script scope variables inexplicably becoming unset

[Richard Gaskin]

Dr. Hawkins wrote:

> I've stripped the stacks to almost nothing, and still have the
> behavior, reproducibly.  There is almost no code left.

The example stack uploaded there is nearly 1 MB.

I often try to reproduce bugs reported to the RQCC to help them along, 
but with this one I couldn't figure out how to do so.  When I open the 
stack a script editor appears, but apparently not because of a 
breakpoint.  The recipe given was difficult for me to follow; I couldn't 
figure out how to make any palette appear at all.

So having given up on the recipe I was at least curious about how many 
objects there might be in this megabyte stack to estimate the level of 
effort I might need to try to reproduce this issue.

I opened the Project Browser to get a look, and something noteworthy 
immediately stood out:

Group "pp_calcids" of card "entry" of substack "dna" is flagged by the 
Project Browser with a bright yellow warning triangle as having a 
behavior object assigned to it that doesn't exist.

This would seem to fit the description of the problem in which some 
behavior-driven things are working while one doesn't.

Please let us know if simply reassigning that behavior resolves the 
issue.

As for the local vars clearing with the palette command, is there any 
code in play which initializes those vars on one of the open messages 
(openStack, preOpenStack, etc.)?

-- 
  Richard Gaskin
  Fourth World Systems
  LiveCode training and consulting: http://www.fourthworld.com
  Webzine for LiveCode developers: http://www.LiveCodeJournal.com
  Follow me on Twitter: http://twitter.com/FourthWorldSys