Hello list,
i'm having trouble getting rid of stack overflow issue. The best object to solve it happens to be [ignore], (change is not going to help int his particular situation).
However when i connect a [+] object to ignore, then add some stuff up, set the ignore to a reasonable timing, say 100 ms, the new value isnt let thru by ignore, even though it is being presented to ignore for a lot longer then 100 ms. Now if i connect a number box directly to the ignore, it works just fine. And if i connect a numberbox directly to the [+] then that works just fine aswell. So somehow, i guess, the results of the equation arent being passed on to the ignore for a very long time.
So, im looking for suggestions to solve this problem as its been frustrating me for nearly a week now.
Thanks.
Anton
Hi Anton, i don't know if it's related to your problem, but i was having troubles with stack overflows as well. To get rid of it i introduced a different stack overflow detection in the devel_0_37 cvs version a while ago. I really cant't imagine how the original code can actually be working across several platforms. You should probably try if that cvs version works for you as well.
best greetings, Thomas
----- Original Message ----- From: "Anton Woldhek" woldhek@xs4all.nl To: PD-list@iem.at Sent: Wednesday, February 25, 2004 3:34 PM Subject: [PD] problem with [ignore]
Hello list,
i'm having trouble getting rid of stack overflow issue. The best object to solve it happens to be [ignore], (change is not going to help int his particular situation).
However when i connect a [+] object to ignore, then add some stuff up, set the ignore to a reasonable timing, say 100 ms, the new value isnt let thru by ignore, even though it is being presented to ignore for a lot longer
then
100 ms. Now if i connect a number box directly to the ignore, it works just fine. And if i connect a numberbox directly to the [+] then that works just fine aswell. So somehow, i guess, the results of the equation arent being passed on to the ignore for a very long time.
So, im looking for suggestions to solve this problem as its been
frustrating
me for nearly a week now.
Thanks.
Anton
PD-list mailing list PD-list@iem.at http://iem.at/cgi-bin/mailman/listinfo/pd-list
Hi Thomas,
i don't know if it's related to your problem, but i was having troubles
with
stack overflows as well. You should probably try if that cvs version works for you as well.
Actually I think my stack overflow is pretty legit, there is an intentional loop in my patch. Basically i have a process which involves 3 different models, of which 2 have mutiple instances. (8 eachto be exact) These need to check how many of their brothers and sisters are "alive" and how many of the other models are alive *and* if they themselves are currently alive. So this introduces a feedback loop, whcih should be solveable (in my understanding with a combined use of ignore and change. (If only change is used when one model decides to be alive, some other model might decide at roughly the same time, then they both decide again that it was a mistake and instantly die again, etc. etc.) Ive allread partly solved this by letting each model check their current state only at a random time (for each model). However that doesnt solve it completly. In order for it to work, a change in status must be stabil for something like 10 ms before it is passed on to the model itself *and* the other models.
If this is very confusing i could attach or upload the whole patch somewhere.
Cheers Anton ps I dont remember from which library the ignore object was unfortunatly
Cheers Anton ps I dont remember from which library the ignore object was unfortunatly
The answer to this should be Maxlib (thank you for pdb). http://www.akustische-kunst.de/puredata/
Cheers Anton
-
Anton Woldhek -
Thomas Grill