Re: [Jack-Devel] Jack shutdown action

On Sat, Mar 17, 2012 at 5:08 PM, Fons Adriaensen <[hidden]> wrote:
> On Sat, Mar 17, 2012 at 04:47:09PM -0400, Paul Davis wrote:
>> On Sat, Mar 17, 2012 at 4:44 PM, Fons Adriaensen <[hidden]> wrote:
>> > A simple question (reading jack.h doesn't answer it):
>> >
>> > If a client does not register a shutdown callback using
>> > jack_on_shutdown(), what is supposed to happen when the
>> > server is terminated ?
>> >
>> > Apparently the process containing the client is terminated
>> > in some way, but can this be relied on ?
>>
>> the server will not terminate any clients. the thread(s) created by
>> libjack will exit at some point when/if they discover that the server
>> is gone, but that may never happen.
>>
>> the whole point of the shutdown callback is that  the client has no
>> way to differentiate between "the server has gone" and "the server
>> hasn't called me in a very, very long time.
>
> Indeed. I'm asking because I noticed that an app (in development
> and not having a shutdown callback) terminates if the server is
> stopped, even while it has other active threads. None of its code
> watching stop conditions is ever called.  Removing the shutdown
> handler in other apps (e.g. jaaa) leads to the same result. Now I'm
> pretty sure that did not happen before. It looks as if Jack's thread
> is calling exit() or something similar when it terminates... Have
> there been recent changes in that area ?

if you run this command:

   find . -name \*.c -o -name \*.h | xargs grep '[^a-zA-Z]exit'

on the JACK1 source tree, you'll find that the only place that the
libjack code calls exit is immediately after a fork() to start the
server (auto-starting).

all other exits from within libjack are via pthread_exit() which
should not take down a process. running inside gdb() with a breakpoint
at _exit and then a backtrace may prove revealing.

1332018928.19495_0.ltw:2,a <CAFa_cKnS46XvbifGABagsjz4ez9U1y6B6EHyrFTKshW-pEir+A at mail dot gmail dot com>