Mailing-List: contact cygwin-help@sourceware.cygnus.com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin@sources.redhat.com>
List-Help: <mailto:cygwin-help@sources.redhat.com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner@sources.redhat.com
Delivered-To: mailing list cygwin@sources.redhat.com
Message-ID: <3A9845AB.E206F0C6@ece.gatech.edu>
Date: Sat, 24 Feb 2001 18:37:15 -0500
From: "Charles S. Wilson" <cwilson@ece.gatech.edu>
X-Mailer: Mozilla 4.75 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Robert Collins <robert.collins@itdomain.com.au>
CC: madhu@quickmonkey.com, cgf@redhat.com, cygwin@cygwin.com
Subject: Re: cygwin with sockscap32
References: <200102222041.f1MKfKj29110@quickmonkey.com> <20010224164002.B6385@redhat.com> <200102242149.f1OLns802613@quickmonkey.com> <00b601c09eb4$e3cc1ac0$0200a8c0@lifelesswks>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit



Robert Collins wrote:
> 
> Madhu,

> I have seen many occasions where software vendors have to release new
> versions of their product when an O/S patch occurs because _they broke
> the rules writing it_. Cygwin 1.1.x has the same ABI as cygwin b20. Most
> ports for B20 run just fine under the current cygwin because cygwin has
> been carefully kept backwarsds compatible. 

Not entirely true.  executables would work on both B20 and v1.1.x, but
objects and libraries had to be recompiled.

> Occams razor suggests that
> this is just another case of a corner cutting software vendor. 

Yep -- they "make assumptions about the application and stack
implementations".  It seems these assumptions were valid for B20
cygwin1.dll and B20 executables, but not for V1.1.x cygwin1.dll and
V1.1.x executables.

> The
> sockscap made use of an unsupported API or ABI feature, and as such is
> now broken.

Not really.  sockscap doesn't link to cygwin, nor does it run under the
cygwin "platform".  It modifies the behavior of the windows networking
stack -- in ways that are apparently incompatible with the cygwin-1.1.x
networking implementation.

> 
> Of course, it might be a cygwin problem, in which case...
> 
> YOU have the cygwin source. YOU are observing the problem in a closed
> source product, YOU need to liase with the software vendor.

Yep.

--Chuck

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

