Date: Tue, 4 Feb 1997 16:19:54 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: SerDevian <scriven AT CS DOT ColoState DOT edu>
cc: djgpp AT delorie DOT com
Subject: Re: fopen(), I think (possible bug?)
In-Reply-To: <32F116E8.2781@CS.ColoState.edu>
Message-ID: <Pine.SUN.3.91.970204161355.35D-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Thu, 30 Jan 1997, SerDevian wrote:

> The crash (when I've been able to trace it) leads into and slightly
> beyond fopen().  However, it more often just goes into an infinite loop
> I can't escape from (that particular program replaces the keyboard
> handler).  In these cases, the crash seems to occur sometimes at fopen()
> an sometimes with findfirst() or findnext().

Either you are overwriting the stack, or you didn't properly lock the 
handler code and the data that it touches.  Lock all the code and data by 
using the appropriate bit of _crt0_startup_flags (documented in the libc 
reference) and see if the crashes go away.  If they do, you should look 
for the code or data that you left unlocked.

> I haven't been able to purposely recreate the problem, though; and I
> have only been able to catch the callback frame a few times.

Run the program through "redir -e file prog ..." and the traceback will 
be written to the named file.  Or press PrtScr when the program crashes.  
The First 4 lines are the most important info, so you might as well just 
copy them from the screen and post here.  Posting a traceback might 
explain what's going on, so do make the effort to save it.