Description of problem: I am using the beakerlib skeleton. Currently if I specify that a test is not for a package by specifying "None" (without the quotes), beaker-wizard takes care of it when it generates the Makefile. However, since the runtest.sh skeleton makes it mandatory to have the PACKAGE variable and rlAssertRpm $PACKAGE, it includes it as PACKAGE="None" and the subsequent rlAssertRpm. Perhaps, the runtest.sh skeleton could be modified to not always include PACKAGE and rlAssertRpm or simply check for None and not include it? Version-Release number of selected component (if applicable): How reproducible: Always. Steps to Reproduce: 1. Run beaker-wizard in interactive mode 2. Edit the default "Run for" to None 3. Create test 4. See the runtest.sh file and Makefile Actual results: PACKAGE="None" and rlAssertRpm $PACKAGE Expected results: If "Run for" has been specified as "None", remove the assertion atleast. Additional info:
Created attachment 926074 [details] proposed fix; removes rlAssertRpm if RunFor is 'None'
Please check if the patch works as expected. Also note that 'Package' will remain intact if 'Run for' is changed. Is it desired that the condition takes the value of 'Run for' into consideration?
Thanks for the patch, Martin! I will take a look. In the meantime, will it be possible for you to submit the patch to Gerrit, as described in https://beaker-project.org/dev/guide/writing-a-patch.html#submitting-your-patch ?
Thanks for the reply! I will try to submit it to Gerrit (probably next week).
(In reply to Martin Žember from comment #3) > Please check if the patch works as expected. > > Also note that 'Package' will remain intact if 'Run for' is changed. Is it > desired that the condition takes the value of 'Run for' into consideration? Yes, I would think so.
Patch submitted to Gerrit.
Gerrit review: http://gerrit.beaker-project.org/#/c/3500/
I will not be able to proceed so I am resetting the assignee. Amit, you might want to review and answer Dan's question in Gerrit.
Once we started digging into the patch more closely and looking at beaker-wizard's behaviour, there are lots of places where it assumes that there is *some* package under test. It doesn't seem worthwhile trying to update all of those in case there is no specific package. So I think we will just have to say, running beaker-wizard without a specific package is not a supported use case.