Hi,
I'm able to parametrize a build with the branch name for checkout. Nevertheless, the build results for all such branch builds are linearly ordered. This means that on the JUnit test result page, the result of build #n is always compared to the result of #n-1. If those builds are not for the same branch, this comparison is quite useless. When building a feature branch before reintegration, I'd like to compare the result with the build just before forking the branch, not with the last build executed. Even worse, the next build of the stable branch is now compared to the last broken build of an arbitrary feature branch.
Everything works fine, if builds are expected to terminate with zero test failures. Unfortunately, when doing test-first-programming (where tests for not yet implemented features exist long before the feature is being implemented) or having lots of regression tests for known bugs that are not yet scheduled, there is a need for precisely monitoring the difference of the test result in comparison to some baseline build - not the test result itself.
Is there a way to do that, planed to be implemented, or regarded a feature worth considering?
Best regards
Bernhard Haumacher