bpo-42238: Doc CI: Disable suspicious checks.

[JulienPalard]

They are slow and raise too many false positive, I'm in the slow
process to try to change this.

I've removed it from PEP 101 too to avoid the burden for RMs.

I've setup a cron to follow this closely / to avoid to forgot to monitor this closely.

https://bugs.python.org/issue42238

[vstinner]

LGTM.

Maybe you can document this unknown command in the devguide? For example, near:
https://devguide.python.org/documenting/#gotchas

[JulienPalard]

For the moment my biggest issue is: people are autonomously and quietly fixing what's reported by make suspicious, so I don't have a list of true positives to work on.

So my current idea is to let people people commit errors¹, fix them myself (I've setup a build bot to monitor this closely) to document true positives and try to implement them in make check (or in Sphinx).

1: It is not expected to happen often, and have the nice side effet of not bothering people with false positives: 9 out of 12 errors reported by make suspicious are false positives.

[bedevere-bot]

GH-98221 is a backport of this pull request to the 3.10 branch.