Mock runner includes a function that runs mock to initialize the mock environment before it actually gets used.

Theoretically the initialization could have been done in the same mock invocation that makes use of it. It seems the reason for the separate initialization was that it could be done without including bind mount configuration, because there was a need to pre-create the mount points inside the chroot before bind mounts could be done.

Since mock 1.4.10, it can now create the mount points for bind mounts on its own, a separate hase for creating mount points is no longer needed and tracks the work to remove it.

It seems that once the mount point creation phase is removed, the separate initialization could be removed as well. Preliminary testing showed that this may be more tricky then anticipated though, as the way 'chroothome' mock configuration option is configured by mock_runner causes issues for mock when bootstrap chroots are used (The reason this is not causing issue for the separate initialization is that this option only gets configured along with the bind mount options only when mock is invoked to do actual work).

To remove the separate initialization phase and reduce the amount of times mock is called from 3 to 2, we need to work around the issue above by either submitting a patch to mock itself or finding a work-around that could be used from mock_runner.sh.