[3.13] gh-146004: fix test_args_from_interpreter_flags on windows (GH-146580)

[miss-islington]

(cherry picked from commit 1af025d)

Co-authored-by: Chris Eibl 138194463+chris-eibl@users.noreply.github.com

• Issue: recent -X options do not auto-propagate to multiprocessing processes #146004

[gpshead]

!buildbot .win.

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @gpshead for commit e0125f6 🤖

Results will be shown at:

https://buildbot.python.org/all/#/grid?branch=refs%2Fpull%2F155965%2Fmerge

The command will test the builders whose names match following regular expression: .*win.*

The builders matched are:

• ARM64 Windows PR
• AMD64 Windows PGO Tailcall PR
• AMD64 Windows11 Non-Debug PR
• ARM64 Windows Non-Debug PR
• AMD64 Windows10 PR
• AMD64 Windows Server 2025 Refleaks PR
• AMD64 Windows Server 2022 NoGIL PR
• AMD64 Windows PGO NoGIL Tailcall PR
• AMD64 Windows PGO PR
• AMD64 Windows Server 2025 Clang PR
• AMD64 Windows PGO NoGIL PR

[chris-eibl]

This is good to merge IMHO:

• I've tested locally before doing the backport
• the failing buildbots are unrelated:
  • tail-calling isn't supported on Windows before 3.15. I don't know why the PR workers were triggered, maybe because branches={MAIN_BRANCH, PR_BRANCH}). IMHO, it should build on main and 3.15, and for PRs against those two branches.
  • the clang-cl failure is known on older branches: the fix did not get backported that far, hence Windows64ClangBuild has BRANCHES.only_since(3, 14). Furthermore, the new Windows clang-cl buildbots from @zware have tags={UNSTABLE, NO_TIER}. I just don't know why it triggered for a 3.13 PR ...

[chris-eibl]

• branches={MAIN_BRANCH, PR_BRANCH}). IMHO, it should build on main and 3.15, and for PRs against those two branches.

@zware: would branches={BRANCHES.only_since(3, 15), PR_BRANCH.only_since(3, 15)} work?

[zware]

See python/buildmaster-config#654; we don't currently have a way to restrict PR builders to particular base branches as we don't currently detect the base branch of a PR at all.