gh-101100: Docs: Check Sphinx warnings and fail if improved#106460
Conversation
|
Marked as draft to make some temporary commits to demonstrate errors. First, the new check finds five files that no longer contain warnings! https://github.com/python/cpython/actions/runs/5468129112/jobs/9955353278?pr=106460 Some of these files were deleted along with PEP 594's dead batteries. Others have been "accidentally" fixed since |
Sorry, something went wrong.
|
Adding a (temporary) warning to Plus the file is annotated. https://github.com/python/cpython/actions/runs/5468217855/jobs/9955568557?pr=106460 (Edit: output updated per 3522e2c) |
Sorry, something went wrong.
|
Updating from main finds a new file that's been fixed! 🎉 |
Sorry, something went wrong.
|
Thanks @hugovk for the PR 🌮🎉.. I'm working now to backport this PR to: 3.12. |
Sorry, something went wrong.
|
Sorry @hugovk, I had trouble checking out the |
Sorry, something went wrong.
|
3.12 doesn't yet have the refactored |
Sorry, something went wrong.
|
Looks like that was landed and this is required to backport #108065 , so retrying the backport. |
Sorry, something went wrong.
|
Thanks @hugovk for the PR 🌮🎉.. I'm working now to backport this PR to: 3.12. |
Sorry, something went wrong.
|
Sorry, @hugovk, I could not cleanly backport this to |
Sorry, something went wrong.
…thon#106460) (cherry picked from commit 806d7c9)
…ved (pythonGH-106460). (cherry picked from commit 806d7c9) Co-authored-by: Hugo van Kemenade <hugovk@users.noreply.github.com>
|
Had to manually backport it due to conflicts in .nitignore, and might need to resolve a few additional discrepancies there once it builds. |
Sorry, something went wrong.
…ved (pythonGH-106460). (cherry picked from commit 806d7c9) Co-authored-by: Hugo van Kemenade <hugovk@users.noreply.github.com>
…H-106460) (#108116) * gh-101100: Docs: Check Sphinx warnings and fail if improved (#106460) (cherry picked from commit 806d7c9) * [3.12] gh-101100: Docs: Check Sphinx warnings and fail if improved (GH-106460). (cherry picked from commit 806d7c9) Co-authored-by: Hugo van Kemenade <hugovk@users.noreply.github.com> --------- Co-authored-by: Hugo van Kemenade <hugovk@users.noreply.github.com>
Recently I improved some colour contrast for Pygments, and they have a file of current contrast ratios. If they're improved by a PR, the CI will fail to make sure the ratio constants are updated, to ensure continual improvement with no regressions.
This is a good idea!
Let's add it here, to check if any files in
.nitignore(which may have Sphinx warnings) switch to having no warnings. Then we can remove them from.nitignoreto make sure they remain free of warnings.I was going to add another Sphinx build, a fourth, to touch the relevant files, so only those would be rebuilt. But four builds is far too many!
Instead, let's build only once and save the warnings to a file. Then we can process the file and do the three phases:
.nitignore(new!)📚 Documentation preview 📚: https://cpython-previews--106460.org.readthedocs.build/