Atomicity violation is a common kind of concurrency bug in real world projects. PENELOPE is a framework that can detect atomicity violations in a single observed trace, without explicitly examining all traces that result from every possible interleaving. This thesis proposes a way to improve performance of PENELOPE’s prediction stage by performing computation directly on grammar compressed execution trace file, leading to a running time linear in the length of compressed file and size of grammar.
【 预 览 】
附件列表
Files
Size
Format
View
Adaptation of grammar based compression in detecting atomicity violations