a differencing algorithm for object oriented programs
play

A Differencing Algorithm for Object-Oriented Programs Taweesup - PowerPoint PPT Presentation

A Differencing Algorithm for Object-Oriented Programs Taweesup (Term) Apiwattanapong Alessandro Orso Mary Jean Harrold College of Computing Georgia Institute of Technology National Science Foundation awards CCR-0306372, CCR-0205422,


  1. A Differencing Algorithm for Object-Oriented Programs Taweesup (Term) Apiwattanapong Alessandro Orso Mary Jean Harrold College of Computing Georgia Institute of Technology National Science Foundation awards CCR-0306372, CCR-0205422, CCR-9988294, CCR-0209322, and SBE-0123532 to Georgia Tech ASE 2004 Sep 22, 2004

  2. Example A Exception A Exception 2c2 float n double n < float n; void m1() void m1() --- E1 > double n; B B E1 E2 void m1() E2 6a7,9 void m2() void m2() > public void m1() { void m3(A a) void m3(A a) > ... > } n = 1/n; n = 1/n; int i=0; 9a13 a.m1(); a.m1(); > int i=0; Original Modified try { try { Version Version … … 21c25 (P) (P’) } } < public class E2 extends E1 { catch (E1 e) { … } catch (E1 e) { … } --- catch (Exception e) { … } catch (Exception e) { … } > public class E2 extends Exception { ASE 2004 Sep 22, 2004

  3. Outline • Introduction • Differencing Algorithm • Representation • Matching • Empirical Studies • Related Work • Conclusions ASE 2004 Sep 22, 2004

  4. Overview of Differencing Algorithm Phase 1 Phase 2 Phase 3 New New New classes and methods statements interfaces P Matched Match statements Match Matched Matched Match Match class and 1. create Differencing classes and method statement methods statements interface Graphs (DiGs) interfaces pairs pairs pairs 2. compare statements P’ Deleted Deleted Deleted classes and methods statements interfaces ASE 2004 Sep 22, 2004

  5. DiG: Extend CFG Entry n_float = 1/ n_float n = 1/n A Exception call a.m1() float n E1 void m1() E2 B void m2() n = 1/n; catch Exception:E1, void m3(A a) catch E1,E2 catch E1 Exception:E1:E2 a.m1(); try { … … catch Exception } … catch (E1 e) { … } catch (Exception e) { … } exit ASE 2004 Sep 22, 2004

  6. DiG: Extend CFG Type in scalar variables’ names Entry Dynamic Dispatch n_float = 1/ n_float Exception Handling Globally-qualified names A Exception call a.m1() float n A B E1 void m1() A.m1() A.m1() E2 B return void m2() n = 1/n; catch Exception:E1, try EX void m3(A a) Exception:E1:E2 a.m1(); EX try { … … catch Exception } … catch (E1 e) { … } catch (Exception e) { … } exit ASE 2004 Sep 22, 2004

  7. DiG: Simplify the extended CFG G (DiG for Entry m3 in P) n_float = 1/ n_float call a.m1() A B A.m1() A.m1() return catch Exception:E1, try EX Exception:E1:E2 EX … catch Exception … exit ASE 2004 Sep 22, 2004

  8. DiG: Simplify the extended CFG G (DiG for Entry m3 in P) n_float = 1/ n_float call a.m1() A B HM1 A.m1() A.m1() return catch Exception:E1, try EX Exception:E1:E2 EX HM2 … catch Exception … exit ASE 2004 Sep 22, 2004

  9. DiG: Simplify the extended CFG G (DiG for G’ (DiG for Entry Entry m3 in P) m3 in P’) n_float = 1/ n_float n_double = 1/ n_double int i_int=0; call a.m1() call a.m1() A B A B HM1 HM3 A.m1() A.m1() A.m1() B.m1() return return catch Exception:E1, try EX try EX catch Exception:E1 Exception:E1:E2 EX EX HM2 HM4 … catch Exception, catch Exception … Exception:E2 … … exit exit ASE 2004 Sep 22, 2004

  10. Matching G G’ Entry Entry Look-ahead unchanged limit: 1 n_float = 1/ n_float n_double = 1/ n_double modified Similarity unchanged int i_int=0; threshold: 0.5 call a.m1() unchanged call a.m1() A B A B A.m1() A.m1() A.m1() B.m1() modified return return 4 unchanged matched (.67) catch Exception:E1, try EX 6 compared try EX catch Exception:E1 Exception:E1:E2 EX EX … … catch Exception, catch Exception Exception:E2 … … exit exit ASE 2004 Sep 22, 2004

  11. Matching public class A { double n; Entry public void m1() { ... } } n_double = 1/ n_double public class B extends A { public void m1() { … int i_int=0; } public void m2() { ... } call a.m1() public void m3() { A B n = 1/n; A.m1() B.m1() int i = 0; a.m1(); try { return ... } try EX catch Exception:E1 catch (E1 e) { ... } catch (Exception e) { ... } EX } … catch Exception, public class E1 extends Exception { Exception:E2 … … } public class E2 extends Exception { exit … ASE 2004 Sep 22, 2004 }

  12. Outline • Introduction • Differencing Algorithm • Representation • Matching • Empirical Studies • Related Work • Conclusions ASE 2004 Sep 22, 2004

  13. Empirical Studies Experimental Setup • JDiff: A Java implementation of our technique • Subject : Jaba • A Java bytecode analysis tool • 60KLOC (550 classes, 2800 methods) • 2 sets of 4 consecutive versions – Low activity: v1, …, v4 (3-20 changes) – High activity: va, …, vd (15-150 changes) Studies 1. Efficiency of our algorithm 2. Effectiveness of our algorithm in matching 3. Effectiveness of our algorithm for a maintenance task ASE 2004 Sep 22, 2004

  14. Study 1: Efficiency of Our Algorithm Goal: Measure the efficiency of our algorithm for various look-ahead limits and hammock similarity thresholds Method: 1. Run JDiff • Low-activity versions (v1-v2, v1-v3, and v1-v4) • Various look-ahead limits (0-50) • Various similarity thresholds (0-1) 2. Collect the running times. ASE 2004 Sep 22, 2004

  15. Study 1: Efficiency of Our Algorithm 450 400 v1-v2 (S>0.2) running time (sec) v1-v3 (S>0.2) v1-v4 (S>0.2) 350 v1-v2 (S=0) v1-v3 (S=0) 300 v1-v4 (S=0) 250 200 0 10 20 30 40 50 60 look-ahead limit ASE 2004 Sep 22, 2004

  16. Study 3: Effectiveness for a Maintenance Task Goal: Assess the effectiveness of our algorithm for a maintenance task (coverage estimation) Method: • Jaba’s regression test suite (~60% coverage) • Both low- and high-activity versions (v1-v2, v1-v3, v1-v4, va-vb, va-vc, and va-vd) • For each pair (vi-vj), 1. Collect coverage for vi 2. Run JDiff on vi-vj to get mappings 3. Get estimated coverage of vj based on mappings 4. Collect actual coverage for vj 5. Compare actual and estimated coverage of vj ASE 2004 Sep 22, 2004

  17. Study 3: Effectiveness for a Maintenance Task Pair Avg. Correctly vi,vj estimated for vj (%) v1,v2 98.57 Low-activity v1,v3 98.46 period v1,v4 98.03 va,vb 96.25 High-activity va,vc 86.08 period va,vd 84.70 ASE 2004 Sep 22, 2004

  18. Outline • Introduction • Differencing Algorithm • Representation • Matching • Empirical Studies • Related Work • Conclusions ASE 2004 Sep 22, 2004

  19. Related Work Textual • E.W. Myers. Algorithmica 1986 (UNIX diff) Control-flow graph based • J. Laski and W. Szermer. ICSM 1992 • Z. Wang, K. Pierce, and S. McFarling. JILP 2000 (BMAT) Dependence graph based • S. Horwitz. PLDI 1990 • D. Binkley. ICSM 1992 Abstract syntax tree based • Raghavan et al. ICSM 2004 (Dex) • Ren et al. Technical Report 2004 (Chianti) Input-output dependence based • D. Jackson. ICSM 1994 (Semantic diff) ASE 2004 Sep 22, 2004

  20. Conclusions Contributions • A differencing algorithm that • Based on a new graph representation which models object- oriented features • Uses several strategies to increase matching capability • A tool that implements our technique (JDiff) • A set of studies that show the efficiency and effectiveness of the approach Future Directions • To improve matching results • Investigate additional heuristics • Use common change patterns • Test-suite augmentation • Create new test cases based on changes in the program ASE 2004 Sep 22, 2004

  21. Questions? ASE 2004 Sep 22, 2004

Recommend


More recommend