PRs ready for testing (and already tested)

Many PR’s are just version bumps, like:

Such PR’s don’t really need a ‘review’, but they do need to be tested. I wonder if we could find people who might not feel comfortable reviewing, but are comfortable testing. We could somehow mark such PR’s so they are easy to find (either with a tag or a discourse topic).

Of course after being tested they still need a final look-over by someone with commit access, but perhaps that becomes easier when you already know the PR has been tested?


Maybe GitHub - timokau/marvin-mk2: Discontinued! See (Previously: "Making sure your PR gets a review and your reviews don't get lost.") could help?

1 Like

When I read this, I thought of the StackOverflow/Exchange triage queue that occasionally pops up with a red icon indicating that there’s a small way the viewer could break off a piece of work to help the ecosystem.

It’s always hard to guess if any given effort pays off, but if Nix prompted me…

Nix, whispering: hey, you; can you help us test an update to this package you’re already using?

…it’s hard to imagine I wouldn’t say yes.

(I bang this drum any time I get a chance, but I feel like there’s potential for tooling around micro-contributions that can represent big value for the community, but that are also tiny things that would be hard for some to justify overhead like PRing or cloning nixpkgs. Things like manual testing, identifying real-world invocations that exercise a package, fishing for new maintainers, etc.)

1 Like

388 Potentially broken qt applications & libraries need your help had a bunch of things that need testing :wink:

1 Like

The PR is already working but it’s blocked by something like a political decision.

It adds a variant of the package with some unfree assets so it’s actually useful but this package is marked as unfree so I can’t see why it can’t be packaged as the org cache would not redistribute it.