From: mlkessle AT cip DOT physik DOT uni-wuerzburg DOT de (Manuel Kessler)
Newsgroups: comp.os.msdos.djgpp
Subject: GASP GPF's
Date: 7 Oct 1996 15:06:33 GMT
Organization: CipPool der Physikalischen Institute, Uni Wuerzburg
Lines: 62
Message-ID: <53b69p$u4d@winx03.informatik.uni-wuerzburg.de>
NNTP-Posting-Host: wpax07.physik.uni-wuerzburg.de
Summary: GASP sometimes crashes in macros being expanded
Keywords: GASP, GPF
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

Hi,
is there anybody out there really using GASP, the Gnu ASsembler 
Preprocessor? Sometimes when i try to preprocess my .gas files to .s files
i get a GPF when it comes to first macro being expanded. I tracked my rather
longish source file to the minimal file i can imagine to reproduce the 
error: 

>.MACRO DOTEST result
>        movl $\result, %eax
>.ENDM
>
>.TEXT
>        DOTEST 1
>
>.END

The output is then:


<    .TEXT
<
<Exiting due to signal SIGSEGV
<General Protection Fault at eip=00001744
<eax=83e58979 ebx=000527e4 ecx=0005bd40 edx=00067040 esi=00000007 edi=83e58955
<ebp=00052764 esp=00052764 cs=00a7 ds=00af es=00af fs=008f gs=00bf ss=00af
<Call frame traceback EIPs:
<  0x00001744
<  0x00004c9b
<  0x000050f8
<  0x000039ea
<  0x00006783
<  0x0000784b

So far so bad. The real problem is, sometimes i have no problems, sometimes 
it crashes. I tried it on three different machines with very different 
configurations (MS-DOS, Novell DOS, CWSDPMI, CWSDPR0, Wfw 3.11) with the
same results, so i don't expect it's a configuration problem, especially
no other djgpp program has problems. After booting the first few invocations
of gasp are ok, no GPF. After that i start e.g. RHIDE to do some editing,
i shell out to DOS and gasp crashes. I exit, exit RHIDE, gasp crashes. After
the first crash i usually cannot get gasp working whatever i do, at least
with reasonably source files, surprisingly enough the little example above 
sometimes gets no GPF, even when bigger source files fail. If i open a DOS 
window in Windows, i get GPF most of the time.
I've checked the gasp sources from binutils and recompiled gasp with no luck.
I can say the GPF is in sb_reset, which gets a pointer far outside the data
space (ptr is in eax), but i cannot tell why this invalid pointer gets there.
Sorry for not giving more information here, but it is really very annoying
to reboot before every compiling, so i hope to get some infos. I suspect a
not initialized pointer anywhere in gasp.c (sometimes it works, sometimes not,
typical not initalized data error), but under linux i never encountered any
problems ?!?
Any help is appreciated, (and sorry for my bad english), many thanks,
	Manuel 


-- 
Manuel Kessler
Graduate Student at the University of Wuerzburg, Germany, Physics Department
SNAIL: Zeppelinstrasse 5, D-97074 Wuerzburg, Germany
EMAIL: mlkessle AT cip DOT physik DOT uni-wuerzburg DOT de
WWW: http://cip.physik.uni-wuerzburg.de/~mlkessle