Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Date: Wed, 12 Nov 2003 00:02:13 -0600
From: Bill Priest <bpriest AT verizon DOT net>
To: jbuehler AT hekimian DOT com
Cc: cygwin AT cygwin DOT com
Subject: Re: Building gcc on cygwin w/ Herman Ten Brugge's bounds checking
 patch
Message-Id: <20031112000213.034deb91.bpriest@verizon.net>
In-Reply-To: <3FAFCAC6.5090601@hekimian.com>
References: <20031108144026 DOT 41321 DOT qmail AT web20807 DOT mail DOT yahoo DOT com>
	<3FAFCAC6 DOT 5090601 AT hekimian DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Authentication-Info: Submitted using SMTP AUTH at out005.verizon.net from [4.47.210.70] at Wed, 12 Nov 2003 00:01:49 -0600

On Mon, 10 Nov 2003 12:28:38 -0500
"Joe Buehler" <jbuehler AT hekimian DOT com> wrote:

> I had downloaded this but not tried it yet.  If you get it
> working it would be "interesting" to recompile Cygwin and
> all of its packages and see what happens.  We have most of
> the core dumps out of our local Cygwin setup but there are
> still some happening from time to time...

Ok. I've got it working for trivial (hello world) type programs
that don't have errors; and it reports errors then seg faults
for a program w/ an array bounds error; similar operation vs.
the Linux version of the same program.

The main problem was that it was checking "char **environ"
which from limited googling appears to be defined in
cygwin1.dll; which explains the "funny" message I was getting
from gdb wrt "ptype environ".

This was w/ 3.3.2 (FSF) w/ the appropriate bounds check patch from
Herman's site, w/ a couple of tweaks that I made.  I don't
anticipate it being difficult to make the "same" changes to the
cygwin version of 3.3.1 and provide a working patch; but we
will see.

For the impatient
change libgcc2.c so that __bounds_checking_on declaration is outside
of #ifndef __CYGWIN__ (keeping #ifdef L__main so there is only one of these.

change init.c so that for (i = 0; environ[i]; ++i)  block is commented out
could change #ifndef __vxworks to #ifndef __CYGWIN__ or equivalent.

This is all I can remember changing off the top of my head.

Of course this checking only works for "C" & not "C++" or any other language
AFAIK.

FYI,

Bill
PS.  I'm sure it should be possible to process the environ; but I'd need so
direction from cygwin experts or a pointer to an example in existing source
code.


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/