Mail Archives: djgpp/1999/10/21/00:58:25
"Peter J. Farley III" wrote:
>
> Hello all,
>
> I'm getting a SIGSEGV in malloc (symified trace pasted below),
> encountered while testing the latest snapshot of Awka. I think it's
> probably a bug somewhere in Awka, but I need help to track it down.
>
> I'm running DJGPP v2.02, gcc 2.8.1 (10-28-98), libc/libg/libm 11-15-98
> in a W95 DOS box, LFN=Y.
>
> Since I kind of doubt there really is a bug in malloc, are there any
> DJGPP memory debuggers available other than MSS and YAMD? I haven't
> had much luck with YAMD (and will be sending Nate Eldridge a few lines
> on my troubles there), and MSS seems to require more source
> manipulation than I wanted to do at this point (though I will if
> that's my only alternative).
There is a package called Fortify, but I can't find it at the moment and
haven't used it enough to know what it's like.
Please do email me regarding YAMD, though.
> I have run the test program with gdb, and tracing up the stack frames
> seems to show that the malloc parameter at the point of failure is
> reasonable, so I'm guessing that memory has been overrun somewhere
> prior to the failure, wrecking malloc's control structures someway.
> Obviously, a memory debugger would help here.
>
> The test program is (of course) one of the more complicated and long
> ones, so a "simple" source is not available. Plus, if the problem is
> in Awka itself (as I suspect), there is the whole Awka library source
> available as potential culprits.
Luckily the traceback is confined to two files, which may simplify
things a bit.
--
Nate Eldredge
neldredge AT hmc DOT edu
- Raw text -