Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/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
Message-ID: <41C3453D.90408@kleckner.net>
Date: Fri, 17 Dec 2004 12:44:45 -0800
From: Jim Kleckner <jek-cygwin1 AT kleckner DOT net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040616
MIME-Version: 1.0
To: cygwin AT cygwin DOT com
CC: Bill Priest <priestwilliaml AT yahoo DOT com>
Subject: gcc bounds checking, valgrind, purify
References: <20041213011810 DOT 91265 DOT qmail AT web20824 DOT mail DOT yahoo DOT com>
In-Reply-To: <20041213011810.91265.qmail@web20824.mail.yahoo.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-IsSubscribed: yes
Note-from-DJ: This may be spam


Bill Priest wrote:

>Jim & All,
>   I've successfully built and used cygwin 3.3.1 with
>Herman Ten Brugge's bounds checking patches.  I've
>submitted the trivial changes required for cygwin back
>to Herman.
>   You can google for Herman Ten Brugge and find a
>reference to the changes I made to get them to work
>w/ cygwin.
>   Of course it only works w/ 'C'; but I used it to
>track down array indexing problems and pointer
>problems.
>  
>
Very interesting.  Checking out those links, it appears
that these patches are very much like a static version
of Valgrind.  Valgrind works dynamically with an x86
code rewriter that appears to be language independent
and seems to implement a lot of the difficult functionality
that the original Purify implemented.  Valgrind ought
to work with C++ where gcc-bounds-checking as you
say is only 'C' (though there are some traces in

These patches have been around for 10 years (!).
This kind of test is so valuable, I'm surprised that
they have never been integrated into gcc.  It seems
like it isn't such a big risk to include them and
config them to be off by default.  This would
hugely improve the maintainability and availability
of the patches.

Did your cygwin patches get into bounds checking
patch for the 3.4.2-100 version of 7-Sep-2004 ?
It does look so from the source.

Would it be reasonable integrate this patch into
the cygwin GCC?  I expect back the answer no,
because GCC needs to be synced with the upstream.
But this stuff would be so much more valuable if
(in order of accessibility):
 - It was just part of GCC as an option
 - It could be selectively installed as a package alternative
 - The patches were integrated so that one could download
    and recompile.

There was a discussion here:
  http://www.cygwin.com/ml/cygwin/2003-11/msg00320.html
about the value of recompiling sort of a "super debug"
version of the cygwin dll and libraries with gcc-bounds-checking
that ran slowly but did a lot of checking.  That would indeed
be valuable. 

Jim


PS:  Please don't put my email address in the subject line.
Spam harvesters, you know.


PPS:  Convenience links for fellow travelers:

Good article giving the context and pointers:
 http://www.network-theory.co.uk/articles/boundschecking.html

GCC extensions page:
 http://gcc.gnu.org/extensions.html

Herman Ten Brugge pages with source:
 http://web.inter.nl.net/hcc/Haj.Ten.Brugge/

Alternate (main?) source of the patches:
 http://sourceforge.net/projects/boundschecking/

Valgrind home page:
 http://valgrind.kde.org/



--
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/