Clang vs. GCC – code coverage checking

Re:Clang vs. GCC – code coverage checking by lantastic on Saturday April 02, 2016 @05:16PM (#11413494)

One use case already pointed out is to check for gaps that your tests have not covered.

Other use case is figuring out what tests to re-execute when you change something - you can skip tests that never touch any code you are changing. Can be a time-saver if you have a big test suite or lack finer-grained tests (component/unit).

[ Parent | on HN ]

Re:Clang vs. GCC – code coverage checking by viraptor on Saturday April 02, 2016 @07:52PM (#11414162)

This sounds a bit dangerous. Code coverage works on code, not on data, so if you're testing something like:

    // globals
    rules_t rules[] = {{1, OK}, {2, FAIL}, {3, THERMONUCLEAR_WAR}};

It sounds like you could change rules 1 and 2 to THERMONUCLEAR_WAR as well and none of the tests would rerun. Is that right?

[ Parent | on HN ]

Re:Clang vs. GCC – code coverage checking by stouset on Saturday April 02, 2016 @08:56PM (#11414385)

You might do something like this for local development, keeping things fast with some confidence you're not grossly breaking things. But you run the full suite through CI.

[ Parent | on HN ]