Admin message

Due to an influx of spam, we have had to require each new account to be manually approved. Please register an account and then write an email to accountsupport@archlinux.org to get it approved. Sorry for the inconvenience.

Feature request: Output a warning on `pkgctl release` if nvchecker intergration is not (properly) setup
We adopted `nvchecker` as our standard tool for packaging version handling and we'd like to rely on its integration in our packages sources (in the form of the `.nvchecker.toml` file) for future automation on that front (some efforts have already started and are currently being tested, see https://gitlab.archlinux.org/archlinux/nvchecker-poc & https://gitlab.archlinux.org/archlinux/packaging/packages/codespell/-/issues/2 for instance). As such, I think it would be great for `pkgctl` to emit a WARNING during `release` if the nvchecker integration is not setup (or not correctly), as a way to remind packagers to set it up and detect eventual issues. The way I imagine it would be a check performed when `pkgctl release` is invoked which would test the following: - Does the `.nvchecker.toml` file exists? If not, output a WARNING that nvchecker integration isn't set up (and maybe hint that `pkgctl version setup` is a thing :stuck_out_tongue:). - If the `.nvchecker.toml` exists, test it. For instance, by running `pkgctl version check` and verifying that the exit code is equal to 0 or 2. If not, output a WARNING that the nvchecker integration might be broken. Some packages (e.g. meta packages) don't really have any upstream to follow in terms of versioning, so the check needs a way to be muted in order to avoid false positives in such cases. A potential solution could be to add a specific property in the `.nvchecker.toml` file that can be checked in order to react properly during `version check` (e.g. skip it). Another thing to keep in mind is that such a check will make `nvchecker` a hard dependency for `devtools` (it's an optional one currently). I'm opening this issue to confirm the usefulness of this idea and to agree of the implementation's specifications before opening a MR for it :)
issue