Source of corruption

[J. Landman Gay]

On 3/24/12 8:11 PM, dunbarx at aol.com wrote:

> I am aware of the ease in which one can accidentally change the name
> of a stack in the inspector when you think you are typing somewhere
> else. This is a bit insidious. But this was the pulldown to set the
> mainstack. Though this property is settable by command, I am positive
> I never did this.

I should have been more clear that the mixup doesn't just happen with 
the message box (which is why I ended up changing standalone settings 
for the wrong stack.) It also happens with the property inspector or any 
other IDE component that needs to look for a topstack. In my case, the 
working stack was modeless and my utility stack was invisible but 
toplevel. The utility stack had precedence in this case and not only the 
standalone settings dialog, but also all property inspectors, were 
targeting my utility stack. I usually noticed the wrong name in the 
inspector but breezed right past it in the standalone settings.

If at any point your wonky stack was displayed at a lower mode than 
another stack, it would be the target stack for IDE tools. If you were 
trying to set the mainstack on a modeless or palette stack, and didn't 
notice your "corrupted" toplevel stack was the target, it would get its 
mainstack changed.

That may not be what happened in your case, but it's been biting me a 
lot lately. I've been working on a modeless stack and I always have my 
utility stack open too. I have to keep remembering to make my modeless 
stack toplevel.


> Why this spurious mainstack reference makes all that trouble, is a
> mystery to me.

I don't know why either, but it makes a sort of sense. If a piece of the 
message path is missing (the parent stack) then who knows what could 
short-circuit in the engine.

-- 
Jacqueline Landman Gay         |     jacque at hyperactivesw.com
HyperActive Software           |     http://www.hyperactivesw.com