15:12:02 <tsoome_> mmm sauna + Kulmbacher Mönchshof... 
18:15:00 <jclulow> alanc: what is postwait_post(3C)?!
18:33:18 <sommerfeld> https://docs.oracle.com/cd/E88353_01/html/E37843/postwait-post-3c.html ?  (unclear what you'd use it for..)
19:06:53 <richlowe> it's how you send letters to Theo
19:06:56 <richlowe> obviously
19:07:08 <jclulow> sommerfeld: yes that one haha
20:02:56 <gitomat> [illumos-gate] 16765 loader: remove unused bzipfs, splitfs, pkgfs, ext2fs -- Toomas Soome <tsoome⊙mc>
21:29:47 <sommerfeld> the postwait_post thing reminds me that I was annoyed that there doesn't seem to be a way to use event ports to wait for a process to exit.
21:35:38 <sommerfeld> there does seem to bne something in there (postwait_post_deferred() with PWPOST_ATEXIT) but it's unclear if that survives exec())
22:08:39 <jclulow> sommerfeld: I was actually just talking to rmustacc about PORT_SOURCE_PID haha
22:09:37 <jclulow> Probably you would associate with a pid and it would return you the event when it was safe to call waitid() without blocking
22:23:11 <sommerfeld> yeah, PORT_SOURCE_PID would be one way to do it.   
22:24:14 <sommerfeld> linux prior art in this space is pidfd_open() (which if we implemented you could then PORT_SOURCE_FD it)
22:30:44 <sommerfeld> presumably with useful guarantees about avoiding trouble from pid reuse (likely with similar caveats for pidfd_open() on linux -- you must call it on a pid you got from fork() before you have reaped it or did anything that would have allowed the kernel to auto-reap it)