. Furthermore, because the STL is
extensible, you can define your own containers, add your own algorithms, and even adapt the built-in containers. The
ability to extend, adapt, and reuse functionality is the essence of the software component.

+ 经 读
Today, as the component software revolution is nearing completion, it is easy to forget that it was the STL that laid
much of the groundwork, including modularized functionality, standardized interfaces, and extensibility through

不
inheritance. The history of computing will chronicle the STL as one of the major milestones in language design.

C+ 典
无 经
 The Programmer Is in Control
There are two competing philosophies of programming languages. One says that the language should prevent a
programmer from causing problems by avoiding features that can cause problems in the first place. Although this
sounds good, it usually results in certain powerful yet potentially risky features being restricted, diluted, or left out
entirely. Two examples of such features are pointers and explicit memory allocation. Pointers are considered risky
because they are often misused by beginning programmers and can (in some cases) be used to breach security
barriers. Explicit memory allocation (such as through new and delete) is considered risky because a programmer might
unwisely allocate large blocks of memory, or forget to release memory that is no longer needed, resulting in a memory
leak. Although both features come with risks, they also give you, the programmer, detailed control and the ability to
create highly efficient code. Fortunately, C++ does not subscribe to this philosophy.

The second philosophy, and the one to which C++ adheres, is that “the programmer is king.” This means that the
programmer is in control. It is not the job of the language to prevent you from being a bad programmer. Rather, the
primary goal of the language is to give the programmer an agile, unobtrusive environment in which to work. If you are
a good programmer, your work will reflect it. If you program poorly, your work will reflect that, too. In essence, C++

籍
gives you the power and then gets out of the way. A C++ programmer never has to “fight the language.”

书
Obviously, most programmers prefer the philosophy of C++!

经 典 读 书
+ + 不
C 典
无 经
 Detailed Control
C++ does more than just put you, the programmer, in control, it gives you detailed control. For example, consider the
increment operator: ++. As you know, C++ defines both a prefix and a postfix version. The prefix version increments
the operand before obtaining its value. The postfix version increments the operand after obtaining its value. This
allows you to gain precise control over how expressions involving the ++ are executed. Another example of detailed
control is the register specifier. By modifying a variable declaration with register, you are telling the compiler to
optimize access to that variable. Thus, you can control which variables receive the highest priority for optimization.
The fine-grained control that C++ gives the programmer is one reason C++ has effectively replaced assembler as the
language of choice for programming system code.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Operator Overloading
One of the most important features in C++ is operator overloading because it supports type extensibility. Type
extensibility is the attribute that enables you to add and fully integrate new data types into the C++ programming
environment. Type extensibility is built on two features. The first is the class, which lets you define a new data type.
The second is operator overloading, which lets you define what the various operators mean relative to a class type.
Through operator overloading and classes, you can create a new data type and then operate on that type in the same
way that you operate on built-in types: through operators.

Type extensibility is very powerful because it makes C++ an open system rather than a closed one. For example,
imagine that you need to manage three-dimensional coordinates. You can do this by creating a new data type called
ThreeD and then defining various operations on objects of that type. For example, you could use + to add two ThreeD
coordinates, or = = to determine if two sets of coordinates are equal. Thus, you can write code that operates on ThreeD
just as you would any of the built-in types, as shown here:
ThreeD a(0, 0, 0), b(1, 2, 3), c(5, 6, 7);

籍
a = b + c;
// ...

书
if(a == c) // ...

Without operator overloading, operations on ThreeD objects would have to be handled by calls to functions, such as

典 书
addThreeD( ) or isEqualThreeD( ), which is a far less pleasing approach.

+ 经 不 读
C+ 典
无 经
 A Clean, Streamlined Object Model
The C++ object model is a masterpiece of brevity! In the ISO standard for C++, the object model is described in less
than a page (six paragraphs, to be precise). Within those few paragraphs, the standard manages to explain the
essence of an object, its lifetime, and polymorphism. For example, the standard gives this definition for an object, “An
object is a region of storage.” It is this sort of fundamental simplicity that makes the C++ object model so remarkable.

Of course, the syntax and semantics necessary to support objects, including their creation, destruction, inheritance,
and so on, consume many, many pages in the standard. However, this is because of the depth of control that C++
gives you over objects, not because of quirks or inconsistencies of the object model. Moreover, because of the
elegance of its design, the C++ object model is the pattern used by Java and C#.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 The Legacy of C++
Invented by Dennis Ritchie in the 1970s, C marked the beginning of a radical transformation in programming. Although
some earlier languages, most notably Pascal, had achieved significant success, it was C that established the
paradigm that would influence a generation of computer languages. With C, the modern age of programming began.

Soon after the creation of C, a new factor emerged: object-oriented programming (OOP). Although we take OOP for
granted today, at the time of its invention, it was a major step forward. The object-oriented philosophy quickly captured
the imagination of programmers because it offered a powerful, new way to approach the job of programming. At that
time, programs were becoming larger and their complexity was growing. Some way to handle this complexity was
needed, and OOP offered a solution. OOP enabled large and complicated programs to be framed in terms of
compartmentalized units of functionality (objects). This allowed a complex system to be reduced to more manageable
parts. The problem was that C did not support objects.

Designed by Bjarne Stroustrup, C++ was built on the foundation of C. To C, Stroustrup added the keywords and
syntax needed for object-oriented programming. By adding object- oriented features to the already popular C

籍
language, Stroustrup made it possible for thousands of programmers to move forward to OOP. With the creation of
C++, the modern age of programming was fully realized. In one masterful stroke, Stroustrup created the world’s most

书
powerful computer language and charted the course for future language development.

Although the legacy of C++ is just beginning, it has already led to the creation of two important languages: Java and

典 书
C#. Despite some minor differences, the syntax, object model, and overall “look and feel” of Java and C# are nearly
identical to that of C++. Furthermore, the libraries of Java and C# reflect the design of those in C++, and the
Collections frameworks of Java and C# are directly derived from the STL. The groundbreaking design of C++

经 读
reverberates loudly throughout computing.

+
The power that C++ puts in the hands of the programmer is why C++ became important. Its far-reaching influence is

不
the reason it remains the preeminent language for programmers worldwide. Thus, the ultimate power of C++ is its

+
legacy.

C 典
无 经
 Chapter 2: A Simple Garbage Collector for C++

Overview
Throughout the history of computing, there has been an ongoing debate concerning the best way to manage the use
of dynamically allocated memory. Dynamically allocated memory is memory that is obtained during runtime from the
heap, which is a region of free memory that is available for program use. The heap is also commonly referred to as
free store or dynamic memory. Dynamic allocation is important because it enables a program to obtain, use, release,
and then reuse memory during execution. Because nearly all real-world programs use dynamic allocation in some
form, the way it is managed has a profound effect on the architecture and performance of programs.

In general, there are two ways that dynamic memory is handled. The first is the manual approach, in which the
programmer must explicitly release unused memory in order to make it available for reuse. The second relies on an
automated approach, commonly referred to as garbage collection, in which memory is automatically recycled when it is

籍
no longer needed. There are advantages and disadvantages to both approaches, and the favored strategy has shifted
between the two over time.

书
C++ uses the manual approach to managing dynamic memory. Garbage collection is the mechanism employed by
Java and C#. Given that Java and C# are newer languages, the current trend in computer language design seems to
be toward garbage collection. This does not mean, however, that the C++ programmer is left on the “wrong side of

典 书
history.” Because of the power built into C++, it is possible—even easy—to create a garbage collector for C++. Thus,
the C++ programmer can have the best of both worlds: manual control of dynamic allocation when needed and

经 读
automatic garbage collection when desired.

This chapter develops a complete garbage collection subsystem for C++. At the outset, it is important to understand

+ 不
that the garbage collector does not replace C++’s built-in approach to dynamic allocation. Rather, it supplements it.

+
Thus, both the manual and garbage collection systems can be used within the same program.

C
Aside from being a useful (and fascinating) piece of code in itself, a garbage collector was chosen for the first example

典
in this book because it clearly shows the unsurpassed power of C++. Through the use of template classes, operator
overloading, and C++’s inherent ability to handle the low-level elements upon which the computer operates, such as

经
memory addresses, it is possible to transparently add a core feature to C++. For most other languages, changing the
way that dynamic allocation is handled would require a change to the compiler itself. However, because of the
unparalleled power that C++ gives the programmer, this task can be accomplished at the source code level.

无
The garbage collector also shows how a new type can be defined and fully integrated into the C++ programming
environment. Such type extensibility is a key component of C++, and it’s one that is often overlooked. Finally, the
garbage collector testifies to C++’s ability to “get close to the machine” because it manipulates and manages pointers.
Unlike some other languages which prevent access to the low-level details, C++ lets the programmer get as close to
the hardware as necessary.
.

Comparing the Two Approaches to Memory Management

Before developing a garbage collector for C++, it is useful to compare garbage collection to the manual method that is
built-in to C++. Normally, the use of dynamic memory in C++ requires a two-step process. First, memory is allocated
from the heap via new. Second, when that memory is no longer needed, it is released by delete. Thus, each dynamic
allocation follows this sequence:

p = new some_object;

// ...

delete p;

籍
In general, each use of new must be balanced by a matching delete. If delete is not used, the memory is not
released, even if that memory is no longer needed by your program.

书
Garbage collection differs from the manual approach in one key way: it automates the release of unused
memory. Therefore, with garbage collection, dynamic allocation is a one-step operation. For example, in Java
and C#, memory is allocated for use by new, but it is never explicitly freed by your program. Instead, the

典 书
garbage collector runs periodically, looking for pieces of memory to which no other object points. When no other
object points to a piece of dynamic memory, it means that there is no program element using that memory.

读
When it finds a piece of unused memory, it frees it. Thus, in a garbage collection system, there is no delete

经
operator, nor a need for one, either.

+ 不
At first glance, the inherent simplicity of garbage collection makes it seem like the obvious choice for managing

+
dynamic memory. In fact, one might question why the manual method is used at all, especially by a language as
sophisticated as C++. However, in the case of dynamic allocation, first impressions prove deceptive because both

C
approaches involve a set of trade-offs. Which approach is most appropriate is decided by the application. The

典
following sections describe some of the issues involved.

经
The Pros and Cons of Manual Memory Management

The main benefit of manually managing dynamic memory is efficiency. Because there is no garbage collector, no time

无
is spent keeping track of active objects or periodically looking for unused memory. Instead, when the programmer
knows that the allocated object is no longer needed, the programmer explicitly frees it and no additional overhead is
incurred. Because it has none of the overhead associated with garbage collection, the manual approach enables more
efficient code to be written This is one reason why it was necessary for C++ to support manual memory management:
it enabled the creation of high-performance code.

Another advantage to the manual approach is control. Although requiring the programmer to handle both the allocation
and release of memory is a burden, the benefit is that the programmer gains complete control over both halves of the
process. You know precisely when memory is being allocated and precisely when it is being released. Furthermore,
when you release an object via delete, its destructor is executed at that point rather than at some later time, as can be
the case with garbage collection. Thus, with the manual method you can control precisely when an allocated object is
destroyed.

Although it is efficient, manual memory management is susceptible to a rather annoying type of error: the memory
leak. Because memory must be freed manually, it is possible (even easy) to forget to do so. Failing to release unused
memory means that the memory will remain allocated even if it is no longer needed. Memory leaks cannot occur in a
garbage collection environment because the garbage collector ensures that unused objects are eventually freed.
Memory leaks are a particularly troublesome problem in Windows programming, where the failure to release unused
resources slowly degrades performance.

Other problems that can occur with C++’s manual approach include the premature releasing of memory that is still in
use, and the accidental freeing of the same memory twice. Both of these errors can lead to serious trouble.
Unfortunately, they may not show any immediate symptoms, making them hard to find.

The Pros and Cons of Garbage Collection

There are several different ways to implement garbage collection, each offering different performance characteristics.
However, all garbage collection systems share a set of common attributes that can be compared against the manual
approach. The main advantages to garbage collection are simplicity and safety. In a garbage collection environment,
you explicitly allocate memory via new, but you never explicitly free it. Instead, unused memory is automatically
recycled. Thus, it is not possible to forget to release an object or to release an object prematurely. This simplifies
programming and prevents an entire class of problems. Furthermore, it is not possible to accidentally free dynamically
allocated memory twice. Thus, garbage collection provides an easy-to-use, error-free, reliable solution to the memory
management problem.

Unfortunately, the simplicity and safety of garbage collection come at a price. The first cost is the overhead incurred by
the garbage collection mechanism. All garbage collection schemes consume some CPU cycles because the
reclamation of unused memory is not a cost-free process. This overhead does not occur with the manual approach.

A second cost is loss of control over when an object is destroyed. Unlike the manual approach, in which an object is

籍
destroyed (and its destructor called) at a known point in time—when a delete statement is executed on that
object—garbage collection does not have such a hard and fast rule. Instead, when garbage collection is used, an
object is not destroyed until the collector runs and recycles the object, which may not occur until some arbitrary time in

书
the future. For example, the collector might not run until the amount of free memory drops below a certain point.
Furthermore, it is not always possible to know the order in which objects will be destroyed by the garbage collector. In

典 书
some cases, the inability to know precisely when an object is destroyed can cause trouble because it also means that
your program can’t know precisely when the destructor for a dynamically allocated object is called.

经 读
For garbage collection systems that run as a background task, this loss of control can escalate into a potentially more
serious problem for some types of applications because it introduces what is essentially nondeterministic behavior into

+
a program. A garbage collector that executes in the background reclaims unused memory at times that are, for all

不
practical purposes, unknowable. For example, the collector will usually run only when free CPU time is available.

+
Because this might vary from one program run to the next, from one computer to next, or from one operating system to

C
the next, the precise point in program execution at which the garbage collector executes is effectively nondeterministic.

典
This is not a problem for many programs, but it can cause havoc with real-time applications in which the unexpected
allocation of CPU cycles to the garbage collector could cause an event to be missed.

经
You Can Have It Both Ways

无
As the preceding discussions explained, both manual management and garbage collection maximize one feature at
the expense of another. The manual approach maximizes efficiency and control at the expense of safety and ease of
use. Garbage collection maximizes simplicity and safety but pays for it with a loss of runtime performance and control.
Thus, garbage collection and manual memory management are essentially opposites, each maximizing the traits that
the other sacrifices. This is why neither approach to dynamic memory management can be optimal for all
programming situations.

Although opposites, the two approaches are not mutually exclusive. They can coexist. Thus, it is possible for the C++
programmer to have access to both approaches, choosing the proper method for the task at hand. All one needs to do
is create a garbage collector for C++, and this is the subject of the rest of this chapter.
 Creating a Garbage Collector in C++
Because C++ is a rich and powerful language, there are many different ways to implement a garbage collector. One
obvious, but limited, approach is to create a garbage collector base class, which is then inherited by classes that want
to use garbage collection. This would enable you to implement garbage collection on a class-by-class basis. This
solution is, unfortunately, too narrow to be satisfying.

A better solution is one in which the garbage collector can be used with any type of dynamically allocated object. To
provide such a solution, the garbage collector must:
1. Coexist with the built-in, manual method provided by C++.

2. Not break any preexisting code. Moreover, it must have no impact whatsoever on existing code.

3. Work transparently so that allocations that use garbage collection are operated on in the same
way as those that don’t.

籍
4. Allocate memory using new in the same way that C++’s built-in approach does.

5. Work with all data types, including the built-in types such as int and double.

书
6. Be simple to use.

典 书
In short, the garbage collection system must be able to dynamically allocate memory using a mechanism and syntax
that closely resemble that already used by C++ and not affect existing code. At first thought, this might seem to be a
daunting task, but it isn’t.

经 读
Understanding the Problem

+ 不
+
The key challenge that one faces when creating a garbage collector is how to know when a piece of memory is
unused. To understand the problem, consider the following sequence:

C 典
int *p;

无 经
p = new int(99);

p = new int(100);

Here, two int objects are dynamically allocated. The first contains the value 99 and a pointer to this value is
stored in p. Next, an integer containing the value 100 is allocated, and its address is also stored in p, thus
overwriting the first address. At this point, the memory for int(99) is not pointed to by p (or any other object) and
can be freed. The question is, how does the garbage collector know that neither p nor any other object points to
int(99)?

Here is a slight variation on the problem:

int *p, *q;

p = new int(99);

q = p; // now, q points to same memory as p

p = new int(100);

In this case, q points to the memory that was originally allocated for p. Even though p is then pointed to a
different piece of memory, the memory that it originally pointed to can’t be freed because it is still in use by q.
The question: how does the garbage collector know this fact? The precise way that these questions are
answered is determined by the garbage collection algorithm employed.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

Choosing a Garbage Collection Algorithm

Before implementing a garbage collector for C++, it is necessary to decide what garbage collection algorithm to use.
The topic of garbage collection is a large one, having been the focus of serious academic study for many years.
Because it presents an intriguing problem for which there is a variety of solutions, a number of different garbage
collection algorithms have been designed. It is far beyond the scope of this book to examine each in detail. However,
there are three archetypal approaches: reference counting, mark and sweep, and copying. Before choosing an
approach, it will be useful to review these three algorithms.

Reference Counting

In reference counting, each dynamically allocated piece of memory has associated with it a reference count. This
count is incremented each time a reference to the memory is added and decremented each time a reference to the
memory is removed. In C++ terms, this means that each time a pointer is set to point to a piece of allocated memory,
the reference count associated with that memory is incremented. When the pointer is set to point elsewhere, the

籍
reference count is decremented. When the reference count drops to zero, the memory is unused and can be released.

The main advantage of reference counting is simplicity—it is easy to understand and implement. Furthermore, it

书
places no restrictions on the organization of the heap because the reference count is independent of an object’s
physical location. Reference counting adds overhead to each pointer operation, but the collection phase is relatively
low cost. The main disadvantage is that circular references prevent memory that is otherwise unused from being

典 书
released. A circular reference occurs when two objects point to each other, either directly or indirectly. In this situation,
neither object’s reference count ever drops to zero. Some solutions to the circular reference problem have been

经 读
devised, but all add complexity and/or overhead.

+
Mark and Sweep

+ 不
Mark and sweep involves two phases. In the first phase, all objects in the heap are set to their unmarked state. Then,

C
all objects directly or indirectly accessible from program variables are marked as “in-use.” In phase two, all of allocated

典
memory is scanned (that is, a sweep of memory is made), and all unmarked elements are released.

There are two main advantages of mark and sweep. First, it easily handles circular references. Second, it adds

经
virtually no runtime overhead prior to collection. It has two main disadvantages. First, a considerable amount of time
might be spent collecting garbage because the entire heap must be scanned during collection. Thus, garbage
collection may cause unacceptable runtime characteristics for some programs. Second, although mark and sweep is

无
simple conceptually, it can be tricky to implement efficiently.

Copying

The copying algorithm organizes free memory into two spaces. One is active (holding the current heap), and the other
is idle. During garbage collection, in-use objects from the active space are identified and copied into the idle space.
Then, the roles of the two spaces are reversed, with the idle space becoming active and the active space becoming
idle. Copying offers the advantage of compacting the heap in the copy process. It has the disadvantage of allowing
only half of free memory to be in use at any one time.

Which Algorithm?

Given that there are advantages and disadvantages to all of the three classical approaches to garbage collection, it
might seem hard to choose one over the other. However, given the constraints enumerated earlier, there is a clear
choice: reference counting. First and most importantly, reference counting can be easily “layered onto” the existing
C++ dynamic allocation system. Second, it can be implemented in a straightforward manner and in a way that does
not affect preexisting code. Third, it does not require any specific organization or structuring of the heap, thus the
standard allocation system provided by C++ is unaffected.

The one drawback to using reference counting is its difficulty in handling circular references. This isn’t an issue for
.

many programs because intentional circular references are not all that common and can usually be avoided. (Even
things that we call circular, such as a circular queue, don’t necessarily involve a circular pointer reference.) Of course,
there are cases in which circular references are needed. It is also possible to create a circular reference without
knowing you have done so, especially when working with third-party libraries. Therefore, the garbage collector must
provide some means to gracefully handle a circular reference, should one exist.

To handle the circular reference problem, the garbage collector developed in this chapter will release any remaining
allocated memory when the program exits. This ensures that objects involved in a circular reference will be freed and
their destructors called. It is important to understand that normally there will be no allocated objects remaining at
program termination. This mechanism is explicitly for those objects that can’t be released because of a circular
reference. (You might want to experiment with other means of handling the circular reference problem. It presents an
interesting challenge.)

Implementing the Garbage Collector

To implement a reference counting garbage collector, there must be some way to keep track of the number of pointers
that point to each piece of dynamically allocated memory. The trouble is that C++ has no built-in mechanism that
enables one object to know when another object is pointing to it. Fortunately, there is a solution: you can create a new
pointer type that supports garbage collection. This is the approach used by the garbage collector in this chapter.

书 籍
To support garbage collection, the new pointer type must do three things. First, it must maintain a list of reference
counts for active dynamically allocated objects. Second, it must keep track of all pointer operations, increasing an
object’s reference count each time a pointer is pointed to that object and decreasing the count each time a pointer is
redirected to another object. Third, it must recycle those objects whose reference counts drop to zero. Aside from

典 书
supporting garbage collection, the new pointer type will look and feel just like a normal pointer. For example, all pointer
operations, such as * and –>, are supported.

经 读
In addition to being a convenient way to implement a garbage collector, the creation of a garbage collection pointer
type satisfies the constraint that the original C++ dynamic allocation system must be unaffected. When garbage

+
collection is desired, garbage collection-enabled pointers are used. When garbage collection is not desired, normal

不
C++ pointers are available. Thus, both types of pointers can be used within the same program.

C+
To Multithread or Not?

典
Another consideration when designing a garbage collector for C++ is whether it should be single-threaded or

经
multithreaded. That is, should the garbage collector be designed as a background process, running in its own thread
and collecting garbage as CPU time permits? Or, should the garbage collector run in the same thread as the process
that uses it, collecting garbage when certain program conditions occur? Both approaches have advantages and

无
disadvantages.

The main advantage to creating a multithreaded garbage collector is efficiency. Garbage can be collected when idle
CPU cycles are available. The disadvantage is, of course, that C++ does not provide any built-in support for
multithreading. This means that any multithreaded approach will depend upon operating system facilities to support
the multitasking. This makes the code nonportable.

The main advantage to using a single-threaded garbage collector is portability. It can be used in situations that do not
support multithreading or in cases in which the price of multithreading is too high. The main disadvantage is that the
rest of the program stops when garbage collection takes place.

In this chapter, the single-threaded approach is used because it works in all C++ environments. Thus, it can be used
by all readers of this book. However, for those readers wanting a multithreaded solution, one is given in Chapter 3,
which deals with the techniques needed to successfully multithread a C++ program in a Windows environment.

When to Collect Garbage?

One final question that needs to be answered before a garbage collector can be implemented: When is garbage
collected? This is less of an issue for a multithreaded garbage collector, which can run continuously as a background
task, collecting garbage whenever CPU cycles are available, than it is for a single-threaded garbage collector, such as
that developed in this chapter, which must stop the rest of the program to collect garbage.
In the real world, garbage collection usually takes place only when there is sufficient reason to do so, such as the case
of memory running low. This makes sense for two reasons. First, with some garbage collection algorithms, such as
mark and sweep, there is no way to know that a piece of memory is unused without actually performing the collection.
(That is, sometimes there is no way to know that there is garbage to be collected without actually collecting it!)
Second, collecting garbage is a time-consuming process which should not be performed needlessly.

However, waiting for memory to run low before initiating garbage collection is not suitable for the purposes of this
chapter because it makes it next to impossible to demonstrate the collector. Instead, the garbage collector developed
in this chapter will collect garbage more frequently so that its actions can easily be observed. As the collector is coded,
garbage is collected whenever a pointer goes out of scope. Of course, this behavior can easily be changed to fit your
applications.

One last point: when using reference-count based garbage collection, it is technically possible to recycle unused
memory as soon as its reference count drops to zero, rather than using a separate garbage collection phase.
However, this approach adds overhead to every pointer operation. The method used by this chapter is to simply
decrement the memory’s reference count each time a pointer to that memory is redirected and let the collection
process handle the actual recycling of memory at more convenient times. This reduces the runtime overhead
associated with pointer operations, which one typically wants to be as fast as possible.

籍
What About auto_ptr?

As many readers will know, C++ defines a library class called auto_ptr. Because an auto_ptr automatically frees the

书
memory to which it points when it goes out of scope, you might think that it would be of use when developing a
garbage collector, perhaps forming the foundation. This is not the case, however. The auto_ptr class is designed for a

典 书
concept that the ISO C++ Standard calls “strict ownership,” in which an auto_ptr “owns” the object to which it points.
This ownership can be transferred to another auto_ptr, but in all cases some auto_ptr will own the object until it is
destroyed. Furthermore, an auto_ptr is assigned an address of an object only when it is initialized. After that, you can’t

经 读
change the memory to which an auto_ptr points, except by assigning one auto_ptr to another. Because of auto_ptr’s
“strict ownership” feature, it is not particularly useful when creating a garbage collector, and it is not used by the

+ 不
garbage collectors in this book.

C+ 典
无 经
.

A Simple C++ Garbage Collector

The entire garbage collector is shown here. As explained, the garbage collector works by creating a new pointer type
that provides built-in support for garbage collection based on reference counting. The garbage collector is
single-threaded, which means that it is quite portable and does not rely upon (or make assumptions about) the
execution environment. This code should be stored in a file called gc.h.

There are two things to notice as you look through the code. First, most of the member functions are quite short and
are defined inside their respective classes in the interest of efficiency. Recall that a function defined within its class is
automatically in-lined, which eliminates the overhead of a function call. Only a few functions are long enough to require
their definition to be outside their class.

Secondly, notice the comment near the top of the file. If you want to watch the action of the garbage collector, simply
turn on the display option by defining the macro called DISPLAY . For normal use, leave DISPLAY undefined.
// A single-threaded garbage collector.

籍
#include <iostream>
#include <list>

书
#include <typeinfo>
#include <cstdlib>

典 书
using namespace std;

// To watch the action of the garbage collector, define DISPLAY.

经 读
// #define DISPLAY

+
// Exception thrown when an attempt is made to

不
// use an Iter that exceeds the range of the

+
// underlying object.
//

C 典
class OutOfRangeExc {
// Add functionality if needed by your application.
};

经
// An iterator-like class for cycling through arrays
// that are pointed to by GCPtrs. Iter pointers

无
// ** do not ** participate in or affect garbage
// collection. Thus, an Iter pointing to
// some object does not prevent that object
// from being recycled.
//
template <class T> class Iter {
T *ptr; // current pointer value
T *end; // points to element one past end
T *begin; // points to start of allocated array
unsigned length; // length of sequence
public:

Iter() {
ptr = end = begin = NULL;
length = 0;
}

Iter(T *p, T *first, T *last) {

ptr = p;
end = last;
begin = first;
 length = last - first;
}

// Return length of sequence to which this

// Iter points.
unsigned size() { return length; }

// Return value pointed to by ptr.

// Do not allow out-of-bounds access.
T &operator*() {
if( (ptr >= end) || (ptr < begin) )
throw OutOfRangeExc();
return *ptr;
}

// Return address contained in ptr.

// Do not allow out-of-bounds access.
T *operator->() {
if( (ptr >= end) || (ptr < begin) )
throw OutOfRangeExc();

籍
return ptr;
}

书
// Prefix ++.
Iter operator++() {

典 书
ptr++;
return *this;
}

ptr--;

+ 经
// Prefix --.

读
Iter operator--() {

不
+
return *this;
}

C 典
// Postfix ++.
Iter operator++(int notused) {

经
T *tmp = ptr;

ptr++;

无
return Iter<T>(tmp, begin, end);
}

// Postfix --.
Iter operator--(int notused) {
T *tmp = ptr;

ptr--;
return Iter<T>(tmp, begin, end);
}

// Return a reference to the object at the

// specified index. Do not allow out-of-bounds
// access.
T &operator[](int i) {
if( (i < 0) || (i >= (end-begin)) )
throw OutOfRangeExc();
return ptr[i];
}

// Define the relational operators.

bool operator==(Iter op2) {
 return ptr == op2.ptr;
}

bool operator!=(Iter op2) {

return ptr != op2.ptr;
}

bool operator<(Iter op2) {

return ptr < op2.ptr;
}

bool operator<=(Iter op2) {

return ptr <= op2.ptr;
}

bool operator>(Iter op2) {

return ptr > op2.ptr;
}

bool operator>=(Iter op2) {

籍
return ptr >= op2.ptr;
}

书
// Subtract an integer from an Iter.
Iter operator-(int n) {

典 书
ptr -= n;
return *this;
}

+ 经 读
// Add an integer to an Iter.
Iter operator+(int n) {
ptr += n;

不
+
return *this;
}

C 典
// Return number of elements between two Iters.
int operator-(Iter<T> &itr2) {

经
return ptr - itr2.ptr;
}

无
};

// This class defines an element that is stored

// in the garbage collection information list.
//
template <class T> class GCInfo {
public:
unsigned refcount; // current reference count

T *memPtr; // pointer to allocated memory

/* isArray is true if memPtr points

to an allocated array. It is false
otherwise. */
bool isArray; // true if pointing to array

/* If memPtr is pointing to an allocated

array, then arraySize contains its size */
unsigned arraySize; // size of array

// Here, mPtr points to the allocated

. memory.s
.

// If this is an array, then size specifies

// the size of the array.
GCInfo(T *mPtr, unsigned size=0) {
refcount = 1;
memPtr = mPtr;
if(size != 0)
isArray = true;
else
isArray = false;

arraySize = size;
}
};

// Overloading operator== allows GCInfos to be compared.

// This is needed by the STL list class.
template <class T> bool operator==(const GCInfo<T> &ob1,
const GCInfo<T> &ob2) {
return (ob1.memPtr == ob2.memPtr);
}

书 籍
// GCPtr implements a pointer type that uses
// garbage collection to release unused memory.
// A GCPtr must only be used to point to memory

典 书
// that was dynamically allocated using new.
// When used to refer to an allocated array,
// specify the array size.

经 读
//
template <class T, int size=0> class GCPtr {

+ 不
// gclist maintains the garbage collection list.

+
static list<GCInfo<T> > gclist;

C 典
// addr points to the allocated memory to which
// this GCPtr pointer currently points.
T *addr;

经
/* isArray is true if this GCPtr points
to an allocated array. It is false

无
otherwise. */
bool isArray; // true if pointing to array

// If this GCPtr is pointing to an allocated

// array, then arraySize contains its size.
unsigned arraySize; // size of the array

static bool first; // true when first GCPtr is created

// Return an iterator to pointer info in gclist.

typename list<GCInfo<T> >::iterator findPtrInfo(T *ptr);

public:

// Define an iterator type for GCPtr<T>.

typedef Iter<T> GCiterator;

// Construct both initialized and uninitialized objects.

GCPtr(T *t=NULL) {

// Register shutdown() as an exit function.

if(first) atexit(shutdown);
.

first = false;

list<GCInfo<T> >::iterator p;

p = findPtrInfo(t);

// If t is already in gclist, then

// increment its reference count.
// Otherwise, add it to the list.
if(p != gclist.end())
p->refcount++; // increment ref count
else {
// Create and store this entry.
GCInfo<T> gcObj(t, size);
gclist.push_front(gcObj);
}

addr = t;
arraySize = size;
if(size > 0) isArray = true;

籍
else isArray = false;
#ifdef DISPLAY
cout << "Constructing GCPtr. ";

书
if(isArray)
cout << " Size is " << arraySize << endl;

典 书
else
cout << endl;
#endif

经 读
}

+
// Copy constructor.

不
GCPtr(const GCPtr &ob) {

+
list<GCInfo<T> >::iterator p;

C 典
p = findPtrInfo(ob.addr);
p->refcount++; // increment ref count

经
addr = ob.addr;
arraySize = ob.arraySize;
if(arraySize > 0) isArray = true;

无
else isArray = false;
#ifdef DISPLAY
cout << "Constructing copy.";
if(isArray)
cout << " Size is " << arraySize << endl;
else
cout << endl;
#endif
}

// Destructor for GCPTr.

~GCPtr();

// Collect garbage. Returns true if at least

// one object was freed.
static bool collect();

// Overload assignment of pointer to GCPtr.

T *operator=(T *t);

// Overload assignment of GCPtr to GCPtr.

GCPtr &operator=(GCPtr &rv);
.

// Return a reference to the object pointed

// to by this GCPtr.
T &operator*() {
return *addr;
}

// Return the address being pointed to.

T *operator->() { return addr; }

// Return a reference to the object at the

// index specified by i.
T &operator[](int i) {
return addr[i];
}

// Conversion function to T *.
operator T *() { return addr; }

// Return an Iter to the start of the allocated memory.

籍
Iter<T> begin() {
int size;

书
if(isArray) size = arraySize;
else size = 1;

典 书
return Iter<T>(addr, addr, addr + size);
}

+ 经
Iter<T> end() {
int size;

不 读
// Return an Iter to one past the end of an allocated array.

+
if(isArray) size = arraySize;

C 典
else size = 1;

return Iter<T>(addr + size, addr, addr + size);

经
}

// Return the size of gclist for this type

无
// of GCPtr.
static int gclistSize() { return gclist.size(); }

// A utility function that displays gclist.

static void showlist();

// Clear gclist when program exits.

static void shutdown();
};

// Creates storage for the static variables

template <class T, int size>
list<GCInfo<T> > GCPtr<T, size>::gclist;

template <class T, int size>

bool GCPtr<T, size>::first = true;

// Destructor for GCPtr.

template <class T, int size>
GCPtr<T, size>::~GCPtr() {
list<GCInfo<T> >::iterator p;
 p = findPtrInfo(addr);
if(p->refcount) p->refcount--; // decrement ref count

#ifdef DISPLAY
cout << "GCPtr going out of scope.\n";
#endif

// Collect garbage when a pointer goes out of scope.

collect();

// For real use, you might want to collect

// unused memory less frequently, such as after
// gclist has reached a certain size, after a
// certain number of GCPtrs have gone out of scope,
// or when memory is low.
}

// Collect garbage. Returns true if at least

// one object was freed.
template <class T, int size>

籍
bool GCPtr<T, size>::collect() {
bool memfreed = false;

书
#ifdef DISPLAY
cout << "Before garbage collection for ";

典 书
showlist();
#endif

经 读
list<GCInfo<T> >::iterator p;
do {

+ + 不
// Scan gclist looking for unreferenced pointers.
for(p = gclist.begin(); p != gclist.end(); p++) {
// If in-use, skip.

C 典
if(p->refcount > 0) continue;

memfreed = true;

经
// Remove unused entry from gclist.
gclist.remove(*p);

无
// Free memory unless the GCPtr is null.
if(p->memPtr) {
if(p->isArray) {
#ifdef DISPLAY
cout << "Deleting array of size "
<< p->arraySize << endl;
#endif
delete[] p->memPtr; // delete array
}
else {
#ifdef DISPLAY
cout << "Deleting: "
<< *(T *) p->memPtr << "\n";
#endif
delete p->memPtr; // delete single element
}
}

// Restart the search.

break;
. }s
.

} while(p != gclist.end());

#ifdef DISPLAY
cout << "After garbage collection for ";
showlist();
#endif

return memfreed;
}

// Overload assignment of pointer to GCPtr.

template <class T, int size>
T * GCPtr<T, size>::operator=(T *t) {
list<GCInfo<T> >::iterator p;

// First, decrement the reference count

// for the memory currently being pointed to.
p = findPtrInfo(addr);
p->refcount--;

书 籍
// Next, if the new address is already
// existent in the system, increment its
// count. Otherwise, create a new entry
// for gclist.

典 书
p = findPtrInfo(t);
if(p != gclist.end())
p->refcount++;

经 读
else {
// Create and store this entry.

+
GCInfo<T> gcObj(t, size);

不
gclist.push_front(gcObj);

+
}

C 典
addr = t; // store the address.

return t;

经
}

// Overload assignment of GCPtr to GCPtr.

无
template <class T, int size>
GCPtr<T, size> & GCPtr<T, size>::operator=(GCPtr &rv) {
list<GCInfo<T> >::iterator p;

// First, decrement the reference count

// for the memory currently being pointed to.
p = findPtrInfo(addr);
p->refcount--;

// Next, increment the reference count of

// the new address.
p = findPtrInfo(rv.addr);
p->refcount++; // increment ref count

addr = rv.addr;// store the address.

return rv;
}

// A utility function that displays gclist.

template <class T, int size>
void GCPtr<T, size>::showlist() {
.

list<GCInfo<T> >::iterator p;

cout << "gclist<" << typeid(T).name() << ", "

<< size << ">:\n";
cout << "memPtr refcount value\n";

if(gclist.begin() == gclist.end()) {
cout << " -- Empty --\n\n";
return;
}

for(p = gclist.begin(); p != gclist.end(); p++) {

cout << "[" << (void *)p->memPtr << "]"
<< " " << p->refcount << " ";
if(p->memPtr) cout << " " << *p->memPtr;
else cout << " ---";
cout << endl;
}
cout << endl;
}

书 籍
// Find a pointer in gclist.
template <class T, int size>
typename list<GCInfo<T> >::iterator
GCPtr<T, size>::findPtrInfo(T *ptr) {

典 书
list<GCInfo<T> >::iterator p;

经 读
// Find ptr in gclist.
for(p = gclist.begin(); p != gclist.end(); p++)

+
if(p->memPtr == ptr)

不
return p;

+
return p;

C 典
// Clear gclist when program exits.

经
template <class T, int size>
void GCPtr<T, size>::shutdown() {

无
if(gclistSize() == 0) return; // list is empty

list<GCInfo<T> >::iterator p;

for(p = gclist.begin(); p != gclist.end(); p++) {

// Set all reference counts to zero
p->refcount = 0;
}

#ifdef DISPLAY
cout << "Before collecting for shutdown() for "
<< typeid(T).name() << "\n";
#endif

collect();

#ifdef DISPLAY
cout << "After collecting for shutdown() for "
<< typeid(T).name() << "\n";
#endif
}
 An Overview of the Garbage Collector Classes

The garbage collector uses four classes: GCPtr, GCInfo , Iter, and OutOfRangeExc. Before examining the code in
detail, it will be helpful to understand the role each class plays.

GCPtr

At the core of the garbage collector is the class GCPtr, which implements a garbage-collection pointer. GCPtr
maintains a list that associates a reference count with each piece of memory allocated for use by a GCPtr. In general,
here is how it works. Each time a GCPtr is pointed at a piece of memory, the reference count for that memory is
incremented. If the GCPtr was previously pointing at a different piece of memory prior to the assignment, the reference
count for that memory is decremented. Thus, adding a pointer to a piece of memory increases the memory’s reference
count. Removing a pointer decreases the memory’s reference count. Each time a GCPtr goes out of scope, the
reference count associated with the memory to which it currently points is decremented. When a reference count
drops to zero, that piece of memory can be released.

GCPtr is a template class that overloads the * and –> pointer operators and the array indexing operator [ ]. Thus, GCPtr
creates a new pointer type and integrates it into the C++ programming environment. This allows a GCPtr to be used in
much the same way that you use a normal C++ pointer. However, for reasons that will be made clear later in this

籍
chapter, GCPtr does not overload the ++, – –, or the other arithmetic operations defined for pointers. Thus, except
through assignment, you cannot change the address to which a GCPtr object points. This may seems like a significant

书
restriction, but it isn’t because the Iter class provides these operations.

For the sake of illustration, the garbage collector runs whenever a GCPtr object goes out of scope. At that time, the

典 书
garbage collection list is scanned and all memory with a reference count of zero is released, even if it was not
originally associated with the GCPtr that went out of scope. Your program can also explicitly request garbage collection
if you need to recycle memory earlier.

GCInfo

+ 经 不 读
+
As explained, GCPtr maintains a list that links reference counts with allocated memory. Each entry in this list is
encapsulated in an object of type GCInfo . GCInfo stores the reference count in its refcount field and a pointer to the

C
memory in its memPtr field. Thus, a GCInfo object binds a reference count to a piece of allocated memory.

典
GCInfo defines two other fields: isArray and arraySize. If memPtr points to an allocated array, then its isArray member
will be true and the length of the array will be stored in its arraySize field.

经
The Iter Class

无
As explained, a GCPtr object allows you to access the memory to which it points by using the normal pointer operators
* and –>, but it does not support pointer arithmetic. To handle situations in which you need to perform pointer
arithmetic, you will use an object of type Iter. Iter is a template class similar in function to an STL iterator, and it defines
all pointer operations, including pointer arithmetic. The main use of Iter is to enable you to cycle through the elements
of a dynamically allocated array. It also provides bounds checking. You obtain an Iter from GCPtr by calling either
begin( ) or end( ), which work much like their equivalents in the STL.

It is important to understand that although Iter and the STL iterator type are similar, they are not the same, and you
cannot use one in place of the other.

OutOfRangeExc

If an Iter encounters an attempt to access memory outside the range of the allocated memory, an OutOfRangeExc
exception is thrown. For the purposes of this chapter, OutOfRangeExc contains no members. It is simply a type that
can be thrown. However, you are free to add functionality to this class as your own applications dictate.
.

GCPtr In Detail
GCPtr is the heart of the garbage collector. It implements a new pointer type that keeps a reference count for objects
allocated on the heap. It also provides the garbage collection functionality that recycles unused memory.

GCptr is a template class with this declaration:

template <class T, int size=0> class GCPtr {

GCPtr requires that you specify the type of the data that will be pointed to, which will be substituted for the generic type
T. If an array is being allocated, you must specify the size of the array in the size parameter. Otherwise, size defaults to
zero, which indicates that a single object is being pointed to. Here are two examples.
GCPtr<int> p; // declare a pointer to a single integer
GCPtr<int, 5> ap; // declare a pointer to an array of 5 integers

Here, p can point to single objects of type int, and ap can point to an array of 5 ints.

籍
In the preceding examples, notice that you do not use the * operator when specifying the name of the GCPtr object.
That is, to create a GCPtr to int, you do not use a statement like this:

书
GCPtr<int> *p; // this creates a pointer to a GCPtr<int> object

This declaration creates a normal C++ pointer called p that can point to a GCPtr<int> object. It does not create a GCPtr

典 书
object that can point to int. Remember, GCPtr defines a pointer type by itself.

Be careful when specifying the type parameter to GCPtr. It specifies the type of the object to which the GCPtr object

经 读
can point. Therefore, if you write a declaration like this:
GCPtr<int *> p; // this creates a GCPtr to pointers to ints

+ 不
you are creating a GCPtr object that points to int * pointers, not a GCPtr to ints.

+
Because of its importance, each member of GCPtr is examined in detail in the following sections.

C 典
GCPtr Data Members

经
GCPtr declares the following data members:

// gclist maintains the garbage collection list.

无
static list<GCInfo<T> > gclist;

// addr points to the allocated memory to which

// this GCPtr pointer currently points.
T *addr;

/* isArray is true if this GCPtr points

to an allocated array. It is false
otherwise. */
bool isArray; // true if pointing to array

// If this GCPtr is pointing to an allocated

// array, then arraySize contains its size.
unsigned arraySize; // size of the array

static bool first; // true when first GCPtr is created

The gclist field contains a list of GCInfo objects. (Recall that GCInfo links a reference count with a piece of allocated
memory.) This list is used by the garbage collector to determine when allocated memory is unused. Notice that gclist
is a static member of GCPtr. This means that for each specific pointer type, there is only one gclist. For example, all
pointers of type GCPtr<int> share one list, and all pointers of type GCPtr<double> share a different list. gclist is an
instance of the STL list class. Using the STL substantially simplifies the code for GCPtr because there is no need for it
.

to create its own set of list-handling functions.

GCPtr stores the address of the memory to which it points in addr. If addr points to an allocated array, then isArray will
be true, and the length of the array will be stored in arraySize.

The first field is a static variable that is initially set to true. It is a flag that the GCPtr constructor uses to know when the
first GCPtr object is created. After the first GCPtr object is constructed, first is set to false. It is used to register a
termination function that will be called to shut down the garbage collector when the program ends.

The findPtrInfo( ) Function

GCPtr declares one private function: findPtrInfo( ). This function searches gclist for a specified address and returns an
iterator to its entry. If the address is not found, an iterator to the end of gclist is returned. This function is used
internally by GCPtr to update the reference counts of the objects in gclist. It is implemented as shown here:
// Find a pointer in gclist.
template <class T, int size>
typename list<GCInfo<T> >::iterator
GCPtr<T, size>::findPtrInfo(T *ptr) {

籍
list<GCInfo<T> >::iterator p;

// Find ptr in gclist.

书
for(p = gclist.begin(); p != gclist.end(); p++)
if(p->memPtr == ptr)
return p;

}
return p;

经 典 读 书
+
The GCIterator typedef

+ 不
At the start of the public section of GCPtr is the typedef of Iter<T> to GCiterator. This typedef is bound to each
instance of GCPtr, thus eliminating the need to specify the type parameter each time an Iter is needed for a specific

C 典
version of GCPtr. This simplifies the declaration of an iterator. For example, to obtain an iterator to the memory pointed
to by a specific GCPtr, you can use a statement like this:
GCPtr<int>::GCiterator itr;

经
The GCPtr Constructor

无
The constructor for GCPtr is shown here:
// Construct both initialized and uninitialized objects.
GCPtr(T *t=NULL) {

// Register shutdown() as an exit function.

if(first) atexit(shutdown);
first = false;

list<GCInfo<T> >::iterator p;

p = findPtrInfo(t);

// If t is already in gclist, then

// increment its reference count.
// Otherwise, add it to the list.
if(p != gclist.end())
p->refcount++; // increment ref count
else {
// Create and store this entry.
GCInfo<T> gcObj(t, size);
.

gclist.push_front(gcObj);
}

addr = t;
arraySize = size;
if(size > 0) isArray = true;
else isArray = false;
#ifdef DISPLAY
cout << "Constructing GCPtr. ";
if(isArray)
cout << " Size is " << arraySize << endl;
else
cout << endl;
#endif
}

GCPtr( ) allows both initialized and uninitialized instances to be created. If an initialized instance is declared, then the
memory to which this GCPtr will point is passed through t . Otherwise, t will be null. Let’s examine the operation of
GCPtr( ) in detail.

籍
First, if first is true, it means that this is the first GCPtr object to be created. If this is the case, then shutdown( ) is
registered as a termination function by calling atexit( ). The atexit( ) function is part of the standard C++ function library,
and it registers a function that will be called when a program terminates. In this case, shutdown( ) releases any

书
memory that was prevented from being released because of a circular reference.

书
Next, a search of gclist is made, looking for any preexisting entry that matches the address in t . If one is found, then

典
its reference count is incremented. If no preexising entry matches t , a new GCInfo object is created that contains this
address, and this object is added to gclist.

经 读
GCPtr( ) then sets addr to the address specified by t and sets the values of isArray and arraySize appropriately.

+
Remember, if you are allocating an array, you must explicitly specify the size of the array when you declare the GCPtr

不
pointer that will point to it. If you don’t, the memory won’t be released correctly, and, in the case of an array of class

+
objects, the destructors won’t be called properly.

C 典
The GCPtr Destructor

The destructor for GCPtr is shown here:

经
// Destructor for GCPtr.
template <class T, int size>

无
GCPtr<T, size>::~GCPtr() {
list<GCInfo<T> >::iterator p;

p = findPtrInfo(addr);
if(p->refcount) p->refcount--; // decrement ref count

#ifdef DISPLAY
cout << "GCPtr going out of scope.\n";
#endif

// Collect garbage when a pointer goes out of scope.

collect();

// For real use, you might want to collect

// unused memory less frequently, such as after
// gclist has reached a certain size, after a
// certain number of GCPtrs have gone out of scope,
// or when memory is low.
}

Garbage collection takes place each time a GCPtr goes out of scope. This is handled by ~GCPtr( ). First, a search of
gclist is made, looking for the entry that corresponds to the address pointed to by the GCPtr being destroyed. Once
.

found, its refcount field is decremented. Next, ~GCptr( ) calls collect( ) to release any unused memory (that is, memory
whose reference count is zero).

As the comment at the end of ~GCPtr( ) states, for real applications, it is probably better to collect garbage less often
than each time a single GCPtr goes out of scope. Collecting less frequently will usually be more efficient. As explained
earlier, collecting each time a GCPtr is destroyed is useful for demonstrating the garbage collector because it clearly
illustrates the garbage collector’s operation.

Collect Garbage with collect( )

The collect( ) function is where garbage collection takes place. It is shown here:
// Collect garbage. Returns true if at least
// one object was freed.
template <class T, int size>
bool GCPtr<T, size>::collect() {
bool memfreed = false;

#ifdef DISPLAY
cout << "Before garbage collection for ";

籍
showlist();
#endif

书
list<GCInfo<T> >::iterator p;
do {

典 书
// Scan gclist looking for unreferenced pointers.
for(p = gclist.begin(); p != gclist.end(); p++) {

经 读
// If in-use, skip.
if(p->refcount > 0) continue;

+ 不
memfreed = true;

+
// Remove unused entry from gclist.

C
gclist.remove(*p);

典
// Free memory unless the GCPtr is null.

经
if(p->memPtr) {
if(p->isArray) {
#ifdef DISPLAY

无
cout << "Deleting array of size "
<< p->arraySize << endl;
#endif
delete[] p->memPtr; // delete array
}
else {
#ifdef DISPLAY
cout << "Deleting: "
<< *(T *) p->memPtr << "\n";
#endif
delete p->memPtr; // delete single element
}
}

// Restart the search.

break;
}

} while(p != gclist.end());

#ifdef DISPLAY
cout << "After garbage collection for ";
.

showlist();
#endif

return memfreed;
}

The collect( ) function works by scanning the contents of gclist, looking for entries that have a refcount of zero. When
such an entry is found, it is removed from gclist by calling the remove( ) function, which is a member of the STL list
class. Then the memory associated with that entry is freed.

Recall that in C++, single objects are freed by delete, but arrays of objects are freed via delete[ ] . Thus, the value of
the entry’s isArray field determines whether delete or delete[ ] is used to free the memory. This is one reason you must
specify the size of an allocated array for any GCPtr that will point to one: it causes isArray to be set to true. If isArray is
not set correctly, it is impossible to properly release the allocated memory.

Although the point of garbage collection is to recycle unused memory automatically, you can take a measure of
manual control if necessary. The collect( ) function can be called directly by user code to request that garbage
collection take place. Notice that it is declared as a static function within GCPtr. This means that it can be invoked
without reference to any object. For example:

籍
GCPtr<int>::collect(); // collect all unused int pointers

This causes gclist<int> to be collected. Because there is a different gclist for each type of pointer, you will need to call

书
collect( ) for each list that you want to collect. Frankly, if you need to closely manage the release of dynamically
allocated objects, you are better off using the manual allocation system provided by C++. Directly calling collect( ) is
best reserved for specialized situations, such as when free memory is running unexpectedly low.

经 典 读 书
The Overloaded Assignment Operators

GCPtr overloads operator=( ) twice: once for the assignment of a new address to a GCPtr pointer, and once for the

+
assignment of one GCPtr pointer to another. Both versions are shown here:

不
// Overload assignment of pointer to GCPtr.

+
template <class T, int size>

C
T * GCPtr<T, size>::operator=(T *t) {

典
list<GCInfo<T> >::iterator p;

// First, decrement the reference count

经
// for the memory currently being pointed to.
p = findPtrInfo(addr);
p->refcount--;

无
// Next, if the new address is already
// existent in the system, increment its
// count. Otherwise, create a new entry
// for gclist.
p = findPtrInfo(t);
if(p != gclist.end())
p->refcount++;
else {
// Create and store this entry.
GCInfo<T> gcObj(t, size);
gclist.push_front(gcObj);
}

addr = t; // store the address.

return t;
}

// Overload assignment of GCPtr to GCPtr.

template <class T, int size>
.

GCPtr<T, size> & GCPtr<T, size>::operator=(GCPtr &rv) {

list<GCInfo<T> >::iterator p;

// First, decrement the reference count

// for the memory currently being pointed to.
p = findPtrInfo(addr);
p->refcount--;

// Next, increment the reference count of

// the new address.
p = findPtrInfo(rv.addr);
p->refcount++; // increment ref count

addr = rv.addr;// store the address.

return rv;
}

The first overload of operator=( ) handles assignments in which a GCPtr pointer is on the left and an address is on the
right. For example, it handles cases like the one shown here:

籍
GCPtr<int> p;
// ...

书
p = new int(18);

Here, the address returned by new is assigned to p. When this happens, operator=(T *t) is called with the new address

典 书
passed to t . First, the entry in gclist for the memory that is currently being pointed to is found, and its reference count
is decremented. Next, a search of gclist is made for the new address. If it is found, its reference count is incremented.
Otherwise, a new GCInfo object is created for the new address, and this object is added to gclist. Finally, the new

经 读
address is stored in the invoking object’s addr, and this address is returned.

+
The second overload of the assignment operator, operator=(GCPtr &rv), handles the following type of assignment:

+ 不
GCPtr<int> p;
GCPtr<int> q;

C
// ...

典
p = new int(88);
q = p;

经
Here, both p and q are GCPtr pointers, and p is assigned to q. This version of the assignment operator works much like
the other. First, the entry in gclist for the memory that is currently being pointed to is found, and its reference count is
decremented. Next, a search of gclist is made for the new address, which is contained in rv.addr, and its reference

无
count is incremented. Then the invoking object’s addr field is set to the address contained in rv.addr. Finally, the
right-hand object is returned. This allows a chain of assignments to take place, such as:
p = q = w = z;

There is one other important point to make about the way that the assignment operators work. As mentioned earlier in
this chapter, it is technically possible to recycle memory as soon as its reference count drops to zero, but doing so puts
an extra overhead on each pointer operation. This is why, in the overloaded assignment operators, the reference count
for the memory previously pointed to by the left-hand operand is simply decremented, and no further action is taken.
Thus, the management overhead associated with actually releasing memory and performing maintenance on gclist is
avoided. These actions are deferred to later, when the garbage collector runs. This approach makes for faster
runtimes for the code that uses a GCPtr. It also lets garbage collection take place at times that are (potentially) more
convenient in terms of runtime performance.

The GCPtr Copy Constructor

Because of the need to keep track of each pointer to allocated memory, the default copy constructor (which makes a
bitwise identical copy) cannot be used. Instead, GCPtr must define is own copy constructor, which is shown here:
// Copy constructor.
GCPtr(const GCPtr &ob) {
.

list<GCInfo<T> >::iterator p;

p = findPtrInfo(ob.addr);
p->refcount++; // increment ref count

addr = ob.addr;
arraySize = ob.arraySize;
if(arraySize > 0) isArray = true;
else isArray = false;
#ifdef DISPLAY
cout << "Constructing copy.";
if(isArray)
cout << " Size is " << arraySize << endl;
else
cout << endl;
#endif
}

Recall that a class’ copy constructor is invoked when a copy of an object is required, such as when an object is
passed as an argument to a function, when an object is returned from a function, or when one object is used to

籍
initialize another. GCPtr’s copy constructor duplicates the information contained in the original object. It also
increments the reference count associated with the memory pointed to by the original object. When the copy goes out

书
of scope, this reference count will be decremented.

Actually, the extra work performed by the copy constructor is not usually necessary because the overloaded

典 书
assignment operators properly maintain the garbage collection list in most cases. However, there are a small number
of cases in which the copy constructor is needed, such as when memory is allocated inside a function and a GCPtr to
that memory is returned.

+ 经 不 读
The Pointer Operators and Conversion Function

Because GCPtr is a pointer type, it must overload the pointer operators * and –>, plus the indexing operator [ ]. This is

+
done by the functions shown here. Given that it is the ability to overload these operators that makes it possible to

C
create a new pointer type, it is amazing how simple they are.

典
// Return a reference to the object pointed
// to by this GCPtr.

经
T &operator*() {
return *addr;
}

无
// Return the address being pointed to.
T *operator->() { return addr; }

// Return a reference to the object at the

// index specified by i.
T &operator[](int i) {
return addr[i];
}

The operator*( ) function returns a reference to the object pointed to by the addr field of the invoking GCPtr, operator– >(
) returns the address contained in addr, and operator[ ] returns a reference to the element specified by the index.
operator[ ] should be used only on GCPtrs that point to allocated arrays.

As mentioned earlier, no pointer arithmetic is supported. For example, neither the ++ nor the – – operator is overloaded
for GCPtr. The reason is that the garbage collection mechanism assumes that a GCPtr is pointing to the start of
allocated memory. If a GCPtr could be incremented, for example, then when that pointer was garbage collected, the
address used with delete would be invalid.

You have two options if you need to perform operations that involve pointer arithmetic. First, if the GCPtr is pointing to
an allocated array, you can create an Iter that will let you cycle through that array. This approach is described later.
.

Second, you can convert a GCPtr into a normal pointer by use of the T * conversion function defined by GCPtr. This
function is shown here:
// Conversion function to T *.
operator T*() { return addr; }

This function returns a normal pointer that points to the address stored in addr. It can be used like this.
GCPtr<double> gcp = new double(99.2);
double *p;

p = gcp; // now, p points to same memory as gcp

p++; // because p is a normal pointer, it can be incremented

In the preceding example, because p is a normal pointer, it can be manipulated in any way that any other pointer can.
Of course, whether such manipulations yield meaningful results is dependent upon your application.

The main advantage of the conversion to T * is that it lets you use GCPtrs in place of normal C++ pointers when
working with preexisting code that requires such pointers. For example, consider this sequence:
GCPtr<char> str = new char[80];
strcpy(str, "this is a test");

籍
cout << str << endl;

Here, str is a GCPtr pointer to char that is used in a call to strcpy( ). Because strcpy( ) is expecting its arguments to be

书
of type char *, the conversion to T * inside GCPtr is automatically invoked because, in this case, T is char. The same
conversion is automatically invoked when str is used in the cout statement. Thus, the conversion function enables

书
GCPtrs to seamlessly integrate with existing C++ functions and classes.

经 典
Keep in mind that the T * pointer returned by this conversion does not participate in or affect garbage collection. Thus,

读
it is possible for the allocated memory to be freed even if a regular C++ pointer is still pointing to it. So, use the
conversion to T * wisely––and infrequently.

+ 不
The begin( ) and end( ) Functions

+
The begin( ) and end( ) functions, shown next, are similar to their counterparts in the STL:

C 典
// Return an Iter to the start of the allocated memory.
Iter<T> begin() {
int size;

无 经
if(isArray) size = arraySize;
else size = 1;

return Iter<T>(addr, addr, addr + size);

// Return an Iter to one past the end of an allocated array.

Iter<T> end() {
int size;

if(isArray) size = arraySize;

else size = 1;

return Iter<T>(addr + size, addr, addr + size);

}

The begin( ) function returns an Iter to the start of the allocated array pointed to by addr. The end( ) function returns an
Iter to one past the end of the array. Although there is nothing that stops these functions from being called on a GCPtr
that points to a single object, their purpose is to support operations on allocated arrays. (Obtaining an Iter to a single
object is not harmful, just pointless.)

The shutdown( ) Function

If a program creates a circular reference of GCPtrs, then when the program ends there will still be dynamically
allocated objects that need to be released. This is important because these objects might have destructors that need
to be called. The shutdown( ) function handles this task. This function is registered by atexit( ) when the first GCPtr is
created, as described earlier. This means that it will be called when the program terminates.

The shutdown( ) function is shown here:

// Clear gclist when program exits.
template <class T, int size>
void GCPtr<T, size>::shutdown() {

if(gclistSize() == 0) return; // list is empty

list<GCInfo<T> >::iterator p;

for(p = gclist.begin(); p != gclist.end(); p++) {

// Set all reference counts to zero
p->refcount = 0;
}

籍
#ifdef DISPLAY
cout << "Before collecting for shutdown() for "
<< typeid(T).name() << "\n";

书
#endif

collect();

经 典
#ifdef DISPLAY

读 书
cout << "After collecting for shutdown() for "
<< typeid(T).name() << "\n";
#endif

+ 不
}

+
First, if the list is empty, which it will normally be, then shutdown( ) simply returns. Otherwise, it sets to zero the

C
reference counts of the entries that still exist in gclist, and then it calls collect( ). Recall that collect( ) releases any

典
object that has a reference count of zero. Thus, setting the reference counts to zero ensures that all objects will be
freed.

经
Two Utility Functions

无
GCPtr ends by defining two utility functions. The first is gclistSize( ), which returns the number of entries currently held
in gclist. The second is showlist( ), which displays the contents of gclist. Neither of these are necessary for the
implementation of a garbage collection pointer type, but they are useful when you want to watch the operation of the
garbage collector.
 GCInfo
The garbage collection list in gclist holds objects of type GCInfo . The GCInfo class is shown here:
// This class defines an element that is stored
// in the garbage collection information list.
//
template <class T> class GCInfo {
public:
unsigned refcount; // current reference count

T *memPtr; // pointer to allocated memory

/* isArray is true if memPtr points

to an allocated array. It is false
otherwise. */
bool isArray; // true if pointing to array

籍
/* If memPtr is pointing to an allocated
array, then arraySize contains its size */

书
unsigned arraySize; // size of array

书
// Here, mPtr points to the allocated memory.

典
// If this is an array, then size specifies
// the size of the array.

经 读
GCInfo(T *mPtr, unsigned size=0) {
refcount = 1;

+
memPtr = mPtr;

不
if(size != 0)

+
isArray = true;
else

C
isArray = false;

};
}

经 典
arraySize = size;

无
As mentioned earlier, each GCInfo object stores a pointer to allocated memory in memPtr and the reference count
associated with that memory in refcount. If the memory pointed to by memPtr contains an array, then the length of that
array must be specified when the GCInfo object is created. In this case, isArray is set to true, and the length of the
array will be stored in arraySize.

GCInfo objects are stored in an STL list. To enable searches on this list, it is necessary to define operator==( ), as
shown here:
// Overloading operator== allows GCInfos to be compared.
// This is needed by the STL list class.
template <class T> bool operator==(const GCInfo<T> &ob1,
const GCInfo<T> &ob2) {
return (ob1.memPtr == ob2.memPtr);
}

Two objects are equal only if both their memPtr fields are identical. Depending upon the compiler you are using, other
operators may need to be overloaded to enable GCInfo s to be stored in an STL list.
.

Iter
The Iter class implements an iterator-like object that can be used to cycle through the elements of an allocated array.
Iter is not technically necessary because a GCPtr can be converted to a normal pointer of its base type, but Iter offers
two advantages. First, it lets you cycle through an allocated array in a fashion similar to the way in which you cycle
through the contents of an STL container. Thus, the syntax for using an Iter is familiar. Second, Iter will not allow
out-of-range accesses. Thus, an Iter is a safe alternative to using a normal pointer. Understand, however, that Iter
does not participate in garbage collection. Thus, if the underlying GCPtr on which an Iter is based goes out of scope,
the memory to which it points will be freed whether or not it is still needed by that Iter.

Iter is a template class defined like this:

template <class T> class Iter {

The type of data to which the Iter points is passed through T.

Iter defines these instance variables:

籍
T *ptr; // current pointer value
T *end; // points to element one past end

书
T *begin; // points to start of allocated array
unsigned length; // length of sequence

典 书
The address to which the Iter currently points is held in ptr . The address to the start of the array is stored in begin, and
the address of an element one past the end of the array is stored in end. The length of the dynamic array is stored in
length.

Iter() {

+ 经 读
Iter defines the two constructors shown here. The first is the default constructor. The second constructs an Iter, given
an initial value for ptr , and pointers to the beginning and end of the array.

不
+
ptr = end = begin = NULL;

C
length = 0;

典
}

Iter(T *p, T *first, T *last) {

经
ptr = p;
end = last;
begin = first;

无
length = last - first;
}

For use by the garbage collector code shown in this chapter, the initial value of ptr will always equal begin. However,
you are free to construct Iters in which the initial value of ptr is a different value.

To enable Iter’s pointer-like nature, it overloads the * and –> pointer operators, and the array indexing operator [ ], as
shown here:
// Return value pointed to by ptr.
// Do not allow out-of-bounds access.
T &operator*() {
if( (ptr >= end) || (ptr < begin) )
throw OutOfRangeExc();
return *ptr;
}

// Return address contained in ptr.

// Do not allow out-of-bounds access.
T *operator->() {
if( (ptr >= end) || (ptr < begin) )
throw OutOfRangeExc();
 return ptr;
}
// Return a reference to the object at the
// specified index. Do not allow out-of-bounds
// access.
T &operator[](int i) {
if( (i < 0) || (i >= (end-begin)) )
throw OutOfRangeExc();
return ptr[i];
}

The * operator returns a reference to the element currently being pointed to in the dynamic array. The –> returns the
address of the element currently being pointed to. The [ ] returns a reference to the element at the specified index.
Notice that these operations do not allow an out-of-bounds access. If one is attempted, an OutOfRangeExc exception
is thrown.

Iter defines the various pointer arithmetic operators, such as ++, – –, and so on, which increment or decrement an Iter.
These operators enable you to cycle through a dynamic array. In the interest of speed, none of the arithmetic
operators perform range checks themselves. However, any attempt to access an out-of-bounds element will cause an
exception, which prevents a boundary error. Iter also defines the relational operators. Both the pointer arithmetic and

籍
relational functions are straightforward and easy to understand.

Iter also defines a utility function called size( ) , which returns the length of the array to which the Iter points.

典 书
As mentioned earlier, inside GCPtr, Iter<T> is typedefed to GCiterator for each instance of GCPtr, which simplifies the

书
declaration of an iterator. This means that you can use the type name GCiterator to obtain the Iter for any GCPtr.

+ 经 不 读
C+ 典
无 经
.

How to Use GCPtr

Using a GCPtr is quite easy. First, include the file gc.h. Then, declare a GCPtr object, specifying the type of the data to
which it will point. For example, to declare a GCPtr object called p that can point to an int, use this kind of declaration:
GCPtr<int> p; // p can point to int objects

Next, dynamically allocate the memory using new and assign the pointer returned by new to p, as shown here:
p = new int; // assign p the address of an int

You can assign a value to the allocated memory using an assignment operation like this:
*p = 88; // give that int a value

Of course, you can combine the three preceding statements like this:
GCPtr<int> p = new int(88); // declare and initialize

You can obtain the value of the memory at the location pointed to by p, as shown here:
int k = *p;

书 籍
As these examples show, in general you use a GCPtr just like a normal C++ pointer. The only difference is that you
don’t need to delete the pointer when you are through with it. The memory allocated to that pointer will be
automatically released when it is no longer needed.

经 典 读
#include <iostream>
#include <new> 书
Here is an entire program that assembles the pieces just shown:

+
#include "gc.h"

+ 不
using namespace std;

C
int main() {

典
GCPtr<int> p;

经
try {
p = new int;
} catch(bad_alloc exc) {

无
cout << "Allocation failure!\n";
return 1;
}

*p = 88;

cout << "Value at p is: " << *p << endl;

int k = *p;

cout << "k is " << k << endl;

return 0;
}

The output from this program with the display option turned on is shown here. (Recall that you can watch the operation
of the garbage collector by defining DISPLAY within gc.h.)
Constructing GCPtr.
Value at p is: 88
k is 88
GCPtr going out of scope.
Before garbage collection for gclist<int, 0>:
.

memPtr refcount value

[002F12C0] 0 88
[00000000] 0 ---

Deleting: 88
After garbage collection for gclist<int, 0>:
memPtr refcount value
-- Empty --

When the program ends, p goes out of scope. This causes its destructor to be called, which causes the reference
count for the memory pointed to by p to be decremented. Because p was the only pointer to this memory, this
operation sets the reference count to zero. Next, p’s destructor calls collect( ), which scans gclist, looking for entries
that have a reference count of zero. Because the entry previously associated with p has a reference count of zero, its
memory is freed.

One other point: notice that prior to garbage collection, a null pointer entry is also in gclist. This null pointer was
created when p was constructed. Recall that if a GCPtr is not given an initial address, the null address (which is zero)
is used. Although it is not technically necessary to store a null pointer in gclist (because it is never freed), doing so
simplifies other parts of GCPtr because it ensures that every GCPtr has a corresponding entry in gclist.

籍
Handling Allocation Exceptions

书
As the preceding program shows, because the garbage collector does not change the way that memory is allocated
via new, you handle allocation failures in the same way as usual, by catching the bad_alloc exception. (Recall that
when new fails, it throws an exception of type bad_alloc.) Of course, the preceding program won’t run out of memory,

典 书
and the try/catch block isn’t really needed, but a real-world program might exhaust the heap. Thus, you should always
check for this possibility.

经 读
In general, the best way to respond to a bad_alloc exception when using garbage collection is to call collect( ) to

+
recycle any unused memory and then retry the allocation that failed. This technique is employed by the load-testing

不
program shown later in this chapter. You can use the same basic technique in your own programs.

+
A More Interesting Example

C 典
Here is a more interesting example that shows the effect of a GCPtr going out of scope before the end of the program:
// Show a GCPtr going out of scope prior to the end

经
// of the program.
#include <iostream>
#include <new>

无
#include "gc.h"

using namespace std;

int main() {
GCPtr<int> p;
GCPtr<int> q;

try {
p = new int(10);
q = new int(11);

cout << "Value at p is: " << *p << endl;

cout << "Value at q is: " << *q << endl;

cout << "Before entering block.\n";

// Now, create a local object.

{ // start a block
GCPtr<int> r = new int(12);
cout << "Value at r is: " << *r << endl;
.

} // end the block, causing r to go out of scope

cout << "After exiting block.\n";

} catch(bad_alloc exc) {
cout << "Allocation failure!\n";
return 1;
}

cout << "Done\n";

return 0;
}

This program produces the following output when the display option is turned on:
Constructing GCPtr.
Constructing GCPtr.
Value at p is: 10
Value at q is: 11
Before entering block.

籍
Constructing GCPtr.
Value at r is: 12

书
GCPtr going out of scope.
Before garbage collection for gclist<int, 0>:
memPtr refcount value

典 书
[002F31D8] 0 12
[002F12F0] 1 11
[002F12C0] 1 10

经 读
[00000000] 0 ---

+
Deleting: 12

不
After garbage collection for gclist<int, 0>:

+
memPtr refcount value
[002F12F0] 1 11

C 典
[002F12C0] 1 10

After exiting block.

经
Done
GCPtr going out of scope.
Before garbage collection for gclist<int, 0>:

无
memPtr refcount value
[002F12F0] 0 11
[002F12C0] 1 10

Deleting: 11
After garbage collection for gclist<int, 0>:
memPtr refcount value
[002F12C0] 1 10

GCPtr going out of scope.

Before garbage collection for gclist<int, 0>:
memPtr refcount value
[002F12C0] 0 10

Deleting: 10
After garbage collection for gclist<int, 0>:
memPtr refcount value
-- Empty --

Examine this program and its output closely. First, notice that p and q are created at the start of main( ), but r is not
created until its block is entered. As you know, in C++, local variables are not created until their block is entered. When
.

r is created, the memory to which it points is given an initial value of 12. This value is then displayed, and the block
ends. This causes r to go out of scope, which means that its destructor is called. This causes r’s reference count in
gclist to be decremented to zero. Then collect( ) is called to collect garbage.

Because the display option is on, when collect( ) begins, it displays the contents of gclist. Notice that it has four
entries. The first is the one that was previously linked to r. Notice that its refcount field is zero, indicating that the
memory pointed to by the memPtr field is no longer in use by any program element. The next two entries are still
active, and they are linked to p and q. Because they are still in use, the memory they point to is not freed at this time.
The final entry represents the null pointer to which p and q originally pointed when they were created. Because it is no
longer in use, it will be removed from the list by collect( ). (Of course, no memory is freed when the null pointer is
removed.)

Because no other GCPtr points to the same memory as r, its memory can be released, as the Deleting: 12 line
confirms. Once this is done, program execution continues after the block. Finally, p and q go out of scope when the
program ends and their memory is released. In this case, q’s destructor is called first, meaning that it is collected first.
Finally, p is destroyed and gclist is empty.

Allocating and Discarding Objects

籍
It is important to understand that memory becomes subject to garbage collection as soon as its reference count drops
to zero (which means that no GCPtr is pointing to it). It is not necessary for the GCPtr that originally pointed to that
object to go out of scope. Thus, you can use a single GCPtr object to point to any number of allocated objects by

书
simply assigning that GCPtr a new value. The discarded memory will eventually be collected. For example:
// Allocate and discard objects.

典 书
#include <iostream>
#include <new>
#include "gc.h"

+
int main() {
经 不 读
using namespace std;

+
try {

C
// Allocate and discard objects.

典
GCPtr<int> p = new int(1);
p = new int(2);
p = new int(3);

经
p = new int(4);

// Manually collect unused objects for

无
// demonstration purposes.
GCPtr<int>::collect();

cout << "*p: " << *p << endl;

} catch(bad_alloc exc) {
cout << "Allocation failure!\n";
return 1;
}

return 0;
}

The output from this program, with the display option turned on, is shown here:
Constructing GCPtr.
Before garbage collection for gclist<int, 0>:
memPtr refcount value
[002F1310] 1 4
[002F1300] 0 3
[002F12D0] 0 2
[002F12A0] 0 1
.

Deleting: 3
Deleting: 2
Deleting: 1
After garbage collection for gclist<int, 0>:
memPtr refcount value
[002F1310] 1 4

*p: 4
GCPtr going out of scope.
Before garbage collection for gclist<int, 0>:
memPtr refcount value
[002F1310] 0 4

Deleting: 4
After garbage collection for gclist<int, 0>:
memPtr refcount value
-- Empty --

In the program, p, a GCPtr to int, is assigned a pointer to four separate chunks of dynamic memory, each being

籍
initialized with a different value. Next, a call is made to collect( ), which forces garbage collection to take place. Notice
the contents of gclist: three of the entries are marked inactive, and only the entry that points to the memory that was
allocated last is still in use. Next, the unused entries are deleted. Finally, the program ends, p goes out of scope, and

书
the final entry is removed.

典 书
Notice that the first three chunks of dynamic memory to which p pointed have reference counts of zero. This is
because of the way the overloaded assignment operator works. Recall that when a GCPtr is assigned a new address,
the reference count for its original value is decremented. Thus, each time p is assigned the address of a new integer,

经 读
the reference count for the old address is reduced.

+
One other point: because p was initialized when it was declared, no null-pointer entry was generated and put on gclist.

不
Remember, a null-pointer entry is created only when a GCPtr is declared without an initial value.

C+
Allocating Arrays

典
If you are allocating an array using new, then you must tell GCPtr this fact by specifying its size when the GCPtr pointer
to that array is declared. For example, here is the way to allocate an array of five doubles:

经
GCPtr<double, 5> pda = new double[5];

无
The size must be specified for two reasons. First, it tells the GCPtr constructor that this object will point to an allocated
array, which causes the isArray field to be set to true. When isArray is true, the collect( ) function frees memory by
using delete[ ] , which releases a dynamically allocated array, rather than delete, which releases only a single object.
Therefore, in this example, when pda goes out of scope, delete[ ] is used and all five elements of pda are freed.
Ensuring that the correct number of objects are freed is especially important when arrays of class objects are
allocated. Only by using delete[ ] can you know that the destructor for each object will be called.

The second reason that the size must be specified is to prevent an out-of-bounds element from being accessed when
an Iter is used to cycle through an allocated array. Recall that the size of the array (stored in arraySize) is passed by
GCPtr to Iter’s constructor whenever an Iter is needed.

Be aware that nothing enforces the rule that an allocated array be operated on only through a GCPtr that has been
specified as pointing to an array. This is solely your responsibility.

Once you have allocated an array, there are two ways you can access its elements. First, you can index the GCPtr that
points to it. Second, you can use an iterator. Both methods are shown here.

Using Array Indexing

The following program creates a GCPtr to a 10-element array of ints. It then allocates that array and initializes it to the
values 0 through 9. Finally, it displays those values. It performs these actions by indexing the GCPtr.
.

// Demonstrate indexing a GCPtr.

#include <iostream>
#include <new>
#include "gc.h"

using namespace std;

int main() {

try {
// Create a GCPtr to an allocated array of 10 ints.
GCPtr<int, 10> ap = new int[10];

// Give the array some values using array indexing.

for(int i=0; i < 10; i++)
ap[i] = i;

// Now, show the contents of the array.

for(int i=0; i < 10; i++)
cout << ap[i] << " ";

cout << endl;

书 籍
} catch(bad_alloc exc) {
cout << "Allocation failure!\n";

典 书
return 1;
}

}
return 0;

+ 经 不 读
+
The output, with the display option off, is shown here:
0123456789

C 典
Because a GCPtr emulates a normal C++ pointer, no array bounds checking is performed, and it is possible to overrun
or under run the dynamically allocated array. So, use the same care when accessing an array through a GCPtr as you

经
do when accessing an array through a normal C++ pointer.

Using Iterators

无
Although array indexing is certainly a convenient method of cycling through an allocated array, it is not the only
method at your disposal. For many applications, the use of an iterator will be a better choice because it has the
advantage of preventing boundary errors. Recall that for GCPtr, iterators are objects of type Iter. Iter supports the full
complement of pointer operations, such as ++. It also allows an iterator to be indexed like an array.

Here is the previous program reworked to use an iterator. Recall that the easiest way to obtain an iterator to a GCPtr is
to use GCiterator, which is a typedef inside GCPtr that is automatically bound to the generic type T.
// Demonstrate an iterator.
#include <iostream>
#include <new>
#include "gc.h"

using namespace std;

int main() {

try {
// Create a GCPtr to an allocated array of 10 ints.
GCPtr<int, 10> ap = new int[10];
.

// Declare an int iterator.

GCPtr<int>::GCiterator itr;

// Assign itr a pointer to the start of the array.

itr = ap.begin();

// Give the array some values using array indexing.

for(unsigned i=0; i < itr.size(); i++)
itr[i] = i;

// Now, cycle through array using the iterator.

for(itr = ap.begin(); itr != ap.end(); itr++)
cout << *itr << " ";

cout << endl;

} catch(bad_alloc exc) {
cout << "Allocation failure!\n";
return 1;
} catch(OutOfRangeExc exc) {

籍
cout << "Out of range access!\n";
return 1;
}

}
return 0;

典 书 书
经 读
On your own, you might want to try incrementing itr so that it points beyond the boundary of the allocated array. Then
try accessing the value at that location. As you will see, an OutOfRangeExc is thrown. In general, you can increment or

+
decrement an iterator any way you like without causing an exception. However, if it is not pointing within the underlying

不
array, attempting to obtain or set the value at that location will cause a boundary error.

C+
Using GCPtr with Class Types

典
GCPtr is used with class types in just the same way it is used with built-in types. For example, here is a short program
that allocates objects of MyClass:

经
// Use GCPtr with a class type.
#include <iostream>

无
#include <new>
#include "gc.h"

using namespace std;

class MyClass {
int a, b;
public:
double val;

MyClass() { a = b = 0; }

MyClass(int x, int y) {
a = x;
b = y;
val = 0.0;
}

~MyClass() {
cout << "Destructing MyClass(" <<
a << ", " << b << ")\n";
}
.

int sum() {
return a + b;
}

friend ostream &operator<<(ostream &strm, MyClass &obj);

};

// An overloaded inserter to display MyClass.

ostream &operator<<(ostream &strm, MyClass &obj) {
strm << "(" << obj.a << " " << obj.b << ")";
return strm;
}

int main() {
try {
GCPtr<MyClass> ob = new MyClass(10, 20);

// Show value via overloaded inserter.

cout << *ob << endl;

籍
// Change object pointed to by ob.
ob = new MyClass(11, 21);

书
cout << *ob << endl;

典 书
// Call a member function through a GCPtr.
cout << "Sum is : " << ob->sum() << endl;

经 读
// Assign a value to a class member through a GCPtr.
ob->val = 98.6;

+
cout << "ob->val: " << ob->val << endl;

+ 不
cout << "ob is now " << *ob << endl;
} catch(bad_alloc exc) {

C 典
cout << "Allocation error!\n";
return 1;
}

经
return 0;
}

无
Notice how the members of MyClass are accessed through the use of the –> operator. Remember, GCPtr defines a
pointer type. Thus, operations through a GCPtr are performed in exactly the same fashion that they are with any other
pointer.

The output from the program, with the display option turned off, is shown here:
(10 20)
(11 21)
Sum is : 32
ob->val: 98.6
ob is now (11 21)
Destructing MyClass(11, 21)
Destructing MyClass(10, 20)

Pay special attention to the last two lines. These are output by ~MyClass( ) when garbage is collected. Even though
only one GCPtr pointer was created, two MyClass objects were allocated. Both of these objects are represented by
entries in the garbage collection list. When ob is destroyed, gclist is scanned for entries having a reference count of
zero. In this case, two such entries are found, and the memory to which they point is deleted.

A Larger Demonstration Program

.

The following program shows a larger example that exercises all of the features of GCPtr:
// Demonstrating GCPtr.
#include <iostream>
#include <new>
#include "gc.h"

using namespace std;

// A simple class for testing GCPtr with class types.

class MyClass {
int a, b;
public:
double val;

MyClass() { a = b = 0; }

MyClass(int x, int y) {
a = x;
b = y;

籍
val = 0.0;
}

书
~MyClass() {
cout << "Destructing MyClass(" <<
a << ", " << b << ")\n";

典 书
}

int sum() {

经 读
return a + b;
}

};

+ 不
friend ostream &operator<<(ostream &strm, MyClass &obj);

+
C 典
// Create an inserter for MyClass.
ostream &operator<<(ostream &strm, MyClass &obj) {
strm << "(" << obj.a << " " << obj.b << ")";

经
return strm;
}

无
// Pass a normal pointer to a function.
void passPtr(int *p) {
cout << "Inside passPtr(): "
<< *p << endl;
}

// Pass a GCPtr to a function.

void passGCPtr(GCPtr<int, 0> p) {
cout << "Inside passGCPtr(): "
<< *p << endl;
}

int main() {

try {
// Declare an int GCPtr.
GCPtr<int> ip;

// Allocate an int and assign its address to ip.

ip = new int(22);
// Display its value.
cout << "Value at *ip: " << *ip << "\n\n";

// Pass ip to a function
passGCPtr(ip);

// ip2 is created and then goes out of scope

{
GCPtr<int> ip2 = ip;
}

int *p = ip; // convert to int * pointer'

passPtr(p); // pass int * to passPtr()

*ip = 100; // Assign new value to ip

// Now, use implicit conversion to int *

passPtr(ip);
cout << endl;

籍
// Create a GCPtr to an array of ints
GCPtr<int, 5> iap = new int[5];

书
// Initialize dynamic array.

典 书
for(int i=0; i < 5; i++)
iap[i] = i;

经 读
// Display contents of array.
cout << "Contents of iap via array indexing.\n";

+
for(int i=0; i < 5; i++)

不
cout << iap[i] << " ";

+
cout << "\n\n";

C 典
// Create an int GCiterator.
GCPtr<int>::GCiterator itr;

经
// Now, use iterator to access dynamic array.
cout << "Contents of iap via iterator.\n";
for(itr = iap.begin(); itr != iap.end(); itr++)

无
cout << *itr << " ";
cout << "\n\n";

// Generate and discard many objects

for(int i=0; i < 10; i++)
ip = new int(i+10);

// Now, manually garbage collect GCPtr<int> list.

// Keep in mind that GCPtr<int, 5> pointers
// will not be collected by this call.
cout << "Requesting collection on GCPtr<int> list.\n";
GCPtr<int>::collect();

// Now, use GCPtr with class type.

GCPtr<MyClass> ob = new MyClass(10, 20);

// Show value via overloaded insertor.

cout << "ob points to " << *ob << endl;

// Change object pointed to by ob.

ob = new MyClass(11, 21);
cout << "ob now points to " <<. *ob << endl;s
.

// Call a member function through a GCPtr.

cout << "Sum is : " << ob->sum() << endl;

// Assign a value to a class member through a GCPtr.

ob->val = 19.21;
cout << "ob->val: " << ob->val << "\n\n";

cout << "Now work with pointers to class objects.\n";

// Declare a GCPtr to a 5-element array

// of MyClass objects.
GCPtr<MyClass, 5> v;

// Allocate the array.

v = new MyClass[5];

// Get a MyClass GCiterator.

GCPtr<MyClass>::GCiterator mcItr;

籍
// Initialize the MyClass array.
for(int i=0; i<5; i++) {
v[i] = MyClass(i, 2*i);

书
}

典 书
// Display contents of MyClass array using indexing.
cout << "Cycle through array via array indexing.\n";
for(int i=0; i<5; i++) {

经 读
cout << v[i] << " ";
}

+
cout << "\n\n";

+ 不
// Display contents of MyClass array using iterator.
cout << "Cycle through array through an iterator.\n";

C 典
for(mcItr = v.begin(); mcItr != v.end(); mcItr++) {
cout << *mcItr << " ";
}

经
cout << "\n\n";

// Here is another way to write the preceding loop.

无
cout << "Cycle through array using a while loop.\n";
mcItr = v.begin();
while(mcItr != v.end()) {
cout << *mcItr << " ";
mcItr++;
}
cout << "\n\n";

cout << "mcItr points to an array that is "

<< mcItr.size() << " objects long.\n";

// Find number of elements between two iterators.

GCPtr<MyClass>::GCiterator mcItr2 = v.end()-2;
mcItr = v.begin();
cout << "The difference between mcItr2 and mcItr is "
<< mcItr2 - mcItr;
cout << "\n\n";

// Can also cycle through loop like this.

cout << "Dynamically compute length of array.\n";
mcItr = v.begin();
mcItr2 = v.end();
.

for(int i=0; i < mcItr2 - mcItr; i++) {

cout << v[i] << " ";
}
cout << "\n\n";

// Now, display the array backwards.

cout << "Cycle through array backwards.\n";
for(mcItr = v.end()-1; mcItr >= v.begin(); mcItr--)
cout << *mcItr << " ";
cout << "\n\n";

// Of course, can use "normal" pointer to

// cycle through array.
cout << "Cycle through array using 'normal' pointer\n";
MyClass *ptr = v;
for(int i=0; i < 5; i++)
cout << *ptr++ << " ";
cout << "\n\n";

籍
// Can access members through a GCiterator.
cout << "Access class members through an iterator.\n";
for(mcItr = v.begin(); mcItr != v.end(); mcItr++) {

书
cout << mcItr->sum() << " ";
}

典 书
cout << "\n\n";

// Can allocate and delete a pointer to a GCPtr

经 读
// normally, just like any other pointer.
cout << "Use a pointer to a GCPtr.\n";

+
GCPtr<int> *pp = new GCPtr<int>();

不
*pp = new int(100);

+
cout << "Value at **pp is: " << **pp;
cout << "\n\n";

C 典
// Because pp is not a garbage collected pointer,
// it must be deleted manually.

经
delete pp;
} catch(bad_alloc exc) {
// A real application could attempt to free

无
// memory by collect() when an allocation
// error occurs.
cout << "Allocation error.\n";
}

return 0;
}

Here is the output with the display option turned off:

Value at *ip: 22

Inside passGCPtr(): 22
Inside passPtr(): 22
Inside passPtr(): 100

Contents of iap via array indexing.

01234

Contents of iap via iterator.

01234
.

Requesting collection on GCPtr<int> list.

ob points to (10 20)
ob now points to (11 21)
Sum is : 32
ob->val: 19.21

Now work with pointers to class objects.

Destructing MyClass(0, 0)
Destructing MyClass(1, 2)
Destructing MyClass(2, 4)
Destructing MyClass(3, 6)
Destructing MyClass(4, 8)
Cycle through array via array indexing.
(0 0) (1 2) (2 4) (3 6) (4 8)

Cycle through array through an iterator.

(0 0) (1 2) (2 4) (3 6) (4 8)

Cycle through array using a while loop.

(0 0) (1 2) (2 4) (3 6) (4 8)

籍
mcItr points to an array that is 5 objects long.
The difference between mcItr2 and mcItr is 3

书
Dynamically compute length of array.

典 书
(0 0) (1 2) (2 4) (3 6) (4 8)

Cycle through array backwards.

经 读
(4 8) (3 6) (2 4) (1 2) (0 0)

+
Cycle through array using 'normal' pointer

不
(0 0) (1 2) (2 4) (3 6) (4 8)

+
Access class members through an iterator.

C
0 3 6 9 12

典
Use a pointer to a GCPtr.

经
Value at **pp is: 100

Destructing MyClass(4, 8)

无
Destructing MyClass(3, 6)
Destructing MyClass(2, 4)
Destructing MyClass(1, 2)
Destructing MyClass(0, 0)
Destructing MyClass(11, 21)
Destructing MyClass(10, 20)

On your own, try compiling and running this program with the display option turned on. (That is, define DISPLAY in
gc.h.) Next, walk through the program, matching the output against each statement. This will give you a good feel for
the way the garbage collector works. Remember, garbage collection occurs whenever a GCPtr goes out of scope. This
happens at various points in the program, such as when a function that receives a copy of a GCPtr returns. In this
case, the copy goes out of scope and garbage collection takes place. Also remember that each type of GCPtr
maintains its own gclist. Thus, collecting garbage from one list does not cause it to be collected from other types of
lists.

Load Testing

The following program load tests GCPtr by repeatedly allocating and discarding objects until free memory is
exhausted. When this occurs, a bad_alloc exception is thrown by new. Inside the exception handler, collect( ) is
explicitly called to reclaim the unused memory, and the process continues. You can use this same technique in your
own programs.
.

// Load test GCPtr by creating and discarding

// thousands of objects.
#include <iostream>
#include <new>
#include <limits>
#include "gc.h"

using namespace std;

// A simple class for load testing GCPtr.

class LoadTest {
int a, b;
public:
double n[100000]; // just to take up memory
double val;

LoadTest() { a = b = 0; }

LoadTest(int x, int y) {
a = x;

籍
b = y;
val = 0.0;
}

};

典 书
friend ostream &operator<<(ostream &strm, LoadTest &obj);

书
// Create an inserter for LoadTest.

经 读
ostream &operator<<(ostream &strm, LoadTest &obj) {
strm << "(" << obj.a << " " << obj.b << ")";

+
return strm;

不
}

+
int main() {

C 典
GCPtr<LoadTest> mp;
int i;

经
for(i = 1; i < 20000; i++) {
try {
mp = new LoadTest(i, i);

无
} catch(bad_alloc xa) {
// When an allocation error occurs, recycle
// garbage by calling collect().
cout << "Last object: " << *mp << endl;
cout << "Length of gclist before calling collect(): "
<< mp.gclistSize() << endl;
GCPtr<LoadTest>::collect();
cout << "Length after calling collect(): "
<< mp.gclistSize() << endl;
}
}

return 0;
}

A portion of the output from the program (with the display option off) is shown here. Of course, the precise output you
see may differ because of the amount of memory available in your system and the compiler that you are using.
Last object: (518 518)
Length of gclist before calling collect(): 518
Length after calling collect(): 1
Last object: (1036 1036)
Length of gclist before calling collect(): 518
Length after calling collect(): 1
Last object: (1554 1554)
Length of gclist before calling collect(): 518
Length after calling collect(): 1
Last object: (2072 2072)
Length of gclist before calling collect(): 518
Length after calling collect(): 1
Last object: (2590 2590)
Length of gclist before calling collect(): 518
Length after calling collect(): 1
Last object: (3108 3108)
Length of gclist before calling collect(): 518
Length after calling collect(): 1
Last object: (3626 3626)
Length of gclist before calling collect(): 518
Length after calling collect(): 1

Some Restrictions

籍
There are a few restrictions to using GCPtr:
1. You cannot create global GCPtrs. Recall that a global object goes out of scope after the rest of the

书
program ends. When a global GCPtr goes out of scope, the GCPtr destructor calls collect( ) to try
to release the unused memory. The trouble is that, depending on how your C++ compiler is

书
implemented, gclist may have already been destroyed. In this case, executing collect( ) will cause

典
a runtime error. Therefore, GCPtr should be used only when creating local objects.

经 读
2. When using dynamically allocated arrays, you must specify the size of the array when you declare
a GCPtr that will point to it. There is no mechanism that enforces this, however, so be careful.

+ 不
3. You must not attempt to release the memory pointed to by a GCPtr by explicitly using delete. If

+
you need to immediately release an object, call collect( ).

C
4. A GCPtr object must point only to memory that is dynamically allocated via new. Assigning to a

典
GCPtr object a pointer to any other memory will cause an error when the GCPtr object goes out of
scope because an attempt will be made to free memory that was never allocated.

经
5. It is best to avoid circular pointer references for reasons described earlier in this chapter. Although
all allocated memory is eventually released, objects containing circular references remain

无
allocated until the program ends, rather than being released when they are no longer used by
another program element.
 Some Things to Try
It is easy to tailor GCPtr to the needs of your applications. As explained earlier, one of the changes that you might want
to try is collecting garbage only after some metric has been reached, such as gclist reaching a certain size, or after a
certain number of GCPtrs have gone out of scope.

An interesting enhancement to GCPtr is to overload new so that it automatically collects garbage when an allocation
failure occurs. It is also possible to bypass the use of new when allocating memory for a GCPtr and use factory
functions defined by GCPtr instead. Doing this lets you carefully control the dynamic allocation of memory, but it makes
the allocation process fundamentally different than it is for C++’s built-in approach.

You might want to experiment with other solutions to the circular reference problem. One way is to implement the
concept of a weak reference, which does not prevent garbage collection from occurring. You would then use a weak
reference whenever a circular reference was needed.

Perhaps the most interesting variation on GCPtr is found in Chapter 3. There, a multithreaded version is created in

籍
which garbage collection takes place automatically, when free CPU time exists.

典 书 书
+ 经 不 读
C+ 典
无 经
 Chapter 3: Multithreading in C++

Overview
Multithreading is becoming an increasingly important part of modern programming. One reason for this is that
multithreading enables a program to make the best use of available CPU cycles, thus allowing very efficient programs
to be written. Another reason is that multithreading is a natural choice for handling event-driven code, which is so
common in today’s highly distributed, networked, GUI-based environments. Of course, the fact that the most widely
used operating system, Windows, supports multithreading is also a factor. Whatever the reasons, the increased use of
multithreading is changing the way that programmers think about the fundamental architecture of a program. Although
C++ does not contain built-in support for multithreaded programs, it is right at home in this arena.

Because of its growing importance, this chapter explores using C++ to create multithreaded programs. It does so by
developing two multithreaded applications. The first is a thread control panel, which you can use to control the

籍
execution of threads within a program. This is both an interesting demonstration of multithreading and a practical tool
that you can use when developing multithreaded applications. The second example shows how to apply multithreading
to a practical example by creating a modified version of the garbage collector from Chapter 2 that runs in a background

书
thread.

This chapter also serves another purpose: it shows how adept C++ is at interfacing directly to the operating system. In

典 书
some other languages, such as Java, there is a layer of processing between your program and the OS. This layer
adds overhead that can be unacceptable for some types of programs, such as those used in a real-time environment.

经 读
In sharp contrast, C++ has direct access to low-level functionality provided by the operating system. This is one of the
reasons C++ can produce higher performance code.

+ + 不
C 典
无 经
 What Is Multithreading?
Before beginning, it is necessary to define precisely what is meant by the term multithreading. Multithreading is a
specialized form of multitasking. In general, there are two types of multitasking: process-based and thread-based. A
process is, in essence, a program that is executing. Thus, process-based multitasking is the feature that allows your
computer to run two or more programs concurrently. For example, it is process-based multitasking that allows you to
run a word processor at the same time you are using a spreadsheet or browsing the Internet. In process-based
multitasking, a program is the smallest unit of code that can be dispatched by the scheduler.

A thread is a dispatchable unit of executable code. The name comes from the concept of a “thread of execution.” In a
thread-based multitasking environment, all processes have at least one thread, but they can have more. This means
that a single program can perform two or more tasks concurrently. For instance, a text editor can be formatting text at
the same time that it is printing, as long as these two actions are being performed by two separate threads. The
differences between process-based and thread-based multitasking can be summarized like this: Process-based
multitasking handles the concurrent execution of programs. Thread-based multitasking deals with the concurrent
execution of pieces of the same program.

籍
In the preceding discussions, it is important to clarify that true concurrent execution is possible only in a multiple-CPU

书
system in which each process or thread has unrestricted access to a CPU. For single CPU systems, which constitute
the vast majority of systems in use today, only the appearance of simultaneous execution is achieved. In a single CPU
system, each process or thread receives a portion of the CPU’s time, with the amount of time determined by several

典 书
factors, including the priority of the process or thread. Although truly concurrent execution does not exist on most
computers, when writing multithreaded programs, you should assume that it does. This is because you can’t know the

读
precise order in which separate threads will be executed, or if they will execute in the same sequence twice. Thus, it’s

经
best to program as if true concurrent execution is the case.

+ 不
Multithreading Changes the Architecture of a Program

+
Multithreading changes the fundamental architecture of a program. Unlike a single-threaded program that executes in

C 典
a strictly linear fashion, a multithreaded program executes portions of itself concurrently. Thus, all multithreaded
programs include an element of parallelism. Consequently, a major issue in multithreaded programs is managing the
interaction of the threads.

经
As explained earlier, all processes have at least one thread of execution, which is called the main thread. The main
thread is created when your program begins. In a multithreaded program, the main thread creates one or more child

无
threads. Thus, each multithreaded process starts with one thread of execution and then creates one or more additional
threads. In a properly designed program, each thread represents a single logical unit of activity.

The principal advantage of multithreading is that it enables you to write very efficient programs because it lets you
utilize the idle time that is present in most programs. Most I/O devices, whether they are network ports, disk drives, or
the keyboard, are much slower than the CPU. Often, a program will spend a majority of its execution time waiting to
send or receive data. With the careful use of multithreading, your program can execute another task during this idle
time. For example, while one part of your program is sending a file over the Internet, another part can be reading
keyboard input, and still another can be buffering the next block of data to send.
 Why Doesn’t C++ Contain Built-In Support for Multithreading?
C++ does not contain any built-in support for multithreaded applications. Instead, it relies entirely upon the operating
system to provide this feature. Given that both Java and C# provide built-in support for multithreading, it is natural to
ask why this isn’t also the case for C++. The answers are efficiency, control, and the range of applications to which
C++ is applied. Let’s examine each.

By not building in support for multithreading, C++ does not attempt to define a “one size fits all” solution. Instead, C++
allows you to directly utilize the multithreading features provided by the operating system. This approach means that
your programs can be multithreaded in the most efficient way supported by the execution environment. Because many
multitasking environments offer rich support for multithreading, being able to access that support is crucial to the
creation of high-performance, multithreaded programs.

Using operating system functions to support multithreading gives you access to the full range of control offered by the
execution environment. Consider Windows. It defines a rich set of thread-related functions that enable finely grained
control over the creation and management of a thread. For example, Windows has several ways to control access to a

籍
shared resource, including semaphores, mutexes, event objects, waitable timers, and critical sections. This level of
flexibility cannot be easily designed into a language because the capabilities of operating systems differ. Thus,

书
language-level support for multithreading usually means offering only a “lowest common denominator” of features.
With C++, you gain access to all the features that the operating system provides. This is a major advantage when
writing high-performance code.

经 典 读 书
C++ was designed for all types of programming, from embedded systems in which there is no operating system in the
execution environment to highly distributed, GUI-based end-user applications and everything in between. Therefore,
C++ cannot place significant constraints on its execution environment. Building in support for multithreading would
have inherently limited C++ to only those environments that supported it and thus prevented C++ from being used to

+ 不
create software for nonthreaded environments.

+
In the final analysis, not building in support for multithreading is a major advantage for C++ because it enables

C
programs to be written in the most efficient way possible for the target execution environment. Remember, C++ is all

典
about power. In the case of multithreading, it is definitely a situation in which “less is more.”

无 经
 What Operating System and Compiler?
Because C++ relies on the operating system to provide support for multithreaded programming, it is necessary to
choose an operating system as the target for the multithreaded applications in this chapter. Because Windows is the
most widely used operating system in the world, it is the operating system used in this chapter. However, much of the
information can be generalized to any OS that supports multithreading.

Because Visual C++ is arguably the most widely used compiler for producing Windows programs, it is the compiler
required by the examples in this chapter. The importance of this is made apparent in the following section. However, if
you are using another compiler, the code can be easily adapted to accommodate it.

Note The examples in this chapter assume a basic, working knowledge of Windows programming.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

An Overview of the Windows Thread Functions

Windows offers a wide array of Application Programming Interface (API) functions that support multithreading. Many
readers will be at least somewhat familiar with the multithreading functions offered by Windows, but for those who are
not, an overview of those used in this chapter is presented here. Keep in mind that Windows provides many other
multithreading-based functions that you might want to explore on your own.

To use Windows’ multithreading functions, you must include <windows.h> in your program.

Creating and Terminating a Thread

To create a thread, the Windows API supplies the CreateThread( ) function. Its prototype is shown here:
HANDLE CreateThread(LPSECURITY_ATTRIBUTES secAttr,
SIZE_T stackSize,
LPTHREAD_START_ROUTINE threadFunc,

籍
LPVOID param,

DWORD flags,

书
LPDWORD threadID);

典 书
Here, secAttr is a pointer to a set of security attributes pertaining to the thread. However, if secAttr is NULL, then the
default security descriptor is used.

经 读
Each thread has its own stack. You can specify the size of the new thread’s stack in bytes using the stackSize
parameter. If this integer value is zero, then the thread will be given a stack that is the same size as the creating

+
thread. In this case, the stack will be expanded, if necessary. (Specifying zero is the common approach taken to

不
thread stack size.)

+
Each thread of execution begins with a call to a function, called the thread function, within the creating process.

C 典
Execution of the thread continues until the thread function returns. The address of this function (that is, the entry point
to the thread) is specified in threadFunc. All thread functions must have this prototype:
DWORD WINAPI threadfunc(LPVOID param);

无 经
Any argument that you need to pass to the new thread is specified in CreateThread( )’s param. This 32-bit value is
received by the thread function in its parameter. This parameter may be used for any purpose. The function returns its
exit status.

The flags parameter determines the execution state of the thread. If it is zero, the thread begins execution
immediately. If it is CREATE_SUSPEND, the thread is created in a suspended state, awaiting execution. (It may be
started using a call to ResumeThread( ), discussed later.)

The identifier associated with a thread is returned in the long integer pointed to by threadID.

The function returns a handle to the thread if successful or NULL if a failure occurs. The thread handle can be explicitly
destroyed by calling CloseHandle( ). Otherwise, it will be destroyed automatically when the parent process ends.

As just explained, a thread of execution terminates when its entry function returns. The process may also terminate the
thread manually, using either TerminateThread( ) or ExitThread( ), whose prototypes are shown here:
BOOL TerminateThread(HANDLE thread, DWORD status);
VOID ExitThread(DWORD status);

For TerminateThread( ), thread is the handle of the thread to be terminated. ExitThread( ) can only be used to terminate
the thread that calls ExitThread( ). For both functions, status is the termination status. TerminateThread( ) returns
nonzero if successful and zero otherwise.

Calling ExitThread( ) is functionally equivalent to allowing a thread function to return normally. This means that the
.

stack is properly reset. When a thread is terminated using TerminateThread( ), it is stopped immediately and does not
perform any special cleanup activities. Also, TerminateThread( ) may stop a thread during an important operation. For
these reasons, it is usually best (and easiest) to let a thread terminate normally when its entry function returns.

The Visual C++ Alternatives to CreateThread( ) and ExitThread( )

Although CreateThread( ) and ExitThread( ) are the Windows API functions used to create and terminate a thread, we
won’t be using them in this chapter! The reason is that when these functions are used with Visual C++ (and possibly
other Windows-compatible compilers), they can result in memory leaks, the loss of a small amount of memory. For
Visual C++, if a multithreaded program utilizes C/C++ standard library functions and uses CreateThread( ) and
ExitThread( ), then small amounts of memory are lost. (If your program does not use the C/C++ standard library, then
no such losses will occur.) To eliminate this problem, you must use functions defined by the Visual C++ runtime library
to start and stop threads rather than those specified by the Win32 API. These functions parallel CreateThread( ) and
ExitThread( ), but do not generate a memory leak.

Note If you are using a compiler other than Visual C++, check its documentation to determine if you need to bypass
CreateThread( ) and ExitThread( ) and how to do so, if necessary.

籍
The Visual C++ alternatives to CreateThread( ) and ExitThread( ) are _beginthreadex( ) and _endthreadex( ). Both
require the header file <process.h>. Here is the prototype for _beginthreadex( ) :

书
uintptr_t _beginthreadex(void *secAttr, unsigned stackSize,
unsigned (__stdcall *threadFunc)(void *),
void *param, unsigned flags,

典 书
unsigned *threadID);

As you can see, the parameters to _beginthreadex( ) parallel those to CreateThread( ). Furthermore, they have the

经 读
same meaning as those specified by CreateThread( ). secAttr is a pointer to a set of security attributes pertaining to the
thread. However, if secAttr is NULL, then the default security descriptor is used. The size of the new thread’s stack, in

+ 不
bytes, is passed in stackSize parameter. If this value is zero, then the thread will be given a stack that is the same size

+
as the main thread of the process that creates it.

C
The address of the thread function (that is, the entry point to the thread) is specified in threadFunc. For

典
_beginthreadex( ) , a thread function must have this prototype:

unsigned __stdcall threadfunc(void * param);

经
This prototype is functionally equivalent to the one for CreateThread( ), but it uses different type names. Any argument
that you need to pass to the new thread is specified in the param parameter.

无
The flags parameter determines the execution state of the thread. If it is zero, the thread begins execution
immediately. If it is CREATE_SUSPEND, the thread is created in a suspended state, awaiting execution. (It may be
started using a call to ResumeThread( ).) The identifier associated with a thread is returned in the double word pointed
to by threadID.

The function returns a handle to the thread if successful or zero if a failure occurs. The type uintptr_t specifies a Visual
C++ type capable of holding a pointer or handle.

The prototype for _endthreadex( ) is shown here:

void _endthreadex(unsigned status);

It functions just like ExitThread( ) by stopping the thread and returning the exit code specified in status.

Because the most widely used compiler for Windows is Visual C++, the examples in this chapter will use
_beginthreadex( ) and _endthreadex( ) rather than their equivalent API functions. If you are using a compiler other than
Visual C++, simply substitute CreateThread( ) and EndThread( ).

When using _beginthreadex( ) and _endthreadex( ), you must remember to link in the multithreaded library. This will
vary from compiler to compiler. Here are some examples. When using the Visual C++ command-line compiler, include
the –MT option. To use the multithreaded library from the Visual C++ 6 IDE, first activate the Project | Settings
.

property sheet. Then, select the C/C++ tab. Next, select Code Generation from the Category list box and then choose
Multithreaded in the Use Runtime Library list box. For Visual C++ 7 .NET IDE, select Project | Properties. Next, select
the C/C++ entry and highlight Code Generation. Finally, choose Multithreaded as the runtime library.

Suspending and Resuming a Thread

A thread of execution can be suspended by calling SuspendThread( ). It can be resumed by calling ResumeThread( ).
The prototypes for these functions are shown here:
DWORD SuspendThread(HANDLE hThread);
DWORD ResumeThread(HANDLE hThread);

For both functions, the handle to the thread is passed in hThread.

Each thread of execution has associated with it a suspend count. If this count is zero, then the thread is not
suspended. If it is nonzero, the thread is in a suspended state. Each call to SuspendThread( ) increments the suspend
count. Each call to ResumeThread( ) decrements the suspend count. A suspended thread will resume only when its
suspend count has reached zero. Therefore, to resume a suspended thread implies that there must be the same
number of calls to ResumeThread( ) as there have been calls to SuspendThread( ).

籍
Both functions return the thread’s previous suspend count or –1 if an error occurs.

Changing the Priority of a Thread

典 书
In Windows, each thread has associated with it a priority setting. A thread’s priority determines how much CPU time a

书
thread receives. Low priority threads receive little time. High priority threads receive a lot. Of course, how much CPU
time a thread receives has a profound impact on its execution characteristics and its interaction with other threads
currently executing in the system.

+ 经 读
In Windows, a thread’s priority setting is the combination of two values: the overall priority class of the process and the
priority setting of the individual thread relative to that priority class. That is, a thread’s actual priority is determined by

不
combining the process’ priority class with the thread’s individual priority level. Each is examined next.

C+
Priority Classes

典
By default, a process is given a priority class of normal, and most programs remain in the normal priority class
throughout their execution lifetime. Although neither of the examples in this chapter changes the priority class, a brief

经
overview of the thread priority classes is given here in the interest of completeness.

Windows defines six priority classes, which correspond to the values shown here, in order of highest to lowest priority:

无
REALTIME_PRIORITY_CLASS

HIGH_PRIORITY_CLASS

ABOVE_NORMAL_PRIORITY_CLASS

NORMAL_PRIORITY_CLASS

BELOW_NORMAL_PRIORITY_CLASS

IDLE_PRIORITY_CLASS

Programs are given the NORMAL_PRIORITY_CLASS by default. Usually, you won’t need to alter the priority class of
your program. In fact, changing a process’ priority class can have negative consequences on the overall performance
of the computer system. For example, if you increase a program’s priority class to REALTIME_PRIORITY_CLASS , it will
dominate the CPU. For some specialized applications, you may need to increase an application’s priority class, but
usually you won’t. As mentioned, neither of the applications in this chapter changes the priority class.

In the event that you do want to change the priority class of a program, you can do so by calling SetPriorityClass( ).
You can obtain the current priority class by calling GetPriorityClass( ). The prototypes for these functions are shown
.

here:
DWORD GetPriorityClass(HANDLE hApp);
BOOL SetPriorityClass(HANDLE hApp, DWORD priority);

Here, hApp is the handle of the process. GetPriorityClass( ) returns the priority class of the application or zero on
failure. For SetPriorityClass( ), priority specifies the process’ new priority class.

Thread Priorities

For any given priority class, each individual thread’s priority determines how much CPU time it receives within its
process. When a thread is first created, it is given normal priority, but you can change a thread’s priority—even while it
is executing.

You can obtain a thread’s priority setting by calling GetThreadPriority( ). You can increase or decrease a thread’s
priority using SetThreadPriority( ) . The prototypes for these functions are shown here:
BOOL SetThreadPriority(HANDLE hThread, int priority);
int GetThreadPriority(HANDLE hThread);

For both functions, hThread is the handle of the thread. For SetThreadPriority( ) , priority is the new priority setting. If an

籍
error occurs, SetThreadPriority( ) returns zero. It returns nonzero otherwise. For GetThreadPriority( ), the current
priority setting is returned. The priority settings are shown here, in order of highest to lowest:

典 书
Thread Priority

书
THREAD_PRIORITY_TIME_CRITICAL

THREAD_PRIORITY_HIGHEST
Value

15

+ 经 不 读
THREAD_PRIORITY_ABOVE_NORMAL

THREAD_PRIORITY_NORMAL
1

+
THREAD_PRIORITY_BELOW_NORMAL -1

C 典
THREAD_PRIORITY_LOWEST -2

THREAD_PRIORITY_IDLE -15

无 经
These values are increments or decrements that are applied relative to the priority class of the process. Through the
combination of a process’ priority class and thread priority, Windows supports 31 different priority settings for
application programs.

GetThreadPriority( ) returns THREAD_PRIORITY_ERROR_RETURN if an error occurs.

For the most part, if a thread has the normal priority class, you can freely experiment with changing its priority setting
without fear of catastrophically affecting overall system performance. As you will see, the thread control panel
developed in the next section allows you to alter the priority setting of a thread within a process (but does not change
its priority class).

Obtaining the Handle of the Main Thread

It is possible to control the execution of the main thread. To do so, you will need to acquire its handle. The easiest way
to do this is to call GetCurrentThread( ), whose prototype is shown here:
HANDLE GetCurrentThread(void);

This function returns a pseudohandle to the current thread. It is called a pseudohandle because it is a predefined
value that always refers to the current thread rather than specifically to the calling thread. It can, however, be used any
place that a normal thread handle can.

Synchronization
.

When using multiple threads or processes, it is sometimes necessary to coordinate the activities of two or more. This
process is called synchronization. The most common use of synchronization occurs when two or more threads need
access to a shared resource that must be used by only one thread at a time. For example, when one thread is writing
to a file, a second thread must be prevented from doing so at the same time. Another reason for synchronization is
when one thread is waiting for an event that is caused by another thread. In this case, there must be some means by
which the first thread is held in a suspended state until the event has occurred. Then the waiting thread must resume
execution.

There are two general states that a task may be in. First, it may be executing (or ready to execute as soon as it obtains
its time slice). Second, a task may be blocked, awaiting some resource or event, in which case its execution is
suspended until the needed resource is available or the event occurs.

If you are not familiar with the synchronization problem or its most common solution, the semaphore, the next section
discusses it.

Understanding the Synchronization Problem

Windows must provide special services that allow access to a shared resource to be synchronized, because without
help from the operating system, there is no way for one process or thread to know that it has sole access to a

籍
resource. To understand this, imagine that you are writing programs for a multitasking operating system that does not
provide any synchronization support. Further imagine that you have two concurrently executing threads, A and B, both
of which, from time to time, require access to some resource R (such as a disk file) that must be accessed by only one

书
thread at a time. As a means of preventing one thread from accessing R while the other is using it, you try the
following solution. First, you establish a variable called flag that is initialized to zero and can be accessed by both

典 书
threads. Then, before using each piece of code that accesses R, you wait for flag to be cleared, then set flag, access
R, and finally, clear flag. That is, before either thread accesses R, it executes this piece of code:

读
while(flag) ; // wait for flag to be cleared

经
flag = 1; // set flag

+ 不
// ... access resource R ...

+
flag = 0; // clear the flag

C 典
The idea behind this code is that neither thread will access R if flag is set. Conceptually, this approach is in the spirit of
the correct solution. However, in actual fact it leaves much to be desired for one simple reason: it won’t always work!
Let’s see why.

经
Using the code just given, it is possible for both processes to access R at the same time. The while loop is, in
essence, performing repeated load and compare instructions on flag or, in other words, it is testing flag’s value. When

无
flag is cleared, the next line of code sets flag’s value. The trouble is that it is possible for these two operations to be
performed in two different time slices. Between the two time slices, the value of flag might have been accessed by the
other thread, thus allowing R to be used by both threads at the same time. To understand this, imagine that thread A
enters the while loop and finds that flag is zero, which is the green light to access R. However, before it can set flag to
1, its time slice expires and thread B resumes execution. If B executes its while, it too will find that flag is not set and
assume that it is safe to access R. However, when A resumes it will also begin accessing R. The crucial aspect of the
problem is that the testing and setting of flag do not comprise one uninterruptible operation. Rather, as just illustrated,
they can be separated by a time slice. No matter how you try, there is no way, using only application-level code, that
you can absolutely guarantee that one and only one thread will access R at one time.

The solution to the synchronization problem is as elegant as it is simple. The operating system (in this case Windows)
provides a routine that in one uninterrupted operation, tests and, if possible, sets a flag. In the language of operating
systems engineers, this is called a test and set operation. For historical reasons, the flags used to control access to a
shared resource and provide synchronization between threads (and processes) are called semaphores. The
semaphore is at the core of the Windows synchronization system.

The Windows Synchronization Objects

Windows supports several types of synchronization objects. The first type is the classic semaphore. When using a
semaphore, a resource can be completely synchronized, in which case one and only one thread or process can
.

access it at any one time, or the semaphore can allow no more than a small number of processes or threads access at
any one time. Semaphores are implemented using a counter that is decremented when a task is granted the
semaphore and incremented when the task releases it.

The second synchronization object is the mutex semaphore, or just mutex, for short. A mutex synchronizes a resource
such that one and only one thread or process can access it at any one time. In essence, a mutex is a special case
version of a standard semaphore.

The third synchronization object is the event object. It can be used to block access to a resource until some other
thread or process signals that it can be used. (That is, an event object signals that a specified event has occurred.)

The fourth synchronization object is the waitable timer. A waitable timer blocks a thread’s execution until a specific
time. You can also create timer queues, which are lists of timers.

You can prevent a section of code from being used by more than one thread at a time by making it into a critical
section using a critical section object. Once a critical section is entered by one thread, no other thread may use it until
the first thread has left the critical section.

The only synchronization object used in this chapter is the mutex, which is described in the following section.
However, all synchronization objects defined by Windows are available to the C++ programmer. As explained, this is

籍
one of the major advantages that results from C++’s reliance on the operating system to handle multithreading: all
multithreading features are at your command.

典 书
Using a Mutex to Synchronize Threads

书
As explained, a mutex is a special-case semaphore that allows only one thread to access a resource at any given
time. Before you can use a mutex, you must create one using CreateMutex( ), whose prototype is shown here:

经 读
HANDLE CreateMutex(LPSECURITY_ATTRIBUTES secAttr,
BOOL acquire,

+ 不
LPCSTR name);

+
Here, secAttr is a pointer to the security attributes. If secAttr is NULL, the default security descriptor is used.

C 典
If the creating thread desires control of the mutex, then acquire must be true. Otherwise, pass false.

The name parameter points to a string that becomes the name of the mutex object. Mutexes are global objects, which

经
may be used by other processes. As such, when two processes each open a mutex using the same name, both are
referring to the same mutex. In this way, two processes can be synchronized. The name may also be NULL, in which
case the semaphore is localized to one process.

无
The CreateMutex( ) function returns a handle to the semaphore if successful or NULL on failure. A mutex handle is
automatically closed when the main process ends. You can explicitly close a mutex handle when it is no longer
needed by calling CloseHandle( ).

Once you have created a semaphore, you use it by calling two related functions: WaitForSingleObject( ) and
ReleaseMutex( ). The prototypes for these functions are shown here:

DWORD WaitForSingleObject(HANDLE hObject, DWORD howLong);

BOOL ReleaseMutex(HANDLE hMutex);

WaitForSingleObject( ) waits on a synchronization object. It does not return until the object becomes available or a
time-out occurs. For use with mutexes, hObject will be the handle of a mutex. The howLong parameter specifies, in
milliseconds, how long the calling routine will wait. Once that time has elapsed, a time-out error will be returned. To
wait indefinitely, use the value INFINITE. The function returns WAIT_OBJECT_0 when successful (that is, when access
is granted). It returns WAIT_TIMEOUT when time-out is reached.

ReleaseMutex( ) releases the mutex and allows another thread to acquire it. Here, hMutex is the handle to the mutex.
The function returns nonzero if successful and zero on failure.

To use a mutex to control access to a shared resource, wrap the code that accesses that resource between a call to
WaitForSingleObject( ) and ReleaseMutex( ), as shown in this skeleton. (Of course, the time-out period will differ from
application to application.)
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT) {
// handle time-out error
}

// access the resource

ReleaseMutex(hMutex);

Generally, you will want to choose a time-out period that will be more than enough to accommodate the actions of your
program. If you get repeated time-out errors when developing a multithreaded application, it usually means that you
have created a deadlock condition. Deadlock occurs when one thread is waiting on a mutex that another thread never
releases.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

Creating a Thread Control Panel

When developing multithreaded programs, it is often useful to experiment with various priority settings. It is also useful
to be able to dynamically suspend and resume a thread, or even terminate a thread. As you will see, it is quite easy,
using the thread functions just described, to create a thread control panel that allows you to accomplish these things.
Further, you can use the control panel while your multithreaded program is running. The dynamic nature of the thread
control panel allows you to easily change the execution profile of a thread and observe the results.

The thread control panel developed in this section is capable of controlling one thread. However, you can create as
many panels as needed, with each controlling a different thread. For the sake of simplicity, the control panel is
implemented as a modeless dialog box that is owned by the desktop, not the application whose thread it controls.

The thread control panel is capable of performing the following actions:

Setting a thread’s priority

籍
Suspending a thread

Resuming a thread

书
Terminating a thread

书
It also displays the current priority setting of the thread. The thread control dialog box is shown in Figure 3-1.

经 典 读
+ + 不
C 典
无 经
Figure 3-1: The Thread Control dialog box

As stated, the control panel is a modeless dialog box. As you know, when a modeless dialog box is activated, the rest
of the application is still active. Thus, the control panel runs independently of the application for which it is being used.

The Thread Control Panel

The code for the thread control panel is shown here. This file is called tcp.cpp.
// A thread control panel.

#include <map>
#include <windows.h>
#include "panel.h"
using namespace std;

const int NUMPRIORITIES = 5;

const int OFFSET = 2;

// Array of strings for priority list box.

char priorities[NUMPRIORITIES][80] = {
"Lowest",
"Below Normal",
"Normal",
"Above Normal",
"Highest"
};

// A Thread Control Panel Class.

class ThrdCtrlPanel {
// Information about the thread under control.
struct ThreadInfo {
HANDLE hThread; // handle of thread

籍
int priority; // current priority
bool suspended; // true if suspended
ThreadInfo(HANDLE ht, int p, bool s) {

书
hThread = ht;
priority = p;

典 书
suspended = s;
}
};

+ 经 读
// This map holds a ThreadInfo for each
// active thread control panel.

不
static map<HWND, ThreadInfo> dialogmap;

+
public:

C 典
// Construct a control panel.
ThrdCtrlPanel(HINSTANCE hInst, HANDLE hThrd);

经
// The control panel's callback function.
static LRESULT CALLBACK ThreadPanel(HWND hwnd, UINT message,

无
WPARAM wParam, LPARAM lParam);
};

// Define static member dialogmap.

map<HWND, ThrdCtrlPanel::ThreadInfo>
ThrdCtrlPanel::dialogmap;

// Create a thread control panel.

ThrdCtrlPanel::ThrdCtrlPanel(HINSTANCE hInst,
HANDLE hThrd)
{
ThreadInfo ti(hThrd,
GetThreadPriority(hThrd)+OFFSET,
false);

// Owner window is desktop.

HWND hDialog = CreateDialog(hInst, "ThreadPanelDB",
NULL,
(DLGPROC) ThreadPanel);

// Put info about this dialog box in the map.

dialogmap.insert(pair<HWND, ThreadInfo>(hDialog,
. ti));s
.

// Set the control panel's title.

char str[80] = "Control Panel for Thread ";
char str2[4];
_itoa(dialogmap.size(), str2, 10);
strcat(str, str2);
SetWindowText(hDialog, str);

// Offset each dialog box instance.

MoveWindow(hDialog, 30*dialogmap.size(),
30*dialogmap.size(),
300, 250, 1);

// Update priority setting in the list box.

SendDlgItemMessage(hDialog, IDD_LB, LB_SETCURSEL,
(WPARAM) ti.priority, 0);

// Increase priority to ensure control. You can

// change or remove this statement based on your
// execution environment.

籍
SetThreadPriority(GetCurrentThread(),
THREAD_PRIORITY_ABOVE_NORMAL);
}

典 书 书
// Thread control panel dialog box callback function.
LRESULT CALLBACK ThrdCtrlPanel::ThreadPanel(HWND hwnd,
UINT message,

经 读
WPARAM wParam,
LPARAM lParam)

+
{

不
int i;

+
HWND hpbRes, hpbSus, hpbTerm;

C 典
switch(message) {
case WM_INITDIALOG:
// Initialize priority list box.

经
for(i=0; i<NUMPRIORITIES; i++) {
SendDlgItemMessage(hwnd, IDD_LB,
LB_ADDSTRING, 0, (LPARAM) priorities[i]);

无
}

// Set suspend and resume buttons for thread.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, true); // enable Suspend
EnableWindow(hpbRes, false); // disable Resume
return 1;
case WM_COMMAND:
map<HWND, ThreadInfo>::iterator p = dialogmap.find(hwnd);

switch(LOWORD(wParam)) {
case IDD_TERMINATE:
TerminateThread(p->second.hThread, 0);

// Disable Terminate button.

hpbTerm = GetDlgItem(hwnd, IDD_TERMINATE); }
EnableWindow(hpbTerm, false); // disable

// Disable Suspend and Resume buttons.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
.

EnableWindow(hpbSus, false); // disable Suspend

EnableWindow(hpbRes, false); // disable Resume

return 1;
case IDD_SUSPEND:
SuspendThread(p->second.hThread);

// Set state of the Suspend and Resume buttons.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, false); // disable Suspend
EnableWindow(hpbRes, true); // enable Resume

p->second.suspended = true;
return 1;
case IDD_RESUME:
ResumeThread(p->second.hThread);

// Set state of the Suspend and Resume buttons.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);

籍
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, true); // enable Suspend
EnableWindow(hpbRes, false); // disable Resume

典 书
p->second.suspended = false;
return 1;
case IDD_LB:

书
// If a list box entry was clicked,

经 读
// then change the priority.
if(HIWORD(wParam)==LBN_DBLCLK) {

+
p->second.priority = SendDlgItemMessage(hwnd,

不
IDD_LB, LB_GETCURSEL,

+
0, 0);
SetThreadPriority(p->second.hThread,

C 典
p->second.priority-OFFSET);
}
return 1;

经
case IDCANCEL:
// If thread is suspended when panel is closed,
// then resume thread to prevent deadlock.

无
if(p->second.suspended) {
ResumeThread(p->second.hThread);
p->second.suspended = false;
}

// Remove this thread from the list.

dialogmap.erase(hwnd);

// Close the panel.

DestroyWindow(hwnd);
return 1;
}
}
return 0;
}

The control panel requires the following resource file, called tcp.rc:
#include <windows.h>
#include "panel.h"

ThreadPanelDB DIALOGEX 20, 20, 140, 110

.

CAPTION "Thread Control Panel"

STYLE WS_BORDER | WS_VISIBLE | WS_POPUP | WS_CAPTION | WS_SYSMENU
{
DEFPUSHBUTTON "Done", IDCANCEL, 55, 80, 33, 14
PUSHBUTTON "Terminate", IDD_TERMINATE, 10, 20, 42, 12
PUSHBUTTON "Suspend", IDD_SUSPEND, 10, 35, 42, 12
PUSHBUTTON "Resume", IDD_RESUME, 10, 50, 42, 12
LISTBOX IDD_LB, 65, 20, 63, 42, LBS_NOTIFY | WS_VISIBLE |
WS_BORDER | WS_VSCROLL | WS_TABSTOP
CTEXT "Thread Priority", IDD_TEXT1, 65, 8, 64, 10
CTEXT "Change State", IDD_TEXT2, 0, 8, 64, 10
}

The control panel uses the following header file called panel.h:
#define IDD_LB 200
#define IDD_TERMINATE 202
#define IDD_SUSPEND 204
#define IDD_RESUME 206
#define IDD_TEXT1 208
#define IDD_TEXT2 209

书 籍
To use the thread control panel, follow these steps:
1. Include tcp.cpp in your program.

2. Include tcp.rc in your program’s resource file.

典 书
3. Create the thread or threads that you want to control.

经 读
4. Instantiate a ThrdCtrlPanel object for each thread.

+
Each ThrdCtrlPanel object links a thread with a dialog box that controls it.

+ 不
For large projects in which multiple files need access to ThrdCtrlPanel, you will need to use a header file called tcp.h
that contains the declaration for ThrdCtrlPanel. Here is tcp.h:

C 典
// A header file for the ThrdCtrlPanel class.

class ThrdCtrlPanel {

经
public:

// Construct a control panel.

无
ThrdCtrlPanel(HINSTANCE hInst, HANDLE hThrd);

// The control panel's callback function.

static LRESULT CALLBACK ThreadPanel(HWND hwnd, UINT message,
WPARAM wParam, LPARAM lParam);
};

A Closer Look at the Thread Control Panel

Let’s take a closer look at the thread control panel. It begins by defining the following global definitions:
const int NUMPRIORITIES = 5;
const int OFFSET = 2;

// Array of strings for priority list box.

char priorities[NUMPRIORITIES][80] = {
"Lowest",
"Below Normal",
"Normal",
"Above Normal",
"Highest"
};
.

The priorities array holds strings that correspond to a thread’s priority setting. It initializes the list box inside the control
panel that displays the current thread priority. The number of priorities is specified by NUMPRIORITIES, which is 5 for
Windows. Thus, NUMPRIORITIES defines the number of different priorities that a thread may have. (If you adapt the
code for use with another operating system, a different value might be required.) Using the control panel, you can set a
thread to one of the following priorities:
THREAD_PRIORITY_HIGHEST
THREAD_PRIORITY_ABOVE_NORMAL
THREAD_PRIORITY_NORMAL
THREAD_PRIORITY_BELOW_NORMAL
THREAD_PRIORITY_LOWEST

The other two thread priority settings:

THREAD_PRIORITY_TIME_CRITICAL
THREAD_PRIORITY_IDLE

are not supported because, relative to the control panel, they are of little practical value. For example, if you want to
create a time-critical application, you are better off making its priority class time-critical.

籍
OFFSET defines an offset that will be used to translate between list box indexes and thread priorities. You should recall
that normal priority has the value zero. In this example, the highest priority is THREAD_PRIORITY_HIGHEST, which is 2.

书
The lowest priority is THREAD_PRIORITY_LOWEST, which is –2. Because list box indexes begin at zero, the offset is
used to convert between indexes and priority settings.

典 书
Next, the ThrdCtrlPanel class is declared. It begins as shown here:
// A Thread Control Panel Class.

读
class ThrdCtrlPanel {

经
// Information about the thread under control.

+
struct ThreadInfo {

不
HANDLE hThread; // handle of thread

+
int priority; // current priority
bool suspended; // true if suspended

C
ThreadInfo(HANDLE ht, int p, bool s) {

典
hThread = ht;
priority = p;
suspended = s;

经
}
};

无
// This map holds a ThreadInfo for each
// active thread control panel.
static map<HWND, ThreadInfo> dialogmap;

Information about the thread under control is contained within a structure of type ThreadInfo. The handle of the thread
is stored in hThread. Its priority is stored in priority. If the thread is suspended, then suspended will be true. Otherwise,
suspended will be false.

The static member dialogmap is an STL map that links the thread information with the handle of the dialog box used to
control that thread. Because there can be more than one thread control panel active at any given time, there must be
some way to determine which thread is associated with which panel. It is dialogmap that provides this linkage.

The ThreadCtrlPanel Constructor

The ThrdCtrlPanel constructor is shown here. The constructor is passed the instance handle of the application and the
handle of the thread being controlled. The instance handle is needed to create the control panel dialog box.
// Create a thread control panel.
ThrdCtrlPanel::ThrdCtrlPanel(HINSTANCE hInst,
HANDLE hThrd)
{
.

ThreadInfo ti(hThrd,
GetThreadPriority(hThrd)+OFFSET,
false);

// Owner window is desktop.

HWND hDialog = CreateDialog(hInst, "ThreadPanelDB",
NULL,
(DLGPROC) ThreadPanel);

// Put info about this dialog box in the map.

dialogmap.insert(pair<HWND, ThreadInfo>(hDialog, ti));

// Set the control panel's title.

char str[80] = "Control Panel for Thread ";
char str2[4];
_itoa(dialogmap.size(), str2, 10);
strcat(str, str2);
SetWindowText(hDialog, str);

// Offset each dialog box instance.

籍
MoveWindow(hDialog, 30*dialogmap.size(),
30*dialogmap.size(),
300, 250, 1);

典 书
// Update priority setting in the list box.

书
SendDlgItemMessage(hDialog, IDD_LB, LB_SETCURSEL,
(WPARAM) ti.priority, 0);

经 读
// Increase priority to ensure control. You can
// change or remove this statement based on your

+
// execution environment.

不
SetThreadPriority(GetCurrentThread(),

+
THREAD_PRIORITY_ABOVE_NORMAL);
}

C 典
The constructor begins by creating a ThreadInfo instance called ti that contains the initial settings for the thread. Notice
that the priority is obtained by calling GetThreadPriority( ) for the thread being controlled. Next, the control panel dialog

经
box is created by calling CreateDialog( ). CreateDialog( ) is a Windows API function that creates a modeless dialog box,
which makes it independent of the application that creates it. The handle of this dialog box is returned and stored in
hDialog. Next, hDialog and the thread information contained in ti are stored in dialogmap. Thus, the thread is linked

无
with the dialog box that controls it.

Next, the title of the dialog box is set to reflect the number of the thread. The number of the thread is obtained based
on the number of entries in dialogmap. An alternative that you might want to try implementing is to explicitly pass a
name for each thread to the ThrdCtrlPanel constructor. For the purposes of this chapter, simply numbering each thread
is sufficient.

Next, the control panel’s position on the screen is offset a bit by calling MoveWindow( ) , another Windows API function.
This enables multiple panels to be displayed without each one fully covering the one before it. The thread’s priority
setting is then displayed in the priority list box by calling the Windows API function SendDlgItemMessage( ).

Finally, the current thread has its priority increased to above normal. This ensures that the application receives enough
CPU time to be responsive to user input no matter what is the priority level of the thread under control. This step may
not be needed in all cases. You can experiment to find out.

The ThreadPanel( ) Function

ThreadPanel( ) is the Windows callback function that responds to user interaction with the thread control panel. Like all
dialog box callback functions, it receives a message each time the user changes the state of a control. It is passed the
handle of the dialog box in which the action occurred, the message, and any additional information required by the
message. Its general mode of operation is the same as that for any other callback function used by a dialog box. The
.

following discussion describes what happens for each message.

When the thread control panel dialog box is first created, it receives a WM_INITDIALOG message, which is handled by
this case sequence:
case WM_INITDIALOG:
// Initialize priority list box.
for(i=0; i<NUMPRIORITIES; i++) {
SendDlgItemMessage(hwnd, IDD_LB,
LB_ADDSTRING, 0, (LPARAM) priorities[i]);
}

// Set Suspend and Resume buttons for thread.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, true); // enable Suspend
EnableWindow(hpbRes, false); // disable Resume
return 1;

This initializes the priority list box and sets the Suspend and Resume buttons to their initial states, which are Suspend
enabled and Resume disabled.

籍
Each user interaction generates a WM_COMMAND message. Each time this message is received, an iterator to this

书
dialog box’s entry in dialogmap is retrieved, as shown here:
case WM_COMMAND:
map<HWND, ThreadInfo>::iterator p = dialogmap.find(hwnd);

经 典 读 书
The information pointed to by p will be used to properly process each action. Because p is an iterator for a map, it
points to an object of type pair , which is a structure defined by the STL. This structure contains two fields: first and
second. These fields correspond to the information that comprises the key and the value, respectively. In this case, the
handle is the key and the thread information is the value.

+ 不
A code indicating precisely what action has occurred is contained in the low-order word of wParam , which is used to

+
control a switch statement that handles the remaining messages. Each is described next.

C 典
When the user presses the Terminate button, the thread under control is stopped. This is handled by this case
sequence:
case IDD_TERMINATE:

经
TerminateThread(p->second.hThread, 0);

无
// Disable Terminate button.
hpbTerm = GetDlgItem(hwnd, IDD_TERMINATE);
EnableWindow(hpbTerm, false); // disable

// Disable Suspend and Resume buttons.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, false); // disable Suspend
EnableWindow(hpbRes, false); // disable Resume

return 1;

The thread is stopped with a call to TerminateThread( ). Notice how the handle for the thread is obtained. As explained,
because p is an iterator for a map, it points to an object of type pair that contains the key in its first field and the value
in its second field. This is why the thread handle is obtained by the expression p->second.hThread. After the thread is
stopped, the Terminate button is disabled.

Once a thread has been terminated, it cannot be resumed. Notice that the control panel uses TerminateThread( ) to
halt execution of a thread. As mentioned earlier, this function must be used with care. If you use the control panel to
experiment with threads of your own, you will want to make sure that no harmful side effects are possible.

When the user presses the Suspend button, the thread is suspended. This is accomplished by the following sequence:
.

case IDD_SUSPEND:
SuspendThread(p->second.hThread);

// Set state of the Suspend and Resume buttons.

hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, false); // disable Suspend
EnableWindow(hpbRes, true); // enable Resume

p->second.suspended = true;
return 1;

The thread is suspended by a call to SuspendThread( ). Next, the state of the Suspend and Resume buttons are
updated such that Resume is enabled and Suspend is disabled. This prevents the user from attempting to suspend a
thread twice.

A suspended thread is resumed when the Resume button is pressed. It is handled by this code:
case IDD_RESUME:
ResumeThread(p->second.hThread);

籍
// Set state of the Suspend and Resume buttons.
hpbSus = GetDlgItem(hwnd, IDD_SUSPEND);

书
hpbRes = GetDlgItem(hwnd, IDD_RESUME);
EnableWindow(hpbSus, true); // enable Suspend
EnableWindow(hpbRes, false); // disable Resume

return 1;

经 典 读 书
p->second.suspended = false;

The thread is resumed by a call to ResumeThread( ), and the Suspend and Resume buttons are set appropriately.

+ 不
To change a thread’s priority, the user double-clicks an entry in the Priority list box. This event is handled as shown

+
next:

C
case IDD_LB:

典
// If a list box entry was double-clicked,
// then change the priority.
if(HIWORD(wParam)==LBN_DBLCLK) {

经
p->second.priority = SendDlgItemMessage(hwnd,
IDD_LB, LB_GETCURSEL,
0, 0);

无
SetThreadPriority(p->second.hThread,
p->second.priority-OFFSET);
}
return 1;

List boxes generate various types of notification messages that describe the precise type of event that occurred.
Notification messages are contained in the high-order word of wParam . One of these messages is LBN_DBLCLK,
which means that the user double-clicked an entry in the box. When this notification is received, the index of the entry
is retrieved by calling the Windows API function SendDlgItemMessage( ), requesting the current selection. This value is
then used to set the thread’s priority. Notice that OFFSET is subtracted to normalize the value of the index.

Finally, when the user closes the thread control panel dialog box, the IDCANCEL message is sent. It is handled by the
following sequence:
case IDCANCEL:
// If thread is suspended when panel is closed,
// then resume thread to prevent deadlock.
if(p->second.suspended) {
ResumeThread(p->second.hThread);
p->second.suspended = false;
}
.

// Remove this thread from the list.

dialogmap.erase(hwnd);

// Close the panel.

DestroyWindow(hwnd);
return 1;

If the thread was suspended, it is restarted. This is necessary to avoid accidentally deadlocking the thread. Next, this
dialog box’s entry in dialogmap is removed. Finally, the dialog box is removed by calling the Windows API function
DestroyWindow( ).

Demonstrating the Control Panel

Here is a program that includes the thread control panel and demonstrates its use. Sample output is shown in Figure
3-2. The program creates a main window and defines two child threads. When started, these threads simply count
from 0 to 50,000, displaying the count in the main window. These threads can be controlled by activating a thread
control panel.

书 籍
经 典 读 书
+ + 不
C 典
无 经
Figure 3-2: Sample output from the thread control panel sample program

To use the program, first begin execution of the threads by selecting Start Threads from the Threads menu (or by
pressing F2) and then activate the thread control panels by selecting Control Panels from the Threads menu (or by
pressing F3). Once the control panels are active, you can experiment with different priority settings and so on.
.

Note It is beyond the scope of this book to teach Windows programming. However, the operation of this sample
program is straightforward and should be easily understood by all Windows programmers.
// Demonstrate the thread control panel.
#include <windows.h>
#include <process.h>
#include "thrdapp.h"
#include "tcp.cpp"

const int MAX = 500000;

LRESULT CALLBACK WindowFunc(HWND, UINT, WPARAM, LPARAM);

unsigned __stdcall MyThread1(void * param);

unsigned __stdcall MyThread2(void * param);

char str[255]; // holds output strings

unsigned tid1, tid2; // thread IDs

HANDLE hThread1, hThread2; // thread handles

书 籍
HINSTANCE hInst; // instance handle

int WINAPI WinMain(HINSTANCE hThisInst, HINSTANCE hPrevInst,

LPSTR args, int winMode)

典 书
{
HWND hwnd;
MSG msg;

经 读
WNDCLASSEX wcl;
HACCEL hAccel;

+ + 不
// Define a window class.
wcl.cbSize = sizeof(WNDCLASSEX);

C 典
wcl.hInstance = hThisInst; // handle to this instance
wcl.lpszClassName = "MyWin"; // window class name
wcl.lpfnWndProc = WindowFunc; // window function

经
wcl.style = 0; // default style

wcl.hIcon = LoadIcon(NULL, IDI_APPLICATION); // large icon

无
wcl.hIconSm = NULL; // use small version of large icon
wcl.hCursor = LoadCursor(NULL, IDC_ARROW); // cursor style

wcl.lpszMenuName = "ThreadAppMenu"; // main menu

wcl.cbClsExtra = 0; // no extra memory needed

wcl.cbWndExtra = 0;

// Make the window background white.

wcl.hbrBackground = (HBRUSH) GetStockObject(WHITE_BRUSH);

// Register the window class.

if(!RegisterClassEx(&wcl)) return 0;

/* Now that a window class has been registered, a window

can be created. */
hwnd = CreateWindow(
wcl.lpszClassName, // name of window class
"Using a Thread Control Panel", // title
WS_OVERLAPPEDWINDOW, // window style - normal
CW_USEDEFAULT, // X coordinate - let Windows decide
CW_USEDEFAULT, // Y coordinate - let Windows decide
.

260, // width
200, // height
NULL, // no parent window
NULL, // no override of class menu
hThisInst, // instance handle
NULL // no additional arguments
);

hInst = hThisInst; // save instance handle

// Load the keyboard accelerators.

hAccel = LoadAccelerators(hThisInst, "ThreadAppMenu");

// Display the window.

ShowWindow(hwnd, winMode);
UpdateWindow(hwnd);

// Create the message loop.

while(GetMessage(&msg, NULL, 0, 0))

籍
{
if(!TranslateAccelerator(hwnd, hAccel, &msg)) {
TranslateMessage(&msg); // translate keyboard messages

书
DispatchMessage(&msg); // return control to Windows
}
}

经 典 读 书
return msg.wParam;

+
/* This function is called by Windows and is passed

不
messages from the message queue.

+
*/
LRESULT CALLBACK WindowFunc(HWND hwnd, UINT message,

C
WPARAM wParam, LPARAM lParam)

典
{
int response;

经
switch(message) {
case WM_COMMAND:

无
switch(LOWORD(wParam)) {
case IDM_THREAD: // create the threads
hThread1 = (HANDLE) _beginthreadex(NULL, 0,
MyThread1, (void *) hwnd,
0, &tid1);
hThread2 = (HANDLE) _beginthreadex(NULL, 0,
MyThread2, (void *) hwnd,
0, &tid2);
break;
case IDM_PANEL: // activate control panel
ThrdCtrlPanel(hInst, hThread1);
ThrdCtrlPanel(hInst, hThread2);
break;
case IDM_EXIT:
response = MessageBox(hwnd, "Quit the Program?",
"Exit", MB_YESNO);
if(response == IDYES) PostQuitMessage(0);
break;
case IDM_HELP:
MessageBox(hwnd,
"F1: Help\nF2: Start Threads\nF3: Panel",
"Help", MB_OK);
break;
.

}
break;
case WM_DESTROY: // terminate the program
PostQuitMessage(0);
break;
default:
return DefWindowProc(hwnd, message, wParam, lParam);
}

return 0;
}

// First thread.
unsigned __stdcall MyThread1(void * param)
{
int i;
HDC hdc;

for(i=0; i<MAX; i++) {

籍
wsprintf(str, "Thread 1: loop # %5d ", i);
hdc = GetDC((HWND) param);
TextOut(hdc, 1, 1, str, lstrlen(str));

书
ReleaseDC((HWND) param, hdc);
}

典 书
return 0;
}

经 读
// Second thread.

+
unsigned __stdcall MyThread2(void * param)

不
{

+
int i;
HDC hdc;

C 典
for(i=0; i<MAX; i++) {
wsprintf(str, "Thread 2: loop # %5d ", i);
hdc = GetDC((HWND) param);

经
TextOut(hdc, 1, 20, str, lstrlen(str));
ReleaseDC((HWND) param, hdc);

无
}

return 0;
}

This program requires the header file thrdapp.h, shown here:

#define IDM_THREAD 100
#define IDM_HELP 101
#define IDM_PANEL 102
#define IDM_EXIT 103

The resource file required by the program is shown here:

#include <windows.h>
#include "thrdapp.h"
#include "tcp.rc"

ThreadAppMenu MENU
{
POPUP "&Threads" {
MENUITEM "&Start Threads\tF2", IDM_THREAD
MENUITEM "&Control Panels\tF3", IDM_PANEL
MENUITEM "E&xit\tCtrl+X", IDM_EXIT
 }
MENUITEM "&Help", IDM_HELP
}

ThreadAppMenu ACCELERATORS
{
VK_F1, IDM_HELP, VIRTKEY
VK_F2, IDM_THREAD, VIRTKEY
VK_F3, IDM_PANEL, VIRTKEY
"^X", IDM_EXIT
}

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

A Multithreaded Garbage Collector

Although controlling threads using the thread control panel is useful when developing multithreaded programs,
ultimately it is using threads that makes them important. Toward this end, this chapter shows a multithreaded version
of the GCPtr garbage collector class originally developed in Chapter 2. Recall that the version of GCPtr shown in
Chapter 2 collected unused memory each time a GCPtr object went out of scope. Although this approach is fine for
some applications, often a better alternative is to have the garbage collector run as a background task, recycling
memory whenever free CPU cycles are available. The implementation developed here is designed for Windows, but
the same basic techniques apply to other multithreaded environments.

To convert GCPtr into a background task is actually fairly easy, but it does involve a number of changes. Here are the
main ones:
1. Member variables that support the thread must be added to GCPtr. These variables include the
thread handle, the mutex handle, and an instance counter that keeps track
of the number of GCPtr objects in existence.

籍
2. The constructor for GCPtr must start the garbage collection thread. The constructor must also
create the mutex that controls synchronization. This must happen only once, when the first GCPtr

书
object is created.

3. Another exception must be defined that will be used to indicate a time-out condition.

经 典 读 书 4. The GCPtr destructor must no longer call collect( ). Garbage collection is handled
by the garbage collection thread.

5. A function called gc( ) that serves as the thread entry point for the garbage collector must be

+
defined.

C+ 不 6. A function called isRunning( ) must be defined. It returns true if the garbage collection is in use.

7. The member functions of GCPtr that access the garbage collection list contained in gclist must be

典
synchronized so that only one thread at a time can access the list.

经
The following sections show the changes.

The Additional Member Variables

无
The multithreaded version of GCPtr requires that the following member variables be added:
// These support multithreading.
unsigned tid; // thread id
static HANDLE hThrd; // thread handle
static HANDLE hMutex; // handle of mutex

static int instCount; // counter of GCPtr objects

The ID of the thread used by the garbage collector is stored in tid. This member is unused except in the call to
_beginthreadex( ) . The handle to the thread is stored in hThrd . The handle of the mutex used to synchronize access to
GCPtr is stored in hMutex. A count of GCPtr objects in existence is maintained in instCount. The last three are static
because they are shared by all instances of GCPtr. They are defined like this, outside of GCPtr:
template <class T, int size>
int GCPtr<T, size>::instCount = 0;

template <class T, int size>

HANDLE GCPtr<T, size>::hMutex = 0;

template <class T, int size>

HANDLE GCPtr<T, size>::hThrd = 0;
.

The Multithreaded GCPtr Constructor

In addition to its original duties, the multithreaded GCPtr( ) must create the mutex, start the garbage collector thread,
and update the instance counter. Here is the updated version:
// Construct both initialized and uninitialized objects.
GCPtr(T *t=NULL) {

// When first object is created, create the mutex

// and register shutdown().
if(hMutex==0) {
hMutex = CreateMutex(NULL, 0, NULL);
atexit(shutdown);
}

if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

list<GCInfo<T> >::iterator p;

籍
p = findPtrInfo(t);

// If t is already in gclist, then

书
// increment its reference count.
// Otherwise, add it to the list.
if(p != gclist.end())

典 书
p->refcount++; // increment ref count
else {

读
// Create and store this entry.

经
GCInfo<T> gcObj(t, size);

+
gclist.push_front(gcObj);

不
}

+
addr = t;

C
arraySize = size;

典
if(size > 0) isArray = true;
else isArray = false;

经
// Increment instance counter for each new object.
instCount++;

无
// If the garbage collection thread is not
// currently running, start it running.
if(hThrd==0) {
hThrd = (HANDLE) _beginthreadex(NULL, 0, gc,
(void *) 0, 0, (unsigned *) &tid);

// For some applications, it will be better

// to lower the priority of the garbage collector
// as shown here:
//
// SetThreadPriority(hThrd,
// THREAD_PRIORITY_BELOW_NORMAL);
}

ReleaseMutex(hMutex);
}

Let’s examine this code closely. First, if hMutex is zero, it means that this is the first GCPtr object to be created and no
mutex has yet been created for the garbage collector. If this is the case, the mutex is created and its handle is
assigned to hMutex. At the same time, the function shutdown( ) is registered as a termination function by calling atexit(
).
.

It is important to note that in the multithreaded garbage collector, shutdown( ) serves two purposes. First, as in the
original version of GCPtr, shutdown( ) frees any unused memory that has not been released because of a circular
reference. Second, when a program using the multithreaded garbage collector ends, it stops the garbage collection
thread. This means that there might still be dynamically allocated objects that haven’t been freed. This is important
because these objects might have destructors that need to be called. Because shutdown( ) releases all remaining
objects, it also releases these objects.

Next, the mutex is acquired by calling WaitForSingleObject( ). This is necessary to prevent two threads from accessing
gclist at the same time. Once the mutex has been acquired, a search of gclist is made, looking for any preexisting
entry that matches the address in t . If one is found, its reference count is incremented. If no preexising entry matches
t , a new GCInfo object is created that contains this address, and this object is added to gclist. Then, addr, arraySize,
and isArray are set. These actions are the same as in the original version of GCPtr.

Next, instCount is incremented. Recall that instCount is initialized to zero. Incrementing it each time an object is
created keeps track of how many GCPtr objects are in existence. As long as this count is above zero, the garbage
collector will continue to execute.

Next, if hThrd is zero (as it is initially), then no thread has yet been created for the garbage collector. In this case,
_beginthreadex( ) is called to begin the thread. A handle to the thread is then assigned to hThrd . The thread entry

籍
function is called gc( ), and it is examined shortly.

Finally, the mutex is released and the constructor returns. It is important to point out that each call to

书
WaitForSingleObject( ) must be balanced by a call to ReleaseMutex( ), as shown in the GCPtr constructor. Failure to
release the mutex will cause deadlock.

典 书
The TimeOutExc Exception

经 读
As you probably noticed in the code for GCPtr( ) described in the preceding section, if the mutex cannot be acquired
after 10 seconds, then a TimeOutExc is thrown. Frankly, 10 seconds is a very long time, so a time-out shouldn’t ever

+
happen unless something disrupts the task scheduler of the operating system. However, in the event it does occur,

不
your application code may want to catch this exception. The TimeOutExc class is shown here:

+
// Exception thrown when a time-out occurs

C
// when waiting for access to hMutex.

典
//
class TimeOutExc {
// Add functionality if needed by your application.

经
};

Notice that it contains no members. Its existence as a unique type is sufficient for the purposes of this chapter. Of

无
course, you can add functionality if desired.

The Multithreaded GCPtr Destructor

Unlike the single-threaded version of the GCPtr destructor, the multithreaded version of ~GCPtr( ) does not call collect(
) . Instead, it simply decrements the reference count of the memory pointed to by the GCPtr that is going out of scope.
The actual collection of garbage (if any exists) is handled by the garbage collection thread. The destructor also
decrements the instance counter, instCount.

The multithreaded version of ~GCPtr( ) is shown here:

// Destructor for GCPtr.
template <class T, int size>
GCPtr<T, size>::~GCPtr() {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

list<GCInfo<T> >::iterator p;

p = findPtrInfo(addr);
if(p->refcount) p->refcount--; // decrement ref count
.

// Decrement instance counter for each object

// that is destroyed.
instCount--;

ReleaseMutex(hMutex);
}

The gc( ) Function

The entry function for the garbage collector is called gc( ), and it is shown here:
// Entry point for garbage collector thread.
template <class T, int size>
unsigned __stdcall GCPtr<T, size>::gc(void * param) {
#ifdef DISPLAY
cout << "Garbage collection started.\n";
#endif

while(isRunning()) {

籍
collect();
}

书
collect(); // collect garbage on way out

// Release and reset the thread handle so

典 书
// that the garbage collection thread can
// be restarted if necessary.
CloseHandle(hThrd);

经 读
hThrd = 0;

+ 不
#ifdef DISPLAY

+
cout << "Garbage collection terminated for "
<< typeid(T).name() << "\n";

C
#endif

}
return 0;

经 典
The gc( ) function is quite simple: it runs as long as the garbage collector is in use. The isRunning( ) function returns
true if instCount is greater than zero (which means that the garbage collector is still needed) and false otherwise.

无
Inside the loop, collect( ) is called continuously. This approach is suitable for demonstrating the multithreaded garbage
collector, but it is probably too inefficient for real-world use. You might want to experiment with calling collect( ) less
often, such as only when memory runs low. You could also experiment by calling the Windows API function Sleep( )
after each call to collect( ). Sleep( ) pauses the execution of the calling thread for a specified number of milliseconds.
While sleeping, a thread does not consume CPU time.

When isRunning( ) returns false, the loop ends, causing gc( ) to eventually end, which stops the garbage collection
thread. Because of the multithreading, it is possible that there will still be an entry on gclist that has not yet been freed
even though isRunning( ) returns false. To handle this case, a final call to collect( ) is made before gc( ) ends.

Finally, the thread handle is released via a call to the Windows API function CloseHandle( ), and its value is set to zero.
Setting hThrd to zero enables the GCPtr constructor to restart the thread if later in the program new GCPtr objects are
created.

The isRunning( ) Function

The isRunning( ) function is shown here:

// Returns true if the collector is still in use.
static bool isRunning() { return instCount > 0; }
.

It simply compares instCount to zero. As long as instCount is greater than 0, at least one GCPtr pointer is still in
existence and the garbage collector is still needed.

Synchronizing Access to gclist

Many of the functions in GCPtr access gclist, which holds the garbage collection list. Access to gclist must be
synchronized to prevent two or more threads from attempting to use it at the same time. The reason for this is easy to
understand. If access were not synchronized, then, for example, one thread might be obtaining an iterator to the end
of the list at the same time that another thread is adding or deleting an element from the list. In this case, the iterator
would be invalid. To prevent such problems, each sequence of code that accesses gclist must be guarded by a
mutex. The copy constructor for GCPtr shown here is one example:
// Copy constructor.
GCPtr(const GCPtr &ob) {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

list<GCInfo<T> >::iterator p;

p = findPtrInfo(ob.addr);

籍
p->refcount++; // increment ref count

书
addr = ob.addr;
arraySize = ob.arraySize;
if(arraySize > 0) isArray = true;

典 书
else isArray = false;

instCount++; // increase instance count for copy

+ 经 不 读
ReleaseMutex(hMutex);

+
Notice that the first thing that the copy constructor does is acquire the mutex. Then it creates a copy of the object and
adjusts the reference count for the memory being pointed to. On its way out, the copy constructor releases the mutex.

C 典
This same basic method is applied to all functions that access gclist.

Two Other Changes

无 经
There are two other changes that you must make to the original version of the garbage collector. First, recall that the
original version of GCPtr defined a static variable called first that indicated when the first GCPtr was created. This
variable is no longer needed because hMutex now performs this function. Thus, remove first from GCPtr. Because it is
a static variable, you will also need to remove its definition outside of GCPtr.

In the original, single-threaded version of the garbage collector, if you defined the DISPLAY macro, you could watch
the garbage collector in action. Most of that code has been removed in the multithreaded version because
multithreading causes the output to be scrambled and unintelligible in most cases. For the multithreaded version,
defining DISPLAY simply lets you know when the garbage collector has started and when it has stopped.

The Entire Multithreaded Garbage Collector

The entire multithreaded version of the garbage collector is shown here. Call this file gcthrd.h.
// A garbage collector that runs as a background task.

#include <iostream>
#include <list>
#include <typeinfo>
#include <cstdlib>
#include <windows.h>
#include <process.h>
.

using namespace std;

// To watch the action of the garbage collector, define DISPLAY.

// #define DISPLAY

// Exception thrown when an attempt is made to

// use an Iter that exceeds the range of the
// underlying object.
//
class OutOfRangeExc {
// Add functionality if needed by your application.
};

// Exception thrown when a time-out occurs

// when waiting for access to hMutex.
//
class TimeOutExc {
// Add functionality if needed by your application.
};

籍
// An iterator-like class for cycling through arrays
// that are pointed to by GCPtrs. Iter pointers
// ** do not ** participate in or affect garbage

书
// collection. Thus, an Iter pointing to
// some object does not prevent that object

典 书
// from being recycled.
//
template <class T> class Iter {

经 读
T *ptr; // current pointer value
T *end; // points to element one past end

+
T *begin; // points to start of allocated array

不
unsigned length; // length of sequence

+
public:

C 典
Iter() {
ptr = end = begin = NULL;
length = 0;

经
}

Iter(T *p, T *first, T *last) {

无
ptr = p;
end = last;
begin = first;
length = last - first;
}

// Return length of sequence to which this

// Iter points.
unsigned size() { return length; }

// Return value pointed to by ptr.

// Do not allow out-of-bounds access.
T &operator*() {
if( (ptr >= end) || (ptr < begin) )
throw OutOfRangeExc();
return *ptr;
}

// Return address contained in ptr.

// Do not allow out-of-bounds access.
T *operator->() {
if( (ptr >= end) || (ptr < begin) )
.

throw OutOfRangeExc();
return ptr;
}

// Prefix ++.
Iter operator++() {
ptr++;
return *this;
}

// Prefix --.
Iter operator--() {
ptr--;
return *this;
}

// Postfix ++.
Iter operator++(int notused) {
T *tmp = ptr;

籍
ptr++;
return Iter<T>(tmp, begin, end);
}

// Postfix --.

典 书 书
Iter operator--(int notused) {
T *tmp = ptr;

经 读
ptr--;
return Iter<T>(tmp, begin, end);

+
}

+ 不
// Return a reference to the object at the
// specified index. Do not allow out-of-bounds

C 典
// access.
T &operator[](int i) {
if( (i < 0) || (i >= (end-begin)) )

经
throw OutOfRangeExc();
return ptr[i];
}

} 无
// Define the relational operators.
bool operator==(Iter op2) {
return ptr == op2.ptr;

bool operator!=(Iter op2) {

return ptr != op2.ptr;
}

bool operator<(Iter op2) {

return ptr < op2.ptr;
}

bool operator<=(Iter op2) {

return ptr <= op2.ptr;
}

bool operator>(Iter op2) {

return ptr > op2.ptr;
}
.

bool operator>=(Iter op2) {

return ptr >= op2.ptr;
}

// Subtract an integer from an Iter.

Iter operator-(int n) {
ptr -= n;
return *this;
}

// Add an integer to an Iter.

Iter operator+(int n) {
ptr += n;
return *this;
}

// Return number of elements between two Iters.

int operator-(Iter<T> &itr2) {
return ptr - itr2.ptr;
}

籍
};

书
// This class defines an element that is stored
// in the garbage collection information list.

典 书
//
template <class T> class GCInfo {
public:

经 读
unsigned refcount; // current reference count

+
T *memPtr; // pointer to allocated memory

+ 不
/* isArray is true if memPtr points
to an allocated array. It is false

C 典
otherwise. */
bool isArray; // true if pointing to array

经
/* If memPtr is pointing to an allocated
array, then arraySize contains its size */
unsigned arraySize; // size of array

无
// Here, mPtr points to the allocated memory.
// If this is an array, then size specifies
// the size of the array.
GCInfo(T *mPtr, unsigned size=0) {
refcount = 1;
memPtr = mPtr;
if(size != 0)
isArray = true;
else
isArray = false;

arraySize = size;
}
};

// Overloading operator== allows GCInfos to be compared.

// This is needed by the STL list class.
template <class T> bool operator==(const GCInfo<T> &ob1,
const GCInfo<T> &ob2) {
return (ob1.memPtr == ob2.memPtr);
}
.

// GCPtr implements a pointer type that uses

// garbage collection to release unused memory.
// A GCPtr must only be used to point to memory
// that was dynamically allocated using new.
// When used to refer to an allocated array,
// specify the array size.
//
template <class T, int size=0> class GCPtr {

// gclist maintains the garbage collection list.

static list<GCInfo<T> > gclist;

// addr points to the allocated memory to which

// this GCPtr pointer currently points.
T *addr;

/* isArray is true if this GCPtr points

to an allocated array. It is false

籍
otherwise. */
bool isArray; // true if pointing to array

书
// If this GCPtr is pointing to an allocated
// array, then arraySize contains its size.

典 书
unsigned arraySize; // size of the array

// These support multithreading.

经 读
unsigned tid; // thread id
static HANDLE hThrd; // thread handle

+
static HANDLE hMutex; // handle of mutex

+ 不
static int instCount; // counter of GCPtr objects

C 典
// Return an iterator to pointer info in gclist.
typename list<GCInfo<T> >::iterator findPtrInfo(T *ptr);

经
public:

// Define an iterator type for GCPtr<T>.

无
typedef Iter<T> GCiterator;

// Construct both initialized and uninitialized objects.

GCPtr(T *t=NULL) {

// When first object is created, create the mutex

// and register shutdown().
if(hMutex==0) {
hMutex = CreateMutex(NULL, 0, NULL);
atexit(shutdown);
}

if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

list<GCInfo<T> >::iterator p;

p = findPtrInfo(t);

// If t is already in gclist, then

// increment its reference count.
// Otherwise, add it to the list.
.

if(p != gclist.end())
p->refcount++; // increment ref count
else {
// Create and store this entry.
GCInfo<T> gcObj(t, size);
gclist.push_front(gcObj);
}

addr = t;
arraySize = size;
if(size > 0) isArray = true;
else isArray = false;

// Increment instance counter for each new object.

instCount++;

// If the garbage collection thread is not

// currently running, start it running.
if(hThrd==0) {
hThrd = (HANDLE) _beginthreadex(NULL, 0, gc,

籍
(void *) 0, 0, (unsigned *) &tid);

// For some applications, it will be better

书
// to lower the priority of the garbage collector
// as shown here:

典 书
//
// SetThreadPriority(hThrd,
// THREAD_PRIORITY_BELOW_NORMAL);

经 读
}

+
ReleaseMutex(hMutex);

不
}

+
// Copy constructor.

C 典
GCPtr(const GCPtr &ob) {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

经
list<GCInfo<T> >::iterator p;

无
p = findPtrInfo(ob.addr);
p->refcount++; // increment ref count

addr = ob.addr;
arraySize = ob.arraySize;
if(arraySize > 0) isArray = true;
else isArray = false;

instCount++; // increase instance count for copy

ReleaseMutex(hMutex);
}

// Destructor for GCPTr.

~GCPtr();

// Collect garbage. Returns true if at least

// one object was freed.
static bool collect();

// Overload assignment of pointer to GCPtr.

T *operator=(T *t);
 // Overload assignment of GCPtr to GCPtr.
GCPtr &operator=(GCPtr &rv);

// Return a reference to the object pointed

// to by this GCPtr.
T &operator*() {
return *addr;
}

// Return the address being pointed to.

T *operator->() { return addr; }

// Return a reference to the object at the

// index specified by i.
T &operator[](int i) {
return addr[i];
}

// Conversion function to T *.

籍
operator T *() { return addr; }

// Return an Iter to the start of the allocated memory.

书
Iter<T> begin() {
int size;

典 书
if(isArray) size = arraySize;
else size = 1;

+ 经 不 读
return Iter<T>(addr, addr, addr + size);

+
// Return an Iter to one past the end of an allocated array.
Iter<T> end() {

C 典
int size;

if(isArray) size = arraySize;

经
else size = 1;

return Iter<T>(addr + size, addr, addr + size);

无
}

// Return the size of gclist for this type

// of GCPtr.
static int gclistSize() {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

unsigned sz = gclist.size();

ReleaseMutex(hMutex);
return sz;
}

// A utility function that displays gclist.

static void showlist();

// The following functions support multithreading.

//
// Returns true if the collector is still in use.
static bool isRunning() { return instCount > 0; }
. s
.

// Clear gclist when program exits.

static void shutdown();

// Entry point for garbage collector thread.

static unsigned __stdcall gc(void * param);
};

// Create storage for the static variables.

template <class T, int size>
list<GCInfo<T> > GCPtr<T, size>::gclist;

template <class T, int size>

int GCPtr<T, size>::instCount = 0;

template <class T, int size>

HANDLE GCPtr<T, size>::hMutex = 0;

template <class T, int size>

HANDLE GCPtr<T, size>::hThrd = 0;

籍
// Destructor for GCPtr.
template <class T, int size>
GCPtr<T, size>::~GCPtr() {

书
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

典 书
list<GCInfo<T> >::iterator p;

经 读
p = findPtrInfo(addr);
if(p->refcount) p->refcount--; // decrement ref count

+ + 不
// Decrement instance counter for each object
// that is destroyed.
instCount--;

C 典
ReleaseMutex(hMutex);

经
// Collect garbage. Returns true if at least
// one object was freed.

无
template <class T, int size>
bool GCPtr<T, size>::collect() {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

bool memfreed = false;

list<GCInfo<T> >::iterator p;
do {

// Scan gclist looking for unreferenced pointers.

for(p = gclist.begin(); p != gclist.end(); p++) {
// If in-use, skip.
if(p->refcount > 0) continue;

memfreed = true;

// Remove unused entry from gclist.

gclist.remove(*p);

// Free memory unless the GCPtr is null.

if(p->memPtr) {
.

if(p->isArray) {
delete[] p->memPtr; // delete array
}
else {
delete p->memPtr; // delete single element
}
}

// Restart the search.

break;
}

} while(p != gclist.end());

ReleaseMutex(hMutex);

return memfreed;
}

// Overload assignment of pointer to GCPtr.

籍
template <class T, int size>
T * GCPtr<T, size>::operator=(T *t) {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)

书
throw TimeOutExc();

典 书
list<GCInfo<T> >::iterator p;

// First, decrement the reference count

经 读
// for the memory currently being pointed to.
p = findPtrInfo(addr);

+
p->refcount--;

+ 不
// Next, if the new address is already
// existent in the system, increment its

C 典
// count. Otherwise, create a new entry
// for gclist.
p = findPtrInfo(t);

经
if(p != gclist.end())
p->refcount++;
else {

无
// Create and store this entry.
GCInfo<T> gcObj(t, size);
gclist.push_front(gcObj);
}

addr = t; // store the address.

ReleaseMutex(hMutex);

return t;
}

// Overload assignment of GCPtr to GCPtr.

template <class T, int size>
GCPtr<T, size> & GCPtr<T, size>::operator=(GCPtr &rv) {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

list<GCInfo<T> >::iterator p;

// First, decrement the reference count

// for the memory currently being pointed to.
.

p = findPtrInfo(addr);
p->refcount--;

// Next, increment the reference count

// of the new object.
p = findPtrInfo(rv.addr);
p->refcount++; // increment ref count

addr = rv.addr;// store the address.

ReleaseMutex(hMutex);

return rv;
}

// A utility function that displays gclist.

template <class T, int size>
void GCPtr<T, size>::showlist() {
if(WaitForSingleObject(hMutex, 10000)==WAIT_TIMEOUT)
throw TimeOutExc();

书 籍
list<GCInfo<T> >::iterator p;

cout << "gclist<" << typeid(T).name() << ", "

<< size << ">:\n";

典 书
cout << "memPtr refcount value\n";

if(gclist.begin() == gclist.end()) {

经 读
cout << " -- Empty --\n\n";
return;

+
}

+ 不
for(p = gclist.begin(); p != gclist.end(); p++) {
cout << "[" << (void *)p->memPtr << "]"

C 典
<< " " << p->refcount << " ";
if(p->memPtr) cout << " " << *p->memPtr;
else cout << " ---";

经
cout << endl;
}
cout << endl;

无
ReleaseMutex(hMutex);

// Find a pointer in gclist.

template <class T, int size>
typename list<GCInfo<T> >::iterator
GCPtr<T, size>::findPtrInfo(T *ptr) {

list<GCInfo<T> >::iterator p;

// Find ptr in gclist.

for(p = gclist.begin(); p != gclist.end(); p++)
if(p->memPtr == ptr)
return p;

return p;
}

// Entry point for garbage collector thread.

template <class T, int size>
unsigned __stdcall GCPtr<T, size>::gc(void * param) {
 #ifdef DISPLAY
cout << "Garbage collection started.\n";
#endif

while(isRunning()) {
collect();
}

collect(); // collect garbage on way out

// Release and reset the thread handle so

// that the garbage collection thread can
// be restarted if necessary.
CloseHandle(hThrd);
hThrd = 0;

#ifdef DISPLAY
cout << "Garbage collection terminated for "
<< typeid(T).name() << "\n";
#endif

}
return 0;

书 籍
// Clear gclist when program exits.

典 书
template <class T, int size>
void GCPtr<T, size>::shutdown() {

经 读
if(gclistSize() == 0) return; // list is empty

+
list<GCInfo<T> >::iterator p;

+ 不
#ifdef DISPLAY
cout << "Before collecting for shutdown() for "

C 典
<< typeid(T).name() << "\n";
#endif

经
for(p = gclist.begin(); p != gclist.end(); p++) {
// Set all remaining reference counts to zero.
p->refcount = 0;

无
}

collect();

#ifdef DISPLAY
cout << "After collecting for shutdown() for "
<< typeid(T).name() << "\n";
#endif
}

Using the Multithreaded Garbage Collector

To use the multithreaded garbage collector, include gcthrd.h in your program. Then, use GCPtr in the same way as
described in Chapter 2. When you compile the program, you must remember to link in the multithreaded libraries, as
explained earlier in this chapter in the section describing _beginthreadex( ) and _endthreadex( ).

To see the effects of the multithreaded garbage collector, try this version of the load test program originally shown in
Chapter 2:
// Demonstrate the multithreaded garbage collector.
#include <iostream>
. s
#include <new>
.

#include "gcthrd.h"

using namespace std;

// A simple class for load testing GCPtr.

class LoadTest {
int a, b;
public:
double n[100000]; // just to take-up memory
double val;

LoadTest() { a = b = 0; }

LoadTest(int x, int y) {
a = x;
b = y;
val = 0.0;
}

friend ostream &operator<<(ostream &strm, LoadTest &obj);

籍
};

// Create an inserter for LoadTest.

书
ostream &operator<<(ostream &strm, LoadTest &obj) {
strm << "(" << obj.a << " " << obj.b << ")";

典 书
return strm;
}

经 读
int main() {
GCPtr<LoadTest> mp;

+
int i;

+ 不
for(i = 1; i < 2000; i++) {
try {

C 典
mp = new LoadTest(i, i);
if(!(i%100))
cout << "gclist contains " << mp.gclistSize()

经
<< " entries.\n";
} catch(bad_alloc xa) {
// For most users, this exception won't

无
// ever occur.
cout << "Last object: " << *mp << endl;
cout << "Length of gclist: "
<< mp.gclistSize() << endl;
}
}

return 0;
}

Here is a sample run. (Of course, your output may vary.) This output was produced with the display option turned on
by defining DISPLAY within gcthrd.h.
Garbage collection started.
gclist contains 42 entries.
gclist contains 35 entries.
gclist contains 29 entries.
gclist contains 22 entries.
gclist contains 18 entries.
gclist contains 11 entries.
gclist contains 4 entries.
gclist contains 51 entries.
gclist contains 47 entries.
gclist contains 40 entries.
gclist contains 33 entries.
gclist contains 26 entries.
gclist contains 19 entries.
gclist contains 15 entries.
gclist contains 10 entries.
gclist contains 3 entries.
gclist contains 53 entries.
gclist contains 46 entries.
gclist contains 42 entries.
Before collecting for shutdown() for class LoadTest
After collecting for shutdown() for class LoadTest

As you can see, because collect( ) is running in the background, gclist never gets very large, even though thousands
of objects are being allocated and abandoned.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Some Things to Try
Creating successful multithreaded programs can be quite challenging. One reason for this is the fact that
multithreading requires that you think of programs in parallel rather than linear terms. Furthermore, at runtime, threads
interact in ways that are often difficult to anticipate. Thus, you might be surprised (or even bewildered) by the actions
of a multithreaded program. The best way to get good at multithreading is to play with it. Toward this end, here are
some ideas that you might want to try.

Try adding another list box to the thread control panel that lets the user adjust the priority class of the thread in
addition to its priority value. Try adding various synchronization objects to the control panel that can be turned on or off
under user control. This will let you experiment with different synchronization options.

For the multithreaded garbage collector, try collecting garbage less often, such as when gclist reaches a certain size
or after free memory drops to a predetermined point. Alternatively, you could use a waitable timer to activate garbage
collection on a regular basis. Finally, you might want to experiment with the garbage collector’s priority class and
settings to find which level is optimal for your use.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

Chapter 4: Extending C++

For the typical C++ programmer, programming is not just a job, it’s a way of life. We programmers appreciate not only
the raw power of the language, but also its nuances and subtleties. Much like a connoisseur savors a vintage wine, we
enjoy the finer points of the language. Of course, our interest in C++ is not confined to simply using it to create
programs. Instead, we find ourselves attracted to the language itself and to the issues surrounding the design and
development of languages in general.

Because of the interest that most C++ programmers have in computer languages, it is a rare C++ programmer who
has not daydreamed about adding a new feature to the language. (How often have you thought, “How great it would
be if C++ had ...”?) The trouble is that most programmers don’t have access to the source code for a full-blown C++
compiler to which they can add their experimental constructs. Fortunately, there is an easy way to experiment with
your own extensions to C++: by creating a translator that converts your experimental ideas into their equivalent C++
code. Such a translator is the topic of this chapter.

籍
Why Use a Translator?

书
You might question why a translator is needed to experiment with a new language feature when there appears to be
an obvious alternative: the standard C++ preprocessor. As all readers know, the preprocessor supports macro
substitutions, in which one text sequence is substituted for another. For years, preprocessor macros have been used

典 书
to add experimental features to C++. For example, one construct commonly implemented by macros is the repeat/
until loop. A repeat/until loop is similar to C++’s do/while loop except that a repeat/until runs until some condition

经 读
becomes true. (Thus, a repeat/until runs while its conditional expression is false.) This differs from the do/while loop,
which runs until some condition becomes false. Because of the similarities between the repeat/until and the do/while, it

+
is possible to use macros to implement a repeat/until, as the following fragment shows:

不
#define repeat do

+
#define until(exp) while(!(exp))

C
// ...

典
int i=0;
repeat {
cout << "i: " << i << endl;

经
i++;
} until(i==10);

无
The macros cause do to be substituted for repeat, and while to be substituted for until. Also, the conditional expression
is reversed in the while. Thus, the repeat/until loop in the preceding fragment is converted by the preprocessor into this
do/while loop:

do {
cout << "i: " << i << endl;
i++;
} while(!(i==10));

As this example shows, the use of macros to implement a repeat/until construct is both an easy and reasonable
approach.

Given the long-standing tradition of using macros to implement experimental features, why would one want to build a
translator? That is, why not just use macros like those used for the repeat/until? The answer is that macro
substitutions cannot be used to add all types of experimental features. Furthermore, even in cases in which a macro
substitution will work, not all are as elegant or conceptually pure as those used by the repeat/until example. Although
programmers have managed some very impressive feats with macros, in many cases they have been achieved by
use of complex and hard-to-understand #define directives. Such “macro tricks” often serve to destructure code, are
hard to verify for correctness, and often lack resiliency. Frankly, many C++ programmers (including this author) avoid
the use of complicated macros for just these reasons.

Fortunately, there is an alternative to complicated macros that can be used to experiment with a wide variety of new
C++ features: you can build a translator that handles, under program control, the conversion of an experimental
construct into its equivalent C++ code. Using such a “pre-preprocessor” enables you to add features to C++ that are
either difficult or impossible to implement using preprocessor macros. Such a translator is also an interesting project in
its own right, and the basic framework can be readily adapted for other purposes.

It is important to understand that although the translator developed in this chapter is capable of performing
sophisticated text substitutions, it still has its limitations. Because the translator implements a single-pass algorithm
that reads the source file only once, it can be used to experiment with only those constructs that can be handled by
single-pass substitution. (You can change this behavior, if you want.) Aside from performing sophisticated text
substitutions, the translator is completely ignorant about the contents of the program. That is, it knows nothing about
the type of a variable, what an operator means, or even what the previously read token was. Thus, the translator is not
a parser for the C++ language. It is simply an engine for specialized text substitutions; in essence, it is a specialized
preprocessor. Despite these limitations, you can use it to experiment with a wide array of ideas.

Note If you’re interested in a parser for C++, seeChapter 9.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

The Experimental Keywords

Before developing the translator, it is necessary to define what translations it performs. The translator developed in this
chapter handles the following experimental extensions to C++:

A foreach loop

A cases statement

A typeof operator

A repeat/until loop

With the exception of the repeat/until loop, which is included simply for illustration, the other keywords use a syntax
that would be difficult (or impossible) to translate using macro substitutions. Each of these experimental keywords is
described here.

籍
The foreach Loop

书
Contemporary language development has embraced the foreach loop. For example, foreach is part of C# and a “for
each” style loop is being added to Java. Furthermore, the “for each” concept is already incorporated into one of the
newest parts of C++ because the STL (Standard Template Library) defines the for_each( ) algorithm, which applies a

典 书
function to each element in a container. However, C++ does not currently define a general-purpose foreach loop.

The foreach was invented to handle a situation that occurs quite often in programming: the need to cycle through the

经 读
elements of an array (or other type of collection of objects) in strictly sequential fashion, from beginning to end. For
example, consider this fragment that computes the average of the values in an array:

+ 不
int n[] = { 1, 7, 3, 11, 5 };

+
double avg=0.0;

C
for(int i=0; i < 5; i++)

典
avg += n[i];

经
cout << "Average: " << avg/5 << endl;

To compute the average of the values in n, each element in the array is read in order, from start to finish. Of course,

无
this is just one instance of a general concept. The same type of loop is used to find the minimum or maximum value in
an array, to sum the array’s values, to compute the least common denominator, or for a hundred other uses.
Moreover, the same type of loop construct is used whenever the contents of an array need to be accessed. The
foreach loop was created to simplify and streamline such loops.

The value of the foreach is that it eliminates the need to manually index the array. Instead, the foreach automatically
cycles through the entire array, obtaining one element at a time, in sequence. For example, here is the preceding
fragment rewritten using a foreach:
int n[] = { 1, 7, 3, 11, 5 };
double avg=0.0;

foreach(int x in n)
avg += x;

cout << "Average: " << avg/5 << endl;

With each pass through the loop, x is automatically given a value equal to the next element in n. Thus, on the first
iteration x contains 1, on the second iteration x contains 7, and so on. Not only is the syntax streamlined, it also
prevents counting errors.

The syntax for the foreach loop is shown here:

.

foreach(type varname in arrayname)

Here, type is the type of the loop variable, its name is specified by varname, and the array being accessed is specified
by arrayname. Furthermore, varname is local to the loop and not known outside of it. (This syntax is borrowed from
C#.) Remember, with each iteration of the loop, varname will contain (in sequential order) the value of the next
element of the specified array.

The translator converts a foreach loop into its equivalent C++ for loop.

The cases Statement

The cases statement allows you to specify a range of values that will be matched by a switch expression. Normally,
when you want the same code sequence to be used by two or more case statements, you must use “stacked” case
statements, as the following example shows:
switch(i) {
case 1:
case 2:
case 3:
case 4:

籍
// do something for cases 1 to 4
break;
case 5:

书
// do something else
break;
// ...

典 书
}

Stacked case statements, such as those for 1 through 4 in the example, are common in programming. They can also

经 读
be tedious to code. Wouldn’t it be convenient to be able to specify a range of values within a single case? The

+
translator enables you to do precisely that!

+ 不
The translator implements a cases statement that lets you define a range of values that are all handled by a single
code sequence. For example, the preceding code can be rewritten as shown here using the cases statement:

C 典
switch(i) {
cases 1 to 4:
// do something for cases 1 to 4

经
break;
case 5:
// do something else

无
break;
// ...
}

Thus, when i is between 1 and 4 inclusive, that value will be matched by the cases statement.

The syntax for cases is shown here:

cases start to end:

Here, start is the first value to match, and end is the last value to match.

The translator converts a cases statement into series of stacked case statements.

The typeof Operator

Runtime type ID has become an important part of most modern programming. Although C++’s built-in support for it is
superb, we programmers are a restless lot and sometimes try to improve on a good thing! The typeof experimental
operator is a case in point. It simply provides an alternative syntax for an operation already supported by C++: the
comparison of two types. Thus, typeof adds no new functionality, but it offers a different perspective on the process.

Normally, the typeid operator is used when a type comparison is required. For example, consider the following
statement, which determines if the object pointed to by ptr1 is of the same type as the object pointed to by ptr2:
if(typeid(*ptr1) == typeid(*ptr2))
cout << "ptr1 points to same type as ptr2\n";

In this statement, the typeid operator obtains the type of the objects pointed to by ptr1 and ptr2. If these two types are
the same, the if statement succeeds. A statement such as this might be used when working with polymorphic classes,
for example. In this case, the type of an object being pointed to by a base class pointer can’t always be known at
compile time, and a runtime check is needed.

Although there is nothing wrong with typeid or the preceding statement, the following approach offers an intriguing
alternative syntax:
if(typeof *ptr1 same as *ptr2)
cout << "ptr1 points to same type as ptr2\n";

In this statement, the typeof operator is used to compare two types. The result is true if the two types are the same
and false otherwise. Although it performs the same operation as the first version, it changes the way in which the
operation is expressed, which lets you see it from a different point of view. It also demonstrates the range of ideas that
the translator lets you experiment with.

籍
The syntax for typeof is shown here:
typeof op1 same as op2

书
Here, op1 and op2 specify either a type identifier (such as int or MyClass) or an object. Thus, typeof can be used to
compare the types of two objects, the type of an object to a known type, or two types.

典 书
The translator converts typeof into its corresponding typeid expression.

经 读
The repeat/until Loop

+
As explained earlier, it is quite easy to use preprocessor macros to experiment with a repeat/until loop. The translator

不
implements repeat/until simply for illustration and because it serves as a model for any other type of loop with which

+
you might want to experiment.

C 典
无 经
.

A Translator for Experimental C++ Features

The entire code for the translator is shown here. To follow along, call this file trans.cpp.
// A translator for experimental C++ extensions.
#include <iostream>
#include <fstream>
#include <cctype>
#include <cstring>
#include <string>

using namespace std;

// Prototypes for the functions that handle

// the extended keywords.
void foreach();
void cases();

籍
void repeat();
void until();
void typeof();

书
// Prototypes for tokenizing the input file.

书
bool gettoken(string &tok);

典
void skipspaces();

经 读
// Indentation padding string.
string indent = "";

+ 不
// The input and output file streams.

+
ifstream fin;
ofstream fout;

C 典
// Exception class for syntax errors.
class SyntaxExc {

经
string what;
public:
SyntaxExc(char *e) { what = string(e); }

无
string geterror() { return what; }
};

int main(int argc, char *argv[]) {

string token;

if(argc != 3) {
cout << "Usage: ep <input file> <output file>\n";
return 1;
}

fin.open(argv[1]);

if(!fin) {
cout << "Cannot open " << argv[1] << endl;
return 1;
}

fout.open(argv[2]);

if(!fout) {
cout << "Cannot open " << argv[2] << endl;
.

return 1;
}

// Write header.
fout << "// Translated from an .exp source file.\n";

try {
// Main translation loop.
while(gettoken(token)) {

// Skip over // comments.

if(token == "//") {
do {
fout << token;
gettoken(token);
} while(token.find('\n') == string::npos);
fout << token;
}

// Skip over /* comments.

籍
else if(token == "/*") {
do {
fout << token;

书
gettoken(token);
} while(token != "*/");

典 书
fout << token;
}

经 读
// Skip over quoted string.
else if(token == "\"") {

+
do {

不
fout << token;

+
gettoken(token);
} while(token != "\"");

C 典
fout << token;
}

经
else if(token == "foreach") foreach();

else if(token == "cases") cases();

无
else if(token == "repeat") repeat();

else if(token == "until") until();

else if(token == "typeof") typeof();

else fout << token;

}
} catch(SyntaxExc exc) {
cout << exc.geterror() << endl;
return 1;
}

return 0;
}

// Get the next token from the input stream.

bool gettoken(string &tok) {
char ch;
char ch2;
static bool trackIndent = true;
.

tok = "";

ch = fin.get();

// Check for EOF and return false if EOF

// is found.
if(!fin) return false;

// Read whitespace.
if(isspace(ch)) {
while(isspace(ch)) {
tok += ch;

// Reset indent counter with each new line.

if(ch == '\n') {
indent = "";
trackIndent = true;
}
else if(trackIndent) indent += ch;

书 籍
ch = fin.get();

fin.putback(ch);
return true;

典 书
}

// Stop tracking indentation after encountering

经 读
// first non-whitespace character on a line.
trackIndent = false;

+ + 不
// Read an identifier or keyword.
if(isalpha(ch) || ch=='_') {
while(isalpha(ch) || isdigit(ch) || ch=='_') {

C 典
tok += ch;
ch = fin.get();
}

经
fin.putback(ch);
return true;
}

无
// Read a number.
if(isdigit(ch)) {
while(isdigit(ch) || ch=='.' ||
tolower(ch) == 'e' ||
ch == '-' || ch =='+') {
tok += ch;
ch = fin.get();
}
fin.putback(ch);
return true;
}

// Check for \"

if(ch == '\\') {
ch2 = fin.get();
if(ch2 == '"') {
tok += ch;
tok += ch2;
ch = fin.get();
} else
fin.putback(ch2);
.

// Check for '"'

if(ch == '\'') {
ch2 = fin.get();
if(ch2 == '"') {
tok += ch;
tok += ch2;
return true;
} else
fin.putback(ch2);
}

// Check for begin comment symbols.

if(ch == '/') {
tok += ch;
ch = fin.get();
if(ch == '/' || ch == '*') {
tok += ch;
}

籍
else fin.putback(ch);
return true;
}

典 书
// Check for end comment symbols.
if(ch == '*') {
tok += ch;

书
ch = fin.get();

经 读
if(ch == '/') {
tok += ch;

+
}

不
else fin.putback(ch);

+
return true;
}

C
tok += ch;

典
经
return true;
}

无
// Translate a foreach loop.
void foreach() {
string token;
string varname;
string arrayname;

char forvarname[5] = "_i";

static char counter[2] = "a";

// Create loop control variable for generated

// for loop.
strcat(forvarname, counter);
counter[0]++;

// Only 26 foreach loops in a file because

// generated loop control variables limited to
// _ia to _iz. This can be changed if desired.
if(counter[0] > 'z')
throw SyntaxExc("Too many foreach loops.");

fout << "int " << forvarname

<< " = 0;\n";
.

// Write beginning of generated for loop.

fout << indent << "for(";

skipspaces();

// Read the (
gettoken(token);
if(token[0] != '(')
throw SyntaxExc("( expected in foreach.");

skipspaces();

// Get the type of the foreach variable.

gettoken(token);
fout << token << " ";

skipspaces();

// Read and save the foreach variable's name.

籍
gettoken(token);
varname = token;

书
skipspaces();

典 书
// Read the "in"
gettoken(token);
if(token != "in")

经 读
throw SyntaxExc("in expected in foreach.");

+
skipspaces();

+ 不
// Read the array name.
gettoken(token);

C 典
arrayname = token;

fout << varname << " = " << arrayname << "[0];\n";

经
// Construct target value.
fout << indent + " " << forvarname << " < "

无
<< "((sizeof " << token << ")/"
<< "(sizeof " << token << "[0]));\n";

fout << indent + " " << forvarname << "++, "
<< varname << " = " << arrayname << "["
<< forvarname << "])";

skipspaces();

// Read the )
gettoken(token);
if(token[0] != ')')
throw SyntaxExc(") expected in foreach.");
}

// Translate a cases statement.

void cases() {
string token;
int start, end;

skipspaces();
.

// Get starting value.

gettoken(token);

if(isdigit(token[0])) {
// is an int constant
start = atoi(token.c_str());
}
else if(token[0] == '\'') {
// is char constant
gettoken(token);

start = (int) token[0];

// discard closing '

gettoken(token);
if(token[0] != '\'')
throw SyntaxExc("' expected in cases.");
}
else
throw SyntaxExc("Constant expected in cases.");

skipspaces();

书 籍
// Read and discard the "to".
gettoken(token);

典 书
if(token != "to")
throw SyntaxExc("to expected in cases.");

经 读
skipspaces();

+
// Get ending value.

不
gettoken(token);

} +
if(isdigit(token[0])) {

C 典
// is an int constant
end = atoi(token.c_str());

经
else if(token[0] == '\'') {
// is char constant
gettoken(token);

无
end = (int) token[0];

// discard closing '

gettoken(token);
if(token[0] != '\'')
throw SyntaxExc("' expected in cases.");
}
else
throw SyntaxExc("Constant expected in cases.");

skipspaces();

// Read and discard the :

gettoken(token);

if(token != ":")
throw SyntaxExc(": expected in cases.");

// Generate stacked case statments.

fout << "case " << start << ":\n";
for(int i = start+1 ; i <= end; i++) {
.

fout << indent << "case " << i << ":";

if(i != end) fout << endl;
}
}

// Translate a repeat loop.

void repeat() {
fout << "do";
}

// Translate an until.
void until() {
string token;
int parencount = 1;

fout << "while";

skipspaces();

// Read and store the (

籍
gettoken(token);
if(token != "(")
throw SyntaxExc("( expected in typeof.");

书
fout << "(";

典 书
// Begin while by reversing and
// parenthesizing the condition.
fout << "!(";

do {

+ 经 不 读
// Now, read the expression.

if(!gettoken(token))

+
throw SyntaxExc("Unexpected EOF encountered.");

C 典
if(token == "(") parencount++;
if(token == ")") parencount--;

经
fout << token;
} while(parencount > 0);
fout << ")";

无
}

// Translate a typeof expression.

void typeof() {
string token;
string temp;

fout << "typeid(";

skipspaces();

gettoken(token);

do {
temp = token;

if(!gettoken(token))
throw SyntaxExc("Unexpected EOF encountered.");

if(token != "same") fout << temp;

} while(token != "same");
 skipspaces();

gettoken(token);

if(token != "as") throw SyntaxExc("as expected.");

fout << ") == typeid(";

skipspaces();

do {
if(!gettoken(token))
throw SyntaxExc("Unexpected EOF encountered.");

fout << token;

} while(token != ")");
fout << ")";
}

void skipspaces() {

籍
char ch;

do {

书
ch = fin.get();
} while(isspace(ch));

典 书
fin.putback(ch);
}

+ 经 不 读
C+ 典
无 经
 Using the Translator
To use the translator, first create a file containing a program that uses the experimental keywords. To keep things
straight, use the .exp file extension for files that use the experimental constructs. Understand that an exp file will
usually contain mostly standard C++ code. It’s just that it also contains one or more experimental features that will be
converted into C++ by the translator. For example, here is an exp file that demonstrates the foreach loop. Notice that
most of the program consists of normal C++ code.
// Demonstrate the foreach loop.
#include <iostream>
using namespace std;

int main() {
int nums[] = { 1, 6, 19, 4, -10, 88 };
int min;

// Find the minimum value.

籍
min = nums[0];
foreach(int x in nums)

书
if(min > x) min = x;

cout << "Minimum is " << min << endl;

}
return 0;

经 典 读 书
To convert an exp file into a standard cpp file, run it through the translator. For example, assuming that this file is

+
called foreach.exp, the following command line will translate it into pure C++ code that can be compiled by a C++

不
compiler:

+
trans foreach.exp foreach.cpp

C 典
After trans runs, foreach.cpp will contain the following C++ program:
// Translated from an .exp source file.
// Demonstrate the foreach loop.

经
#include <iostream>
using namespace std;

无
int main() {
int nums[] = { 1, 6, 19, 4, -10, 88 };
int min;

// Find the minimum value.

min = nums[0];
int _ia = 0;
for(int x = nums[0];
_ia < ((sizeof nums)/(sizeof nums[0]));
_ia++, x = nums[_ia])
if(min > x) min = x;

cout << "Minimum is " << min << endl;

return 0;
}

The remainder of this chapter describes how the translation works.

.

How the Translator Works

The translator operates in a straightforward manner. It simply reads the input file and writes it to the output file. In the
process, when it finds an experimental keyword, it substitutes the equivalent C++ code. This inherent simplicity is why
the translator is useful for experimentation. It takes no great effort to implement an extension and then try it. The
following sections describe the action of each part of the translator in detail.

The Global Declarations

The translator begins by defining the following global variables and class:
// Indentation padding string.
string indent = "";

// The input and output file streams.

ifstream fin;

籍
ofstream fout;

// Exception class for syntax errors.

书
class SyntaxExc {
string what;
public:

典 书
SyntaxExc(char *e) { what = string(e); }
string geterror() { return what; }
};

+ 经 读
The current whitespace sequence used to indent a line is stored in indent. This string is used to add the proper
amount of indentation when multiple lines of code are substituted for an experimental construct.

不
+
The input file stream is stored in fin; the output file stream is held in fout. When the program begins, the filenames
specified on the command line are linked to fin and fout.

C 典
Syntax errors encountered during the translation of an experimental construct are reported by throwing a SyntaxExc
object. As it is written, SyntaxExc contains only a string that describes the error, but you can add functionality, if

经
desired.

The main( ) Function

无
The main( ) function performs two duties. First, it opens the input and output files specified on the command line. This
will be familiar code to all C++ programmers. Second, it runs the main loop of the translator, which handles the
translation of the experimental keywords. The main loop is shown here:
try {
// Main translation loop.
while(gettoken(token)) {

// Skip over // comments.

if(token == "//") {
do {
fout << token;
gettoken(token);
} while(token.find('\n') == string::npos);
fout << token;
}

// Skip over /* comments.

else if(token == "/*") {
do {
fout << token;
.

gettoken(token);
} while(token != "*/");
fout << token;
}

// Skip over quoted string.

else if(token == "\"") {
do {
fout << token;
gettoken(token);
} while(token != "\"");
fout << token;
}

else if(token == "foreach") foreach();

else if(token == "cases") cases();

else if(token == "repeat") repeat();

籍
else if(token == "until") until();

else if(token == "typeof") typeof();

典 书
else fout << token;

书
} catch(SyntaxExc exc) {
cout << exc.geterror() << endl;

经 读
return 1;
}

+ 不
With each iteration of the loop, a token is read from the input file. If the token does not require translation, it is written

+
to the output file. However, if the token contains one of the experimental keywords, then the proper function is called to
translate that keyword into its corresponding C++ code. Notice that the loop skips over comments and quoted strings,

C 典
copying them to the output file without testing their contents for the experimental keywords. This is necessary to
prevent the translation of a keyword that is contained in a comment or quoted string.

经
If a syntax error is detected in the translation of an experimental feature, then a SyntaxExc object is thrown by the
code doing the translating, and it is caught by the catch in main( ), which simply displays the error. You can enhance
the error report to include a line number or other information if you desire.

无
The gettoken( ) and skipspaces( ) Functions

In order for the translator to convert an experimental keyword into its equivalent C++ code, it must be able to know
when a keyword has been found. To enable this, the input file must be processed on a token by token basis. Here, the
term token is used in a very loose sense. It simply means a piece of text. The translator does not require finely grained
tokenization. It simply needs to recognize identifiers (including keywords), numbers, and whitespace. Most other
characters can be processed as single characters.

The gettoken( ) function is shown here:

// Get the next token from the input stream.
bool gettoken(string &tok) {
char ch;
char ch2;
static bool trackIndent = true;

tok = "";

ch = fin.get();

// Check for EOF and return false if EOF

// is found.
if(!fin) return false;

// Read whitespace.
if(isspace(ch)) {
while(isspace(ch)) {
tok += ch;

// Reset indent counter with each new line.

if(ch == '\n') {
indent = "";
trackIndent = true;
}
else if(trackIndent) indent += ch;

ch = fin.get();
}
fin.putback(ch);
return true;
}

籍
// Stop tracking indentation after encountering
// first non-whitespace character on a line.

书
trackIndent = false;

典 书
// Read an identifier or keyword.
if(isalpha(ch) || ch=='_') {
while(isalpha(ch) || isdigit(ch) || ch=='_') {

经 读
tok += ch;
ch = fin.get();

+
}

不
fin.putback(ch);

+
return true;
}

C 典
// Read a number.
if(isdigit(ch)) {

经
while(isdigit(ch) || ch=='.' ||
tolower(ch) == 'e' ||
ch == '-' || ch =='+') {

无
tok += ch;
ch = fin.get();
}
fin.putback(ch);
return true;
}

// Check for \"

if(ch == '\\') {
ch2 = fin.get();
if(ch2 == '"') {
tok += ch;
tok += ch2;
ch = fin.get();
} else
fin.putback(ch2);
}

// Check for '"'

if(ch == '\'') {
ch2 = fin.get();
if(ch2 == '"') {
.

tok += ch;
tok += ch2;
return true;
} else
fin.putback(ch2);
}

// Check for begin comment symbols.

if(ch == '/') {
tok += ch;
ch = fin.get();
if(ch == '/' || ch == '*') {
tok += ch;
}
else fin.putback(ch);
return true;
}

// Check for end comment symbols.

if(ch == '*') {

籍
tok += ch;
ch = fin.get();
if(ch == '/') {

书
tok += ch;
}

典 书
else fin.putback(ch);
return true;
}

tok += ch;

+
return true;
经 不 读
+
}

C
The gettoken( ) function has one parameter, a string called tok, that is passed a reference to a string object. This

典
object will contain the token when the function returns. The function returns true if a token was read and false if the
end of the file is encountered.

经
The gettoken( ) function begins by setting tok to the null string. It then reads the next character from fin and stores it in
ch. If this read encounters the end of the file, then false is returned. Otherwise, ch is tested against a number of

无
possibilities.

First, if ch is a whitespace, then a loop is entered that reads whitespace characters until the first nonwhitespace
character is read. The whitespace characters are appended to tok. The nonwhitespace character is returned to the
input stream with a call to fin.putback( ). At the end of the loop, tok contains all of the whitespace characters that were
read, and this is the token returned to the calling routine.

There is one other thing to notice in the whitespace loop. If a newline is read, indent is set to the null string and the
trackIndent variable is set to true. When trackIndent is true, whitespace will be stored in indent. After the whitespace
loop, trackIndent is set to false. Thus, only the leading whitespace on a line is stored in indent.

If ch is not a whitespace character, other possibilities are tested. If the next token is an identifier or keyword, it begins
with a letter or underscore and is read by the next loop in gettoken( ). Otherwise, if ch is a digit, then a number is read.

If ch is not a whitespace, letter, underscore, or digit, then five special conditions are checked for. The first is the \”
sequence. As explained, the translator performs no translations of the text in quoted strings. When the translator finds
an opening quote, it simply copies all text between it and the closing quote. However, it is necessary to prevent an
embedded quote from being mistaken as the beginning or end of a quoted string. Thus, the backslash sequence \”
must be handled as a pair. The same situation applies when a character constant specifies a quote. The comment
symbols must also be handled as pairs of characters, so gettoken( ) checks for //, /*, and */.

Occasionally during translation it is necessary to discard the spaces present in an experimental statement because
.

they have no relationship to the C++ code being produced. The skipspaces( ) function shown here accomplishes this.
It simply reads and discards spaces.
void skipspaces() {
char ch;

do {
ch = fin.get();
} while(isspace(ch));
fin.putback(ch);
}

Translating the foreach Loop

The most challenging translation is the foreach loop. The function that handles this is foreach( ), and it is shown here:
// Translate a foreach loop.
void foreach() {
string token;
string varname;
string arrayname;

籍
char forvarname[5] = "_i";
static char counter[2] = "a";

书
// Create loop control variable for generated

典 书
// for loop.
strcat(forvarname, counter);
counter[0]++;

+ 经 读
// Only 26 foreach loops in a file because
// generated loop control variables limited to

不
// _ia to _iz. This can be changed if desired.

+
if(counter[0] > 'z')
throw SyntaxExc("Too many foreach loops.");

C 典
fout << "int " << forvarname
<< " = 0;\n";

经
// Write beginning of generated for loop.
fout << indent << "for(";

无
skipspaces();

// Read the (
gettoken(token);
if(token[0] != '(')
throw SyntaxExc("( expected in foreach.");

skipspaces();

// Get the type of the foreach variable.

gettoken(token);
fout << token << " ";

skipspaces();

// Read and save the foreach variable's name.

gettoken(token);
varname = token;

skipspaces();
.

// Read the "in"

gettoken(token);
if(token != "in")
throw SyntaxExc("in expected in foreach.");

skipspaces();

// Read the array name.

gettoken(token);
arrayname = token;

fout << varname << " = " << arrayname << "[0];\n";

// Construct target value.

fout << indent + " " << forvarname << " < "
<< "((sizeof " << token << ")/"
<< "(sizeof " << token << "[0]));\n";

fout << indent + " " << forvarname << "++, "

籍
<< varname << " = " << arrayname << "["
<< forvarname << "])";

书
skipspaces();

典 书
// Read the )
gettoken(token);
if(token[0] != ')')

经 读
throw SyntaxExc(") expected in foreach.");
}

+ 不
The translator converts a foreach loop into its equivalent for loop. This involves handling two rather complicated

+
issues. To understand them, consider the following example:

C
double nums[] = { 1.1, 2.2, 3.3 };

典
double sum = 0.0;

foreach(double v in nums)

经
sum += v;

The foreach loop is translated into the following C++ code:

无
int _ia = 0;
for(double v = nums[0];
_ia < ((sizeof nums)/(sizeof nums[0]));
_ia++, v = nums[_ia])
sum += v;

First, notice that the for loop requires a loop control variable, called _ia in this case, but no such variable is present in
the foreach statement. Remember, the variable declared in the foreach, which is v in this case, receives the values of
the elements of the array. It is not a loop counter. This means that a loop control variable must be created. This
variable cannot be declared inside the for because this declaration needs to be reserved for the declaration of the
foreach variable, v, which must be local to the loop. Thus, it is necessary to declare the loop control variable outside
the for loop. This requires the generation of a unique integer variable name that won’t conflict with any other variables
used within its scope.

Second, the for must cycle through all elements in the array (nums, in this case), but the size of the array is not part of
the foreach specification. This means that the number of elements in the array must be computed. Fortunately, this can
be accomplished through the use of sizeof , as described shortly.

Now, let’s go through each step in the translation process. The foreach translation begins by constructing a (hopefully)
unique variable that will be used as the for loop control variable. This variable’s name begins with _i , to which a letter
between a and z is added. The first generated variable in a file is called _ia, the second _ib, and so on. Up to 26 such
.

variables are allowed. This isn’t the best way to generate nonconflicting variable names, but it has the advantage of
simplicity and is sufficient for the purposes of experimenting with a foreach loop. This variable is given a type of int and
its declaration is written to the output file.

Next, the for loop is begun by writing for( to the output file. Then the initialization portion of the for is constructed. First,
the type and name of the foreach variable is read and copied to the output file. Next, the keyword in is read and
discarded. The array name is then read and used to construct an initialization that sets the foreach variable to the
value of the first element in the array.

Next, the conditional portion of the for is generated. In this section, the loop control variable is tested against a value
that represents the number of elements in the array. This value is computed by dividing the size of the array by the
size of an individual element in the array. This division is necessary because sizeof returns the size in bytes of an
array, not the number of elements in the array.

Finally, the iteration portion of the for is created. It increments the loop control variable and sets the foreach variable to
the value of the next element in the array. Before returning, foreach( ) confirms that the foreach statement ends with a
).

One last point: In C#, the foreach can be used with arrays and collections (which are much like STL containers). For
simplicity, this version will work only with arrays. However, you can try adapting it so that it can be used to cycle

籍
through STL containers.

书
Translating the cases Statement

As explained earlier, the cases statement provides a single-statement alternative to specifying a series of stacked case

典 书
statements. It is translated by the cases( ) function, shown here:
// Translate a cases statement.

经 读
void cases() {
string token;

+
int start, end;

+ 不
skipspaces();

C 典
// Get starting value.
gettoken(token);

经
if(isdigit(token[0])) {
// is an int constant
start = atoi(token.c_str());

无
}
else if(token[0] == '\'') {
// is char constant
gettoken(token);

start = (int) token[0];

// discard closing '

gettoken(token);
if(token[0] != '\'')
throw SyntaxExc("' expected in cases.");
}
else
throw SyntaxExc("Constant expected in cases.");

skipspaces();

// Read and discard the "to".

gettoken(token);
if(token != "to")
throw SyntaxExc("to expected in cases.");
 skipspaces();

// Get ending value.

gettoken(token);

if(isdigit(token[0])) {
// is an int constant
end = atoi(token.c_str());
}
else if(token[0] == '\'') {
// is char constant
gettoken(token);

end = (int) token[0];

// discard closing '

gettoken(token);
if(token[0] != '\'')
throw SyntaxExc("' expected in cases.");

籍
}
else
throw SyntaxExc("Constant expected in cases.");

skipspaces();

典 书 书
// Read and discard the :
gettoken(token);

+ 经 读
if(token != ":")
throw SyntaxExc(": expected in cases.");

不
+
// Generate stacked case statements.
fout << "case " << start << ":\n";

C 典
for(int i = start+1 ; i <= end; i++) {
fout << indent << "case " << i << ":";
if(i != end) fout << endl;

经
}
}

无
Although it looks like a lot of code, it’s actually a simple translation. Recall that a cases statement lets you specify a
range of values that will match a switch expression. The translator simply converts that range into a series of stacked
case statements. For example, this switch statement:

switch(val) {
cases 0 to 3:
cout << "val is 0, 1, 2, or 3\n";
break;
case 4:
cout << "val is 4\n";
}

is translated into this equivalent C++ code:

switch(val) {
case 0:
case 1:
case 2:
case 3:
cout << "val is 0, 1, 2, or 3\n";
break;
case 4:
cout << "val is 4\n";
.

As you can see, individual case statements representing the range of values specified by the cases statement have
been produced.

The cases( ) function works like this. First, the beginning value in the range is read. This value will be either an integer
or character constant. If it is an integer constant, it will begin with a digit. If it is a character constant, it will begin with a
single quote. Either way, this value is stored as an integer value in start. Next, the to is read and discarded. Then the
ending value is read. It too might be either an integer or character constant. Its value is stored in the integer variable
end. Finally, a series of stacked case statements are output, one for each value in the range.

Translating the typeof Operator

The typeof operator supports an alternative syntax for comparing two types. As explained earlier, its general form is
typeof op1 same as op2

Here, op1 and op2 specify either a type identifier (such as int or MyClass) or an object. Thus, typeof can be used to
compare the types of two objects, the type of an object to a known type, or two types. Its outcome is true if the types
are the same and false otherwise. The typeof operator is converted into a C++ expression that compares the results of

籍
applying the typeid operator to each operand.

The translation of typeof is handled by the typeof( ) function, shown here:

书
// Translate a typeof expression.
void typeof() {

典 书
string token;
string temp;

经 读
fout << "typeid(";

+
skipspaces();

+ 不
gettoken(token);

C 典
do {
temp = token;

经
if(!gettoken(token))
throw SyntaxExc("Unexpected EOF encountered.");

无
if(token != "same") fout << temp;
} while(token != "same");

skipspaces();

gettoken(token);

if(token != "as") throw SyntaxExc("as expected.");

fout << ") == typeid(";

skipspaces();

do {
if(!gettoken(token))
throw SyntaxExc("Unexpected EOF encountered.");

fout << token;

} while(token != ")");
fout << ")";
}
.

The operation of typeof( ) is easy to follow. It reads the first operand and outputs a typeid expression containing that
type. It then reads and discards the same and as keywords. Then it reads the second operand and outputs the
corresponding typeid expression. Thus, this typeof statement:
if(typeof obj same as Myclass)
cout << "obj is a MyClass object.\n";

is translated into this C++ code:

if(typeid(obj) == typeid(Myclass))
cout << "obj is a MyClass object.\n";

Translating a repeat/until Loop

As mentioned earlier, the translator handles a repeat/until loop mostly for the sake of illustration, since it is easier to
use macros to handle this construct. However, having the translator perform this conversion does save you the effort
of defining repeat and until macros in every program that uses them. It also serves as a model that can be adapted for
other types of loops.

The repeat keyword is translated by the repeat( ) method, shown here, which simply substitutes do:

籍
// Translate a repeat loop.
void repeat() {
fout << "do";

书
}

The until keyword is handled by until( ), which is shown next:

典 书
// Translate an until.
void until() {
string token;

经 读
int parencount = 1;

+ 不
fout << "while";

+
skipspaces();

C 典
// Read and store the (
gettoken(token);
if(token != "(")

经
throw SyntaxExc("( expected in typeof.");
fout << "(";

无
// Begin while by reversing and
// parenthesizing the condition.
fout << "!(";

// Now, read the expression.

do {
if(!gettoken(token))
throw SyntaxExc("Unexpected EOF encountered.");

if(token == "(") parencount++;

if(token == ")") parencount--;

fout << token;

} while(parencount > 0);
fout << ")";
}

The until( ) function substitutes the keyword while for until and then reverses the conditional expression. (Remember, a
repeat/until loop runs until something becomes true. A do/while loop runs until something becomes false.)

Although conceptually simple, reversing the conditional expression requires a bit of work on the part of the translator.
The reason is that it can’t simply add a ! to the start of the expression. It must also parenthesize the expression that
the ! precedes. To understand why, consider this repeat/until loop:
int i=0;

repeat {
cout << i << " ";
i++;
} until (i==10);

It is translated into the following do/while:

int i=0;

do {
cout << i << " ";
i++;
} while(!(i==10));

The parentheses in the conditional expression of the while ensure that the loop will run as long as i does not equal 10.
However, if the parentheses were removed, as shown here:

籍
} while(!i==10);

the loop would run while !i was equal to 10, which is a completely different condition!

典 书
To parenthesize the until expression, the translator must add an opening parenthesis to the start of the expression
and a closing parenthesis to the end of it. The question is, how does the translator know when the conditional

书
expression has ended? It determines this by counting closing parentheses. Recall that the expression in the until is
contained within a pair of parentheses. The parenthesis count is stored in the parencount variable, which is initialized

经 读
to 1 (which corresponds to the opening parenthesis of the until expression). When copying the conditional expression,
the translator increments parencount each time an opening parenthesis is found and decrements it each time a closing

+
parenthesis is encountered. Thus, when parencount drops to zero, the end of the expression has been reached, and

不
the final closing parenthesis can be added.

C+ 典
无 经
.

A Demonstration Program
The following program demonstrates the experimental features supported by the translator:
// Demonstrate all of the experimental features
// supported by the translator.
#include <iostream>
using namespace std;

// Create a polymorphic base class.

class A {
public:
virtual void f() { };
};

// And a concrete subclass.

class B: public A {

籍
public:
void f() {}
};

int main() {

书 书
int n[] = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 };

典
double dn[] = {1.1, 2.2, 3.3, 4.4 };

经 读
cout << "Using a foreach loop.\n";

+
/* Keywords, such as foreach or typeof

不
are ignored when inside comments

+
or quoted strings. */

C
// A foreach loop.

典
foreach(int x in n )
cout << x << ' ';

经
cout << "\n\n";

无
cout << "Using nested foreach loops.\n";

// A foreach loop with a block.

foreach(double f in dn) {
cout << f << ' ';
cout << f*f << ' ';

// A nested foreach loop.

foreach(double f in dn)
cout << f/3 << " ";

cout << endl;

}

cout << endl;

cout << "Demonstrate cases statement.\n";

cout << "A cases statement that uses integer constants:\n";

// Demonstrate cases statement that uses

// integer constants.
.

for(int i=0; i < 12; i++)

switch(i) {
case 0:
cout << "case 0\n";
break;
cases 1 to 6:
cout << "cases 1 to 6\n";
break;
case 7:
cout << "case 7\n";
break;
cases 8 to 10:
cout << "cases 8 to 10\n";
break;
default:
cout << "case 11\n";
}

cout << "\n";

籍
cout << "A cases statement that uses character constants:\n";

// Demonstrate a cases statement that uses

书
// character constants.
for(char ch='a'; ch <= 'e'; ch++)

典 书
switch(ch) {
case 'a':
cout << "case a\n";

经 读
break;
cases 'b' to 'd':

+
cout << "cases b to d\n";

不
break;

+
case 'e':
cout << "case e\n";

C 典
}

cout << endl;

经
cout << "A repeat/until loop.\n";

无
// Demonstrate a repeat/until loop.
int k = 0;
repeat {
k++;
cout << "k: " << k << " ";
} until(k==10);

cout << "\n\n";

cout << "Use typeof.\n";

// Demonstrate typeof.
A *aPtr;
B *bPtr, bObj;

// Assign a base pointer to derived object.

aPtr = &bObj;
bPtr = &bObj;

if(typeof *aPtr same as *bPtr)

cout << "aPtr points to same type of object as bPtr\n";
.

if(typeof *aPtr same as B)

cout << "aPtr points to B object\n";

return 0;
}

When this program is run through the translator, the following C++ code is produced:
// Translated from an .exp source file.
// Demonstrate all of the experimental features
// supported by the translator.
#include <iostream>
using namespace std;

// Create a polymorphic base class.

class A {
public:
virtual void f() { };
};

// And a concrete subclass.

籍
class B: public A {
public:

书
void f() {}
};

典 书
int main() {
int n[] = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 };
double dn[] = {1.1, 2.2, 3.3, 4.4 };

+ 经 不 读
cout << "Using a foreach loop.\n";

/* Keywords, such as foreach or typeof

+
are ignored when inside comments
or quoted strings. */

C 典
// A foreach loop.
int _ia = 0;

经
for(int x = n[0];
_ia < ((sizeof n)/(sizeof n[0]));
_ia++, x = n[_ia])

无
cout << x << ' ';

cout << "\n\n";

cout << "Using nested foreach loops.\n";

// A foreach loop with a block.

int _ib = 0;
for(double f = dn[0];
_ib < ((sizeof dn)/(sizeof dn[0]));
_ib++, f = dn[_ib]) {
cout << f << ' ';
cout << f*f << ' ';

// A nested foreach loop.

int _ic = 0;
for(double f = dn[0];
_ic < ((sizeof dn)/(sizeof dn[0]));
_ic++, f = dn[_ic])
cout << f/3 << " ";
.

cout << endl;

}

cout << endl;

cout << "Demonstrate cases statement.\n";

cout << "A cases statement that uses integer constants:\n";

// Demonstrate cases statement that uses

// integer constants.
for(int i=0; i < 12; i++)
switch(i) {
case 0:
cout << "case 0\n";
break;
case 1:
case 2:
case 3:
case 4:

籍
case 5:
case 6:
cout << "cases 1 to 6\n";

书
break;
case 7:

典 书
cout << "case 7\n";
break;
case 8:

经 读
case 9:
case 10:

+
cout << "cases 8 to 10\n";

不
break;

+
default:
cout << "case 11\n";

C 典
}

cout << "\n";

经
cout << "A cases statement that uses character constants:\n";

无
// Demonstrate a cases statement that uses
// character constants.
for(char ch='a'; ch <= 'e'; ch++)
switch(ch) {
case 'a':
cout << "case a\n";
break;
case 98:
case 99:
case 100:
cout << "cases b to d\n";
break;
case 'e':
cout << "case e\n";
}

cout << endl;

cout << "A repeat/until loop.\n";

// Demonstrate a repeat/until loop.

int k = 0;
.

do {
k++;
cout << "k: " << k << " ";
} while(!(k==10));

cout << "\n\n";

cout << "Use typeof.\n";

// Demonstrate typeof.
A *aPtr;
B *bPtr, bObj;

// Assign a base pointer to derived object.

aPtr = &bObj;
bPtr = &bObj;

if(typeid(*aPtr) == typeid(*bPtr))
cout << "aPtr points to same type of object as bPtr\n";

籍
if(typeid(*aPtr) == typeid(B))
cout << "aPtr points to B object\n";

书
return 0;
}

典 书
The C++ version can be compiled by any modern C++ compiler and produces the following output:
Using a foreach loop.

经 读
1 2 3 4 5 6 7 8 9 10

+
Using nested foreach loops.

不
1.1 1.21 0.366667 0.733333 1.1 1.46667

+
2.2 4.84 0.366667 0.733333 1.1 1.46667
3.3 10.89 0.366667 0.733333 1.1 1.46667

C 典
4.4 19.36 0.366667 0.733333 1.1 1.46667

Demonstrate cases statement.

经
A cases statement that uses integer constants:
case 0
cases 1 to 6

无
cases 1 to 6
cases 1 to 6
cases 1 to 6
cases 1 to 6
cases 1 to 6
case 7
cases 8 to 10
cases 8 to 10
cases 8 to 10
case 11

A cases statement that uses character constants:

case a
cases b to d
cases b to d
cases b to d
case e

A repeat/until loop.
k: 1 k: 2 k: 3 k: 4 k: 5 k: 6 k: 7 k: 8 k: 9 k: 10
Use typeof.
aPtr points to same type of object as bPtr
aPtr points to B object

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Some Things to Try
As explained at the start of this chapter, the purpose of the translator is to enable you to experiment with your own
ideas for new language features or intriguing alternatives to existing features. To add your own experimental keyword
to the translator, first create a function that handles the translation. Then, inside the main loop, add another else if
statement that calls that function when the keyword is encountered.

The constructs with which you can experiment are limited only by your imagination. Here are some ideas to get you
started:

A breakon statement, which breaks from a loop when some condition is true. It might use a syntax like
this:
breakon(x == 99);

A breakto statement, which transfers control out of a loop or switch to the specified label. It might use
a syntax like this:

籍
breakto jmp12;

An ignore statement, which causes the early iteration of a loop when a certain value

书
is encountered. Thus, it streamlines an if...continue statement. It might use a syntax like this:
ignore(n == 12);

典 书
Of course, not all experimental features will turn out to be good ones. That’s what makes them experimental—and
enjoyable to play with!

经 读
One last point: If you are interested in designing your own language, you will find Chapter 9 especially valuable

+
because it develops an interpreter for a small subset of C++. You can use this interpreter as the starting point for

不
language features more advanced than can be easily handled by the translator found in this chapter.

C+ 典
无 经
 Chapter 5: An Internet File Downloader

Overview
The Internet has irreversibly altered the course of programming. Prior to the rise of the Internet, most applications
executed in isolation on individual machines, or made use of small, local area networks. The Internet changed this.
Today, most computers are Internet-enabled, and many programs make use of the vast resources of the World Wide
Web. For the modern programmer, incorporating Internet functionality into an application is no longer an option, it is a
necessity.

Despite the importance of the Internet, C++ does not provide built-in support for it. This is because C++ came to
maturity before the Internet reached critical mass in the late 1990s. However, this is not a hindrance. In fact, it’s
actually a benefit. Instead of dictating a single approach that all programmers must use, C++ lets you access the
Internet capabilities of the operating system. Not only does this offer the potential for greater efficiency and flexibility by

籍
letting you choose the best way to interface to the Internet, it also lets you write an Internet-enabled program in a way
that is most compatible with the underlying execution environment. Therefore, if you want to create high-performance,
Internet-enabled code, C++ is your first choice.

典 书
To illustrate the ease with which C++ handles the Internet, this chapter develops a file download subsystem that can
be integrated into a number of different Internet-based applications. The file downloader retrieves a file from the

书
Internet, given that file’s URL. It is a self-contained subsystem that handles all the details of opening an Internet
connection, reading the file, and then closing the connection. The file downloader has one special feature: it is

经 读
restartable. For example, if in the middle of a long download, the connection to the Internet is lost, the download can
be resumed at the point at which it was interrupted when the connection is reestablished. This is especially useful

+
when downloading very long files over a slow modem connection!

+ 不
After developing the download subsystem, two applications that use it are created. The first is a very simple,
console-based application that simply demonstrates the downloader. The second is a GUI application that you can use

C 典
to download files.

Because Internet support is provided by the operating system, it is necessary to choose one. The one operating

经
system that is available to virtually all readers of this book is Windows, and it is used in this chapter. However, the
basic techniques can be translated into other environments.

无
Note This chapter assumes that the reader has a general understanding of the Internet and a working knowledge of
Windows programming. It is far beyond the scope of this book to teach either.
 The WinINet Library
To access the Internet, this chapter takes advantage of an easy-to-use library provided by Windows. This library is
called Windows Internet, or WinINet for short. The WinINet API offers a rich assortment of high-level functions that
handle the various protocols, such as HTTP and FTP, in a consistent, stable way. Windows handles the low-level
details for you. (This is somewhat similar to the way that <fstream> presents a consistent interface for file operations
while handling the details for you.) As you will see, if you follow a few rules, it is easy to add Internet access to any
Windows application.

Although WinINet is a large library, you need to use only a few of its functions:

InternetAttemptConnect Checks if an Internet connection is available.

InternetOpen Opens an Internet connection and returns a handle to it.

InternetOpenUrl Opens a URL and returns a handle to it.

籍
HttpQueryInfo Obtains information from the last response header.

书
InternetReadFile Reads bytes from an open URL.

InternetCloseHandle Closes an Internet handle.

经 典 读 书
Each of these will be described in detail during the discussion of the code for the file downloader.

To use WinINet, you must include wininet.h in your program and link wininet.lib with your application.

+ + 不
C 典
无 经
.

The File Downloader Subsystem

The entire code for the file downloader subsystem is shown here. Put this code in a file called dl.cpp.
// A file download subsystem.
#include <iostream>
#include <windows.h>
#include <wininet.h>
#include <fstream>
#include <cstdio>

using namespace std;

const int MAX_ERRMSG_SIZE = 80;

const int MAX_FILENAME_SIZE = 512;
const int BUF_SIZE = 1024;

书 籍
// Exception class for download errors.
class DLExc {
char err[MAX_ERRMSG_SIZE];
public:

典 书
DLExc(char *exc) {
if(strlen(exc) < MAX_ERRMSG_SIZE)
strcpy(err, exc);

经 读
}

+
// Return a pointer to the error message.

不
const char * geterr() {

+
return err;
}

C 典
};

// A class for downloading files from the Internet.

经
class Download {
static bool ishttp(char *url);
static bool httpverOK(HINTERNET hIurl);

无
static bool getfname(char *url, char *fname);
static unsigned long openfile(char *url, bool reload,
ofstream &fout);
public:
static bool download(char *url, bool restart=false,
void (*update)(unsigned long, unsigned long)=NULL);
};

// Download a file.
//
// Pass the URL of the file to url.
//
// To reload a file, pass true to reload.
//
// To specify an update function that is called after
// each buffer is read, pass a pointer to that
// function as the third parameter. If no update
// function is desired, then let the third parameter
// default to null.
bool Download::download(char *url, bool reload,
void (*update)(unsigned long, unsigned long)) {
.

ofstream fout; // output stream

unsigned char buf[BUF_SIZE]; // input buffer
unsigned long numrcved; // number of bytes read
unsigned long filelen; // length of file on disk
HINTERNET hIurl, hInet; // Internet handles
unsigned long contentlen;// length of content
unsigned long len; // length of contentlen
unsigned long total = 0; // running total of bytes received
char header[80]; // holds Range header

try {
if(!ishttp(url))
throw DLExc("Must be HTTP url.");

// Open the file specified by url.

// The open stream will be returned
// in fout. If reload is true, then
// any preexisting file will be truncated.
// The length of any preexisting file (after

籍
// possible truncation) is returned.
filelen = openfile(url, reload, fout);

书
// See if Internet connection available.
if(InternetAttemptConnect(0) != ERROR_SUCCESS)

典 书
throw DLExc("Can't connect.");

// Open Internet connection.

经 读
hInet = InternetOpen("downloader",
INTERNET_OPEN_TYPE_DIRECT,

+
NULL, NULL, 0);

+ 不
if(hInet == NULL)
throw DLExc("Can't open connection.");

C 典
// Construct header requesting range of data.
sprintf(header, "Range:bytes=%d-", filelen);

经
// Open the URL and request range.
hIurl = InternetOpenUrl(hInet, url,

无
header, -1,
INTERNET_FLAG_NO_CACHE_WRITE, 0);

if(hIurl == NULL) throw DLExc("Can't open url.");

// Confirm that HTTP/1.1 or greater is supported.

if(!httpverOK(hIurl))
throw DLExc("HTTP/1.1 not supported.");

// Get content length.

len = sizeof contentlen;
if(!HttpQueryInfo(hIurl,
HTTP_QUERY_CONTENT_LENGTH |
HTTP_QUERY_FLAG_NUMBER,
&contentlen, &len, NULL))
throw DLExc("File or content length not found.");

// If existing file (if any) is not complete,

// then finish downloading.
if(filelen != contentlen && contentlen)
do {
// Read a buffer of info.
.

if(!InternetReadFile(hIurl, &buf,
BUF_SIZE, &numrcved))
throw DLExc("Error occurred during download.");

// Write buffer to disk.

fout.write((const char *) buf, numrcved);
if(!fout.good())
throw DLExc("Error writing file.");

total += numrcved; // update running total

// Call update function, if specified.

if(update && numrcved > 0)
update(contentlen+filelen, total+filelen);

} while(numrcved > 0);

else
if(update)
update(filelen, filelen);

籍
} catch(DLExc) {
fout.close();
InternetCloseHandle(hIurl);

书
InternetCloseHandle(hInet);

典 书
throw; // rethrow the exception for use by caller
}

经 读
fout.close();
InternetCloseHandle(hIurl);

+
InternetCloseHandle(hInet);

+
return true;

C 不
典
// Return true if HTTP version of 1.1 or greater.
bool Download::httpverOK(HINTERNET hIurl) {

经
char str[80];
unsigned long len = 79;

无
// Get HTTP version.
if(!HttpQueryInfo(hIurl, HTTP_QUERY_VERSION, &str, &len, NULL))
return false;

// First, check major version number.

char *p = strchr(str, '/');
p++;
if(*p == '0') return false; // can't use HTTP 0.x

// Now, find start of minor HTTP version number.

p = strchr(str, '.');
p++;

// Convert to int.
int minorVerNum = atoi(p);

if(minorVerNum > 0) return true;

return false;
}

// Extract the filename from the URL. Return false if

// the filename cannot be found.
.

bool Download::getfname(char *url, char *fname) {

// Find last /.
char *p = strrchr(url, '/');

// Copy filename after the last /.

if(p && (strlen(p) < MAX_FILENAME_SIZE)) {
p++;
strcpy(fname, p);
return true;
}
else
return false;
}

// Open the output file, initialize the output

// stream, and return the file's length. If
// reload is true, first truncate any preexisting
// file.
unsigned long Download::openfile(char *url,
bool reload,

籍
ofstream &fout) {
char fname[MAX_FILENAME_SIZE];

书
if(!getfname(url, fname))
throw DLExc("File name error.");

典 书
if(!reload)
fout.open(fname, ios::binary | ios::out |

经 读
ios::app | ios::ate);
else

+
fout.open(fname, ios::binary | ios::out |

不
ios::trunc);

+
if(!fout)

C 典
throw DLExc("Can't open output file.");

// Get current file length.

经
return fout.tellp();
}

无
// Confirm that the URL specifies HTTP.
bool Download::ishttp(char *url) {
char str[5] = "";

// Get first four characters from URL.

strncpy(str, url, 4);

// Convert to lowercase
for(char *p=str; *p; p++) *p = tolower(*p);

return !strcmp("http", str);

}

Notice that this file defines two classes. The first is DLExc, which is a class that encapsulates exceptions thrown by the
downloader. The DLExc constructor is passed a pointer to a string that describes the exception, and it stores that
string. A pointer to the error message is obtained by calling the member function geterr( ).

The second class is Download, which is the class that handles the downloading of files. Notice that the Download class
consists solely of static functions. Thus, putting the downloader into a class is more of an organizational technique
than it is a means of encapsulation. In fact, it would have been possible to use a namespace rather than a class for
this purpose. However, by using a class, it is possible to make several of the functions private, thus preventing their
.

use by other code. Also, using a class facilitates the addition of new features at a later date.

The following sections examine each part of Download in detail.

General Theory of Operation

Before examining the various pieces of Download, it will be helpful to understand its general theory of operation. To
download a file, the URL of that file is passed to the download( ) function. If no preexisting file by that name exists on
disk, then the file is downloaded in its entirety. However, if a partially downloaded file by that name is found, then only
the remainder of the file is downloaded. It is the ability to download the remainder of a file that makes it possible to
resume a download that has been interrupted.

To perform partial downloads, Download relies upon a feature provided by HTTP version 1.1 (or greater) that allows a
range of data to be downloaded: the Range header. Therefore, the downloader only works with URLs that support
HTTP version 1.1 or above. It is the Range header that allows an interrupted download to be restarted from the point
at which it left off. Because HTTP, version 1.1 is required, only HTTP downloads are supported by the downloader.

As it relates to an HTTP request, a header is information that accompanies a request. The Range header is a string
that has this general form:

籍
Range:bytes=start-end

书
Here, start specifies the beginning of the range, and end specifies the end. If end is left off, then the range runs from
start to the end of the file.

典 书
Download defines one public function, download( ), which is the function called to download a file. Thus, it is the entry
point for downloading. Download defines four support functions:

ishttp

+
httpverOK
经 不 读 Determines if the URL specifies an HTTP request.

Determines if HTTP version 1.1 or greater is being used.

+
getfname Obtains the filename portion of the URL.

C
openfile

典
Opens the file and returns its length, which will be greater than zero if a partial download
already exits.

经
These support functions are declared private within Download.

无
The download( ) Function

The download( ) function is the only public function defined by Download. It is called by user code and handles the
downloading of a file. Thus, to download a file, you call download( ). Because of its importance, we will examine it
line-by-line. It begins with
bool Download::download(char *url, bool reload,
void (*update)(unsigned long, unsigned long)) {

The download( ) function has three parameters. The first, url, is a pointer to a string that contains the complete URL for
the file (including the filename). The filename is assumed to be the name that follows the last / in the URL. For
example, in this URL the filename is MyFile.dat:
http://www.SoeSite.com/SomeDir/MyFile.dat

Keep in mind that you must specify the entire URL using the canonical form, which includes http://.

The second parameter, reload, determines if a file that has already been downloaded will be downloaded again. If
reload is true, then the file is downloaded in its entirety, whether or not it already exists on disk (either completely or
partially). If reload is false, then only the remaining portion of the file (if any) is downloaded. Thus, if you want to
retrieve a fresh copy of a file that you have already downloaded, pass true to reload.

The third parameter is update. It is a pointer to a function that will be periodically called by download( ) as a means of
.

keeping user code apprised of the download progress. If this parameter is null, then no update function is used. We
will examine the requirements for the update function parameter a bit later.

Next, the following local variables are declared:

ofstream fout; // output stream
unsigned char buf[BUF_SIZE]; // input buffer
unsigned long numrcved; // number of bytes read
unsigned long filelen; // length of file on disk
HINTERNET hIurl, hInet; // Internet handles
unsigned long contentlen;// length of content
unsigned long len; // length of contentlen
unsigned long total = 0; // running total of bytes received
char header[80]; // holds Range header

Notice the variables hIurl and hInet. These hold handles to the URL being downloaded and the Internet connection,
respectively. One or the other of these handles is used by the WinINet API functions.

The main function code begins as shown here. Notice that the entire download code is wrapped in a try/catch block,
which handles communication and/or file errors.

籍
try {
if(!ishttp(url))
throw DLExc("Must be HTTP url.");

书
The first thing that download( ) does is call ishttp( ) to determine if the URL specifies an HTTP request by confirming
that it begins with “http”. As explained, only HTTP requests are handled by the downloader.

经 典 读 书
Next, the receiving file is opened and its length is obtained.
// Open the file specified by url.
// The open stream will be returned
// in fout. If reload is true, then

+ 不
// any preexisting file will be truncated.

+
// The length of any preexisting file (after
// possible truncation) is returned.

C
filelen = openfile(url, reload, fout);

经 典
As the comments explain, if the file does not exist or is being reloaded, then any preexisting file by the specified name
will be truncated to zero. In this case, the length of the file returned by openfile( ) will be zero. Otherwise, if the file
already exists, its length will be returned. Remember, if a download was interrupted, then a partial file may already
exist when the download is restarted.

无
After the file is opened, a connection with the Internet is established by the following code:
// See if Internet connection available.
if(InternetAttemptConnect(0) != ERROR_SUCCESS)
throw DLExc("Can't connect.");

// Open Internet connection.

hInet = InternetOpen("downloader",
INTERNET_OPEN_TYPE_DIRECT,
NULL, NULL, 0);

if(hInet == NULL)
throw DLExc("Can't open connection.");

First, an attempt is made to connect to the Internet by calling the WinINet API function InternetAttemptConnect( ). This
function returns ERROR_SUCCESS if it is possible to establish a connection. Its only argument is reserved and must be
zero. If the computer is not currently connected, then (depending upon your computer’s settings) you will see a dialog
box asking if you want to connect. If no connection is available, InternetAttemptConnect( ) returns an error, which
causes an exception to be thrown.

Assuming that a connection is available, InternetOpen( ) is called to open the connection. It is another WinINet function
and has this prototype:
.

HINTERNET InternetOpen(LPCTSTR agent, DWORD access,

LPCTSTR proxy, LPCTSTR proxopt,
DWORD options)

Here, agent specifies the application name, and access specifies what type of access is used. For the downloader, the
access is INTERNET_OPEN_TYPE_DIRECT . Neither proxy nor proxopt are used when access is
INTERNET_OPEN_TYPE_DIRECT and must be null. The options parameter specifies any options, none of which are
needed by the downloader. It returns a handle to the open connection. Null is returned if the connection can’t be
opened.

Once the connection is open, a Range header is generated:

// Construct header requesting range of data.
sprintf(header, "Range:bytes=%d-", filelen);

This header is then passed to the WinINet function InternetOpenUrl( ), shown next, which opens the URL specified in
url:

// Open the URL and request range.

hIurl = InternetOpenUrl(hInet, url,
header, -1,

籍
INTERNET_FLAG_NO_CACHE_WRITE, 0);

if(hIurl == NULL) throw DLExc("Can't open url.");

典 书
The InternetOpenUrl( ) function has this prototype:

书
HINTERNET InternetOpenUrl(HINTERNET hInet, LPCTSTR url,
LPCTSTR headerstr, DWORD headlen,
DWORD options, LPDWORD extra);

+ 经 读
The Internet handle obtained from InternetOpen( ) is passed in hInet. A pointer to the string containing the URL to open
is passed in url. A pointer to a string containing one or more additional headers to be incorporated into the request is

不
passed in headerstr. The length of headerstr is passed in headlen, which can be –1 if headerstr points to a

+
null-terminated string, as it does for the downloader. Various option flags can be specified in options. The one used by

C
the downloader is

典
INTERNET_FLAG_NO_CACHE_WRITE

which prevents the downloaded file from being cached. Any additional, application-specific information can be passed

经
in extra. Because the downloader does not require any, zero is used. The function returns a handle to the open URL,
or null on failure.

无
Each response from the server carries with it a header. One component of this header contains the HTTP version.
Because the downloader requires that the server support HTTP, version 1.1 or higher, a check of the HTTP version is
necessary before proceeding. This is accomplished by calling httpverOK( ) , as shown here:
// Confirm that HTTP/1.1 or greater is supported.
if(!httpverOK(hIurl))
throw DLExc("HTTP/1.1 not supported.");

Assuming that the HTTP version is acceptable, the next step obtains the amount of content that is being downloaded.
This is accomplished by calling the WinINet function HttpQueryInfo( ), as shown here:
// Get content length.
len = sizeof contentlen;
if(!HttpQueryInfo(hIurl,
HTTP_QUERY_CONTENT_LENGTH |
HTTP_QUERY_FLAG_NUMBER,
&contentlen, &len, NULL))
throw DLExc("File or content length not found.");

When HttpQueryInfo( ) returns, the content length (in bytes) will be in contentlen. The content length automatically
takes into account the range requested by the Range header. Therefore, if the entire file is being downloaded, the
requested range will be “0-” (which specifies the entire file) and the content length will equal the length of the file.
.

When an interrupted download is being resumed, the range value will specify a starting point somewhere in the middle
of the file and the content length will equal the number of bytes remaining.

HttpQueryInfo( ) obtains information from a header. It has the following prototype:

BOOL HttpQueryInfo(HINTERNET hIurl, DWORD what, LPVOID buf,

LPDWORD buflen, LPDWORD index)

The handle of the request, which in this case is the one returned by InternetOpenUrl( ), is passed to hIurl. A value
specifying the item of information to obtain is passed in what. For the downloader, two values are needed. The first is
HTTP_QUERY_CONTENT_LENGTH

which obtains the length of the content being requested (which in this case is the file), and
HTTP_QUERY_FLAG_NUMBER

which requests that this value be represented as an integer value. A pointer to a buffer that will receive the information
is passed in buf, and a pointer to an integer that specifies the length of that buffer is passed in buflen. The index
parameter allows you to specify the index of a header, but the downloader does not require this feature, so null is
passed. The function returns true if it can obtain the requested information. The content length may not be available, in
which case HttpQueryInfo( ) returns false.

籍
Assuming that the content length is available, if the length of any preexisting file on disk is less than the content length,

书
and the content length does not equal zero, the file (or the remaining portion of the file) is downloaded by the following
lines of code:
// If existing file (if any) is not complete,

典 书
// then finish downloading.
if(filelen != contentlen && contentlen)
do {

经 读
// Read a buffer of info.
if(!InternetReadFile(hIurl, &buf,

+ 不
BUF_SIZE, &numrcved))

+
throw DLExc("Error occurred during download.");

C
// Write buffer to disk.

典
fout.write((const char *) buf, numrcved);
if(!fout.good())
throw DLExc("Error writing file.");

经
total += numrcved; // update running total

无
// Call update function, if specified.
if(update && numrcved > 0)
update(contentlen+filelen, total+filelen);

} while(numrcved > 0);

else
if(update)
update(filelen, filelen);

This code uses the WinINet function InternetReadFile( ) to read the file, one buffer at a time. Each time through the
loop, the function pointed to by update is called, unless update is null. As explained, the function pointed to by update is
used to report on the progress of the file download.

InternetReadFile( ) is a very useful function because it lets you read a file from the Internet in much the same way that
you read a file from disk. Its prototype is shown here:
BOOL InternetReadFile(HINTERNET hIurl, LPVOID buf, DWORD numbytes,
LPDWORD numrcvd);

The handle of the file is passed in hIurl. For the downloader, this is the handle returned by InternetOpenUrl( ). A pointer
to a buffer that will receive the data is passed in buf, and the number of bytes to read is passed in numbytes. This
value must not exceed the size of buf. The number of bytes actually read is returned in the variable pointed to by
.

numrcvd. This value will be zero when there are no more bytes to retrieve. The function returns true if successful and
false otherwise.

The download( ) function ends with the following lines:

} catch(DLExc) {
fout.close();
InternetCloseHandle(hIurl);
InternetCloseHandle(hInet);

throw; // rethrow the exception for use by caller

}

fout.close();
InternetCloseHandle(hIurl);
InternetCloseHandle(hInet);

return true;
}

Assuming successful completion, the download( ) function ends by closing the file and Internet handles and returning

籍
true. If an error has occurred, the handles are still closed and then the exception is rethrown. This allows user code to
respond to the error.

典 书
The ishttp( ) Function

书
As explained, only HTTP file downloads are supported by the downloader. Download uses the ishttp( ) function, shown
here, to confirm that the URL of the file specifies the HTTP protocol:

经 读
// Confirm that the URL specifies HTTP.
bool Download::ishttp(char *url) {

+
char str[5] = "";

+ 不
// Get first four characters from URL.
strncpy(str, url, 4);

C 典
// Convert to lowercase
for(char *p=str; *p; p++) *p = tolower(*p);

经
return !strcmp("http", str);
}

无
The operation of ishttp( ) is straightforward. It simply checks that the first four characters in the URL contain the string
“http”. It returns true if this is the case and false otherwise.

The httpverOK( ) Function

The httpverOK( ) function, shown here, confirms that the server handling the request supports HTTP version 1.1:
// Return true if HTTP version of 1.1 or greater.
bool Download::httpverOK(HINTERNET hIurl) {
char str[80];
unsigned long len = 79;

// Get HTTP version.

if(!HttpQueryInfo(hIurl, HTTP_QUERY_VERSION, &str, &len, NULL))
return false;

// First, check major version number.

char *p = strchr(str, '/');
p++;
if(*p == '0') return false; // can't use HTTP 0.x
.

// Now, find start of minor HTTP version number.

p = strchr(str, '.');
p++;

// Convert to int.
int minorVerNum = atoi(p);

if(minorVerNum > 0) return true;

return false;
}

A handle to the URL is passed to httpverOK( ) . Next, the HTTP version is obtained (in string form) by calling
HttpQueryInfo( ) on this handle, specifying HTTP_QUERY_VERSION, which requests the version information. This
query stores the version string in str. For HTTP version 1.1, the string will have this form:
HTTP/1.1

The function then sets p to point to the / character by calling the standard library function strchr( ). It then increments p
and confirms that the digit it points to is nonzero. Next, a pointer to the period separating the major and minor version
numbers is found, and then the pointer is advanced to point to the start of the numeric value that follows the period.

籍
Finally, this value is converted into an integer through use of the standard library function atoi( ). If the minor version
number is greater than or equal to 1, then the HTTP version is greater than or equal to 1.1.

书
The getfname( ) Function

The getfname( ) function, shown next, extracts the filename from a URL:

经 典 读 书
// Extract the filename from the URL. Return false if
// the filename cannot be found.
bool Download::getfname(char *url, char *fname) {
// Find last /.

+
char *p = strrchr(url, '/');

+ 不
// Copy filename after the last /.
if(p && (strlen(p) < MAX_FILENAME_SIZE)) {

C 典
p++;
strcpy(fname, p);
return true;

经
}
else
return false;

无
}

The function is passed a pointer to a string that holds a URL and a pointer to a character array that will hold the
filename upon the function’s return. The filename is assumed to be that portion of the URL that extends from the last /
to the end. The function returns true if successful and false if the filename cannot be found.

The openfile( ) Function

The openfile( ) function opens a file on disk to which the downloaded file will be written. It also returns the length of any
preexisting file. It is shown here:
// Open the output file, initialize the output
// stream, and return the file's length. If
// reload is true, first truncate any existing
// file.
unsigned long Download::openfile(char *url,
bool reload,
ofstream &fout) {
char fname[MAX_FILENAME_SIZE];

if(!getfname(url, fname))
throw DLExc("File name error.");
 if(!reload)
fout.open(fname, ios::binary | ios::out |
ios::app | ios::ate);
else
fout.open(fname, ios::binary | ios::out |
ios::trunc);

if(!fout)
throw DLExc("Can't open output file.");

// Get current file length.

return fout.tellp();
}

The openfile( ) function has three parameters: url, reload, and fout. A pointer to the string that contains the URL is
passed to url. A value that determines if any preexisting file of the same name is truncated is passed in reload. A
pointer to a variable that will contain the opened file stream when openfile( ) returns is passed in fout.

First, openfile( ) obtains the filename from the string pointed to by url. Then it checks the value of reload. If reload is

籍
true, the file is opened for output, with the contents of any preexisting file being truncated to zero. If reload is false, it
opens the file without truncating any preexisting contents and specifies that all output occur at the end of the file. It
then advances the file pointer to the end of the file. Next, the length of the file is obtained by calling tellp( ). For a new

书
file, this value will be zero. However, if a preexisting file was found, this value gives the length of the file because the
file pointer was advanced to the end of the file when it was opened. The length of the file is returned.

经 典 读 书
The update( ) Function

As explained earlier, if you want to track the progress of the download, you must pass to download( )’s update
parameter a pointer to a function that will receive the tracking information. (For silent operation, update can simply be

+ 不
allowed to default to null.) The tracking function must have this prototype. (Of course, the name of the function can be

+
different.)
void update(unsigned long total, unsigned long part);

C 典
When update( ) is called, the total length of the file is passed in total and the amount currently downloaded is passed in
part. You can use this information to monitor the progress of the download and to keep the user informed.

无 经
 The Download Header File

The following header file must be included in any file that uses the downloader. Call this file dl.h.
// Header file for downloader. Call this file dl.h.
#include <iostream>
#include <string>
#include <windows.h>
#include <wininet.h>
#include <fstream>
using namespace std;

const int MAX_ERRMSG_SIZE = 80;

// Exception class for downloading errors.

class DLExc {

籍
char err[MAX_ERRMSG_SIZE];
public:
DLExc(char *exc) {

书
if(strlen(exc) < MAX_ERRMSG_SIZE)
strcpy(err, exc);

典 书
}

const char * geterr() {

经 读
return err;
}

+
};

+ 不
class Download {
static bool httpverOK(HINTERNET hIurl);

C 典
static bool getfname(char *url, char *fname);
static unsigned long openfile(char *url, bool reload,
ofstream &fout);

经
public:
static bool download(char *url, bool restart=false,
void (*update)(unsigned long, unsigned long)=NULL);

无
};
.

Demonstrating the File Downloader

The following program demonstrates the file downloader by creating a simple console-based application that
downloads a file. To follow along, call this file dltest.cpp.
// A Sample program that uses Download.
#include <iostream>
#include "dl.h"

// This function displays the download progress as a percentage.

void showprogress(unsigned long total, unsigned long part) {
int val = (int) ((double) part/total*100);
cout << val << "%" << endl;
}

int main(int argc, char *argv[])

{

char url[] =

书 籍
// This URL is for demonstration purposes only. Substitute
// the URL of the file that you want to download.

"http://www.osborne.com/products/0072226803/0072226803_code.zip";

典 书
bool reload = false;

经 读
if(argc==2 && !strcmp(argv[1], "reload"))
reload = true;

+ 不
cout << "Beginning download.\n";

try {

+
C 典
if(Download::download(url, reload, showprogress))
cout << "Download Complete\n";
} catch(DLExc exc) {

经
cout << exc.geterr() << endl;
cout << "Download Interrupted\n";
}

无
return 0;

There are three things of interest in this program. First, notice that it includes a hard-coded URL. This URL specifies
the file that contains the free, online code for another of my books, C++: The Complete Reference, 4th Edition. This
file makes a convenient test of the downloader. Of course, you can specify a URL of your own choosing.

Next, notice the showprogress( ) function. A pointer to this function is passed to download( )’s update parameter. This
means that it is called each time a buffer of data is downloaded. It displays a value that indicates the percentage of
completion.

Finally, notice that dltest accepts one command-line argument. If this argument is “reload”, then the file is downloaded
in its entirety whether or not it already exists on disk. If this argument is not present, the file (or the remaining portion of
the file) is downloaded only if the file on disk is not complete.

To compile dltest.cpp, you must remember to link in wininet.lib. For example, if you are compiling from the command
line using Visual C++, then you will use this command line to compile this program:
cl -GX dl.cpp dltest.cpp wininet.lib

If you are using an IDE, then you must remember to add wininet.lib to the link.
.

A GUI-Based Downloader
The Download class can be used whenever you need to download a file from the Internet from within program code.
For example, you might use the downloader to obtain a weekly sales report. If you add a user interface front end, it
can also be used as a stand-alone Internet utility. To illustrate this second use, this chapter concludes by developing a
complete, GUI-based file downloader for Windows called WinDL.

WinDL displays a dialog box that allows the user to enter the URL of the file to download. It has a progress bar that
shows the status of the download. It also includes a check box that lets the user request that the file be fully reloaded.
Figure 5-1 shows the file download dialog box.

书 籍
经 典 读 书
+ + 不
C 典
Figure 5-1: The File Download dialog box

The WinDL Code

经
WinDL consists of two main parts. The first is, of course, the code for Download, which was described earlier. The

无
second is the code that handles the GUI interface. This code is shown here. To follow along, call this file windl.cpp.
// WinDL: A GUI-based file download utility.

#include <windows.h>
#include <commctrl.h>
#include <cstring>
#include <cstdio>
#include "windl.h"
#include <process.h>
#include "dl.h"

const int URL_BUF_SIZE = 1024;

LRESULT CALLBACK WindowFunc(HWND, UINT, WPARAM, LPARAM);

BOOL CALLBACK DialogFunc(HWND, UINT, WPARAM, LPARAM);

void showprogress(unsigned long total,

unsigned long part);
void resetprogress();
unsigned __stdcall dlstart(void * reload);
.

char szWinName[] = "Download"; // name of window class

HINSTANCE hInst; // instance handle

HWND hwnd; // handle of main window
HWND hProgWnd; // handle of progress bar

HANDLE hThrd = 0; // thread handle

unsigned long Tid; // thread ID

// Progress counters.
int percentdone = 0;
int oldpercentdone = 0;

// A small struct for passing info to dlstart().

struct ThrdInfo {
char *url; // pointer to URL string
int reload; // reload flag
HWND hPBStart; // handle of Start button
};

籍
int WINAPI WinMain(HINSTANCE hThisInst, HINSTANCE hPrevInst,
LPSTR lpszArgs, int nWinMode)
{

书
MSG msg;
WNDCLASSEX wcl;

典 书
INITCOMMONCONTROLSEX cc;

// Define a window class.

经 读
wcl.cbSize = sizeof(WNDCLASSEX);

+
wcl.hInstance = hThisInst; // handle to this instance

不
wcl.lpszClassName = szWinName; // window class name

+
wcl.lpfnWndProc = WindowFunc; // window function
wcl.style = 0; // default style

C 典
wcl.hIcon = LoadIcon(NULL, IDI_APPLICATION); // large icon
wcl.hIconSm = NULL; // use small version of large icon

经
wcl.hCursor = LoadCursor(NULL, IDC_ARROW); // cursor style

wcl.lpszMenuName = NULL; // no menu

无
wcl.cbClsExtra = 0; // no extras
wcl.cbWndExtra = 0;

wcl.hbrBackground = NULL; // not used

// Register the window class.

if(!RegisterClassEx(&wcl)) return 0;

// Create a main window that won't be visible.

hwnd = CreateWindow(
szWinName, // name of window class
"File Downloader", // title
0, // no style needed
0, 0, 0, 0, // no dimensions
NULL, // no parent window
NULL, // no menu
hThisInst, // instance handle
NULL // no additional arguments
);

hInst = hThisInst; // save the current instance handle

.

// Initialize the common controls. This is

// needed because of the progress bar.
cc.dwSize = sizeof(INITCOMMONCONTROLSEX);
cc.dwICC = ICC_PROGRESS_CLASS;
InitCommonControlsEx(&cc);

// Show the window minimized.

ShowWindow(hwnd, SW_SHOWMINIMIZED);

// Create the download dialog box.

DialogBox(hInst, "DLDB", hwnd, (DLGPROC) DialogFunc);

// Create the message loop.

while(GetMessage(&msg, NULL, 0, 0))
{
TranslateMessage(&msg); // translate keyboard messages
DispatchMessage(&msg); // return control to Windows
}

籍
return msg.wParam;
}

书
// Window function.
LRESULT CALLBACK WindowFunc(HWND hwnd, UINT message,

典 书
WPARAM wParam, LPARAM lParam)
{
switch(message) {

经 读
case WM_DESTROY:
PostQuitMessage(0); // terminate the program

+
break;

不
default:

+
return DefWindowProc(hwnd, message, wParam, lParam);
}

C 典
return 0;
}

经
// Downloader Dialog function.
BOOL CALLBACK DialogFunc(HWND hdwnd, UINT message,
WPARAM wParam, LPARAM lParam)

无
{
// Here, url is initialized with a sample url for
// demonstration purposes only.
static char url[URL_BUF_SIZE] =
"http://www.osborne.com/products/0072226803/0072226803_code.zip";

static ThrdInfo ti;

switch(message) {
case WM_INITDIALOG:
// Initialize edit box with URL.
SetDlgItemText(hdwnd, IDD_EB1, url);

// Create progress bar.

hProgWnd = CreateWindow(PROGRESS_CLASS,
"",
WS_CHILD | WS_VISIBLE | WS_BORDER,
4, 64, 320, 12,
hdwnd, NULL, hInst, NULL);

// Set step increment to 1.

SendMessage(hProgWnd, PBM_SETSTEP, 1, 0);
.

return 1;
case WM_COMMAND:
switch(LOWORD(wParam)) {
case IDCANCEL:
EndDialog(hdwnd, 0);
PostQuitMessage(0);

return 1;

case IDD_START: // start download

// Set position to zero.
SendMessage(hProgWnd, PBM_SETPOS, 0, 0);

// Get URL from edit box.

GetDlgItemText(hdwnd, IDD_EB1, url, URL_BUF_SIZE);
ti.url = url;

// Get reload status.

ti.reload = SendDlgItemMessage(hdwnd, IDD_CB1,

籍
BM_GETCHECK, 0, 0);

// Get handle to Start button.

书
ti.hPBStart = GetDlgItem(hdwnd, IDD_START);

典 书
// Reset progress counters.
resetprogress();

经 读
// Start download thread.
if(!hThrd)

+
hThrd = (HANDLE) _beginthreadex(NULL, 0, dlstart,

不
(void *) &ti, 0, (unsigned *) &Tid);

}
}

C+
return 1;

典
经
return 0;
}

无
// Show progress in the progress bar. This is called
// by the download() function.
void showprogress(unsigned long total,
unsigned long part) {

percentdone = (part*100)/total;

if(percentdone > oldpercentdone) {

for(int i= oldpercentdone; i < percentdone; i++) {
// Advance the progress bar.
SendMessage(hProgWnd, PBM_STEPIT, 0, 0);
}
oldpercentdone = percentdone;
}
}

// Reset the progress counters.

void resetprogress() {
percentdone = 0;
oldpercentdone = 0;
}
.

// Thread entry function that begins downloading.

unsigned __stdcall dlstart(void * param) {
ThrdInfo *tip = (ThrdInfo *) param;

// Disable Start button.

EnableWindow(tip->hPBStart, 0);

try {
if(tip->reload == BST_CHECKED)
Download::download(tip->url, true, showprogress);
else
Download::download(tip->url, false, showprogress);
} catch(DLExc exc) {
MessageBox(hwnd, exc.geterr(),
"Download Error", MB_OK);
}

// Enable Start button.

EnableWindow(tip->hPBStart, 1);

籍
CloseHandle(hThrd); // close the thread handle
hThrd = 0; // set thread handle to inactive

书
return 0;
}

典 书
WinDL uses this resource file:
// Resources for file downloader.

经 读
#include <windows.h>
#include "windl.h"

+ + 不
DLDB DIALOGEX 18, 18, 164, 100
CAPTION "Download a File"
STYLE DS_MODALFRAME | WS_POPUP | WS_CAPTION |

C 典
WS_SYSMENU | WS_VISIBLE
{
PUSHBUTTON "Start", IDD_START, 42, 80, 30, 14

经
PUSHBUTTON "Cancel", IDCANCEL, 90, 80, 30, 14

CTEXT "Download Progress", IDD_TEXT2, 2, 40, 160, 12

}
无
CTEXT "Enter URL", IDD_TEXT1, 2, 16, 160, 12
EDITTEXT IDD_EB1, 2, 1, 160, 12, ES_LEFT | WS_CHILD |
WS_VISIBLE | WS_BORDER | ES_AUTOHSCROLL

AUTOCHECKBOX "Reload", IDD_CB1, 62, 56, 36, 14

Both the resource file and program code require the header file windl.h, shown here:
#define IDD_START 100

#define IDD_CB1 200

#define IDD_EB1 300

#define IDD_TEXT1 401

#define IDD_TEXT2 402

To compile WinDL, you must create a project that contains these files:

dl.cpp windl.cpp windl.rc

The header files dl.h and windl.h must also be available. You must also remember to link in wininet.lib and
comctl32.lib. (The comctl32.lib library is required by the progress control.) Finally, because download( ) runs in its own
thread, you must link in the multithreaded libraries.

How WinDL Works

WinDL provides a visual front end to the Download class that handles user input and displays the progress of the
download. WinDL begins by creating a minimized main window and then displays the download dialog box. Thus,
WinDL is a dialog box–based application that never displays a main window. As mentioned, it is beyond the scope of
this book to describe the parts of WinDL that support the basic Windows framework common to all Windows
programs. However, those portions of the code relating specifically to the downloader are described next.

The function that handles user interaction with the dialog box is DialogFunc( ). It declares two static variables. The first
is an array called url that provides the backing array for the edit box. The second is ti, which is a small structure that
contains information that will be passed to the thread entry function.

When the dialog box is created, it initializes the edit box and the progress bar. As the comments suggest, the edit box
is initialized with a URL simply for the sake of demonstration. (Normally, this edit box will not contain an initial string.)
The progress bar increment is set to 1. By default, progress bars have a range of 0 to 100, so by setting the increment

籍
to 1, each increment of the bar equals 1 percent.

To download a file, enter the URL of the desired file into the edit box and then click the Start button. This initiates the

书
following sequence of events. First, the progress bar is reset to zero. Next, a string containing the URL is obtained
from the edit box, the reload status is obtained from the check box, and a handle to the Start pushbutton is obtained.

典 书
These are stored in ti, which will be passed to the thread entry function. Next, the progress counters are reset to zero.
These are the global variables percentdone and oldpercentdone. They are used to update the position of the progress
bar during the download. Finally, a new thread that will handle the download is started.

+ 经 读
It is necessary to run download( ) in a thread of its own because the Windows messaging system assumes that control
will pass back to Windows relatively quickly. (That is, DialogFunc( ) can’t undertake some extended operation that

不
prevents new messages from being processed.) The entry function for the thread is dlstart( ). As explained in Chapter

+
3, in Windows all thread entry functions take precisely one void * parameter. This parameter can be passed anything

C
needed by the function. In this case, it is passed a pointer to the ti structure, which contains three fields: url, reload,

典
and hPBStart. The URL to download is pointed to by the url field. The status of the Reload check box is in the reload
field. It determines if the file will be reloaded in its entirety. The Start button handle is in the hPBStart field. It is used by
dlstart( ) to disable the Start button when a download begins and to enable it when the download ends.

closed.

无 经
The dlstart( ) function calls download( ) to handle the downloading of the file. When it does so, it passes the address of
showprogress to the update parameter. When download( ) ends, the Start button is enabled and the thread handle is

The showprogress( ) function is called by download( ) to display the progress of the download. It simply increments the
progress bar each time another percent of the file is downloaded.
 Some Things to Try
Several enhancements can be added to Download. One of the first things you might want to try is adding the ability to
download files given an FTP address. Because InternetOpenUrl( ) supports FTP, this is a relatively simple matter.
Although HTTP version 1.0 is becoming increasingly rare, you might also want to try adding support for it. This is easy
to do because it just means that the file will always be downloaded in its entirety. Another useful enhancement to
Download is to enable it to retrieve a list of files. One way to do this is to have it read the URLs from a disk file. Finally,
you might want to add an automatic retry feature that attempts to finish a download that was interrupted.

The general file download mechanism has applications beyond being the basis of a file downloader. The Download
class can be used whenever a file must be retrieved. For example, you could use Download to create a remote data
acquisition program that downloads a data file, such as an inventory report, from a remote site on a regular basis.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Chapter 6: Financial Calculations in C++

Overview
Despite all of the large, sophisticated applications, such as compilers, Internet browsers, word processors,
databases, and accounting packages that dominate much of the computing landscape, there has remained a class of
programs that are both popular and small. These perform various financial calculations, such as computing the regular
payments on a loan, the future value of an investment, or the remaining balance on a loan. None of these calculations
are very complicated or require much code, yet they yield information that is quite useful.

Because C++ excels in the creation of high-powered system software, it is often overlooked for financial applications,
but this is a mistake. C++ has excellent capabilities in this area, supporting a complete range of math functions and
efficient floating-point arithmetic. Furthermore, because of its ability to produce extremely fast executable code, C++ is
a perfect fit for programs performing sophisticated financial analysis and modeling.

籍
To illustrate the ease with which C++ handles financial calculations, this chapter develops a number of small programs
that perform the financial computations shown here:

书
1. Regular payments on a loan

2. Future value of an investment

经 典 读 书
3. Initial investment needed to attain a desired future value

4. Investment necessary for a desired annuity

+
5. The maximum annuity from an investment

C+ 不 6. Remaining balance on a loan

These programs can be used as-is, or tweaked to fit your specific need. Frankly, even though these are the simplest

典
programs in this book, they are also the ones that you may find yourself using the most!

无 经
.

Finding the Payments for a Loan

Perhaps the most popular financial calculation is the one that computes the regular payments on a loan, such as a car
or house loan. The payments on a loan are found by using the following formula:
Payment = (intRate * (principal / payPerYear)) /
(1 – ((intRate / payPerYear) + 1) – payPerYear * numYears)

where intRate specifies the interest rate, principal contains the starting balance, payPerYear specifies the number of
payments per year, and numYears specifies the length of the loan in years.

In the following program, the function called regpay( ) uses the preceding formula to compute the payments on a loan.
It is passed the principal, the interest rate, the term of the loan in years, and the number of payments per year. It
returns the payment.
#include <iostream>
#include <cmath>

籍
#include <iomanip>
#include <locale>

书
using namespace std;

// Compute the regular payments on a loan.

典 书
double regpay(double principal, double intRate,
int numYears, int payPerYear) {
double numer;

经 读
double denom;
double b, e;

+ 不
intRate /= 100.0; // convert percentage to fraction

+
C
numer = intRate * principal / payPerYear;

典
e = -(payPerYear * numYears);
b = (intRate / payPerYear) + 1.0;

经
denom = 1.0 - pow(b, e);

无
return numer / denom;
}

int main() {
double p, r;
int y, ppy;

// Set locale to English. Adjust as necessary

// for your language and/or region.
cout.imbue(locale("english"));

cout << "Enter principal: ";

cin >> p;

cout << "Enter interest rate (as a percentage): ";

cin >> r;

cout << "Enter number years: ";

cin >> y;

cout << "Enter number of payments per year: ";

 cin >> ppy;

cout << "\nPayment: " << fixed << setprecision(2)

<< regpay(p, r, y, ppy) << endl;

return 0;
}

To compute a loan payment, simply enter the information requested by the prompts. Here is a sample run:
Enter principal: 1000
Enter interest rate (as a percentage): 9
Enter number years: 5
Enter number of payments per year: 12

Payment: 20.76

There are a couple of points of interest to notice in main( ). First, the locale associated with cout is set to English. This
is accomplished by calling the imbue( ) member function and passing the locale object for English. This ensures that
monetary values are displayed using the conventions defined by English, which is the comma as the thousands
separator and the period as a decimal point. Second, before the loan payment is displayed, the number format is

籍
changed to fixed with a precision of 2. This causes the output to have two decimal places, with appropriate rounding.
The decimal places are also padded with zeros as necessary. All of the financial programs use this same approach.

书
To change the format to reflect a different language or region, simply change the language/region of the locale object
passed to imbue( ).

经 典 读 书
+ + 不
C 典
无 经
.

Finding the Future Value of an Investment

Another popular financial calculation finds the future value of an investment given the initial investment, the rate of
return, the number of compounding periods per year, and the number of years the investment is held. For example,
you might want to know what your retirement account will be worth in 12 years if it currently contains $98,000 and has
an average annual rate of return of 6 percent. The program shown here will supply the answer.

To compute the future value, use the following formula:

Future Value = principal * ((rateOfRet / compPerYear) + 1) compPerYear * numYears

where rateOfRet specifies the rate of return, principal contains the initial value of the investment, compPerYear
specifies the number of compounding periods per year, and numYears specifies the length of the investment in years.
If you use an annualized rate of return for rateOfRet, then the number of compounding periods is 1.

In the following program, the futval( ) function uses the preceding formula to compute the future value of an
investment. It is passed the principal, the rate of return, the term of the investment in years, and the number of

籍
compoundings per year. It returns the future value.
#include <iostream>

书
#include <cmath>
#include <iomanip>
#include <locale>

经 典
using namespace std;

读 书
// Compute the future value of an investment.
double futval(double principal, double rateOfRet,

+
int numYears, int compPerYear) {

不
double b, e;

C+
rateOfRet /= 100.0; // convert percentage to fraction

典
b = (1 + rateOfRet/compPerYear);
e = compPerYear * numYears;

无
int main() {经
return principal * pow(b, e);

double p, r;
int y, cpy;

// Set locale to English. Adjust as necessary

// for your language and/or region.
cout.imbue(locale("english"));

cout << "Enter principal: ";

cin >> p;

cout << "Enter rate of return (as a percentage): ";

cin >> r;

cout << "Enter number years: ";

cin >> y;

cout << "Enter number of compoundings per year: ";

cin >> cpy;
 cout << "\nFuture value: " << fixed << setprecision(2)
<< futval(p, r, y, cpy) << endl;

return 0;
}

Here is a sample run:

Enter principal: 10000
Enter rate of return (as a percentage): 6
Enter number years: 5
Enter number of compoundings per year: 12

Future value: 13,488.50

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

Finding the Initial Investment Required to Achieve a Future Value

Sometimes you will want to know how large of an initial investment is required to achieve some future value. For
example, if you are saving for your child’s college education and you know that you will need $75,000 in five years,
how much money do you need to invest at 7 percent to reach that goal? The program developed here can answer that
question.

The formula to compute an initial investment is shown here:

Initial Investment = targetValue / (((rateOfRet / compPerYear) + 1) compPerYear * numYears)

where rateOfRet specifies the rate of return, targetValue contains the desired future value, compPerYear specifies the
number of compounding periods per year, and numYears specifies the length of the investment in years. If you use an
annualized rate of return for rateOfRet, then the number of compounding periods is 1.

In the following program, the initval( ) function uses the preceding formula to compute the initial investment required to
reach a desired future value. It is passed the target value, the rate of return, the term of the investment in years, and

籍
the number of compoundings per year. It returns the initial investment required to reach the target value.
#include <iostream>

书
#include <cmath>
#include <iomanip>
#include <locale>

经 典
using namespace std;

读 书
// Compute the initial investment necessary for
// a specified future value.

+
double initval(double targetValue, double rateOfRet,

不
int numYears, int compPerYear) {

+
double b, e;

C 典
rateOfRet /= 100.0; // convert percentage to fraction

b = (1 + rateOfRet/compPerYear);

经
e = compPerYear * numYears;

return targetValue / pow(b, e);

无
}

int main() {
double p, r;
int y, cpy;

// Set locale to English. Adjust as necessary

// for your language and/or region.
cout.imbue(locale("english"));

cout << "Enter desired future value: ";

cin >> p;

cout << "Enter rate of return: ";

cin >> r;

cout << "Enter number years: ";

cin >> y;

cout << "Enter number of compoundings per year: ";

cin >> cpy;
 cout << "\nInitial investment required: "
<< fixed << setprecision(2)
<< initval(p, r, y, cpy) << endl;

return 0;
}

Here is a sample run:

Enter desired future value: 75000
Enter rate of return (as a percentage): 7
Enter number years: 5
Enter number of compoundings per year: 4

Initial investment required: 53,011.84

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

Finding the Initial Investment Needed for a Desired Annuity

Another common financial calculation computes the amount of money that one must invest so that a desired annuity,
in terms of a regular withdrawal, can be paid. For example, you might decide that you need $5,000 per month at
retirement, and that you will need that amount for 20 years. The question is how much will you need to invest to
secure that annuity? The answer can be found using the following formula:
Initial Investment = ((regWD * wdPerYear) / rateOfRet) *
(1 – (1 / (rateOfRet / wdPerYear + 1) wdPerYear * numYears ))

where rateOfRet specifies the rate of return, regWD contains the desired regular withdrawal, wdPerYear specifies the
number of withdrawals per year, and numYears specifies the length of the annuity in years.

In the next program, the annuity( ) function computes the initial investment required to produce the desired annuity. It is
passed the desired withdrawal, the rate of return, the term of the annuity in years, and the number of compoundings
per year. It returns the minimum investment needed to produce the annuity.

籍
#include <iostream>
#include <cmath>
#include <iomanip>

书
#include <locale>

using namespace std;

经 典 读 书
// Compute the initial investment necessary for
// a desired annuity. In other words, it finds
// the initial amount needed to allow the regular
// withdrawals of a desired amount over a period

+ 不
// of time.

+
double annuity(double regWD, double rateOfRet,
int numYears, int numPerYear) {

C 典
double b, e;
double t1, t2;

经
rateOfRet /= 100.0; // convert percentage to fraction

无
t1 = (regWD * numPerYear) / rateOfRet;

b = (1 + rateOfRet/numPerYear);
e = numPerYear * numYears;

t2 = 1 - (1 / pow(b, e));

return t1 * t2;
}

int main() {
double wd, r;
int y, wpy;

// Set locale to English. Adjust as necessary

// for your language and/or region.
cout.imbue(locale("english"));

cout << "Enter desired withdrawal: ";

cin >> wd;

cout << "Enter rate of return (as a percentage): ";

 cin >> r;

cout << "Enter number years: ";

cin >> y;

cout << "Enter number of withdrawals per year: ";

cin >> wpy;

cout << "\nInitial investment required: "

<< fixed << setprecision(2)
<< annuity(wd, r, y, wpy) << endl;

return 0;
}

Here is a sample run.

Enter desired withdrawal: 5000
Enter rate of return (as a percentage): 6
Enter number years: 20
Enter number of withdrawals per year: 12

籍
Initial investment required: 697,903.86

书
经 典 读 书
+ + 不
C 典
无 经
.

Finding the Maximum Annuity for a Given Investment

Another annuity calculation computes the maximum annuity (in terms of a regular withdrawal) available from a given
investment over a specified period of time. For example, if you have $500,000 in a retirement account, how much can
you take out each month for 20 years, assuming a 6 percent rate of return? The formula that computes the maximum
withdrawal is shown here:
Maximum Withdrawal = principal * ( ( (rateOfRet / wdPerYear) /
(–1 + ((rateOfRet / wdPerYear) + 1) wdPerYear * numYears ) )
+ (rateOfRet / wdPerYear) )

where rateOfRet specifies the rate of return, principal contains the value of the initial investment, wdPerYear specifies
the number of withdrawals per year, and numYears specifies the length of the annuity in years.

The maxwd( ) function, shown in the following program, computes the maximum periodic withdrawals that can be
made over a specified length of time for an assumed rate of return. It is passed the principal, the rate of return, the

籍
term of the annuity in years, and the number of compoundings per year. It returns the maximum annuity.
#include <iostream>

书
#include <cmath>
#include <iomanip>
#include <locale>

经 典
using namespace std;

读 书
// Compute the maximum annuity that can

+
// be withdrawn from an investment over

不
// a period of time.

+
double maxwd(double principal, double rateOfRet,
int numYears, int numPerYear) {

C 典
double b, e;
double t1, t2;

经
rateOfRet /= 100.0; // convert percentage to fraction

无
t1 = rateOfRet / numPerYear;

b = (1 + t1);
e = numPerYear * numYears;

t2 = pow(b, e) - 1;

return principal * (t1/t2 + t1);

}

int main() {
double p, r;
int y, wpy;

// Set locale to English. Adjust as necessary

// for your language and/or region.
cout.imbue(locale("english"));

cout << "Enter principal: ";

cin >> p;

cout << "Enter rate of return (as a percentage): ";

 cin >> r;

cout << "Enter number years: ";

cin >> y;

cout << "Enter number of withdrawals per year: ";

cin >> wpy;

cout << "\nMaximum withdrawal: " << fixed << setprecision(2)

<< maxwd(p, r, y, wpy) << endl;

return 0;
}

Here is a sample run:

Enter principal: 500000
Enter rate of return (as a percentage): 6
Enter number years: 20
Enter number of withdrawals per year: 12

籍
Maximum withdrawal: 3,582.16

典 书 书
+ 经 不 读
C+ 典
无 经
.

Finding the Remaining Balance on a Loan

Often you will want to know the remaining balance on a loan. This is easily calculated if you know the original principal,
the interest rate, the loan length, and the number of payments made. To find the remaining balance, you must sum the
payments, subtracting from each payment the amount allocated to interest, and then subtract that result from the
principal.

The balance( ) function shown here, finds the remaining balance of a loan. It is passed the principal, the interest rate,
the payment, the number of payments per year, and the number of payments that have been made. It returns the
remaining balance.
#include <iostream>
#include <cmath>
#include <iomanip>
#include <locale>

using namespace std;

籍
// Find the remaining balance on a loan.

书
double balance(double principal, double intRate,
double payment, int payPerYear,
int numPayments) {

典 书
double bal = principal;
double rate = intRate / payPerYear;

经 读
rate /= 100.0; // convert percentage to fraction

+ 不
for(int i = 0; i < numPayments; i++)

+
bal -= payment - (bal * rate);

C 典
return bal;
}

经
int main() {
double p, r, pmt;
int ppy, npmt;

无
// Set locale to English. Adjust as necessary
// for your language and/or region.
cout.imbue(locale("english"));

cout << "Enter original principal: ";

cin >> p;

cout << "Enter interest rate (as a percentage): ";

cin >> r;

cout << "Enter payment: ";

cin >> pmt;

cout << "Enter number of payments per year: ";

cin >> ppy;

cout << "Enter number of payments made: ";

cin >> npmt;

cout << "Remaining balance: " << fixed << setprecision(2)

 << balance(p, r, pmt, ppy, npmt) << endl;

return 0;
}

A sample run is shown here:

Enter original principal: 10000
Enter interest rate (as a percentage): 9
Enter payment: 207.58
Enter number of payments per year: 12
Enter number of payments made: 30

Remaining balance: 5,558.19

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Some Things to Try
There are many other financial calculations that you might find useful. For example, the rate of return of an investment
or the amount of a regular deposit over time to reach a future value would be useful additions. You could also print a
loan amortization chart.

You might want to try creating a larger application that offers all of the calculations presented in this chapter, allowing
the user to select the desired calculation from a menu.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Chapter 7: AI-Based Problem Solving

Overview
This chapter examines a topic from an interesting discipline of programming: artificial intelligence (AI). One of the
goals of this book is to show the range and versatility of C++. Perhaps nothing demonstrates this better than its
application to the demanding realm of artificial intelligence.

The field of artificial intelligence is comprised of several fascinating areas, but fundamental to many AI-based
applications is problem solving. Essentially, there are two types of problems. The first type can be solved through the
use of some sort of deterministic procedure that is guaranteed success, such as the computation of the sine of an
angle or the square root of a value. These types of problems are easily translated into algorithms that a computer can
execute. In the real world, however, few problems lend themselves to such straightforward solutions. Instead, many
problems can be solved only by searching for a solution. It is this type of problem solving with which AI is concerned. It

籍
is also the type of searching that is explored in this chapter.

To understand why searching is so important to AI, consider the following. One of the early goals of AI research was

书
the creation of a general problem solver. A general problem solver is a program that can produce solutions to all sorts
of different problems about which it has no specific, designed-in knowledge. It is an understatement to say that such a
program would be highly desirable. Unfortunately, a general problem solver is as difficult to realize as it is tantalizing.

典 书
One trouble is the sheer size and complexity of many real-world situations. Because a general problem solver must
search for a solution through what might be a very large, maze-like universe of possibilities, finding ways to search

经 读
such an environment was a priority. Although we won’t attempt something as ambitious as a general problem solver in
this chapter, we will explore several AI-based search techniques that are applicable to a wide variety of problems.

+ 不
C++ is a language well suited to the AI developer. The reason is because C++ offers significant support for the

+
program elements that AI-based applications typically use: recursion, lists, and stacks. As you know, C++ handles
recursion with both ease and efficiency. Add to that the power of the various STL containers, and you have a

C 典
programming environment that streamlines AI development.

无 经
.

Representation and Terminology

Imagine that you have lost your car keys. You know that they are somewhere in your house, which looks like this:

书 籍
You are standing at the front door (where the X is). As you begin your search, you check the living room. Then you go
down the hall to the first bedroom, through the hall to the second bedroom, back to the hall, and to the master

典 书
bedroom. Not having found your keys, you backtrack further by going back through the living room. Finally, you find
your keys in the kitchen. This situation is easily represented by a graph, as shown in Figure 7-1. Representing search
problems in a graphical form is helpful because it provides a convenient way to depict the way a solution was found.

+ 经 不 读
C+ 典
无 经
Figure 7-1: The solution path to find the missing keys

With the preceding discussion in mind, consider the following terms, which will be used throughout this chapter:

Node A discrete point

Terminal A node that ends a path

node

Search The set of all nodes

space

Goal The node that is the object of the search

Heuristics Information about whether any specific node is a better next choice than another

Solution path A directed graph of the nodes visited en route to the goal

In the example of the lost keys, each room in the house is a node; the entire house is the search space; the goal, as it
turns out, is the kitchen; and the solution path is shown in Figure 7-1. The bedrooms, kitchen, and the bath are
terminal nodes because they lead nowhere. Heuristics are not represented on a graph. Rather, they are techniques
that you might employ to help you better choose a path.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

Combinatorial Explosions

Given the preceding example, you may think that searching for a solution is easy—you start at the beginning and work
your way to the conclusion. In the extremely simple case of the lost keys, this is not a bad approach because the
search space is so small. But for many problems (especially those for which you would want to use a computer), the
number of nodes in the search space is very large, and as the search space grows, so does the number of possible
paths to the goal. The trouble is that often adding another node to the search space adds more than one path. That is,
the number of potential pathways to the goal can increase in a nonlinear fashion as the size of the search space
grows. In a nonlinear situation, the number of possible paths can quickly become very large.

For instance, consider the number of ways three objects—A, B, and C—can be arranged on a table. The six possible
permutations are

A B C

籍
A C B

B C A

典 书 书
A

A
C

+ 经 不 读
You can quickly prove to yourself that these six are the only ways that A, B, and C can be arranged. However, you can
derive the same number by using a theorem from the branch of mathematics called combinatorics—the study of the

+
way that things can be combined. According to the theorem, the number of ways that N objects can be arranged is
equal to N! (N factorial). The factorial of a number is the product of all whole numbers equal to or less than itself down

C
to 1. Therefore, 3! is 3 x 2 x 1, or 6. If you had four objects to arrange, there would be 4!, or 24, permutations. With five

典
objects, the number is 120, and with six it is 720. With 1000 objects the number of possible permutations is huge! The
graph in Figure 7-2 gives you a visual feel for what is sometimes referred to as a combinatoric explosion. Once there

经
are more than a handful of possibilities, it very quickly becomes difficult to examine (indeed, even to enumerate) all the
arrangements.

无
 书 籍
经 典 读 书
+ + 不
C 典
经
Figure 7-2: A combinatoric explosion with factorials

无
This same sort of combinatorial explosion can occur in paths through search spaces. Because of this, only the
simplest of problems lend themselves to exhaustive searches. An exhaustive search is one that examines all nodes.
Thus, it is a “brute-force” technique. Brute force always works, but it is not often practical for large problems because it
consumes too much time, too many computing resources, or both. For this reason, AI-based search techniques were
developed.
 Search Techniques

There are several ways to search for a solution. The four most fundamental are

Depth first

Breadth first

Hill climbing

Least cost

In the course of this chapter, each of these searches is examined.

Evaluating a Search

籍
Evaluating the performance of an AI-based search technique can be complicated. Fortunately, for the purposes of this
chapter, we are concerned with only these two measurements:

书
How quickly is a solution found?

How good is the solution?

经 典 读 书
There are several types of problems for which all that matters is that a solution, any solution, be found with the
minimum effort. For these problems, the first measurement is especially important. In other situations, the quality of the
solution is more important.

+ 不
The speed of a search is affected both by the size of the search space and by the number of nodes actually traversed

+
in the process of finding the solution. Because backtracking from dead ends is wasted effort, you want a search that
seldom retraces its steps.

C 典
In AI-based searching, there is a difference between finding the best solution and finding a good solution. Finding the
best solution can require an exhaustive search because sometimes this is the only way to know that the best solution

经
has been found. Finding a good solution, in contrast, means finding a solution that is within a set of constraints—it
does not matter if a better solution might exist.

无
As you will see, the search techniques described in this chapter all work better in certain situations than in others. It is
difficult to say whether one search method is always superior to another, but some search techniques have a greater
probability of being better for the average case. In addition, the way a problem is defined can sometimes help you
choose an appropriate search method.
.

The Problem
Now, let us consider the problem that we will use various searches to solve. Imagine that you are a travel agent, and a
rather quarrelsome customer wants you to book a flight from New York to Los Angeles with XYZ Airlines. You try to tell
the customer that XYZ does not have a direct flight from New York to Los Angeles, but the customer insists that XYZ is
the only airline that he will fly. Thus, you must find connecting flights between New York and Los Angeles. You consult
XYZ’s scheduled flights, shown here:

Flight Distance

New York to Chicago 900 miles

Chicago to Denver 1000 miles

New York to Toronto 500 miles

籍
New York to Denver 1800 miles

Toronto to Calgary 1700 miles

书
Toronto to Los Angeles 2500 miles

Toronto to Chicago 500 miles

经 典
Denver to Urbana

读 书
Denver to Houston
1000 miles

1000 miles

+
Houston to Los Angeles 1500 miles

+ 不
Denver to Los Angeles 1000 miles

C
You quickly see that there are connections that enable your customer to fly from New York to Los Angeles. The

典
problem is to write C++ programs that do the same thing that you just did in your head!

经
A Graphic Representation

The flight information in XYZ’s schedule can be translated into the directed graph shown in Figure 7-3. A directed

无
graph is simply a graph in which the lines connecting each node include an arrow to indicate the direction of motion. In
a directed graph, you cannot travel against the direction of the arrow.
 籍
Figure 7-3: A directed graph of XYZ’s flight schedule

To make things easier to understand, this graph is redrawn in a tree-like fashion in Figure 7-4. Refer to this version for

书
the rest of this chapter. The goal, Los Angeles, is circled. Also notice that various cities appear more than once to
simplify the construction of the graph. Thus, the tree-like representation does not depict a binary tree. It is simply a
visual convenience.

经 典 读 书
+ + 不
C 典
无 经
Figure 7-4: A tree version of XYZ’s flight schedule

Now we are ready to develop the various search techniques that will find routes from New York to Los Angeles.
.

The FlightInfo Structure and the Search Class

Writing a program to find a route from New York to Los Angeles requires a database that contains the flight
information. Each entry in the database must contain the departure and destination cities, the distance between them,
and a flag that aids in backtracking. This information is held in a structure called FlightInfo, shown here:
// Flight information.
struct FlightInfo {
string from; // departure city
string to; // destination city
int distance; // distance between from and to
bool skip; // used in backtracking

FlightInfo() {
from = "";
to = "";

籍
distance = 0;
skip = false;

书
}

FlightInfo(string f, string t, int d) {

典 书
from = f;
to = t;
distance = d;

经 读
skip = false;
}

+
};

+ 不
This structure will be used by all of the search techniques described in the remainder of the chapter.

C
The AI-based searches are encapsulated within a class called Search . The precise implementation of this class will

典
change from one search technique to the next. However, all will share the same general framework, shown here:
// An AI-based search class.

经
class Search {
// This vector holds the flight information.
vector<FlightInfo> flights;

无
// This stack is used for backtracking.
stack<FlightInfo> btStack;

// If there is a flight between from and to,

// store the distance of the flight in dist.
// Return true if the flight exists and,
// false otherwise.
bool match(string from, string to, int &dist);

// Given from, find any connection.

// Return true if a connection is found,
// and false otherwise.
bool find(string from, FlightInfo &f);

public:

// Put flights into the database.

void addflight(string from, string to, int dist) {
flights.push_back(FlightInfo(from, to, dist));
}
 // Show the route and total distance.
void route();

// Determine if there is a route between from and to.

void findroute(string from, string to);

// Return true if a route has been found.

bool routefound() {
return !btStack.empty();
}
};

Search declares two private instance variables. The first is a vector of FlightInfo objects, called flights, that holds the
flight data. (Recall that a vector is an STL container that implements a dynamic array.) The second is a stack, called
btStack, that is used for backtracking. As you will see, the backtrack stack is very important to all of the search
techniques.

Search contains two in-line functions: addflight( ) and routefound( ). When a Search object is first created, its flights
vector is empty. Connections are added by repeatedly calling addflight( ), specifying the departure city, the destination

籍
city, and the distance between the two. The addflight( ) function simply pushes each connection onto the end of the
flights vector, which automatically expands in size to accommodate each new entry.

书
The routefound( ) function determines if a route exists between the departure and destination cities based on the
contents of the backtrack stack. If the backtrack stack is empty, no route exists between the two cities. Otherwise, the

典 书
backtrack stack contains the route. (The process of constructing the route varies from one search technique to the
next.)

经 读
The implementation of the remaining members of Search are described in the sections that follow. The following table
briefly explains their purpose:

+ +
match( )

不
Determines if there is a direct connection between two cities

C
find( ) Attempts to find a connection from a specified city to anywhere else

典
findroute( Attempts to construct a route from the city of departure to the city of destination
)

经
route( ) Displays the route

无
.

The Depth-First Search

The depth-first search explores each possible path to its conclusion before another path is tried. To understand exactly
how this works, consider the tree that follows. F is the goal.

书 籍
A depth-first search traverses the graph in the following order: ABDBEBACF. If you are familiar with trees, you
recognize this type of search as an inorder tree traversal. That is, the path goes left until a terminal node is reached or
the goal is found. If a terminal node is reached, the path backs up one level, goes right, and then left until either the

典 书
goal or a terminal node is encountered. This procedure is repeated until the goal is found or the last node in the search
space has been examined.

经 读
As you can see, a depth-first search is certain to find the goal because in the worst case it degenerates into an
exhaustive search. In this example, an exhaustive search would result if G were the goal.

+ 不
The entire depth-first search program follows:

+
// Search for a route.

C
#include <iostream>

典
#include <stack>
#include <string>
#include <vector>

经
using namespace std;

无
// Flight information.
struct FlightInfo {
string from; // departure city
string to; // destination city
int distance; // distance between from and to
bool skip; // used in backtracking

FlightInfo() {
from = "";
to = "";
distance = 0;
skip = false;
}

FlightInfo(string f, string t, int d) {

from = f;
to = t;
distance = d;
skip = false;
}
};
.

// Find connections using a depth-first search.

class Search {
// This vector holds the flight information.
vector<FlightInfo> flights;

// This stack is used for backtracking.

stack<FlightInfo> btStack;

// If there is a flight between from and to,

// store the distance of the flight in dist.
// Return true if the flight exists and
// false otherwise.
bool match(string from, string to, int &dist);

// Given from, find any connection.

// Return true if a connection is found,
// and false otherwise.
bool find(string from, FlightInfo &f);

籍
public:

// Put flights into the database.

书
void addflight(string from, string to, int dist) {
flights.push_back(FlightInfo(from, to, dist));

典 书
}

// Show the route and total distance.

经 读
void route();

+
// Determine if there is a route between from and to.

不
void findroute(string from, string to);

} +
// Return true if a route has been found.

C 典
bool routefound() {
return !btStack.empty();

经
};

// Show the route and total distance.

无
void Search::route()
{
stack<FlightInfo> rev;
int dist = 0;
FlightInfo f;

// Reverse the stack to display route.

while(!btStack.empty()) {
f = btStack.top();
rev.push(f);
btStack.pop();
}

// Display the route.

while(!rev.empty()) {
f = rev.top();
rev.pop();
cout << f.from << " to ";
dist += f.distance;
}

cout << f.to << endl;

.

cout << "Distance is " << dist << endl;

}

// If there is a flight between from and to,

// store the distance of the flight in dist.
// Return true if the flight exists and,
// false otherwise.
bool Search::match(string from, string to, int &dist)
{
for(unsigned i=0; i < flights.size(); i++) {
if(flights[i].from == from &&
flights[i].to == to && !flights[i].skip)
{
flights[i].skip = true; // prevent reuse
dist = flights[i].distance;
return true;
}
}

return false; // not found

籍
}

// Given from, find any connection.

书
// Return true if a connection is found,
// and false otherwise.

典 书
bool Search::find(string from, FlightInfo &f)
{
for(unsigned i=0; i < flights.size(); i++) {

经 读
if(flights[i].from == from && !flights[i].skip) {
f = flights[i];

+
flights[i].skip = true; // prevent reuse

+ 不
return true;

C 典
}

return false;

经
}

// Depth-first version.

无
// Determine if there is a route between from and to.
void Search::findroute(string from, string to)
{
int dist;
FlightInfo f;

// See if at destination.
if(match(from, to, dist)) {
btStack.push(FlightInfo(from, to, dist));
return;
}

// Try another connection.

if(find(from, f)) {
btStack.push(FlightInfo(from, to, f.distance));
findroute(f.to, to);
}
else if(!btStack.empty()) {
// Backtrack and try another connection.
f = btStack.top();
btStack.pop();
findroute(f.from, f.to);
.

}
}

int main() {
char to[40], from[40];
Search ob;

// Add flight connections to database.

ob.addflight("New York", "Chicago", 900);
ob.addflight("Chicago", "Denver", 1000);
ob.addflight("New York", "Toronto", 500);
ob.addflight("New York", "Denver", 1800);
ob.addflight("Toronto", "Calgary", 1700);
ob.addflight("Toronto", "Los Angeles", 2500);
ob.addflight("Toronto", "Chicago", 500);
ob.addflight("Denver", "Urbana", 1000);
ob.addflight("Denver", "Houston", 1000);
ob.addflight("Houston", "Los Angeles", 1500);
ob.addflight("Denver", "Los Angeles", 1000);

籍
// Get departure and destination cities.
cout << "From? ";

书
cin.getline(from, 40);
cout << "To? ";

典 书
cin.getline(to, 40);

经 读
// See if there is a route between from and to.
ob.findroute(from, to);

+ + 不
// If there is a route, show it.
if(ob.routefound())
ob.route();

C
return 0;

典
经
Notice that main( ) prompts you for both the city of origin and the city of destination. This means that you can use the
program to find routes between any two cities. However, the rest of this chapter assumes that New York is the origin

无
and Los Angeles is the destination. When the program is run with New York as the origin and Los Angeles as the
destination, the following solution is displayed:
From? New York
To? Los Angeles
New York to Chicago to Denver to Los Angeles
Distance is 2900

If you refer to Figure 7-5, you see that this is indeed the first solution that would be found by a depth-first search. It is
also a fairly good solution.
.

书 籍
典 书
Figure 7-5: The depth-first path to a solution

As main( ) illustrates, to use Search , first create a Search object. Then load the flight database with the connections.

经 读
Next, call findroute( ) to attempt to find a route between the point of departure and the destination. To determine if
such a route has been found, call routefound( ). It returns true if a route exists. To display the route, call route( ). Let’s

+ 不
now examine each piece of the depth-first search in detail.

+
The match( ) Function

C 典
The match( ) function, shown here, determines if there is a flight between two cities, specified by from and to:
// If there is a flight between from and to,

经
// store the distance of the flight in dist.
// Return true if the flight exists and
// false otherwise.

无
bool Search::match(string from, string to, int &dist)
{
for(unsigned i=0; i < flights.size(); i++) {
if(flights[i].from == from &&
flights[i].to == to && !flights[i].skip)
{
flights[i].skip = true; // prevent reuse
dist = flights[i].distance;
return true;
}
}

return false; // not found

}

The match( ) function operates by searching the flights vector, looking for any entry in which the departure and
destination cities match those passed through from and to. It returns false if no such flight exists. It returns true if there
is a flight, in which case it also obtains the distance between the two cities and stores this value in the variable
referenced by its dist parameter.

Notice that match( ) ignores connections that have the skip field set to true. Also, if a connection is found, its skip field
.

is set. This manages backtracking from dead ends, preventing the same connections from being tried over and over
again.

The find( ) Function

The find( ) function, shown here, searches the database for any connection from a given departure city:
// Given from, find any connection.
// Return true if a connection is found,
// and false otherwise.
bool Search::find(string from, FlightInfo &f)
{
for(unsigned i=0; i < flights.size(); i++) {
if(flights[i].from == from && !flights[i].skip) {
f = flights[i];
flights[i].skip = true; // prevent reuse

return true;
}
}

}
return false;

书 籍
The departure city is passed to from. If a connecting flight is found, find( ) returns true, and the FlightInfo object
associated with that connection is stored in the variable referenced by the second parameter, f . Otherwise, find( )

典 书
returns false. The difference between match( ) and find( ) is that match( ) determines if there is a flight between two
specific cities; find( ) determines if there is a flight from a given city to any other city.

经 读
Like match( ), find( ) also use the skip field to manage backtracking from dead ends.

+ 不
The findroute( ) Function

+
The code that actually finds the connecting flights is contained in the findroute( ) function, the key routine in finding a

C 典
route between two cities. It is called with the names of the departure and destination cities.
// Determine if there is a route between from and to
// by using depth-first searching.

经
void Search::findroute(string from, string to)
{
int dist;

无
FlightInfo f;

// See if at destination.
if(match(from, to, dist)) {
btStack.push(FlightInfo(from, to, dist));
return;
}

// Try another connection.

if(find(from, f)) {
btStack.push(FlightInfo(from, to, f.distance));
findroute(f.to, to);
}
else if(!btStack.empty()) {
// Backtrack and try another connection.
f = btStack.top();
btStack.pop();
findroute(f.from, f.to);
}
}
.

Let’s examine findroute( ) closely. First, the flight database is checked by match( ) to see if there is a flight between
from and to. If there is, the goal has been reached, the connection is pushed onto the stack, and the function returns.
Otherwise, findroute( ) calls find( ) to find a connection between from and any place else. The find( ) function returns
true if such a connection is found, in which case it stores the FlightInfo object describing the connection in f . If no
connecting flight is available, find( ) returns false. If there is a connecting flight, the current flight is pushed onto the
backtrack stack, and findroute( ) is called recursively, with the city in f.to becoming the new departure city. Otherwise,
backtracking takes place. The previous node is removed from the stack and findroute( ) is called recursively. This
process continues until the goal is found, or the database is exhausted.

For example, if findroute( ) is called with New York and Chicago, the first if would succeed and findroute( ) would
terminate because there is a direct flight from New York to Chicago. The situation is more complex when findroute( ) is
called with New York and Calgary. In this case, the first if would fail because there is no direct flight connecting these
two cities. Next, the second if is tried by attempting to find a connection between New York and any other city. In this
case, find( ) first finds the New York to Chicago connection, this connection is pushed onto the backtrack stack, and
findroute( ) is called recursively with Chicago as the starting point. Unfortunately, there is no path from Chicago to
Calgary and several false paths are followed. Eventually, after several recursive calls to findroute( ) and substantial
backtracking, the connection from New York to Toronto is found, and Toronto connects to Calgary. This causes
findroute( ) to return, unraveling all recursive calls in the process. Finally, the original call to findroute( ) returns. You
might want to try adding cout statements in findroute( ) to see precisely how it works with various other departure and

籍
destination cities.

书
It is important to understand that findroute( ) does not actually return the solution—it generates it. Upon exit from
findroute( ), the backtrack stack contains the route between the point of departure and the destination. That is, the
solution is contained in btStack. Furthermore, the success or failure of findroute( ) is determined by the state of the

典 书
stack. An empty stack indicates failure; otherwise, the stack holds a solution.

In general, backtracking is a crucial ingredient in AI-based search techniques. In Search , backtracking is accomplished

经 读
through the use of recursion and a backtrack stack. (This is why C++’s support for recursion and the STL make it good
choice for AI development.) Almost all backtracking situations are stack-like in operation—that is, they are first-in,

+ 不
last-out. As a path is explored, nodes are pushed onto the stack as they are encountered. At each dead end, the last

+
node is popped off the stack and a new path, from that point, is tried. This process continues until either the goal is
reached or all paths have been exhausted.

C 典
Displaying the Route

经
The route( ) function displays the path and the total distance. It is shown here:
// Show the route and total distance.
void Search::route()

无
{
stack<FlightInfo> rev;
int dist = 0;
FlightInfo f;

// Reverse the stack to display route.

while(!btStack.empty()) {
f = btStack.top();
rev.push(f);
btStack.pop();
}

// Display the route.

while(!rev.empty()) {
f = rev.top();
rev.pop();
cout << f.from << " to ";
dist += f.distance;
}

cout << f.to << endl;

 cout << "Distance is " << dist << endl;
}

Notice the use of a second stack called rev. The solution stored in btStack is in reverse order, with the top of the stack
holding the last connection, the bottom of the stack holding the first connection. Thus, it must be reversed in order to
display the connection in the proper sequence. To put the solution into its proper order, the connections are popped
from btStack and pushed onto rev.

An Analysis of the Depth-First Search

The depth-first approach found a good solution. Also, relative to this specific problem, depth-first searching found this
solution on its first try with no backtracking—this is very good. However, had the data been organized differently,
finding a solution might have involved considerable backtracking. Thus, the outcome of this example cannot be
generalized. Moreover, the performance of depth-first searches can be quite poor when a particularly long branch with
no solution at the end is explored. In this case, a depth-first search wastes time not only exploring this chain, but also
backtracking to the goal.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

The Breadth-First Search

The opposite of the depth-first search is the breadth-first search. In this method, each node on the same level is
checked before the search proceeds to the next deeper level. This traversal method is shown here with C as the goal:

书 籍
Although thinking in terms of a binary tree-structured search space makes it easy to describe the actions of a
breadth-first search, many search spaces, including the flight example, are not binary trees. Therefore, precisely what
constitutes “breadth” is a bit subjective, being defined by the problem at hand. As it relates to the flight example, the

典 书
breadth-first approach is implemented by checking if any flight leaving the departure city connects to a flight that
reaches the destination. In other words, before advancing to another level, the destinations of all connections of

读
connecting flights are checked.

+ 经
To make Search perform a breadth-first search, you need to make an alteration to findroute( ), as shown here:

不
// Breadth-first version.

+
// Determine if there is a route between from and to.
void Search::findroute(string from, string to)

C
{

典
int dist;
FlightInfo f;

经
// This stack is needed by the breadth-first search.
stack<FlightInfo> resetStck;

无
// See if at destination.
if(match(from, to, dist)) {
btStack.push(FlightInfo(from, to, dist));
return;
}

// Following is the first part of the breadth-first

// modification. It checks all connecting flights
// from a specified node.
while(find(from, f)) {
resetStck.push(f);
if(match(f.to, to, dist)) {
resetStck.push(FlightInfo(f));
btStack.push(FlightInfo(from, f.to, f.distance));
btStack.push(FlightInfo(f.to, to, dist));
return;
}
}

// The following code resets the skip fields set by

// preceding while loop. This is also part of the
.

// breadth-first modification.
while(!resetStck.empty()) {
resetSkip(resetStck.top());
resetStck.pop();
}

// Try another connection.

if(find(from, f)) {
btStack.push(FlightInfo(from, to, f.distance));
findroute(f.to, to);
}
else if(!btStack.empty()) {
// Backtrack and try another connection.
f = btStack.top();
btStack.pop();
findroute(f.from, f.to);
}
}

Two changes have been made. First, the for loop checks all flights leaving from the departure city (from) to see if they

籍
connect with flights that arrive at the destination city. Second, if the destination is not found, the skip fields of those
connecting flights are cleared by calling resetSkip( ), which is a new function that must be added to Search . The
connections that need to be reset are stored on their own stack, called resetStck, which is local to findroute( ).

书
Resetting the skip flags is necessary to enable alternative paths that might involve those connections.

典 书
The resetSkip( ) function is shown here:
// Reset the skip fields in flights vector.
void Search::resetSkip(FlightInfo f) {

经 读
for(unsigned i=0; i < flights.size(); i++)
if(flights[i].from == f.from &&

+
flights[i].to == f.to)

不
flights[i].skip = false;

+
}

C 典
You need to add this function (and its prototype) to Search .

To try the breadth-first search, substitute the new version of findroute( ) into the preceding search program and then

经
add the resetSkip( ) function. When run, it produces the following solution:
From? New York
To? Los Angeles

无
New York to Toronto to Los Angeles
Distance is 3000

Figure 7-6 shows the breadth-first path to the solution.

 书 籍
典 书
Figure 7-6: The breadth-first path to a solution

An Analysis of the Breadth-First Search

+ 经 不 读
In this example, the breadth-first search performed fairly well, finding a reasonable solution. As before, this result
cannot be generalized because the first path to be found depends upon the physical organization of the information.

+
The example does illustrate, however, how depth-first and breadth-first searches often find different paths through the
same search space.

C 典
Breadth-first searching works well when the goal is not buried too deeply in the search space. It works poorly when the
goal is several layers deep. In this case, a breadth-first search expends substantial effort during the backtrack stage.

无 经
.

Adding Heuristics
Neither the depth-first nor the breadth-first search attempts to make any educated guesses about whether one node in
the search space is closer to the goal than another. Instead, they simply move from one node to the next using a
prescribed pattern until the goal is finally found. This may be the best you can do for some situations, but often a
search space contains information that you can use to increase the probability that a search will reach its goal faster.
To take advantage of such information, you must add heuristic capabilities to the search.

Heuristics are simply rules that increase the likelihood that a search is proceeding in the correct direction. For
example, imagine that you are lost in the woods and need a drink of water. The woods are so thick that you cannot
see far ahead, and the trees are too big to climb to get a look around. However, you know that rivers, streams, and
ponds are most likely in valleys; that animals frequently make paths to their watering places; that when you are near
water it is possible to “smell” it; and that you can hear running water. So, you begin by moving downhill because water
is unlikely to be uphill. Next you come across a deer trail that also runs downhill. Knowing that this may lead to water,
you follow it. You begin to hear a slight rushing off to your left. Knowing that this may be water, you cautiously move in
that direction. As you move, you begin to detect the increased humidity in the air; you can smell the water. Finally, you

籍
find a stream and have your drink. In this situation, the heuristic information used to find the water did not guarantee
success, but it did increase the probability of an early success. In general, heuristics improve the odds in favor of

书
quickly finding the goal.

Most often, heuristic search methods are based on maximizing or minimizing some constraint. In the problem of

典 书
scheduling a flight from New York to Los Angeles, there are two possible constraints that a passenger may want to
minimize. The first is the number of connections that have to be made. The second is the length of the route.

读
Remember, the shortest route does not necessarily imply the fewest connections, or vice versa. In this section, two

经
heuristic searches are developed. The first minimizes the number of connections. The second minimizes the length of

+
the route. Both heuristic searches are built upon the depth-first search framework.

+ 不
The Hill-Climbing Search

C 典
A search algorithm that attempts to find a route that minimizes the number of connections uses the heuristic that the
longer the length of the flight, the greater the likelihood that it takes the traveler closer to the destination; therefore, the
number of connections is minimized. In the language of AI, this is an example of hill climbing.

经
The hill-climbing algorithm chooses as its next step the node that appears to place it closest to the goal (that is,
farthest away from the current position). It derives its name from the analogy of a hiker being lost in the dark, halfway

无
up a mountain. Assuming that the hiker’s camp is at the top of the mountain, even in the dark the hiker knows that
each step that goes up is a step in the right direction.

Working only with the information contained in the flight-scheduling database, here is how to incorporate the
hill-climbing heuristic into the routing program: Choose the connecting flight that is as far away as possible from the
current position in the hope that it will be closer to the destination. To do this, begin with the depth-first version of
Search and modify the find( ) routine, as shown here:

// Hill-climbing version.
// Given from, find the farthest away connection.
// Return true if a connection is found,
// and false otherwise.
bool Search::find(string from, FlightInfo &f)
{
int pos = -1;
int dist = 0;

for(unsigned i=0; i < flights.size(); i++) {

if(flights[i].from == from && !flights[i].skip) {
// Use the longest flight.
if(flights[i].distance > dist) {
pos = i;
.

dist = flights[i].distance;
}
}
}

if(pos != -1) {
f = flights[pos];
flights[pos].skip = true; // prevent reuse

return true;
}

return false;
}

The find( ) routine now searches the entire database, looking for the connection that is farthest away from the
departure city.

For the sake of clarity, the entire hill-climbing program follows:

// Search for a connection by hill climbing.

籍
#include <iostream>
#include <stack>
#include <string>

书
#include <vector>

典 书
using namespace std;

// Flight information.

经 读
struct FlightInfo {
string from; // departure city

+
string to; // destination city

不
int distance; // distance between from and to

+
bool skip; // used in backtracking

C 典
FlightInfo() {
from = "";
to = "";

经
distance = 0;
skip = false;
}

}
to = t;
distance = d;
skip = false;
无
FlightInfo(string f, string t, int d) {
from = f;

};

// This version of search finds connections

// through the heuristic of hill climbing.
class Search {
// This vector holds the flight information.
vector<FlightInfo> flights;

// This stack is used for backtracking.

stack<FlightInfo> btStack;

// If there is a flight between from and to,

// store the distance of the flight in dist.
// Return true if the flight exists and
// false otherwise.
.

bool match(string from, string to, int &dist);

// Hill-climbing version.
// Given from, find the farthest away connection.
// Return true if a connection is found,
// and false otherwise.
bool find(string from, FlightInfo &f);

public:

// Put flights into the database.

void addflight(string from, string to, int dist) {
flights.push_back(FlightInfo(from, to, dist));
}

// Show the route and total distance.

void route();

// Determine if there is a route between from and to.

void findroute(string from, string to);

书 籍
// Return true if a route has been found.
bool routefound() {
return btStack.size() != 0;

典 书
};

// Show the route and total distance.

经 读
void Search::route()
{

+
stack<FlightInfo> rev;

不
int dist = 0;

+
FlightInfo f;

C 典
// Reverse the stack to display route.
while(!btStack.empty()) {
f = btStack.top();

经
rev.push(f);
btStack.pop();
}

无
// Display the route.
while(!rev.empty()) {
f = rev.top();
rev.pop();
cout << f.from << " to ";
dist += f.distance;
}

cout << f.to << endl;

cout << "Distance is " << dist << endl;
}

// If there is a flight between from and to,

// store the distance of the flight in dist.
// Return true if the flight exists and
// false otherwise.
bool Search::match(string from, string to, int &dist)
{
for(unsigned i=0; i < flights.size(); i++) {
if(flights[i].from == from &&
flights[i].to == to && !flights[i].skip)
.

{
flights[i].skip = true; // prevent reuse
dist = flights[i].distance;
return true;
}
}

return false; // not found

}

// Hill-climbing version.
// Given from, find the farthest away connection.
// Return true if a connection is found,
// and false otherwise.
bool Search::find(string from, FlightInfo &f)
{
int pos = -1;
int dist = 0;

for(unsigned i=0; i < flights.size(); i++) {

籍
if(flights[i].from == from && !flights[i].skip) {
// Use the longest flight.
if(flights[i].distance > dist) {

书
pos = i;
dist = flights[i].distance;

典 书
}
}
}

+ 经
if(pos != -1) {
f = flights[pos];

不 读
flights[pos].skip = true; // prevent reuse

+
return true;

C
return false;

典
经
}

// Determine if there is a route between from and to.

无
void Search::findroute(string from, string to)
{
int dist;
FlightInfo f;

// See if at destination.
if(match(from, to, dist)) {
btStack.push(FlightInfo(from, to, dist));
return;
}

// Try another connection.

if(find(from, f)) {
btStack.push(FlightInfo(from, to, f.distance));
findroute(f.to, to);
}
else if(!btStack.empty()) {
// Backtrack and try another connection.
f = btStack.top();
btStack.pop();
findroute(f.from, f.to);
}
.

int main() {
char to[40], from[40];
Search ob;

// Add flight connections to database.

ob.addflight("New York", "Chicago", 900);
ob.addflight("Chicago", "Denver", 1000);
ob.addflight("New York", "Toronto", 500);
ob.addflight("New York", "Denver", 1800);
ob.addflight("Toronto", "Calgary", 1700);
ob.addflight("Toronto", "Los Angeles", 2500);
ob.addflight("Toronto", "Chicago", 500);
ob.addflight("Denver", "Urbana", 1000);
ob.addflight("Denver", "Houston", 1000);
ob.addflight("Houston", "Los Angeles", 1500);
ob.addflight("Denver", "Los Angeles", 1000);

// Get departure and destination cities.

籍
cout << "From? ";

cin.getline(from, 40);

书
cout << "To? ";

典 书
cin.getline(to, 40);

// See if there is a route between from and to.

经 读
ob.findroute(from, to);

+
// If there is a route, show it.

不
if(ob.routefound())

+
ob.route();

C 典
return 0;
}

经
When the program is run, the solution is
From? New York
To? Los Angeles

无
New York to Denver to Los Angeles
Distance is 2800

This is quite good! The route contains the minimal number of stops on the way (only one), and it is the shortest route.
Thus, it found the best possible route.

However, if the Denver to Los Angeles connection did not exist, the solution would not be quite so good. It would be
New York to Denver to Houston to Los Angeles—a distance of 4300 miles! In this case, the solution climbs a “false
peak,” because the connection to Houston does not take us closer to the goal of Los Angeles. Figure 7-7 shows the
first solution as well as the path to the false peak.
 书 籍
典 书
Figure 7-7: The hill-climbing path to a solution and to a false peak

读
An Analysis of Hill Climbing

+ 经
Hill climbing provides fairly good solutions in many circumstances because it tends to reduce the number of nodes that

不
need to be visited before a solution is found. However, it can suffer from three maladies. First, there is the problem of

+
false peaks, as just described. In this case, extensive backtracking may result. The second problem relates to
plateaus, a situation in which all next steps look equally good (or bad). In this case, hill climbing is no better than

C 典
depth-first searching. The final problem is that of a ridge. In this case, hill climbing really performs poorly because the
algorithm causes the ridge to be crossed several times as backtracking occurs. In spite of these potential troubles, hill
climbing often increases the probability of finding a good solution.

无 经
.

The Least-Cost Search

The opposite of a hill-climbing search is a least-cost search. This strategy is similar to standing in the middle of a street
on a big hill while wearing roller skates. You have the definite feeling that it’s a lot easier to go down rather than up! In
other words, a least-cost search takes the path of least resistance.

Applying a least-cost search to the flight-scheduling problem implies that the shortest connecting flight is taken in all
cases so that the route found has a good chance of covering the shortest distance. Unlike hill climbing, which attempts
to minimize the number of connections, a least-cost search attempts to minimize the number of miles.

To use a least-cost search, you must alter find( ) in the previous program, as shown here:
const int MAXDIST = 100000;

// Least-cost version.
// Given from, find the closest connection.

籍
// Return true if a connection is found,
// and false otherwise.
bool Search::find(string from, FlightInfo &f)

书
{
int pos = -1;

典 书
int dist = MAXDIST; // longer than longest flight

for(unsigned i=0; i < flights.size(); i++) {

经 读
if(flights[i].from == from && !flights[i].skip) {
// Use the shortest flight.

+
if(flights[i].distance < dist) {

不
pos = i;

+
dist = flights[i].distance;
}

C 典
}
}

经
if(pos != -1) {
f = flights[pos];
flights[pos].skip = true; // prevent reuse

}
}

无
return true;

return false;

Using this version of find( ), the solution is

From? New York
To? Los Angeles
New York to Toronto to Los Angeles
Distance is 3000

As you can see, the search found a good route—not the best, but acceptable. Figure 7-8 shows the least-cost path to
the goal.
 书 籍
典 书
Figure 7-8: The least-cost path to a solution

An Analysis of the Least-Cost Search

+ 经 不 读
Least-cost searches and hill climbing have the same advantages and disadvantages, but in reverse. In the least-cost
search there can be false valleys, lowlands, and gorges. In this specific case, it worked about as well as hill climbing.

C+ 典
无 经
.

Finding Multiple Solutions

Sometimes it is valuable to find several solutions to the same problem. This is not the same as finding all solutions (an
exhaustive search). Instead, multiple solutions offer a representative sample of the solutions present in the search
space.

There are several ways to generate multiple solutions, but only two are examined here. The first is path removal, and
the second is node removal. As their names imply, generating multiple solutions without redundancy requires that
already found solutions be removed from the system. Remember that neither of these techniques attempts to find all
solutions. Finding all solutions is a different problem that is usually not attempted because it implies an exhaustive
search.

Path Removal

The path-removal method of generating multiple solutions removes all nodes from the database that form the current
solution and then attempts to find another solution. In essence, path removal prunes limbs from the tree. To find

籍
multiple solutions by using path removal, you just need to alter main( ), as shown here:
// Path removal version.

书
int main() {
char to[40], from[40];
Search ob;

经 典 读 书
// Add flight connections to database.
ob.addflight("New York", "Chicago", 900);
ob.addflight("Chicago", "Denver", 1000);

+
ob.addflight("New York", "Toronto", 500);

不
ob.addflight("New York", "Denver", 1800);

+
ob.addflight("Toronto", "Calgary", 1700);
ob.addflight("Toronto", "Los Angeles", 2500);

C
ob.addflight("Toronto", "Chicago", 500);

典
ob.addflight("Denver", "Urbana", 1000);
ob.addflight("Denver", "Houston", 1000);
ob.addflight("Houston", "Los Angeles", 1500);

经
ob.addflight("Denver", "Los Angeles", 1000);

无
// Get departure and destination cities.
cout << "From? ";

cin.getline(from, 40);
cout << "To? ";

cin.getline(to, 40);

// Find multiple solutions.

for(;;) {
// See if there is a connection.
ob.findroute(from, to);

// If no new route was found, then end.

if(!ob.routefound()) break;

ob.route();
}

return 0;
}
.

Here, a for loop is added that iterates until the backtrack stack is empty. Recall that when the backtrack stack is
empty, no solution (in this case, no additional solution) has been found. No other modifications are needed because
any connection that is part of a solution will have its skip field marked. Consequently, such a connection can no longer
be found by find( ) and cannot be part of the next solution; it cannot be refound.

If you use the original depth-first version of Search (shown first in this chapter) with the preceding main( ), the following
solutions will be found:
From? New York
To? Los Angeles
New York to Chicago to Denver to Los Angeles
Distance is 2900
New York to Toronto to Los Angeles
Distance is 3000
New York to Denver to Houston to Los Angeles
Distance is 4300

The search found the three solutions. Notice, however, that none are the best solution.

Node Removal

书 籍
The second way to force the generation of additional solutions, node removal, simply removes the last node in the
current solution path and tries again. To do this, main( ) is changed so that it removes the last connection from the
flight database, clears all of the skip fields, and obtains a new, empty stack which is used to hold the next solution.
The updated main( ) function is shown here:

典 书
// Node removal version
int main() {
char to[40], from[40];

经 读
Search ob;
FlightInfo f;

+ 不
// Add flight connections to database.

+
ob.addflight("New York", "Chicago", 900);

C
ob.addflight("Chicago", "Denver", 1000);

典
ob.addflight("New York", "Toronto", 500);
ob.addflight("New York", "Denver", 1800);
ob.addflight("Toronto", "Calgary", 1700);

经
ob.addflight("Toronto", "Los Angeles", 2500);
ob.addflight("Toronto", "Chicago", 500);
ob.addflight("Denver", "Urbana", 1000);

无
ob.addflight("Denver", "Houston", 1000);
ob.addflight("Houston", "Los Angeles", 1500);
ob.addflight("Denver", "Los Angeles", 1000);

// Get departure and destination cities.

cout << "From? ";

cin.getline(from, 40);
cout << "To? ";

cin.getline(to, 40);

// Find multiple solutions.

for(;;) {
// See if there is a connection.
ob.findroute(from, to);

// If no new route was found, then end.

if(!ob.routefound()) break;

// Save the flight on top-of-stack.

.

f = ob.getTOS();

ob.route(); // display the current route.

ob.resetAllSkip(); // reset the skip fields

// Remove last flight in previous solution

// from the flight database.
ob.remove(f);
}

return 0;
}

To remove the last connection of the previous solution from the flight database, main( ) first obtains that connection by
calling getTOS( ), which is declared in-line within the Search class, as shown here:
// Return flight on top of stack.
FlightInfo getTOS() {
return btStack.top();
}

书 籍
It returns the flight information found on the top of the backtrack stack, which is the last connection in the route.

To actually remove this connection, main( ) calls remove( ), shown here:

// Remove a connection.

典 书
void Search::remove(FlightInfo f) {
for(unsigned i=0; i< flights.size(); i++)
if(flights[i].from == f.from &&

经 读
flights[i].to == f.to)
flights[i].from = "";

+
}

+ 不
Removing a connection is accomplished by assigning a zero-length string to the name of the departure city.

C
To clear the skip fields, main( ) calls resetAllSkip( ), shown next:

典
// Reset all skip fields.
void Search::resetAllSkip() {

经
for(unsigned i=0; i< flights.size(); i++)
flights[i].skip = false;
}

无
It simply sets the skip fields to false. (Remember to add the prototypes for resetAllSkip( ) and remove( ) to the Search
class declaration.)

Because so many changes are required, the entire node-removal program is shown here for the sake of clarity. Notice
that it too uses the original depth-first version of Search .
// Search for multiple routes by use of node removal.
#include <iostream>
#include <stack>
#include <string>
#include <vector>

using namespace std;

// Flight information.
struct FlightInfo {
string from; // departure city
string to; // destination city
int distance; // distance between from and to
bool skip; // used in backtracking
.

FlightInfo() {
from = "";
to = "";
distance = 0;
skip = false;
}

FlightInfo(string f, string t, int d) {

from = f;
to = t;
distance = d;
skip = false;
}
};

// Find multiple solutions via node removal.

class Search {
// This vector holds the flight information.
vector<FlightInfo> flights;

籍
// This stack is used for backtracking.
stack<FlightInfo> btStack;

书
// If there is a flight between from and to,
// store the distance of the flight in dist.

典 书
// Return true if the flight exists and
// false otherwise.
bool match(string from, string to, int &dist);

+ 经 读
// Given from, find any connection.
// Return true if a connection is found,

不
// and false otherwise.

+
bool find(string from, FlightInfo &f);

C 典
public:

// Put flights into the database.

经
void addflight(string from, string to, int dist) {
flights.push_back(FlightInfo(from, to, dist));
}

无
// Show the route and total distance.
void route();

// Determine if there is a route between from and to.

void findroute(string from, string to);

// Return true if a route has been found.

bool routefound() {
return btStack.size() != 0;
}

// Return flight on top of stack.

FlightInfo getTOS() {
return btStack.top();
}

// Reset all skip fields.

void resetAllSkip();

// Remove a connection.
void remove(FlightInfo f);
.

};

// Show the route and total distance.

void Search::route()
{
stack<FlightInfo> rev;
int dist = 0;
FlightInfo f;

// Reverse the stack to display route.

while(!btStack.empty()) {
f = btStack.top();
rev.push(f);
btStack.pop();
}

// Display the route.

while(!rev.empty()) {
f = rev.top();
rev.pop();

籍
cout << f.from << " to ";
dist += f.distance;
}

典 书
cout << f.to << endl;

书
cout << "Distance is " << dist << endl;

经 读
// If there is a flight between from and to,
// store the distance of the flight in dist.

+
// Return true if the flight exists and

不
// false otherwise.

+
bool Search::match(string from, string to, int &dist)
{

C 典
for(unsigned i=0; i < flights.size(); i++) {
if(flights[i].from == from &&
flights[i].to == to && !flights[i].skip)

经
{
flights[i].skip = true; // prevent reuse
dist = flights[i].distance;

无
return true;
}
}

return false; // not found

}

// Given from, find any connection.

// Return true if a connection is found,
// and false otherwise.
bool Search::find(string from, FlightInfo &f)
{
for(unsigned i=0; i < flights.size(); i++) {
if(flights[i].from == from && !flights[i].skip) {
f = flights[i];
flights[i].skip = true; // prevent reuse

return true;
}
}

return false;
.

// Determine if there is a route between from and to.

void Search::findroute(string from, string to)
{
int dist;
FlightInfo f;

// See if at destination.
if(match(from, to, dist)) {
btStack.push(FlightInfo(from, to, dist));
return;
}

// Try another connection.

if(find(from, f)) {
btStack.push(FlightInfo(from, to, f.distance));
findroute(f.to, to);
}
else if(!btStack.empty()) {

籍
// Backtrack and try another connection.
f = btStack.top();
btStack.pop();

书
findroute(f.from, f.to);
}

典 书
}

// Reset all skip fields.

经 读
void Search::resetAllSkip() {
for(unsigned i=0; i< flights.size(); i++)

+
flights[i].skip = false;

不
}

+
// Remove a connection.

C 典
void Search::remove(FlightInfo f) {
for(unsigned i=0; i< flights.size(); i++)
if(flights[i].from == f.from &&

经
flights[i].to == f.to)
flights[i].from = "";
}

无
// Node removal version.
int main() {
char to[40], from[40];
Search ob;
FlightInfo f;

// Add flight connections to database.

ob.addflight("New York", "Chicago", 900);
ob.addflight("Chicago", "Denver", 1000);
ob.addflight("New York", "Toronto", 500);
ob.addflight("New York", "Denver", 1800);
ob.addflight("Toronto", "Calgary", 1700);
ob.addflight("Toronto", "Los Angeles", 2500);
ob.addflight("Toronto", "Chicago", 500);
ob.addflight("Denver", "Urbana", 1000);
ob.addflight("Denver", "Houston", 1000);
ob.addflight("Houston", "Los Angeles", 1500);
ob.addflight("Denver", "Los Angeles", 1000);

// Get departure and destination cities.

 cout << "From? ";

cin.getline(from, 40);
cout << "To? ";

cin.getline(to, 40);

// Find multiple solutions.

for(;;) {
// See if there is a connection.
ob.findroute(from, to);

// If no new route was found, then end.

if(!ob.routefound()) break;

// Save the flight on top-of-stack.

f = ob.getTOS();

ob.route(); // display the current route.

籍
ob.resetAllSkip(); // reset the skip fields

// Remove last flight in previous solution

书
// from the flight database.
ob.remove(f);

典 书
}

return 0;

经 读
}

+
This program finds the following routes:

不
From? New York

+
To? Los Angeles
New York to Chicago to Denver to Los Angeles

C 典
Distance is 2900
New York to Chicago to Denver to Houston to Los Angeles
Distance is 4400

经
New York to Toronto to Los Angeles
Distance is 3000

无
In this case, the second solution is the worst possible route, but two fairly good solutions are also found. Notice that
the set of solutions found by the node-removal method differs from that found by the path-removal approach. Different
approaches to generating multiple solutions can often yield different results.
.

Finding the “Optimal” Solution

All of the previous search techniques were concerned, first and foremost, with finding a solution—any solution. As you
saw with the heuristic searches, efforts can be made to improve the likelihood of finding a good solution. But no
attempt was made to ensure that the optimal solution was found. However, at times you may want only the optimal
solution. Keep in mind that “optimal,” as it is used here, simply means the best route that can be found by using one of
the multiple-solution generation techniques—it may not actually be the best solution. (Finding the best solution would,
of course, require the prohibitively time-consuming exhaustive search.)

Before leaving the well-used flight scheduling example, consider a program that finds the optimal route given the
constraint that distance is to be minimized. To do this, the program employs the path-removal method of generating
multiple solutions and uses a least-cost search to minimize distance. The key to finding the shortest path is to keep a
solution that is shorter than the previously generated solution. When there are no more solutions to generate, the
optimal solution remains.

The entire “optimal solution” program is shown here. Notice that the program creates an additional stack, called

籍
optimal, which holds the optimal solution, and an instance variable, called minDist, which keeps track of the distance.
There are also changes to route( ) and some minor modifications to main( ).

书
// Find an "optimal" solution using least-cost with path removal.
#include <iostream>
#include <stack>

典 书
#include <string>
#include <vector>

经 读
using namespace std;

+ 不
// Flight information.
struct FlightInfo {

+
string from; // departure city

C
string to; // destination city

典
int distance; // distance between from and to
bool skip; // used in backtracking

经
FlightInfo() {
from = "";
to = "";

无
distance = 0;
skip = false;
}

FlightInfo(string f, string t, int d) {

from = f;
to = t;
distance = d;
skip = false;
}
};

const int MAXDIST = 100000;

// Find connections using least cost.

class Optimal {
// This vector holds the flight information.
vector<FlightInfo> flights;

// This stack is used for backtracking.

stack<FlightInfo> btStack;
.

// This stack holds the optimal solution.

stack<FlightInfo> optimal;

int minDist;

// If there is a flight between from and to,

// store the distance of the flight in dist.
// Return true if the flight exists and
// false otherwise.
bool match(string from, string to, int &dist);

// Least-cost version.
// Given from, find the closest connection.
// Return true if a connection is found,
// and false otherwise.
bool find(string from, FlightInfo &f);

public:

籍
// Constructor
Optimal() {
minDist = MAXDIST;

书
}

典 书
// Put flights into the database.
void addflight(string from, string to, int dist) {
flights.push_back(FlightInfo(from, to, dist));

经 读
}

+
// Show the route and total distance.

不
void route();

+
// Display the optimal route.

C 典
void Optimal::showOpt();

// Determine if there is a route between from and to.

经
void findroute(string from, string to);

// Return true if a route has been found.

无
bool routefound() {
return btStack.size() != 0;
}
};

// Show the route and total distance.

void Optimal::route()
{
stack<FlightInfo> optTemp;
int dist = 0;
FlightInfo f;

// Reverse the stack to display route.

while(!btStack.empty()) {
f = btStack.top();
optTemp.push(f);
btStack.pop();
dist += f.distance;
}

// If shorter, keep this route.

if(minDist > dist) {
.

optimal = optTemp;
minDist = dist;
}
}

// Display the optimal route.

void Optimal::showOpt()
{
FlightInfo f;
int dist = 0;

cout <<"Optimal solution is:\n";

// Display the optimal route.

while(!optimal.empty()) {
f = optimal.top();
optimal.pop();
cout << f.from << " to ";
dist += f.distance;
}

}
籍
cout << f.to << endl;
cout << "Distance is " << dist << endl;

书
典 书
// If there is a flight between from and to,
// store the distance of the flight in dist.
// Return true if the flight exists and

经 读
// false otherwise.
bool Optimal::match(string from, string to, int &dist)

+
{

不
for(unsigned i=0; i < flights.size(); i++) {

+
if(flights[i].from == from &&
flights[i].to == to && !flights[i].skip)

C 典
{
flights[i].skip = true; // prevent reuse
dist = flights[i].distance;

经
return true;
}
}

无
return false; // not found

// Least-cost version.
// Given from, find the closest connection.
// Return true if a connection is found,
// and false otherwise.
bool Optimal::find(string from, FlightInfo &f)
{
int pos = -1;
int dist = MAXDIST; // longer than longest flight

for(unsigned i=0; i < flights.size(); i++) {

if(flights[i].from == from && !flights[i].skip)
{
// Use the shortest flight.
if(flights[i].distance < dist) {
pos = i;
dist = flights[i].distance;
}
}
.

if(pos != -1) {
f = flights[pos];
flights[pos].skip = true; // prevent reuse

return true;
}

return false;
}

// Determine if there is a route between from and to.

void Optimal::findroute(string from, string to)
{
int dist;
FlightInfo f;

// See if at destination.
if(match(from, to, dist)) {

籍
btStack.push(FlightInfo(from, to, dist));
return;
}

典 书
// Try another connection.
if(find(from, f)) {

书
btStack.push(FlightInfo(from, to, f.distance));
findroute(f.to, to);

经 读
}
else if(!btStack.empty()) {

+
// Backtrack and try another connection.

不
f = btStack.top();

+
btStack.pop();
findroute(f.from, f.to);

C 典
}
}

经
// Find "optimal" solution by using least-cost with path removal.
int main() {
char to[40], from[40];

无
Optimal ob;

// Add flight connections to database.

ob.addflight("New York", "Chicago", 900);
ob.addflight("Chicago", "Denver", 1000);
ob.addflight("New York", "Toronto", 500);
ob.addflight("New York", "Denver", 1800);
ob.addflight("Toronto", "Calgary", 1700);
ob.addflight("Toronto", "Los Angeles", 2500);
ob.addflight("Toronto", "Chicago", 500);
ob.addflight("Denver", "Urbana", 1000);
ob.addflight("Denver", "Houston", 1000);
ob.addflight("Houston", "Los Angeles", 1500);
ob.addflight("Denver", "Los Angeles", 1000);

// Get departure and destination cities.

cout << "From? ";

cin.getline(from, 40);
cout << "To? ";

cin.getline(to, 40);
 // Find multiple solutions.
for(;;) {
// See if there is a connection.
ob.findroute(from, to);

// If no route found, then end.

if(!ob.routefound()) break;

ob.route();
}

// Display optimal solution.

ob.showOpt();

return 0;

The output from the program is shown here:

籍
From? New York
To? Los Angeles

书
Optimal solution is:
New York to Chicago to Denver to Los Angeles
Distance is 2900

典 书
In this case, the “optimal” solution is not quite the very best one, but it is still a very good one. As explained, when
using AI-based searches, the best solution to be found by one search technique will not always be the very best one

经 读
that might exist. You might want to try substituting another search technique in the preceding program, observing what
type of “optimal” solution it finds.

+ + 不
The one inefficiency in the preceding method is that all paths are followed to their conclusion. An improved method
would stop following a path as soon as the length equaled or exceeded the current minimum. You might want to
modify this program to accommodate such an enhancement.

C 典
无 经
.

Back to the Lost Keys

To conclude this chapter on problem solving, it seems only fitting to provide a C++ program that finds the lost car keys
described in the first example. The following code employs the same techniques used in the problem of finding a route
between two cities, so the program is presented without further explanation:
// Search for the lost keys.
#include <iostream>
#include <stack>
#include <string>
#include <vector>

using namespace std;

// Room information.
struct RoomInfo {
string from;

籍
string to;
bool skip;

书
RoomInfo() {
from = "";

典 书
to = "";
skip = false;
}

+ 经
from = f;
to = t;

不 读
RoomInfo(string f, string t) {

+
skip = false;
}

C 典
};

// Find the keys using a depth-first search.

经
class Search {
// This vector holds the room information.
vector<RoomInfo> rooms;

无
// This stack is used for backtracking.
stack<RoomInfo> btStack;

// Return true if a path exists between

// from and to. Return false otherwise.
bool match(string from, string to);

// Given from, find any path.

// Return true if a path is found,
// and false otherwise.
bool find(string from, RoomInfo &f);

public:

// Put rooms into the database.

void addroom(string from, string to)
{
rooms.push_back(RoomInfo(from, to));
}

// Show the route taken.

.

void route();

// Determine if there is a path between from and to.

void findkeys(string from, string to);

// Return true if the keys have been found.

bool keysfound() {
return !btStack.empty();
}
};

// Show the route.

void Search::route()
{
stack<RoomInfo> rev;
RoomInfo f;

// Reverse the stack to display route.

while(!btStack.empty()) {
f = btStack.top();

籍
rev.push(f);
btStack.pop();
}

典 书
// Display the route.

书
while(!rev.empty()) {
f = rev.top();
rev.pop();

经 读
cout << f.from << " to ";
}

+ 不
cout << f.to << endl;

+
C 典
// Return true if a path exists between
// from and to. Return false otherwise.
bool Search::match(string from, string to)

经
{
for(unsigned i=0; i < rooms.size(); i++) {
if(rooms[i].from == from &&

无
rooms[i].to == to && !rooms[i].skip)
{
rooms[i].skip = true; // prevent reuse
return true;
}
}

return false; // not found

}

// Given from, find any path.

// Return true if a path is found,
// and false otherwise.
bool Search::find(string from, RoomInfo &f)
{
for(unsigned i=0; i < rooms.size(); i++) {
if(rooms[i].from == from && !rooms[i].skip) {
f = rooms[i];
rooms[i].skip = true; // prevent reuse

return true;
}
 }

return false;
}

// Find the keys.

void Search::findkeys(string from, string to)
{
RoomInfo f;

// See if keys are found.

if(match(from, to)) {
btStack.push(RoomInfo(from, to));
return;
}

// Try another room.

if(find(from, f)) {
btStack.push(RoomInfo(from, to));
findkeys(f.to, to);

籍
}
else if(!btStack.empty()) {
// Backtrack and try another path.

书
f = btStack.top();
btStack.pop();

典 书
findkeys(f.from, f.to);
}
}

int main() {

+
Search ob;

经 不 读
+
// Add rooms to database.
ob.addroom("front_door", "lr");

C 典
ob.addroom("lr", "bath");
ob.addroom("lr", "hall");
ob.addroom("hall", "bd1");

经
ob.addroom("hall", "bd2");
ob.addroom("hall", "mb");
ob.addroom("lr", "kitchen");

无
ob.addroom("kitchen", "keys");

// Find the keys.

ob.findkeys("front_door", "keys");

// If keys are found, show the path.

if(ob.keysfound())
ob.route();

return 0;
}
 Some Things to Try
Because AI-based searching is an engaging, yet challenging area of computer science, it is enjoyable to experiment
with. Here are some things you might want to try. First, experiment with the heuristic searches by substituting the
breadth-first technique and observe the results. Second, substitute breadth-first searching when finding multiple
solutions or when finding an “optimal” solution. Finally, try handling other real-world search situations.

书 籍
经 典 读 书
+ + 不
C 典
无 经
 Chapter 8: Building a Custom STL Container

Overview
A central theme of this book is the raw power that C++ puts into the hands of the programmer. Perhaps nothing
illustrates this better than the Standard Template Library (STL), which is reshaping the way that programs are written.
The STL is a sophisticated set of template classes and functions that implement many popular and commonly used
data structures and algorithms. Because the STL offers off-the-shelf solutions to a variety of programming problems
that involve the storage and retrieval of data, STL-based code is becoming commonplace. For example, the garbage
collection subsystem in Chapter 2 and the thread control panel in Chapter 3 both make use of the STL, which greatly
simplifies their code. In the past, if a data structure such as a list, stack, or queue was needed, it would have to be
coded manually. Today, the programmer can simply use one provided by the STL.

At the core of the STL is the container. A container is an object that holds other objects. The STL offers several built-in

籍
containers that support such data structures as stacks, queues, lists, and vectors. Because containers are
templatized, they can hold any type of object, including objects of classes that you create.

书
Although the built-in containers are quite useful, you are not limited to them. One of the most impressive aspects of the
STL is that it allows you to create your own containers. Thus, the STL is extensible. Once you have created a
container, it is automatically fully compatible with the rest of the STL.

经 典 读 书
In this chapter, you will see how to create your own custom STL container. Although custom containers are not difficult
to create, many programmers are initially intimidated by the prospect. One reason for this is that the STL’s
template-based syntax can be a bit overwhelming at first glance. In actuality the STL is a conceptually clean and
easy-to-understand subsystem. If you follow a few rules, you will have no trouble creating your own containers.

+ 不
The custom container developed here implements a range-selectable dynamic array called RangeArray. When using

+
RangeArray, you can specify the beginning and ending indices. For example, RangeArray lets you create an array that

C
runs from –10 to 10.

经 典
无
 A Brief Review of the STL
The STL is a very large topic that would require hundreds of pages to describe in detail, and a complete description of
the STL is far beyond the scope of this chapter. Therefore, this chapter assumes that you have a basic working
knowledge of the STL. The foregoing not withstanding, a brief review of the main STL terms and constituents is given
here so that all readers will be on an even footing.

The STL is based on three main features: containers, iterators, and algorithms. As mentioned, containers are objects
that hold other objects. Iterators are pointer-like objects that enable you to cycle through a container in much the same
way that you cycle through an array using a pointer. Algorithms operate on containers, modifying, copying, or
otherwise manipulating their elements.

In addition to containers, algorithms, and iterators, the STL relies upon several other standard components for support,
such as allocators, adaptors, predicates, and function objects. The following sections take a closer look at the main
STL constituents.

Containers
籍
Note For a complete discussion of the STL, I recommend my book STL Programming From the Ground Up,
McGraw-Hill/Osborne.

书
经 典 读 书
The STL defines two types of containers: sequence and associative. Sequence containers hold a linear list of objects.
The STL provides several built-in sequence containers, including vector, which defines a dynamic array, deque, which
creates a double-ended queue, and list, which implements a linked list. The associative containers store key/value
pairs. Thus, the associative containers allow efficient retrieval of values based on keys. For example, a map provides

+ 不
access to values with unique keys. Therefore, a map stores a key/value pair and allows a value to be retrieved given

+
its key.

C
Each container class defines a set of functions that may be applied to the container. For example, a list container

典
includes functions that insert, delete, and merge elements. A stack includes functions that push and pop values.

Algorithms

经
Algorithms act on containers. They provide the means by which you can manipulate the contents of containers. Their

无
capabilities include initialization, sorting, searching, and transforming the contents of containers. Many algorithms
operate on a range of elements within a container rather than the entire container. Examples of algorithms include
copy( ), remove( ), replace( ), and find( ).

Iterators

Iterators are objects that are similar to pointers. You can increment and decrement them. You can apply the * operator
to them. Iterators are declared using the iterator type defined by the various containers.

The STL also supports reverse iterators. Reverse iterators move through a sequence in the reverse direction. Thus, if
a reverse iterator points to the end of a sequence, incrementing that iterator will cause it to point to one element before
the end.
 Other STL Entities
In addition to containers, iterators, and algorithms, there are several other STL elements that play important roles:
allocators, function objects, adaptors, predicates, binders, and negators.

Each container has defined for it an allocator. Allocators manage memory allocation for a container. The default
allocator is an object of class allocator, but you can define your own allocator if one is needed by a specialized
application. For most uses, the default allocator is sufficient.

Function objects are classes that define operator( ). There are several predefined function objects, such as less( ),
greater( ) , plus( ), minus( ), multiplies( ), and divides( ). Perhaps the most widely used function object is less( ), which
determines when one object is less than another. Function objects can be used in place of function pointers when
using STL algorithms.

In the most general sense, an adaptor transforms one thing into another. There are container adaptors, iterator
adaptors, and function adaptors. An example of a container adaptor is queue, which adapts the deque container for use

籍
as a standard queue. Adaptors simplify a number of difficult situations.

Several of the algorithms and containers use a special type of function called a predicate. There are two variations of

书
predicates: unary and binary. A unary predicate takes one argument; a binary predicate takes two arguments. These
functions return true/false results, but the precise conditions that make them return true or false are defined by you.
Some algorithms use a special type of binary predicate that compares two elements. These comparison functions

典 书
return true if their first argument is less than their second.

读
Two other entities that populate the STL are binders and negators. A binder binds an argument to a function object.

经
There are two binders defined by the STL: bind2nd( ) and bind1st( ). A negator returns the complement of a predicate.

+
There are two negators: not1( ) and not2( ). Both binders and negators increase the versatility of the STL.

C+ 不
经 典
无
.

Custom Container Requirements

Before designing a custom container, it is necessary to know precisely what it must provide. According to Standard
C++, all containers must supply a prescribed set of types, operators, and member functions. In addition to these
common elements, both sequence and associative containers add several specific requirements. If you follow all the
rules, then your custom container can make use of all of the other entities defined by the STL, including such things as
allocators, function objects, predicates, and binders. Thus, a properly constructed custom container will be fully
integrated into the STL landscape. The following sections describe the requirements for a container.

General Requirements

Every container (sequence or associative) must manage memory through an allocator rather than by using new and
delete. Thus, the container must use the allocator’s member functions to allocate and release memory. The allocator is
specified as a parameter when a container is created. However, a default allocator class, called allocator, is provided
by C++, and the allocator parameter can be allowed to default to an object of type allocator, which is what all of the

籍
standard containers do. All allocators must provide the same member functions as allocator (but, of course, may vary
in implementation). This way, a container can use any allocator that the programmer supplies. The allocator functions
used by the container in this chapter are shown in Table 8-1.

Function

典 书
Table 8-1: The allocator Functions Used in This Chapter

书 Description

经 读
pointer allocate(size_type num, typename Returns a pointer to allocated memory that is large
allocator<void>::const_pointer h = 0); enough to hold num objects of type T. The value of

+ 不
h is a hint to the function that can be used to help

+
satisfy the request or ignored.

C
void construct(pointer ptr, const_reference val); Constructs an object of type T atptr.

经 典
void deallocate(pointer ptr, size_type num);

void destroy(pointer ptr);

Deallocates num objects of type T starting atptr.
The value of ptr must have been obtained from
allocate( ).

Destroys the object at ptr. Its destructor is

无
automatically called.

size_type max_size( ) const throw( ); Returns the maximum number of objects of type T
that can be allocated.

Every container must provide these types:

iterator const_iterator reference const_reference

value_type size_type difference_type

A reversible container (one that supports bidirectional iterators) must also supply these types:

reverse_iterator const_reverse_iterator

All containers must provide a default constructor, which creates a zero-length container, and a copy constructor.
Various parameterized constructors are also needed, the precise form of which differs between sequence and
associative containers. A destructor is also required.
.

The following member functions must be supported:

begin( ) clear( ) empty( ) end( )

erase( ) insert( ) max_size( ) rbegin( )

rend( ) size( ) swap( )

Of course, rbegin( ) and rend( ) are needed only by reversible containers. Some functions have overloaded forms.

Note All of the built-in STL containers supply a function called get_allocator( ), but this is not a requirement for custom
containers.

The fact that a container must define the iterator functions, such as begin( ), implies that the container must support all
required iterator operations.

The following operators must be supported by all containers:

籍
= == != >

< <= >=

典 书
Additional Sequence Container Requirements

书
In addition to a default constructor and a copy constructor, a sequence container must supply a constructor that
creates and initializes a specified number of elements. It must also provide a constructor that creates and initializes an

经 读
object given a range of elements. That is, the following forms of constructors must be supported:
Cnt( )

+ 不
Cnt(c)

+
Cnt(num, val)
Cnt(start, end)

C 典
Here, c is an object of type Cnt, num is an integer specifying a count, val is a value compatible with the type of objects
stored in Cnt, and start and end are iterators to a range of elements that will be used to initialize the container. A
custom container can specify additional constructors.

无 经
Note With the exception of the copy constructor, the constructors for the built-in STL sequence containers accept an
argument that specifies an allocator (which defaults to allocator), but this is not a requirement.

Standard C++ defines the following optional member functions for sequence containers:

at( ) back( ) front( ) pop_back( )

pop_front( ) push_back( ) push_front( )

The subscript operator [ ] is also optional. Of course, you can also add other member functions of your own design.

Associative Container Requirements

All associative containers must define these additional types:

key_compare key_type value_compare

Along with a default constructor and a copy constructor, an associative container must supply constructors that allow
you to specify a comparison function. You must also define a constructor that creates and initializes an object given a
range of elements. One version of this constructor must use the default comparison function. Another must allow the
user to specify the comparison function. That is, the following forms of constructors must be supported:
Cnt( )

Cnt(c)
Cnt(comp)
Cnt(start, end)
Cnt(start, end, comp)

Here, c is an object of type Cnt, start and end are iterators to a range of elements that will be used to initialize the
container, and comp is a comparison function. A custom container can specify additional constructors.

Note With the exception of the copy constructor, the constructors for the built-in STL associative containers accept an
argument that specifies an allocator (which defaults to allocator), but this is not a requirement.

Associative containers must provide these additional member functions:

count( ) equal_range( ) find( ) key_comp( )

籍
lower_bound( ) upper_bound( ) value_comp( )

典 书 书
+ 经 不 读
C+ 典
无 经
.

Creating a Range-Selectable, Dynamic Array Container

The remainder of this chapter develops a custom sequence container called RangeArray that provides a
range-selectable dynamic array. While the example shown here illustrates the creation of a sequence container, most
of the concepts can be applied to the implementation of associative containers as well.

How RangeArray Works

As you know, in C++ all arrays begin at zero, and negative indexes are not allowed. However, some applications
would benefit from an array that allows the programmer to specify different endpoints. Consider the Cartesian
coordinate plane: each axis is a line that has both positive and negative values. A convenient way to represent such a
line in a program would be to use an array that allows negative as well as positive indices. For example, given a line
that runs from –5 to 5, you would like to use an array that could be indexed as shown here:

书 籍
The RangeArray container developed here enables you to create such an array. More generally, RangeArray lets you
specify any arbitrary lower and upper bounds for the array indices.

RangeArray is a dynamic container that allows the array to grow in both the positive and negative directions. In this

典 书
regard, it is similar to the built-in container vector. RangeArray supports all of the required sequence container
operations, plus the optional [ ] operator and the optional functions at( ) , push_front( ), pop_front( ), and so on.

经 读
Using RangeArray, you can write code like this:

+
// This creates an array that runs from -3 to 4

不
// and is initialized to zero.

+
RangeArray<int> ob(-3, 4, 0);

C
// Load the values -3 to 4 into ob.

典
for(int i = -3; i < 5; i++) ob[i] = i;
// ...
cout << ob[-2];

经
// ...
ob[2] = ob[-1] % 2;

无
As you might surmise, the first line constructs a RangeArray object that runs from –3 to 4, with each element in the
array initialized to zero. The other lines of code illustrate that the array can be accessed via indices ranging from –3 to
4.

In general, RangeArray allows you to construct objects by specifying the lower and upper bounds of the array and the
value that is used to initialize each element in the array. That is, RangeArray supports the following form of constructor:
RangeArray(lowerbound, upperbound, initvalue)

Once the array has been constructed, it can be indexed by any value that is within its range. This means that arbitrary
indices, including negative indices, are allowed.

Because RangeArray is dynamic, it allows elements to be inserted or removed. When either of these operations occur,
the array grows or shrinks as needed. The key point, however, is that the array can grow or shrink in either direction. If
an element is added to the negative side, the negative side grows. If an element is removed from the positive side, the
positive shrinks.

Before starting, it is necessary to point out that the implementation of the RangeArray container is, by intent, not
optimized for high performance. Rather, it is optimized for clarity. Its purpose is to clearly show the steps required to
create a custom container. As such, it is designed for ease of understanding and straightforwardness of
implementation. You might find it enjoyable to enhance its performance on your own.
.

The Entire RangeArray Class

To begin, the entire code for RangeArray is shown here. As you will find when you create your own container classes,
even a simple container turns into a fairly large class. The reason for this is, of course, that several requirements must
be met. While no single requirement is difficult, combined they present a fair bit of code. Don’t be intimidated; each
part is examined piece by piece in the sections that follow.
// A custom container that implements a
// range-selectable array.
//
// Call this file ra.h
//
#include <iostream>
#include <iterator>
#include <algorithm>
#include <cstdlib>
#include <stdexcept>

using namespace std;

籍
// An exception class for RangeArray.
class RAExc {
string err;

书
public:

书
RAExc(string e) {

典
err = e;
}

经 读
string geterr() { return err; }

+
};

+ 不
// A range-selectable array container.
template<class T, class Allocator = allocator<T> >

C
class RangeArray {

典
T *arrayptr; // pointer to array that underlies the container

经
unsigned len; // holds length of the container
int upperbound; // lower bound
int lowerbound; // upper bound

无
Allocator a; // allocator
public:
// Required typedefs for container.
typedef T value_type;
typedef Allocator allocator_type;
typedef typename Allocator::reference reference;
typedef typename Allocator::const_reference const_reference;
typedef typename Allocator::size_type size_type;
typedef typename Allocator::difference_type difference_type;
typedef typename Allocator::pointer pointer;
typedef typename Allocator::const_pointer const_pointer;

// Forward iterators.
typedef T * iterator;
typedef const T * const_iterator;

// Note: This container does not support reverse

// iterators, but you can add them if you like.

// ***** Constructors and Destructor *****

.

// Default constructor.
RangeArray()
{
upperbound = lowerbound = 0;
len = 0;
arrayptr = a.allocate(0);
}

// Construct an array of the specified range

// with each element having the specified initial value.
RangeArray(int low, int high, const T &t);

// Construct zero-based array of num elements

// with the value t. This constructor is required
// for STL compatibility.
RangeArray(int num, const T &t=T());

// Construct from range of iterators.

RangeArray(iterator start, iterator stop);

籍
// Copy constructor.
RangeArray(const RangeArray &o);

书
// Destructor.

典 书
~RangeArray();

经 读
// ***** Operator Functions *****

+
// Return reference to specified element.

不
T &operator[](int i)

+
{
return arrayptr[i - lowerbound];

C 典
}

// Return const references to specified element.

经
const T &operator[](int i) const
{
return arrayptr[i - lowerbound];

无
}

// Assign one container to another.

RangeArray &operator=(const RangeArray &o);

// ***** Insert Functions *****

// Insert val at p.
iterator insert(iterator p, const T &val);

// Insert num copies of val at p.

void insert(iterator p, int num, const T &val)
{
for(; num>0; num--) p = insert(p, val) + 1;
}

// Insert range specified by start and stop at p.

void insert(iterator p, iterator start, iterator stop)
{
while(start != stop) {
p = insert(p, *start) + 1;
.

start++;
}
}

// ***** Erase Functions *****

// Erase element at p.
iterator erase(iterator p);

// Erase specified range.

iterator erase(iterator start, iterator stop)
{
iterator p = end();

for(int i=stop-start; i > 0; i--)

p = erase(start);

return p;
}

书 籍
// ***** Push and Pop Functions *****

// Add element to end.

典 书
void push_back(const T &val)
{
insert(end(), val);

经 读
}

+
// Remove element from end.

不
void pop_back()

+
{
erase(end()-1);

C 典
}

// Add element to front.

经
void push_front(const T &val)
{
insert(begin(), val);

无
}

// Remove element from front.

void pop_front()
{
erase(begin());
}

// ***** front() and back() functions *****

// Return reference to first element.

T &front()
{
return arrayptr[0];
}

// Return const reference to first element.

const T &front() const
{
return arrayptr[0];
}
.

// Return reference to last element.

T &back()
{
return arrayptr[len-1];
}

// Return const reference to last element.

const T &back() const
{
return arrayptr[len-1];
}

// ***** Iterator Functions *****

// Return iterator to first element.

iterator begin()
{
return &arrayptr[0];

籍
}

// Return iterator to last element.

书
iterator end()
{

典 书
return &arrayptr[upperbound - lowerbound];
}

经 读
// Return const iterator to first element.
const_iterator begin() const

+
{

不
return &arrayptr[0];

+
}

C 典
// Return const iterator to last element.
const_iterator end() const
{

经
return &arrayptr[upperbound - lowerbound];
}

无
// ***** Misc. Functions *****

// The at() function performs a range check.

// Return a reference to the specified element.
T &at(int i)
{
if(i < lowerbound || i >= upperbound)
throw out_of_range("Index Out of Range");

return arrayptr[i - lowerbound];

}

// Return a const reference to the specified element.

const T &at(int i) const
{
if(i < lowerbound || i >= upperbound)
throw out_of_range("Index Out of Range");

return arrayptr[i - lowerbound];

}
.

// Return the size of the container.

size_type size() const
{
return end() - begin();
}

// Return the maximum size of a RangeArray.

size_type max_size()
{
return a.max_size();
}

// Return true if container is empty.

bool empty()
{
return size() == 0;
}

// Exchange the values of two containers.

void swap(RangeArray &b)

籍
{
RangeArray<T> tmp;

书
tmp = *this;
*this = b;

典 书
b = tmp;
}

经 读
// Remove and destroy all elements.
void clear()

+
{

不
erase(begin(), end());

+
}

C 典
// ***** Non-STL functions *****

// Return endpoints.

经
int getlowerbound()
{
return lowerbound;

无
}

int getupperbound()
{
return upperbound;
}

};

// ***** Implementations of non-inline functions *****

// Construct an array of the specified range

// with each element having the specified initial value.
template <class T, class A>
RangeArray<T, A>::RangeArray(int low, int high,
const T &t)
{
if(high <= low) throw RAExc("Invalid Range");

high++;
.

// Save endpoints.
upperbound = high;
lowerbound = low;

// Allocate memory for the container.

arrayptr = a.allocate(high - low);

// Save the length of the container.

len = high - low;

// Construct the elements.

for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], t);
}

// Construct zero-based array of num elements

// with the value t. This constructor is required
// for STL compatibility.
template <class T, class A>
RangeArray<T, A>::RangeArray(int num, const T &t) {

籍
// Save endpoints.
upperbound = num;

书
lowerbound = 0;

典 书
// Allocate memory for the container.
arrayptr = a.allocate(num);

经 读
// Save the length of the container.
len = num;

+ + 不
// Construct the elements.
for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], t);

C 典
经
// Construct zero-based array from range of iterators.
// This constructor is required for STL compatibility.
template <class T, class A>

无
RangeArray<T, A>::RangeArray(iterator start,
iterator stop)
{
// Allocate sufficient memory.
arrayptr = a.allocate(stop - start);

upperbound = stop - start;

lowerbound = 0;

len = stop - start;

// Construct the elements using those

// specified by the range of iterators.
for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], *start++);
}

// Copy constructor.
template <class T, class A>
RangeArray<T, A>::RangeArray(const RangeArray<T, A> &o)
{
// Allocate memory for the copy.
.

arrayptr = a.allocate(o.size());

upperbound = o.upperbound;
lowerbound = o.lowerbound;
len = o.len;

// Make the copy.

for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], o.arrayptr[i]);
}

// Destructor.
template <class T, class A>
RangeArray<T, A>::~RangeArray()
{
// Call destructors for elements in the container.
for(size_type i=0; i < size(); i++)
a.destroy(&arrayptr[i]);

// Release memory.

籍
a.deallocate(arrayptr, size());
}

书
// Assign one container to another.
template <class T, class A> RangeArray<T, A> &

典 书
RangeArray<T, A>::operator=(const RangeArray<T, A> &o)
{
// Call destructors for elements in target container.

经 读
for(size_type i=0; i < size(); i++)
a.destroy(&arrayptr[i]);

+ 不
// Release original memory.

+
a.deallocate(arrayptr, size());

C 典
// Allocate memory for new size.
arrayptr = a.allocate(o.size());

经
upperbound = o.upperbound;
lowerbound = o.lowerbound;
len = o.len;

}
无
// Make copy.
for(size_type i=0; i < size(); i++)
arrayptr[i] = o.arrayptr[i];

return *this;

// Insert val at p.
template <class T, class A>
typename RangeArray<T, A>::iterator
RangeArray<T, A>::insert(iterator p, const T &val)
{
iterator q;
size_type i, j;

// Get sufficient memory.

T *tmp = a.allocate(size() + 1);

// Copy existing elements to new array,

// inserting new element if possible.
for(i=j=0; i < size(); i++, j++) {
.

if(&arrayptr[i] == p) {
tmp[j] = val;
q = &tmp[j];
j++;
}
tmp[j] = arrayptr[i];
}

// Otherwise, the new element goes on end.

if(p == end()) {
tmp[j] = val;
q = &tmp[j];
}

// Adjust len and bounds.

len++;
if(p < &arrayptr[abs(lowerbound)])
lowerbound--;
else

籍
upperbound++;

// Call destructors for elements in old container.

书
for(size_type i=0; i < size()-1; i++)
a.destroy(&arrayptr[i]);

典 书
// Release memory for old container.
a.deallocate(arrayptr, size()-1);

return q;

+ 经
arrayptr = tmp;

不 读
+
}

C 典
// Erase element at p.
template <class T, class A>
typename RangeArray<T, A>::iterator

经
RangeArray<T, A>::erase(iterator p)
{
iterator q = p;

无
// Destruct element being erased.
if(p != end()) a.destroy(p);

// Adjust len and bounds.

len--;
if(p < &arrayptr[abs(lowerbound)])
lowerbound++;
else
upperbound--;

// Compact remaining elements.

for( ; p < end(); p++)
*p = *(p+1);

return q;
}

// ******** Relational Operators **************

template<class T, class Allocator>

.

bool operator==(const RangeArray<T, Allocator> &a,

const RangeArray<T, Allocator> &b)
{
if(a.size() != b.size()) return false;

return equal(a.begin(), a.end(), b.begin());

}

template<class T, class Allocator>

bool operator!=(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{
if(a.size() != b.size()) return true;

return !equal(a.begin(), a.end(), b.begin());

}

template<class T, class Allocator>

bool operator<(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)

籍
{
return lexicographical_compare(a.begin(), a.end(),
b.begin(), b.end());

书
}

典 书
template<class T, class Allocator>
bool operator>(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)

经 读
{
return b < a;

+
}

+ 不
template<class T, class Allocator>
bool operator<=(const RangeArray<T, Allocator> &a,

C 典
const RangeArray<T, Allocator> &b)
{
return !(a > b);

经
}

template<class T, class Allocator>

无
bool operator>=(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{
return !(a < b);
}

As the comment at the beginning of the code indicates, you should put all of this code into a file called ra.h . It will be
used by the sample programs shown later.

The preceding code contains two classes. The first is the RAExc exception class. This is the type of exception that is
thrown if there is an attempt to create a RangeArray with invalid bounds, such as if the lower bound is greater than the
upper bound. The second is the RangeArray container class, which we will examine closely in the sections that follow.

The RangeArray Class in Detail

Like all of the built-in STL sequence containers, RangeArray begins with the following template specification:
template<class T, class Allocator = allocator<T> >

T is the type of data stored in the container and Allocator is the allocator, which defaults to the standard allocator.

The Private Members

.

The RangeArray array class begins with the following private declarations:
T *arrayptr; // pointer to array that underlies the container

unsigned len; // holds length of the container

int upperbound; // lower bound
int lowerbound; // upper bound

Allocator a; // allocator

The pointer arrayptr stores a pointer to the memory that will contain an array of elements of type T. This memory will
store the elements held by an object of type RangeArray. This array is indexed as usual, from zero. Indexes to a
RangeArray object will be translated into zero-based indexes into the array pointed to by arrayptr.

The current length of a RangeArray is stored in len. The upper and lower bounds of the array are stored in upperbound
and lowerbound, respectively. For the purposes of RangeArray, zero is counted as a positive value. The allocator for
the container is stored in a.

The Required Type Definitions

籍
After the private members, RangeArray defines the various typedefs required by all sequence containers, as shown
here:
// Required typedefs for container.

书
typedef T value_type;
typedef Allocator allocator_type;

典 书
typedef typename Allocator::reference reference;
typedef typename Allocator::const_reference const_reference;
typedef typename Allocator::size_type size_type;

经 读
typedef typename Allocator::difference_type difference_type;
typedef typename Allocator::pointer pointer;

+
typedef typename Allocator::const_pointer const_pointer;

+ 不
// Forward iterators.
typedef T * iterator;

C 典
typedef const T * const_iterator;

These typedefs are similar to the ones used by the built-in STL containers. Notice that the forward iterators are simply

经
pointers to objects of type T. This is sufficient for the purposes of RangeArray, but it might not be for more complicated
containers. Also, reverse iterators are not provided, but you might want to try adding them as an interesting exercise.

无
The RangeArray Constructors and Destructor

To create a compliant container, you must support the four constructors defined for sequence containers described
earlier. RangeArray also defines a fifth constructor, which enables you to construct an array of a specified range. The
constructors are shown here:
// Default constructor.
RangeArray()
{
upperbound = lowerbound = 0;
len = 0;
arrayptr = a.allocate(0);
}

// Construct an array of the specified range

// with each element having the specified initial value.
template <class T, class A>
RangeArray<T, A>::RangeArray(int low, int high,
const T &t)
{
if(high <= low) throw RAExc("Invalid Range");
.

high++;

// Save endpoints.
upperbound = high;
lowerbound = low;

// Allocate memory for the container.

arrayptr = a.allocate(high - low);

// Save the length of the container.

len = high - low;

// Construct the elements.

for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], t);
}

// Construct zero-based array of num elements

// with the value t. This constructor is required
// for STL compatibility.

籍
template <class T, class A>
RangeArray<T, A>::RangeArray(int num, const T &t) {

书
// Save endpoints.
upperbound = num;

典 书
lowerbound = 0;

// Allocate memory for the container.

经 读
arrayptr = a.allocate(num);

+
// Save the length of the container.

不
len = num;

+
// Construct the elements.

C 典
for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], t);

经
}

// Construct zero-based array from range of iterators.

无
// This constructor is required for STL compatibility.
template <class T, class A>
RangeArray<T, A>::RangeArray(iterator start,
iterator stop)
{
// Allocate sufficient memory.
arrayptr = a.allocate(stop - start);

upperbound = stop - start;

lowerbound = 0;

len = stop - start;

// Construct the elements using those

// specified by the range of iterators.
for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], *start++);
}

// Copy constructor.
template <class T, class A>
RangeArray<T, A>::RangeArray(const RangeArray<T, A> &o)
.

{
// Allocate memory for the copy.
arrayptr = a.allocate(o.size());

upperbound = o.upperbound;
lowerbound = o.lowerbound;
len = o.len;

// Make the copy.

for(size_type i=0; i < size(); i++)
a.construct(&arrayptr[i], o.arrayptr[i]);
}

The first constructor, which is the default constructor, creates an empty object. One of the constraints specified by the
STL is that the result of calling size( ) on a default object must be zero. Therefore, the default constructor sets the
lower and upper bounds to zero. It also sets len to zero and constructs a zero-length array using the allocate( ) function
provided by the allocator. The last two steps ensure that a fully formed object exists in all cases.

The second constructor is unique to RangeArray. It creates an object with a specified range, with each element given
the specified initial value. The lower bound is specified in the first parameter; the upper bound is specified in the

籍
second. Notice that a RAExc is thrown if the upper bound is less than or equal to the lower bound. The third parameter
is passed the initialization value. Thus, this statement:

书
RangeArray<char> ch(-2, 10, 'X');

creates an array of characters that begins at –2 and runs through 10, with each element having the initial value of X.

典 书
The array is allocated using allocate( ), which is a member function of the allocator class. To be a compliant container,
you must use the allocator functions, rather than new, to obtain memory required by the container. Once allocated,
each element in the array is constructed using the value passed as the third parameter. (Although it would be more

经 读
convenient if the initialization value was allowed to default so that it would not always have to be specified, doing so

+
would cause ambiguity with the third constructor, which is required by the STL.) The construction is accomplished

不
using the allocator function construct( ), another allocator function.

+
The third constructor creates a zero-based array that has a specified number of elements, each initialized with a

C
specified value. The initialization value can be allowed to default. This constructor is not particularly useful for

典
RangeArray but is required by the STL container specification.

The fourth constructor creates a zero-based array from a range of values. The range is specified by passing an iterator

经
to the beginning and end of the range. This constructor is also not particularly useful for RangeArray but is required by
the STL container specification.

无
Last is the copy constructor. It allocates memory for the new object and then copies the bounds, the length, and
elements from the source object. Thus, the copy has its own memory but is otherwise identical to the source.

The RangeArray destructor is shown here:

// Destructor.
template <class T, class A>
RangeArray<T, A>::~RangeArray()
{
// Call destructors for elements in the container.
for(size_type i=0; i < size(); i++)
a.destroy(&arrayptr[i]);

// Release memory.
a.deallocate(arrayptr, size());
}

It first destroys each element in the array by calling the allocator’s destroy( ) function. It then frees the memory used by
the array by calling the allocator’s deallocate( ) function.

Remember: the constructors for RangeArray do not use new to allocate memory. Nor does the destructor use delete to
release memory. Instead, they use the appropriate member functions provided by the allocator. This approach lets the
.

user specify an alternative means of managing memory for the container.

The RangeArray Operator Functions

There are three member operator functions defined by RangeArray. The first two are the operator[ ]( ) functions shown
here:
// Return reference to specified element.
T &operator[](int i)
{
return arrayptr[i - lowerbound];
}

// Return const references to specified element.

const T &operator[](int i) const
{
return arrayptr[i - lowerbound];
}

For a fully formed container, both const and non-const versions of the [ ] operator are required. These functions

籍
provide the mechanism by which the RangeArray can be indexed. Pay special attention to the way that arrayptr is
indexed. Remember that arrayptr points to a standard C++ array. Thus, the index passed in i must be converted into a
zero-based index into arrayptr. Recall that lowerbound holds the value that corresponds to the lowest index within the

书
RangeArray. Thus, to obtain a zero-based index, lowerbound is subtracted from i.

One other point: notice that no bounds checking is performed by operator[ ]( ). Bounds checking by this operator is not

典 书
required by the STL, and none is included here. This is reasonable because, aside from having a user-specifiable
range, a RangeArray will behave just a like a normal array in this regard. (Remember, normal C++ arrays do not

经 读
provide boundary checks.) Of course, if you want, you can add bounds checks to operator[ ] ( ).

+
The operator=( ) function shown next is a bit more complex:

不
// Assign one container to another.

+
template <class T, class A> RangeArray<T, A> &
RangeArray<T, A>::operator=(const RangeArray<T, A> &o)

C 典
{
// Call destructors for elements in target container.
for(size_type i=0; i < size(); i++)

经
a.destroy(&arrayptr[i]);

// Release original memory.

无
a.deallocate(arrayptr, size());

// Allocate memory for new size.

arrayptr = a.allocate(o.size());

upperbound = o.upperbound;
lowerbound = o.lowerbound;
len = o.len;

// Make copy.
for(size_type i=0; i < size(); i++)
arrayptr[i] = o.arrayptr[i];

return *this;
}

It first destroys and frees any preexisting objects in the target object. It then allocates sufficient memory to hold the
contents of the source object. Next, it sets the member variables appropriately and then copies the elements. Finally, a
reference to the target object is returned.

The Insert Functions

.

The STL requires that a sequence container support three forms of insert( ): one that inserts a value, one that inserts
multiple copies of a value, and one that inserts a range. The first version of insert( ) is shown here:
// Insert val at p.
template <class T, class A>
typename RangeArray<T, A>::iterator
RangeArray<T, A>::insert(iterator p, const T &val)
{
iterator q;
size_type i, j;

// Get sufficient memory.

T *tmp = a.allocate(size() + 1);

// Copy existing elements to new array,

// inserting new element if possible.
for(i=j=0; i < size(); i++, j++) {
if(&arrayptr[i] == p) {
tmp[j] = val;
q = &tmp[j];

籍
j++;
}
tmp[j] = arrayptr[i];

书
}

// Otherwise, the new element goes on end.

典 书
if(p == end()) {
tmp[j] = val;
q = &tmp[j];

经 读
}

+ 不
// Adjust len and bounds.
len++;

+
if(p < &arrayptr[abs(lowerbound)])

C
lowerbound--;

典
else
upperbound++;

经
// Call destructors for elements in old container.
for(size_type i=0; i < size()-1; i++)
a.destroy(&arrayptr[i]);

无
// Release memory for old container.
a.deallocate(arrayptr, size()-1);

arrayptr = tmp;

return q;
}

The first version inserts an element at the location specified by the iterator passed as its first parameter. It returns an
iterator to the inserted element. It operates by allocating a new segment of memory that is large enough to hold all of
the existing elements plus the new one. It then copies the existing elements to the newly allocated memory, inserting
the new element at the proper location. It then updates len and upperbound or lowerbound appropriately. Next, it
destroys and deallocates the objects from the original RangeArray and assigns to arrayptr the address of the new
memory. Finally, it returns a pointer to the inserted object.

Now, look closely at the code that updates the lowerbound or upperbound variable. Remember that the array can grow
in either the positive or negative direction, based upon whether the new element is inserted on the positive side or the
negative size of the array. Thus it is necessary to determine where the insertion takes place and change the proper
value. This is determined by comparing the iterator passed in p to the address of the element at arrayptr[lowerbound].
If the iterator is less than that element, the negative side is expanding; otherwise, the positive end is growing.
.

The second two versions of insert( ), shown next, are defined in terms of the first:
// Insert num copies of val at p.
void insert(iterator p, int num, const T &val)
{
for(; num>0; num--) p = insert(p, val) + 1;
}

// Insert range specified by start and stop at p.

void insert(iterator p, iterator start, iterator stop)
{
while(start != stop) {
p = insert(p, *start) + 1;
start++;
}
}

The Erase Functions

Sequence containers must support two erase( ) functions. The first removes the element pointed to by an iterator. It

籍
returns an iterator to the element immediately after the one removed, or end( ) if the last element is removed. This
version is shown here:
// Erase element at p.

书
template <class T, class A>
typename RangeArray<T, A>::iterator

典 书
RangeArray<T, A>::erase(iterator p)
{
iterator q = p;

+ 经 读
// Destruct element being erased.
if(p != end()) a.destroy(p);

不
+
// Adjust len and bounds.
len--;

C 典
if(p < &arrayptr[abs(lowerbound)])
lowerbound++;
else

经
upperbound--;

// Compact remaining elements.

无
for( ; p < end(); p++)
*p = *(p+1);

return q;
}

This version operates by destroying the deleted element, adjusting the lower and upper bounds appropriately, and
then compacting the remaining elements.

The second version, shown next, is framed in terms of the first. It removes a range of elements. It returns an iterator to
the last element in the range (that is, the one pointed to by stop). Thus, it removes the elements between start and
stop–1, inclusive.

// Erase specified range.

iterator erase(iterator start, iterator stop)
{
iterator p = end();

for(int i=stop-start; i > 0; i--)

p = erase(start);

return p;
.

The Push and Pop Functions

RangeArray implements push_back( ), pop_back( ), push_front( ), and pop_front( ), shown next. As you can see, they
are implemented in terms of insert( ) and erase( ), and their operation is straightforward.
// Add element to end.
void push_back(const T &val)
{
insert(end(), val);
}

// Remove element from end.

void pop_back()
{
erase(end()-1);
}

// Add element to front.

籍
void push_front(const T &val)
{
insert(begin(), val);

书
}

// Remove element from front.

典 书
void pop_front()
{
erase(begin());

经 读
}

+ 不
The front( ) and back( ) Functions

+
The front( ) and back( ) functions, shown here, simply return an iterator to the start or end of the array, respectively.

C
Their implementation is straightforward. Notice, however, that both const and non-const versions are required.

典
// Return reference to first element.
T &front()

经
{
return arrayptr[0];
}

}
无
// Return const reference to first element.
const T &front() const

return arrayptr[0];

// Return reference to last element.

T &back()
{
return arrayptr[len-1];
}

// Return const reference to last element.

const T &back() const
{
return arrayptr[len-1];
}

The Iterator Functions

.

Because iterators for the RangeArray container are simply pointers into the memory pointed to by arrayptr, the iterator
functions begin( ) and end( ) are trivial. Notice that both const and non-const versions are required.
// Return iterator to first element.
iterator begin()
{
return &arrayptr[0];
}

// Return iterator to last element.

iterator end()
{
return &arrayptr[upperbound - lowerbound];
}

// Return const iterator to first element.

const_iterator begin() const
{
return &arrayptr[0];
}

籍
// Return const iterator to last element.
const_iterator end() const

书
{
return &arrayptr[upperbound - lowerbound];
}

经 典 读 书
Miscellaneous Functions

All sequence containers must implement size( ) , max_size( ), empty( ), swap( ), and clear( ). These functions are shown

+
here:

不
// Return the size of the container.

+
size_type size() const
{

C 典
return end() - begin();
}

经
// Return the maximum size of a RangeArray.
size_type max_size()
{

无
return a.max_size();
}

// Return true if container is empty.

bool empty()
{
return size() == 0;
}

// Exchange the values of two containers.

void swap(RangeArray &b)
{
RangeArray<T> tmp;

tmp = *this;
*this = b;
b = tmp;
}

// Remove and destroy all elements.

void clear()
{
.

erase(begin(), end());
}

In general, the operation of these functions is intuitive. However, max_size( ) deserves a few words. It returns the
number of elements of type T that can be held by the largest container that can be created. This value is obtained by
calling the max_size( ) function defined by the allocator.

RangeArray also implements the optional at( ) function, shown next:

// The at() function performs a range check.

// Return a reference to the specified element.
T &at(int i)
{
if(i < lowerbound || i >= upperbound)
throw out_of_range("Index Out of Range");

return arrayptr[i - lowerbound];

}

// Return a const reference to the specified element.

const T &at(int i) const

籍
{
if(i < lowerbound || i >= upperbound)

书
throw out_of_range("Index Out of Range");

return arrayptr[i - lowerbound];

典 书
}

The at( ) function returns a reference to the element located at the specified index. It differs from operator[ ]( ) only in

经 读
that it performs a range check on the index. If the index is out of bounds, at( ) throws an out_of_range exception. This
exception is defined by C++ in the header <stdexcept>.

+ 不
RangeArray also provides the non-STL functions getlowerbound( ) and getupperbound( ). These obtain the lower and

+
upper bounds, respectively, of the RangeArray. They are shown here:

C
// Return endpoints.

典
int getlowerbound()
{
return lowerbound;

经
}

int getupperbound()

无
{
return upperbound;
}

The Relational Operators

The relational operators defined for RangeArray are shown here:

template<class T, class Allocator>
bool operator==(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{
if(a.size() != b.size()) return false;

return equal(a.begin(), a.end(), b.begin());

}

template<class T, class Allocator>

bool operator!=(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{
if(a.size() != b.size()) return true;
.

return !equal(a.begin(), a.end(), b.begin());

}

template<class T, class Allocator>

bool operator<(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{
return lexicographical_compare(a.begin(), a.end(),
b.begin(), b.end());
}

template<class T, class Allocator>

bool operator>(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{
return b < a;
}

template<class T, class Allocator>

籍
bool operator<=(const RangeArray<T, Allocator> &a,
const RangeArray<T, Allocator> &b)
{

书
return !(a > b);
}

典 书
template<class T, class Allocator>
bool operator>=(const RangeArray<T, Allocator> &a,

经 读
const RangeArray<T, Allocator> &b)
{

+
return !(a < b);

不
}

+
Both operator=( ) and operator!=( ) use the equal( ) algorithm to determine equality. As defined by equal( ), two objects

C
are the same if each contains the same elements in the same order.

典
The < operator uses lexicographical_compare( ) to determine when one object is less than another. The use of this

经
function is recommended by Standard C++. It operates by comparing corresponding elements in the two sequences,
searching for the first mismatch. If a mismatch is found, it returns true if the element from the first range is less than
the element from the second range. It returns false otherwise.

无
Some RangeArray Sample Programs

To demonstrate RangeArray, three sample programs are shown. The first one, shown next, exercises the various
member functions. It also employs three algorithms, a function object, and a binder. The use of these elements
demonstrates that RangeArray is a fully functional container that is compatible with the rest of the STL.
// Demonstrate basic RangeArray operations.
#include <iostream>
#include <algorithm>
#include <functional>
#include "ra.h"
using namespace std;

// Display integers -- for use by for_each.

void display(int v)
{
cout << v << " ";
}

int main()
{
.

RangeArray<int> ob(-5, 5, 0);

RangeArray<int>::iterator p;
int i, sum;

cout << "Size of ob is: " << ob.size() << endl;

cout << "Initial contents of ob:\n";

for(i=-5; i <= 5; i++) cout << ob[i] << " ";
cout << endl;

// Give ob some values.

for(i=-5; i <= 5; i++) ob[i] = i;

cout << "New values for ob: \n";

p = ob.begin();
do {
cout << *p++ << " ";
} while (p != ob.end());
cout << endl;

籍
// Display sum of negative indexes.
sum = 0;
for(i = ob.getlowerbound(); i < 0; i++)

书
sum += ob[i];
cout << "Sum of values with negative subscripts is: ";

典 书
cout << sum << "\n\n";

// Use copy() algorithm to copy one object to another.

经 读
cout << "Copy ob to ob2 using copy() algorithm.\n";

+
RangeArray<int> ob2(-5, 5, 0);

不
copy(ob.begin(), ob.end(), ob2.begin());

+
// Use for_each() algorithm to display ob2.

C 典
cout << "Contents of ob2: \n";
for_each(ob2.begin(), ob2.end(), display);
cout << "\n\n";

经
// Use replace_copy_if() algorithm to remove values less than zero.
cout << "Replace values less than zero with zero.\n";

无
cout << "Put the result into ob3.\n";
RangeArray<int> ob3(ob.begin(), ob.end());

// The next line uses the function object less() and

// the binder bind2nd().
replace_copy_if(ob.begin(), ob.end(), ob3.begin(),
bind2nd(less<int>(), 0), 0);
cout << "Contents of ob3: \n";
for_each(ob3.begin(), ob3.end(), display);
cout << "\n\n";

cout << "Swap ob and ob3.\n";

ob.swap(ob3); // swap ob and ob3
cout << "Here is ob3:\n";
for_each(ob3.begin(), ob3.end(), display);
cout << endl;
cout << "Swap again to restore.\n";
ob.swap(ob3); // restore
cout << "Here is ob3 after second swap:\n";
for_each(ob3.begin(), ob3.end(), display);
cout << "\n\n";
.

// Use insert() member functions.

cout << "Element at ob[0] is " << ob[0] << endl;
cout << "Insert values into ob.\n";
ob.insert(ob.end(), -9999);
ob.insert(&ob[1], 99);
ob.insert(&ob[-3], -99);
for_each(ob.begin(), ob.end(), display);
cout << endl;
cout << "Element at ob[0] is " << ob[0] << "\n\n";

cout << "Insert -7 three times to front of ob.\n";

ob.insert(ob.begin(), 3, -7);
for_each(ob.begin(), ob.end(), display);
cout << endl;
cout << "Element at ob[0] is " << ob[0] << "\n\n";

// Use push_back() and pop_back().

cout << "Push back the value 40 onto ob.\n";
ob.push_back(40);
for_each(ob.begin(), ob.end(), display);

籍
cout << endl;
cout << "Pop back two values from ob.\n";
ob.pop_back(); ob.pop_back();

书
for_each(ob.begin(), ob.end(), display);
cout << "\n\n";

典 书
// Use push_front() and pop_front().
cout << "Push front the value 19 onto ob.\n";

经 读
ob.push_front(19);
for_each(ob.begin(), ob.end(), display);

+
cout << endl;

不
cout << "Pop front two values from ob.\n";

+
ob.pop_front(); ob.pop_front();
for_each(ob.begin(), ob.end(), display);

C 典
cout << "\n\n";

// Use front() and back()

经
cout << "ob.front(): " << ob.front() << endl;
cout << "ob.back(): " << ob.back() << "\n\n";

无
// Use erase().
cout << "Erase element at 0.\n";
p = ob.erase(&ob[0]);
for_each(ob.begin(), ob.end(), display);
cout << endl;
cout << "Element at ob[0] is " << ob[0] << endl;
cout << endl;

cout << "Erase many elements in ob.\n";

p = ob.erase(&ob[-2], &ob[3]);
for_each(ob.begin(), ob.end(), display);
cout << endl;
cout << "Element at ob[0] is " << ob[0] << endl;
cout << endl;

cout << "Insert ob4 into ob.\n";

RangeArray<int> ob4(0, 2, 0);
for(i=0; i < 3; i++) ob4[i] = i+100;
ob.insert(&ob[0], ob4.begin(), ob4.end());
for_each(ob.begin(), ob.end(), display);
cout << endl;
cout << "Element at ob[0] is " << ob[0] << endl;
.

cout << endl;

cout << "Here is ob shown with its indices:\n";

for(i=ob.getlowerbound(); i<ob.getupperbound(); i++)
cout << "[" << i << "]: " << ob[i] << endl;
cout << endl;

// Use the at() function.

cout << "Use the at() function.\n";
for(i=ob.getlowerbound(); i < ob.getupperbound(); i++)
ob.at(i) = i * 11;

for(i=ob.getlowerbound(); i < ob.getupperbound(); i++)

cout << ob.at(i) << " ";
cout << "\n\n";

// Use the clear() function.

cout << "Clear ob.\n";
ob.clear();
for_each(ob.begin(), ob.end(), display); // no effect!

籍
cout << "Size of ob after clear: " << ob.size()
<< "\nBounds: " << ob.getlowerbound()
<< " to " << ob.getupperbound() << "\n\n";

典 书
// Create a copy of an object.

书
cout << "Make a copy of ob2.\n";
RangeArray<int> ob5(ob2);
for_each(ob5.begin(), ob5.end(), display);

经 读
cout << "\n\n";

+
// Construct a new object from a range.

不
cout << "Construct object from a range.\n";

+
RangeArray<int> ob6(&ob2[-2], ob2.end());
cout << "Size of ob6: " << ob6.size() << endl;

C 典
for_each(ob6.begin(), ob6.end(), display);
cout << endl;

经
return 0;
}

无
The output from this program is shown here:
Size of ob is: 11
Initial contents of ob:
00000000000
New values for ob:
-5 -4 -3 -2 -1 0 1 2 3 4 5
Sum of values with negative subscripts is: -15

Copy ob to ob2 using copy() algorithm.

Contents of ob2:
-5 -4 -3 -2 -1 0 1 2 3 4 5

Replace values less than zero with zero.

Put the result into ob3.
Contents of ob3:
00000012345

Swap ob and ob3.

Here is ob3:
-5 -4 -3 -2 -1 0 1 2 3 4 5
Swap again to restore.
.

Here is ob3 after second swap:

00000012345

Element at ob[0] is 0
Insert values into ob.
-5 -4 -99 -3 -2 -1 0 99 1 2 3 4 5 -9999
Element at ob[0] is 0

Insert -7 three times to front of ob.

-7 -7 -7 -5 -4 -99 -3 -2 -1 0 99 1 2 3 4 5 -9999
Element at ob[0] is 0

Push back the value 40 onto ob.

-7 -7 -7 -5 -4 -99 -3 -2 -1 0 99 1 2 3 4 5 -9999 40
Pop back two values from ob.
-7 -7 -7 -5 -4 -99 -3 -2 -1 0 99 1 2 3 4 5

Push front the value 19 onto ob.

19 -7 -7 -7 -5 -4 -99 -3 -2 -1 0 99 1 2 3 4 5
Pop front two values from ob.

籍
-7 -7 -5 -4 -99 -3 -2 -1 0 99 1 2 3 4 5

ob.front(): -7

书
ob.back(): 5

典 书
Erase element at 0.
-7 -7 -5 -4 -99 -3 -2 -1 99 1 2 3 4 5
Element at ob[0] is 99

+ 经 读
Erase many elements in ob.
-7 -7 -5 -4 -99 -3 3 4 5

不
Element at ob[0] is 3

+
Insert ob4 into ob.

C 典
-7 -7 -5 -4 -99 -3 100 101 102 3 4 5
Element at ob[0] is 100

经
Here is ob shown with its indices:
[-6]: -7
[-5]: -7

无
[-4]: -5
[-3]: -4
[-2]: -99
[-1]: -3
[0]: 100
[1]: 101
[2]: 102
[3]: 3
[4]: 4
[5]: 5

Use the at() function.

-66 -55 -44 -33 -22 -11 0 11 22 33 44 55

Clear ob.
Size of ob after clear: 0
Bounds: 0 to 0

Make a copy of ob2.

-5 -4 -3 -2 -1 0 1 2 3 4 5

Construct object from a range.

.

Size of ob6: 8
-2 -1 0 1 2 3 4 5

The next sample program demonstrates the relational operators:

// Demonstrate the relational operators.
#include <iostream>
#include "ra.h"

using namespace std;

// Display integers -- for use by for_each.

void display(int v)
{
cout << v << " ";
}

int main()
{
RangeArray<int> ob1(-3, 2, 0), ob2(-3, 2, 0), ob3(-4, 4, 0);
int i;

籍
// Give ob1 and ob2 some values.

书
for(i = -3; i < 3; i++) {
ob1[i] = i;
ob2[i] = i;

典 书
}

cout << "Contents of ob1 and ob2:\n";

经 读
for(i=-3; i < 3; i++)
cout << ob1[i] << " ";

+
cout << endl;

+ 不
for(i=-3; i < 3; i++)
cout << ob2[i] << " ";

C 典
cout << "\n\n";

if(ob1 == ob2) cout << "ob1 == ob2\n";

经
if(ob1 != ob2) cout << "error\n";
cout << endl;

无
cout << "Assign ob1[-1] the value 99\n";
ob1[-1] = 99;
cout << "Contents of ob1 are now:\n";
for(i=-3; i < 3; i++)
cout << ob1[i] << " ";
cout << endl;

if(ob1 == ob2) cout << "error\n";

if(ob1 != ob2) cout << "ob1 != ob2\n";
cout << endl;

if(ob1 < ob2) cout << "ob1 < ob2\n";

if(ob1 <= ob2) cout << "ob1 <= ob2\n";
if(ob1 > ob2) cout << "ob1 > ob2\n";
if(ob1 >= ob2) cout << "ob1 >= ob2\n";

if(ob2 < ob1) cout << "ob2 < ob1\n";

if(ob2 <= ob1) cout << "ob2 <= ob1\n";
if(ob2 > ob1) cout << "ob2 > ob1\n";
if(ob2 >= ob1) cout << "ob2 >= ob1\n";
cout << endl;
.

// Compare objects of differing sizes.

if(ob3 != ob1) cout << "ob3 != ob1\n";
if(ob3 == ob1) cout << "ob3 == ob1\n";

return 0;
}

Its output is shown next:

Contents of ob1 and ob2:
-3 -2 -1 0 1 2
-3 -2 -1 0 1 2

ob1 == ob2

Assign ob1[-1] the value 99

Contents of ob1 are now:
-3 -2 99 0 1 2
ob1 != ob2

籍
ob1 > ob2
ob1 >= ob2

书
ob2 < ob1
ob2 <= ob1

典 书
ob3 != ob1

The following program stores class objects in a RangeArray. It also illustrates when the constructor and destructor

经 读
functions are called when various operations take place.
// Store class objects in a RangeArray.

+
#include <iostream>

不
#include "ra.h"

C+
using namespace std;

class test {
public:

典
经
int a;

test() { cout << "Constructing\n"; a=0; }

}
无
test(const test &o) {
cout << "Copy Constructor\n";
a = o.a;

~test() { cout << "Destructing\n"; }

};

int main()
{
RangeArray<test> t(-3, 1, test());
int i;

cout << "Original contents of t:\n";

for(i=-3; i < 2; i++) cout << t[i].a << " ";
cout << endl;

// Give t some new values.

for(i=-3; i < 2; i++) t[i].a = i;
.

cout << "New contents of t:\n";

for(i=-3; i < 2; i++) cout << t[i].a << " ";
cout << endl;

// Copy to new container.

RangeArray<test> t2(-7, 3, test());
copy(t.begin(), t.end(), &t2[-2]);

cout << "Contents of t2:\n";

for(i=-7; i < 4; i++) cout << t2[i].a << " ";
cout << endl;

RangeArray<test> t3(t.begin()+1, t.end()-1);

cout << "Contents of t3:\n";
for(i=t3.getlowerbound(); i < t3.getupperbound(); i++)
cout << t3[i].a << " ";
cout << endl;

t.clear();

籍
cout << "Size after clear(): " << t.size() << endl;

// Assign container objects.

书
t = t3;
cout << "Contents of t:\n";

典 书
for(i=t.getlowerbound(); i<t.getupperbound(); i++)
cout << t[i].a << " ";
cout << endl;

}
return 0;

+ 经 不 读
+
The output from this program is shown here:
Constructing

C 典
Copy Constructor
Copy Constructor
Copy Constructor

经
Copy Constructor
Copy Constructor
Destructing

无
Original contents of t:
00000
New contents of t:
-3 -2 -1 0 1
Constructing
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Copy Constructor
Destructing
Contents of t2:
0 0 0 0 0 -3 -2 -1 0 1 0
Copy Constructor
Copy Constructor
Copy Constructor
Contents of t3:
-2 -1 0
Destructing
Destructing
Destructing
Destructing
Destructing
Size after clear(): 0
Contents of t:
-2 -1 0
Destructing
Destructing
Destructing
Destructing
Destructing
Destructing
Destructing
Destructing
Destructing

籍
Destructing
Destructing
Destructing

书
Destructing
Destructing

典 书
Destructing
Destructing
Destructing

+ 经 不 读
Some Things to Try

You might want to try enhancing and experimenting with the RangeArray class. For example, you can add reverse

+
iterators and the rbegin( ) and rend( ) functions. Here are some other ideas. Try optimizing the container. As stated, the

C
code in RangeArray was designed for transparency of operation, not for speed, so it can be easily improved. For

典
example, try allocating a little more memory than needed when constructing objects so that not all insert operations
force a re-allocation. Try creating member functions that convert a RangeArray into a standard, zero-based array. Try
adding a constructor that takes a standard array and an index as a parameter. Have the constructor convert the array

经
into a RangeArray. Use the index as the location of zero. Finally, try using a vector, rather than a standard array, to
hold the elements of the RangeArray. See if it simplifies the implementation. (You will be pleasantly surprised.)

无
 Chapter 9: A Mini C++ Interpreter

Overview
The preceding chapters have shown the richness and power of the C++ language, focusing on the wide range of
tasks to which it can be successfully applied. This, the final chapter of the book, examines C++ from a different point of
view: the artistry of its design. At the base of C++’s power lies a profoundly elegant, logically consistent architecture. It
is this solid framework that binds the individual features of C++ into a cohesive whole. It is also what gives C++ that
rare, almost intangible, formal beauty that all programmers intuitively recognize.

To explore the elegant design of C++, this chapter develops a program that understands a portion of the language.
There are two types of programs that are capable of doing this: compilers and interpreters. Although creating a
compiler for C++ is far beyond the scope of this book, creating an interpreter is a more manageable task. Therefore,
this chapter develops an interpreter for a subset of C++. This interpreter is called Mini C++. Because Mini C++

籍
interprets the C++ language, its implementation illustrates in concrete terms the reasons and rationale behind the C++
syntax.

书
In addition to demonstrating the logically consistent nature of C++, Mini C++ provides a second benefit: it gives you an
execution engine that you can enhance or adapt as desired. For example, you can expand Mini C++ to interpret a
larger subset of C++, or you can adapt it to interpret a different computer language altogether. You can even use it as

典 书
a test bench for trying out a computer language of your own design. Its uses are limited only by your imagination.

+ 经 不 读
C+ 典
无 经
 Interpreters versus Compilers
Before we begin, it is necessary to explain what an interpreter is and how it differs from a compiler. Although both
compilers and interpreters take as input the source code for a program, what they do with that source code differs
significantly.

A compiler converts the source code of a program into an executable form. Often, as is the case with C++, this
executable form consists of actual CPU instructions that are directly executed by the computer. In other cases, the
output of a compiler is a portable intermediate form, which is then executed by a runtime system. For example, Java
compiles to an intermediate form called bytecode, which is then executed by the Java Virtual Machine.

An interpreter works in a completely different way. It reads the source code to a program, executing each statement as
it is encountered. Thus, an interpreter does not translate the source code into executable code. Instead, the interpreter
directly executes the program. The main drawback to interpretation is loss of execution speed. A program that is
interpreted executes more slowly than it does in its compiled form.

籍
Sometimes the term interpreter is used in situations other than that just described. For example, the original Java
Virtual Machine was called a bytecode interpreter. This is not the same type of interpreter developed in this chapter.
The interpreter developed here is a source code interpreter, which means that it executes a program simply by reading

书
its source code.

Although compiled programs execute faster than interpreted ones, interpreters are still commonly used in

典 书
programming for the following reasons. First, they can provide a truly interactive environment in which program
execution can be paused and resumed through user interaction. Such an interactive environment is well suited to

经 读
robotics, for example. Second, because of the nature of language interpreters, they are especially well suited for
interactive debugging. Third, interpreters are excellent for “script languages,” such as query languages for databases.

+
Fourth, they allow the same program to run on a variety of different platforms. Only the interpreter’s runtime package

不
must be implemented for each new environment.

+
There is another reason that interpreters are interesting: they are easy to modify, alter, or enhance. This means if you

C 典
want to create, experiment with, or control your own language, it is easier to do so with an interpreter than with a
compiler. Interpreters make great language prototyping environments because you can make a change to the
language and quickly see the results.

经
Of course, for the purposes of this chapter, the greatest benefit of an interpreter is that it enables you to see the inner
workings of C++ firsthand.

无
 An Overview of Mini C++
Mini C++ consists of a large amount of code. Before beginning, it will be helpful for you to understand in general how
Mini C++ is organized. Mini C++ contains three major subsystems: the expression parser, which handles numeric
expressions; the interpreter, which actually executes the program; and a set of library functions.

The expression parser evaluates a numeric expression and returns the result. The expression can contain constants
such as 10 or 88, variables, function calls, and, of course, operators. For example:
x + num / 32

is an expression.

The interpreter module executes a C++ program. The interpreter works by reading the source code one token at a
time. When it encounters a keyword, it does whatever that keyword requests. For example, when the interpreter reads
an if, it processes the condition specified by the if. If that condition is true, the interpreter executes the block of code
associated with the if. Interpretation continues until the end of the program is reached.

书 籍
The library subsystem defines those functions that are built into Mini C++. Thus, they constitute the Mini C++
“standard library.” Only a handful of library functions are included in this chapter, but you can easily add as many
others as you like. In general, the Mini C++ library functions serve the same purpose as the real C++ library functions.

Mini C++ organizes these subsystems into the files shown here:

Subsystem

Parser

经 典 读 书 File Name

parser.cpp

+ 不
Interpreter minicpp.cpp

+
Library libcpp.cpp

C 典
Mini C++ also uses the header file mccommon.h.

无 经
.

The Mini C++ Specifications

Despite the fact that C++ has relatively few keywords, it is a very rich and complicated language. It would take far
more than a single chapter to describe and implement an interpreter for all of C++. Instead, Mini C++ understands a
fairly narrow subset of the language. However, this particular subset includes many of C++’s most important features.
For example, Mini C++ supports recursive functions, global and local variables, nested scopes, and most control
statements. A list of all supported features is shown here:

Parameterized functions with local variables

Nested scopes

Recursion

The if statement

The switch statement

书 籍 The do-while, while, and for loops

The break statement

Local and global variables of type int and char

经 典 读 书Function parameters of type int and char

Integer and character constants

+
String constants (limited implementation)

C+ 不 The return statement, both with and without a value

A handful of standard library functions

经 典 The operators +, –, *, /, %, <, >, <=, >=, ==, !=, ++, – –, unary –, and unary +

Functions returning integers

/* and // comments

无 Console I/O via cin and cout

Even though this list may seem short, it takes a relatively large amount of code to implement it. One reason for this is
that a substantial “price of admission” must be paid when interpreting a language such as C++. Although the features
just listed describe only a small subset of the C++ language, they do enable the interpreter to handle the core of the
language, including its basic syntax, control statements, expressions, and function-call mechanisms. Thus, Mini C++
handles what one might call the “backbone” of the language.

As you no doubt noticed, Mini C++ does not support class. The reason for this is simply one of practicality. Supporting
class implies that the interpreter also supports user-defined types, object instantiation, and the dot (.) operator.
Furthermore, class requires that the interpreter understand public and private. Although these features are not in
themselves difficult for an interpreter to handle, collectively they cause the code for the interpreter to grow to a size that
is far too big to present in a single chapter. Once you understand how the interpreter works, you might find it enjoyable
to add support for class on your own.

Other features not supported include function and operator overloading, templates, namespaces, exceptions, the
preprocessor, structures, unions, and bitfields. Again, it is not difficult to interpret these features, but doing so causes
the code to grow unacceptably long for use in a book format. Adding support for these features does make for an
interesting project that you might enjoy on your own, however.
 Some Mini C++ Restrictions

Even by implementing a relatively small subset of C++, the source code for Mini C++ is still quite long. In order to
prevent it from growing even longer, a few restrictions have been imposed on the C++ grammar. The first is that the
targets of if, while, do, and for must be blocks of code surrounded by beginning and ending braces. You cannot use a
single statement. For example, Mini C++ will not correctly interpret code such as this:
for(a=0; a < 10; a++)
for(b=0; b < 10; b++)
for(c=0; c < 10; c++)
cout << "hi";

if(...)
if(...) x = 10;

Instead, you must write the code like this:

for(a=0; a < 10; a++) {
for(b=0; b < 10; b++) {
for(c=0; c < 10; c++) {
cout << "hi";

籍
}
}
}

if(...) {
if(...) {

}
x = 10;;

典 书 书
经 读
}

+
This restriction makes it easier for the interpreter to find the end of the code that forms the target to one of these

不
program control statements. However, because the objects of the program control statements are often blocks of code

+
anyway, this restriction does not seem too harsh. (With a little effort, you can remove this restriction, if you like.)

C
Another restriction is that prototypes are not supported. All functions are assumed to return an integer type (char

典
return types are allowed but elevated to int), and no parameter type checking is performed. Furthermore, because
function overloading is not supported, there must be one and only one version of a function.

经
The switch statement is fully implemented except that it does not support the default statement. This restriction was
made to reduce the size and complexity of the code needed to interpret a switch statement. (The switch is one of the

无
more complex statements to interpret.) Support for default is something that you might want to try adding on your own.

Finally, all functions must be preceded by either an int or char type specifier. Therefore, the Mini C++ interpreter does
not support a void return type. Thus, this declaration is valid:
int f()
{
// ...
}

But this one isn’t:

void f()
{
// ...
}
.

An Informal Theory of C++

In order to best understand the operation of Mini C++, it is necessary to understand how the C++ language is
structured. If you have ever seen a formal specification for the C++ language (such as that found in the ANSI/ISO C++
standard), you know that it is quite long and filled with rather cryptic statements. Don’t worry—you won’t need to deal
this formally with the C++ language to understand Mini C++. The parts of C++ that Mini C++ interprets are easy to
understand. That said, you will still need a basic understanding of how the C++ language is defined. For our purposes,
the informal discussion that follows is sufficient. Keep in mind, however, that it intentionally simplifies a few concepts.

A C++ program consists of a collection of one or more functions, plus global variables (if any exist). A function is
composed of a function name, its parameter list, and its body, which is a block of code. A block begins with a { , is
followed by one or more statements, and ends with a }. A block (also called a compound statement) creates a scope.
In general, a statement is either an expression, a nested code block, or begins with a keyword, such as if, for , or int. A
nested block creates a nested scope.

籍
Note Although the preceding categories of statements (keyword, block, and expression) are sufficient for the purposes
of understanding Mini C++, it is interesting to note that the ANSI/ISO standard for C++ uses a more finely grained

书
definition. It defines statements as labeled, expression, compound, selection, iteration, jump, declaration, and
try-block.

典 书
A C++ program begins with a call to main( ). It ends when either the last } or a return has been encountered in main(
) —assuming that exit( ) or abort( ) has not been called elsewhere. Any other functions contained in the program must
either be directly or indirectly called by main( ). Thus, when Mini C++ executes a program, it simply starts at the main( )

经 读
function and stops when main( ) ends.

+ 不
C++ Expressions

+
At the foundation of C++ is the expression, because much of the action in a program takes place through one. To

C
understand why, recall that C++ has three main categories of statements: keyword, block, and expression. This

典
implies that any statement that is not based on a keyword or does not define a block is, by definition, an expression
statement. Therefore, the following statements are all expressions:

经
count = 100; // line 1
sample = i / 22 * (c-10); // line 2
num = abs(count) * 2; // line 3

无
strcpy(str1, str2); // line 4

Let’s look more closely at each of these expression statements. In C++, the equal sign is an assignment operator. This
is important. C++ does not treat assignment as a separate type of statement as some languages do. Instead, the
equal sign is an assignment operator, and the value produced by the assignment operation is equal to that produced
by the right side of the expression. Therefore, an assignment statement is actually an assignment expression in C++.
Because it is an expression, it has a value. This is why it is legal to write expressions such as the following:
a = b = c = 100;
if( (a=4+5) == 0 ) ...;

The reason these work is that an assignment is an operation that produces a value.

Line 2 shows a more complex assignment expression. In this case, the value of the expression on the right is
computed and assigned to sample.

In line 3, abs( ) is called to return the absolute value of its argument. Thus, the use of a function in an expression
causes that function to be called.

As line 4 shows, even a function call, by itself, is an expression. In this case, strcpy( ) is called to copy the contents of
one string to another. Even though the value returned by strcpy( ) is ignored, the call to strcpy( ) still constitutes an
expression.
.

Defining an Expression

To handle the evaluation of expressions, Mini C++ uses an expression parser. Before you can understand the
operation of the expression parser, you need to know how expressions are constructed. In virtually all computer
languages, expressions are described using a set of production rules, which define the interaction of operands with
operators. The production rules that define Mini C++ expressions are shown here:

Here, rel-op refers to any of C++’s relational operators. The terms lvalue and rvalue refer to objects that can occur on
the left side and right side of an assignment statement, respectively. The arrow means “produces.” One thing that you

籍
should be aware of is that the precedence of the operators is built into the production rules. The higher the
precedence, the further down the list the operator will be. As the production rules show, Mini C++ supports the

书
following operators:

书
+ – * / % ++ ––

<

经 典 读
<= > >=

To see how these rules work let’s evaluate this C++ expression:
== != ()

+
count = 10 - 5 * 3;

+ 不
First, apply rule 1, which dissects the expression into these three parts:

C
经 典
无
Since there are no relational operators in the “rvalue” part of the subexpression, the term production rule is invoked.

Of course, the second term is composed of the factors 5 and 3. These two factors are constants and represent the
lowest level of the production rules. Next, begin moving back up the rules to compute the value of the expression.
First, multiply 5*3, which yields 15. Next, subtract that value from 10, yielding –5. Finally, this value is assigned to
count and is also the value of the entire expression. The Mini C++ expression parser evaluates expressions in the
same way. It simply implements the production rules.

Because the expression parser is so important to Mini C++, it is where we will begin.
.

The Expression Parser

The piece of code that reads and analyzes expressions is called an expression parser. Without a doubt, the
expression parser is the single most important subsystem needed by Mini C++. Because C++ defines expressions
more broadly than do many other languages, a substantial amount of the code that constitutes a C++ program is
executed by the expression parser.

There are several different ways to design an expression parser for C++. Many commercial compilers use a
table-driven parser, which is usually created by a parser-generator program. While table-driven parsers are generally
faster than other methods, they are hard to create by hand. Mini C++ uses a recursive-descent parser, which
implements in logic the production rules discussed in the previous section.

A recursive-descent parser is essentially a collection of mutually recursive functions that process an expression. If the
parser is used in a compiler, then it generates the proper object code that corresponds to the source code. However,
in an interpreter, the parser evaluates a given expression. In this section, the Mini C++ parser is developed.

籍
Note I have written extensively about expression parsing over the years. For an in-depth look at the process as it
relates to C++, refer to my book C++: The Complete Reference, McGraw-Hill/Osborne.

书
The Parser Code

经 典 读 书
The entire code for the Mini C++ recursive-descent parser is shown here. This code should be put into the file
parser.cpp. The operation of the parser is explained in the sections that follow.

// Recursive descent parser for integer expressions.

//

+ 不
#include <iostream>

+
#include <cstring>
#include <cstdlib>

C
#include <cctype>

典
#include "mccommon.h"

using namespace std;

无 经
// Keyword lookup table.
// Keywords must be entered lowercase.
struct commands {
char command[20];
token_ireps tok;
} com_table[] = {
"if", IF,
"else", ELSE,
"for", FOR,
"do", DO,
"while", WHILE,
"char", CHAR,
"int", INT,
"return", RETURN,
"switch", SWITCH,
"break", BREAK,
"case", CASE,
"cout", COUT,
"cin", CIN,
"", END // mark end of table
};

// This structure links a library function name

.

// with a pointer to that function.

struct intern_func_type {
char *f_name; // function name
int (*p)(); // pointer to the function
} intern_func[] = {
"getchar", call_getchar,
"putchar", call_putchar,
"abs", call_abs,
"rand", call_rand,
"", 0 // null terminate the list
};

// Entry point into parser.

void eval_exp(int &value)
{
get_token();

if(!*token) {
throw InterpExc(NO_EXP);
}

}
return;

书 籍
if(*token == ';') {
value = 0; // empty expression

典 书
eval_exp0(value);

经 读
putback(); // return last token read to input stream
}

+ + 不
// Process an assignment expression.
void eval_exp0(int &value)
{

C 典
// temp holds name of var receiving the assignment.
char temp[MAX_ID_LEN+1];

经
tok_types temp_tok;

if(token_type == IDENTIFIER) {

无
if(is_var(token)) { // if a var, see if assignment
strcpy(temp, token);
temp_tok = token_type;
get_token();
if(*token == '=') { // is an assignment
get_token();
eval_exp0(value); // get value to assign
assign_var(temp, value); // assign the value
return;
}
else { // not an assignment
putback(); // restore original token
strcpy(token, temp);
token_type = temp_tok;
}
}
}
eval_exp1(value);
}

// Process relational operators.

void eval_exp1(int &value)
.

{
int partial_value;
char op;
char relops[] = {
LT, LE, GT, GE, EQ, NE, 0
};

eval_exp2(value);

op = *token;
if(strchr(relops, op)) {
get_token();
eval_exp2(partial_value);

switch(op) { // perform the relational operation

case LT:
value = value < partial_value;
break;
case LE:
value = value <= partial_value;

籍
break;
case GT:
value = value > partial_value;

书
break;
case GE:

典 书
value = value >= partial_value;
break;
case EQ:

经 读
value = value == partial_value;
break;

+
case NE:

不
value = value != partial_value;

+
break;
}

C 典
}
}

经
// Add or subtract two terms.
void eval_exp2(int &value)
{

无
char op;
int partial_value;
char okops[] = {
'(', INC, DEC, '-', '+', 0
};

eval_exp3(value);

while((op = *token) == '+' || op == '-') {

get_token();

if(token_type == DELIMITER &&

!strchr(okops, *token))
throw InterpExc(SYNTAX);

eval_exp3(partial_value);

switch(op) { // add or subtract

case '-':
value = value - partial_value;
break;
case '+':
.

value = value + partial_value;

break;
}
}
}

// Multiply or divide two factors.

void eval_exp3(int &value)
{
char op;
int partial_value, t;
char okops[] = {
'(', INC, DEC, '-', '+', 0
};

eval_exp4(value);

while((op = *token) == '*' || op == '/'

|| op == '%') {
get_token();

籍
if(token_type == DELIMITER &&
!strchr(okops, *token))

书
throw InterpExc(SYNTAX);

典 书
eval_exp4(partial_value);

switch(op) { // mul, div, or modulus

经 读
case '*':
value = value * partial_value;

+
break;

不
case '/':

+
if(partial_value == 0)
throw InterpExc(DIV_BY_ZERO);

C 典
value = (value) / partial_value;
break;
case '%':

经
t = (value) / partial_value;
value = value - (t * partial_value);
break;

无
}
}
}

// Is a unary +, -, ++, or --.

void eval_exp4(int &value)
{
char op;
char temp;

op = '\0';
if(*token == '+' || *token == '-' ||
*token == INC || *token == DEC)
{
temp = *token;
op = *token;
get_token();
if(temp == INC)
assign_var(token, find_var(token)+1);
if(temp == DEC)
assign_var(token, find_var(token)-1);
}
.

eval_exp5(value);
if(op == '-') value = -(value);
}

// Process parenthesized expression.

void eval_exp5(int &value)
{

if((*token == '(')) {
get_token();

eval_exp0(value); // get subexpression

if(*token != ')')
throw InterpExc(PAREN_EXPECTED);
get_token();
}
else
atom(value);

籍
}

// Find value of number, variable, or function.

书
void atom(int &value)
{

典 书
int i;
char temp[MAX_ID_LEN+1];

经 读
switch(token_type) {
case IDENTIFIER:

+
i = internal_func(token);

不
if(i != -1) {

+
// Call "standard library" function.
value = (*intern_func[i].p)();

C 典
}
else if(find_func(token)) {
// Call programmer-created function.

经
call();
value = ret_value;
}

无
else {
value = find_var(token); // get var's value
strcpy(temp, token); // save variable name

// Check for ++ or --.

get_token();
if(*token == INC || *token == DEC) {
if(*token == INC)
assign_var(temp, find_var(temp)+1);
else
assign_var(temp, find_var(temp)-1);
} else putback();
}

get_token();
return;
case NUMBER: // is numeric constant
value = atoi(token);
get_token();

return;
case DELIMITER: // see if character constant
.

if(*token == '\'') {
value = *prog;
prog++;
if(*prog!='\'')
throw InterpExc(QUOTE_EXPECTED);

prog++;
get_token();

return ;
}
if(*token==')') return; // process empty expression
else throw InterpExc(SYNTAX); // otherwise, syntax error
default:
throw InterpExc(SYNTAX); // syntax error
}
}

// Display an error message.

void sntx_err(error_msg error)

籍
{
char *p, *temp;
int linecount = 0;

典 书
static char *e[]= {
"Syntax error",

书
"No expression present",
"Not a variable",

经 读
"Duplicate variable name",
"Duplicate function name",

+
"Semicolon expected",

不
"Unbalanced braces",

+
"Function undefined",
"Type specifier expected",

C 典
"Return without call",
"Parentheses expected",
"While expected",

经
"Closing quote expected",
"Division by zero",
"{ expected (control statements must use blocks)",

无
"Colon expected"
};

// Display error and line number.

cout << "\n" << e[error];
p = p_buf;
while(p != prog) { // find line number of error
p++;
if(*p == '\r') {
linecount++;
}
}
cout << " in line " << linecount << endl;

temp = p;
while(p > p_buf && *p != '\n') p--;

// Display offending line.

while(p <= temp)
cout << *p++;

cout << endl;

.

// Get a token.
tok_types get_token()
{

char *temp;

token_type = UNDEFTT; tok = UNDEFTOK;

temp = token;
*temp = '\0';

// Skip over white space.

while(isspace(*prog) && *prog) ++prog;

// Skip over newline.

while(*prog == '\r') {
++prog;
++prog;

籍
// Again, skip over white space.
while(isspace(*prog) && *prog) ++prog;
}

典 书
// Check for end of program.

书
if(*prog == '\0') {
*token = '\0';
tok = END;

经 读
return (token_type = DELIMITER);
}

+ + 不
// Check for block delimiters.
if(strchr("{}", *prog)) {
*temp = *prog;

C 典
temp++;
*temp = '\0';
prog++;

经
return (token_type = BLOCK);
}

无
// Look for comments.
if(*prog == '/')
if(*(prog+1) == '*') { // is a /* comment
prog += 2;
do { // find end of comment
while(*prog != '*') prog++;
prog++;
} while (*prog != '/');
prog++;
return (token_type = DELIMITER);
} else if(*(prog+1) == '/') { // is a // comment
prog += 2;
// Find end of comment.
while(*prog != '\r' && *prog != '\0') prog++;
if(*prog == '\r') prog +=2;
return (token_type = DELIMITER);
}

// Check for double-ops.

if(strchr("!<>=+-", *prog)) {
switch(*prog) {
case '=':
.

if(*(prog+1) == '=') {
prog++; prog++;
*temp = EQ;
temp++; *temp = EQ; temp++;
*temp = '\0';
}
break;
case '!':
if(*(prog+1) == '=') {
prog++; prog++;
*temp = NE;
temp++; *temp = NE; temp++;
*temp = '\0';
}
break;
case '<':
if(*(prog+1) == '=') {
prog++; prog++;
*temp = LE; temp++; *temp = LE;
}

籍
else if(*(prog+1) == '<') {
prog++; prog++;
*temp = LS; temp++; *temp = LS;

书
}
else {

典 书
prog++;
*temp = LT;
}

经 读
temp++;
*temp = '\0';

+
break;

不
case '>':

+
if(*(prog+1) == '=') {
prog++; prog++;

C 典
*temp = GE; temp++; *temp = GE;
} else if(*(prog+1) == '>') {
prog++; prog++;

经
*temp = RS; temp++; *temp = RS;
}
else {

无
prog++;
*temp = GT;
}
temp++;
*temp = '\0';
break;
case '+':
if(*(prog+1) == '+') {
prog++; prog++;
*temp = INC; temp++; *temp = INC;
temp++;
*temp = '\0';
}
break;
case '-':
if(*(prog+1) == '-') {
prog++; prog++;
*temp = DEC; temp++; *temp = DEC;
temp++;
*temp = '\0';
}
break;
.

if(*token) return(token_type = DELIMITER);

}

// Check for other delimiters.

if(strchr("+-*^/%=;:(),'", *prog)) {
*temp = *prog;
prog++;
temp++;
*temp = '\0';
return (token_type = DELIMITER);
}

// Read a quoted string.

if(*prog == '"') {
prog++;
while(*prog != '"' && *prog != '\r' && *prog) {
// Check for \n escape sequence.
if(*prog == '\\') {

籍
if(*(prog+1) == 'n') {
prog++;
*temp++ = '\n';

书
}
}

典 书
else if((temp - token) < MAX_T_LEN)
*temp++ = *prog;

经 读
prog++;
}

+
if(*prog == '\r' || *prog == 0)

不
throw InterpExc(SYNTAX);

+
prog++; *temp = '\0';
return (token_type = STRING);

C 典
}

// Read an integer number.

经
if(isdigit(*prog)) {
while(!isdelim(*prog)) {
if((temp - token) < MAX_ID_LEN)

无
*temp++ = *prog;
prog++;
}
*temp = '\0';
return (token_type = NUMBER);
}

// Read identifier or keyword.

if(isalpha(*prog)) {
while(!isdelim(*prog)) {
if((temp - token) < MAX_ID_LEN)
*temp++ = *prog;
prog++;
}
token_type = TEMP;
}

*temp = '\0';

// Determine if token is a keyword or identifier.

if(token_type == TEMP) {
tok = look_up(token); // convert to internal form
.

if(tok) token_type = KEYWORD; // is a keyword

else token_type = IDENTIFIER;
}

// Check for unidentified character in file.

if(token_type == UNDEFTT)
throw InterpExc(SYNTAX);

return token_type;
}

// Return a token to input stream.

void putback()
{
char *t;

t = token;
for(; *t; t++) prog--;
}

籍
// Look up a token's internal representation in the
// token table.
token_ireps look_up(char *s)

书
{
int i;

典 书
char *p;

// Convert to lowercase.

经 读
p = s;
while(*p) { *p = tolower(*p); p++; }

+ + 不
// See if token is in table.
for(i=0; *com_table[i].command; i++) {
if(!strcmp(com_table[i].command, s))

C 典
return com_table[i].tok;
}

经
return UNDEFTOK; // unknown command
}

无
// Return index of internal library function or -1 if
// not found.
int internal_func(char *s)
{
int i;

for(i=0; intern_func[i].f_name[0]; i++) {

if(!strcmp(intern_func[i].f_name, s)) return i;
}
return -1;
}

// Return true if c is a delimiter.

bool isdelim(char c)
{
if(strchr(" !:;,+-<>'/*%^=()", c) || c == 9 ||
c == '\r' || c == 0) return true;
return false;
}

The parser uses the following global variables and enumerations (which are declared within the mccommon.h header
file shown later in this chapter):
.

const int MAX_T_LEN = 128; // max token length

const int MAX_ID_LEN = 31; // max identifier length

// Enumeration of token types.

enum tok_types { UNDEFTT, DELIMITER, IDENTIFIER,
NUMBER, KEYWORD, TEMP, STRING, BLOCK };

// Enumeration of internal representation of tokens.

enum token_ireps { UNDEFTOK, ARG, CHAR, INT, SWITCH,
CASE, IF, ELSE, FOR, DO, WHILE, BREAK,
RETURN, COUT, CIN, END };

// Enumeration of two-character operators, such as <=.

enum double_ops { LT=1, LE, GT, GE, EQ, NE, LS, RS, INC, DEC };

// These are the constants used when throwing a

// syntax error exception.
//
// NOTE: SYNTAX is a generic error message used when
// nothing else seems appropriate.

籍
enum error_msg
{ SYNTAX, NO_EXP, NOT_VAR, DUP_VAR, DUP_FUNC,
SEMI_EXPECTED, UNBAL_BRACES, FUNC_UNDEF,

书
TYPE_EXPECTED, RET_NOCALL, PAREN_EXPECTED,
WHILE_EXPECTED, QUOTE_EXPECTED, DIV_BY_ZERO,

典 书
BRACE_EXPECTED, COLON_EXPECTED };

extern char *prog; // current location in source code

经 读
extern char *p_buf; // points to start of program buffer

+
extern char token[MAX_T_LEN+1]; // string version of token

不
extern tok_types token_type; // contains type of token

+
extern token_ireps tok; // internal representation of token

C 典
extern int ret_value; // function return value

// Exception class for Mini C++.

经
class InterpExc {
error_msg err;
public:

无
InterpExc(error_msg e) { err = e; }
error_msg get_err() { return err; }
};

The location in the source code at which execution is currently occurring is pointed to by prog. Thus, prog holds the
address at which the next piece of the program will be read by the interpreter. The p_buf pointer is unchanged by the
interpreter and always points to the start of the program being interpreted. The current token is stored in token. (A
token is an indivisible unit of program code. It is more precisely defined by the following section.) The type of a token is
stored in token_type. The tok variable holds the internal format of the token if that token is a keyword.

The enumeration tok_types declares the types of tokens recognized by Mini C++. The token_ireps enumeration
specifies the internal format of tokens that represent keywords. Values that represent operators comprised of two
characters, such as < = , are specified by the double_ops enumeration. Various error codes are enumerated by
error_msg. Finally, InterpExc is the exception class used to report syntax errors.

Tokenizing the Source Code

Fundamental to all interpreters (and compilers, for that matter) is a mechanism that reduces the source code into its
constituent parts, which are referred to as tokens. A token is an indivisible piece of a program that represents one
logical unit. For example, the {, + , = =, and if are all tokens. Even though the equality operator = = is comprised of two
characters, they cannot be separated without changing the meaning. Thus, = = is a single logical unit, and therefore, a
.

single token. Similarly, if is a token, but neither i nor f has any special meaning to C++.

The ANSI/ISO standard for C++ defines the following types of tokens:

keywords identifiers literal

operator punctuator

The keywords are those tokens that make up the C++ language, such as while. Identifiers are the names of variables,
functions, and so on. Literals include strings and constant values, such as the number 10. Operators are self
explanatory. Punctuators include such things as semicolons, commas, braces, and parentheses. (Some punctuators
are also operators, depending upon their use.)

Although Mini C++ recognizes the same tokens as those specified by the ANSI/ISO C++ standard, it categorizes them
a bit differently in order to make interpretation a bit easier. The token categories used by Mini C++ are shown here:

Token Type Includes

delimiter punctuators and operators

籍
keyword keywords

书
string quoted strings

identifier variable and function names

典 书
number numeric constant

block { or }

+ 经 读
Let’s work through an example that illustrates how these token types are applied. Given the statement

不
for(x=0; x<10; x++) {

+
num = num + x;
}

C 典
the following tokens are produced:

经
Token Category

for keyword

0
无 delimiter

identifier

delimiter

number

; delimiter

x identifier

< delimiter

10 number

; delimiter

x identifier

++ delimiter

) delimiter
.

Token Category

{ block

num identifier

= delimiter

num identifier

+ delimiter

x identifier

; delimiter

} block

The function that returns tokens from the source code for the Mini C++ interpreter is called get_token( ), and it is
shown here:

籍
// Get a token.
tok_types get_token()
{

书
char *temp;

典 书
token_type = UNDEFTT; tok = UNDEFTOK;

temp = token;

经 读
*temp = '\0';

+
// Skip over white space.

不
while(isspace(*prog) && *prog) ++prog;

C+
// Skip over newline.

++prog;
++prog;

典
while(*prog == '\r') {

经
// Again, skip over white space.
while(isspace(*prog) && *prog) ++prog;
}

}
无
// Check for end of program.
if(*prog == '\0') {
*token = '\0';
tok = END;
return (token_type = DELIMITER);

// Check for block delimiters.

if(strchr("{}", *prog)) {
*temp = *prog;
temp++;
*temp = '\0';
prog++;
return (token_type = BLOCK);
}

// Look for comments.

if(*prog == '/')
if(*(prog+1) == '*') { // is a /* comment
prog += 2;
.

do { // find end of comment

while(*prog != '*') prog++;
prog++;
} while (*prog != '/');
prog++;
return (token_type = DELIMITER);
} else if(*(prog+1) == '/') { // is a // comment
prog += 2;
// Find end of comment.
while(*prog != '\r' && *prog != '\0') prog++;
if(*prog == '\r') prog +=2;
return (token_type = DELIMITER);
}

// Check for double-ops.

if(strchr("!<>=+-", *prog)) {
switch(*prog) {
case '=':
if(*(prog+1) == '=') {
prog++; prog++;

籍
*temp = EQ;
temp++; *temp = EQ; temp++;
*temp = '\0';

书
}
break;

典 书
case '!':
if(*(prog+1) == '=') {
prog++; prog++;

经 读
*temp = NE;
temp++; *temp = NE; temp++;

+
*temp = '\0';

不
}

+
break;
case '<':

C 典
if(*(prog+1) == '=') {
prog++; prog++;
*temp = LE; temp++; *temp = LE;

经
}
else if(*(prog+1) == '<') {
prog++; prog++;

无
*temp = LS; temp++; *temp = LS;
}
else {
prog++;
*temp = LT;
}
temp++;
*temp = '\0';
break;
case '>':
if(*(prog+1) == '=') {
prog++; prog++;
*temp = GE; temp++; *temp = GE;
} else if(*(prog+1) == '>') {
prog++; prog++;
*temp = RS; temp++; *temp = RS;
}
else {
prog++;
*temp = GT;
}
temp++;
.

*temp = '\0';
break;
case '+':
if(*(prog+1) == '+') {
prog++; prog++;
*temp = INC; temp++; *temp = INC;
temp++;
*temp = '\0';
}
break;
case '-':
if(*(prog+1) == '-') {
prog++; prog++;
*temp = DEC; temp++; *temp = DEC;
temp++;
*temp = '\0';
}
break;
}

籍
if(*token) return(token_type = DELIMITER);
}

书
// Check for other delimiters.
if(strchr("+-*^/%=;:(),'", *prog)) {

典 书
*temp = *prog;
prog++;
temp++;

经 读
*temp = '\0';
return (token_type = DELIMITER);

+
}

+ 不
// Read a quoted string.
if(*prog == '"') {

C 典
prog++;
while(*prog != '"' && *prog != '\r' && *prog) {
// Check for \n escape sequence.

经
if(*prog == '\\') {
if(*(prog+1) == 'n') {
prog++;

无
*temp++ = '\n';
}
}
else if((temp - token) < MAX_T_LEN)
*temp++ = *prog;

prog++;
}
if(*prog == '\r' || *prog == 0)
throw InterpExc(SYNTAX);
prog++; *temp = '\0';
return (token_type = STRING);
}

// Read an integer number.

if(isdigit(*prog)) {
while(!isdelim(*prog)) {
if((temp - token) < MAX_ID_LEN)
*temp++ = *prog;
prog++;
}
*temp = '\0';
 return (token_type = NUMBER);
}

// Read identifier or keyword.

if(isalpha(*prog)) {
while(!isdelim(*prog)) {
if((temp - token) < MAX_ID_LEN)
*temp++ = *prog;
prog++;
}
token_type = TEMP;
}

*temp = '\0';

// Determine if token is a keyword or identifier.

if(token_type == TEMP) {
tok = look_up(token); // convert to internal form
if(tok) token_type = KEYWORD; // is a keyword
else token_type = IDENTIFIER;

籍
}

// Check for unidentified character in file.

书
if(token_type == UNDEFTT)
throw InterpExc(SYNTAX);

典 书
return token_type;
}

经 读
The get_token( ) function begins by skipping over all white space, including carriage returns and line feeds. Because

+
no token contains a space (except for a quoted string or character constant), spaces must be bypassed. The

不
get_token( ) function also skips over comments. Then the next token in the program is read. Notice how each category

+
of token is handled separately. For example, if the next character in the program is a digit, a number is read; if the next
character is a letter, an identifier or keyword is obtained, and so on.

C 典
The string representation of the token is placed into token. Once read, the token’s type (as enumerated by the
tok_types enumeration) is put into token_type and, if the token is a keyword, its internal representation (as enumerated

经
by token_ireps) is assigned to tok via the look_up( ) function. (The reason for the internal representation of keywords
will be explained later.) As you can see, get_token( ) converts C++’s two-character operators, such as <= or ++, into
their corresponding enumeration values. Although not technically necessary, this step makes the parser easier to

无
implement.

Displaying Syntax Errors

If the parser encounters a syntax error, it throws an InterpExc, specifying an enumerated value that corresponds to the
type of error found. (The other parts of Mini C++ also use InterpExcs to report errors.) The handler for InterpExc
exceptions is found in main( ), which is part of the main interpreter file, minicpp.cpp, described later. This handler calls
the sntx_err( ) function, shown here, to report the error:
// Display an error message.
void sntx_err(error_msg error)
{
char *p, *temp;
int linecount = 0;

static char *e[]= {

"Syntax error",
"No expression present",
"Not a variable",
"Duplicate variable name",
"Duplicate function name",
.

"Semicolon expected",
"Unbalanced braces",
"Function undefined",
"Type specifier expected",
"Return without call",
"Parentheses expected",
"While expected",
"Closing quote expected",
"Division by zero",
"{ expected (control statements must use blocks)",
"Colon expected"
};

// Display error and line number.

cout << "\n" << e[error];
p = p_buf;
while(p != prog) { // find line number of error
p++;
if(*p == '\r') {
linecount++;

籍
}
}
cout << " in line " << linecount << endl;

temp = p;

典 书 书
while(p > p_buf && *p != '\n') p--;

// Display offending line.

经 读
while(p <= temp)
cout << *p++;

+ 不
cout << endl;

+
C
Notice that sntx_err( ) displays a string that describes the error and the line number in which the error was detected

典
(which may be one line after the error actually occurred). It also displays the line in which the error was detected.

经
Evaluating an Expression

The functions that begin with eval_exp and the atom( ) function implement the production rules for Mini C++

无
expressions. To understand exactly how the parser evaluates an expression, work through the following expression.
(Assume that prog points to the start of the expression.)
10 – 3 * 2

When eval_exp( ), the entry point into the parser, is called, it gets the first token. If the token is null, an exception is
thrown that indicates that no expression was present. However, in this case, the token contains the number 10. Since
the token is not null, eval_exp0( ) is called, which checks for an assignment operator. Because no assignment is
present, eval_exp1( ) is called, which in turn calls eval_exp2( ). Next, eval_exp2( ) calls eval_exp3( ) and eval_exp3( )
calls eval_exp4( ). The eval_exp4( ) handles the unary + and –, and the prefix ++ and – –. It then calls eval_exp5( ). At
this point eval_exp5( ) either recursively calls eval_exp0( ) (in the case of a parenthesized expression) or atom( ) to find
a value. Because the token is not a left parenthesis, atom( ) is executed and value is assigned the value 10. Next,
another token is retrieved, and the functions begin to return up the chain. The token is now the operator –, and the
functions return up to eval_exp2( ).

What happens next is very important. Because the token is –, it is saved in op. The parser then gets the next token,
which is 3, and the descent down the chain begins again. As before, atom( ) is entered. The value 3 is returned in
value and the operator * is read. This causes a return back up the chain to eval_exp3( ), where the final token 2 is read.
At this point, the first arithmetic operation occurs—the multiplication of 2 and 3. The result is returned to eval_exp2( )
and the subtraction is performed. The subtraction yields the answer 4. Although this process may seem complicated at
first, working through some other examples on your own will clarify the parser’s operation.
Now, let’s look a bit more closely at the atom( ) function. It finds the value of an integer constant or variable, a function,
or a character constant. (It also handles postfix ++ and – –.) There are two kinds of functions that may be present in the
source code: user-defined or library. If a user-defined function is encountered, its code is executed by the interpreter in
order to determine its return value. (How Mini C++ actually achieves a function call is discussed in the next section.) If
the function is a library function, first its address is looked up by the internal_func( ) function, and then it is accessed
via its interface function. The library functions and the addresses of their interface functions are held in the intern_func
array, shown here:
// This structure links a library function name
// with a pointer to that function.
struct intern_func_type {
char *f_name; // function name
int (*p)(); // pointer to the function
} intern_func[] = {
"getchar", call_getchar,
"putchar", call_putchar,
"abs", call_abs,
"rand", call_rand,
"", 0 // null terminate the list
};

Functions.?)

书 籍
As you can see, Mini C++ knows only a few library functions, but it is easy to add others that you might need. (The
actual interface functions are contained in a separate file, which is discussed in the section ?The Mini C++ Library

One final point about the routines in the expression parser file: To correctly parse the C++ syntax, one-token

典 书
look-ahead is occasionally required. For example, consider the following statement:
alpha = count();

+ 经 读
In order for Mini C++ to know that count is a function and not a variable, it must read both count and the next token. If
the next token is a parenthesis, as it is in this case, then the parser knows that count refers to a function. However, if

不
the statement was

+
alpha = count * 10;

C 典
then the next token after count is the *. Because it is not a parenthesis, it means that count must be a variable.
Without one-token look-ahead, the parser has no way to make this determination. In both cases, the look-ahead token
needs to be returned to the input stream for subsequent processing. For this reason, the expression parser file

经
includes the putback( ) function, which returns the last token read to the input stream.

无
.

The Mini C++ Interpreter

The engine that actually executes a C++ program is the interpreter. Before jumping right into the code, it will be helpful
for you to understand in general how the interpreter operates. In many ways, the code for the interpreter is easier to
understand than the expression parser because, conceptually, the act of interpreting a C++ program can be summed
up by the following algorithm:
while(tokens_present) {
get next token;
take appropriate action;
}

This algorithm may seem unbelievably simple when compared to the expression parser, but this is exactly what all
interpreters do! One thing to keep in mind is that the “take appropriate action” step may also involve reading additional
tokens from the input stream and taking further appropriate action. Thus, the “take appropriate action” step can be
recursive.

籍
To understand how the algorithm actually works, let’s manually interpret the following code fragment:
int a;

a = 10 * num;

if(a < 100) {

cout << a;

典 书 书
读
}

+ 经
Following the algorithm, read the first token, which is int. The appropriate action given this token is to read the next

不
token in order to find out the name of the variable being declared (which, in this case, is a) and then to store its name.

+
The next token is the semicolon that ends the line. The appropriate action here is to ignore it. Now, get the next token,
which is a. Because this statement does not begin with a keyword, it must begin an expression. Here, the appropriate

C
action is to evaluate the expression using the parser. This process eats up all the tokens in that statement. Next, read

典
the if token. This signals the beginning of an if statement, so the appropriate action is to process the if, which means
evaluating the conditional expression. If the outcome of the expression is true, then the block of code associated with

经
the if is interpreted. Otherwise, it is skipped.

The process just described will take place for any C++ program. Interpretation stops only when the last token has

无
been read or when a syntax error is encountered. With this basic algorithm in mind, let’s examine the interpreter. The
entire interpreter file, called minicpp.cpp, is shown here. The sections that follow examine it in detail.
// A Mini C++ interpreter.

#include <iostream>
#include <fstream>
#include <new>
#include <stack>
#include <vector>
#include <cstring>
#include <cstdlib>
#include <cctype>
#include "mccommon.h"

using namespace std;

char *prog; // current execution point in source code

char *p_buf; // points to start of program buffer

// This structure encapsulates the info

// associated with variables.
.

struct var_type {
char var_name[MAX_ID_LEN+1]; // name
token_ireps v_type; // data type
int value; // value
};

// This vector holds info for global variables.

vector<var_type> global_vars;

// This vector holds info for local variables

// and parameters.
vector<var_type> local_var_stack;

// This structure encapsulates function info.

struct func_type {
char func_name[MAX_ID_LEN+1]; // name
token_ireps ret_type; // return type
char *loc; // location of entry point in program
};

籍
// This vector holds info about functions.
vector<func_type> func_table;

书
// Stack for managing function scope.
stack<int> func_call_stack;

典 书
// Stack for managing nested scopes.
stack<int> nest_scope_stack;

+ 经 读
char token[MAX_T_LEN+1]; // current token
tok_types token_type; // token type

不
token_ireps tok; // internal representation

C+
int ret_value; // function return value

典
bool breakfound = false; // true if break encountered

经
int main(int argc, char *argv[])
{
if(argc != 2) {

无
cout << "Usage: minicpp <filename>\n";
return 1;
}

// Allocate memory for the program.

try {
p_buf = new char[PROG_SIZE];
} catch (bad_alloc exc) {
cout << "Could Not Allocate Program Buffer\n";
return 1;
}

// Load the program to execute.

if(!load_program(p_buf, argv[1])) return 1;

// Set program pointer to start of program buffer.

prog = p_buf;

try {
// Find the location of all functions
// and global variables in the program.
prescan();
.

// Next, set up the call to main().

// Find program starting point.

prog = find_func("main");

// Check for incorrect or missing main() function.

if(!prog) {
cout << "main() Not Found\n";
return 1;
}

// Back up to opening (.
prog--;

// Set the first token to main

strcpy(token, "main");

// Call main() to start interpreting.

call();

籍
}
catch(InterpExc exc) {
sntx_err(exc.get_err());

书
return 1;
}

典 书
catch(bad_alloc exc) {
cout << "Out Of Memory\n";
return 1;

经 读
}

+
return ret_value;

不
}

{
+
// Load a program.

C
int i=0;

典
bool load_program(char *p, char *fname)

经
ifstream in(fname, ios::in | ios::binary);
if(!in) {

无
cout << "Cannot Open file.\n";
return false;
}

do {
*p = in.get();
p++; i++;
} while(!in.eof() && i < PROG_SIZE);

if(i == PROG_SIZE) {
cout << "Program Too Big\n";
return false;
}

// Null terminate the program. Skip any EOF

// mark if present in the file.
if(*(p-2) == 0x1a) *(p-2) = '\0';
else *(p-1) = '\0';

in.close();

return true;
.

// Find the location of all functions in the program

// and store global variables.
void prescan()
{
char *p, *tp;
char temp[MAX_ID_LEN+1];
token_ireps datatype;
func_type ft;

// When brace is 0, the current source position

// is outside of any function.
int brace = 0;

p = prog;

do {
// Bypass code inside functions
while(brace) {

籍
get_token();
if(tok == END) throw InterpExc(UNBAL_BRACES);
if(*token == '{') brace++;

书
if(*token == '}') brace--;
}

典 书
tp = prog; // save current position
get_token();

+ 经 读
// See if global var type or function return type.
if(tok==CHAR || tok==INT) {

不
datatype = tok; // save data type

+
get_token();

C 典
if(token_type == IDENTIFIER) {
strcpy(temp, token);
get_token();

经
if(*token != '(') { // must be global var
prog = tp; // return to start of declaration

无
decl_global();
}
else if(*token == '(') { // must be a function

// See if function already defined.

for(unsigned i=0; i < func_table.size(); i++)
if(!strcmp(func_table[i].func_name, temp))
throw InterpExc(DUP_FUNC);

ft.loc = prog;
ft.ret_type = datatype;
strcpy(ft.func_name, temp);
func_table.push_back(ft);

do {
get_token();
} while(*token != ')');
// Next token will now be opening curly
// brace of function.
}
else putback();
}
.

}
else {
if(*token == '{') brace++;
if(*token == '}') brace--;
}
} while(tok != END);
if(brace) throw InterpExc(UNBAL_BRACES);
prog = p;
}

// Interpret a single statement or block of code. When

// interp() returns from its initial call, the final
// brace (or a return) in main() has been encountered.
void interp()
{
int value;
int block = 0;

do {
// Don't interpret until break is handled.

籍
if(breakfound) return;

token_type = get_token();

典 书
// See what kind of token is up.

书
if(token_type == IDENTIFIER ||
*token == INC || *token == DEC)

经 读
// Not a keyword, so process expression.
putback(); // restore token to input stream for

+
// further processing by eval_exp()

不
eval_exp(value); // process the expression

+
if(*token != ';') throw InterpExc(SEMI_EXPECTED);
}

C 典
else if(token_type==BLOCK) { // block delimiter?
if(*token == '{') { // is a block
block = 1; // interpreting block, not statement

经
// Record nested scope.
nest_scope_stack.push(local_var_stack.size());
}

无
else { // is a }, so reset scope and return
// Reset nested scope.
local_var_stack.resize(nest_scope_stack.top());
nest_scope_stack.pop();
return;
}
}
else // is keyword
switch(tok) {
case CHAR:
case INT: // declare local variables
putback();
decl_local();
break;
case RETURN: // return from function call
func_ret();
return;
case IF: // process an if statement
exec_if();
break;
case ELSE: // process an else statement
find_eob(); // find end of else block
.

// and continue execution

break;
case WHILE: // process a while loop
exec_while();
break;
case DO: // process a do-while loop
exec_do();
break;
case FOR: // process a for loop
exec_for();
break;
case BREAK: // handle break
breakfound = true;

// Reset nested scope.

local_var_stack.resize(nest_scope_stack.top());
nest_scope_stack.pop();
return;
case SWITCH: // handle a switch statement
exec_switch();

籍
break;
case COUT: // handle console output
exec_cout();

书
break;
case CIN: // handle console input

典 书
exec_cin();
break;
case END:

经 读
exit(0);
}

+
} while (tok != END && block);

不
return;

+
}

C 典
// Return the entry point of the specified function.
// Return NULL if not found.
char *find_func(char *name)

经
{
unsigned i;

无
for(i=0; i < func_table.size(); i++)
if(!strcmp(name, func_table[i].func_name))
return func_table[i].loc;

return NULL;
}

// Declare a global variable.

void decl_global()
{
token_ireps vartype;
var_type vt;

get_token(); // get type

vartype = tok; // save var type

// Process comma-separated list.

do {
vt.v_type = vartype;
vt.value = 0; // init to 0
get_token(); // get name
.

// See if variable is a duplicate.

for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, token))
throw InterpExc(DUP_VAR);

strcpy(vt.var_name, token);
global_vars.push_back(vt);

get_token();
} while(*token == ',');

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

}

// Declare a local variable.

void decl_local()
{
var_type vt;

籍
get_token(); // get var type
vt.v_type = tok; // store type

书
vt.value = 0; // init var to 0

典 书
// Process comma-separated list.
do {
get_token(); // get var name

+ 经 读
// See if variable is already the name
// of a local variable in this scope.

不
if(!local_var_stack.empty())

+
for(int i=local_var_stack.size()-1;
i >= nest_scope_stack.top(); i--)

C 典
{
if(!strcmp(local_var_stack[i].var_name, token))
throw InterpExc(DUP_VAR);

经
}

strcpy(vt.var_name, token);

无
local_var_stack.push_back(vt);
get_token();
} while(*token == ',');

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

}

// Call a function.
void call()
{
char *loc, *temp;
int lvartemp;

// First, find entry point of function.

loc = find_func(token);

if(loc == NULL)
throw InterpExc(FUNC_UNDEF); // function not defined
else {
// Save local var stack index.
lvartemp = local_var_stack.size();
.

get_args(); // get function arguments

temp = prog; // save return location

func_call_stack.push(lvartemp); // push local var index

prog = loc; // reset prog to start of function

get_params(); // load the function's parameters with
// the values of the arguments

interp(); // interpret the function

prog = temp; // reset the program pointer

if(func_call_stack.empty()) throw InterpExc(RET_NOCALL);

// Reset local_var_stack to its previous state.

local_var_stack.resize(func_call_stack.top());
func_call_stack.pop();
}
}

void get_args()
{

书 籍
// Push the arguments to a function onto the local
// variable stack.

典 书
int value, count, temp[NUM_PARAMS];
var_type vt;

经 读
count = 0;
get_token();

+
if(*token != '(') throw InterpExc(PAREN_EXPECTED);

+ 不
// Process a comma-separated list of values.
do {

C 典
eval_exp(value);
temp[count] = value; // save temporarily
get_token();

经
count++;
} while(*token == ',');
count--;

}
无
// Now, push on local_var_stack in reverse order.
for(; count>=0; count--) {
vt.value = temp[count];
vt.v_type = ARG;
local_var_stack.push_back(vt);

// Get function parameters.

void get_params()
{
var_type *p;
int i;

i = local_var_stack.size()-1;

// Process comma-separated list of parameters.

do {
get_token();
p = &local_var_stack[i];
if(*token != ')' ) {
.

if(tok != INT && tok != CHAR)

throw InterpExc(TYPE_EXPECTED);

p->v_type = tok;
get_token();

// Link parameter name with argument already on

// local var stack.
strcpy(p->var_name, token);
get_token();
i--;
}
else break;
} while(*token == ',');

if(*token != ')') throw InterpExc(PAREN_EXPECTED);

}

// Return from a function.

void func_ret()

籍
{
int value;

书
value = 0;

典 书
// Get return value, if any.
eval_exp(value);

经 读
ret_value = value;
}

+ + 不
// Assign a value to a variable.
void assign_var(char *vname, int value)
{

C 典
// First, see if it's a local variable.
if(!local_var_stack.empty())
for(int i=local_var_stack.size()-1;

经
i >= func_call_stack.top(); i--)
{
if(!strcmp(local_var_stack[i].var_name,

无
vname))
{
if(local_var_stack[i].v_type == CHAR)
local_var_stack[i].value = (char) value;
else if(local_var_stack[i].v_type == INT)
local_var_stack[i].value = value;
return;
}
}

// Otherwise, try global vars.

for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, vname)) {
if(global_vars[i].v_type == CHAR)
global_vars[i].value = (char) value;
else if(global_vars[i].v_type == INT)
global_vars[i].value = value;
return;
}

throw InterpExc(NOT_VAR); // variable not found

}
.

// Find the value of a variable.

int find_var(char *vname)
{
// First, see if it's a local variable.
if(!local_var_stack.empty())
for(int i=local_var_stack.size()-1;
i >= func_call_stack.top(); i--)
{
if(!strcmp(local_var_stack[i].var_name, vname))
return local_var_stack[i].value;
}

// Otherwise, try global vars.

for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, vname))
return global_vars[i].value;

throw InterpExc(NOT_VAR); // variable not found

}

{
int cond;

书 籍
// Execute an if statement.
void exec_if()

典 书
eval_exp(cond); // get if expression.

经 读
if(cond) { // if true, process target of IF
// Confirm start of block.

+
if(*token != '{')

不
throw InterpExc(BRACE_EXPECTED);

}
else {
+
interp();

C 典
// Otherwise skip around IF block and

经
// process the ELSE, if present.

find_eob(); // find start of next line

无
get_token();

if(tok != ELSE) {
// Restore token if no ELSE is present.
putback();
return;
}

// Confirm start of block.

get_token();
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);
putback();

interp();
}
}

// Execute a switch statement.

void exec_switch()
{
int sval, cval;
.

int brace;

eval_exp(sval); // Get switch expression.

// Check for start of block.

if(*token != '{')
throw InterpExc(BRACE_EXPECTED);

// Record new scope.

nest_scope_stack.push(local_var_stack.size());

// Now, check case statements.

for(;;) {
brace = 1;
// Find a case statement.
do {
get_token();
if(*token == '{') brace++;
else if(*token == '}') brace--;
} while(tok != CASE && tok != END && brace);

籍
// If no matching case found, then skip.
if(!brace) break;

书
if(tok == END) throw InterpExc(SYNTAX);

典 书
// Get value of the case statement.
eval_exp(cval);

+ 经
get_token();

不 读
// Read and discard the :

+
if(*token != ':')
throw InterpExc(COLON_EXPECTED);

C 典
// If values match, then interpret.
if(cval == sval) {

经
brace = 1;
do {
interp();

无
if(*token == '{') brace++;
else if(*token == '}') brace--;
} while(!breakfound && tok != END && brace);

// Find end of switch statement.

while(brace) {
get_token();
if(*token == '{') brace++;
else if(*token == '}') brace--;
}
breakfound = false;

break;
}
}
}

// Execute a while loop.

void exec_while()
{
int cond;
.

char *temp;

putback(); // put back the while

temp = prog; // save location of top of while loop

get_token();
eval_exp(cond); // check the conditional expression

// Confirm start of block.

if(*token != '{')
throw InterpExc(BRACE_EXPECTED);

if(cond)
interp(); // if true, interpret
else { // otherwise, skip to end of loop
find_eob();
return;
}

prog = temp; // loop back to top

书 籍
// Check for break in loop.
if(breakfound) {
// Find start of loop block.
do {

典 书
get_token();
} while(*token != '{' && tok != END);

经 读
putback();
breakfound = false;

+
find_eob(); // now, find end of loop

不
return;

+
}
}

C 典
// Execute a do loop.
void exec_do()

经
{
int cond;
char *temp;

无
// Save location of top of do loop.
putback(); // put back do
temp = prog;

get_token(); // get start of loop block

// Confirm start of block.

get_token();
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);
putback();

interp(); // interpret loop

// Check for break in loop.

if(breakfound) {
prog = temp;
// Find start of loop block.
do {
get_token();
.

} while(*token != '{' && tok != END);

// Find end of while block.

putback();
find_eob();

// Now, find end of while expression.

do {
get_token();
} while(*token != ';' && tok != END);
if(tok == END) throw InterpExc(SYNTAX);

breakfound = false;
return;
}

get_token();
if(tok != WHILE) throw InterpExc(WHILE_EXPECTED);

eval_exp(cond); // check the loop condition

}
籍
// If true loop; otherwise, continue on.
if(cond) prog = temp;

书
典 书
// Execute a for loop.
void exec_for()
{

经 读
int cond;
char *temp, *temp2;

+
int paren ;

+ 不
get_token(); // skip opening (
eval_exp(cond); // initialization expression

C 典
if(*token != ';') throw InterpExc(SEMI_EXPECTED);
prog++; // get past the ;

经
temp = prog;

for(;;) {

无
// Get the value of the conditional expression.
eval_exp(cond);

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

prog++; // get past the ;
temp2 = prog;

// Find start of for block.

paren = 1;
while(paren) {
get_token();
if(*token == '(') paren++;
if(*token == ')') paren--;
}

// Confirm start of block.

get_token();
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);
putback();

// If condition is true, interpret

.

if(cond)
interp();
else { // otherwise, skip to end of loop
find_eob();
return;
}

prog = temp2; // go to increment expression

// Check for break in loop.

if(breakfound) {
// Find start of loop block.
do {
get_token();
} while(*token != '{' && tok != END);

putback();
breakfound = false;
find_eob(); // now, find end of loop
return;

籍
}

// Evaluate the increment expression.

书
eval_exp(cond);

典 书
prog = temp; // loop back to top
}
}

void exec_cout()
{

+ 经 不 读
// Execute a cout statement.

+
int val;

C 典
get_token();
if(*token != LS) throw InterpExc(SYNTAX);

经
do {
get_token();

无
if(token_type==STRING) {
// Output a string.
cout << token;
}
else if(token_type == NUMBER ||
token_type == IDENTIFIER) {
// Output a number.
putback();
eval_exp(val);
cout << val;
}
else if(*token == '\'') {
// Output a character constant.
putback();
eval_exp(val);
cout << (char) val;
}

get_token();
} while(*token == LS);

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

.

// Execute a cin statement.

void exec_cin()
{
int val;
char chval;
token_ireps vtype;

get_token();
if(*token != RS) throw InterpExc(SYNTAX);

do {
get_token();
if(token_type != IDENTIFIER)
throw InterpExc(NOT_VAR);

vtype = find_var_type(token);

if(vtype == CHAR) {

籍
cin >> chval;
assign_var(token, chval);
}

书
else if(vtype == INT) {
cin >> val;

典 书
assign_var(token, val);
}

经 读
get_token();
} while(*token == RS);

+ 不
if(*token != ';') throw InterpExc(SEMI_EXPECTED);

+
C 典
// Find the end of a block.
void find_eob()
{

经
int brace;

get_token();

无
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);

brace = 1;

do {
get_token();
if(*token == '{') brace++;
else if(*token == '}') brace--;
} while(brace && tok != END);

if(tok==END) throw InterpExc(UNBAL_BRACES);

}

// Determine if an identifier is a variable. Return

// true if variable is found; false otherwise.
bool is_var(char *vname)
{
// See if vname is a local variable.
if(!local_var_stack.empty())
for(int i=local_var_stack.size()-1;
i >= func_call_stack.top(); i--)
.

{
if(!strcmp(local_var_stack[i].var_name, vname))
return true;
}

// See if vname is a global variable.

for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, vname))
return true;

return false;
}

// Return the type of variable.

token_ireps find_var_type(char *vname)
{
// First, see if it's a local variable.
if(!local_var_stack.empty())
for(int i=local_var_stack.size()-1;
i >= func_call_stack.top(); i--)

籍
{
if(!strcmp(local_var_stack[i].var_name, vname))
return local_var_stack[i].v_type;

书
}

典 书
// Otherwise, try global vars.
for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, vname))

经 读
return local_var_stack[i].v_type;

+
return UNDEFTOK;

不
}

+
The main( ) Function

C 典
The main( ) function begins the interpretation of the program that you specify on the command line. It is shown here:
int main(int argc, char *argv[])

经
{
if(argc != 2) {

无
cout << "Usage: minicpp <filename>\n";
return 1;
}

// Allocate memory for the program.

try {
p_buf = new char[PROG_SIZE];
} catch (bad_alloc exc) {
cout << "Could Not Allocate Program Buffer\n";
return 1;
}

// Load the program to execute.

if(!load_program(p_buf, argv[1])) return 1;

// Set program pointer to start of program buffer.

prog = p_buf;

try {
// Find the location of all functions
// and global variables in the program.
prescan();
.

// Next, set up the call to main().

// Find program starting point.

prog = find_func("main");

// Check for incorrect or missing main() function.

if(!prog) {
cout << "main() Not Found\n";
return 1;
}

// Back up to opening (.
prog--;

// Set the first token to main

strcpy(token, "main");

// Call main() to start interpreting.

call();

籍
}
catch(InterpExc exc) {
sntx_err(exc.get_err());

书
return 1;
}

典 书
catch(bad_alloc exc) {
cout << "Out Of Memory\n";
return 1;

经 读
}

+
return ret_value;

不
}

+
The main( ) function begins by allocating memory to hold the program being interpreted. Notice that the largest

C
program that can be interpreted is specified by the constant PROG_ SIZE. This value is arbitrarily set at 10,000, but you

典
can increase it if you want. Next, the program is loaded by calling load_program( ). After the program has been loaded,
main( ) performs three primary actions:

经
1. It calls prescan( ), the interpreter prescanner.

2. It readies the interpreter for the call to main( ) by finding its location in the program.

无 3. It executes call( ), which begins execution of the program at the start of main( ).

The main( ) function also handles all InterpExc exceptions generated by Mini C++. This includes those thrown by the
parser, too.

The following sections describe the key components of the interpreter in detail.

The Interpreter Prescan

Before the interpreter can start executing a program, two important clerical tasks must be performed:
1. All global variables must be found and initialized.

2. The location of each function defined in the program must be found.

These tasks are carried out by the interpreter prescan.

In Mini C++, all executable code exists inside functions, so there is no reason for the Mini C++ interpreter to go outside
a function once execution has begun. However, global variable declaration statements lie outside all functions. Thus, it
is necessary for the prescanner to handle these declarations. There is no other (efficient) way for the interpreter to
know about them.
.

In the interest of execution speed, it is important (although not technically necessary) that the location of each function
defined in the program be known so that a call to a function can be reasonably fast. If this step is not performed, a
lengthy sequential search of the source code would be needed to find the entry point to a function each time it is
called.

Finding the entry point of all functions also serves a second purpose. As you know, a C++ program does not begin
execution at the physical top of the program. Instead, it begins at the main( ) function. Furthermore, there is no
requirement that main( ) be the first function in the program. It is therefore necessary to find the location of the main( )
function within the program’s source code so that execution can begin at that point. (Remember also that global
variables may precede main( ), so even if it is the first function, it is not necessarily the first line of code.) Because the
prescanner finds the entry point of all functions, it also finds the entry point of main( ).

The function that performs the prescan is called prescan( ). It is shown here:
// Find the location of all functions in the program
// and store global variables.
void prescan()
{
char *p, *tp;
char temp[MAX_ID_LEN+1];

籍
token_ireps datatype;
func_type ft;

书
// When brace is 0, the current source position
// is outside of any function.
int brace = 0;

p = prog;

do {

经 典 读 书
+
// Bypass code inside functions

不
while(brace) {

+
get_token();
if(tok == END) throw InterpExc(UNBAL_BRACES);

C
if(*token == '{') brace++;

典
if(*token == '}') brace--;
}

经
tp = prog; // save current position
get_token();

无
// See if global var type or function return type.
if(tok==CHAR || tok==INT) {
datatype = tok; // save data type
get_token();

if(token_type == IDENTIFIER) {
strcpy(temp, token);
get_token();

if(*token != '(') { // must be global var

prog = tp; // return to start of declaration
decl_global();
}
else if(*token == '(') { // must be a function

// See if function already defined.

for(unsigned i=0; i < func_table.size(); i++)
if(!strcmp(func_table[i].func_name, temp))
throw InterpExc(DUP_FUNC);

ft.loc = prog;
.

ft.ret_type = datatype;
strcpy(ft.func_name, temp);
func_table.push_back(ft);

do {
get_token();
} while(*token != ')');
// Next token will now be opening curly
// brace of function.
}
else putback();
}
}
else {
if(*token == '{') brace++;
if(*token == '}') brace--;
}
} while(tok != END);
if(brace) throw InterpExc(UNBAL_BRACES);
prog = p;

籍
}

The prescan( ) function works like this. Each time an opening curly brace is encountered, brace is incremented.

书
Whenever a closing curly brace is read, brace is decremented. Therefore, whenever brace is greater than zero, the
current token is being read from within a function. However, if brace equals zero when a variable is found, then the

典 书
prescanner knows it must be a global variable. By the same method, if a function name is encountered when brace
equals zero, then it must be that function’s definition. (Remember, Mini C++ does not support function prototypes.)

经 读
Global variables are stored in a vector called global_vars, which holds structures of type var_type, shown here:
// This structure encapsulates the info

+
// associated with variables.

不
struct var_type {

+
char var_name[MAX_ID_LEN+1]; // name

C
token_ireps v_type; // data type

典
int value; // value
};

经
This structure stores the name, value, and type of a variable.

Global variables are entered into the global_vars vector by decl_global( ), shown next:

无
// Declare a global variable.
void decl_global()
{
token_ireps vartype;
var_type vt;

get_token(); // get type

vartype = tok; // save var type

// Process comma-separated list.

do {
vt.v_type = vartype;
vt.value = 0; // init to 0
get_token(); // get name

// See if variable is a duplicate.

for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, token))
throw InterpExc(DUP_VAR);
.

strcpy(vt.var_name, token);
global_vars.push_back(vt);

get_token();
} while(*token == ',');

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

}

In essence, decl_global( ) obtains the type and name of a variable, initializes it to zero and then puts it on the end of
global_vars. However, first a check is performed to ensure that a variable with that name has not already been
declared.

The location of each user-defined function is put into a vector called func_table, which stores structures of type
func_type, shown here:

// This structure encapsulates function info.

struct func_type {
char func_name[MAX_ID_LEN+1]; // name
token_ireps ret_type; // return type
char *loc; // location of entry point in program

籍
};

Each function entry contains the return type, name, and location in the source code of the function’s entry point.

书
Before a function is entered into the table, prescan( ) checks that one by the same name does not already exist. Notice
that no parameter information is stored. This information is obtained at runtime, when a function is actually called.

经 典 读 书
The interp( ) Function

The interp( ) function is the heart of the interpreter. It is the function that decides what action to take based on the next
token in the input stream. The function is designed to interpret one unit of code and then return. If the unit consists of a

+ 不
single statement, then that statement is interpreted and interp( ) returns. However, if an opening brace is read, then all

+
statements within that block are interpreted. This is managed by the block flag, which is set to 1 if an opening brace is
read. This causes the function to continue to interpret statements until a closing brace is read. The interp( ) function is

C
shown here:

典
// Interpret a single statement or block of code. When
// interp() returns from its initial call, the final

经
// brace (or a return) in main() has been encountered.
void interp()
{

无
int value;
int block = 0;

do {
// Don't interpret until break is handled.
if(breakfound) return;

token_type = get_token();

// See what kind of token is up.

if(token_type == IDENTIFIER ||
*token == INC || *token == DEC)
{
// Not a keyword, so process expression.
putback(); // restore token to input stream for
// further processing by eval_exp()
eval_exp(value); // process the expression
if(*token != ';') throw InterpExc(SEMI_EXPECTED);
}
else if(token_type==BLOCK) { // block delimiter?
if(*token == '{') { // is a block
.

block = 1; // interpreting block, not statement

// Record nested scope.
nest_scope_stack.push(local_var_stack.size());
}
else { // is a }, so reset scope and return
// Reset nested scope.
local_var_stack.resize(nest_scope_stack.top());
nest_scope_stack.pop();
return;
}
}
else // is keyword
switch(tok) {
case CHAR:
case INT: // declare local variables
putback();
decl_local();
break;
case RETURN: // return from function call
func_ret();

籍
return;
case IF: // process an if statement
exec_if();

书
break;
case ELSE: // process an else statement

典 书
find_eob(); // find end of else block
// and continue execution
break;

经 读
case WHILE: // process a while loop
exec_while();

+
break;

不
case DO: // process a do-while loop

+
exec_do();
break;

C 典
case FOR: // process a for loop
exec_for();
break;

经
case BREAK: // handle break
breakfound = true;

无
// Reset nested scope.
local_var_stack.resize(nest_scope_stack.top());
nest_scope_stack.pop();
return;
case SWITCH: // handle a switch statement
exec_switch();
break;
case COUT: // handle console output
exec_cout();
break;
case CIN: // handle console input
exec_cin();
break;
case END:
exit(0);
}
} while (tok != END && block);
return;
}

Calls to functions like exit( ) excepted, a C++ program ends when the last brace (or a return) in main( ) is
encountered—not necessarily at the last line of source code. This is one reason that interp( ) executes only a
.

statement or a block of code and not the entire program. Also, conceptually, C++ consists of blocks of code.
Therefore, interp( ) is called each time a new block of code is encountered. This includes function blocks and blocks
begun by various C++ statements, such as if. In the process of executing a program, Mini C++ may call interp( )
recursively.

The interp( ) function works like this. First, it checks if a break statement has been encountered in the program. If one
has, the global variable breakfound will be true. This variable remains set until it is cleared by another part of the
interpreter, as you will see when the interpretation of the control statements are described.

Assuming that breakfound is false, interp( ) reads the next token from the program. If the token is an identifier, the
statement must be an expression, so the expression parser is called. Because the expression parser expects to read
the first token in the expression itself, the token is returned to the input stream via a call to putback( ). When eval_exp(
) returns, token will hold the last token read by the expression parser, which must be a semicolon if the statement is
syntactically correct. If token does not contain a semicolon, an error is reported.

If the next token from the program is an opening brace, block is set to 1 and the current size of local_var_stack is
pushed onto nest_scope_stack. local_var_stack holds all local variables, including those declared in nested block
scopes. Conversely, if the next token is a closing brace, local_var_stack is truncated to the size specified by the top of
nest_scope_ stack. This effectively removes any local variables that may have been declared inside that scope. Thus,

籍
a { causes the size of the local variable stack to be saved, and a } causes the local variable stack to be reset to its
previous size. This mechanism supports nested, local scopes.

书
Finally, if the token is a keyword, the switch statement is executed, calling the appropriate routine to handle the
statement. The reason that keywords are given integer equivalents by get_token( ) is to allow the use of the switch
statement instead of requiring a sequence of if statements involving string comparisons (which are quite slow).

经 典 读 书
Handling Local Variables

When the interpreter encounters an int or char keyword, it calls decl_local( ) to create storage for a local variable. As

+
stated earlier, no global variable declaration statement will be encountered by the interpreter once the program is

不
executing because only code within a function is executed. Therefore, if a variable declaration statement is found, it

+
must be for a local variable (or a parameter, which will be discussed in the next section). As a general rule, local

C
variables are stored on a stack. If the language is compiled, the system stack is generally used; however, in an

典
interpreted mode, the stack for local variables must be maintained by the interpreter.

In Mini C++, local_var_stack provides the stack for local variables. Somewhat unexpectedly, local_var_stack is a

经
vector rather than a stack container. The reason for this is that although local variables are managed in a stack-like
fashion, the container must be able to be searched sequentially when access to a variable’s value is needed. The
stack container does not conveniently allow such a sequential search, but vector does.

无
The decl_local( ) function is shown here:
// Declare a local variable.
void decl_local()
{
var_type vt;

get_token(); // get var type

vt.v_type = tok; // store type

vt.value = 0; // init var to 0

// Process comma-separated list.

do {
get_token(); // get var name

// See if variable is already the name

// of a local variable in this scope.
if(!local_var_stack.empty())
for(int i=local_var_stack.size()-1;
i >= nest_scope_stack.top(); i--)
.

{
if(!strcmp(local_var_stack[i].var_name, token))
throw InterpExc(DUP_VAR);
}

strcpy(vt.var_name, token);
local_var_stack.push_back(vt);
get_token();
} while(*token == ',');

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

}

Each time a local variable is encountered, its name, type, and value (initially zero) are pushed onto the stack. The
process works like this. The decl_local( ) function first reads the type of the variable or variables being declared and
sets the initial value to zero. Next, it enters a loop, which reads the names of a comma-separated list of identifiers.
Each time through the loop, the information about each variable is pushed onto the local variable stack. In the process,
a check is made to ensure that a variable by the same name is not already existent in the current scope. (Variables
declared within the current scope are found between the current top of local_var_stack and the index saved on the top
of nest_scope_stack.) At the end, the final token is checked to make sure that it contains a semicolon.

书 籍
Calling User-Defined Functions

Probably the most difficult part of implementing an interpreter for C++ is managing the execution of user-defined
functions. Not only does the interpreter need to begin reading the source code at a new location and then return to the

典 书
calling routine after the function terminates, but it must also deal with these three tasks: the passing of arguments, the
allocation of parameters, and the return value of the function.

经 读
All function calls (except the initial call to main( )) take place through the expression parser from the atom( ) function by

+
a call to call( ). It is the call( ) function that actually handles the details of calling a function. The call( ) function is shown

不
here. Let’s examine it closely:

+
// Call a function.
void call()

C 典
{
char *loc, *temp;
int lvartemp;

无 经
// First, find entry point of function.
loc = find_func(token);

if(loc == NULL)
throw InterpExc(FUNC_UNDEF); // function not defined
else {
// Save local var stack index.
lvartemp = local_var_stack.size();

get_args(); // get function arguments

temp = prog; // save return location

func_call_stack.push(lvartemp); // push local var index

prog = loc; // reset prog to start of function

get_params(); // load the function's parameters with
// the values of the arguments

interp(); // interpret the function

prog = temp; // reset the program pointer

if(func_call_stack.empty()) throw InterpExc(RET_NOCALL);

.

// Reset local_var_stack to its previous state.

local_var_stack.resize(func_call_stack.top());
func_call_stack.pop();
}
}

The first thing that call( ) does is find the location of the entry point in the source code to the specified function by
calling find_func( ). Next, it saves the current size of the local variable stack into the lvartemp variable. Then it calls
get_args( ) to process any function arguments. The get_args( ) function reads a comma-separated list of expressions
and pushes them onto the local variable stack in reverse order. (The expressions are pushed in reverse order so that
they can be more easily matched with their corresponding parameters when the function is interpreted.) When the
values are pushed, they are not given names. The names of the parameters are given to them by the get_params( )
function, which will be discussed in a moment.

Once the function arguments have been processed, the current value of prog is saved in temp. This location is the
return point of the function. Next, the value of lvartemp is pushed onto the function call stack, func_call_stack. Its
purpose is to store the index of the top of the local variable stack each time a function is called. This value represents
the starting point on the local variable stack for variables (and parameters) relative to the function being called. The

籍
value on the top of the function call stack is used to prevent a function from accessing any local variables other than
those it declares.

书
The next two lines of code set the program pointer to the start of the function and link the name of its formal
parameters with the values of the arguments already on the local variable stack with a call to get_params( ). To do this,
get_params( ) reads each parameter and copies its name to the corresponding argument already stored in

典 书
local_var_stack.

The actual execution of the function is performed through a call to interp( ). When interp( ) returns, the program pointer

经 读
(prog) is reset to its return point, and the local variable stack index is reset to its value before the function call. This
final step effectively removes all of the function’s local variables and parameters from the stack.

+ 不
If the function being called contains a return statement, then interp( ) calls func_ret( ) prior to returning to call( ). This

+
function processes any return value. It is shown here:

C
// Return from a function.

典
void func_ret()
{

经
int value;

value = 0;

}
无
// Get return value, if any.
eval_exp(value);

ret_value = value;

The global variable ret_value is an integer that holds the return value of a function. At first glance you might wonder
why the local variable value receives the return value from eval_exp( ), rather than ret_value. The reason is that
functions can be recursive and eval_exp( ) may need to call the same function in order to obtain its value. In this
situation, a global variable cannot be used to receive the value because it will be overwritten.

Assigning Values to Variables

Let’s return briefly to the expression parser. When an assignment statement is encountered, the value of the right side
of the expression is computed, and this value is assigned to the variable on the left using a call to assign_var( ).
However, C++ supports various scopes, including the global scope (formally called namespace scope) and local
scope. Furthermore, local scopes can be nested. This is important because it affects the way that Mini C++ finds the
value of variables. To understand why, consider the following Mini C++ program:
int count;
.

int main()
{
int count, i;

count = 100;

i = f();

return 0;
}

int f()
{
int count;

count = 99;

return count;
}

籍
When count is assigned a value, how does the assign_var( ) function know which variable is being referred to? Is it the
global count or one of the local counts? The answer is simple. In C++, local variables take priority over global
variables of the same name. Furthermore, local variables are not known outside their own block. To see how these

书
rules can be used to resolve the preceding assignments, examine the assign_var( ) function, shown here:
// Assign a value to a variable.

典 书
void assign_var(char *vname, int value)
{
// First, see if it's a local variable.

经 读
if(!local_var_stack.empty())
for(int i=local_var_stack.size()-1;

+
i >= func_call_stack.top(); i--)

不
{

+
if(!strcmp(local_var_stack[i].var_name,

C
vname))

典
{
if(local_var_stack[i].v_type == CHAR)
local_var_stack[i].value = (char) value;

经
else if(local_var_stack[i].v_type == INT)
local_var_stack[i].value = value;
return;

无
}
}

// Otherwise, try global vars.

for(unsigned i=0; i < global_vars.size(); i++)
if(!strcmp(global_vars[i].var_name, vname)) {
if(global_vars[i].v_type == CHAR)
global_vars[i].value = (char) value;
else if(global_vars[i].v_type == INT)
global_vars[i].value = value;
return;
}

throw InterpExc(NOT_VAR); // variable not found

}

As explained in the previous section, each time a function is called, the current index of the top of the local variable
stack (local_var_stack) is pushed onto the function call stack (func_call_stack). This means that any local variables (or
parameters) defined by the function will be pushed onto the stack above that point. Therefore, the assign_var( )
function first searches local_var_stack, beginning with the current top-of-stack and stopping when the index reaches
that saved by the latest function call. This mechanism ensures that only those variables local to the function are
.

examined. (It also helps support recursive functions because the current top-of-stack value is saved each time a
function is invoked.) Therefore, in the example, the line
count = 100;

in main( ) causes assign_var( ) to find the local variable count inside main( ). In f( ), the statement
count = 99;

causes assign_var( ) to find the count declared within f( ).

If no local variable matches the name of a variable, then the global variable list is searched. If neither the local or
global variable list contains the variable, a syntax error results.

Now, consider the case in which a variable is declared within a nested scope, such as shown here:
int f(int n)
{
int count;

count = 99

籍
if(n > 0) {
int count; // this count is local to the if

书
count = 100; // refers to count in if block
// ...
}

经 典 读 书
return count; // refers to outer count.

In this situation, the outer count (which is declared at the start of the function code) is pushed onto local_var_stack

+ 不
first. Then the count that is local to the if block is pushed onto local_var_stack. Thus, when the following line executes:

+
count = 100; // refers to count in if block

C
assign_var( ) looks up count and finds the copy of count that is local to the if block first, just as it should.

经 典
One last point, within the interp( ) function, each time a block is left, local_var_stack is truncated to the size is was
before entering the block. This effectively removes all local variables declared within that block from the stack.
local_var_stack is also truncated each time a function returns, which removes all local variables and parameters
associated with the function.

无
Executing an if Statement

Now that the basic structure of the Mini C++ interpreter has been described, it is time to see how the control
statements are implemented. In general, each time a keyword statement is encountered, interp( ) calls a function that
processes that statement. The functions that interpret the various control statements all begin with the prefix exec_.
For example, a for is interpreted by exec_for( ), a switch is interpreted by exec_switch( ) , and so on.

One of the easiest control statements to interpret is the if. The if statement is processed by exec_if( ) , shown here:
// Execute an if statement.
void exec_if()
{
int cond;

eval_exp(cond); // get if expression.

if(cond) { // if true, process target of IF

// Confirm start of block.
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);
.

interp();
}
else {
// Otherwise skip around IF block and
// process the ELSE, if present.

find_eob(); // find start of next line

get_token();

if(tok != ELSE) {
// Restore token if no ELSE is present.
putback();
return;
}

// Confirm start of block.

get_token();
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);
putback();

}
}
interp();

书 籍
Let’s look closely at the operation of this function.

经 典 读 书
First, exec_if( ) calls eval_exp( ) to compute the value of the conditional expression. If the condition is true (nonzero),
the function calls interp( ), which executes the code within the if block. If the condition is false, find_eob( ) is called,
which advances the program pointer to the location immediately after the end of the if block. If an else is present, the

+
block of code associated with the else is executed. Otherwise, execution simply begins with the next line of code.

+ 不
If the if block executes and there is an else block present, there must be some way for the else block to be bypassed.
This is accomplished in interp( ) when an else statement is encountered. In this case, interp( ) simply calls find_eob( )

C
to bypass the else block. After that, execution resumes with the first statement after the else block. Remember, the only

典
time an else will be processed by interp( ) (in a syntactically correct program) is after an if block is executed.

经
One other point: notice that exec_if( ) confirms that the code that is the target of the if and else is contained within a
block. As explained, to simplify the interpreter, the targets of all control statements must be blocks of code. This
restriction streamlines the interpreter code.

无
The switch and break Statements

Interpreting the switch statement requires more work than does the if statement. One reason is that its syntax is more
complicated. Another reason is that it relies on the break statement. Thus, to handle a switch statement implies that
break must also be handled. Both are examined here.

Handling break is actually fairly easy because it performs the same action whether it is used in a switch statement or in
a loop: it causes the block associated with that statement to be exited. Mini C++ handles break by use of a global flag
variable called breakfound, which is initially set to false. When a break is found, breakfound is set to true. This variable
is then checked by the switch statement (and the loop statements described later) to determine if a break has been
executed. If breakfound is true, the current block is terminated and breakfound is reset to false.

The break statement is handled inside interp( ) by the code shown here:
case BREAK: // handle break
breakfound = true;

// Reset nested scope.

local_var_stack.resize(nest_scope_stack.top());
nest_scope_stack.pop();
return;
.

As you can see, in addition to setting breakfound to true, the local variable stack is reset to its state prior to the start of
the block being terminated. Recall that variables can be declared within any block and are local to that block. Thus,
when break is found, any local variables associated with the enclosing block must be removed.

The switch statement is handled by the exec_switch( ) function shown here:

// Execute a switch statement.
void exec_switch()
{
int sval, cval;
int brace;

eval_exp(sval); // Get switch expression.

// Check for start of block.

if(*token != '{')
throw InterpExc(BRACE_EXPECTED);

// Record new scope.

nest_scope_stack.push(local_var_stack.size());

籍
// Now, check case statements.
for(;;) {

书
brace = 1;
// Find a case statement.
do {

典 书
get_token();
if(*token == '{') brace++;
else if(*token == '}') brace--;

经 读
} while(tok != CASE && tok != END && brace);

+
// If no matching case found, then skip.

不
if(!brace) break;

C+
if(tok == END) throw InterpExc(SYNTAX);

典
// Get value of the case statement.
eval_exp(cval);

无 经
// Read and discard the :
get_token();

if(*token != ':')
throw InterpExc(COLON_EXPECTED);

// If values match, then interpret.

if(cval == sval) {
brace = 1;
do {
interp();

if(*token == '{') brace++;

else if(*token == '}') brace--;
} while(!breakfound && tok != END && brace);

// Find end of switch statement.

while(brace) {
get_token();
if(*token == '{') brace++;
else if(*token == '}') brace--;
}
breakfound = false;
.

break;
}
}
}

First, exec_switch( ) obtains the value of the switch expression and stores this value in sval. Next, it checks for the
start of the switch block and stores the top of the local variable stack on nest_scope_stack. This step is necessary
because the switch statement creates a nested scope. Next, the value of each case statement is examined until either
a match with the value in sval is found, or the end of the switch is reached. (Recall that, for the sake of simplicity, Mini
C++ does not support the default statement.) If a match is found, the statements associated with that case are
executed until a break statement is encountered (that is, until breakfound is true), or until the end of the switch block is
found. After a break is encountered, exec_switch( ) ends by finding the end of the switch block and then setting
breakfound to false.

Processing a while Loop

A while loop is quite easy to interpret. The function that performs this task is exec_while( ), shown here:
// Execute a while loop.

籍
void exec_while()
{
int cond;

书
char *temp;

典 书
putback(); // put back the while
temp = prog; // save location of top of while loop

经 读
get_token();
eval_exp(cond); // check the conditional expression

+ +
if(*token != '{')

不
// Confirm start of block.

throw InterpExc(BRACE_EXPECTED);

if(cond)

C 典
interp(); // if true, interpret

经
else { // otherwise, skip to end of loop
find_eob();
return;

无
}

prog = temp; // loop back to top

// Check for break in loop.

if(breakfound) {
// Find start of loop block.
do {
get_token();
} while(*token != '{' && tok != END);

putback();
breakfound = false;
find_eob();
return;
}
}

The exec_while( ) function works like this. First, the while token is put back into the input stream, and the location of
the while in the program is saved into the temp pointer. This address is used to allow the interpreter to loop back to the
top of the while. Next, the while is reread to remove it from the input stream, and eval_exp( ) is called to compute the
.

value of the while’s conditional expression. If the conditional expression is true, interp( ) is called to interpret the while
code block. When interp( ) returns, prog (the program pointer) is loaded with the location of the start of the while loop,
which causes program execution to resume at the top of the loop when control returns to interp( ). This results in the
next iteration of the loop. However, if interp( ) returns because of a break statement being found within the loop,
iteration stops, the end of the while block is found, and exec_while( ) returns. When the conditional expression is false,
the end of the while block is found and the function returns.

Processing a do-while Loop

A do-while loop is processed much like the while. When interp( ) encounters a do statement, it calls exec_do( ), shown
here:
// Execute a do loop.
void exec_do()
{
int cond;
char *temp;

// Save location of top of do loop.

籍
putback(); // put back do
temp = prog;

书
get_token(); // get start of loop block

典 书
// Confirm start of block.
get_token();
if(*token != '{')

经 读
throw InterpExc(BRACE_EXPECTED);
putback();

+ 不
interp(); // interpret loop

+
// Check for break in loop.

C 典
if(breakfound) {
prog = temp;
// Find start of loop block.

经
do {
get_token();
} while(*token != '{' && tok != END);

无
// Find end of while block.
putback();
find_eob();

// Now, find end of while expression.

do {
get_token();
} while(*token != ';' && tok != END);
if(tok == END) throw InterpExc(SYNTAX);

breakfound = false;
return;
}

get_token();
if(tok != WHILE) throw InterpExc(WHILE_EXPECTED);

eval_exp(cond); // check the loop condition

// If true loop; otherwise, continue on.

.

if(cond) prog = temp;

}

The main difference between the do-while and the while loops is that the do-while always executes its block of code at
least once because the conditional expression is at the bottom of the loop. Therefore, exec_do( ) first saves the
location of the top of the loop into temp and then calls interp( ) to interpret the block of code associated with the loop.
When interp( ) returns, the corresponding while is retrieved and the conditional expression is evaluated. If the condition
is true, prog is reset to the top of the loop; otherwise, execution will continue on. If a break is encountered, iteration
stops and the end of the loop block is found.

The for Loop

The interpretation of the for loop poses a more difficult challenge than the other loops. Part of the reason for this is that
the structure of the C++ for is definitely designed with compilation in mind. The main trouble is that the conditional
expression of the for must be checked at the top of the loop, but the increment portion occurs at the bottom of the
loop. Therefore, even though these two pieces of the for loop occur next to each other in the source code, their
interpretation is separated by the block of code being iterated. With a little effort, though, the for can be correctly
interpreted.

籍
When interp( ) encounters a for statement, exec_for( ) is called. This function is shown here:
// Execute a for loop.
void exec_for()

书
{
int cond;

典 书
char *temp, *temp2;
int paren ;

经 读
get_token(); // skip opening (
eval_exp(cond); // initialization expression

+ + 不
if(*token != ';') throw InterpExc(SEMI_EXPECTED);
prog++; // get past the ;
temp = prog;

C
for(;;) {

典
// Get the value of the conditional expression.

经
eval_exp(cond);

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

无
prog++; // get past the ;
temp2 = prog;

// Find start of for block.

paren = 1;
while(paren) {
get_token();
if(*token == '(') paren++;
if(*token == ')') paren--;
}

// Confirm start of block.

get_token();
if(*token != '{')
throw InterpExc(BRACE_EXPECTED);
putback();

// If condition is true, interpret

if(cond)
interp();
else { // otherwise, skip to end of loop
.

find_eob();
return;
}

prog = temp2; // go to increment expression

// Check for break in loop.

if(breakfound) {
// Find start of loop block.
do {
get_token();
} while(*token != '{' && tok != END);

putback();
breakfound = false;
find_eob();
return;
}

// Evaluate the increment expression.

籍
eval_exp(cond);

prog = temp; // loop back to top

书
}
}

典 书
This function begins by processing the initialization expression in the for . The initialization portion of the for is
executed only once and does not form part of the loop. Next, the program pointer is advanced to a point immediately

经 读
after the semicolon that ends the initialization statement, and this address is assigned to temp. This location is the
start of the conditional expression. An infinite loop is then entered that checks the conditional portion of the loop and

+
assigns temp2 the address of the start of the increment expression. Then, the beginning of the loop code is found.

不
Finally, if the conditional expression is true, the loop block is interpreted. Otherwise, the end of the block is found and

+
execution continues on after the for loop. Assuming that the loop executes, when the call to interp( ) returns, the

C
increment expression is evaluated, and the process repeats. Of course, this process stops if a break is encountered

典
within the loop block.

Handling cin and cout Statements

经
Because I/O via cin and cout is a fundamental part of C++, it seems only proper for Mini C++ to support them.

无
However, Mini C++ does not handle I/O through cin and cout the way that it is handled by a commercial compiler. As
you know, cin and cout are predefined identifiers that correspond to streams that are linked to standard input and
standard output. They are used to input and output information at the console through the use of the I/O operators <<
and >>. Thus, << and >> are overloaded for I/O. However, Mini C++ does not support operator overloading. In fact, to
keep Mini C++ as simple as possible, it does not even support the << or >> shift operators! (However, get_token( ) will
recognize these operators.) Despite these restrictions, it is still quite easy to interpret cin and cout statements.

Console output through cout is handled by the exec_cout( ) function, shown here:
// Execute a cout statement.
void exec_cout()
{
int val;

get_token();
if(*token != LS) throw InterpExc(SYNTAX);

do {
get_token();

if(token_type==STRING) {
// Output a string.
.

cout << token;

}
else if(token_type == NUMBER ||
token_type == IDENTIFIER) {
// Output a number.
putback();
eval_exp(val);
cout << val;
}
else if(*token == '\'') {
// Output a character constant.
putback();
eval_exp(val);
cout << (char) val;
}

get_token();
} while(*token == LS);

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

籍
}

When a cout identifier is encountered, the next token is read. If it is not a <<, then a syntax error is reported.

书
Otherwise, a loop is entered that obtains and then outputs the value of the string or expression on the right side of the
<<. This process continues until the end of the cout statement is reached.

典 书
The cin statement is handled by exec_cin( ), shown next:
// Execute a cin statement.

经 读
void exec_cin()
{

+
int val;

不
char chval;

+
token_ireps vtype;

C 典
get_token();
if(*token != RS) throw InterpExc(SYNTAX);

经
do {
get_token();
if(token_type != IDENTIFIER)

无
throw InterpExc(NOT_VAR);

vtype = find_var_type(token);

if(vtype == CHAR) {
cin >> chval;
assign_var(token, chval);
}
else if(vtype == INT) {
cin >> val;
assign_var(token, val);
}

get_token();
} while(*token == RS);

if(*token != ';') throw InterpExc(SEMI_EXPECTED);

}

When a cin identifier is encountered, the next token is read. If it is not a >>, then a syntax error is reported. Otherwise,
a loop is entered that obtains the variable that will receive input, reads input from the console, and stores that input
into the variable. Notice that the type of the variable determines whether integer or character data is read. This
process continues until the end of the cout statement is reached.

书 籍
经 典 读 书
+ + 不
C 典
无 经
.

The Mini C++ Library Functions

Because the programs executed by Mini C++ are never compiled and linked, any library routines they use must be
handled directly by Mini C++. The best way to do this is to create an interface function that Mini C++ calls when a
library function is encountered. This interface function sets up the call to the actual library function and handles any
return values.

Because of space limitations, Mini C++ supports only four “library” functions: getchar( ), putchar( ), abs( ) , and rand( ).
These functions are translated into calls to the actual library functions of the same name. The Mini C++ library routines
are in the file libcpp.cpp. It is shown here:
// ***** Internal Library Functions *****

// Add more of your own, here.

#include <iostream>
#include <cstdlib>

籍
#include <cstdio>
#include "mccommon.h"

书
using namespace std;

典 书
// Read a character from the console.
// If your compiler supplies an unbuffered
// character intput function, feel free to

经 读
// substitute it for the call to cin.get().
int call_getchar()

+
{

不
char ch;

+
ch = getchar();

C 典
// Advance past ()
get_token();

经
if(*token != '(')
throw InterpExc(PAREN_EXPECTED);

无
get_token();
if(*token != ')')
throw InterpExc(PAREN_EXPECTED);

return ch;
}

// Write a character to the display.

int call_putchar()
{
int value;

eval_exp(value);

putchar(value);

return value;
}

// Return absolute value.

int call_abs()
{
int val;

eval_exp(val);

val = abs(val);

return val;
}

// Return a random integer.

int call_rand()
{

// Advance past ()
get_token();
if(*token != '(')
throw InterpExc(PAREN_EXPECTED);

get_token();

籍
if(*token != ')')
throw InterpExc(PAREN_EXPECTED);

书
return rand();
}

典 书
To add more library functions of your own choosing, first enter their names and the addresses of their interface
functions into the intern_func array (which is declared in parser.cpp). Next, following the lead of the functions shown

经 读
previously, create appropriate interface functions. Finally, add their prototypes to mccommon.h.

+ + 不
C 典
无 经
.

The mccommon.h Header File

All three source files for Mini C++, minicpp.cpp, parser.cpp, and libcpp.cpp, include the header file mccommon.h,
shown here:
// Common declarations used by parser.cpp, minicpp.cpp,
// or libcpp.cpp, or by other files that you might add.
//
const int MAX_T_LEN = 128; // max token length
const int MAX_ID_LEN = 31; // max identifier length
const int PROG_SIZE = 10000; // max program size
const int NUM_PARAMS = 31; // max number of parameters

// Enumeration of token types.

enum tok_types { UNDEFTT, DELIMITER, IDENTIFIER,
NUMBER, KEYWORD, TEMP, STRING, BLOCK };

籍
// Enumeration of internal representation of tokens.
enum token_ireps { UNDEFTOK, ARG, CHAR, INT, SWITCH,
CASE, IF, ELSE, FOR, DO, WHILE, BREAK,

书
RETURN, COUT, CIN, END };

典 书
// Enumeration of two-character operators, such as <=.
enum double_ops { LT=1, LE, GT, GE, EQ, NE, LS, RS, INC, DEC };

经 读
// These are the constants used when throwing a
// syntax error exception.

+
//

不
// NOTE: SYNTAX is a generic error message used when

+
// nothing else seems appropriate.
enum error_msg

C 典
{ SYNTAX, NO_EXP, NOT_VAR, DUP_VAR, DUP_FUNC,
SEMI_EXPECTED, UNBAL_BRACES, FUNC_UNDEF,
TYPE_EXPECTED, RET_NOCALL, PAREN_EXPECTED,

经
WHILE_EXPECTED, QUOTE_EXPECTED, DIV_BY_ZERO,
BRACE_EXPECTED, COLON_EXPECTED };

无
extern char *prog; // current location in source code
extern char *p_buf; // points to start of program buffer

extern char token[MAX_T_LEN+1]; // string version of token

extern tok_types token_type; // contains type of token
extern token_ireps tok; // internal representation of token

extern int ret_value; // function return value

extern bool breakfound; // true if break encountered

// Exception class for Mini C++.

class InterpExc {
error_msg err;
public:
InterpExc(error_msg e) { err = e; }
error_msg get_err() { return err; }
};

// Interpreter prototypes.
void prescan();
void decl_global();
void call();
void putback();
void decl_local();
void exec_if();
void find_eob();
void exec_for();
void exec_switch();
void get_params();
void get_args();
void exec_while();
void exec_do();
void exec_cout();
void exec_cin();
void assign_var(char *var_name, int value);
bool load_program(char *p, char *fname);
int find_var(char *s);
void interp();
void func_ret();
char *find_func(char *name);
bool is_var(char *s);

籍
token_ireps find_var_type(char *s);

// Parser prototypes.

书
void eval_exp(int &value);
void eval_exp0(int &value);

典 书
void eval_exp1(int &value);
void eval_exp2(int &value);
void eval_exp3(int &value);

经 读
void eval_exp4(int &value);
void eval_exp5(int &value);

+
void atom(int &value);

不
void sntx_err(error_msg error);

+
void putback();
bool isdelim(char c);

C 典
token_ireps look_up(char *s);
int find_var(char *s);
tok_types get_token();

经
int internal_func(char *s);
bool is_var(char *s);

无
// "Standard library" prototypes.
int call_getchar();
int call_putchar();
int call_abs();
int call_rand();
 Compiling and Linking the Mini C++ Interpreter
To use Mini C++, you must compile and link minicpp.cpp, parser.cpp, and libcpp.cpp. You can use just about any
modern C++ compiler, including Borland C++ and Visual C++. For example, for Visual C++, you can use the following
command line:
cl -GX minicpp.cpp parser.cpp libcpp.cpp

For Borland C++, this command line can be used:

bcc32 minicpp.cpp parser.cpp libcpp.cpp

If you use a different compiler, simply follow the instructions that come with it.

Note For some older versions of Visual C++, Mini C++ may not be given sufficient stack space. You can use the /Fsize
option to increase the stack.

籍
To run a program, specify its name after minicpp on the command line. For example, this runs a program called
test.cpp :

书
minicpp test.cpp

经 典 读 书
+ + 不
C 典
无 经
.

Demonstrating Mini C++

In this section, several C++ programs are shown that demonstrate the features and capabilities of Mini C++. The first
demonstrates all features supported by Mini C++:
/* Mini C++ Demonstration Program #1.

This program demonstrates all features

of C++ that are recognized by Mini C++.
*/

int i, j; // global vars

char ch;

int main()
{
int i, j; // local vars

籍
// Call a "standard library' function.
cout << "Mini C++ Demo Program.\n\n";

书
// Call a programmer-defined function.

典 书
print_alpha();

cout << "\n";

do {

+ 经 读
// Demonstrate do and for loops.
cout << "Use loops.\n";

不
+
cout << "Enter a number (0 to quit): ";
cin >> i;

C 典
// Demonstrate the if
if(i < 0 ) {

经
cout << "Numbers must be positive, try again.\n";
}
else {

无
for(j = 0; j <= i; ++j) {
cout << j << " summed is ";
cout << sum(j) << "\n";
}
}
} while(i != 0);

cout << "\n";

// Demonstrate the break in a loop.

cout << "Break from a loop.\n";
for(i=0; i < 100; i++) {
cout << i << "\n";
if(i == 5) {
cout << "Breaking out of loop.\n";
break;
}
}

cout << "\n";

// Demonstrate the switch

.

cout << "Use a switch.\n";

for(i=0; i < 6; i++) {
switch(i) {
case 1: // can stack cases
case 0:
cout << "1 or 0\n";
break;
case 2:
cout << "two\n";
break;
case 3:
cout << "three\n";
break;
case 4:
cout << "four\n";
cout << "4 * 4 is "<< 4*4 << "\n";
// break; // this break is optional
// no case for 5
}
}

籍
cout << "\n";

cout << "Use a library function to generate "

书
<< "10 random integers.\n";
for(i=0; i < 10; i++) {

典 书
cout << rand() << " ";
}

经 读
cout << "\n";
cout << "Done!\n";

+
return 0;

+ 不
C 典
// Sum the values between 0 and num.
// This function takes a parameter.
int sum(int num)

经
{
int running_sum;

无
running_sum = 0;

while(num) {
running_sum = running_sum + num;
num--;
}
return running_sum;
}

// Print the alphabet.

int print_alpha()
{
cout << "This is the alphabet:\n";

for(ch = 'A'; ch<='Z'; ch++) {

putchar(ch);
}
cout << "\n";

return 0;
}
.

Here is a sample run:

Mini C++ Demo Program.

This is the alphabet:

ABCDEFGHIJKLMNOPQRSTUVWXYZ

Use loops.
Enter a number (0 to quit): 10
0 summed is 0
1 summed is 1
2 summed is 3
3 summed is 6
4 summed is 10
5 summed is 15
6 summed is 21
7 summed is 28
8 summed is 36
9 summed is 45
10 summed is 55

籍
Enter a number (0 to quit): 0
0 summed is 0

书
Break from a loop.
0
1

典 书
2
3
4

经 读
5
Breaking out of loop.

+
Use a switch.
1 or 0

+ 不
C
1 or 0

典
two
three
four

经
4 * 4 is 16

Use a library function to generate 10 random integers.

无
130 10982 1090 11656 7117 17595 6415 22948 31126 9004
Done!

The next example demonstrates nested loops:

// Nested loop example.
int main()
{
int i, j, k;

for(i = 0; i < 5; i = i + 1) {
for(j = 0; j < 3; j = j + 1) {
for(k = 3; k ; k = k - 1) {
cout << i <<", ";
cout << j << ", ";
cout << k << "\n";
}
}
}

cout << "done";

.

return 0;
}

A portion of its output is shown here:

0, 0, 3
0, 0, 2
0, 0, 1
0, 1, 3
0, 1, 2
0, 1, 1
0, 2, 3
0, 2, 2
0, 2, 1
.
.
.

The next program exercises the assignment operator:

// Assignments as operations.

籍
int main()
{
int a, b;

书
a = b = 5;

典 书
cout << a << " " << b << "\n";

while(a=a-1) {

经 读
cout << a << " ";
do {

+
cout << b << " ";

不
} while((b=b-1) > -5);

+
cout << "\n";

C
}

}
return 0;

经 典
The output from this program is shown next:
55

无
4 5 4 3 2 1 0 -1 -2 -3 -4
3 -5
2 -6
1 -7

Recursive functions are demonstrated by the next program. In it, the function factr( ) computes the factorial of a
number.
// This program demonstrates a recursive function.

// A recursive function that returns the

// factorial of i.
int factr(int i)
{
if(i<2) {
return 1;
}
else {
return i * factr(i-1);
}
}
.

int main()
{
cout << "Factorial of 4 is: ";
cout << factr(4) << "\n";

cout << "Factorial of 6 is: ";

cout << factr(6) << "\n";

return 0;
}

The output is shown here:

Factorial of 4 is: 24
Factorial of 6 is: 720

The next program fully demonstrates function arguments:

// A more rigorous example of function arguments.

int f1(int a, int b)

籍
{
int count;

书
cout << "Args for f1 are ";
cout << a << " " << b << "\n";

典 书
count = a;
do {
cout << count << " ";

经 读
} while(count=count-1);

+
cout << a << " " << b

不
<< " " << a*b << "\n";

C+
return a*b;

典
int f2(int a, int x, int y)

经
{
cout << "Args for f2 are ";
cout << a << " " << x << " "

无
<< y << "\n";
cout << x / a << " ";
cout << y*x << "\n";

return 0;
}

int main()
{
f2(10, f1(10, 20), 99);

return 0;
}

The output from this program is shown here:

Args for f1 are 10 20
10 9 8 7 6 5 4 3 2 1 10 20 200
Args for f2 are 10 200 99
20 19800

The following program exercises each loop statement:

.

// Exercise the loop statements.

int main()
{
int a;
char ch;

// The while.
cout << "Enter a number: ";
cin >> a;
while(a) {
cout << a*a << " ";
--a;
}
cout << "\n";

// The do-while.
cout << "\nEnter characters, 'q' to quit.\n";
do {
// Use two "standard library" functions.
ch = getchar();

籍
putchar(ch);
} while(ch != 'q');
cout << "\n\n";

// the for.

典 书 书
for(a=0; a<10; ++a) {
cout << a << " ";

+
return 0;
经 不 读
cout << "\n\nDone!\n";

+
}

C
Here is a sample run:

典
Enter a number: 10
100 81 64 49 36 25 16 9 4 1

无 经
Enter characters, 'q' to quit.
This is a test. q
This is a test. q

0123456789

Done!

Notice that the built-in library function getchar( ) is line-buffered in this run, which causes no characters to be displayed
by putchar( ) until ENTER has been pressed. This behavior is the result of the real getchar( ) function that is called by
Mini C++. As you know, most compilers implement getchar( ) as line-buffered. The point is that the built-in, Mini C++
function exhibits the same behavior as the library function that underlies it.

The final program demonstrates the use of nested scopes. In it, the variable x is declared three times: first as a global
variable, then as a variable local to the if block, and finally, again within the while block. All three are separate and
distinct from each other.
// Demonstrate nested scopes.

int x; // global x

int main()
{
int i;
 i = 4;

x = 99; // global x is 99

if(i == 4) {
int x; // local x
int num; // local to if statement

x = i * 2;
cout << "Outer local x before loop: "
<< x << "\n";

while(x--) {
int x; // another local x

x = 18;
cout << "Inner local x: " << x << "\n";
}

籍
cout << "Outer local x after loop: "
<< x << "\n";
}

典 书
// Can't refer to num here because it is local

书
// to the preceding if block.
// num = 10;

经 读
cout << "Global x: " << x << "\n";
}

+ 不
The output is shown here. Notice that all three x’s are separate.

+
Outer local x before loop: 8
Inner local x: 18

C 典
Inner local x: 18
Inner local x: 18
Inner local x: 18

经
Inner local x: 18
Inner local x: 18
Inner local x: 18

无
Inner local x: 18
Outer local x after loop: -1
Global x: 99
 Improving Mini C++
Mini C++ was designed with transparency of operation in mind. The goal was to develop an interpreter that could be
easily understood with the least amount of effort. It was also designed in such a way that it could be easily expanded.
As such, Mini C++ is not particularly fast or efficient; however, the basic structure of the interpreter is in place, and you
can increase its speed of execution by following these steps.

Virtually all commercial interpreters expand the role of the prescanner. The entire source program being interpreted is
converted from its human-readable form into an internal format. In this internal form, all but quoted strings and
constants are transformed into integer tokens, much the way that Mini C++ converts the C++ keywords into integer
tokens. It may have occurred to you that Mini C++ performs a number of string comparisons. For example, each time a
variable or function is searched for, several string comparisons take place. String comparisons are very costly in terms
of time; however, if each identifier in the source program is converted into an integer, then much faster integer
comparisons can be used. In general, the conversion of the source program into an internal form is the single most
important change you can make to Mini C++ to improve its efficiency. The increase in speed will be dramatic.

籍
Another area for improvement, meaningful mostly for large programs, is the lookup routines for variables and
functions. Even if you convert these items into integer tokens, the current approach to searching for them relies upon a

书
sequential search. You could, however, substitute some other faster method. For example, you might try using a map
container, or perhaps some sort of hashing method or tree structure.

典 书
As stated earlier, one restriction that Mini C++ has relative to the full C++ grammar is that the targets of the control
statements, such as if, must be blocks of code enclosed between curly braces. The reason for this is that it greatly
simplifies the find_eob( ) function, which is used to find the end of a block of code after one of the control statements

经 读
executes. The find_eob( ) function need look only for a closing curly brace to match the one that starts the block. You
might find it an interesting exercise to remove this restriction.

+ + 不
C 典
无 经
 Expanding Mini C++
There are two general areas in which you can expand and enhance the Mini C++ interpreter: C++ features and
ancillary features. Some of these are discussed briefly in the following sections.

Adding New C++ Features

There are two basic categories of statements you can add to Mini C++. The first is additional action statements, such
as the goto and continue. You might also want to add support for the default statement to the switch. You should have
little trouble adding any of these if you study the way that Mini C++ interprets the other statements. If something
doesn’t work the first time, try finding the problem by displaying the contents of each token as it is processed.

The second category of statement you can add is support for additional data types. Mini C++ already contains the
basic “hooks” for additional data types. For example, the var_type structure already contains a field for the type of
variable. To add other built-in types (for example, float, double, and long), simply increase the size of the value field to
the size of the largest element you wish to hold.

籍
The addition of classes poses a greater challenge. First, you will need to design ways to instantiate objects. To do this,

书
you will need to allocate memory large enough to hold the data members of the class and store a reference to this
memory in another new field that you will need to add to var_type. You will also need to handle the concepts of public
and private.

经 典 读 书
Supporting pointers is no more difficult than supporting any other data type. However, you will need to add support for
the pointer operators to the expression parser. Once you have implemented pointers, arrays will be easy. Space for an
array should be allocated dynamically using new, and a pointer to the array should be stored in a new field added to
var_type for this purpose.

+ 不
To handle different return types for functions, you will need to make use of the ret_type field in the func_type structure.

+
This field defines what type of data a function returns. It is currently set but otherwise unused.

C 典
One easy addition you might try adding is support for #include. This preprocessor directive can be easily handled
during the prescan process.

经
One final thought: if you like to experiment with language constructs, don’t be afraid to add a non-C++ extension. For
example, you can easily add a foreach loop like the one described in Chapter 4.

无
Adding Ancillary Features

Interpreters give you the opportunity to add several interesting and useful features. For example, you can add a trace
facility that displays each token as it is executed. You can also add the ability to display the contents of each variable
as the program executes. Another feature you might want to add is an integrated editor so that you can “edit and go”
instead of having to use a separate editor to create your C++ programs.
 Index

Adaptors, STL, 256, 258

addflight( ), 210

AI-based problem solving, 202-253

using breadth-first search, 220-224

using depth-first search, 211-220

书 籍
and finding multiple solutions, 233-242

and finding the “optimal” solution, 243-249

典 书
using hill-climbing search, 224-231

读
using least-cost search, 231-233

+ 经
See also Search(es), AI-based

不
+
Algorithms, STL, 256, 257

C 典
allocate( ), 259, 278

经
allocator, 257, 258

无
Allocator, STL, 256, 257

functions, 258-259

annuity( ), 194

Annuity for given investment, maximum

formula for computing, 196

program for finding, 196-197

Annuity, initial investment needed for desired

formula for computing, 194

program for finding, 194-195

Artificial intelligence and problem solving. See AI-based problem solving

assign_var( ), 366-368

at( ) STL sequence container function, 260, 286

atexit( ), 99

atoi( ), 174

atom( ), 331, 332, 364

auto_ptr class, 15

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

back( ) STL sequence container function, 260, 284

bad_alloc exception, 46, 62

balance( ), 198

begin( ) STL sequence container function, 260, 284

籍
_beginthreadex( ), 72-73, 97, 99

书
Binders, STL, 258

Bytecode, 302

经 典 读 书
+ + 不
C 典
无 经
Index

C++

characteristics, overview of, 2-6

and direct access to the operating system, 3, 68, 69-70, 160

elegant design of, 302

informal theory of, 306-309

籍
and the Internet, 160

书
C# and C++, 5, 6

典 书
call( ), 364-365

经 读
case statements, stacked, 123, 147, 148

+ 不
cases experimental statement, 123-124

+
translating a , 145-148

C 典
clear( ) STL sequence container function, 259, 285

经
CloseHandle( ), 71, 78, 101

无
collect( ), 33-35, 46, 48-49, 62, 96, 101, 118

Combinatoric explosions, 204-205

Combinatorics, 204

Component, software, 3-4

construct( ), 259, 278

Container, creating a custom STL, 256-299

requirements for, 258-261

Containers, STL, 3, 256, 257

memory management for, 258, 278, 279

Control statement execution in Mini C++, 368-376

do-while loop, 373-375

for loop, 375-376

if, 368-370

switch and break, 370-372

while loop, 372-373

Copying (garbage collection algorithm), 12, 13

CreateDialog( ), 87

CreateMutex( ), 78

书 籍
CreateThread( ), 71, 72, 73

Critical section object, 77

经 典 读 书
+ + 不
C 典
无 经
Index

Deadlock, 79

decl_global( ), 358-359

decl_local( ), 362-363

delete, 4, 9

and garbage collection, 10

书 籍
and GCPtr, 64

and STL containers, 258, 279

deque, 257

经 典 读 书
destroy( ), 259, 279

+ 不
DialogFunc( ), 185

+
C
DLExc Internet downloader exception class, 166

经 典
dlstart( ), 185-186

Download class, 166-176

无
download( ), 166, 167-172

and GUI-based downloader, 185-186

Dynamic allocation. See Memory management, dynamic

Dynamic memory, 8
Index

empty( ) STL sequence container function, 259, 285

end( ) STL sequence container function, 260, 284

_endthreadex( ), 72, 73

equal( ) STL algorithm, 288

籍
erase( ) STL sequence container function, 259, 282, 283

书
eval_exp... Mini C++ functions, 331-333

Event object, 77

经 典
exec_cin( ), 378

读 书
exec_cout( ), 377-378

+ + 不
exec_do( ), 373-374

C 典
exec_for( ), 375-376

经
exec_if( ), 368-370

无
exec_switch( ), 370-372

exec_while( ), 372-373

ExitThread( ), 71-72

Expressions, C++, 306, 307

parsing, 309

production rules of, 308-309

Index

Financial calculations, C++ programs for, 188-199

find( ), 210, 217-218, 219, 225, 231-232

find_eob( ), 369, 393

find_func( ), 365

籍
findroute( ), 210, 216, 218-219, 221-222, 223

书
FlightInfo structure, 209

for loop, 123, 144-145

经 典 读 书
foreach experimental loop, 122-123

demonstration program, 135-136

+ 不
translating a, 142-144

+
C 典
Free store memory, 8

front( ) STL sequence container function, 260, 284

无 经
func_ret( ), 365

Function objects, STL, 256, 258

futval( ), 190
Index

Garbage collection, 8, 9-11

timing of, 15, 33

Garbage collector, creating a C++, 11-15

algorithm for, 12-14

constraints for, 11

书 籍
pointer type for, 14. See also GCPtr

Garbage collector, multithreaded C++, 14-15, 96-118

典 书
code listing for, 102-116

读
load-testing demonstration program, 116-118

+ 经
Garbage collector, single-threaded C++, 15-65

不
+
code listing for, 16-28

C 典
gc( ), 99, 100-101

经
GCInfo garbage collector class, 28-29, 41-102, 99

无
GCiterator, 31, 44, 52

GCPtr multithreaded garbage collector class, 96-97, 102

constructor, 96, 97-99

copy constructor, 102

destructor, 96, 100

GCPtr single-threaded garbage collector class, 28, 29-41, 42, 43, 44

constructor, 31-32

copy constructor, 37-38

destructor, 32

GCPtr, using single-threaded, 44-64

 and arrays, 51-53, 64

and class types, 53-55

demonstration program, 55-62

and load testing, 62-64

and objects, 49-51

restrictions to, 64

get_args( ), 365

GetCurrentThread( ), 75-76

籍
getfname( ), 167, 174

getlowerbound( ), 286-287

典 书
get_params( ), 365

书
GetPriorityClass( ), 74

经 读
GetThreadPriority( ), 75, 87

+ 不
+
get_token( ), 325-330, 377

C 典
gettoken( ) translator function, 139-142

经
getupperbound( ), 286-287

无
Index

Heap, 8, 12, 13, 46

Heuristics, 203, 224

HttpQueryInfo( ), 161, 170-171, 174

httpverOK( ), 167, 170, 173-174

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

imbue( ), 190

initval( ), 192

insert( ) STL sequence container function, 259, 280

Internet file downloader, 160-186

code listing for, 161-166

书 籍
demonstration program, 177-178

GUI-based, 178-186

典 书
header file, 176-177

经 读
InternetAttemptConnect( ), 161, 169

+ 不
InternetOpen( ), 161, 169, 170

C+
InternetOpenUrl( ), 161, 169-170, 171, 172, 186

典
InternetReadFile( ), 161, 172

无 经
interp( ), 359-362, 365

InterpExc interpreter exception class, 324, 330, 356

Interpreter(s)

advantages of, 302-303

Mini C++. See Mini C++ interpreter

operation of, 333-334

versus compilers, 302-303

Interpreter portion of Mini C++, 303, 333-378

and assigning values to variables, 366-368

code listing, 333-354

and executing control statements. See Control statement execution in Mini C

 and global variables, 358-359

and handling cin and cout statements, 377-378

interp( ), function, 359-362, 365

interpreter prescan, 356-359

and local variables, 362-363

main( ) function, 354-356

and user-defined functions, 364-365

Investment, future value of

formula for computing, 190

书 籍
program for finding, 190-192

Investment required to achieve a future value, initial

典 书
formula for computing, 192

经 读
program for finding, 192-193

+
ishttp( ), 167, 168, 172-173

+ 不
isRunning( ), 101

C
经 典
Iter garbage collector class, 29, 31, 40, 42-44, 51, 52

iterator, 257

无
Iterator, STL, 256, 257

functions, 259, 284

Index

Java and C++, 5, 6

Java Virtual Machine, 302

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

lexicographical_compare( ) STL algorithm, 288

Libraries, C++, 3

list, 257

loadprogram( ), 355

籍
Loan balance, program to find, 197-199

书
Loan payments

formula for calculating, 188

典 书
program to compute, 188-190

经 读
+ + 不
C 典
无 经
Index

Macro substitutions, 120-121, 122

main( ), 307

map, 257

Mark and sweep, 12, 13

籍
match( ), 210, 217, 218, 219

书
max_size( ) STL sequence container function, 259, 285

maxwd( ), 196

经 典 读 书
mccommon.h Mini C++ header file, 303, 322, 381-383

Memory leak, 4, 9-10, 72

+ + 不
Memory management, dynamic

C 典
manual, 8-10, 11

and garbage collection, 10

无 经
Mini C++ interpreter, 302-394

compiling and linking, 383

demonstration programs, 383-393

expression parser, 303, 307, 309-333

interpreter portion of. See Interpreter portion of Mini C++

library functions, 303, 379-380

specifications, 304-306

suggestions for improving and expanding, 393-394

MoveWindow( ), 88

Multitasking, 68
Multithreading, 68-118

advantage of, 69

and C++, 69-70

definition of, 68-69

and synchronization, 76-79

and Windows, 70-79

Mutex semaphore, 77-79

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

Negator, STL, 258

new, 4, 8-9, 11, 45, 46

and garbage collection, 10

and GCPtr, 64

and STL containers, 258, 278, 279

Node, 203

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

Object model, 5

OOP (Object-Oriented Programming), 6

operator[ ]( ) function for STL container, 260, 262, 279-280

OutOfRangeExc garbage collector exception class, 28-29, 44, 53

籍
openfile( ), 167, 168, 174-175

书
Operator overloading, 5

经 典 读 书
+ + 不
C 典
无 经
Index

Parser

for Mini C++, 309-333

recursive descent, 309

table-driven, 309

pop_back( ) STL sequence container function, 260, 283

书 籍
pop_front( ) STL sequence container function, 260, 283

Predicates, STL, 256, 258

经
Process, 68
典 读 书
prescan( ), 355, 356-359

+ 不
<process.h>, 72

+
C
Production rules of expressions, 308-309

经 典
push_back( ) STL sequence container function, 260, 283

push_front( ) STL sequence container function, 260, 283

无
putback( ), 333, 362
Index

Range header, 167, 169

RangeArray custom STL sequence container, 261-299

code listing, 263-274

constructors, 262, 276-279

copy constructor, 279

书 籍
destructor, 279

erase functions, 282-283

典 书
front and back functions, 284

insert functions, 280-282

经 读
iterator functions, 284-285

+ 不
+
miscellaneous functions, 285-286

C
operator functions, 279-280

经 典
push and pop functions, 283-284

sample programs, 288-299

无
rbegin( ) STL sequence container function, 259, 299

Recursion, 202, 219

Reference counting, 12-13

and the circular reference problem, 12-13, 14, 40-41, 64

regpay( ), 188

ReleaseMutex( ), 78-79, 99

remove( ), 236-237

rend( ), STL sequence container function, 259, 299

repeat/until experimental loop, 125

 translating a, 149-151

resetAllSkip( ), 237

resetSkip( ), 222-223

ResumeThread( ), 71, 73, 90

Ritchie, Dennis, 6

route( ), 210, 216, 219-220, 243

routefound( ), 210, 216

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

Search class, general framework of, 209-210

Search space, 203

Search(es), AI-based

and backtrack stack, 210, 219

breadth-first, 220-224

书 籍
depth-first, 211-220

evaluating, 206

典 书
exhaustive, 205, 206, 211, 233, 243

读
to find the “optimal” solution, 243-249

+ 经
hill-climbing, 224-231

不
+
least-cost, 231-233

C 典
using node removal, 233, 235-242

using path removal, 233, 234-235

无 经
Semaphores, 77

SendDlgItemMessage( ), 88, 90

SetPriorityClass( ), 74

SetThreadPriority( ), 75

showprogress( ), 186

shutdown( ) garbage collector function

multithreaded, 99

single-threaded, 40-41

size( ) STL sequence container function, 259, 285

skipspaces( ), 142
sntx_err( ), 330-331

Software component, 3-4

Solution path, 203

stack, 257

Statements, C++, 306, 307

STL (Standard Template Library), 3-4

overview of, 256-258

Stroustrup, Bjarne, 6

书 籍
SuspendThread( ), 73, 90

swap( ) STL sequence container function, 259, 285

典 书
Synchronization, 76-79

经 读
Syntax, C++, 2-3

+ 不
SyntaxExc translator exception class, 137, 138

C+ 典
无 经
Index

tellp ( ), 175

Terminal node, 203

TerminateThread( ), 71-72, 89

Test and set operation, 77

籍
Thread

functions, Windows, 70-79

典 书
overview of, 68-69

书
priorities, 74-75, 79, 85-86

经 读
Thread control panel, 79-96

+ 不
code listing for, 80-85

C+
demonstration program, 91-96

典
ThreadCtrlPanel

经
class, 86

无
constructor, 86-88

threadfunc( ) prototypes for a thread function, 71, 72-73

ThreadPanel( ), 88-90

TimeOutExc garbage collector exception class, 99

Tokens, 324

Translator, C++, 120-138

code listing for, 125-135

demonstration program, 151-158

versus preprocessor, 120-121

Type extensibility, 5, 8

typeid operator, 124, 148, 149

typeof experimental operator, 124

translating a, 148-149

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

vector, 210, 257

Visual C++ compiler, 70

and thread functions, 72-73

书 籍
经 典 读 书
+ + 不
C 典
无 经
Index

Waitable timer, 77

WaitForSingleObject( ), 78-79, 99

WinDL GUI-based Internet file downloader, 178-186

Windows

Internet library, 160-161

书 籍
and multithreading, 70-79

WinINet library, 160-161

经 典 读 书
+ + 不
C 典
无 经
List of Figures

Chapter 3: Multithreading in C++

Figure 3-1: The Thread Control dialog box

Figure 3-2: Sample output from the thread control panel sample program

Chapter 5: An Internet File Downloader

Figure 5-1: The File Download dialog box

Chapter 7: AI-Based Problem Solving

书 籍
Figure 7-1: The solution path to find the missing keys

Figure 7-2: A combinatoric explosion with factorials

Figure 7-3: A directed graph of XYZ’s flight schedule

经 典 读 书
Figure 7-4: A tree version of XYZ’s flight schedule

Figure 7-5: The depth-first path to a solution

+
Figure 7-6: The breadth-first path to a solution

C+ 不
Figure 7-7: The hill-climbing path to a solution and to a false peak

Figure 7-8: The least-cost path to a solution

经 典
无
List of Tables

Chapter 8: Building a Custom STL Container

Table 8-1: The allocator Functions Used in This Chapter

书 籍
经 典 读 书
+ + 不
C 典
无 经