gh-101047: Remove LIBTOOL_CRUFT

[indygreg]

This variable has been unused since commit
1919983 in 2011.

I noticed it because the arch logic breaks when cross-compiling on macOS.

• Issue: Remove LIBTOOL_CRUFT from configure #101047

[sobolevn]

Please, run make regen-configure to fix the CI :)

[indygreg]

I saw the configure failure and was kinda hoping some automation would suggest a fix since I didn't feel like shaving the yak of installing autoconf 2.69. But now that I'm enlightened about the existence of a make rule that pulls a container image with autoconf 2.69, there's no yak to shave. Nice!

Force pushed with the configure regen.

[ned-deily]

Thanks for the PR. It looks OK to me. But, if we're going to clean up vestigial libtool code from the early days of MacOS X, we should also remove the few references to OTHER_LIBTOOL_OPT in configure.ac, configure, and Makefile.pre.in.

[bedevere-bot]

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

[indygreg]

Thanks for the PR. It looks OK to me. But, if we're going to clean up vestigial libtool code from the early days of MacOS X, we should also remove the few references to OTHER_LIBTOOL_OPT in configure.ac, configure, and Makefile.pre.in.

I agree with you on the desire for some ancient macOS support code to get purged. But one of the things I learned from being a maintainer of Firefox's build system for several years (and submitting + reviewing thousands of commits to the build system) is that scope bloating build system changes is a recipe for unexpected regressions. Due to the nuance and sensitivity of build systems, I've learned it is best to keep changes small and tightly scoped.

So I'm gently pushing back on the request to scope bloat an otherwise looks OK PR.

(FWIW I'm sitting on a pile of patches to CPython's build system that I'd love to get upstreamed. Many of them involve Apple OS support code. I could potentially add OTHER_LIBTOOL_OPT to that set.)

[ned-deily]

I agree with the desire to avoid "scope bloat" but I think the key here is what is the reason for making this change at all. In general, we don't make code cleanup changes without a reason. But this PR identifies a good reason, that is, the commit you cited, 1919983, was incomplete; since it removed the last use of libtool for macOS builds, it should have removed the remaining references to it. The PR as it stands removes one of those; a quick grep of the source identifies the other, the OTHER_LIBTOOL_OPT variable (also seen in the cited commit). So the reason for making this change is to remove libtool cruft, not just LIBTOOL_CRUFT, and, if we make such a change, we need to finish the job not just leave it partially done. It's pretty clear that this is a low-risk change but we would further mitigate risk of problems by only making such a change in the main branch for the next feature release; i.e. we would not backport it to existing releases as it fixes no bugs.

[indygreg]

I refreshed the commit to also delete code related to OTHER_LIBTOOL_OPT.

[ned-deily]

Thanks for making the changes!