On Wed Jul 10 19:23:10 2024 +0000, Brendan Shanks wrote:
Yeah that's better. I guess it's still possible that one thread could be inside sysctlnametomib() and writing to mib while another one is trying to call sysctl() for it, but that seems pretty remote.
Yeah, it's still not thread-safe. I don't think adding a `get_pcblist_mib()` (or whatever) helper that used `pthread_once` would be too horrible.