On Dec 20, 2007 5:52 AM, Armando Di Cianno <armando_AT_goodship.net> wrote:
> I'm going to travel soon, and I didn't want this issue to get lost to
> the memory hole, so please excuse the irc copy and paste.
>
> I'm guessing that these extra processes hanging around isn't by design,
> and should be considering a bug.
Yeah, pretty sure they're harmless though. Something along the way
(wmii? rc?) is setting up a signal handler for SIGPIPE, which is of
course inherited across forks. Combined with the fact that "ixpc read"
doesn't check the return value of its write to stdout and you've got a
process that won't die (at least, until you close wmii - it does check
the return on the reading side).
At a guess, I'd say this is p9p related - part of turning signals
into notes? But I'm not familiar with the details there, and never
tracked down the cause.
-sqweek
Received on Thu Dec 20 2007 - 04:17:40 UTC
This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:32:47 UTC