Linter comment workflow breaks if the repo fork is renamed

[freundTech]

Bug:

I just noticed that the make comment for staged-recipes linter workflow breaks when the head repository has been renamed.
In my case, I renamed my fork from freundTech/staged-recipes to freundTech/conda-forge-staged-recipes, but other users have also done so.
Here is an example of a failing workflow run.
The problem seems to be that the python script always assumes the head repo to be {args.head_repo_owner}/staged-recipes in line 44.

@conda-forge/staged-recipes

[Varun786223]

@freundTech @conda-forge/staged-recipes

Hi FreundTech, thanks for catching this—it’s definitely a pesky bug! I took a look, and you’re right: the "make comment for staged-recipes linter" workflow is tripping up because the Python script hardcodes the head repo as {head_repo_owner}/staged-recipes on line 44. When you renamed your fork to freundTech/conda-forge-staged-recipes, the script couldn’t find the repo, causing the workflow to fail. It’s a classic case of assuming fork names never change, which doesn’t hold up since renaming forks is pretty common.

I’d like to contribute a fix: instead of hardcoding, we can grab the actual head repo name from the GitHub event (github.event.pull_request.head.repo.full_name). Here’s the tweak I’m proposing for the script:

# Before
repo = f"{args.head_repo_owner}/staged-recipes"

# After
import os
repo = os.getenv("GITHUB_HEAD_REPO", f"{args.head_repo_owner}/staged-recipes")  # Fallback for safety

And in the workflow, we’d pass the head repo name like this:

- name: Run linter script
  env:
    GITHUB_HEAD_REPO: ${{ github.event.pull_request.head.repo.full_name }}
  run: python make_comment.py --head-repo-owner ${{ github.event.pull_request.head.user.login }}

This should make the workflow handle renamed forks smoothly. I also suggest adding a log warning if the repo name differs from the expected staged-recipes, just to keep an eye on things.

I’d be happy to open a PR with this fix—should unblock your workflow and help others hitting the same issue. @conda-forge/staged-recipes, any thoughts on this approach? FreundTech, great catch again—let me know if you’d like to test the fix together!

[carterbox]

Seems reasonable to me.