โ
For Homebrew formulae, there's a command that encourages users to contribute version bumps for formulae they use: brew bump-formula-pr.
โ
This works great in Homebrew/homebrew-core for macOS, but not so great in Homebrew/linuxbrew-core as the version bump mechanism for the majority of Homebrew on Linux formulae is via merges from the macOS repo into the Linux repo, roughly once per day, done by maintainers.
โ
Currently, Homebrew on Linux users who employ brew bump-formula-pr as their tool of choice raise their formulae PRs against Homebrew/linuxbrew-core as that's the default tap on Linux. This is a bad user experience, as users who try to be helpful end up with their pull-request closed by maintainers, because they've unknowingly submitted it to the wrong repo.
If the tap's remote is Homebrew/linuxbrew-core, brew bump-formula-pr could somehow:
โ
Homebrew/homebrew-core, then either:โCurrently, users who submit formulae updates are redirected by maintainers to submit their changes to Homebrew/homebrew-core in most cases, or they realise that themselves. This takes up maintainers' time. It's also potentially discouraging for the submitter, because their PR is closed through no fault of their own.
The tools that we advertise as being good entry points for contributing should work according to our workflows, across both macOS and Linux.
Evidence:
โ
A _lot_ of users update formulae with brew bump-formula-pr in Homebrew/homebrew-core, and as Homebrew on Linux gains more users, we'll run into this problem more and more.
โ
โ
brew bump-formula-pr while running Homebrew on Linux. This has the side effect of discouraging initial contributions from users, decreasing the likelihood that they will get more involved in the future.Whoops, it looks like you created an issue without filling out the checklist and providing all the needed information from one of the issue templates: https://github.com/Homebrew/brew/issues/new/choose
As we need all this information to help you we're closing this issue. If you add the necessary information to this issue (don't create a new or duplicate issue, please) and comment to let us know, a Homebrew maintainer will check it out when they get the chance. If it's deemed to be a valid issue then this issue may be reopened.
Sorry!
This makes sense to me. Good idea @issyl0!
- Not allowing users to
brew bump-formula-prwhile running Homebrew on Linux. This has the side effect of discouraging initial contributions from users, decreasing the likelihood that they will get more involved in the future.
This may make sense to do as a quick PR while the other one is in development.
[Not allowing users to
brew bump-formula-prwhile running Linux] may make sense to do as a quick PR while the other one is in development.
I was going to do a quick odie "Unsupported on Linux." unless OS.mac? edit to address this, then I remembered that brew bump-formula-pr is used for Linux-specific taps other than linuxbrew-core (for example linuxbrew/extra and xorg), and that would disable it for _everyone_ no matter what formula they were trying to update in which tap. Due to this, I feel we should just wait for the actual improvements rather than dealing with these edge cases temporarily. :crossed_fingers: :pray:
I just run into this issue. Good it is already tracked.
Can I ask if there are any ideas on how to make bump-formula-pr work on linux too?
Can I ask if there are any ideas on how to make
bump-formula-prwork on linux too?
Yes, this is under the header "A detailed description of the proposed feature" in the original post. Please feel free to open a pull request working on this; I'd be happy to review it.
I'd like to tackle this
@morganrconnolly, please go ahead! ๐
@morganrconnolly any progress on this?
@dawidd6 apologies for the delay. Here's a first stab https://github.com/Homebrew/brew/pull/6538 . I'm not sure what the appropriate development workflow is for brew commands, e.g. how do modify a command, rebuild brew and then test the modified command. Is there any documentation about this you can point me towards?
Most helpful comment
I'd like to tackle this