Fundamentals of Symbian C++

Two-Phase Construction and Object

Destruction

This work is licensed under the Creative Commons Attribution-Share Alike 2.0 UK: England & Wales License.

To view a copy of this license, visit http://creativecommons.org/licenses/bysa/2.0/uk/ or send a letter to Creative Commons, 171 Second Street, Suite 300, San Francisco, California,
94105, USA.
 Two Phase Construction
and Object Destruction

Introduction

Symbian OS is designed to perform well on devices with

limited memory
•  It uses memory-management models such as the cleanup stack to
ensure that memory is not leaked, even under error conditions or in
exceptional circumstances, such as when there is insufficient free
memory to complete an allocation
•  Two-phase construction is an idiom used extensively in Symbian
OS code to provide a means by which heap-based objects may be
constructed fully initialized, even when that initialization code may
leave

Fundamentals of Symbian C++ 2

 Two Phase Construction
and Object Destruction

Two-Phase Construction and Object

Destruction

Two-Phase Construction
•  Know why code should not leave inside a constructor
•  Recognize that two-phase construction is used to avoid the
accidental creation of objects with undefined state
•  Understand that constructors and second-phase ConstructL()
methods are given private or protected access specifiers in classes
which use two-phase construction, to prevent their inadvertent use
•  Understand how to implement two-phase construction, and how to
construct an object which derives from a base class which also
uses a two-phase method of initialization
•  Know the Symbian OS types (C classes) which typically use two-
phase construction

Fundamentals of Symbian C++ 3

 Two Phase Construction
and Object Destruction

No Code within a C++ Constructor should ever Leave

Consider this example:
class CFoo : public CBase
{... //Details omitted for clarity
public:
CFoo(TInt aVal)
{
iVal=aVal;
iBar = new(ELeave) CBar(aVal);
}
private:
CBar* iBar;
};
...

CFoo* foo = new(ELeave) CFoo(42);

Memory is allocated for foo and its constructor is called

But what if iBar = new(ELeave) CBar(aVal); in the CFoo constructor leaves?
Memory has already been allocated for the CFoo object
Because of the leave there is no valid pointer for the partially-constructed CFoo object
Without a valid pointer there is no way to clean up the CFoo object

Fundamentals of Symbian C++ 4

 Two Phase Construction
and Object Destruction

Two-Phase Construction

Two-phase construction breaks object construction into two parts

1.  A basic C++ constructor which cannot leave

•  It is this constructor which is called by the new operator.
•  It implicitly calls base-class constructors
•  It may also invoke functions that cannot leave and/or initialize
member variables with default values or those supplied as
arguments to the constructor

Fundamentals of Symbian C++ 5

 Two Phase Construction
and Object Destruction

Two-Phase Construction

Two-phase construction breaks object construction into two

parts
2.  A class method (typically called ConstructL())
•  This method may be called once the allocated and constructed object has
been pushed onto the cleanup stack
•  It completes construction of the object and may safely perform operations that
may leave.
•  If a leave does occur, the cleanup stack calls the destructor to free any
resources that had already been successfully allocated, and destroys the
memory allocated for the object itself

Fundamentals of Symbian C++ 6

 Two Phase Construction
Two-Phase Construction and Object Destruction

Splitting the construction allows a more atomic approach to handling

leaves.
•  The memory is allocated for foo and safely placed on the cleanup stack
•  The leaving function ConstructL is called separately
class CFoo : public CBase
{... //Details omitted for clarity
public:
CFoo(TInt aVal){iVal=aVal;}

void ConstructL()
{
iBar = new (ELeave) CBar(iVal);
}
private:
CBar* iBar;
};
...
CFoo* foo = new (ELeave)CFoo(42);
CleanupStack::PushL(foo);
foo->ConstructL();
CleanupStack::Pop(foo);

Fundamentals of Symbian C++ 7

 Two Phase Construction
and Object Destruction

Introducing NewL and NewLC

The previous example can introduce programming errors

•  Some callers may forget to call ConstructL() after instantiating the object
•  As it is not a standard C++ requirement for construction

The idiom
•  Is best implemented if the class itself makes the call to the second-phase
construction function, rather than the caller
•  Obviously the code cannot do this from within the simple constructor, since this
takes it back to the original problem of calling a method which may leave

Fundamentals of Symbian C++ 8

 Two Phase Construction
and Object Destruction

Introducing NewL and NewLC

The Symbian OS idiom

•  Is to provide a static function which wraps both phases of construction
•  Providing a simple and easily identifiable means to instantiate objects of
a class on the heap.
•  The function is typically called NewL()
•  A NewLC()function is often provided too and leaves the constructed
object on the cleanup stack for convenience (as discussed previously in
the cleanup stack section)

Fundamentals of Symbian C++ 9

 Two Phase Construction
and Object Destruction

Introducing NewL and NewLC

The NewL()and NewLC functions are static

•  Thus they can be called without first having an existing instance of the class

The non-leaving constructors and second-phase ConstructL()

functions have been made private
•  A caller cannot instantiate objects of the class except through NewL()/NewLC
•  Prevents duplicate calls to ConstructL() after the object has been fully
constructed

Fundamentals of Symbian C++ 10

 Two Phase Construction
Example Implementation and Object Destruction

/*static*/ CFoo* CFoo::NewLC(TInt aVal)

{
CFoo*self = new(ELeave) CFoo(aVal); // First-phase construction
CleanupStack::PushL(self);
self->ConstructL(aVal); // Second-phase construction
return self;
}

/*static*/ CFoo* CFoo::NewL(TInt aVal)

{
CFoo* self = CFoo::NewLC(aVal);
CleanupStack::Pop(this);
return self;
}

•  The NewL() function is implemented in terms of the NewLC() function

rather than the other way around (which would be slightly less efficient
since this would require an extra PushL() call on the cleanup stack)
•  Note the use of the Symbian OS overload of operator new(ELeave) this
implementation will leave if it fails to allocate the required amount of
memory.
•  This means that there is no need to check for a NULL pointer after a call to
new(ELeave)

Fundamentals of Symbian C++ 11

 Two Phase Construction
and Object Destruction

Deriving from a Class which uses Two-Phase

Construction

If any class is to be sub-classed

•  The default constructor should be made protected rather than private
so the compiler can construct the derived classes
•  C++ will ensure that the first-phase constructor of a base class is called
prior to calling the derived-class constructor

Two-phase construction is not part of standard C++ construction

•  The second-phase constructor of a base class will not be called
automatically when constructing a derived class
•  The second-phase ConstructL() method should be made private if it is
not be called by a derived class, or protected if it does need to be called -
document this in your code

Fundamentals of Symbian C++ 12

 Two Phase Construction
and Object Destruction

Deriving from a Class which uses Two-Phase

Construction

A class intended
•  For extension through inheritance which uses the two-phase
construction pattern typically supplies a protected method to do this
•  Called BaseConstructL() rather than ConstructL()

The derived class

•  Calls this method in its own ConstructL() method to ensure that
the base-class object is fully initialized

Fundamentals of Symbian C++ 13

 Two Phase Construction
and Object Destruction

Discussion Point: Using PushL in a Constructor

Two-stage construction for a class could be avoided

•  By including a CleanupStack::PushL(this) at the start of the
class's C++ constructor
•  This would achieve the same effect as using ConstructL() (since
PushL() will always store the this pointer safely for later cleanup)

If the class is to be used as a base class

•  the constructor of any classes derived from it will incur the overhead
of one push and pop in the constructor called at each level in the
inheritance hierarchy

This is less efficient

•  Than one pop and push in its own NewLC()
•  Better for C++ to automatically initialize the base classes, while the
developer manages the resource allocation of base class objects in
the most derived class’s NewLC()
Fundamentals of Symbian C++ 14
 Two Phase Construction
and Object Destruction

Two-phase construction and Symbian OS

types

Two-phase construction is typically used only for C classes

•  T classes do not usually require complex construction code as they
do not contain heap-based member data
•  R classes are usually created uninitialized requiring their callers to
call Connect() or Open() to associate the R-class object with a
particular resource

Fundamentals of Symbian C++ 15

 Two Phase Construction
and Object Destruction

Two-Phase Construction and Object

Destruction

Object Destruction
•  Know that it is neither efficient nor necessary to set a pointer to
NULL after deleting it in destructor code
•  Understand that a destructor must check before dereferencing a
pointer in case it is NULL, but need not check if simply calling
delete on that pointer

Fundamentals of Symbian C++ 16

 Two Phase Construction
and Object Destruction

Object Destruction
When implementing
•  The standard Symbian OS two-phase construction idiom it is
important to consider the destructor code carefully
•  A destructor must be coded to release all the resources that an
object owns
•  The destructor may be called to clean up partially constructed
objects if a leave occurs in the second-phase ConstructL()
function

The memory for a CBase-derived object

•  Is guaranteed to be set to binary zeroes on first construction
•  It is safe for a destructor to call delete on a NULL pointer

But ...

Fundamentals of Symbian C++ 17

 Introduction

Object Destruction

The destructor code

•  Cannot assume that the object is fully initialized
•  Should beware of calling functions on pointers which may not yet
be set to point to valid objects
•  Should beware of attempting to free other resources - by
dereferencing them - without checking whether the handle or
pointer which refers to them is valid:

CExample::~CExample()
{
if (iMyAllocatedMember) // iMyAllocatedMember may be NULL
{
iMyAllocatedMember->DoSomeCleanupPreDestruction();
delete iMyAllocatedMember; // No need to set it to NULL
}
}

Fundamentals of Symbian C++ 18

 Two Phase Construction
and Object Destruction

Two-Phase Construction and Object Destruction

  Two-Phase Construction

  Object Destruction

Fundamentals of Symbian C++ 19