From: pavenis AT lanet DOT lv
Message-ID: <B0000097379@stargate.astr.lu.lv>
To: Michel de Ruiter <Michel AT smr DOT nl>, djgpp-workers AT delorie DOT com
Date: Fri, 6 Aug 1999 16:49:36 +0300
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: Gcc crash
In-reply-to: <B0FEA00E82A7D1118BFB00A0CC99027821336A@ARGON>
X-mailer: Pegasus Mail for Win32 (v3.11)
Reply-To: djgpp-workers AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

At first You can try DJGPP port of gcc-2.95 release instead of 
snapshot (timestamp Jul 15 01:09 of Your binary). I didn't keep 
binaries of all snapshots I built. 

As this problem doesn't repeat after rebuilding perhaps posting 
preprocessed source will not help much

Andris

On 6 Aug 99, at 11:59, Michel de Ruiter wrote:

> Hi, workers,
> 
> My GCC 2.95 crashed again. I tested all my memory very thoroughly,
> under DOS and Linux and irrespective of both using memtest (?), so
> that is not the problem. It might be my CPU or motherboard, I'm not
> sure how to test those. When I try to compiling Emacs-20.4, the
> following happens:
> 
> ls -l j:/lib/gcc-lib/djgpp/2.95/cc1.exe
> -rwxr-xr-x   1 mdruiter root      1861026 Jul 15 01:09
> j:/lib/gcc-lib/djgpp/2.95/cc1.exe
> 
> make install
> ...
> gcc -c  -Demacs -DHAVE_CONFIG_H   -I. -I.             -O2 -gcoff w16select.c
> Exiting due to signal SIGSEGV
> General Protection Fault at eip=000c0ba2
> eax=011a7140 ebx=003212cc ecx=00000061 edx=00000006 esi=00469c8c
> edi=00469c8c
> ebp=003212bc esp=00321274 program=j:/lib/gcc-lib/djgpp/2.95/cc1.exe
> cs: sel=0167  base=10290000  limit=0047ffff
> ds: sel=016f  base=10290000  limit=0047ffff
> es: sel=016f  base=10290000  limit=0047ffff
> fs: sel=014f  base=00011b20  limit=0000ffff
> gs: sel=017f  base=00000000  limit=0010ffff
> ss: sel=016f  base=10290000  limit=0047ffff
> App stack: [00323104..001a3104]  Exceptn stack: [001a3010..001a10d0]
> 
> Call frame traceback EIPs:
>   0x000c0ba2 _extended_count+6238
>   0x000b84b9 _combine_instructions+25381
>   0x000b6898 _combine_instructions+18180
>   0x000b67cf _combine_instructions+17979
>   0x000b67cf _combine_instructions+17979
>   0x000b3925 _combine_instructions+6033
>   0x000b24bf _combine_instructions+811
>   0x0000ae03 _rest_of_compilation+2715
>   0x001624a4 _finish_function+188
>   0x0014ed0e _yyparse+3366
>   0x00009ae1 _check_global_declarations+3357
>   0x0000d8f3 _main+4199
>   0x00175504 ___crt1_startup+204
> 
> I do not have the any gcc sources, but I doubt they would help a lot.
> 
> If I type `make w16select.o' after the crash, or just `make', gcc does
> *not* crash on w16select.c any more, and the build succeeds. The crash
> only happens when I run `make' from start. This makes it very hard to
> track down the precise problem.
> 
> If I change the maximum size of virtual memory using cwsparam, the
> crash goes away in this particular case. I did *not* set up cwsdpmi to
> use near 255Mb, having heard about some bugs.
> 
> It must be something particular to my setup. I run MS-DOS 7.10 (Win98
> without GUI) with cwsdpmi-r4.
> 
> If anyone has a vague guess or idea of what might be going on, please
> help.
> 
> Groeten, Michel.
>