Bugs item #2621932, was opened at 2009-02-21 01:44 Message generated for change (Comment added) made by sistisette You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=2621932...

Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: gop multiplicity

Initial Comment:

1. Open patch.pd 2. Delete the subpatch. 3. Either a) create a new my_abs object or b) undo the previous delete 4. Repeat steps 2-3 n times.

Now when you open the subpatch, all actions occur n+1 times, e.g.: * creating a new object box makes a chain of n+1 object boxes * patch is saved n+1 times * paste creates n+1 new copies

If you don't save and try to close patch.pd, it repeats the message "discard changes to patch.pd" and sometimes crashes after two or three times.

The behavior persists until closing pd completely. So if you go through the steps above 3 times, close the patch, and reopen it and go through them 3 more times, you'll get 7 extra objects every time you create something from the "put" menu.

I only get this behavior when using gop, and with an object visible in the gop window. Same for abstractions. For each extra object created in the subpatch, I get a "consistency check failed: canvas_vis" error to the console.

Using pd-0.42-4 on windows xp sp3

-Jonathan jancsika@yahoo.com

----------------------------------------------------------------------

Comment By: Matteo Sisti Sette (sistisette) Date: 2010-03-18 23:52

Message: Do i understand correctly: the last comment mean that the fix was not a fix?

Would you agree on raising the priority of this bug? This is a crasher, and even before the crash it renders Pd unusable _very_ soon when editing gop patches....

----------------------------------------------------------------------

Comment By: Hans-Christoph Steiner (eighthave) Date: 2010-01-18 07:07

Message: I removed commit r12831, canvas_destory_editor() is already being called in the canvas_free() process within canvas_vis(x, 0). I don't know the source of this bug tho, I'll look into it more.

----------------------------------------------------------------------

Comment By: Mathieu Bouchard (matju) Date: 2009-12-13 04:45

Message: i just fixed it in revision 12831, pd-extended/0.42

----------------------------------------------------------------------

Comment By: Mathieu Bouchard (matju) Date: 2009-12-13 04:29

Message: editor_new gets called for a gop, but editor_free does not.

----------------------------------------------------------------------

Comment By: Mathieu Bouchard (matju) Date: 2009-12-13 04:24

Message: t_guiconnect registers a receive-symbol for any canvas but deleting a gop does not unregister the receive-symbol.

----------------------------------------------------------------------

Comment By: Matteo Sisti Sette (sistisette) Date: 2009-12-11 15:21

Message: Oh shit, it only disappears with the reporter's test patch, but I have experimented the bug also with -noautopatch

----------------------------------------------------------------------

Comment By: Matteo Sisti Sette (sistisette) Date: 2009-12-11 15:19

Message: I have tested with -noautopatch, which disables the belowmentioned feature, and the bug disappears. So it is definitely related to that feature, dunnow how.

Hope this helps m.

P.S. can anybody confirm?

----------------------------------------------------------------------

Comment By: Matteo Sisti Sette (sistisette) Date: 2009-12-11 13:26

Message: Hi, This bug seem to have APPEARED ON VERSION 0.42.0 At least, I tested it with the reporter's patch on 0.41.4 and it works fine, while on 0.42.0 the bug shows up 100% of the time.

That is the same version in which the (very annoying) feature was introduced that when you create a new object and another one is selected, a connection is automatically created. May the bug be related to the introduction of that feature?

----------------------------------------------------------------------

Comment By: Matteo Sisti Sette (sistisette) Date: 2009-12-11 12:48

Message: I confirm I can 100% reproduce it on Windows Vista with 0.42-4. Unfortunately I am only a user so I cannot help.

I have also encountered the same problem; my patch is too big and complex to attach it, but in my case I can reproduce the problem by just: A - loading the main patch B - closing it C - loading it again And I get the "mulyiplicity" in the main patch, not any subpatch. But not with your patch.

----------------------------------------------------------------------

Comment By: Jonathan Wilkes (jancsika1) Date: 2009-11-27 01:44

Message: I forgot to say I'm using winxp sp3 with pd vanille 0.42-5 to produce the behavior in the attached patch

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody) Date: 2009-04-03 11:07

Message: In step #3 above, please ignore "a) create a new my_abs object" and simply undo the previous deletion of the subpatch. Also, the multiplicity bug happens only within the subpatch window.

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig) Date: 2009-04-02 16:26

Message: could not reproduce with 0.42-4 on linux

----------------------------------------------------------------------

You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=2621932...