Hello,

I use a Ibook 12" on os x with Pd-0.39.2-extended-RC1-macosx104-powerpc disrtibution.

I try to make a patch to use the output datas from the great app called "motion".
You can find it here : http://blogs.23.nu/c0re/stories/7894/
It show the x, y, z values of the apple motion sensor on the ibooks or macbooks.

So, i tried to use the "shell" object included in the Pd-0.39.2-extended-RC1-macosx104-powerpc disrtibution with this patch :

#N canvas 214 92 345 511 10;
#X obj 30 165 shell;
#X msg 13 6 1;
#X msg 49 5 0;
#X obj 32 46 metro 1000;
#X obj 31 200 unpack 0 0 0;
#X floatatom 20 237 5 0 0 0 - - -;
#X floatatom 70 237 5 0 0 0 - - -;
#X floatatom 114 237 5 0 0 0 - - -;
#X msg 31 101 ~/Desktop/motion/motion ibook;
#X connect 0 0 4 0;
#X connect 1 0 3 0;
#X connect 2 0 3 0;
#X connect 3 0 8 0;
#X connect 4 0 5 0;
#X connect 4 1 6 0;
#X connect 4 2 7 0;
#X connect 8 0 0 0;

As you can see, i have the "motion" app on my desktop.

HERE comes the problem : when i start the patch, this error message appear :
"old process still running". The patch do the job : i have my x, y, z ! great !
BUT suddenly my os x session get out and i am in front of the login window !

After this crash, i try same thing but i check the activity doing a "top" on terminal.app. i saw each time, the "crashdump" processus doing something and crash the session.

I saw a start for an answer here : http://lists.puredata.info/pipermail/pd-list/2005-11/032614.html
It seems to be a bug from the "shell" object so is there any alternative to using it or can it be fixed ?

Thanks in advance !!

it might be that top is an open ended process. [shell] requires calls that return 0; I haven't disected top, but I have a feeling you could flag the process to return the data you wanted, and then end.

I'd probably >man top; but I'm just gonna hit the hay.

I don't think top itself is the problem Sunji, as Alex is just using this to diagnose. The problem is with how [shell] gets used though.

Without knowing a lot about the program it's hard to say, but it looks like you're creating new instances over and over until the process table fills up. Either that or the second call to 'motion' causes a segmentation fault becuase you have some fixed resource that gets stomped on.

[shell] isn't such a good way to call a program repeatedly. The way I have gotten around this before is to create a socket and use [netrecieve] to pick up output.

Another puzzle is how to manage the one onstance of motion that you run. The ideal way is to get a PID and when you're done issue a "kill PID". A more brutal but easier way (if you have no ther programs called 'motion' is to issue a "killall motion".