Building firefox with clang

Hello! For whatever reasons, both firefox (28) and firefox-esr (24) compiled here using GCC 4.9 kept crashing with SIGBUS -- possibly some sort of misunderstanding between different binaries (firefox and dependencies) compiled with different compilers. My system is a 9.2-STABLE/i386. Unfortunatel

Why use jemalloc bundled with Mozilla?

The system malloc on FreeBSD-10.x is jemalloc-3.4.x. Why do Mozilla ports build with "--enable-jemalloc" turned on despite that? Is it just a typo in bsd.gecko.mk and the fix as simple as: --- /usr/ports/Mk/bsd.gecko.mk (revision 390575) +++ /usr/ports/Mk/bsd.gecko.mk (working copy)

Non-prefix safe ports

Building out a new machine from scratch using /opt (rather than the usual /usr/local) as LOCALBASE, I found the following two ports, which break: * multimedia/openh264 - the files built by the port get installed into stage/usr/local/... instead of stage/opt -- I had to manually move the

enabling INTEGER_SAMPLES seems to break sound

Hi! Fed up with the through-the-roof CPU consumption during video playback, I decided to experiment and rebuilt soundtouch and firefox (and thunderbird) with INTEGER_SAMPLES on. Upon restarting, firefox started to produce garbled sound on various video-containing pages (such as this one). The CPU

Using xulrunner (Re: USE_MOZILLA=yes for firefox and thunderbird)

Picking this up 4 years later... Why don't we simply install xulrunner and make all the applications using it into wrappers? Would not that be terrific? Smaller packages, less run-time memory consumption for millions of people, shorter build-times, no patch-duplications? -mi On 11.07.2012 22:

Is "PGO" still a valid option for firefox?

Hi! I enabled the PGO-option, but the build failed with the below error: gmake[3]: входимо до каталогу <> gmake[3]: *** Нема правила для створення цілі "pgo-profile-run". Зупинка. (no rule to make target pgo-profile-run) Is it even still a valid option, or did Mozilla allow

Firefox' plugin-container crashes on i386

Hello! One of my users has a consistent problem with many sites (including Weather.com and BBC.co.uk) -- firefox disappears immediately or shortly after opening them with its plugin-container leaving a core-dump in his home-directory. Disabling plugin-container by removing executable bit from it

Question about gecko ports

Gentlemen! Why is it, that in order to build the Java-plugin for firefox-49 per the procedure officially documented here: https://www.freebsd.org/doc/handbook/desktop-browsers.html one must download and compile firefox-45 (ESR)? Is this a known bug awaiting a fix, or is it considered normal

Re: Firefox-3.6 can't talk to https sites...

17.03.2010 15:38, Jeremy Messenger написав(ла): On Mon, 01 Mar 2010 23:32:53 -0600, Mikhail T. wrote: Hello! After upgrading to firefox-3.6 (from 3.5.8) I am no longer able to connect to anything over SSL -- the spinner spins forever, but no content is loaded. Other browsers (Konqueror

Re: Firefox-3.6 can't talk to https sites...

17.03.2010 16:54, Jeremy Messenger написав(ла): Default nss from ports tree or your own patch? The default one... The only "outstanding" patches I have now deal with the NSS' self-tests, not with the library code. Same with nspr. -mi ___ freebs

Re: Firefox-3.6 can't talk to https sites...

17.03.2010 16:54, Jeremy Messenger написав(ла): Default nss from ports tree or your own patch? The default one... The only "outstanding" patches I have now deal with the NSS' self-tests, not with the library code. Same with nspr. -mi ___ freebs

Re: Firefox-3.6 can't talk to https sites...

On 17.03.2010 18:43, Jeremy Messenger wrote: Looked in the www/firefox. It looks like www/firefox no longer depends on nss port, which it uses nss in its own tarball? Indeed... Running ldd on both lib/thunderbird-3.0.3/thunderbird-bin and firefox3/firefox-bin shows, they don't link with nss (not

Re: Re: FreeBSD Port: firefox-3.6.3, 1 (trying to load certain pages causes the browser to hang then crash)

-10.01.-28163 14:59, Joey Mingrone написав(ла): Nothing loads at all. It just remains on the current page and says something like "connected to mail.google.com" in the status bar and hangs for a minute or two then crashes/closes. Sometimes it seg. faults and other times it closes, but the proce

mms:// and firefox

Hello! I'm trying to "register" the mms-protocol with firefox (3.6.3), and it is not working :-( I'm following the advice found on: http://kb.mozillazine.org/Register_protocol Unfortunately, neither of the listed methods change anything about firefox' behavior. I'd like to use the KDE's Dra

sqlite3 and other already-installed components

Hello! As always, I'd like to see more of the already installed components used by the Mozilla applications in preference of building their own versions... The patch below adds sqlite3 to the _ALL_DEPENDS list in bsd.gecko.mk and changes mail/thunderbird to build using that (other application

building lightning as part of thunderbird

It turns out, Lightning can already be built together with the rest of thunderbird with the simple --enable-calendar configure-option. The attached patch does that unconditionally, but adding the OPTI

need help with www/helixplugin

Hello! I've just updated the port to build using libxul instead of seamonkey. However, I noticed, that the resulting plugin is not seen in firefox4... Years ago, the browser would complain on stderr, if it could not find a symbol in the library, but now firefox4 is perfectly content -- yet,

Re: sqlite3 and other already-installed components

Any updates on this? I see, that the firefox-5 is still being built with most 3rd-party components (including even JPEG) taken from firefox, rather than ports :( -mi On 25.05.2011 16:35, Mikhail T. wrote: As always, I'd like to see more of the already installed components used by

Re: sqlite3 and other already-installed components

Attached are the patches for Mk/bsd.gecko.mk and firefox and thunderbird ports, which build both applications with all the 3rd-party components coming from other ports. This presumes, the png is built with APNG enabled, which, hopefully, will soon become the default: http://www.freebsd.or

Building www/firefox with PGO enabled fails

A freshly csup-ed ports tree. No local modifications: [...] /usr/local/bin/zip -r9D /home/ports/www/firefox/work/mozilla-release/obj-i386-unknown-freebsd8.2/browser/locales/../../dist/install/firefox-6.0.2.en-US.langpack.xpi install.rdf chrome chrome.manifest -x chrome/en-US.manifest updating

Re: Building www/firefox with PGO enabled fails

On 14.09.2011 05:04, Florian Smeets wrote: I think that looks like the error you get when the PGO build does not have access to your X server, did you pay attention to the message which was shown? See the Makefile if you didn't see it. I just verified that it works with a vanilla ports tree. I

New nss and firefox

Hello! The following simple update: RCS file: /home/pcvs/ports/security/nss/Makefile,v retrieving revision 1.59 diff -U 2 -r1.59 Makefile --- Makefile7 Oct 2011 20:40:40 - 1.59 +++ Makefile26 Dec 2011 00:53:24 - @@ -22,8 +22,8 @@ _MAJOR=3

Re: New nss and firefox

On 26.12.2011 02:07, Florian Smeets wrote: replying only to the nss part of you email. We have nss 3.13.1 in our svn tree (not the with CKBI 1.88 version yet, I'll get to it) and we decided to wait with updating the version in ports until after FreeBSD 9.0-RELEASE. Is there a ports-freeze in ef

Re: New nss and firefox

On 28.12.2011 03:16, Florian Smeets wrote: The ports slush is still not completley lifted. We are still in a partial ports thaw. Well, the update right now will not even touch anything "major" -- firefox builds with its own. And the new version is completely compatible. What is it that you need

Re: New nss and firefox

Thanks. But if you bothered anyway, why not update to the latest? Are there known problems? -- Sent from my BlackBerry -Original Message- From: Florian Smeets Date: Wed, 28 Dec 2011 23:18:07 To: Mikhail T. Cc: Subject: Re: New nss and firefox On 28.12.11 16:51, Mikhail T. wrote

Re: New nss and firefox

On 29.12.2011 03:36, Florian Smeets wrote: Mikhail, i'll try to explain our rationale one more time. Thank you very much for your patience. a) Sweeping commits are still not allowed as the 9.0-RELEASE process is NOT finished yet. I think, this is the key to our disagreement -- I do not thi

Re: New nss and firefox

On 29.12.2011 03:36, Florian Smeets wrote: a) Sweeping commits are still not allowed as the 9.0-RELEASE process is I have just replaced the older nss (3.12.x) with the new 3.13.1 (patch attached). I then restarted (without rebuilding) the three nss-using programs on my system: pidgin, firefox

ca_root_nss fails to install if /etc/ssl/cert.pem alredy exists

Hello! The simple patch below fixes the scenario, when one is updating the port -- and it fails to finish installing, because the sym- -link already exists: Index: Makefile === RCS file: /home/pcvs/ports/security/ca_root_nss/Makefile

Re: ca_root_nss fails to install if /etc/ssl/cert.pem alredy exists

On 27.02.2012 17:19, Florian Smeets wrote: maybe I'm missing something but this should not happen because of %%ETCSYMLINK%%@cwd / %%ETCSYMLINK%%etc/ssl/cert.pem Not sure... I had the update fail for me because the link was still there... Maybe, it is because I did not have the ETCSYMLINK option

USE_MOZILLA=yes for firefox and thunderbird

Hello! Attached is a fairly small patch, that changes the two ports to use ports of all of the components currently supported by bsd.gecko.mk. Please, consider committing soon so the work on bringing-in other components (such as hunspell, sqlite3, or xulrunner) can begin. Thank you! Yours,

Re: USE_MOZILLA=yes for firefox and thunderbird

On 11.07.2012 22:51, Jan Beich wrote: "Mikhail T." writes: Hello! Attached is a fairly small patch, that changes the two ports to use ports of all of the components currently supported by bsd.gecko.mk. I'm afraid this is already fixed in the experimental branch for *all* gec

Re: [lini...@freebsd.org: ports/170946: [patch] mark certain ports broken on ARM]

On 23.08.2012 19:27, Mark Linimon wrote: ge...@freebsd.org devel/nspr m...@aldan.algebra.com devel/qmake off...@freebsd.org devel/icu po...@freebsd.org lang/tcl86 I'd give a shot at repairing the above four, if I had access to an ARM system... None are present on the current list

Re: [lini...@freebsd.org: ports/170946: [patch] mark certain ports broken on ARM]

On 23.08.2012 21:47, Eitan Adler wrote: NOT and ONLY are meant for ports which by definition will never work another arch (think x86info). In this case the ports are just broken. That's a fine distinction, that makes little sense to me -- the total number of ports, which can not ever work on so

Can't build www/firefox on 8.3

Running: % make CC=gcc46 CXX=g++46 /mi/ports/www/firefox/work/mozilla-release/content/base/src/nsContentUtils.cpp: In static member function 'static bool nsContentUtils::GetSVGGlyphExtents(nsContentUtils::Element*, const gfxMatrix&, gfxRect*)': /mi/ports/www/firefox/wo

firefox crashes

Hello! Firefox-20 would not build with the regular compiler, so I installed the shiny new gcc-4.8 This builds (though the attached patch is required -- at least, on i386), but dies on startup from "illegal instruction". If I start under gdb, instead of SIGILL, I see segmentation fault -- inside

Re: firefox crashes

22.04.2013 05:25, Jan Beich ???(??): > It should have been fixed by files/patch-bug824317. > > cf. https://bugzilla.mozilla.org/show_bug.cgi?id=824317 Well, it does not. Try it on an i386 system. > > but dies on startup from "illegal instruction". > > > > If I start under gdb, instead of SIGILL

Re: firefox crashes

23.04.2013 12:49, Jan Beich ???(??): >>> http://www.freebsd.org/portmgr/policies_eol.html >> > According to this link, 8.x is not EOLed until April of 2014 -- a year >> > from now... > The tip of /stable/8 is, not an arbitrary old revision with multiple > regressions in-between supported releas

Allow thunderbird port to build with any compiler

Hello! The "fix" of requiring a special gcc for Thunderbird (and Seamonkey) on i386 is an abomination. If a library is reporting a missing symbol, it needs to be linked with a library, that defines the symbol (-lnspr4 in this case). Careful reading of ldap/sdks/c-sdk/ldap/libraries/libprldap/Makef

Re: [lini...@freebsd.org: svn commit: r426199 - in head: mail/jmba mail/thunderbird math/gap math/openlibm misc/e2fsprogs-libblkid multimedia/assimp multimedia/bino multimedia/libva-intel-driver multi

On 11/15/16 19:50, Mark Linimon wrote: FYI. The aarch64 logs may be found at: http://thunderx1.nyi.freebsd.org/build.html?mastername=110arm64-default&build=426002 mcl Is there an aarch64 host, where a committer can login and "play"? -mi ___ fr

Videos have a pink hint in Firefox

Hello! This is a bizarre new problem I noticed after a recent upgrade of Firefox -- not sure, when exactly it happened, because I don't view videos on my computer very often. Below is an example of what I'm talking about -- a screenshot from YouTube. But the problem is the same on other sites t

Re: Videos have a pink hint in Firefox

On 03.09.2017 15:46, Jan Beich wrote: Which KMS, DDX and DRI driver? If in doubt try using vesa (BIOS) or scfb (UEFI) with llvmpipe. I can't help much with x11@ issues due to lack of hardware. The box is FreeBSD-10.4-BETA1/amd64. It has a Radeon FireGL V3200 (RV380) card. The problem went aw

Firefox unstable on i386

Although wonderful on amd64, the recent versions of Firefox have been remarkably unstable on the two i386 systems I maintain. Fortunately, a crash in a tab does not bring down the entire program. Unfortunately, some sites - such as, for example, http://gazeta.ru/sport - are quite unusable for t

Re: Firefox unstable on i386

2/28/2018 7:15 PM, Mikhail T. пише: Although wonderful on amd64, the recent versions of Firefox have been remarkably unstable on the two i386 systems I maintain. Fortunately, a crash in a tab does not bring down the entire program. Unfortunately, some sites - such as, for example, http

Re: Firefox unstable on i386

As I wrote earlier, I have no problems with Firefox on the amd64 systems. It is the i386 ones, where the individual tabs keep crashing for no obvious reasons... -mi 3/15/2018 1:34 PM, Thierry Thomas пише: Le jeu. 15 mars 18 à 18:04:06 +0100, Mikhail T. écrivait : Although wonderful

Re: Firefox unstable on i386

How would I build Firefox with clang-4? I wonder, if that can solve my problem... Thanks! -- Sent from mobile device, please, pardon shorthand. 15 бер. 2018 р. о 13:45 Mikhail T. пише: > As I wrote earlier, I have no problems with Firefox on the amd64 systems. It > is the i386 ones,

Re: Firefox unstable on i386

On 19.03.2018 21:54, Mikhail T. wrote: How would I build Firefox with clang-4? I wonder, if that can solve my problem... Thanks! Ok, this did not help either. The subprocesses responsible for various tabs keep crashing. I gave up on trying to build a usable version myself and installed a

New firefox not building

Hi! Trying the port on a just-updated 11-stable/amd64 clang7 I get: /opt/bin/clang++70 -o nsImageModule.o -c -I/symbion/ports/www/firefox/work/.build/dist/stl_wrappers -I/symbion/ports/www/firefox/work/.build/dist/system_wrappers -include /symbion/ports/www/firefox/work/firefox-65

Firefox and llvm

Gentlemen, do you realize, building Firefox on FreeBSD today one has to wait for the entire LLVM to be recompiled -- /twice/: once because gecko, inexplicably, wants LLVM itself, and also because rust builds its own bundled version... What's wrong with the binaries provided by the base OS?.. Y

Re: Firefox and llvm

On 16.06.19 21:31, Jan Beich wrote: "Mikhail T." writes: Gentlemen, do you realize, building Firefox on FreeBSD today one has to wait for the entire LLVM to be recompiled -- /twice/: once because gecko, inexplicably, wants LLVM itself, and also because rust builds its own bundl

Re: Firefox and llvm

On 16.06.19 22:25, Jan Beich wrote: clang-sys didn't support llvm80 when gecko@ switched to it. I'm not sure myself why but maybe bindgen uses a subset of bindings that're stable Perhaps, this is something the port's maintainers should research? Along with the possibility of switching lang/rus

Re: Firefox and llvm

On 17.06.19 00:19, Jan Beich wrote: Good luck. Make sure to test in a clean environment e.g., via poudriere Jan, this is the job of the port's maintainer... The current situation -- requiring a rebuild of LLVM twice -- is ridiculous, should never have come about, and should not remain for long

rust and llvm (Re: Firefox and llvm)

On 17.06.19 00:10, Charlie Li wrote: Rust will not revert back to providing an option to use an external LLVM. Rust needs to use its own LLVM, which not only tracks upstream trunk, but has been forked and a few API bits added (which will not be upstreamed) for Rust use. Some of these API bits a

Restoring seamonkey

Hello! I'm dressing up a fairly low-powered machine, and having both browser and e-mail client in one process would be useful. The port-deleting commit cited poor update frequency, but the upstream is now at version 2.53.1 -- which means, they've made seve

Re: Restoring seamonkey

On 27.03.20 21:15, Jan Beich wrote: Good luck: - 2.53.1 is still vulnerable - Upstream has unstable release cadence - ESR60 engine may not build with new dependencies - Expecting someone else to do the work What, I wonder, made you think, I am expecting someone else to do the work? My question

Re: Restoring seamonkey

On 28.03.20 20:47, Jan Beich wrote: Lack of the homework. I really don't understand this, Jan... Let's replay: 1. I wanted to install Seamonkey on a system I'm dressing up, and found, that the port is no longer available. 2. I looked for the final commit-message, and found: 1. it was de

Shared libxul.so (was Re: Restoring seamonkey)

[Forking the Seamonkey thread, because the topic is different] On 28.03.20 20:47, Jan Beich wrote: libxul.so is no longer the same between various Gecko-based projects. It doesn't have to be /the same/ -- just needs to still be compatible enough for a motivated developer to patch the rest. An

Re: Restoring seamonkey

On 29.03.20 15:13, Jan Beich wrote: "Mikhail T." writes: 4. I then wrote you an e-mail inquiring, if the port can be restored... Did you try to build 2.53.1 before writing the email? No, of course not -- not anticipating such reaction, I thought, I'd ask first... Do the 2.

Firefox playing videos takes too much CPU

Hello! I noticed recently, how much CPU playing a video takes... A firefox-process, playing a modest 1080p clip (such as this ) for example, will show in top as taking between 150-160% CPU (and mine are E5-1620 0 @ 3.60GHz), with Xorg taking /another