Gabriel Scherer

I think that you do not need the CAMLparam and CAMLreturn here. In the first function, `ret` has type `long` so technically it should be `CAMLreturnT(long, ret)` I think --...

I think that this can be closed following #12212, although of course there are other sources of false sharing than atomic values that could degrade program performance.

Note: when I say "this introduces a configuration change, so I (configure and make again)", what happens in practice is that when I `make -j` again, I see this (this...

> Looking only at the second failure, and it may now be hard to answer, but are you certain that you'd re-run configure at this stage? Apologies for being unclear....

Thanks for the work in this direction! Minor remarks in case it helps: - The "configuration issues could warn properly" is secondary I think (... or really I should have...

If I understand correctly, both issues reported have been fixed in #11987 and #11988. Thanks @dra27 and @shindere! Closing.

No, for example because we don't fully trust `recommended_domain_count` to actually return a truthful/correct/accurate value on all systems. Remember that Domain is intended to be a first building block for...

@dclements I agree. Would you like to submit PRs to improve the documentation in various places? There is no recommended pooling library because early 5.x releases were designed as "early...

I don't buy the argument. My working directory changes *all the time* when I program, for example when I `git rebase -i` to cleanup the commit history -- I will...

@dra27 you had an experimental branch that makes this work. Have you managed to upstream this, or part of it?