Impact analysis using Static Execute after in WebKit

Judit Jász, Lajos Schrettner, A. Beszédes, Csaba Osztrogonác, T. Gyimóthy

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

Insufficient propagation of changes causes the majority of regression errors in heavily evolving software systems. Impact analysis of a particular change can help identify those parts of the system that also need to be investigated and potentially propagate the change. A static code analysis technique called Static Execute After can be used to automatically infer such impact sets. The method is safe and comparable in precision to more detailed analyses. At the same time it is significantly more efficient, hence we could apply it to different large industrial systems, including the open source WebKit project. We overview the benefits of the method, its existing implementations, and present our experiences in adapting the method to such a complex project. Finally, using this particular analysis on the WebKit project, we verify whether applying the method we can actually predict the required change propagation and hence reduce regression errors. We report on the properties of the resulting impact sets computed for the change history, and their relationship to the actual fixes required. We looked at actual defects provided by the regression test suite along with their fixes taken from the version control repository, and compared these fixes to the predicted impact sets computed at the changes that caused the failing tests. The results show that the method is applicable for the analysis of the system, and that the impact sets can predict the required changes in a fair amount of cases, but that there are still open issues for the improvement of the method.

Original languageEnglish
Title of host publicationProceedings of the European Conference on Software Maintenance and Reengineering, CSMR
Pages95-104
Number of pages10
DOIs
Publication statusPublished - 2012
Event2012 16th European Conference on Software Maintenance and Reengineering, CSMR 2012 - Szeged, Hungary
Duration: Mar 27 2012Mar 30 2012

Other

Other2012 16th European Conference on Software Maintenance and Reengineering, CSMR 2012
CountryHungary
CitySzeged
Period3/27/123/30/12

Fingerprint

Static analysis
Defects

Keywords

  • Change impact analysis
  • Regression testing
  • Source code analysis
  • Static Execute After

ASJC Scopus subject areas

  • Software

Cite this

Jász, J., Schrettner, L., Beszédes, A., Osztrogonác, C., & Gyimóthy, T. (2012). Impact analysis using Static Execute after in WebKit. In Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR (pp. 95-104). [6178857] https://doi.org/10.1109/CSMR.2012.20

Impact analysis using Static Execute after in WebKit. / Jász, Judit; Schrettner, Lajos; Beszédes, A.; Osztrogonác, Csaba; Gyimóthy, T.

Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR. 2012. p. 95-104 6178857.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Jász, J, Schrettner, L, Beszédes, A, Osztrogonác, C & Gyimóthy, T 2012, Impact analysis using Static Execute after in WebKit. in Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR., 6178857, pp. 95-104, 2012 16th European Conference on Software Maintenance and Reengineering, CSMR 2012, Szeged, Hungary, 3/27/12. https://doi.org/10.1109/CSMR.2012.20
Jász J, Schrettner L, Beszédes A, Osztrogonác C, Gyimóthy T. Impact analysis using Static Execute after in WebKit. In Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR. 2012. p. 95-104. 6178857 https://doi.org/10.1109/CSMR.2012.20
Jász, Judit ; Schrettner, Lajos ; Beszédes, A. ; Osztrogonác, Csaba ; Gyimóthy, T. / Impact analysis using Static Execute after in WebKit. Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR. 2012. pp. 95-104
@inproceedings{f61a0adea4ed4f3697257130c9dd26a7,
title = "Impact analysis using Static Execute after in WebKit",
abstract = "Insufficient propagation of changes causes the majority of regression errors in heavily evolving software systems. Impact analysis of a particular change can help identify those parts of the system that also need to be investigated and potentially propagate the change. A static code analysis technique called Static Execute After can be used to automatically infer such impact sets. The method is safe and comparable in precision to more detailed analyses. At the same time it is significantly more efficient, hence we could apply it to different large industrial systems, including the open source WebKit project. We overview the benefits of the method, its existing implementations, and present our experiences in adapting the method to such a complex project. Finally, using this particular analysis on the WebKit project, we verify whether applying the method we can actually predict the required change propagation and hence reduce regression errors. We report on the properties of the resulting impact sets computed for the change history, and their relationship to the actual fixes required. We looked at actual defects provided by the regression test suite along with their fixes taken from the version control repository, and compared these fixes to the predicted impact sets computed at the changes that caused the failing tests. The results show that the method is applicable for the analysis of the system, and that the impact sets can predict the required changes in a fair amount of cases, but that there are still open issues for the improvement of the method.",
keywords = "Change impact analysis, Regression testing, Source code analysis, Static Execute After",
author = "Judit J{\'a}sz and Lajos Schrettner and A. Besz{\'e}des and Csaba Osztrogon{\'a}c and T. Gyim{\'o}thy",
year = "2012",
doi = "10.1109/CSMR.2012.20",
language = "English",
isbn = "9780769546667",
pages = "95--104",
booktitle = "Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR",

}

TY - GEN

T1 - Impact analysis using Static Execute after in WebKit

AU - Jász, Judit

AU - Schrettner, Lajos

AU - Beszédes, A.

AU - Osztrogonác, Csaba

AU - Gyimóthy, T.

PY - 2012

Y1 - 2012

N2 - Insufficient propagation of changes causes the majority of regression errors in heavily evolving software systems. Impact analysis of a particular change can help identify those parts of the system that also need to be investigated and potentially propagate the change. A static code analysis technique called Static Execute After can be used to automatically infer such impact sets. The method is safe and comparable in precision to more detailed analyses. At the same time it is significantly more efficient, hence we could apply it to different large industrial systems, including the open source WebKit project. We overview the benefits of the method, its existing implementations, and present our experiences in adapting the method to such a complex project. Finally, using this particular analysis on the WebKit project, we verify whether applying the method we can actually predict the required change propagation and hence reduce regression errors. We report on the properties of the resulting impact sets computed for the change history, and their relationship to the actual fixes required. We looked at actual defects provided by the regression test suite along with their fixes taken from the version control repository, and compared these fixes to the predicted impact sets computed at the changes that caused the failing tests. The results show that the method is applicable for the analysis of the system, and that the impact sets can predict the required changes in a fair amount of cases, but that there are still open issues for the improvement of the method.

AB - Insufficient propagation of changes causes the majority of regression errors in heavily evolving software systems. Impact analysis of a particular change can help identify those parts of the system that also need to be investigated and potentially propagate the change. A static code analysis technique called Static Execute After can be used to automatically infer such impact sets. The method is safe and comparable in precision to more detailed analyses. At the same time it is significantly more efficient, hence we could apply it to different large industrial systems, including the open source WebKit project. We overview the benefits of the method, its existing implementations, and present our experiences in adapting the method to such a complex project. Finally, using this particular analysis on the WebKit project, we verify whether applying the method we can actually predict the required change propagation and hence reduce regression errors. We report on the properties of the resulting impact sets computed for the change history, and their relationship to the actual fixes required. We looked at actual defects provided by the regression test suite along with their fixes taken from the version control repository, and compared these fixes to the predicted impact sets computed at the changes that caused the failing tests. The results show that the method is applicable for the analysis of the system, and that the impact sets can predict the required changes in a fair amount of cases, but that there are still open issues for the improvement of the method.

KW - Change impact analysis

KW - Regression testing

KW - Source code analysis

KW - Static Execute After

UR - http://www.scopus.com/inward/record.url?scp=84860508734&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84860508734&partnerID=8YFLogxK

U2 - 10.1109/CSMR.2012.20

DO - 10.1109/CSMR.2012.20

M3 - Conference contribution

AN - SCOPUS:84860508734

SN - 9780769546667

SP - 95

EP - 104

BT - Proceedings of the European Conference on Software Maintenance and Reengineering, CSMR

ER -