Restore default role check in make check.

[ezio-melotti]

This PR restores the check for the default role (see GH-92289) by doing two things. First, it sets the severity to 0 adding --severity=0, so that failing the default role check results in an error. This however also enables the long lines check, so it explicitly disabled it with --disable='line too long'. There are currently no other checks with severity=0.

This PR only fixes Makefile. If this approach is sound, make.bat should be fixed too.

Using --enable instead might be a better approach, once sphinx-contrib/sphinx-lint#27 lands.

As a side note, it took me a bit to figure out that --disable= expected a message rather than the name of a checker.

cc @JulienPalard

[ezio-melotti]

With this PR, the check successfully failed:

PATH=./venv/bin:$PATH sphinx-lint -i tools -i ./venv -i README.rst --severity=0 --disable='line too long'
No problems found.
PATH=./venv/bin:$PATH sphinx-lint ../Misc/NEWS.d/next/ --severity=0 --disable='line too long'
[0] ../Misc/NEWS.d/next/Library/2022-04-26-18-02-44.gh-issue-91928.V0YveU.rst:1: default role used (hint: for inline code, use double backticks)
[0] ../Misc/NEWS.d/next/Documentation/2022-04-24-22-09-31.gh-issue-91888.kTjJLx.rst:1: default role used (hint: for inline code, use double backticks)
2 problems with severity 0 found.
make: *** [Makefile:217: check] Error 1
##[error]Bash exited with code '2'.

[ezio-melotti]

I added the same changes to make.bat, but I don't have a Windows machine to test them (cc @python/windows-team).

make.bat seems outdated compared to Makefile, so I made a couple of changes to make them converge:

• b973542 added the sphinx-lint check to Misc/NEWS.d/next (see Deprecate suspicious.py? #86404 (comment))
• b3f1f59 added README.rst to the ignored files, but I don't see a reason to ignore it, so I removed it from Makefile to align it with make.bat instead (see
  Rename Doc/README.txt to Doc/README.rst and add formatting #104 (comment))

In addition, bfba2cd added the venv dir to the ignored dirs, but make.bat doesn't seem to use venv, so no change was required.

[vstinner]

The CI fails:

Error: [0] ../Misc/NEWS.d/next/Documentation/2022-04-24-22-09-31.gh-issue-91888.kTjJLx.rst:1: default role used (hint: for inline code, use double backticks)
Error: [0] ../Misc/NEWS.d/next/Library/2022-04-26-18-02-44.gh-issue-91928.V0YveU.rst:1: default role used (hint: for inline code, use double backticks)

[ezio-melotti]

Yes, I'm leaving the failures so that people can test if they are detected on Windows too. If they are, I'll fix them and merge the PR.

[AlexWaygood]

Tried the patch out on my Windows machine:

C:\Users\alexw\coding\cpython>"Doc/make.bat" check
Using py -3.10 (found 3.10 with py.exe)
Error: path too does not exist
Error: path too does not exist

(I'll be honest, I usually don't bother with the .bat file when I'm writing a docs patch. I usually use sphinx-build to build the docs locally, git diff to check for trailing whitespace, and sphinx-lint in CPython's CI to check for markup errors.)

[ezio-melotti]

Thanks everyone for the feedback and testing! I'll wait for the feedback of @JulienPalard before fixing the 2 failures and merging the PR.

[JulienPalard]

@ezio-melotti Don't you prefer waiting for sphinx-contrib/sphinx-lint#27 to be merged, avoiding two commits on cpython doing the same thing?

It would look like sphinx-lint --enable default-role, which is more transparent than --severity 0 --disable 'line too long'.

[ezio-melotti]

The PR should be ready to be merged.
@AlexWaygood / @zware: can you check if it works on Windows? (you might have to update sphinx-lint)

[ezio-melotti]

Looks like it works on WIndows too.

[miss-islington]

Thanks @ezio-melotti for the PR 🌮🎉.. I'm working now to backport this PR to: 3.9.
🐍🍒⛏🤖

[miss-islington]

Thanks @ezio-melotti for the PR 🌮🎉.. I'm working now to backport this PR to: 3.10.
🐍🍒⛏🤖

[miss-islington]

Thanks @ezio-melotti for the PR 🌮🎉.. I'm working now to backport this PR to: 3.11.
🐍🍒⛏🤖

[miss-islington]

Sorry @ezio-melotti, I had trouble checking out the 3.9 backport branch.
Please backport using cherry_picker on command line.
cherry_picker 953ab0795243900ccccaaca069d932730a86fc20 3.9

[miss-islington]

Sorry, @ezio-melotti, I could not cleanly backport this to 3.10 due to a conflict.
Please backport using cherry_picker on command line.
cherry_picker 953ab0795243900ccccaaca069d932730a86fc20 3.10

[miss-islington]

Sorry @ezio-melotti, I had trouble checking out the 3.11 backport branch.
Please backport using cherry_picker on command line.
cherry_picker 953ab0795243900ccccaaca069d932730a86fc20 3.11

[bedevere-bot]

GH-92821 is a backport of this pull request to the 3.11 branch.

[JulienPalard]

Thanks @ezio-melotti!