Date: Tue, 17 Mar 1998 10:43:42 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Martin Stromberg <Martin DOT Stromberg AT lu DOT erisoft DOT se>
cc: djgpp AT delorie DOT com, spector AT EnchantedLearning DOT com
Subject: Re: Problem with bash
In-Reply-To: <6ejeco$4m8$1@antares.lu.erisoft.se>
Message-ID: <Pine.SUN.3.91.980317104324.17844K-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On 16 Mar 1998, Martin Stromberg wrote:

> I have a bash compiled with -g. If you want I can mail it to you. So you
> can run symify on the tracebacks (which have to be redone, of course).

I'd be surprised if `symify' will say anything useful.  As you can
see, the tracebacks are either bogus or come from the Windows DPMI
server: DS has 0000 as its selector, and SS is invalid.  Disassembly
near 0x000420e5 (that's the EIP where it crashed) might or might not
help (if the crash is inside Windows, that EIP won't show anything
interesting).

Running Bash under a debugger and trying to reproduce the problem
would be the first step.  Personally, I don't trust DOS/Windows
debuggers when it comes to debugging complicated programs which launch
other programs, redirect standard streams etc., let alone mess with
selectors like Bash does, so I won't be surprised if running under a
debugger will prove useless.  But it's worth trying.

If a debugger doesn't help, the good ole `printf' will have to take
charge.