Document the reasons for using Mock / key features

[praiskup]

We frequently explain why plain "rpmbuild call in podman/buildah" isn't a replacement for Mock. It would be nice to have documentation documenting the key benefits of Mock from this perspective (with some technical details).

/note for myself: ADR

[xsuchy]

Notes for later:

• Mock prepares directory structure for rpmbuild https://github.com/rpm-software-management/mock/blob/main/mock/py/mockbuild/buildroot.py#L783
• Mock prepares fuse device https://github.com/rpm-software-management/mock/blob/main/mock/py/mockbuild/buildroot.py#L810 (important for rpm-ostree builds)
• Mock setup structure for build-specific macros https://github.com/rpm-software-management/mock/blob/main/mock/py/mockbuild/buildroot.py#L763 and other places
• Mock prepares special devices that are needed for various packages to succeed the build https://github.com/rpm-software-management/mock/blob/main/mock/py/mockbuild/buildroot.py#L833
• Mock correctly setup time settings https://github.com/rpm-software-management/mock/blob/main/mock/py/mockbuild/buildroot.py#L509

[jafd]

There probably should be a mention that mock goes out of its way to have a shared cache of downloaded RPMs, which is significant when you build a lot of packages. It (and abovementioned things) kinda sorta can be implemented with lots of bash and docker/podman, but then you essentially end up reimplementing mock badly, and mock is already using containers when it can.

I've been thinking of forgoing mock because of some warts it has that make my workflow inconvenient sometimes, thinking I would just implement a subset of what mock does, but in the end I've ended up discovering lots of edge cases mock already handles capably.