From: Martin Str|mberg <ams AT ludd DOT luth DOT se>
Message-Id: <199902181951.UAA05608@father.ludd.luth.se>
Subject: Re: make crash
To: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date: Thu, 18 Feb 1999 20:51:59 +0100 (MET)
Cc: djgpp AT delorie DOT com (DJGPP)
In-Reply-To: <Pine.SUN.3.91.990207114337.15342K-100000@is> from Eli Zaretskii at "Feb 7, 99 11:44:18 am"
X-Mailer: ELM [version 2.4ME+ PL15 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Reply-To: djgpp AT delorie DOT com

According to Eli Zaretskii:
> 
> On 6 Feb 1999, Martin Str|mberg wrote:
> 
> > Ok. I know what is causing the problem. My environment variable SHELL
> > is set to "f:/djgpp/bin/-bash.exe".
> 
> Why would you need that?  AFAIK, -bash is only needed in interactive
> invocation, to make it a login shell, but should never be necessary
> (and could be even dangerous) after that.

Because it's the right thing to do(tm)! Because DOZE has no sense of
logging in, every invocation of the shell must be a log in shell, and
the easiest way to arrange that is having -BASH.EXE point to
BASH.EXE.

Here's a symified traceback:
bash$ make
make all-recursive
make.exe[1]: Entering directory `f:/hackery/tar/gnu/tar-1.12'
g:/dj100000: g:/dj100000: line 2: syntax error: unexpected end of file
Abort!
Exiting due to signal SIGABRT
Raised at eip=0001c792
eax=000b723c ebx=00000120 ecx=00000000 edx=00000000 esi=00004000 edi=00000000
ebp=000b72e8 esp=000b7238 program=F:\HACKERY\LIBC\BIN\MAKE.EXE
cs: sel=0f57  base=838b8000  limit=000dffff
ds: sel=0f5f  base=838b8000  limit=000dffff
es: sel=0f5f  base=838b8000  limit=000dffff
fs: sel=011f  base=00016840  limit=0000ffff
gs: sel=0faf  base=00000000  limit=0010ffff
ss: sel=0f5f  base=838b8000  limit=000dffff
App stack: [000b84f8..000384f8]  Exceptn stack: [00037ea8..00035f68]
 
Call frame traceback EIPs:
  0x0001c6b0   ___djgpp_traceback_exit+48
  0x0001c792   _raise+94
  0x0001f403   _abort+27
  0x00026f22   __put_path2+58
  0x00026ecd   __put_path+13
  0x0002a40e   __chmod+58
  0x00021ddd   _remove+61
  0x0001b51f   _system+699
  0x0000582d   _start_job_command+665, line 985 of job.c
  0x00005982   _start_waiting_job+98, line 1096 of job.c
  0x00005c1f   _new_job+495, line 1273 of job.c
  0x00004b4e   _execute_file_commands+90, line 359 of commands.c
  0x0000dd46   _remake_file+74, line 962 of remake.c
  0x0000d706   _update_file_1+2102, line 652 of remake.c
  0x0000ca67   _update_file+35, line 310 of remake.c
  0x0000c730   _update_goal_chain+316, line 150 of remake.c
  0x00002ef3   _main+3411, line 1549 of main.c
  0x0001a2f2   ___crt1_startup+174
make.exe: *** [all-recursive-am] Error -1
bash$ symify -o /hackery/tar.sym /hackery/libc/bin/make.exe

As far as I could trace it, it's the call to system() that
crashes. How do I step into system with gdb? My usual trick of copying
the .c file and adding it to the makefile doesn't work.


New Model Army, The Love of Hopeless Causes,

							MartinS