Thanks for the input. If the error is occurring in the function named
in the crash log, it looks like I'm writing to the MemoryBlock safely.
The function creates a new MemoryBlock of size X (depending on the
length of the string it's going to store) and then loops over the
MemoryBlock setting byte values, checking that it's writing it to byte
< x each step of the way. When it finishes it returns
mb.StringValue(0, x).

If this is happening somewhere else in the application, I'm going to
have to go on a real fishing expedition.

I realized there's another twist I didn't mention in my original
message. This beta tester has been using an earlier production version
of this application which contains the exact same code, for at least a
couple of years. I moved from RB 5.5.5 for that version to 2007R2 for
this new version of my application. I wonder if that is a factor.

On 5/16/07, M Pulis <[EMAIL PROTECTED]> wrote:
> This crash appears as if osx choked trying to log an error (levels
> 0-3) while trying to free  a block (4).
>
> Your guess is likely close. The block being released is either
> damaged (the memory manager pointers trashed) or was already released.
>
> Look carefully at the use of memory blocks. Writing beyond the end of
> a block can destroy an adjacent, unrelated, innocent block. We call
> this "lighting the fuse" because the system will continue to run and
> all looks good until the system frees the block that was destroyed,
> then - BANG! Very, very nasty bug to squash because the actual crash
> happens at some time later than when the overwrite bug happened. If
> you have global memory blocks, the crash may not happen until quit.
>
> Check his system and console logs; these may contain other messages
> that did make it to the log.
>
> Ask these questions: How is the one user different from the others?
> Larger files? Runs for hours vs minutes? Uses a combination of
> features others haven't? Plays games/music/videos in the background?
> ad nauseum,.....  the point is, something is different....
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>

Search the archives:
<http://support.realsoftware.com/listarchives/lists.html>

Reply via email to