On Tue, 1 Aug 2006 17:51:55 -0400, Jeff Kell <[log in to unmask]> wrote:

>Steve Cooper wrote:
>> This is actually quite common.  My guess is that there is an 
uninitialized
>> variable or perhaps a parameter which is thought to be longer than it
>> actually is.  When you do the display first, different junk is left on 
the stack,
>> thereby "initializing" your unitialized variable to a value that does not
>> crash the program.
>Yes, and having a DISPLAY in the CM code versus not having it causes a
>sizeable fixed buffer to be allocated for the DISPLAY.  This shuffles
>the relative locations of your variables in the stack, not to mention
>making it larger (I've seen DISPLAYs "fix" bounds violations on occasion).
>
>Jeff
>
I didn't know the reason, but I am also familiar with this result.
Specially the "fixing" of bounds violation. 
Usually was an non-initialized field or the linkage-buffer didn't match.

Michael

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *