Are you using exit 0 at the end of your shell script (assuming it is a script)?
Ive also encountered some funny things with [shell] but in a bit more convoluted scenario and am curious as to what everyones thoughts are. Namely, I am using a shell script to perfectly pair wiimotes, so a [shell] object spits out a message connect <MAC ADDRESS> that is sent forward to disis_wiimote object (this component of the external is very similar to regular wiimote object except for the fact that disis_wiimote is multithreaded). This works *perfectly* when using regular pd-l2ork. But when running pd-l2ork through gdb (in hope of hunting down those sporadic and elusive segfaults), the connection fails whenever the message is passed from shell. If it is passed from a regular message inside gdb it works fine but when it is coming out of [shell] it fails (by fails I mean the connection immediately fails with the external reporting inability to find a wiimote, while pd-l2ork remains stable and operational). Valgrind does not have this problem but it in turn kills jackd due to incredibly slow start-up...
Any thoughts?
From: pd-list-bounces@iem.at [mailto:pd-list-bounces@iem.at] On Behalf Of michael noble Sent: Wednesday, November 07, 2012 9:34 PM To: PD send List Subject: [PD] Issue with [shell] and jack_connect/disconnect
hi all,
I'm trying to use the [shell] object to automate JACK connections under Linux by calling jack_connect or jack_disconnect. While this works at first, it seems that the ultimate result is that pd spawns a number of processes that remain in the background, and very soon jack_connect simply stops responding. Has anyone experienced a similar issue, or can anyone suggest an alternative to [shell] for this purpose?
thanks
Michael