iceweasel issues (update)
- To: Michael Cree <mcree@orcon.net.nz>
- Cc: debian-alpha@lists.debian.org
- Subject: iceweasel issues (update)
- From: Bob Tracy <rct@gherkin.frus.com>
- Date: Thu, 26 Apr 2012 19:09:44 -0500
- Message-id: <[🔎] 20120427000944.GA30710@gherkin.frus.com>
- In-reply-to: <D441E976-83BE-44FD-9EE8-FF6D6CB421EE@orcon.net.nz>
- References: <20120225042843.GB3812@gherkin.frus.com> <20120225050417.GA4191@gherkin.frus.com> <4F486CE4.8090300@orcon.net.nz> <20120225054429.GA4297@gherkin.frus.com> <20120225121614.GA5763@gherkin.frus.com> <4F498B28.2010205@orcon.net.nz> <20120227041350.GA12834@gherkin.frus.com> <D441E976-83BE-44FD-9EE8-FF6D6CB421EE@orcon.net.nz>
On Mon, Feb 27, 2012 at 05:33:34PM +1300, Michael Cree wrote:
> On 27/02/2012, at 5:13 PM, Bob Tracy wrote:
> >On Sun, Feb 26, 2012 at 02:30:16PM +1300, Michael Cree wrote:
> >>I've also just upgraded to iceweasel 10.0.2-1+alpha and have it
> >>running
> >>on the xfce4 desktop. It worked for a while reading a few web
> >>pages, but
> >>has now locked up and it using whatever spare CPU it can take. I can
> >>still use the desktop ok---just a little bit sluggish.
> >>
> >>Yep, there's a problem alright.
> >
> >Just completed monolithic build of 10.0.2 with optimization: same
> >behavior as iceweasel-10.0.2-1. I'll try a build with optimization
> >disabled. Did you get anywhere with the debugger?
>
> Um, yes, I did try it and it always came down to being in one routine,
> but can't remember what it was. Will post later when I can switch on
> the Alpha.
>
> I am hoping that optimisation makes no difference now. We are no
> longer subjected to the nasty optimisation bug in gcc-4.4 and I've
> just got another optimisation bug introduced in gcc-4.6 fixed.
A few moons later, we've got an updated lower bound (release that
works): 9.0.1. Built it with optimization, and it works fine -- no
lockups. The first release version that *doesn't* work is 10.0, so
we're closing in on the problem. All releases >= 10.0 are currently
broken on Alpha.
As far as optimization vs. no optimization, the only difference I
currently note is that the old "nsThreadUtils*" patches are required to
get a non-optimized build to succeed. I have no explanation for this,
but without those patches, a non-optimized build will fail. To say the
least, this is bothersome...
--Bob
Reply to: