we have enough power to handle running builds in parallel.
there is another ticket that talks about optimizing build params to save time.

I think it will just prolong significantly our output from experimental flows without a major benefit.

we now have enough slaves and jenkins power to handle both builds in parallel so I don't see benefit in doing this, we should invest in improving our infra and adding slaves if needed.

also, there is another ticket that talks about not building ALL permutations on build_artifacts for ovirt-engine and only doing that for official builds.
see https://ovirt-jira.atlassian.net/browse/OVIRT-416 for more info

The issue with this is that it effectively duplicates the run time, as most projects for now compile and package twice (see ovirt-engine) the time from merge to built and tested will change from ~1:30 to ~3h if it does not fail (as it's expected not to fail in most cases).

I think this should go only after we can reuse artifacts from job to job