serg adding an install test to protected branches will most likely add an overhead of at least 20 mins. The reson for this is the following: for each push, we trigger source file creation. This build triggers build + MTR, that in turn trigger packages creation that finally triggers install. Is this acceptable? Theoretically, we can trigger a package creation build directly, but install will still be triggered after.

Ok, I see. I still think we need to take install tests in branch protection, but let's do lower-hanging fruits first. Meanwhile I'll try to think how to make install tests faster