Closed GoogleCodeExporter closed 9 years ago
This was originally added in 032912ce0c3a1cbf3a14395a8a621f48544ddc71
but trond changed part of it in a bug fix:
af48b1005639d3cb1dea4e08dd9d2e33a2289a95
in trond's change he silently removed the need for the -d flag to create the
pidfile, but nothing deletes it on exit. You'll also never see the errors from
it.
I may punt on this for 1.4.12 since it might be necessary to do a little more
refactoring to get this "correct": open the pidfile pre-fork, write any error
about trying to open it. fork, then listen or die?
Anyway, will fix it. maybe just not right now.
Original comment by dorma...@rydia.net
on 12 Jan 2012 at 1:23
Pinging back here...
The reason why a local pidfile doesn't work with daemonize is because it
chdir's to / after fork'ing, unless you specify a nonzero maxcore.
So giving it a full path will still work... and -P already works without -d.
I'm trying to convince myself to not punt for another release, but I might.
Doing this correctly is a fair amount of rewiring and we're about to look at
the 1.6 tree again, so I'd like to avoid writing even more core code to port.
IE: We'd have to keep the parent process around listening on a pipe and shuttle
errors/commands around, which means we can't call perror anymore, blah blah.
Original comment by dorma...@rydia.net
on 30 Jul 2012 at 1:03
HELLO WORLD
Original comment by shaoyong...@gmail.com
on 5 Jan 2015 at 6:59
closing old issues. I think we have a general TODO item for keeping the parent
alive until the child is done setting up.
Original comment by dorma...@rydia.net
on 5 Jul 2015 at 1:40
Original issue reported on code.google.com by
br...@tangent.org
on 21 Dec 2011 at 2:55