Our project uses autoconf/automake and automake's "make check" target to do regression testing. I currently use a simple "Script" task to run this sequence of commands:
autoreconf -fis
./configure
make
make check
The build is labeled "testless build" and the results of "make check" are treated as compilation errors.
Is that the recommended way of doing this? Should I not run "make check" as a separate task? But how do I then accesss the binary produced in the "make" step? And where is the "script" task in the "tests" task type so I can run "make check"?
Thanks already,
nick
Depends on your end goal. What does make check do? Does it supply test results in a format that can be understood by Bamboo (for a list, have a look at Test task type in the wizard)? If it does (or it can be configured to), you could add a test parsing task as a finalising task in your job. In this way, you'd get the test results into Bamboo.
If you have a new format, you can write a new test parser plugin - it's fairly simple and you can you use existing test parses as templates. If you don't want to do that, you can file an enhancement request for your test format at http://jira.atlassian.com
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.