Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!zaphod.mps.ohio-state.edu!think.com!linus!linus!sdimax2!jrv
From: jrv@sdimax2.mitre.org (VanZandt)
Newsgroups: comp.os.msdos.programmer
Subject: Re: TD 2.01 bug
Message-ID: <127293@linus.mitre.org>
Date: 14 Dec 90 17:28:30 GMT
References: <1990Dec12.213319.20466@maytag.waterloo.edu>
Sender: usenet@linus.mitre.org
Organization: The MITRE Corp., Bedford, MA
Lines: 21

In article <1990Dec12.213319.20466@maytag.waterloo.edu> dmurdoch@watstat.waterloo.edu (Duncan Murdoch) writes:
>Here's a bug in the new version of Borland's debugger that may have you
>scratching your head for a while.
>
>If you set a conditional breakpoint, TD will pop up a window telling you
>when the condition has been met.  Nice and friendly.
>
>However, if your condition has too many characters in it ...

I tried to set a conditional breakpoint with a long expression and, nearly
as I could figure, the expression was being evaluated incorrectly (triggering
the breakpoint when it shouldn't).  Sounds to me like a buffer overflow that
can manifest itself in more than one way.

Incidently, I recently ran into a floating point divide by zero bug which
would appear only when output was being redirected to a file.  Later, it
crashed the debugger (error window popped up, but further keyboard input
was ignored), forcing a power down reset.  Is Borland's handling of
'87 exceptions up to spec?

                           - Jim Van Zandt