Visualizing massively pruned execution traces to facilitate trace exploration

摘要: Execution traces provide precise pictures of the inner workings software systems. They therefore support programmers in performing various maintenance tasks. However, exploring is difficult due to their size. typically consist thousands participating functions and millions control flow events. When traces, it particularly time-consuming identify those time ranges within trace that are relevant for current task. In this paper, we propose a technique supports first prunes less calls from then provides condensed repetition-aware visualizations facilitate fast accurate navigation even very large traces. Repetitions detected by novel metrics measure similarity between function fuzzy adjustable way. The helps outlier repetitive call sequences guides on paths being likely comprehension implemented prototypical analysis tool copes with C/C++ We demonstrate concepts means case study our industrial partner.