USER’S GUIDE

TCAT for Java/Windows

Version 3.2

Test Coverage Analysis Tool

For Java on
Windows 2000/XP

SOFTWARE RESEARCH, INC.

 This document property of:

Name:_______________________________
Company:____________________________
Address:_____________________________
______________________________
Phone________________________________

SOFTWARE RESEARCH, INC.

1663 Mission Street, Suite 400
San Francisco, CA 94103
Tel: (415) 861-2800
Toll Free: (800) 942-SOFT
Fax: (415) 861-9801
E-mail: [email protected]
http://www.soft.com

ALL RIGHTS RESERVED. No part of this document may be reproduced, stored in a

retrieval system or transmitted, in any form or by any means, photocopying, recording
or otherwise without prior written consent of Software Research, Inc. While every pre-
caution has been taken in the preparation of this document, Software Research, Inc.
assumes no responsibility for errors or omissions. This publication and features
described herein are subject to change without notice.
TOOL TRADEMARKS: CAPBAK/MSW, CAPBAK/UNIX, CAPBAK/X,
CBDIFF, EXDIFF, SMARTS, SMARTS/MSW, S-TCAT, STW/Advisor, STW/
Coverage, STW/Coverage for Windows, STW/Regression, STW/Regression for
Windows, STW/Web, TCAT, TCAT C/C++ for Windows, TCAT-PATH, TCAT for
JAVA, TCAT for JAVA/Windows, TDGEN, TestWorks, T-SCOPE, Xdemo, Xflight,
and Xvirtual are trademarks or registered trademarks of Software Research, Inc.
Other trademarks are owned by their respective companies. METRIC is a
trademark of SET Laboratories, Inc. and Software Research, Inc. and STATIC is a
trademark of Software Research, Inc. and Gimpel Software.
Copyright  1999−2006 by Software Research, Inc.
Table of Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix

CHAPTER 1 TCAT for Java/Windows Overview . . . . . . . . . . . . . . . . . 1

1.1 The QA Problem . . . . . . . . . . . . . . . . . . . . . . . . . . .1
1.2 The Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
1.3 SR's Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
1.4 Testing and TCAT for Java/Windows . . . . . . . . . . . . . . .5
1.5 Software Test Methods . . . . . . . . . . . . . . . . . . . . . . .6
1.5.1 Manual Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
1.5.2 Static Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
1.5.3 Dynamic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
1.6 Single- and Multiple-Module Testing . . . . . . . . . . . . . . . .8
1.6.1 Bottom-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8
1.6.2 Top-Down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8
1.7 TCAT for Java's Cost Benefits . . . . . . . . . . . . . . . . . . .9
1.7.1 Improved Error Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
1.7.2 Earlier Error Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
1.7.3 More Efficient Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12
1.7.4 Minimal Test Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13
1.7.5 Assessment of Progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

CHAPTER 2 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1 System Requirements . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Installation Procedure . . . . . . . . . . . . . . . . . . . . . . 16
2.3 File List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

iii
TABLE OF CONTENTS

CHAPTER 3 Quick Start. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

3.1 Getting Acquainted with TCAT for Java/Windows . . . . . . . . 23
3.1.1 Step 1 - Preparing and Instrumenting TicTacToe . . . . . . . . . . . . . . . 24
Setup Environment Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Instrument Using WinIJava . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.1.2 Step 2 - Executing the Instrumented Application . . . . . . . . . . . . . . . 28
3.1.3 Step 3 - Viewing Coverage Reports Using Cover. . . . . . . . . . . . . . . 29
3.1.4 Step 4 - Viewing Directed Graphs with DiGraph. . . . . . . . . . . . . . . . 31
3.1.5 Step 5 - Viewing Source Code from a Digraph . . . . . . . . . . . . . . . . . 34
3.1.6 Step 6 - Viewing a Calltree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.7 Step 7 - Viewing the Directed Graph Associated With a Calltree Node
37
3.1.8 Step 8 - Viewing the Source Code Associated With a Calltree . . . . 38
3.1.9 Step 9 - Closing TCAT for Java/Windows . . . . . . . . . . . . . . . . . . . . . 39
3.2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

CHAPTER 4 Java Instrumentor Engine . . . . . . . . . . . . . . . . . . . . . . 41

4.1 Instrumentor Description . . . . . . . . . . . . . . . . . . . . . 41
4.1.1 Files Generated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 WinIJava Main Window . . . . . . . . . . . . . . . . . . . . . . 43
4.3 Instrumenting the Application Under Test . . . . . . . . . . . . 46
4.3.1 Options and Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.3.2 Instrumentation Function Names . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.3.3 Instrumentor Inline Directives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4 Database File Formats . . . . . . . . . . . . . . . . . . . . . . 50
4.5 Runtime Classes . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.5.1 Runtime Support Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.5.2 Performance Gain With Buffering . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

CHAPTER 5 Cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.1 Cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.2 Trace File and Archive File Formats . . . . . . . . . . . . . . . 53
5.3 Cover Main Window . . . . . . . . . . . . . . . . . . . . . . . . 54
5.3.1 Tool Bar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.3.2 File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.3.3 View Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3.4 Window Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3.5 Help Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3.6 Status Bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.4 File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

iv
 TCAT for Java/Windows User’s Guide

5.4.1 Open. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58

5.4.2 Print . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59
5.5 Window Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.5.1 Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
5.5.2 Tile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
5.5.3 Arrange Icons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
5.5.4 Window List Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
5.6 Create/Update an Archive File . . . . . . . . . . . . . . . . . . 61
5.7 Analysis of Coverage Reports . . . . . . . . . . . . . . . . . . 62

CHAPTER 6 DiGraph. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1 Purpose and Overview . . . . . . . . . . . . . . . . . . . . . . 67
6.2 Directed Graph File Format . . . . . . . . . . . . . . . . . . . 67
6.3 DiGraph Main Window . . . . . . . . . . . . . . . . . . . . . . 69
6.3.1 Tool Bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70
6.3.2 File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71
6.3.3 Zoom Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .72
6.3.4 View Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73
6.3.5 Options Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
6.3.6 Window Menu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
6.3.7 Help Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
6.3.8 Status Bar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
6.4 File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.4.1 Open. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75
6.4.2 Print . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76
6.5 View Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6.5.1 Viewing Associated Source Code . . . . . . . . . . . . . . . . . . . . . . . . . . .77
6.6 Options Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.6.1 The Digraph Options Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . .78
6.7 Window Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.7.1 Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81
6.7.2 Tile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82
6.7.3 Arrange Icons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83
6.7.4 Window List Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83

CHAPTER 7 CallTree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.1 Calltree Overview . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.2 Generating and Viewing Calltrees . . . . . . . . . . . . . . . . 86
7.3 Calltree File Format . . . . . . . . . . . . . . . . . . . . . . . . 87

v
TABLE OF CONTENTS

7.4 CallTree Window Overview . . . . . . . . . . . . . . . . . . . . 87

7.4.1 Tool Bar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.4.2 File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.4.3 View Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.4.4 Window Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.4.5 Options Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.4.6 Help Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.4.7 Status Bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.5 File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.5.1 Open . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.5.2 Print Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7.6 View Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.6.1 Viewing Associated Source Code. . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.6.2 Viewing a Directed Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.7 Window Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
7.7.1 Cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
7.7.2 Tile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.7.3 Arrange Icons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.7.4 Window List Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.8 Options Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

APPENDIX A Java Instrumentor Engine Database Files . . . . . . . . . 99

APPENDIX B Example Instrumentation Database Files. . . . . . . . . 105

APPENDIX C cover9 —TCAT for Java’s Coverage Analyzer . . . . . 108

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

vi
List of Figures

FIGURE 1 TCAT for Java/Windows Dependency Chart . . . . . . . . . . . . . . . . . . . . . . . . . . .4

FIGURE 2 Stages in Software Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
FIGURE 3 Cost Benefit Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
FIGURE 4 Increase in Cost-to-Fix Throughout Life-cycle . . . . . . . . . . . . . . . . . . . . . . . .11
FIGURE 5 Program Group for TCAT for Java/Windows . . . . . . . . . . . . . . . . . . . . . . . . . .20
FIGURE 6 Files for TCAT for Java/Windows in Windows 95 / NT . . . . . . . . . . . . . . . . . .22
FIGURE 7 WinIJava Window. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
FIGURE 8 Testing TicTacToe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28
FIGURE 9 Cover Main Window Displaying Coverage Report on TicTacToe . . . . . . . . .29
FIGURE 10 DiGraph Open Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
FIGURE 11 Select MDF ID Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
FIGURE 12 Directed Graph of TicTacToe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
FIGURE 13 Source Code Associated with Segment 3 of Digraph of TicTacToe::sta-
tus(int)34
FIGURE 14 Select Function ID Box:. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
FIGURE 15 Displaying a Calltree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36
FIGURE 16 Calltree of TicTacToe::mouseup(boolean) and Digraph of Its Possible Pro-
gram Flows37
FIGURE 17 Source Code Window Displayed from Calltree . . . . . . . . . . . . . . . . . . . . . . . .38
FIGURE 18 WinIJava . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43
FIGURE 19 Select File(s) to Instrument. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
FIGURE 20 IJava Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
FIGURE 21 Other IJava Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
FIGURE 22 Cover Main Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
FIGURE 23 Tool Bar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
FIGURE 24 Cover Open Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58

vii
LIST OF FIGURES

FIGURE 25 Print Dialog Window in Cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

FIGURE 26 Save Archive File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
FIGURE 27 Coverage Report Showing C1 Coverage of 66.67% on the Function TicTac-
Toe::myMove(boolean)62
FIGURE 28 Calltree and Digraph of TicTacToe::myMove(boolean). . . . . . . . . . . . . . . . . 63
FIGURE 29 Calltree and Source Code Associated with One Callpair . . . . . . . . . . . . . . . 64
FIGURE 30 Digraph and Source Code Associated with One of Its Segments . . . . . . . . 65
FIGURE 31 Program Edges as Represented in a Digraph . . . . . . . . . . . . . . . . . . . . . . . . 68
FIGURE 32 Directed Graph of TicTacToe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
FIGURE 33 Tool Bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
FIGURE 34 DiGraph Open Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
FIGURE 35 Print Dialog Box in DiGraph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
FIGURE 36 View Source Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
FIGURE 37 Digraph Options Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
FIGURE 38 Cascading Windows in DiGraph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
FIGURE 39 Tiled Windows in DiGraph. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
FIGURE 40 CallTree Main Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
FIGURE 41 Tool Bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
FIGURE 42 CallTree Open Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
FIGURE 43 Print Dialog Box in CallTree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
FIGURE 44 View Source Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
FIGURE 45 Directed Graph Option. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
FIGURE 46 Cascading Windows in CallTree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
FIGURE 47 Tiled Windows in CallTree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
FIGURE 48 CallTree Options Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

viii
Preface

Congratulations!

By choosing the TestWorks suite of testing tools, you have taken the first
step in bringing your application to the highest possible level of quality.
Software testing and quality assurance, while increasingly important in
today’s competitive marketplace, can dominate your resources and delay
your product release. By automating the testing process, you can assure
the quality of your product without needlessly depleting your resources.
Software Research, Inc. believes strongly in automated software testing. It
is our goal to bring your product as close to flawlessness as possible. Our
leading-edge testing techniques and coverage assurance methods are
designed to give you the greatest insight into your source code.
TCAT for Java/Windows is a quick and easy way to detect weaknesses in
your code. Easily accessible click-and-point reports find the segments
that need further testing. Digraphs and calltrees visualize the location,
allowing you to make immediate improvements to the structure and
performance of your software.
TestWorks is the most complete solution available, and the peace of mind
it provides our customers is our most valued feature.
Thank you for choosing TestWorks.

Audience
This manual is intended for software testers who are using TCAT for Java/
Windows. You should be familiar with the Microsoft Windows System and
your workstation.

ix
PREFACE

Typefaces
Typographical conventions that are used throughout this manual:
boldface Introduces or emphasizes a term that refers to
TestWorks’ window, its submenus and its options.
italics Indicates the names of files, directories, pathnames,
variables, and attributes. Italics is also used for man-
ual, chapter, and book titles.
”Double Quotation Marks”
Indicates chapter titles and sections. Words with
special meanings can also be set apart with double
quotation marks the first time they are used.
courier Indicates system output such as error messages,
system hints, file output, and CAPBAK/MSW’s
keysave file language.
Boldface Courier
Indicates any command or data input that you are
directed to type. For example, prompts and invoca-
tion commands are in this text. (stw, for instance,
invokes TestWorks.)

x
CHAPTER 1

TCAT for Java/Windows

Overview
This chapter is a conceptual introduction to coverage tools, and explains how to use them
most advantageously.

1.1 The QA Problem

It is a sad fact of the software engineering world that on average, without
coverage analysis tools, only around 50% of source code is actually tested
before release. With little more than half of the logic covered, many bugs
go unnoticed until after release. Worse still, the actual percentage of logic
covered is unknown to SQA management, making any informed
decisions impossible.
Questions such as when to stop testing or how much more testing is
required are answered not on the basis of data, but on ad hoc comments
and sketchy impressions. Software developers are forced to gamble with
the quality of the released software and to make plans based on
inadequate data.
A related problem is that test case development is done in an inefficient
manner; that is, many test cases are redundant. Test suites become
cluttered with cases that repeatedly test the same logic, to the exclusion of
other cases that would examine previously unexplored logic. Often,
testers are unsure of which direction to take, and can waste SQA time
devising the wrong tests.

1
CHAPTER 1: TCAT for Java/Windows Overview

1.2 The Solution

The primary purpose of testing is to ensure the reliability of a software
program before it is released to the end user. The software should be
thoroughly tested with a variety of input to provide statistically verifiable
means of demonstrating reliability. In other words, a suite of test cases
should in some way cover all the possible situations in which the pro-
gram will be used.
It is a worthy goal to imagine every possible use, and to develop and run
corresponding test data. However, achieving this goal is extremely com-
plicated and time-consuming. A more realistic goal is to test every part of
the program. According to industry studies, achieving this goal yields
significant improvement in overall software quality. Coverage analysis
improves the quality of your software beyond conventional levels.

2
 TCAT for Java/Windows User’s Guide

1.3 SR's Solution

Software Research, Inc. offers a solution: TCAT for Java/Windows. This
product ensures tests that are more diverse than those chosen by refer-
ence to functional specification alone or those based on a programmer's
intuition. It ensures that they are as complete as possible by measuring
against a range of high-quality test metrics:
• Coverage at the logical branch (or segment) level and the call-
graph level, employing the C1 metric
You can choose to test a single module, multiple modules, or the
entire program using the C1 metric.
• Coverage at the call-pair level employing the S1 metric
After individual modules have been tested, you can test all the
interfaces of the system using the S1 metric.
• Dynamic visualization of test attainment during unit testing and
system integration
This visually demonstrates, in real time, such things as segments
and call-pairs hit/not hit.Java.

3
CHAPTER 1: TCAT for Java/Windows Overview

Below is a TCAT for Java/Windows flow chart.

User

Source TCAT for Java/

File Windows
Reports
Archive File

FIGURE 1 TCAT for Java/Windows Dependency Chart

4
 TCAT for Java/Windows User’s Guide

1.4 Testing and TCAT for Java/Windows

TCAT for Java/Windows instruments your program. During instrumen-
tation, TCAT for Java/Windows inserts function calls (special markers) at
every logical branch (segment) in each program module. Instrumentation
also creates a reference listing file, which is a version of your program
which has logical branch marking comments added to it in a manner sim-
ilar to the code added to the instrumented version. Extensive logical
branch notation and sequence numbers are also listed.
This instrumented program is then compiled and run. By running it, you
are exercising logical branches in the program. The more tests in your test
suite, the higher the coverage. This test information is then written to a
trace file. From the information stored in the trace file, you can generate
coverage reports. In general, the reports give the following information:
• Reports included in the current iteration
• A summary of past coverage runs
• Current and cumulative coverage statistics
• A list of logical branches that have been hit
Recommended coverage is >85%. If reports indicate that you have less
than this amount, you can identify unexercised logical branches by study-
ing the coverage reports, and looking at the source code associated with
the untested functions. When you identify the troubled areas, you can
then create new test cases and re-execute the program.
TCAT for Java/Windows can help you reach your goal of creating the
most extensive test cases possible.

5
CHAPTER 1: TCAT for Java/Windows Overview

1.5 Software Test Methods

Coverage analysis as implemented through TCAT for Java/Windows is a
powerful testing technique which can save you much money and time, in
addition to greatly improving software quality. It is not the only testing
technique in existence, and we recommend that you use it along with
other techniques.
Testing methods vary from shop to shop, but most successful techniques
fall into a few general categories. The most common ones are described
below in the sequence they usually occur.

1.5.1 Manual Analysis

Programs are manually inspected for conformance to in-house rules of
style, format, and content as well as for correctly producing the antici-
pated output and results. This process is sometimes called “code inspec-
tion,” “structured review,” or “formal inspection.”

1.5.2 Static Analysis

Once a program has passed through manual testing steps, it can be tested
more extensively. Automated tools are used to check the design rules
applied in a program. Static analysis validates the software allegations
about the program's static properties, such as the global properties of its
data structures and the application of variable type rules. Such testing can
remove 20-30% of the latent software defects in your program. Static
analyzers include the following:
• Tools for detecting data element misuse
• Complexity measurement tools, which estimate the difficulty of
testing and help identify hard-to-test modules with a statistic
• Conformance measure tools, which flag confusing or inefficient
code

6
 TCAT for Java/Windows User’s Guide

1.5.3 Dynamic Analysis

Dynamic analysis tests the dynamic properties of the software under real
or simulated operating conditions. The software is executed under con-
trolled circumstances with specific expected results. In this phase, it is
important to test as many paths and branches in the program as possible.
Doing so ensures that the tests you run have the greatest diversity, hence
the best chance of discovering defects.
To obtain statistics on the application under test can be very difficult.
Dynamic analysis can uncover 85-90% of the potential remaining soft-
ware defects. TCAT for Java/Windows produces data on what has been
validated and what has been left out of your testing.

Source
Program

Manual
Analysis

Supporting Static
Documents Analysis Archived Test
Files

Archived Test
Dynamic Documents
Analysis

FIGURE 2 Stages in Software Testing

7
CHAPTER 1: TCAT for Java/Windows Overview

1.6 Single- and Multiple-Module Testing

Another consideration in getting the most out of TCAT for Java involves
determining the scope of your tests: whether a single program module,
multiple modules, or even an entire system should be tested. You can pre-
pare or “instrument” many modules with logical branch markers and
run tests on them as a group. TCAT for Java keeps track of each module
by name.
There are two approaches to multiple-module testing: bottom-up or top-
down. Because TCAT for Java is able to track many modules simulta-
neously, it supports either approach. The route you choose depends on
your individual needs and testing style.

1.6.1 Bottom-Up
In the bottom-up approach, testing begins at the lowest level in the
system hierarchy; that is, modules that invoke no other module. Each
bottom-level module is tested individually with special test data. Mod-
ules at each subsequent level of the hierarchy are tested using already-
tested lower-level modules. The process continues until all modules have
been thoroughly exercised. Thus, you can control testing carefully as you
progress up the system hierarchy.

1.6.2 Top-Down
In the top-down approach, testing begins at the highest level in the
system hierarchy. Sometimes module “stubs”' are used to simulate
invoked modules to check the high-level logic of the program. As an
alternative to using module stubs, use a complete program with only a
few selected modules instrumented. TCAT for Java ignores uninstru-
mented modules as it traces test coverage through the program.
In top-down analysis, the tester is chiefly concerned with the combination
of modules to form a larger system. TCAT for Java focuses specifically on
function calls within the system, so that the tester can verify each inter-
connection.

8
 TCAT for Java/Windows User’s Guide

1.7 TCAT for Java's Cost Benefits

TCAT for Java will save your organization much time and effort; the eco-
nomics of coverage analysis are extremely favorable. Here are some ways
it can save you money. TCAT for Java can save you money in the follow-
ing ways.

9
CHAPTER 1: TCAT for Java/Windows Overview

1.7.1 Improved Error Detection

TCAT for Java provides increased error detection. Software Engineering
literature indicates that an average error rate is 40 defects per 1,000 lines
of code (KLOC). With no coverage analysis, 50% of the code is exercised,
leaving the product with 20 defects per KLOC. Assuming a uniform dis-
tribution of errors throughout the source code, the simple act of raising
the coverage rate can uncover many errors. According to the experience
of SR in advanced industrial projects and reports from customers, cover-
age analysis can eliminate another 75% of the errors.

Without TCAT for Java With TCAT for Java

40 defects/KLOC 40 defects/KLOC

50% Coverage 85-90% Coverage

20 defects/KLOC 5 defects/KLOC

FIGURE 3 Cost Benefit Analysis

The economic value of increased error detection varies from organization
to organization. One estimate of the worth of coverage analysis comes
from what software consulting firms charge to find and remove errors, a
price established in the open market. The software testing industry, sized
at $50 million in 1986 by Fortune magazine, typically charges $1,000 per
error fixed.
Applying this to TCAT for Java, you could save $15,000 or more per
thousand lines of code. In practical terms, this means that a large project
with over 20,000 lines of code might save $300,000.

10
 TCAT for Java/Windows User’s Guide

1.7.2 Earlier Error Detection

Not only are more errors detected with TCAT for Java, they are also dis-
covered earlier. The earlier you catch and fix an error, the cheaper. Over
and over, managers, vendors and gurus have shown us figures and charts
that detail how much less it costs to rectify an early detected defect. The
chart below, by Barry Boehm, illustrates this concept.

1000
Larger Software Projects
500
200
100
80%
50 Median (TRW Survey)
20%
20
10
5 Smaller Software
Projects
2
1
Requirements Design Code Dev. Tests Acceptance Tests Operation

Phase in which error was detected and corrected

FIGURE 4 Increase in Cost-to-Fix Throughout Life-cycle

Your organization can reduce its cost-to-fix ratio by a factor of ten by
using TCAT for Java to find errors before system integration. In the dia-
gram, it costs $5,000 to $15,000 to fix errors after they have left the devel-
oper. The developer or the Software Quality Engineer (SQE) can identify
and fix problems more inexpensively than the beta site or independent
testing organization. This is not to say that beta sites or IV&V (indepen-
dent verification and validation) are not needed; but instead, there is a
great cost advantage in letting detailed unit-testing find more errors for
less expense.

11
CHAPTER 1: TCAT for Java/Windows Overview

1.7.3 More Efficient Testing

Using TCAT for Java, you can improve test case development. In general,
the tool can be used to identify previously untested features. This infor-
mation can direct the addition of new test cases.
For example, a software test engineer from a super-minicomputer manu-
facturer used TCAT for Java to reduce the time to test by a factor of eight.
As detailed in a technical article available from SR, the engineer was in
charge of testing a C compiler and used TCAT for Java to identify the fea-
tures missed by commercially-available test suites. The engineer speci-
fied the language elements that were not tested to a software engineer,
who completed the test suite. Overall, the compiler was fully tested in six
weeks rather than the expected one year.

12
 TCAT for Java/Windows User’s Guide

1.7.4 Minimal Test Set

TCAT for Java can be used to develop the minimal covering test suite for
a system. It is useful for a tester to have the smallest test suite that exer-
cises all the logic of a system, since test sets require much time and many
resources to execute.
We recommend the use of SMARTS, CAPBAK, and CBDIFF (from our
Regression/MSW tool suite) to automate test suite execution, evaluation,
and analysis steps. These tools can significantly reduce the cost of test
suite execution and analysis. TCAT for Java can be used to identify and
eliminate redundant test cases. With the coverage reports described in
this manual, it is possible to determine how much each new test case
adds to the total coverage of a test suite.
If a new test adds less than a specified amount to the overall coverage
(e.g. 5%) it might be reasonable to discard it. Having done so, the tester
ends up with more efficient, easier-to-run test suite.

13
CHAPTER 1: TCAT for Java/Windows Overview

1.7.5 Assessment of Progress

Coverage analysis with TCAT for Java can be valuable to important SQA
decisions, such as when to ship a product or how much further product
testing is needed. A coverage value of C1 > 85% has been the traditional
threshold for proper coverage. Generally, one should stop improving test
coverage when the marginal cost of adding a new test is greater than the
cost to visually and rigorously inspect the associated code passage. Other
considerations you can weigh are the added test cost and the risk of
defects.
Coverage analysis data are important for reliability modeling and pre-
dicting error rates. By tracking error rates and number of errors discov-
ered as a function of overall test effort, it is possible to predict eventual
latent defect rates. We encourage SQA managers to keep careful records
of errors found and corresponding coverage values.

14
CHAPTER 2

Installation
This chapter describes the system requirements and the step-by-step installation
procedure for TCAT for Java/Windows

2.1 System Requirements

Your computer system must have the following hardware configuration
to install and run TCAT for Java/Windows.
• Windows 95 or NT
• 486 microprocessor or better
• 7.7 MB free disk space
• 16+ MB RAM recommended
Java compiler must be installed.

15
CHAPTER 2: Installation

2.2 Installation Procedure

1. Insert the diskette labeled Disk 1 in your diskette drive (these
instructions assume A:).
2. Activate setup.exe. :
In Windows 95, or NT 4.0:
a. Using either the My Computer icon (on the desktop) or
Windows Explorer (on the Start menu, Programs submenu),
display the contents of the X: drive ( X: the floppy dirive or
CD-ROM drive ).
b. Double-click setup.exe.

16
 TCAT for Java/Windows User’s Guide

setup.exe presents you with a series of dialog boxes, beginning with the
Welcome box shown below. Each box is a step in the installation process,
and when you are satisfied with the options offered in a box you should
click Next to go on to the next step.

3. Click Next in the Welcome box.

The Choose Destination dialog box asks you where you would like to
store the executables and the supporting files for TCAT for Java/
Windows.

17
CHAPTER 2: Installation

4. To select a path, do one of the following:

• Click on Next if you want to use the Path indicated and to con-
tinue the installation.
• Edit the default path to your own path, then click Next to con-
tinue the installation.
• Click Cancel to end the installation.

18
 TCAT for Java/Windows User’s Guide

After selecting Next, the Setup Type dialog box pops up and asks you
what kind of installation you prefer. It is highly recommended that you
select Custom installation, which allows you to install the FrameReader
software that allows you to read the online help that accompanies TCAT
for Java/Windows. (Be aware that the FrameReader software will occupy
approximately 9 MB of your computer’s memory.)

5. In the Setup Type dialog box, do one of the following:

• Click Next if the Setup Type is the one you prefer.
• Click a different Setup Type, then click Next to continue the
installation.
• Click Back to review or change previous dialog box queries.
• Click Cancel to end installation.

19
CHAPTER 2: Installation

6. Select the components that you want copied.

During copying, a bar gauge names the files being copied.
C:\Program Files\Software Research\Coverage\TCAT-Java directory or
the path you indicated is created. TCAT for Java/Windows automati-
cally stores your files to this directory unless you selected otherwise.
The installation script also creates a program group where TCAT for
Java/Windows and its utilities are installed:

FIGURE 5 Program Group for TCAT for Java/Windows

20
 TCAT for Java/Windows User’s Guide

7. When the installation is complete, include the Coverage pathname in

your system environment variable.
8. To uninstall, use the following:
a. Double click the Add/Remove Programs icon in the Control
Panel.
b. Click the Remove button.

21
CHAPTER 2: Installation

2.3 File List

The following files are written to your computer during the installation.
The locations for these files are given for installation to a directory called
C:\Program Files\Software Research\Coverage\TCAT-Java.

FIGURE 6 Files for TCAT for Java/Windows in Windows 95 / NT

22
CHAPTER 3

Quick Start
This chapter explains getting started with TCAT for Java/Windows using a demonstra-
tion test case. It then describes the main features of the product.

3.1 Getting Acquainted with TCAT for Java/Windows

This section will familiarize you with the main activities involved in
using TCAT for Java/Windows, including instrumenting, compiling, and
running the target program, and finally, looking at the resulting coverage
reports, calltree graphs and digraphs.
The applet used to illustrate the operation of TCAT for Java/Windows in
Windows is TicTacToe, which you will prepare and instrument as a test
application. You can then exercise various logical branches or segments of
TicTacToe, creating trace files from which the coverage reports are gener-
ated. It is recommended that you complete the TicTacToe example before
continuing.
If you are using TCAT for Java/Windows for the first time, you will bene-
fit most if you refer to chapters 4 through 7 for in-depth operational
instructions and detailed explanation of functionality. If you are an inter-
mediate user, you’ll only have to refer to those menu definitions which
need further explanation.

23
CHAPTER 3: Quick Start

3.1.1 Step 1 - Preparing and Instrumenting TicTacToe

3.1.1.1 Setup Environment Variables

For the first time user, check your Java manual to see how the environ-
ment variables are set. Add $TCAT-Java_DIR\Program\ to CLASSPATH.
( e.g. set CLASSPATH=.;C:\jdk1.1.4\lib\classes.zip; C:\Program Files\
Software Research\Coverage\TCAT-Java\Program) )

24
 TCAT for Java/Windows User’s Guide

3.1.1.2 Instrument Using WinIJava

WinIJava instruments the application under test so that any tests can
produce trace files.
To instrument the example application:
1. Start up WinIJava.

FIGURE 7 WinIJava Window

2. Select TicTacToe.java using the Select button.

Note: More than one file can be selected and instrumented, and
instrumenting multiple files results in more thorough coverage.

25
CHAPTER 3: Quick Start

3. Select Instrument.
A copyright box pops up before the instrumentation of each file if
the license is invalid. During instrumentation, a command-line win-
dow displays messages and warnings. The instrumentor parses the
applet’s source code, looking for logical branches or segments and
inserting markers (function calls).
Instrumenting a program does not change its functionality. When
compiled, and executed, the instrumented application behaves as it
normally does, except that it writes coverage data to a trace file. For
more information on TCAT for Java/Windows’ instrumentor, refer to
Chapter 4.
4. When instrumentation is complete, select Close from the WinIJava
window.

26
 TCAT for Java/Windows User’s Guide

Instrumenting TicTacToe.java produces the following files in the TicTacToe

directory:
• TicTacToe.i — the instrumented version of the source file
This file is updated during the instrumentation process.
• TicTacToe.dg — a Directed Graph Listing file
Each instrumented file should have its own .dg file.
• TicTacToe.cg — a Calltree Graph Listing file
Each instrumented file should have its own .cg file.
• Prj_Name.mdf — a Module Definition file
This file contains information about segments and callpairs in all
the processed files.
• mdf.pro — a profile of the applet for useing on your Web server.

27
CHAPTER 3: Quick Start

3.1.2 Step 2 - Executing the Instrumented Application

During instrumentation, TCAT for Java inserted function calls at each
logical branch it found. In order to later determine the C1 coverage, you
must run the applet.
By running TicTacToe and playing the game, you are exercising segments
of the TicTacToe program. Because you have instrumented the applet, the
exercise will create a trace file and allow you to view coverage informa-
tion on the exercise.
To run the instrumented applet:
1. Open a DOS window. From DOS prompt, CD to TicTacToe directory.
2. Type appletviewer TicTacToe.html.
3. The appletviewer and TicTacToe applet will appear. Play the game.
4. When you are finished playing, select the Applet and choose “Quit”.

When the TicTacToe is running, your display should look like this:

FIGURE 8 Testing TicTacToe

28
 TCAT for Java/Windows User’s Guide

3.1.3 Step 3 - Viewing Coverage Reports Using Cover

1. From the Program menu, select Testworks for Win32 folder.
2. From the resulting window, select Cover icon.
3. From the File menu, select Open.
4. In the Open dialogue, click on the filename Trace.trc from the
tcat_db\Prj_Name directory.
A coverage report of the test you ran on the example program appears.

FIGURE 9 Cover Main Window Displaying Coverage Report on TicTacToe

29
CHAPTER 3: Quick Start

Cover displays trace and coverage information on your development

project in a treelike list. Clicking on a branch of the list expands the
branch and shows its contents, and also contracts it. The several fields in
the report have the following meanings:
Hits The number of times the segment and call pair were executed
during the test
Count The number of segments and call pairs within the function
C1 The percentage of branch coverage for each function
S1 The percentage of call pair coverage for the function
For detailed information about Cover, see Chapter 5.

30
 TCAT for Java/Windows User’s Guide

3.1.4 Step 4 - Viewing Directed Graphs with DiGraph

To view a directed graph (digraph) of possible program flows of a
function:
1. From the TCAT Program Group, select DiGraph.
2. Using the File menu, select Open.
3. A selection box asks for the name of the directed graph to view. For
this example, find the TicTacToe.dg file under the
tcat_db\Prj_Name\d_graph directory.
A selection box asks for the name of the module definition file.
4. Find the TicTacToe.mdf file under the tcat_db\Prj_Name directory (one
level up from the TicTacToe.dg file).

FIGURE 10 DiGraph Open Dialog Box

A selection box asks which function to display.

31
CHAPTER 3: Quick Start

5. For this example, select TicTacToe::status(int).

FIGURE 11 Select MDF ID Box

32
 TCAT for Java/Windows User’s Guide

A directed graph depicting possible program flows of the function

TicTacToe::status(int) appears.

FIGURE 12 Directed Graph of TicTacToe

The digraph shows the set of conditions and paths that make up a
function. The next step shows how to look at the code that the digraph
displays as numbered segments.

33
CHAPTER 3: Quick Start

3.1.5 Step 5 - Viewing Source Code from a Digraph

To view the source code represented by a particular segment of the
function TicTacToe::status(int):
1. Click near the number of the segment.
2. From the tool bar, select the View Source Code button.

FIGURE 13 Source Code Associated with Segment 3 of Digraph of TicTacToe::status(int)

34
 TCAT for Java/Windows User’s Guide

3.1.6 Step 6 - Viewing a Calltree

To view a calltree of TicTacToe:
1. From the TCAT Program Group, select Calltree.
2. In the File menu, select Open.

You are prompted for the name of the calltree to view.

3. Find TicTacToe.cg file under the tcat_db\Prj_Name\ c_graph directory.
You are prompted for the name of the database file.
4. Find the TicTacToe.mdf file under the tcat_db\Prj_Name directory.
A window appears asking you which function to display.
5. For this example, select TicTacToe::mouseup(boolean).

FIGURE 14 Select Function ID Box:

35
CHAPTER 3: Quick Start

A calltree depicting the selected function appears.

FIGURE 15 Displaying a Calltree

The calltree shows all of the callpairs associated with the function
TicTacToe::mouseup(boolean).

The next step shows how to look at digraphs of the possible program
flows belonging to this function.

36
 TCAT for Java/Windows User’s Guide

3.1.7 Step 7 - Viewing the Directed Graph Associated With a Calltree Node
To display a directed graph of any callpair shown in the calltree:
1. Select a node by clicking on it.

Notice that the View Digraph button on the toolbar now has a red
arrow, indicating that it is available.
2. To display a directed graph of the selected function, click the View
DiGraph button.

FIGURE 16 Calltree of TicTacToe::mouseup(boolean) and Digraph of Its Possible Program

Flows

37
CHAPTER 3: Quick Start

3.1.8 Step 8 - Viewing the Source Code Associated With a Calltree

You can view the source code associated with any node in a calltree by
clicking on the corresponding edge.
Notice that the Source Code button on the Tool Bar has a red arrow.
1. To display the associated source code, click the Source Code button.

The code is displayed in a separate window with the calling

statement highlighted in red.

FIGURE 17 Source Code Window Displayed from Calltree

38
 TCAT for Java/Windows User’s Guide

3.1.9 Step 9 - Closing TCAT for Java/Windows

After looking at the source code, select one of the following options to
complete the session.
To close TCAT for Java/Windows:
• Select File|Exit from the menu bar of each open program.
• In Windows NT: double-click on the frame window Close Box of
each program.
• In Windows 95: click on the frame window Close Box of each
program.
You have now seen all the main features of TCAT for Java/Windows.

39
CHAPTER 3: Quick Start

3.2 Summary
If you have completed the proceeding steps successfully, you have seen
and practised the basic skills you need to use TCAT for Java/Windows
productively. You should have learned how to invoke TCAT for Java/
Windows, how to instrument, compile, and run a program, and how to
look at the coverage reports.
For best learning you may want to:
• Repeat STEPS 1 - 9 without the manuall and experiment by run-
ning the applet several times and looking at the amount of cover-
age your test input receives.
• Repeat STEPS 1 - 9 with your applet.
• Review the chapters on system operation where you had difficul-
ties. The table of contents can help you locate the topic you want.

40
CHAPTER 4

Java Instrumentor Engine

This chapter discusses the TCAT for Java/Windows integrated Java instrumentor. This
chapter applies to all editions of TCAT for Java/Windows.

4.1 Instrumentor Description

WinIJava instruments the source code of the application under test by
inserting function calls at each logical branch and call pair. The instru-
mentation does not affect the functionality of the program. When com-
piled, and executed, the instrumented program will behave normally,
but writes coverage data to a trace file.
There is some performance overhead related to the data collection
process, but the overhead varies with the choice of the runtime used.
The trace files are processed by several kinds of report generators.
There is a single version of the instrumentor engine for Java programs.

41
CHAPTER 4: Java Instrumentor Engine

4.1.1 Files Generated

In operation, the WinIJava instrumentor parses candidate source code
looking for logical branches and/or call pairs and generates auxiliary
files that are used by other parts of the system. TCAT for Java/Windows
uses and produces the following files:
Instrumenting TicTacToe.java produces the following files in the Example
directory:
• TicTacToe.i — the instrumented version of the source file
This file is updated during the instrumentation process.
• TicTacToe.dg — a Directed Graph Listing file
Each instrumented file should have its own .dg file.
• TicTacToe.cg — a Calltree Graph Listing file
Each instrumented file should have its own .cg file.
• Prj_Name.mdf — a Module Definition file
This file contains information about segments and callpairs in all
the processed files.
• mdf.pro — a profile of the applet for useing on your Web server.
• Trace.trc — produced when the instrumented application is
executed
This file contains coverage information for the current test.

42
 TCAT for Java/Windows User’s Guide

4.2 WinIJava Main Window

FIGURE 18 WinIJava
WinIJava drives the instrumentor, IJava, according to selections made by
the user.
Select Click a file to select it for instrumentation, control-
click to select several files, or shift-click to select a
series of files.
Instrument Instruments the selected file(s). During instrumenta-
tion, a command-line box gives informational and
warning messages.
Options Selects among code languages and modes of instru-
mentation.
Close Exits WinIJava.

43
CHAPTER 4: Java Instrumentor Engine

FIGURE 19 Select File(s) to Instrument

FIGURE 20 IJava Options

44
 TCAT for Java/Windows User’s Guide

Figure 20 shows the default options for IJava.

On Windows 95/NT systems, any alterations generated here are written
to the Registry key HKEY_CURRENT_USER\ Software\Software
Research\Coverage\TCAT-Java\1.2\WinIJava, from which WinIJava reads
them. The Defaults button retrieves the contents of Registry key
HKEY_LOCAL_MACHINE\ SOFTWARE\Software Research\Coverage\
TCAT-Java\1.2\WinIJava to this box.

FIGURE 21 Other IJava Options

For the Instrumentation options, the usual assumption is that more cover-
age is better. Note that S0 coverage requires S1 coverage and cannot be
selected unless S1 coverage is also selected.
Selecting the Keep Instrumented File option means that the *.i file created
during instrumentation is retained. Should the instrumentation fail, this
file can be debugged for information.
Selecting the Instrument Only option prevents IJava from compiling and
producing an *.i file.
The Interactive option makes the instrumentation more visible. The inter-
activity means that the IJava command line window, which is present
during instrumentation, waits for the user to exit from it before closing
down to begin instrumentation of the next file or to return to WinIJava.
This ensures that the user can read the messages and warnings in the
window.

45
CHAPTER 4: Java Instrumentor Engine

4.3 Instrumenting the Application Under Test

4.3.1 Options and Parameters

The syntax for command line invocation of IJava is as follows:
IJava <<option>> file.ext
[-TCAT-A]
[-TCAT-B]
[-TCAT-Cmd driver]
[-TCAT-C1]
[-TCAT-G]
[-TCAT-H]
[-TCAT-PD name]
[-TCAT-PN name]
[-TCAT-S0]
[-TCAT-S1]
These commands instrument submitted “Java” language file(s).
The directory specified with the -TCAT-PD switch becomes the project
directory for the instrumentation. Within this directory, the tcat_db
directory is automatically created. The directory name specified with the
-TCAT-PN switch is created under the tcat_db directory, and contains the
trace file, the module definition file, and the c_graph and d_graph directo-
ries. These lowest directories contain the *.cg and *.dg files, respectively.

46
 TCAT for Java/Windows User’s Guide

If you invoke IJava with the switches -TCAT-PD c:\AAA and -TCAT-
PN XXX on the file example.c, the directory tree created during instru-
mentation is as follows:

c:\
AAA
tcat_db
XXX
c_graph
example.cg
d_graph
example.dg
XXX.mdf

The following instrumentor switches may be used to vary the processing

and reports generated by the instrumentor. The instrumentor switches
are listed in alphabetical order. The trace file (trace.tre) will be created in
the Java Program directory after execution.
Note that the commands are prefixed with -TCAT. This is done because
all other switches are passed to the “Java” compiler. The prefix indicates
that these switches are for TCAT processing.

47
CHAPTER 4: Java Instrumentor Engine

file.ext Instrumented File Specification(s); File(s) to be

instrumented
The extension is java.
If there are multiple files, each one is processed in the
order presented, and they are treated as if they have
been concatenated together.
-TCAT-B Non-Interactive Instrumentation Switch
Instrumentation does not require any input from test-
ed even if more than one file is being instrumented.
-TCAT-Cmd driver Compiler Driver Command Switch
Default driver is cc. For Microsoft Visual C, use
cl.exe.-TCAT-C1 C1 Instrumentation Switch
If this switch is present, then the instrumentor inserts
a function call in each segment, or logical branch. This
is the preset default.

48
 TCAT for Java/Windows User’s Guide

-TCAT-G Instrumented File Disposition Switch

Normally the instrumentor does not keep the
instrumented file, because it has already been used to
produce the instrumented output. When this switch
is present the instrumented files are retained.
-TCAT-Help Help Message Switch
This switch prints out the set of valid switches.
-TCAT-i Instrumentation Only Switch
WinIJava instruments the target application but does
not generate an object file. -TCAT-i overrides the
-TCAT-cmd switch.
-TCAT-PD name Project Directory Switch
This switch specifies the location of the “project”
directory.
-TCAT-PN name Project Name Switch
This switch specifies the project name.
-TCAT-S0 S0 Instrumentation Switch
If this switch is present, then the instrumentor inserts
a function call in each module. This tells you which
functions are actually called during the invocation of
the program, but it does not indicate the callee func-
tions. To do this, you need to use the -S1 switch.
-TCAT-S1 S1 Instrumentation Switch
If this switch is present, then the instrumentor inserts
a function call in each call pair.

49
CHAPTER 4: Java Instrumentor Engine

4.3.2 Instrumentation Function Names

Instrumentation involves inserting function names into the source pro-
gram. The function names for TCAT-instrumented programs are:
Testworks.Runtime.SegHit(); For entry segment,
switch segments
Testworks.Runtime.CprHit();For S1 coverage of call
pairs
ExpHit(); For C1 coverage if's, while's and
for's

4.3.3 Instrumentor Inline Directives

It is possible to control instrumentation from within the processed “java”
file, using the following instrumentor directives to turn off/on all instru-
mentation (but keep the segments and call pairs numbered
correctly):
/* TCAT OFF */
/* TCAT ON */

4.4 Database File Formats

For information on the format of WinIJava output files, see Appendix A,
“Java Instrumentor Engine Database Files.”

50
 TCAT for Java/Windows User’s Guide

4.5 Runtime Classes

This section is a guide to TCAT for Java/Windows’™ Runtime Options
applies to all editions of the product.

4.5.1 Runtime Support Options

Provided with TCAT for Java/Windows is a package containing the
TCAT for Java/Windows runtime classes.
By default, this package is installed into the $SR/program directory, where
$SR is your TCAT for Java/Windows installed directory. The actual
classes in the package are installed into $SR/program/testworks/runtime
directory,
Your instrumented Java classes must be able to find this package. There-
fore, during instrumentation and execution your CLASSPATH environ-
mental variable must point to the top of this package. You can either
append the program directory to your CLASSPATH or copy the testworks
directory to one of the CLASSPATH directories.
Inside the $SR/program/testworks/runtime directory are several classes that
implement various levels of trace buffering.
The default level of buffering is one. This is effectively no buffering, since
as each trace hit occurs it is written to the trace file. To increase the size of
buffering, replace jrun.class by one of the desired class from the chart
below before executing your instrumented applet or application.

51
CHAPTER 4: Java Instrumentor Engine

Buffering level
Class Name

jrun1.class None
jrun10.class 10 Trace Hits
jrun100.class 100 Trace Hits
jrun10000.class 10000 Trace Hits
jrunInt.class Infinite, Trace hits are not written to trace
file until termination of application.

TABLE 1 Buffering level for runtime class

Note: that if your applet or application terminates abnormally that up to

one buffer full of trace data will be lost.

4.5.2 Performance Gain With Buffering

The larger the trace buffer the better the performance of your instrumen-
tal application. However, it should be noted that the trace records up to
the size of a buffer may be lost if the program is terminated abnormally.

52
CHAPTER 5

Cover
This chapter discusses Cover, the TCAT for Java/Windows complete TCAT Java analyzer
for branch (C1) and callpair (S1) metrics. This chapter applies to all editions of the prod-
uct.

5.1 Cover
Cover analyzes the trace files created when an instrumented program is
executed, and generates reports based on the trace file data.These cover-
age reports can be tailored to show a variety of data, including:
• segments hit
• segments not-hit
• past-test and cumulative coverage percentages
Cover makes the following assumptions:
• A [possibly empty] archive file and a current [possibly empty]
trace file exist.
• There is a file containing the names of the files in the project.
• The actual update of trace + archive --> archive is optional at end
of a session.
The package maintains its usual rules for precedence of archive over
trace, and displays warning messages when it finds size differences
between archive and trace file.

5.2 Trace File and Archive File Formats

For information on the format of trace files and archive files, see
Appendix A, “Java Instrumentor Engine Database Files.”

53
CHAPTER 5: Cover
5.3 Cover Main Window
Once you have built an instrumented version of your application and
exercised it, follow these steps to display a coverage report:
1. From the Programs menu, select TCAT for Java folder.
2. From the resulting window, select Cover icon.
3. From the File menu, select Open.
4. In the Open dialogue box, click on the filename Trace.trc in the tcat_db
directory.
A coverage report on the application appears.

FIGURE 22 Cover Main Window

54
 TCAT for Java/Windows User’s Guide

5.3.1 Tool Bar

The options available from the Tool Bar are the frequently used Cover
features.

FIGURE 23 Tool Bar

Open This option brings up the Open dialog box.
Print Button This button brings up the Print dialog box.
Help This button brings up a brief description of Cover.

55
CHAPTER 5: Cover

5.3.2 File Menu

This menu displays the file management and printing options that are
available in Cover.
Open This option brings up the Open dialog box.
Print This option brings up a the Print dialog box.
Print Preview This option displays an image of what prints when
you select the Print option.
Print Setup This option displays a standard Windows printer set-
up dialog box.
Exit To end your Cover session, select the Exit option.

56
 TCAT for Java/Windows User’s Guide

5.3.3 View Menu

This menu provides two options for configuring the Cover display.
Toolbar This toggle allows you to hide the Tool Bar in order to
give your report more vertical display space or to re-
display it.
Status Bar This toggle allows you to hide or re-display the status
bar at the bottom of the Cover window.

5.3.4 Window Menu

This menu allows you to manipulate the Cover windows using the
Cascade, Tile and Arrange Icons options, and the Window list box.

5.3.5 Help Menu

The first help option currently offers a brief description of Cover. The
second option, About, displays the program’s version number and copy-
right information.

5.3.6 Status Bar

This section of the window (appearing at the bottom left) displays
messages regarding the functionality and operation of the Cover options.

57
CHAPTER 5: Cover

5.4 File Menu

This menu is typical of Windows interfaces and provides access to file-
manipulation options.

5.4.1 Open

FIGURE 24 Cover Open Dialog Box

This option brings up a file selection dialog box. Typical of Windows
interfaces, this dialog allows you to browse the directory tree and select
files to open. Since all trace files are usually saved as trace.trc, each project
has only one trace file.
File Name This box lists the files in the current directory that
match the filter.
Directory This box lists the available directories.
When you have found the desired file, click OK, and the coverage report
is displayed. Cancel closes the dialog box without opening a report.

58
 TCAT for Java/Windows User’s Guide

5.4.2 Print

FIGURE 25 Print Dialog Window in Cover

59
CHAPTER 5: Cover

5.5 Window Menu

This menu provides four options to manipulate the Cover windows. By
default the active window entirely overlaps all others.

5.5.1 Cascade
This option arranges your windows in a cascade, with the active window
top-most and highlighted.

5.5.2 Tile
This option arranges the windows so that a portion of each window is
displayed. The active window is highlighted.

5.5.3 Arrange Icons

When you have minimized windows, this option arranges them neatly at
the bottom of the Cover window.

5.5.4 Window List Box

This area of the pull down-menu lists all the windows open in Cover. The
active window is indicated by a check mark. To activate a new window,
especially if the windows are fully overlapping, select it from this list.

60
 TCAT for Java/Windows User’s Guide

5.6 Create/Update an Archive File

If no archive file is loaded, this option creates one by copying the current
*.trc file as an *arh file. Updating combines the information from the
current *.trc file with that of the selected *.arh file.

FIGURE 26 Save Archive File

61
CHAPTER 5: Cover

5.7 Analysis of Coverage Reports

In the following analysis, a coverage report shows that a certain function,
TicTacToe::myMove(boolean), has been tested 66.67%.

FIGURE 27 Coverage Report Showing C1 Coverage of 66.67% on the Function

TicTacToe::myMove(boolean)
The function consists of three segments and one callpair. This coverage
report shows that segments 1 and 3 were hit twenty four times each and
segment 2 not once. The two callpairs were each exercised twenty seven
times.
The following few pages show graphical views of these numerical results.

62
 TCAT for Java/Windows User’s Guide

In Figure 28, TCAT for Java/Windows graphs

TicTacToe::myMove(boolean) and its relations. The calltree shows the
callpairs in TicTacToe::myMove(boolean), and the digraph shows possible
program flows through TicTacToe::myMove(boolean) divided into seg-
ments.

FIGURE 28 Calltree and Digraph of TicTacToe::myMove(boolean)

Note that the calltree shows two callpairs: these are the same two call-
pairs registered by the coverage report in Figure 15 as having been exer-
cised seven times each by the test of TicTacToe. The coverage report
shows that the percentage of S1 coverage (coverage of call pairs) was
100% for this function.
Note that the digraph shows three segments. The coverage report shown
in Figure 32 registered that the test of TicTacToe hit two of these segments
seven times each and one of them not once. The coverage report shows
that the percentage of C1 coverage (branch coverage) was 66.67%.

63
CHAPTER 5: Cover

FIGURE 29 Calltree and Source Code Associated with One Callpair

To look at source code associated with callpairs, highlight the graphic
lines connecting the functions shown in the calltree.

64
 TCAT for Java/Windows User’s Guide

To look more closely at the segments, highlight one of the graphic lines in
the digraph by clicking on it close to the number. Then use the Source
Code button to display the associated source code.

FIGURE 30 Digraph and Source Code Associated with One of Its Segments

65
CHAPTER 5: Cover

66
CHAPTER 6

DiGraph
This chapter provides details on viewing and using directed graphs in TCAT for Java/
Windows.

6.1 Purpose and Overview

Directed graphs (digraphs) graphically display a program’s structure and
flow to help developers isolate flaws and bottlenecks.
TCAT for Java/Windows draws digraphs based on archive files that are
created during instrumentation. Digraphs are composed of edges and
nodes. Edges are derived from segments (also known as logical branches)
representing sets of consecutive program statements or a program’s
“actions” (see Figure 31). Nodes are the places or “states” where the
actions occur.
6.2 Directed Graph File Format
For information regarding the format of a directed graph chart file, see
Appendix A, “Java Instrumentor Engine Database Files.”

67
CHAPTER 6: DiGraph

Succession
A Statement:

statement A;
B statement B;

Alteration
Statement:
A statement A;
if condition then
statement B;

B C else
statement C;
end if;

Iteration
Statement:
A statement A;
while condition
B statement B;
C
loop;
statement C;

Case Statement:

case element is
when value-1--
Statement A;
A B C when value-2---
Statement B;
when value-3---
statement C;
end case;

FIGURE 31 Program Edges as Represented in a Digraph

68
 TCAT for Java/Windows User’s Guide

6.3 DiGraph Main Window

In order to explore all the options available, open a directed graph of the
example program. In order to do this, you must first instrument the
example application, which is discussed in Sections 4.1, 4.2, and 4.3,
“Using IJava.”
When you have an instrumented executable:
1. From the Programs menu, select the TCAT for Java folder.
2. Select the Digraph icon from the resulting window.
3. Use the File pull down menu and select Open.
You are prompted for the name of the directed graph to view.
4. Find the TicTacToe.dg file under the tcat_db\name\d_graph directory.
You are prompted for the name of the database file.
5. Find the TicTacToe.mdf file under the tcat_db\name directory.
A window pops up listing the available functions.
6. Select TicTacToe::status(int).

A directed graph depicting the main[int,int,char*] function appears.

FIGURE 32 Directed Graph of TicTacToe

69
CHAPTER 6: DiGraph

The following sections discuss the options available in DiGraph. Several

options are discussed in more detail in later sections.

6.3.1 Tool Bar

The options available from this Tool Bar are the frequently used DiGraph
features. When available, they appear highlighted.

FIGURE 33 Tool Bar

Open This button brings up the Open dialog box.
Print This button brings up the Print dialog box.
ZoomIn This button Zooms in magnification factors of the
current open window.
ZoomOut This button Zooms out magnification factors of the
current open window.
Source This button brings up a window which contains the
source code for the currently selected edge.
Help This button brings up a brief description of DiGraph.

70
 TCAT for Java/Windows User’s Guide

6.3.2 File Menu

This menu displays the file management and printing options that are
available in DiGraph.
Open This option brings up the Open dialog box.
Print This option brings up a the Print dialog box.
Print Preview This option displays an image of what will print
when you select the Print option.
Print Setup This option displays a standard Windows printer set-
up dialog box.
Exit To end your DiGraph session, select the Exit option.

71
CHAPTER 6: DiGraph

6.3.3 Zoom Menu

This menu contains two options for scaling the digraph’s display. For
information on setting the zoom scale, see Section 6.6.1, “The Digraph
Options Dialog Box.”
In This option allows you to enlarge a portion of the di-
graph so that you can see it in more detail. There is a
limit to how far you can zoom in, determined by your
computer’s display resolution.
Out This option allows you to see a wider portion of the
digraph at a reduced magnification. Again, limits ap-
ply to how far you can zoom out.

72
 TCAT for Java/Windows User’s Guide

6.3.4 View Menu

This menu provides three options for configuring the digraph’s display.
Source This option allows you to display the source code for
the selected function in the current directed graph.
Tool Bar This toggle allows you to hide the Tool Bar in order to
give your digraph more vertical display space or to
re-display it.
Status Bar This toggle allows you to hide or re-display the status
bar at the bottom of the DiGraph window.

73
CHAPTER 6: DiGraph

6.3.5 Options Menu

This menu provides access to two dialog boxes where you can set global
display options for DiGraph.
Digraph Options This option displays a dialog box allowing you to
choose the characteristics of the nodes and edges dis-
played in the digraph, as well as the increments for
the Zoom In and Zoom Out options.

6.3.6 Window Menu

This menu allows you to manipulate the DiGraph windows using the
Cascade, Tile, and Arrange Icons options, and the Window list box.

6.3.7 Help Menu

The first help option currently offers a brief description of DiGraph. The
second option, About, displays the program’s version number and copy-
right information.

6.3.8 Status Bar

This section of the window (appearing at the bottom left) displays mes-
sages regarding the functionality and operation of the DiGraph options.

74
 TCAT for Java/Windows User’s Guide

6.4 File Menu

This menu is typical of Windows interfaces, and provides access to file-
manipulation options.

6.4.1 Open

FIGURE 34 DiGraph Open Dialog Box

This option brings up a file selection dialog box. Typical of Windows
interfaces, this dialog box allows you to browse the directory tree, and
select files to open.
File Name This box lists the files in the current directory that
match the filter.
Directory This box lists the available directories.
When you have found the desired file, click OK, and the directed graph is
displayed. Cancel closes the dialog box without opening a graph.

75
CHAPTER 6: DiGraph

6.4.2 Print

FIGURE 35 Print Dialog Box in DiGraph

76
 TCAT for Java/Windows User’s Guide

6.5 View Menu

The most critical option on this menu is the View Source option.

6.5.1 Viewing Associated Source Code

FIGURE 36 View Source Option

This option displays the source code for the program depicted in the
digraph. If you click on an edge segment number in the digraph’s main
window, and the View Source option, the source code associated with
that edge is displayed.
The arrow (triangle) symbols on the right-hand side (and bottom, when
appropriate) of the window are scroll bars, which you can use to move
vertically (or horizontally) in this window.

77
CHAPTER 6: DiGraph

6.6 Options Menu

The options available from this menu allow you to configure certain
aspects of the DiGraph display.

6.6.1 The Digraph Options Dialog Box

Characteristics

FIGURE 37 Digraph Options Dialog Box

This dialog box allows you to choose the magnification step used for the
Zoom In and Zoom Out commands, the shape and size of the digraph’s
nodes, and the colors of the digraph’s edges.
Zoom Increment This sets the magnification interval for the Zoom In
and Zoom Out options. The default setting is .1
meaning a 10% reduction or enlargement in scale
each time these buttons are used. To change the set-
ting, move the slider left or right. Each 0.1 represents
10%, so if you slide the rule to .3, for example, the re-
duction and enlargement is 30% each time.
Eccentricity This determines the curvature of the generated dis-
play. The default value is .3; bigger values make the
picture wider, and smaller values narrower.

78
 TCAT for Java/Windows User’s Guide

Node You can choose different sizes and shapes for the di-
graph’s nodes. In this window, you can change the
space between nodes and their height-to-width ratio.
You have four choices for shapes: Circle, Box, Oval or
Outlined (the circle is drawn but not filled). The de-
fault setting is Circle.
• You can choose the size of the circle, box or
oval. The default size is 1.0.
• You can change the amount of space
between nodes. The default setting is 1.0.
• You can change the height-to-width ratio (for
ovals or box shapes only). The default set-
ting is 1.0.

79
CHAPTER 6: DiGraph

Edge This area provides options to change the appearance

of edges on your directed graph.
• There are three choices for Unhighlighted
Edge: Fulltone, Halftone (dashes) or Blank
(no visible lines). The default setting is Full-
tone.
• Default Color is the basic color of the
digraph’s edges and nodes. The default set-
ting is blue.
OK If you click on the OK button, all the current settings
in the Options window are applied to the digraph.
Cancel If you click on the Cancel button, any changes you
have made since opening the Options window are
discarded.
Close If you click on the Close button, you exit the Options
window.

80
 TCAT for Java/Windows User’s Guide

6.7 Window Menu

This menu provides four options to manipulate the DiGraph windows.
The default arrangement is that the active window entirely overlaps all
others.

6.7.1 Cascade
This option arranges your windows in a cascade, with the active window
top-most and highlighted.

FIGURE 38 Cascading Windows in DiGraph

81
CHAPTER 6: DiGraph

6.7.2 Tile
This option arranges the windows so that a portion of each window is
displayed. The active window is highlighted.

FIGURE 39 Tiled Windows in DiGraph

82
 TCAT for Java/Windows User’s Guide

6.7.3 Arrange Icons

When you have minimized windows, this option arranges them neatly at
the bottom of the DiGraph window.

6.7.4 Window List Box

This area of the pull down menu lists all the open windows available in
DiGraph. The active window is indicated by a check mark. To activate a
new window, especially if the windows are fully overlapping, select it
from this list.

83
CHAPTER 6: DiGraph

84
CHAPTER 7

CallTree

This chapter provides details about using calltrees in TCAT for Java/Windows.

7.1 Calltree Overview

A calltree displays a program’s caller–callee dependency structure. TCAT
for Java/Windows generates a calltree graph for each segment of your
executable during instrumentation and stores it in a separate archive file.
Once the instrumented application has been exercised, you can display a
calltree window for a specified program segment by opening the target
application’s *.cg file.

85
CHAPTER 7: CallTree

7.2 Generating and Viewing Calltrees

You generate calltrees for your application by instrumenting your source-
code files, as described in Sections 4.2 and 4.3 .
To Launch CallTree:
1. Select the TCAT for Java Program Group.
2. Click CallTree.

To View a calltree of the example program:

1. Pull down the File menu.
2. Select Open.

You are prompted for the name of the calltree to view.

3. Find the EXAMPLE.cg file under the tcat_db\name\c_graph directory.
You are prompted for the name of the database file.
4. Find the Prj_Name.mdf file under the tcat_db\name directory.
5. Select a function ID from the presented list.
A calltree depicting the selected function appears. This first node of the
calltree is called the root, as it is never called from within the program.
The second (and lower) tier of nodes are the called functions, as they are
called by nodes above them. The final tier of a calltree consists of called
functions which never call other functions.

86
 TCAT for Java/Windows User’s Guide

7.3 Calltree File Format

For information on the format of calltree files, see Appendix A, “Java
Instrumentor Engine Database Files.”
7.4 CallTree Window Overview

FIGURE 40 CallTree Main Window

This window allows you to view the calltree. This section briefly
describes the menus available from CallTree. Several of the menus are
discussed in more detail in later sections.

87
CHAPTER 7: CallTree

7.4.1 Tool Bar

The options available from this Tool Bar are the frequently-used CallTree
features. When unavailable, they appear grayed out.

FIGURE 41 Tool Bar

Open This button brings up the Open dialog box.
Print This button brings up the Print dialog box.
ZoomIn This button Zooms in magnification factors of the
current open window.
ZoomOut This button Zooms out magnification factors of the
current open window.
Source This button brings up a window which contains the
source code for the currently selected edge.
Digraph This button brings up a digraph of the associated
function.
Help This button brings up a brief description of CallTree.

88
 TCAT for Java/Windows User’s Guide

7.4.2 File Menu

This menu displays the file management options available for CallTree.
Open This option calls up the Open dialog box.
Close This option closes the currently selected calltree.
Exit If you wish to end your CallTree session, drag the
mouse to Exit.
Print This option brings up a the Print dialog box.
Print Preview This option displays an image of what prints when
you select the Print option.
Print Setup This option displays a standard Windows printer set-
up dialog box.

89
CHAPTER 7: CallTree

7.4.3 View Menu

This menu provides three options (Select Function, Source and Directed
Graph) allowing alternate views of the program segment displayed in the
calltree.

7.4.4 Window Menu

This menu allows you to manipulate any open CallTree windows using
the Cascade, Tile and Arrange Icons options and the Window list box.

7.4.5 Options Menu

In this menu, a dialog box pops up where you can set the size, aspect
ratio, and vertical spacing of the calltree, as well as the increments for the
Zoom In and Zoom Out options.

7.4.6 Help Menu

This menu currently offers only one option, About, which displays the
program’s version number and copyright information.

7.4.7 Status Bar

This section of the window (appearing at the bottom left) displays
messages regarding the functionality and operation of the CallTree.

90
 TCAT for Java/Windows User’s Guide

7.5 File Menu

The File menu is typical of Windows applications.

7.5.1 Open

FIGURE 42 CallTree Open Dialog Box

This option brings up a file selection dialog box. It allows you to browse
the directory tree and select files to open.
File Name This box lists the files in the current directory that
match the filter.
Directory This box lists the available directories. When you
have found the desired file, click OK, and the calltree
is displayed.
Cancel closes the dialog box without opening a calltree.

91
CHAPTER 7: CallTree

7.5.2 Print Menu

FIGURE 43 Print Dialog Box in CallTree

92
 TCAT for Java/Windows User’s Guide

7.6 View Menu

From CallTree, you can view source code and directed graphs of your
program using the options on this menu.

7.6.1 Viewing Associated Source Code

FIGURE 44 View Source Option

This option displays the source code for the program depicted in the call-
tree. If you click on an edge segment in the calltree’s main window, and
select the View Source option, the source code associated with that edge
is displayed. If no call pair was selected, the display is positioned at the
first call pair in the module. You can also select the Source button on the
Tool Bar.
The arrow (triangle) symbols on the right-hand side and bottom of the
window are scroll bars, which you can use to move vertically or horizon-
tally in this window.

93
CHAPTER 7: CallTree

7.6.2 Viewing a Directed Graph

FIGURE 45 Directed Graph Option

This option allows you to view the detailed structure of a function in the
current calltree. If you click on a node and select the Directed Graph
option, a directed graph depicting that node appears. You can also select
the Directed Graph button on the Tool Bar.
From this new window, you can view the source code in terms of edges
and nodes rather than call pairs. To do so, click on an element of the
directed graph and select View Source either from the View menu or
from the Tool Bar.

94
 TCAT for Java/Windows User’s Guide

7.7 Window Menu

This menu provides four options used to manipulate the CallTree
windows.The default arrangement is that the active window entirely
overlaps all others.

7.7.1 Cascade
This option arranges your windows in a cascade, with the active window
top-most and highlighted.

FIGURE 46 Cascading Windows in CallTree

95
CHAPTER 7: CallTree

7.7.2 Tile
This option arranges the windows so that a portion of each window is
displayed. The active window is highlighted.

FIGURE 47 Tiled Windows in CallTree

96
 TCAT for Java/Windows User’s Guide

7.7.3 Arrange Icons

When you have minimized windows, this option arranges them neatly at
the bottom of the CallTree window.

7.7.4 Window List Box

This area of the pull-down menu lists all the open windows available in
CallTree. The active window is indicated by a check mark. To activate a
new window, especially if the windows are fully overlapping, select it
from this list.

97
CHAPTER 7: CallTree

7.8 Options Menu

This menu brings up a dialog box from which several display options are
available.

FIGURE 48 CallTree Options Dialog Box

This dialog box allows you to choose the magnification step used for the
Zoom In and Zoom Out commands, the shape and size of the digraph’s
nodes, and the colors of the digraph’s edges.
Zoom Increment This sets the magnification interval for the Zoom In
and Zoom Out options. The default setting is .1
meaning a 10% reduction or enlargement in scale
each time these buttons are used. To change the set-
ting, move the slider left or right. Each 0.1 represents
10%, so if you slide the rule to .3, for example, the re-
duction and enlargement is 30% each time.
Vertical Spacing This alters the vertical distance between members of
callpairs.
Aspect Ratio This alters the distance between and the width of the
boxes.

OK If you click on the OK button, all the current settings

in the Options window are applied to the calltree.
Cancel If you click on the Cancel button, any changes you
have made since opening the Options window are
discarded.

98
APPENDIX A

Java Instrumentor Engine

Database Files
This file lists examples of WinIC9’s output files. This appendix applies to all editions of
Coverage for Windows.

A.1 Instrumentation Database Definitions

This section outlines the files that are used in the instrumentation data-
base stored in the tcat_db directory. This information is used throughout
Coverage for Windows.

99
APPENDIX A: Java Instrumentor Engine Database Files

A.1.1 d_graph Files

The digraphs for each function are put into files which are named with
the same basename as the file from which they originated, with any file-
name suffix stripped off.
The format of each d_graph file is a set of blank delimited (white space
delimited) lines composed as follows:

tail head edge fun_id type filename

lbeg lend byte_beg byte_end string
result [byte1 byte2]
where the fields have the following meanings:
tail The tail node number (string)
head The head node number (string)
edge The IJava assigned edge number (string), also known
as the seg ID
fun_id The number of the function, whose name is found in
the mdf file
type The type of statement which gave rise to the edge
filename The filename where the original text of the program
was found
lbeg The beginning line number, in the named file, where
the tail node is found
lend The ending line number, in the named file, where the
head node is found
byte_beg The beginning byte number, in the named file, where
the tail node is found
byte_end The ending byte number, in the named file, where the
head node is found
string The text string associated with the logical expression
that headed the segment
result The result corresponding to this edge, e.g. T or F or 36
(for switch outcome)
[byte1 byte2] Currently “0 0”; reserved for expansion
A sample d_graph file is listed in Section A.2.1

100
 TCAT for Java/Windows User’s Guide

A.1.2 c_graph Files

The calltrees for each processed file are put into files which are named
with the same basename as the file from which they originated, with any
filename suffix stripped off.
The format of each c_graph file is as a set of blank delimited (white space
delimited) lines composed as follows:
file.caller callee callpair_id module_id
source_file line 0 0 Segment_id
where the fields have the following meanings:
file.caller The file name (given as a prefix up to the rightmost
“.” in the token, and the name of the calling function
(the “caller”)
callee The name of the called function
callpair_id The assigned identification number of the call pair
module_id The assigned identification number of the module.
This number points into the mdf file
source_file The name of the source file that gave rise to the call
pair
line The line number of the source file where the call pair
exists
00 These two fields are pre-set to be “0 0”
segment_id (Reserved for future releases)
An example c_graph file is given in Section A.2.2.

101
APPENDIX A: Java Instrumentor Engine Database Files

A.1.3 Module Definition Files (mdf)

The mdf file contains basic information about the location of text frag-
ments for every segment and every call pair in all processed files.
The mdf file has the following format:
project-name #segs #CPs [#rels]
file.name.function_id type #segs #CPs
[#rels]
file.name.function_id type #segs #CPs
[#rels]
file.name.function_id type #segs #CPs
[#rels]
...
where the first line identifies:
project-name This is the name of the “project” from which the data
is taken.
#segs This is the total number of segments in the project.
#CPs This is the total number of call pairs in the project.
The subsequent lines' fields have the following meanings:
file.name This token contains, first, the name of the file in which
the function name was found, and second, after the
rightmost “.”, the name of the function.
function_id This is the unique numeric identifier for that function,
as found in the filename, which prefixes the function
name.
type This is the type of function that was processed ac-
cording to the key: 84 = static function; 111 = member
function. Note: These numbers are implementation
specific. Additional function types and different
codes will be added in the future. At present this
function type information is not used.
#segs This is the number of segments in the function.
#CPs This is the number of call pairs in the function.
An example mdf file is given in Section A.2.3.

102
 TCAT for Java/Windows User’s Guide

A.1.4 Trace Files and Archive Files

The format described is the Type 3.0 variation that produces trace files
that are “self describing'' and need no other files to be processed correctly.
The assumption is that the assignment of numbers to modules is done by
a runtime lookup of each module's name.
The format for an Archive File is identical except that the records are
arranged in the “natural” order.
The trace file format is universal for all types of runtimes used and for
either trace files or archive files. The record definitions have the following
meanings:
#Format number Trace file Format Type Record
Defines the type of the current trace file. This line
MUST appear as the first line of the trace file:
#Format 3.0
If it does not then this trace file is assumed to be one
using a prior set of definitions.
# comment Comment Line Record
The entire line is treated as a comment. Any blank
line in the trace file is ignored. Tabs and extra spaces
are treated as singleton blanks (i.e. as white space).
The trace file line can be any length (subject to system
constraints).
@ date Creation Date Record
This is the time and date stamp for the trace file, out-
put taken from date.
p filename F X The Project
The first argument is project name. The first number
represents the number of functions.

103
APPENDIX A: Java Instrumentor Engine Database Files

n"M N nsegments Module Definition Record. The module name M has

been entered, and it has been assigned run-time iden-
tification number N for the duration of this trace file.
The module has nsegments segments and ncallpairs
call pairs. The function name is listed with the path-
name and file name preceding it.
This line is written out only the first time that the
module was executed in the current test. (Second in-
stances of this record can be ignored by the coverage
analyzer.)
c “N M [ntimes]” Call Pair Hit Record
Call pair M in module N has been hit [ntimes times].
This record is used to support S1 coverage measure-
ments.
In an archive file the ntimes show the total number of
times this call pair was hit. If a call pair was not hit,
the record need not appear for that segment.
s “N M [ntimes]” Logical Segment Hit Record. Segment M in module N
has been hit [ntimes times]. This record is used to sup-
port C1 coverage measurements, and also is used to
support S0 coverage measurements.
In an archive file the ntimes show the total number of
times this segment was hit. If a segment was not hit
the record need not appear for that segment.
A sample trace file is listed in Section A.2.4.

104
APPENDIX B

Example Instrumentation Database Files

Here are some examples of database files:

A.2.1 Id_graph File

This is a typical d_graph file:

0 1 1 0 0 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 82 0 0 0 (1) 0 0 0
1 2 2 0 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 82 0 0 0 <DONE 0 0 0
2 1 4 0 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 83 0 0 0 ((i&pos)==pos) 0 0 0
2 1 5 0 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 86 0 0 0 ((i&pos)==pos) 0 0 0
1 3 3 0 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 87 0 0 0 <DONE 0 0 0
0 1 1 1 0 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 109 0 0 0 (1) 0 0 0
1 2 2 1 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 112 0 0 0 <9 0 0 0
2 3 4 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 114 0 0 0 (((white&(1<<mw))==0)&&((black&(1<<mw))==0)) 0 0 0
3 4 6 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 116 0 0 0 (won[pw]) 0 0 0
3 4 7 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 120 0 0 0 (won[pw]) 0 0 0
4 5 8 1 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 120 0 0 0 <9 0 0 0
5 6 10 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 121 0 0 0 (((pw&(1<<mb))==0)&&((black&(1<<mb))==0)) 0 0 0
6 4 12 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 123 0 0 0 (won[pb]) 0 0 0
6 4 13 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 127 0 0 0 (won[pb]) 0 0 0
5 4 11 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 128 0 0 0 (((pw&(1<<mb))==0)&&((black&(1<<mb))==0)) 0 0 0
4 7 9 1 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 130 0 0 0 <9 0 0 0
7 1 14 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 130 0 0 0 (bestmove==-1) 0 0 0
7 1 15 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 133 0 0 0 (bestmove==-1) 0 0 0
2 1 5 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 134 0 0 0 (((white&(1<<mw))==0)&&((black&(1<<mw))==0)) 0 0 0
1 8 3 1 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 135 0 0 0 <9 0 0 0
8 9 16 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 135 0 0 0 (bestmove!=-1) 0 0 0
8 9 17 1 1 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 140 0 0 0 (bestmove!=-1) 0 0 0
9 10 18 1 2 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 140 0 0 0 <9 0 0 0
 B.1.1 c_graph Files
This is a typical c_graph file:
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::myMove(boolean) bestMove(int,int,int) 1 3 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 174 0 0 3
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) status(int,TicTacToe&) 1 7 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 249 0 0 1
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) yourMove(boolean,TicTacToe&,int) 2 7 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 267 0 0 6
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) status(int,TicTacToe&) 3 7 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 270 0 0 7
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) myMove(boolean,TicTacToe&) 4 7 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 280 0 0 12
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) status(int,TicTacToe&) 5 7 C:\JAVA\TEST\TICTAC~1\TicTacToe.java 282 0 0 13

B.1.2 mdf Files

This is a typical mdf file:

C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::isWon(void) 0 111 5 0
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::bestMove(int,int) 1 111 21 0
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::yourMove(boolean,int) 2 111 5 0
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::myMove(boolean) 3 111 3 1
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::status(int) 4 111 7 0
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::init(void) 5 111 1 0
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::paint(void) 6 111 9 0
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) 7 111 18 5
C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::getAppletInfo(int) 8 111 1 0
 B.1.3 Trace Files and Archive Files
This is a typical trace file or archive file:

#Format 3.0
# TCAT for Java (tm) Version 1.2
# Copyright 1997 by Software Research, Inc.
# 11-18-1997
# Profile for project 'Prj_Name':
p Prj_Name 9 8
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::isWon(void) 0 111 5 0
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::bestMove(int,int) 1 111 21 0
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::yourMove(boolean,int) 2 111 5 0
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::myMove(boolean) 3 111 3 1
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::status(int) 4 111 7 0
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::init(void) 5 111 1 0
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::paint(void) 6 111 9 0
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::mouseUp(boolean) 7 111 18 5
n C:\JAVA\TEST\TICTAC~1\TicTacToe.TicTacToe::getAppletInfo(int) 8 111 1 0
# End of profile for project 'Prj_Name'.
s 0 1 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 4 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
s 0 2 1
s 0 5 1
APPENDIX C

cover9 —TCAT for Java’s Coverage

Analyzer
This notes explains options for invoking and customizing the “cover9” coverage analyzer.
This notes applies to all editions of TCAT C/C+ and TCAT for Java.

These are the options on how to invoke cover9. This command, used inside the
TCAT for Java graphical user interface, is used to produce a coverage report which,
optionally, can report results in a Reference Listing. The Reference Listing report allows
you to look up a segment in order to identify the actual unexecuted code, and plan new
test cases.
C.1 Command Line Invocation
The complete syntax for calls to cover9 is listed below. Items enclosed in
[brackets] are to be included zero or more times.

cover9 [tracefile [tracefile]]

[-a old-archive]
[-b file]
[-c]
[-C1]
[-d name [name]]
[-DI deinst-file]
[-DL]
[-f new-archive]
[-h | -h name [name]]
[-html | -html filename]
[-H]
[-N]
[-n]
[-nl namefile]
[-NH]
[-m]
[-l | -l name]
[-p]

108
APPENDIX C: cover9 —TCAT for Java’s Coverage Analyzer

[-P0]
[-P1]
[-q]
[-r report]
[-S0]
[-S1]
[-s]
[-SU]
[-T [threshold]]
[-w width]]
C.2 Cover9 Switch Definitions
The options may be used to vary the processing and reports generated by
cover9. The options are listed in alphabetical order.
[tracefile [tracefile]] These are the names of the trace files that you wish to
process. If there are no trace files then cover9 looks for
data in the default trace file name Trace.trc.
If there are no names given, and Trace.trc is not
present then an error message is issued.
If there are multiple trace files, each trace file is pro-
cessed in the order presented.
Caution: The list of trace files must be the first set of argu-
ments. The list is ended by the first symbol that appears
with a '-', i.e. by the first optional switch.
-a old-archive Old Archive File Name Switch.You can include data
from an old archive file in your reports. On the stan-
dard cumulative coverage report, this data will be in-
cluded in the “Cumulative Summary” test results,
but not under the column “Test”. To test iteratively,
progressing through a structured series of tests to-
wards higher C1 values, each run of cover should in-
clude the cumulative archive file from the previous
test.
If you do not include an archive file, the “Cumulative
Summary” figures will be the same as those for
“Test”. Alternatively, if no -a option is given, the file
Archive is used by default.
The -a option interacts with the other report options
discussed below.

109
 TCAT for Java’s Coverage Analyzer

-b file Banner File Name Switch. This allows you to include

specific text, taken from the first line of the file named
title as a title for your reports. A maximum of 80 char-
acters is allowed for titles.
-c Cumulative Report Switch. This option prints the Cu-
mulative report only.
-C1 Branch Coverage Reporting Switch. Turns on reporting
of C1 or branch coverage.
Note: Unless at least one of -C1, -S1, or -S0 is turned on,
no coverage report will be generated.
-d name Module Name Delete Switch. If this switch is present
then the named modules, if found in the current exe-
cution, are deleted from the generated Archive file.
Subsequently, cover9 will never have heard about
these names. This switch is useful in updating an ex-
tensive test record that would otherwise be lost due
to the complexity of editing the Archive file.
-DI deinst-file De-instrument Switch. Allows the user to specify a list
of modules that are to be excluded from coverage re-
porting. Only the list of module names found in the
specified deinst-file is to be excluded from cov-
erage reporting. The module names can be specified
in any format. White space (such as tabs, spaces) is ig-
nored. deinst-file is also the file where new mod-
ules that pass the coverage threshold value (see the -
T switch) will be written.
-DL De-instrument Module List Switch. Allows the user to
see which modules are excluded from coverage re-
porting. This switch is used along with the -DI
switch. The list of excluded modules is printed at the
end of the coverage report
-f new-archive New Archive File Name Switch. Newly accumulated
test coverage data will be placed in this file. If you do
not include a different name with this switch, the ac-
cumulated test data will be placed in the default
name Archive.

Caution: Each time you run cover9, you will write over the contents of
the Archive file unless you use the -f switch to direct the Archive file to
another place. You may wish to remove the filename before starting a
new test sequence.

110
APPENDIX C: cover9 —TCAT for Java’s Coverage Analyzer

-h | -h [name] Linear Histogram Report Switch (-h).

-html [filename] HTML Switch. If present, the current coverage report
in html format will be generated. Normally the report
is written to the file Coverage.htm (the default name),
but you can rename the file with this switch. CAU-
TION: You will overwrite any file you name with this
switch.
-l | -l [name] Logarithmic Histogram Report Switch (-l).
These two options produce two “histogram” reports
that graph the frequency distribution of the segments
exercised in a single module. The histograms provide
a module-by-module analysis of testing coverage,
combining current trace file data with archive date in-
cluded through the -a option or using the default Ar-
chive file. If the optional name argument is present,
then the corresponding histogram for only the named
module is produced; otherwise, cover9 produces his-
tograms for all modules found. There can be multiple
names in the argument if you want histograms of sev-
eral modules. Also, the names can be mixed between
linear and logarithmic histograms.
-H Hit Report Switch. Lists the segments that have been
hit one or more times in current or past tests. This re-
port analyzes the cumulative effect of the current
trace file and any archive data included through the
use of the -a option or using the default Archive file.
-m Minimal Output Switch. When present, cover9 sup-
presses banner information, list of current options
and trace file descriptions. The coverage report con-
tains only the reports requested.
-N, -n Not Hit Report Switch. This option produces the “Not
Hit” report which lists segments that have not been
exercised. This report analyzes the cumulative effect
of the current trace file and any archive data included
through the use of the -a option or using the default
Archive file.
-NH Newly Hit Report Switch. Shows the segments by mod-
ule that were hit in the current execution that were
not hit previously. Thus this gives the user an assess-
ment of the value of the most-recently added test(s).
This shows what the current test “gained”. Output is
the complement of the “Newly Missed” report.

111
 TCAT for Java’s Coverage Analyzer

-nl namefile Name List Switch. This switch specifies that only the
list of module names found in the specified namefile
file is to be reported on in the current coverage report.
Coverage on other module names that may appear in
the archive or supplied trace files are ignored; howev-
er, the data is accumulated in the archive file.
The names used must be specified one name per line.
White space (tabs, spaces, etc.) on the line is ignored.
The following reports are affected by the existence of
a namefile:
•Cumulative Report
•Past Report
•Not Hit Report
•Hit Report
•Newly Hit Report
•Newly Missed Report.
The histogram outputs are not affected. There is a
separate name mechanism that can be used to pro-
duce individual histogram reports.
-NM Newly Missed Report Switch. This option produces the
Newly Missed report. Shows which segments, by
module, hit in any prior test that were not hit in the
current test. This shows what the current test “lost”.
This output is the complement of the Newly Hit re-
port.
-p Past Report Switch. Print only the Past Test report; this
option should be used in conjunction with the -a op-
tion when you want to analyze the overall perfor-
mance of a set of past tests.
-q Quiet Output Switch. Suppress printout of current ver-
sion and release information (this can be used to facil-
itate running cover9 in batch mode).
-r report Coverage Report File Name Switch. Normally the report
is written to the file Coverage (the default name), but
you can rename the file with this switch. CAUTION:
You will overwrite any file you name with this
switch.
-S1 Call-Pair Coverage Switch. If present, the report will
show call pair coverage.

112
APPENDIX C: cover9 —TCAT for Java’s Coverage Analyzer

-S0 Module Coverage Switch. If present, the report will

show module coverage.

NOTE: Unless at least one of -C1, -S1, or -S0 is turned on, no coverage
report will be generated. However, not both -S1 and -S0 can be present; if
they are then only -S1 is assumed.

-s Sort Switch. This option produces output reports with

module names sorted alphabetically.
-SU Suppress Update Switch. During processing, cover9
will suppress updating of the archive file, either the
default Archive or the file named by the -f switch.
cover9 will read the data in the archive file to form the
basis for the “past test” information.
-T threshold Coverage Threshold Switch. Threshold is a real number
that specifies threshold value. Any module with a
coverage percentage greater than or equal to this
threshold value will be written to the de-instrument-
ed file (see the -DI deinst-file switch). If no
threshold is specified, then the default value of 85
percent is assumed.
-w width Report Width Switch. Normally the reports generated
by cover9 are wide enough to accommodate module
names up to 21 characters in length. The internal limit
on name length is, however, 128 characters. You can
use this switch to force cover9 system to generate re-
ports that are wide enough to accommodate the full
128 character module names.
The width factor is the number of additional charac-
ters to be added to the report. The default value is ze-
ro. Maximum width is 128 - 21 = 107. WARNING:
Reports with high values for the -w option may con-
tain long lines and may not be suitable for printing di-
rectly.
C.3 Error Processing
In case there is an error, cover9 gives a response line (usage line) indicat-
ing the set of switches and options. This response is the same as the -help
response.

113
Index

Symbols calltrees 36, 63, 64

CAPBAK 13
.dg file 27, 42 closing TCAT Java/Windows 39
code inspection 6
code language selection 43
A compiling & running 5
cost benefit analysis 10–14
application under test 25, 41, 46 Cover 30, 53
archive file 53, 61 file selection dialog box 58
sample 107 tool bar 55
archive file format 103, 107 Cover9 Release Notes 108
cover9, coverage analyzer 108–113
B command line syntax 108
invoking 108
coverage analysis 6
Bottom-Up 8 tools 1
bottom-up testing 8 coverage data 41
branch (C1) metrics 53 coverage report 5, 29, 53, 54
sample analysis 62–65
C coverage threshold 14

c_graph file 101 D

sample 106
C1 coverage 63 d_graph file 100
C1 metric 3 sample 105
call pair 41 data structures 6
call tree window 85 database file format 51, 53
called functions 86 DiGraph 31, 67–83
callercallee dependency structure 85 file format 67
calling statement 38 file menu 75
call-pair 41 options menu 78–80
callpair 3, 36, 42, 62, 63 print dialog box 76
(S1) metrics 53 tool bar 70
viewing associated source code 64 view menu 77
CallTree 35, 85–98 viewing associated source code 77
options menu 98 window menu 81
viewing source code 38 digraph 63, 65
calltree 63 digraph edges 67

114
INDEX

digraph nodes 67
directed graph sample 106
viewing from Calltree 37 multiple-module testing 8
Directed Graph Listing 27, 42
directed graphs 31, 33, 67 O
dynamic analysis 7
online documentation
E FrameReader 19

error rate prediction 14 P

EXDIFF 13
percent coverage recommended 5
F possible program flow 63
Prj_Name.mdf 27
font prj_Name.mdf 42
italics x
italix x Q
font, bold face x
font, courier x
Quick Start 23–39
function calls 5, 41

H R
reference listing file 5
hardware configuration 15
reliability modeling 14

I S
IJava 43
S0 coverage 45
command line invocation 46
S1 coverage 45, 63
instrumentation 5, 25, 41, 43
S1 metric 3
function names 50
segment 62, 63
instrumenting module(s) 8
viewing associated source code 65
interactive option 45
segments hit 53
modes 43
segments not-hit 53
instrumentor directives 50
setup.exe 16, 17
instrumentor switches 46–49
SMARTS 13
software reliability 2
J source code 42
viewing from DiGraph 34
Java 15 SQA 1, 14
static analysis 6
static properties (of software) 6
L
logical branch 3, 5, 41, 67 T
TCAT for Java/Windows
M closing 39
editing the default path 18
manual analysis 6 uninstall 21
module definition file (mdf) 102, 106 TCAT Java/Windows
program group 20

115
 TCAT Java User’s Guide

TCAT.mdf 86
tcat_db directory 46, 69, 86, 99
test cases 5
testing methods 6
text
"double quotation marks" x
boldface x
italics x
text, boldface x
text, courier x
text, italix x
TicTacToe.cg 27, 35, 42
TicTacToe.cpp 25
TicTacToe.dg 27, 31, 42, 69
TicTacToe.i 27, 42
TicTacToe.mdf 31, 69
top-down testing 8
trace file 5, 26, 41, 53
format 103, 107
sample 107
Trace.trc 29, 42
tutorial 15, 23

V
variable type rules 6

W
Windows 95 16, 39
Windows Explorer 16
Windows NT 39
WinIJava 25, 26, 41, 42, 43, 50

116