New comment by ericonr on void-packages repository

https://github.com/void-linux/void-packages/pull/29635#issuecomment-808853504

Comment:
Two quick nits on the current PR: the `pari` commit messages should be squashed together, and the `giac` commit should include `update to $version`, instead of just `update` :)

@tornaria, thanks for the thorough analysis.

> I think the reason is that TLS is slow.

This shouldn't affect things so much, unless they made the terrible decision of using `__thread` globals for everything? Or worse, normal global with `pthread_mutex` calls around accesses...

> I wonder if it would make sense to offer different versions of pari (single thread / multi thread). At the very least, I'd make pthread a build option so that one can easily compile one or the other as needed.

That sounds reasonable, especially since there seems to be such a difference.

> Is there a way void could support something like "subarchitectures"? Say x86_64-avx where some packages are compiled for this architecture and the rest default to the packages for x86_64.

Maybe as a build option, at most. I don't believe we wish to carry such packages, because it can increase the maintenance burden a lot. If the threaded version becomes necessary for sage (as a system package) but the performance loss is too high, we can consider it then, but I'm against it when it comes to architecture. Ideally, they would compile opimitized versions of some files and choose between them at runtime...