From owner-freebsd-gecko@FreeBSD.ORG Mon Nov 30 21:34:40 2009 Return-Path: Delivered-To: freebsd-gecko@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 5BDF11065670; Mon, 30 Nov 2009 21:34:40 +0000 (UTC) (envelope-from beat@FreeBSD.org) Received: from marvin.chruetertee.ch (marvin.chruetertee.ch [217.150.245.55]) by mx1.freebsd.org (Postfix) with ESMTP id F04F08FC08; Mon, 30 Nov 2009 21:34:39 +0000 (UTC) Received: from daedalus.network.local (cafe.interxion.wlan.hotspot.nexellent.net [217.147.222.218]) (authenticated bits=0) by marvin.chruetertee.ch (8.14.3/8.14.3) with ESMTP id nAULYbfi013333 (version=TLSv1/SSLv3 cipher=DHE-DSS-AES256-SHA bits=256 verify=NO); Mon, 30 Nov 2009 21:34:37 GMT (envelope-from beat@FreeBSD.org) Message-ID: <4B143A9D.3040302@FreeBSD.org> Date: Mon, 30 Nov 2009 22:35:25 +0100 From: Beat Gaetzi User-Agent: Thunderbird 2.0.0.23 (X11/20090821) MIME-Version: 1.0 To: Gary Palmer References: <4AF1FE62.9070607@FreeBSD.org> <20091123213740.GB69348@in-addr.com> <20091124090123.GL16834@bsdcrew.de> <4B0BC60D.6010201@freebsd.org> <4B0BD67F.7060406@FreeBSD.org> <20091128202822.GA41724@in-addr.com> <20091128224728.GB41724@in-addr.com> <20091129043919.GC41724@in-addr.com> In-Reply-To: <20091129043919.GC41724@in-addr.com> X-Enigmail-Version: 0.95.7 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: freebsd-gecko@FreeBSD.org Subject: Re: Call for tester/reviewer: SeaMonkey 2.0 X-BeenThere: freebsd-gecko@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Gecko Rendering Engine issues List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 30 Nov 2009 21:34:40 -0000 Hi Gary, Gary Palmer wrote: > On Sat, Nov 28, 2009 at 05:47:28PM -0500, Gary Palmer wrote: >> On Sat, Nov 28, 2009 at 03:28:22PM -0500, Gary Palmer wrote: >>> On Tue, Nov 24, 2009 at 01:50:07PM +0100, Beat Gaetzi wrote: >>>> Gary Palmer wrote: >>>>> Martin Wilke wrote: >>>>>> On Mon, Nov 23, 2009 at 04:37:40PM -0500, Gary Palmer wrote: >>>>>> >>>>>>> On Wed, Nov 04, 2009 at 11:21:22PM +0100, Beat Gaetzi wrote: >>>>>>> >>>>>>>> Hi, >>>>>>>> >>>>>>>> A few days ago SeaMonkey 2.0 has been released. To install SeaMonkey >>>>>>>> 2.0 >>>>>>>> a modified bsd.gecko.mk is needed. Therefore we are not able to commit >>>>>>>> it to the ports tree until the tree is unfreezed. So now we are looking >>>>>>>> for tester/reviewer of the SeaMonkey 2.0 port. >>>>>>>> >>>>>>>> The port is available in our SVN repository: >>>>>>>> # svn co >>>>>>>> http://trillian.chruetertee.ch/svn/freebsd-gecko/branches/experimental/www/seamonkey-devel >>>>>>>> >>>>>>>> >>>>>>>> The modified bsd.gecko.mk could be downloaded here: >>>>>>>> http://trillian.chruetertee.ch/svn/freebsd-gecko/branches/experimental/Mk/bsd.gecko.mk >>>>>>>> >>>>>>>> >>>>>>>> A screenshot of SeaMonkey 2.0 running on FreeBSD is available here: >>>>>>>> http://tmp.chruetertee.ch/seamonkey20.png >>>>>>>> >>>>>>>> Many thanks to Florian Smeets, Andreas Tobler and miwi@ for their work. >>>>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> I installed the test port on my FreeBSD 6.4 desktop (don't ask) and >>>>>>> while >>>>>>> initial indications looked positive, the browser reproducibly crashed >>>>>>> when accessing: >>>>>>> >>>>>>> - http://www.slashdot.org/ >>>>>>> - http://www.bhphotovideo.com/ >>>>>>> >>>>>>> The former I don't care about. The latter I do. I don't entirely trust >>>>>>> the backtraces, but one crash seemed to be in js_ConcatStrings in >>>>>>> libmozjs.so. >>>>>>> I tried enabling debugging symbols in the 'make config' section and >>>>>>> rebuilding, but the libmozjs.so library was still stripped before >>>>>>> installation so I wasn't able to do more investigation. >>>>>>> >>>>>> >>>>>> Kernel module "sem" was loaded? >>>>>> >>>>> Hi Martin, >>>>> >>>>> Its not in kldstat but apparently it is in my kernel configuration >>>>> >>>>> % kldload sysvsem >>>>> kldload: can't load sysvsem: File exists >>>>> >>>>> options SYSVSHM # SYSV-style shared memory >>>>> options SYSVMSG # SYSV-style message queues >>>>> options SYSVSEM # SYSV-style semaphores >>>>> >>>>> I also ended up clearing out my .mozilla directory (after taking a >>>>> backup) and also the contents of /usr/local/lib/seamonkey, reinstalling >>>>> seamonkey 2.0, and the crashes persisted so I do not believe it was >>>>> something left behind, although I may have missed something. >>>> It looks like the sem kernel module is not available on FreeBSD 6. Could >>>> you please add "options P1003_1B_SEMAPHORES" to your kernel config as >>>> newer firefox and probably also seamonkey versions require POSIX-style >>>> semaphores. >>> Hi, >>> >>> I've recompiled my kernel with P1003_1B_SEMAPHORES, but seamonkey still >>> crashes going to www.bhphotovideo.com. >>> >>> Thanks, >>> >>> Gary >> The backtrace I've been able to get so far is >> >> #0 js_ConcatStrings (cx=0xbfbf9ab0, left=0x4c6dcb20, right=0x0) >> at jsstr.cpp:136 >> #1 0x4a788f24 in ?? () >> #2 0x00000000 in ?? () >> #3 0xbfbf9a94 in ?? () >> #4 0x4a788d9a in ?? () >> #5 0x00000000 in ?? () >> #6 0x00000000 in ?? () >> #7 0x4c5f20e0 in ?? () >> #8 0x4c644800 in ?? () >> #9 0x4c64315c in ?? () >> #10 0x4c45444c in ?? () >> #11 0x4c644800 in ?? () >> #12 0x48ef1949 in fflush () from /lib/libc.so.6 >> #13 0xbfbfcc78 in ?? () >> #14 0x481edf60 in __JCR_LIST__ () from /usr/local/lib/seamonkey/libmozjs.so >> #15 0x4a788d10 in ?? () >> #16 0xbfbf9ab0 in ?? () >> #17 0x48186c8b in js_ExecuteTree (cx=0x8ac2800, f=0x8e4b200, >> inlineCallCount=@0x4c6430e0, innermostNestedGuardp=0x0) >> at jstracer.cpp:4580 >> #18 0x4819cec5 in js_MonitorLoopEdge (cx=0x8ac2800, >> inlineCallCount=@0xbfbfd0b4) at jstracer.cpp:4936 >> #19 0x4810d99f in js_Interpret (cx=0x8ac2800) at jsinterp.cpp:3875 >> #20 0x4811232f in js_Execute (cx=0x8ac2800, chain=0x4c6c6de0, >> script=0x8e07a80, down=0x0, flags=0, result=0x0) at jsinterp.cpp:1622 >> #21 0x480b2a8e in JS_EvaluateUCScriptForPrincipals (cx=0x8ac2800, >> obj=0x4c6c6de0, principals=0x8d2e884, chars=0x8ef9408, length=120, >> filename=0x8d2a208 "http://www.bhphotovideo.com/", lineno=54, rval=0x0) >> at jsapi.cpp:5145 >> >> If you notice the 'right' parameter to js_ConcatStrings is 0, which is >> what is causing the crash. At this point I'm wondering if its a compiler >> issue. I notice FreeBSD 7 has gcc 4.2.1 while FreeBSD 6.4 has gcc 3.4.6. >> Its just a hunch. Not sure what else I can do to debug the problem >> as the stack backtrace appears ... odd. > > Just FYI, I installed a FreeBSD 6.4 VM and built seamonkey-devel from > scratch in the VM with a clean ports tree and it crashes there > when I go to bhphotovideo.com also, so I suspect its not something > specific to my desktop. The backtrace is a little different > but js_ConcatStrings is still in there. Thanks for all your tests. I setup a box with FeeeBSD 6.4 and was able to reproduce the problem. It really looks like a compiler issue. To test this I build seamonkey 2 with GCC 4.2.5 and the browser didn't crash anymore when visiting www.bhphotovideo.com. Setting USE_GCC= 4.2+ in the Makefile isn't the best solution (it also requires another patch to fix a build issue on FreeBSD 6 with gcc4.2) so maybe we will find a better solution for this. Beat