IDENTIFYING SEMANTIC DIFFERENCES IN ASPECTJ PROGRAMS Martin Görg and Jianjun Zhao Computer Science Department, Shanghai Jiao Tong University

Outline  Motivation and Background  Difference Analysis Algorithm  Evaluation of Quality and Feasibility  Conclusions

Motivation and Background

Static Semantic Difference Analysis  static: source code analysis at compile time  semantic: differences in behavior PP’ modified

Why solve the Problem?  Motivation  Reduce testing costs  Produce correct software  Possible applications  Debugging support  Regression test selection  Program slicing

AOP and AspectJ  AOP encapsulates crosscutting concerns  AspectJ  implementation of AOP  extension to Java public class C { int i; void m1() { } … } Base Code aspect A { double C.d; before() : … { } …. } Aspect Code Introduce Advise

AspectJ Example 1 aspect Constraints { 2 public boolean Shape.immovable = false; 2 void around(Shape s) : execution (public Shape+.set*(..)) && target(s) { 3 if (!s.immovable) {proceed( ) ; } } } ITD around advice

Hammocks  Single entry  Single exit  For any directed graph S E

Motivational Example 1 public class Point extends Shape { 2 private int x, y; 3 public void setX(int i){ 4 x = i; } 5 public void setY(int i){ 6 y = i; } 1 public class Point extends Shape { 2 private int x, y; 3 public void setX(int i){ 4 x = i; } 5 void setY(int i){ 6 y = i; } a change in visibility alters program execution

for AspectJ Programs Difference Analysis Algorithm

Algorithm Outline 1. For every module in P find a matching module in P’ (module-level matching) 2. Build extended CFGs for all modules in P and P’ and identify hammocks 3. Perform node-by-node comparison on every pair of hammock graphs (node-level matching)

Matching at Module Level  Signature matching  Disjunctive matching  Obtain best match from multiple candidates public void p1.C1.add(int, Object) P:P: boolean p1.C1.add(int, Object) P’: public void p1.C1.add(int, Object) P:P: boolean p1.C1.add(double, Object) P’:

1. Matching at Module Level  Problem: Not every AspectJ construct has a signature (most importantly: advices)  Solution: a)Define new AspectJ signatures (e.g. [strictfp] before (Formals) : [throws TypeList] : Pointcut {Body} ) b)Define disjunctive patterns for these signatures

2. Build CFGs and Hammocks

3. Matching at Node Level  Simultaneous graph traversal  Node-by-node comparison  Recursive  Two user inputs  Similarity threshold (S)  Maximum lookahead (LH)

3. Matching at Node Level X E Y H P X E Y H P’ S UV e P S UV’ e P’ Similarity Threshold S = 0.5; Lookahead LH = 1

Quality and Feasibility Tests and Evaluation

ProgramLOCDiffsAffectedMatchedErrors ants (100%)0 bean (100%)0 cona-stack (100%)0 dcm (100%)0 figure (100%)0 introduction (98.3%)6 nullcheck (98.2%)78 quicksort (100%)0 spacewar (100%)0 tracing (100%)0  Signature definitions and disjunctive matching  Minimal change with maximal effect  Deficits: some combinations and swapped statements

 S < 0.6: LH has only minor impact  LH  20: within one minute  S  0.6: slow for LH > 20, but not needed

New Findings and Open Tasks Conclusions

What we did  New signatures for AspectJ modules  Disjunctive matching  for AspectJ and Java modules  a solution for modified signatures  Application of hammock algorithm  from OO to AO  Evaluations using a tool implementation

Conclusions  Disjunctive matching is a good idea  modules are correctly matched  more work for node-level matching  eliminates work of type-level matching  replaces user interaction  Type-level matching is not required  Hammock graph matching can be applied given: a) correctly matched modules b) appropriately modeled and labeled CFGs

Future Work  Improve disjunctive matching patterns  Extend CFG representations  Solve the swapping problem  Handle dynamic pointcuts

Questions? Thank You for listening