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:	Tue, 13 Jan 2004 12:14:27 -0800 (PST)
From:	"Peter A. Castro" <doctor AT fruitbat DOT org>
To:	cygwin AT cygwin DOT com
Subject:	Re: rebaseall breaks zsh?
In-Reply-To:	<20040112214058.GA24046@redhat.com>
Message-ID:	<Pine.LNX.4.53.0401131209540.1104@gremlin.fruitbat.org>
References:	<Law9-OE64jNIFVoaad5000074d3 AT hotmail DOT com> <20031222151939 DOT GI1560 AT tishler DOT net> <Pine DOT LNX DOT 4 DOT 53 DOT 0312221015110 DOT 7948 AT gremlin DOT fruitbat DOT org> <20040112214058 DOT GA24046 AT redhat DOT com>
MIME-Version:	1.0
X-IsSubscribed:	yes

On Mon, 12 Jan 2004, Christopher Faylor wrote:

> On Mon, Dec 22, 2003 at 10:21:28AM -0800, Peter A. Castro wrote:
> >On Mon, 22 Dec 2003, Jason Tishler wrote:
> >
> >> Rafael,
> >>
> >> Sorry for the delay...
> >>
> >> On Wed, Dec 10, 2003 at 01:26:07AM -0800, Rafael Kitover wrote:
> >> > I noticed that the rebaseall scripts rebases /usr/bin/libzsh-4.0.4.dll
> >> > and the modules in /usr/lib/zsh/4.1.1/zsh/*.dll, and that this breaks
> >> > zsh. Rebasing libzsh stops zsh from starting, and rebasing the modules
> >> > stops them from loading.
> >> >
> >> > If this is the case, and not just something messed up on my system,
> >> > adding:
> >> >
> >> > -e '/\/libzsh-.*\.dll$/d' -e '/\/lib\/zsh/d'
> >> >
> >> > To the sed filter on the script should fix it. Or perhaps making an
> >> > /etc/rebase.conf with exclusion filters. Want me to make a patch?
> >>
> >> Let's wait to see what the zsh maintainer comes up with.
> >
> >I've reproduced the problem under a debugger and it's very interesting.
> >The segfault occurs before main even gets control.  From the looks of it,
> >the cygwin runtime faults trying to resolve some entry point that still
> >points to the old image base address, which is below the rebased image
> >base address (so no wonder it's faulting).  I'm trying to come up with a
> >simple test case, but I'm about to leave for the holidays, so I'll get
> >right back on it after the new year.
>
> Whatever happened here?  I've just noticed this myself so I'm wondering
> if there is a fix (to either zsh or rebase) in the queue.

Sorry for the delay.  I've been on vacation and I'm still getting caught
up at work (being at the tail end of a new release tends to dominate my
time), let along working on anything else.  The last I looked into this
it appeared to be related to an offset which either gcc or the ld is
hiding in some section which isn't being rebased.  Gimme a few more days
to get my head above water and I'll get more details on it.

> cgf

-- 
Peter A. Castro <doctor AT fruitbat DOT org> or <Peter DOT Castro AT oracle DOT com>
	"Cats are just autistic Dogs" -- Dr. Tony Attwood

--
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/

- Raw text -