Selecting Test and Identifying Test Coverage Requirement for Modified Software

1
Selecting Test and Identifying Test Coverage
Requirement for Modified Software

• Author G.Rothermel and M.J Harrold
• Source International Symposium on Software
  Testing and Analysis Proceedings of the 1994 ACM
  SIGSOFT international symposium on Software
  testing and analysis
• Presented by Chen Hung-Hsiang

2
Progress

• G.Rothermel and M.J.Harrold , "Analyzing
  Regression Test Selection Techniques," IEEE
  Transaction on Software Engineering,August 1996
• Regression selection techniques
• Linear Equation (intra)
• Symbolic execution
• Data Flow
• Modification Based
• Cluster identification
• Graph Walk

• Linear Equation(inter)
• Path analysis
• Program Dependence Graph
• Firewall
• Slicing
• Modified entity

CDG?PDG?SDG
3
Outline

• Introduction
• Background
• PDG
• IntraProcedural Regression Testing
• InterProcedural Regression Testing
• Conclusion

4
Introduction

• Coverage problem
• Coverage techniques use structural coverage
  criteria to select existing tests that exercise
  changed or affected program components.
• Safe technique problem
• Safe techniques, which select every test that
  will cause a modified program to produce
  different output than its original version.
• Author present an approach to regression testing
  that handles two important tasks
• selecting tests from the existing test suite that
  should be rerun
• identifying portions of the code that must be
  covered by tests.

5
Introduction

• Solution
• Deal with individual procedure
• Use Program Dependence Graph
• Deal for groups of procedure or entire programs
• Use System Dependence Graph
• Advantages
• Safe and more precise selects technique
• identify test coverage requirement
• Handles structural and nonstructural program
  modifications, and processes multiple
  modification with a single application of the
  algorithm

6
What is PDG?
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
Background

• Terminology
• Control dependence edge
• cd-predecessor
• cd-successor
• Data dependence edge
• dd-predecessor
• dd-successor

14
Background
dd-predecessor
cd-predecessor
15
IntraProcedural Regression Testing
16
IntraProcedural Regression Testing
17
IntraProcedural Regression Testing
18
IntraProcedural Regression Testing
Form O(n2) to O(n3)
19
IntraProcedural Regression Testing
20
IntraProcedural Regression Testing
R5
both R2,R3 history is T2
21
IntraProcedural Regression Testing
Data dependence edge
22
IntraProcedural Regression Testing

• Find set of affected definition-use pairs in G
• Every du-pair(S1,S2) in P has a corresponding
  data dependence edge (N1,N2) in G.
• (1)N1 is new or modified
• (2)N2 is new or modified
• (3)N1 is control or data dependent on some other
  new or modified node.

23
IntraProcedural Regression Testing
Type 1 and 3
Type 2
24
InterProcedural Regression Testing

• Edges defined
• 1.Control dependence edges
• 2.Intraprocedural data dependence edges
• 3.Parameter control dependence edges
• 4.Parameter-in edges
• 5.Parameter-out edges
• 6.Call edges
• 7.Interprocedural dependence edges
• 8.Transitive dependence edges

25
InterProcedural Regression Testing
26
InterProcedural Regression Testing
27
InterProcedural Regression Testing
28
Conclusion

• Algorithm selects those tests from an existing
  test suite that may exhibit different output in
  the modified program
• It identifies parts of the program that are
  affected by modifications and should be covered.
• The algorithm handles multiple changes, does not
  require prior knowledge of the location of
  changes, and works on both single procedures and
  entire programs.