Would it make sense to have the "rebuild with ssh access" call mock with "--no-clean-after"? Without that, there is not that much more to inspect than the chroot itself, is there? Or is the idea that we run mock again?
The idea, is that you will be able to execute the command in the
chroot.
So instead of building packages in mode "try this" and "try that", you will be able to run e.g. gcc in that chroot. Maybe put there some print() for debugging. To find the culprit of failure faster. Especially on non-intel architectures.
-- Miroslav Suchy, RHCA Red Hat, Manager, Packit and CPT, #brno, #fedora-buildsys