make patchcheck: Allow Tabs in Some Cases

[ericsnowcurrently]

Currently Tools/patchcheck/reindent.py¹ replaces all tabs with spaces. However, in Tools/c-analyzer/cpython._parser.py there are embedded tab-based tables. Up to now I've basically ignored CI failures (from make patchcheck) in _parser.py, but I'd rather reindent.py be a little smarter (even if just a whitelist). ² Tools/scripts/untabify.py sounds like it may have a similar story.

Footnotes

1. reindent.check() is used by Tools/patchcheck/patchcheck.py (AKA make patchcheck). ↩

2. Addressing trailing-tabs-in-str-literals is even more relevant since I enabled the check-c-globals CI check. For example, see https://github.com/python/cpython/pull/102735#issuecomment-1470516458. ↩

[serhiy-storchaka]

Why not to use \t instead of literal tab?

[ezio-melotti]

I think this is the mentioned table:

cpython/Tools/c-analyzer/cpython/_parser.py

Lines 130 to 141 in d20bb33

	MACROS = clean_lines('''
	# @begin=tsv@
	glob name value
	Include/internal/*.h Py_BUILD_CORE 1
	Python/**/*.c Py_BUILD_CORE 1
	Parser/**/*.c Py_BUILD_CORE 1
	Objects/**/*.c Py_BUILD_CORE 1
	Modules/_asynciomodule.c Py_BUILD_CORE 1

Here \ts would make it less readable. Would it make sense to move that to a stand-alone .tsv file?

[ericsnowcurrently]

Here \ts would make it less readable.

Correct.

Would it make sense to move that to a stand-alone .tsv file?

I considered that but didn't have time to try it out. I'd rather have that particular data stay in the current file if possible though. One alternative is to switch to csv. I don't recall that these specific tables need tsv.

[ericsnowcurrently]

CC @AlexWaygood