I am a contributor to fedora-license-data upstream, and I investigated what will be the easiest way to build a new version of the Fedora package. I ended up using Packit for that. With just one command, it creates a pull request in src.fedoraproject.org.
At the start, we have to configure the upstream. I run:
$ packit init
in the upstream repository. This command created .packit.yaml with
# See the documentation for more information:
# https://packit.dev/docs/configuration/
specfile_path: fedora-license-data.spec
# add or remove files that should be synced
files_to_sync:
- fedora-license-data.spec
- .packit.yaml
# name in upstream package repository or registry (e.g. in PyPI)
upstream_package_name: fedora-license-data
# downstream (Fedora) RPM package name
downstream_package_name: fedora-license-data
The upstream uses tags to mark a release. The commits look like this:
* e960257 (tag: v1.0) Merge branch 'noarch' into 'master'
I put in .packit.yaml one line to configure how we do a release:
upstream_tag_template: "v{version}"
The packit.yaml is highly configurable and you can see result of my .packit.yaml.
Now I can run command:
# packit propose-downstream
If you did not git-push it, you have to run:
# packit propose-downstream --local-content
For some unknown reason, Packit still cannot figure out the version automatically. So I had to run:
# packit propose-downstream --local-content . 1.0 --dist
This command creates PR for all branches. To limit for just some branch(es) you can use the command line option --dist-git-branch rawhide,f36.
You can see the resulting PR here
By the way, the Packit team is working on a feature where you can configure this workflow in dist-git, and you will not need to modify the upstream - if you are not a direct upstream contributor.