Code coverage analysis of large-scale applications

Session Type: 
Standard [25 minutes]

Code coverage has become the industry standard of measuring test completeness. On the first sight code coverage tools are well-represented in the Eclipse world, but as it turns out most of them are meant for Java, and support only weak criteria such as function and statement coverage. C/C++ developers willing to use Eclipse and stronger coverage criteria are left without much support.

Our talk will present our own solution for code coverage visualization using condition/decision criterion, including merging and comparison of coverage data from test suites distinguished in their scope and level. The need of such coverage combining and comparison is especially felt in large-scale industrial projects.

The talk aims both C/C++ developers, who want to use stronger coverage criteria than simple statement coverage, and quality assurance managers, who are willing to simplify the minimization and prioritization of their test suites.

The solution is based on several technologies such as Bullseye Coverage for gathering code coverage, EMF for storing and manipulating coverage data and CDT for visualization.

There are no prerequisites for this talk, but a general familiarity with code coverage and interest in improving code quality are desired.