[PD] Weird Tcl Invalid command name error message (and understanding them)

[Benoît Fortier]

Dear list,
I have to debug and optimize a rather complex patch that was made by a friend. Some parts of the patch are fine, other parts are edgy. Right now, when the dsp is on, the patch seems to be working fine most of the time but there are some specific manipulations that tends to cause tcl Invalid command name error messages in the console. I'm rather good at patching but I lack the under the hood knowledge that some of you seems to have in order to understand exactly what those error messages are about. Although I intuitively solved some of those errors, there are other errors for which I will need help.
First of all, this one is very intriguing (I'm using Pd extended 0.43.4 on a 2014 Mac mini running osx 10.9.) : 
(Tcl) NOM DE COMMANDE INVALIDE : invalid command name ".x929b540.c"    while executing".x929b540.c delete graph9375b40o0"    ("uplevel" body line 2569)    invoked from within"uplevel #0 $cmds_from_pd".
and then a bunch of normal error messages like this one appear in succession :x2257570: no such object
It happens when I try to close the main pd windows of the patch, but only when I've been working on it for a certain time with the dsp on. The main pd windows I'm trying to close actually stays there and freezes. I wish I could be more specific but maybe someone can help already?
Also, I found quite a few nested GOP abstractions in this patch and I noticed that can cause problems (and tcl Invalid command name errors). I searched on forums but I found only rather technical and specific discussions about this issue. Is it recommended to avoid all nested GOP or are there guidelines to follow that makes them work well? Or maybe the problem is caused by something else? I'll post the specific error message that I get the next time is shows up, I havent't got it for a while because I removed a lot of the nested GOP abstractions. I remember the "while executing" line was mentioning something about $select_color. 
Generally speaking, I wish I could understand a little bit more what the tcl invalid command name error messages are trying to tell me. What type of errors are they for specifically? A lot of them seems to show up because of bad patching.... how can I tell that? Is there a way to quickly narrow down what I should be looking for in the patch to solve the problem? How can I interpret those messages in order to help me try to intuitively solve those error? Which of those error message can I safely ignore if any? Sorry, so many questions... I guess you get the Idea. 
For your curiosity it's a pitch tracking device that will be used soon in concert with a string quartet in Quebec City. The patch will be used to "enhance" the sound of each instrument.  Many thanks in advance.

Benoît Fortier

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20141023/da085f53/attachment.html>