Joel Dice

@skywhale Thanks so much for reporting this, and sorry for the delayed response. I'll take a look at this when I get a chance, probably this weekend.

I've refactored the code a bit so it's using `pthread_cond_wait` instead of `pthread_cond_timedwait`, so that at least eliminates the `timespec` as a factor. However, it's still failing occasionally: ``` thread...

I've also noticed that tests::aribitrary_case in the main test suite frequently deadlocks forever (with all threads waiting forever and never being notified). I'm guessing that's related even if the symptom...

Update: I went ahead and modified the tests and benchmarks to fork processes instead of spawning threads. Unfortunately, that didn't address the issue -- `pthread_cond_wait` still returns `EINVAL` unpredictably. I...

This comment indicates that PTHREAD_PROCESS_SHARED has been broken in MacOS since Lion: https://github.com/bitcoin/bitcoin/issues/19411#issuecomment-712153417.

Yes, I was wondering about some Mach-specific way to do locking, or else using named pipes just for locking (but still using the ring buffer for actually moving data around)....

@ui4j: https://en.wikipedia.org/wiki/Bionic_(software)

@Wetitpig: This project might be useful to you: https://github.com/bigfatbrowncat/avian-pack. It provides a way to build Avian to use the Android class library, which is itself based on bionic.

Yes, that would be nice. What would need to be changed and/or added?

I've updated the directory layout to start: https://github.com/ReadyTalk/avian/commit/64002319c56a5f3bcf8a6664ac5d66520933c69e