I think I’ve fixed the hang after fork() problem I reported earlier by having the mach clock service reinitialized after fork(). I also discovered that fs_event callbacks weren’t working on kqueue() based backends. I think I’ve fixed this as well, however while the _Stdio testsuite passes on its own, the kqueue tests fail when run in the scope of a full make verify. I’m not sure what to make of that, but I think it’s an edge case. I don’t think it’s related to fork()ing, as the poll device gets reopened after fork.
If all goes well with pike farm, I’ll cherry pick these changes to master.
I also back ported the fix for CritBit grubba mentioned a few days ago, and that seems to have fixed the problem on Solaris/illumos with old gmp.
Bill