Hello to everybody,
I have this problem: I am programming a "big" piece of program - reactable simulation, and use a lot of patches and libraries. So i've reached to a point that even in editing mode a simple "copy-paste" or a "delete" crashes down the pc and forses pd to close automatically without saving anything. Then i restart the program and then again the same problem.
I' ve been watching CPU usage from [dsp] and it's average around 8% when simple editing, 14% ~ 18% with dsp on or reactivision on, and around 23% with both dsp and reactivision on, But the problem occurs even in simple editing mode.
I have a MacBook 2.1 - Mac OS X 10.6.8 with 2GB ram at 667 MHz, 2GHz Intel Core Duo Processor.
Does anybody knows why is this happening and how i can face this problem?
-
Pd - mac crash
-
I have read that also pd crusses in mac situations. Did anybody had these kind of problems by using pd in linux? Cause one idea is to set a pc in linux and run pd.
A second thought is to make my patches a library but i don't know how and if it will resolve the problem... -
get rid of ALL gui objects that are not absolutely necessary. Generally when pd is crashing without warning, that is the cause.
number boxes, graph-on-parent patches, anything like that...delete them unless the are 100% necessary for the patch to run.
if you still have problems, start looking through your externals. Try deleting them and see which one causes your patch to crash.
If you get further than that, then make sure that you don't have messages being sent to [vline~] while dsp is turned off.
Bug tracking is a process of subtraction. Make a copy of your patch, and then gradually take away parts of it, saving the new copy, until you find which part is causing your trouble.
my money is on the GUI objects, if you're using them. And then if it's not that, then perhaps some glitch in the coding of an external.
-
Thanx Mod
As far by now, i did what you said and i thinkthe problem is in [expr if()] atom which has many "if-s". Every patch that includes this thing or connects to another patch or subpatch via inlet - outlet and etc. that includes that, causes thiw problem by now.
I have replaced it with [#expr...] of GridFlow library and replaced or deleted all gui objects, but i have not checked it yet, because i cannot delete old [expr if(..)]. And even though i have disconnected all of then, the problem still exists because the code is been read by the system.
Just think that if i delete the whole patch, pd crashes down, if i delete only the [expr if()] atom and save it. Console will replay me that is been saved but on my next anything on the patch, pd will crash and turned off, as privious and nothing will have been saved from before...
So 2 questions. How i fix this problem in order to get rid off the false [expr], and second, Toggle is been refered as gui object. Do i have to replace it with some piece of code or toggles and bangs is ok ([bang] is replaced with message [bang( , if this really matters.. -
I think I remember reading on the list a while ago that [expr] does run into problems if there are two many nested if()s in it. If [#expr] is working fine, you could try opening the patch in a text editor and doing a search-and-replace to change all the [expr]s to [#expr].
I've attached an abstraction that I think behaves the same way [tgl] does without the gui stuff.
