Re: Segmentation fault

From: George Greer (greerga@circlemud.org)
Date: 09/14/01


On Fri, 14 Sep 2001, Gerald Richter wrote:

>George Greer wrote:
>> That is a typical symptom of random memory corruption.  The differing
>> compiler and/or library can cause subtle changes in what gets trashed.  It
>> could also just be the compiler miscompiling your code but that's not
>> likely unless you're pushing its capabilities.
>
>Heh. As I said, it's a redhat quirk.

I meant YOUR code is likely broken.  The different code alignment between
the compilers (due to optimization differences not brokenness) would make
your code overwrite your own memory in your own random places for each.
When the compiler takes liberties with C code that it is allowed to, that's
not its fault.

Most of what is RH7's gcc-2.96 became gcc-3. Not that I've checked
CircleMUD for -fstrict-aliasing compliance...but that's on by default in
gcc3 not gcc2.96.

--
George Greer
greerga@circlemud.org

--
   +---------------------------------------------------------------+
   | FAQ: http://qsilver.queensu.ca/~fletchra/Circle/list-faq.html |
   | Archives: http://post.queensu.ca/listserv/wwwarch/circle.html |
   +---------------------------------------------------------------+



This archive was generated by hypermail 2b30 : 12/06/01 PST