scope puzzle with menus

[Charles Hartman]

I've got that incremental-search-in-a-sorted-field routine working  
the way I want it (thanks to all the help from this list.) So I  
decided to move it to my mainstack's script to avoid having a dozen  
copies in the scripts for the scrollable fields themselves. In the  
mainstack script there's an incrementalSearch handler with parameters  
'theKey' for the keystroke, 'theField' for the name of the field from  
which it's being called and whose text is being searched (scrolled,  
selected). Each field that wants to use the facility includes a  
little keyUp handler to call it. No problem.

Also in the mainstack script are the script-local variables for the  
seconds of the last keystroke and the accumulating string of  
keystrokes. Perfect use for script-locals (since they're  
nonvolatile); it works fine.

But *other* keystrokes get into the script-local variable too.  
Specifically, if cmd-U has called up the substack whose field is to  
be searched, then the field starts off scrolled to 'U'; a field  
called up by cmd-S is scrolled to 'S'. These are keystrokes bound to  
menu items.

So two questions. (1) Why are they getting into the script-local  
variable?? It's supposed to be accessible only from within the  
script. The menus (built with MenuBuilder) are owned by card 1 of the  
mainstack, but they are not referenced in any way in the mainstack  
script. Is something behind the scenes violating the script-local  
definition? (2) What's the best way around it? I could build a kludge  
filter (maybe, though it would be easier if I knew exactly how those  
keystrokes got in there), but there must be a better way.

Any advice & enlightenment much appreciated as always.

Charles Hartman