odd stack corruption

[Jeff Reynolds]

ok, this is one of the few times i have ever run into stack corruption  
in all my years with hypercard, metacard, rev, livecode...

I have a application for an exhibit that started way back in the early  
metacard days thats been running and evolving for years now. this last  
update i also reved the stack up to 461 as well with a major upgrade  
to one of the exhibit. then i found a really odd situation with one  
card where two groups were interacting oddly. a click in one strange  
area on one group was causing the other hidden group to be show. there  
were no buttons or scripts that could get fired in that area to do  
this show at all. it was also an odd rectangle that lined up with  
nothing in the interact or would show up in the application browser as  
some old thing left behind.

then i tried to go to edit the group that was causing the other to be  
show and in edit mode in select objects or select grouped if you  
clicked on any element of the group it would not select it, but it  
would behave as if in run mode. i could select the group or any of its  
objects from the application browser. i tired ungrouping and then  
regrouping, but any regrouping combination i tried with various parts  
of the old group would create a new group that would instantly be  
given the old group name and behavior. finally i just reconstructed  
the group from newly created objects and that new group behaved  
normally. when i went to delete the old ungrouped objects of the old  
group i found they all behaved as normal objects except one that  
behaved like the old group, not being able to select it with the  
pointer and it still in run mode when the when i was in authoring mode.

once all the old stuff was gone and the new group there and working  
everything seem to then work fine. i went back through old versions  
over the last 7 or so years in earlier versions and i found the  
corruption seems to have been there for a long time with the non  
editable nature, but the odd showing of the other group on clicking in  
one odd area only comes in when the stack moves to 461.

question is should  i be worried about the rest of the stack? has  
anyone seen behavior like this before? should i just keep letting it  
solder on until something else breaks?

this application has been a trooper since the mid 90s running the  
auditorium presentations at the monterey bay aquarium. its grown from  
one exhibit on deep sea exploration to 3 over the years and worked  
with all sorts of equipment, never failing. want to keep this moving  
into the future, but am loathe to recreate the whole software as it  
represents a large amount of development time over the years and also  
runs very specific equipment that is hard to test if im not on site.

suggestions welcome.

thanks

jeff