Anyone has a good workflow for compiling a Fedora package from git? There’s a lot of ceremony involved the few times I tried (set up rpm tree, update spec, fedpkg, maybe try a COPR, argh…) and I often have a need to test a new patch from upstream.
I used to use Arch Linux and the process is about 2 orders of magnitude simpler. Download the PKGBUILD, replace the original unpacked source with a git repo, makepkg and install the package.
There’s a pipewire patch I’d like to test on Fedora and I wouldn’t even know how to get started, and funnily enough, I’ve created RPMs a dozen times, but it’s so convoluted I immediately forget. And spec files seem only to work with stable tarballs, and trying to get it to work off a git repo is not very well documented.
Lastly, there is a lot of documentation for very old versions of Fedora and it’s really hard to find something that’s up to date. Packaging RPMs seems to be a dark art nobody seems interested in simplifying… and it’s sad because I’d like to contribute to Fedora.
In a perfect world I’d have one script to
- Download spec from src.fedoraproject.org
- Modify the spec to point to a custom source
- Build and create an RPM
But it doesn’t seem to exist.
Does anyone have a good, simple guide/script that can automate the process of building an RPM for an existing package, but from the upstream git repo? How do packagers deal with this stuff?