gh-101614: don't treat python3_d.dll as a Python dll

[davidhewitt]

Tentative fix for #101614.

I've verified locally that this fixes the issue for me, I'm unsure what testing strategy may exist in-tree for this.

• Issue: Cannot import extensions linked against python3_d.dll debug build on Windows #101614

[zooba]

Seems fine, though I wish I understood this code better. That += 20 is... intriguing? Concerning? I'm not quite sure

We should add a NEWS entry for this. Something like Correctly handle extensions built against debug binaries that reference ``python3_d.dll``. in the Windows section.

[arhadthedev]

Seems fine, though I wish I understood this code better. That += 20 is... intriguing? Concerning? I'm not quite sure

It's repeated in the end of the loop too:

cpython/Python/dynload_win.c

Lines 158 to 164 in d8e29ba

	}
	import_data += 20;
	}
	}
	return NULL;
	}

Judging by variable names, I guess it's a length of an import symbol entry and we skip the first one of python3.dll for some reason.

[davidhewitt]

20-byte import symbol structures sounds correct based on the comment just above.

cpython/Python/dynload_win.c

Line 110 in d8e29ba

empty values) of structures (20 bytes each), which contains (at

Will push news shortly.

[miss-islington]

Thanks @davidhewitt for the PR, and @zooba for merging it 🌮🎉.. I'm working now to backport this PR to: 3.10, 3.11.
🐍🍒⛏🤖

[bedevere-bot]

GH-101690 is a backport of this pull request to the 3.11 branch.

[bedevere-bot]

GH-101691 is a backport of this pull request to the 3.10 branch.