This is much better than with dnf4 variant, where we had to decide server-side (copr-frontend), and we didn't have good enough information about the client machine (is that Alma? is that CentOS Stream? RHEL? What user prefers?).
I'm curious if we could provide some "hint" like info in the /api_3/rpmrepo/<owner>/<project>/ call, saying that some chroots are good enough fallbacks for other chroots (like that epel-10 is a good fallback for rhel+epel-10 and vice versa).
Maybe let the Copr owner decide it.
At this point in time,
dnf5 copr enable
implements the fallback mechanism itself: https://github.com/rpm-software-management/dnf5/blob/117bc3e552147925ac469179d4c8f79f06a1eaca/dnf5-plugins/copr_plugin/helpers.cpp#L8-L15This is much better than with dnf4 variant, where we had to decide server-side (copr-frontend), and we didn't have good enough information about the client machine (is that Alma? is that CentOS Stream? RHEL? What user prefers?).
I'm curious if we could provide some "hint" like info in the
/api_3/rpmrepo/<owner>/<project>/
call, saying that some chroots are good enough fallbacks for other chroots (like thatepel-10
is a good fallback forrhel+epel-10
and vice versa). Maybe let the Copr owner decide it.