2016 International Conference on Frontiers of Information Technology (FIT) (2016)
Dec. 19, 2016 to Dec. 21, 2016
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/FIT.2016.044
Crashing of program is an annoying experience for users. Whenever a program crashes, an event log is generated. Sometimes built in crash reporting programs send crash reports automatically to developing site whereas sometimes, user is presented with an option to report the crash himself. This reporting is often useful for the development team to diagnose and fix the problem. It happens quite often that code that crashes a program is not the reason for the crash itself but crash is due to the use of some faulty function or library of some other program. We propose a method to identify cross platform bug correlation to detect faulty functions, using function call stack information given in bug reports. We collect and process bug reports from multiple platforms to compute similarity based on our similarity metric between occurrence sequences of the function calls within different bug reports. For the solved bug we extract information about faulty function by analyzing its bug report to propose fix of the similar bug with same function within correlated bug report. If there exists fix of one bug in one application, it can be used to resolve similar bug in some other application. Using our technique we found similar bug reports. We were able to find cases for similar bug reports that have same reason for the cause of bug and were using same fix.
Computer bugs, Measurement, Correlation, Data mining, Software, Libraries
M. A. Ghafoor and J. H. Siddiqui, "Cross Platform Bug Correlation Using Stack Traces," 2016 International Conference on Frontiers of Information Technology (FIT), Islamabad, Pakistan, 2016, pp. 199-204.