@fmeum looks like it needs a repository alias:

(05:30:40) ERROR: error loading package under directory 'graalvm': error loading package '@rules_graalvm~0.9.0//graalvm': Unable to find package for @[unknown repo 'io_bazel_stardoc' requested from @rules_graalvm~0.9.0]//stardoc:stardoc.bzl: The repository '@[unknown repo 'io_bazel_stardoc' requested from @rules_graalvm~0.9.0]' could not be resolved: No repository visible as '@io_bazel_stardoc' from repository '@rules_graalvm~0.9.0'.

I've only applied the @rules_java dependency for users here, so it should not affect users downstream. Can I just push an update to this PR with a fix as a patch, or what do I do?

Thank you for approving.

Yes, you can make any changes to this PR you want, modules are only immutable after merge.

I looked into https://github.com/sgammon/rules_graalvm/blob/main/graalvm/BUILD.bazel and would recommend moving the stardoc targets to a dedicated docs package. Users may want to reference the bzl_library targets, but that would fail due to stardoc being a dev dependency only (as it should be).

@fmeum I've fixed the doc targets, but I'm guessing I should issue a new release and return to this PR with a force push on 0.9.1? While modules are immutable, I'd have to re-tag and re-issue the release to change the tarball that is downloaded, right?

I suppose I could avoid re-tagging and only upload a new stable release tarball, with fixed doc targets. That feels like cheating but I suppose it would work, right?

You could either close this PR and open a new one or force push over it - modules are only immutable after having been merged.

I suppose I could avoid re-tagging and only upload a new stable release tarball, with fixed doc targets. That feels like cheating but I suppose it would work, right?

While this would work, it's of course more transparent if there is a tag corresponding to that tar ball.

@fmeum this should finally be ready to go 😄

Ahhh, it looks like the other one was merged instead, auto-sent by the PR bot (#912). That's bad because the integrity value has been updated since that PR -- I was unable to prevent that PR since it was filed by the GitHub App (I had disabled my workflow thinking that was it, but it crept through anyway).

I will publish a fix release immediately, because I figure BCR modules can't be edited after merge @fmeum.

Very sorry about this mess of all these PRs. After this release, it looks like BCR publishing is working smoothly.

@fmeum (unless there is a way to rollback? i am open to your advice and will prepare the fix release but hold off until i hear)

@sgammon If the 0.10.0 release in the BCR is (and always has been) broken, I think that we could make an exception and edit it.

@meteorcloudy What do you think?.

@fmeum it was the overlap of these two PRs, unfortunately, because I couldn't close the other one. the hash listed on main now is incorrect, but it is correct in this PR.

i can also push a fix release as 0.10.1 and yank 0.10.0, that might be an alternative option? it won't let me build here with an update to a module which exists.

@fmeum it was the overlap of these two PRs, unfortunately, because I couldn't close the other one. the hash listed on main now is incorrect, but it is correct in this PR.

i can also push a fix release as 0.10.1 and yank 0.10.0, that might be an alternative option? it won't let me build here with an update to a module which exists.

That's certainly the easiest solution. You can also try submitting a PR that just deletes the module, but yanking is cleaner. Sorry for the inconvenience!

@fmeum All good, this release was feeling too quiet anyway 😆. I do have control over this PR, so I will push the fix via the PR bot since that is working smoothly now, tag you in it for review, and close this one. In that PR I will also yank this module. That should clean this up nicely and leave only one PR to merge 👍🏻

Thank you for your patience with me!

EDIT: Just kidding, I will file a different PR to yank because I won't have control over that one.

Yanking the broken one sounds fine to me!