Date: Sun, 25 Jul 1999 11:06:32 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Richard Dawe <richdawe AT bigfoot DOT com>
cc: DJGPP newsgroup <djgpp AT delorie DOT com>
Subject: Re: Bizarre debugging format problem
In-Reply-To: <37971021.D7D5CA69@tudor21.net>
Message-ID: <Pine.SUN.3.91.990725110455.27488E-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


On Thu, 22 Jul 1999, Richard Dawe wrote:

> Eli Zaretskii wrote:
> > Well, I never upgraded to 2.8.1, since it has so many bugs.
> 
> Is there a web page that lists these bugs?

I'm not aware of such a page, but I've seen my share of problems with
code compiled by 2.8.1 which eventually turned out to be some too wize
optimizations.  And when people who work on EGCS tell that 2.8.1 is
``hopelessly buggy'', that's a clear sign to run for your life...

Most of the blatant bugs are in the C++ compiler, btw.

> I think maybe I will downgrade to gcc 2.7.2.1.

Or maybe upgrade to EGCS.

> Exiting due to signal SIGINT
> Control-Break Pressed at eip=00010ce6
> eax=00000000 ebx=00000021 ecx=00000000 edx=0000000e esi=00000054
> edi=000643ac
> ebp=0006439c esp=00064390 program=D:\DEVELOP\TEST\DBGINFO\TESTIT
> cs: sel=00ef  base=83231000  limit=0007ffff
> ds: sel=00f7  base=83231000  limit=0007ffff
> es: sel=00ff  base=83231000  limit=0007ffff
> fs: sel=0117  base=00015bf0  limit=000028ff
> gs: sel=0107  base=00000000  limit=ffffffff
> ss: sel=00f7  base=83231000  limit=0007ffff
> 
> Call frame traceback EIPs:
>   0x00010ce6   ___dpmi_int+114
>   0x0000da99   _getch+49
>   0x00001de1   _main+29
>   0x0000d0aa   ___crt1_startup+138

This seems to indicate that there's a real problem with line-number
info in this program: symify didn't print them as well.  Could you
please make this binary available for downloading, or send it to me?
I want to look into it some more.

This traceback also shows that you still use v2.01 library.  Why?

> BTW is the result of running symify a symification?

You mean, as opposed to symifization?

> > I can let let you download the latest binary of GDB 4.18 that
> > I'm using, to see whether it does any better; or I can try it myself
> > if you tell me where can I download the binary of your terst program.
> 
> I think I'd prefer to download a new version of gdb, because I'd like to
> use command-line editing, etc. Also, the test progs are rather large
> (0.5MB for testit.exe).

I can make the gdb.exe binary available for anon ftp, if you want.
This is not the latest version, so signal support doesn't work 100%;
but it is better than 4.16 even in this regard.  And there are lots of
other improvements.

As for the test program, it seems that the modified program, without
libsocket, also exhibits the problem, and it is probably much
smaller.  So perhaps you could send it to me.

Btw, how did you compile the program (with what switches)?

> It seems the problem isn't with gdb anyway, from what you say about line
> numbers.

Yep, it seems so.  Weird...