If you follow the links on the PR, you'll see that it was a surprising fix given the failing job output, but it was the only difference I could find with the template and other builds.
Some follow-ups that would be nice:
Testing this revert PR to make sure that this really was the issue, and that there wasn't some coincidental fix: https://github.com/openedx/devstack/pull/1285. For this, it would be nice if you could test a branch in readthedocs, and not literally revert and revert-revert.
Once we have all the info, it would be great to report this to readthedocs. I'm not even sure who gets the edX readthedocs emails (I'm trying to figure that out), but it might be simpler for Axim to do this.
I have no idea if we want to communicate this out in some way, or check if this is an issue in other repos?
I made this change to fix a readthedocs problem:
If you follow the links on the PR, you'll see that it was a surprising fix given the failing job output, but it was the only difference I could find with the template and other builds.
Some follow-ups that would be nice:
Thank you.