What started as a relatively simple step (repoHasTests()) turned into a quite complicated thing that is trying to do way too much.
The behavior of the step is not even fully documented -- but this is on purpose, i.e. we don't want to fix the docs.
It would be better to split the functionality into 2 or 3 separate steps:
clone step
extract information step(s)
This way we could clone the repository and then run any number of steps that would extract all the info we need.
Currently one of the hidden (undocumented) features of the step is that it also reads ci.fmf file and checks if a specific configuration option is there (resultsdb-testcase).
What started as a relatively simple step (
repoHasTests()) turned into a quite complicated thing that is trying to do way too much. The behavior of the step is not even fully documented -- but this is on purpose, i.e. we don't want to fix the docs.It would be better to split the functionality into 2 or 3 separate steps:
This way we could clone the repository and then run any number of steps that would extract all the info we need. Currently one of the hidden (undocumented) features of the step is that it also reads
ci.fmffile and checks if a specific configuration option is there (resultsdb-testcase).