Arch Linux User Repository

Starting with the PKGBUILD for 91.7.1, Firefox ESR will use a separate profile directory ($HOME/.mozilla/firefox-esr) so it can be installed side-by-side with other Firefox versions and avoid profile conflicts and downgrade warnings.

The package won't compile anymore against recent libevent: 7:57.07 /build/firefox-esr/src/firefox-102.1.0/ipc/chromium/src/third_party/libevent/./arc4random.c:488:1: error: static declaration of 'arc4random_buf' follows non-static declaration

There's a fix available - please pull it from firefox-nightly.

You need to actually apply the new patch. The build is still breaking.

Build currently breaks with cbindgen >= 0.24.2. This patch fixes it: https://raw.githubusercontent.com/canonical/firefox-snap/5622734942524846fb0eb7108918c8cd8557fde3/patches/fix-ftbfs-newer-cbindgen.patch

Does this mean it pretty much breaks all plugin packages?

@jonathon The language pack files for Firefox ESR should now be placed in /usr/lib/firefox-esr/browser/extensions/, otherwise it will not be recognized

SamLukeYes, that's invalid for this PKGBUILD, it already installs to the correct place.

Edit: Maybe that was a comment from the previous separate i18n package that was merged into this one... :/

No longer works since 91.7.1 update. See https://github.com/archlinuxcn/repo/issues/2712

Sorry, I just noticed a hint that Firefox ESR 91.7.1 will use a separate config directory...

Please ignore the suggestion I just made. XD

installing firefox-esr (91.7.1-1) breaks dependency 'firefox' required by firefox-noscript

@jonathon Maybe re-add provides and conflicts to PKGBUILD .

Starting with the PKGBUILD for 91.7.1, Firefox ESR will use a separate profile directory ($HOME/.mozilla/firefox-esr) so it can be installed side-by-side with other Firefox versions and avoid profile conflicts and downgrade warnings.

Having some issues compiling 91.6.0-1

26:58.39 error: could not compile `gkrust`      
26:58.40 Caused by:                             
26:58.40   process didn't exit successfully: `/usr/bin/rustc --crate-name gkrust toolkit/library/rust/lib.rs --error-format=json --json=diagnostic-rendered-ansi --crate-type staticlib --emit=d
ep-info,link -C opt-level=2 -C panic=abort -C embed-bitcode=no -Clto --cfg 'feature="cubeb-remoting"' --cfg 'feature="cubeb_pulse_rust"' --cfg 'feature="gecko_profiler"' --cfg 'feature="gecko_
profiler_parse_elf"' --cfg 'feature="glean_with_gecko"' --cfg 'feature="moz_memory"' --cfg 'feature="moz_places"' --cfg 'feature="oxidized_breakpad"' --cfg 'feature="quantum_render"' --cfg 'fe
ature="remote_agent"' --cfg 'feature="simd-accel"' --cfg 'feature="webgpu"' --cfg 'feature="webrtc"' --cfg 'feature="with_dbus"' -C metadata=eb0fff5eca01a79e -C extra-filename=-eb0fff5eca01a79
e --out-dir /home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/deps --target x86_64-unknown-linux-gnu -C linker=/home/user/.cache/yay/firefox-esr/src/fir
efox-91.6.0/build/cargo-linker -C incremental=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/incremental -L dependency=/home/user/.cache/yay/firefox-
esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/deps -L dependency=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/release/deps --extern gkrust_shared=/home/user/.cache/ya
y/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/deps/libgkrust_shared-e0d410bb55e8b370.rlib --extern mozglue_static=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/o
bj/x86_64-unknown-linux-gnu/release/deps/libmozglue_static-e34df1e0e6fbe4f3.rlib --extern mozilla_central_workspace_hack=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown
-linux-gnu/release/deps/libmozilla_central_workspace_hack-fa2a1abc60b1960f.rlib --extern swgl=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/deps/lib
swgl-e8038a9e4280e8a8.rlib -C debuginfo=2 --cap-lints warn -Cembed-bitcode=yes -C panic=abort -C profile-generate=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj -C codegen-units=1 -L
 native=/usr/lib -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/build/audioipc-dc225dbe18278497/out -L native=/home/user/.cache/yay/firefox
-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/build/lmdb-rkv-sys-73dd8d703979bc22/out -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu
/release/build/mozglue-static-42f2508774824a15/out -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/dist/bin -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/ob
j/security/nss/lib/nss/nss_nss3 -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/security/nss/lib/ssl/ssl_ssl3 -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/
obj/config/external/nspr/pr -L native=/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/obj/x86_64-unknown-linux-gnu/release/build/swgl-3c3323110db895b7/out` (signal: 9, SIGKILL: kill)
26:58.43 make[4]: *** [/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/config/makefiles/rust.mk:405: force-cargo-library-build] Error 101
26:58.43 make[3]: *** [/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/config/recurse.mk:72: toolkit/library/rust/target] Error 2
26:58.43 make[2]: *** [/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/config/recurse.mk:34: compile] Error 2
26:58.43 make[1]: *** [/home/user/.cache/yay/firefox-esr/src/firefox-91.6.0/config/rules.mk:355: default] Error 2
26:58.43 make: *** [client.mk:65: build] Error 2                                                
26:58.44 165 compiler warnings present.         
26:58.53 Failed to parse ccache stats output: Summary:

You don't seem to use official build tools with current build flags in a clean chroot. PKGBUILD misses !lto option to actually build.

Hi, firefox-esr-bin recently got added conflicts:firefox, probably in aligning to this package; is it really conflicting? Thanks

Hi, it seems like the 0002-Bug-1745560-Add-missing-stub-for-wl_proxy_marshal_fl.patch isn't actually applied in the PKGBUILD so the build still fails.

@AndyRTR fixed :)

for a clean chroot build !lto is required and 0002-Bug-1745560-Add-missing-stub-for-wl_proxy_marshal_fl.patch from the official package. Please add this to fix the build.

!lto is now required. You should always do a clean chroot build and check against changes the devs apply to recent FF in the repos.

47:19.94 toolkit/library/build/libxul.so
47:24.31 ld.lld: error: undefined symbol: wl_proxy_marshal_flags
47:24.32 >>> referenced by wayland-client-protocol.h:3672 (/usr/include/wayland-client-protocol.h:3672)
47:24.33 >>>               /home/devninja/archpackages/firefox-esr/src/firefox-91.4.1/obj/toolkit/library/build/../../../gfx/gl/Unified_cpp_gfx_gl0.o:(mozilla::gl::DeleteWaylandGLSurface(void*))
47:24.33 >>> referenced by wayland-client-protocol.h:3672 (/usr/include/wayland-client-protocol.h:3672)
47:24.33 >>>               /home/devninja/archpackages/firefox-esr/src/firefox-91.4.1/obj/toolkit/library/build/../../../gfx/gl/Unified_cpp_gfx_gl0.o:(mozilla::gl::WaylandGLSurface::~WaylandGLSurface())
47:24.33 >>> referenced by wayland-client-protocol.h:1291 (/usr/include/wayland-client-protocol.h:1291)
47:24.33 >>>               /home/devninja/archpackages/firefox-esr/src/firefox-91.4.1/obj/toolkit/library/build/../../../gfx/gl/Unified_cpp_gfx_gl0.o:(mozilla::gl::GLContextEGL::CreateWaylandBufferSurface(mozilla::gl::EglDisplay&, void*, mozilla::gfx::IntSizeTyped<mozilla::gfx::UnknownUnits>&))
47:24.33 >>> referenced 69 more times

The build fails with recent wayland 1.20 and needs the fix from extra/firefox. Please add it.

@dctxmei

Fixed now thanks

@bittin You have not updated .SRCINFO content, which is still displayed as v91.3.0-1 .

Did my first update of this package now :)

@bittin thanks for taking care

I'm not interested in maintain this package anymore. Any volunteer to adopt it?

@figue precisely, i3core/4G RAM with 16G swap works

@rifl yeah, it uses a lot of RAM now... I can't build in a 12GB laptop... So I have to use my work machine.

@figue thanks for updates gonna work rust compiling went on right way or after I lift up swap memory since rust compiling crashed for eating 90 of my memory

@dctxmei thanks. Package updated.

@figue I made some modifications based on extra/firefox, you can execute:

$ curl -o firefox-esr.91.0.patch https://fars.ee/4glz
$ git apply --reject firefox-esr.91.0.patch

It will work normally.

Thanks for the update. I think there is one final thing. The build errored for me at "crashreporter-symbols-full.zip" step at the end. Looking at the Arch PKGBUILD, it appears this is called "crashreporter-symbols-full.tar.zst" now. I have not had time to double check that yet.

@dudemanguy thanks. I don't have access to my main build machine, and I didn't saw this file is not when it should. Hope it's fixed now.

I believe this is missing the identity-icons-brand.svg file.

@figue I'm not that good on or familiar with details inside of firefox-esr though on code level, maybe I was wrong the recently version of it that feature remaining is firefox-esr60.0. If you confused of what I'm supposed it to be, I have just one line written to .xinitrc: "exec emacs" -- it works without a window manager as emacs takes over anything, except that the external app launched in emacs where window-manager part is missing. It's trivial thing for others who have IDEs with handling the window stuff.

@rifl I don't know. Firefox from [extra] goes well? Do you know if that feature can be enabled in some way in the compilation time?

@figue no it goes well. As precisely I mean the drage-and-move window feature has been removed from official firefox-esr but not aur version, thus I cant drag-and-move window around in the case without window-manager. My question is that is it possible to get it back to work?

@rifl is this some kind of issue of my PKGBUILD code?

@figue I just need feature likes dragging and moving window around wherein os absent window manager, which is missed in latest version of firefox/icecat.

@bittin Seems it was pushed today. I'll update ASAP. Thanks

78.12 Stable: https://pastebin.com/eRwwwNL5

@CommodoreCrunch thanks! Key verified:

gpg --keyserver hkp://keys.openpgp.org --recv-keys EBE41E90F6F12F6D

@figue Refresh the existing key on keys.openpgp.org and you'll get the new subkey. A lot of people, Mozilla included, started using that instead of the keyserver pool due to newish vulnerabilities.

New version is signed with this key EBE41E90F6F12F6D but seems not available to public keyservers. Anyone can confirm please?

https://ftp.mozilla.org/pub/firefox/releases/78.11.0esr/KEY

Key seems to be renewed already...

Todays update: https://pastebin.com/1yTeH1xC

Todays update: https://pastebin.com/YtMVDqGw

Todays update: https://pastebin.com/myMfheA2

Also you can low security.sandbox.content.level to 1 as a workaround, so you don't need to disable it completely.

In case you want to rebuild firefox-esr with crashreporter (and so, report the tab crash) you can use this PKGBUILD:

https://mirror.lnx.im/archlinux/firefox-esr/78.8.0/PKGBUILD

If videos still crash, try to launch firefox with MOZ_DISABLE_CONTENT_SANDBOX=1 variable.

@AndyRTR sorry, forgot to push the new patch. It's fixed now.

I see Youtube crashing here also with 78.7.1-2. The new release seems to have broken checksums. Can you please check the sha256sums? I get here d32c87c4526e897d64453914da43f99366d1d0b7d71e43b4027a6cb5aa274040 rust_1.48.patch.gz after a fresh checkout.

WARNING: this release has a bug in html5 videos, tabs chash. A possible workaround is to change these properties in about:config (multi-process tabs):

browser.tabs.remote.autostart = false
browser.tabs.remote.autostart.2 = false

Until I figure out why it's crashing and can patch it, stay with previous version or use the workaround.

@sp1d3rmxn yeah, it looks like something you describe... I don't know if that port can be changed to another...

In build(), just before "./mach python build/pgo/profileserver.py" you can add this line for instance: sed -e 's/8888/8989/g' -i build/pgo/profileserver.py

and test if it works...

@figure I found what it was, I had to use "watch -n1" to catch it though, yes it was conflicting with port 8888 as you identified in the post. Unfortunately program I am/was using sets up proxy on that port to check updates, etc. but, only when it is setup to use a socks5 proxy.

TLDR; So on my end, a program I was running is intermittently using the port for maintenance to communicate and happened to be running when the build process reached the point of needing to setup its own service on that port.

@figure I was able to build in the chroot and on a virtual machine, so must be something going on with that particular desktop. I will need to dig further on it, probably something weird like bad memory, etc. Apologies for wasting time but, thanks for replying it is appreciated.

@sp1d3rmxn that's odd... I said 8080 from memory, but seems 8888, see:

https://fossies.org/linux/firefox/build/pgo/profileserver.py

In any case, I've built firefox-esr hundreds of times, so if 8888 is free there shouldn't be any problem.

@figue This is all I have running that is listening (used "sudo netstat -anp | grep LISTEN):

tcp        0      0 127.0.0.1:7080          0.0.0.0:*               LISTEN      327788/ssh          
tcp        0      0 0.0.0.0:5900            0.0.0.0:*               LISTEN      371319/x0vncserver  
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      982/sshd: /usr/bin/ 
tcp        0      0 127.0.0.1:20887         0.0.0.0:*               LISTEN      327788/ssh          
tcp        0      0 127.0.0.1:20888         0.0.0.0:*               LISTEN      420898/autossh.bin

I have seen something similar like this in the past where I have had to use sudo or root privileges to build because of the port assignment being low which I am only thinking because maybe it went random? I am not really sure never saw this before in the years of using this package. As I type this I am building in a chrooted environment to see if the port thing comes up again, either way I will report the result.

@sp1d3rmxn I think it uses the default port 8080 (it's hardcoded). Check if you have some process listening on it.

Kinda weird problem wondering if anyone else is getting this:

Traceback (most recent call last):
  File "build/pgo/profileserver.py", line 80, in <module>
httpd.start(block=False)
  File "/home/upgrade/firefox-esr/src/firefox-78.7.1/testing/mozbase/mozhttpd/mozhttpd/mozhttpd.py", line 269, in start
    self.httpd = EasyServer((self.host, self.port), self.handler_class)
  File "/usr/lib64/python2.7/SocketServer.py", line 420, in __init__
    self.server_bind()
  File "/usr/lib64/python2.7/BaseHTTPServer.py", line 108, in server_bind
    SocketServer.TCPServer.server_bind(self)
  File "/usr/lib64/python2.7/SocketServer.py", line 434, in server_bind
    self.socket.bind(self.server_address)
  File "/usr/lib64/python2.7/socket.py", line 228, in meth
    return getattr(self._sock,name)(*args)
socket.error: [Errno 98] Address already in use

I cannot figure out what it is trying to bind to that may already be in use because there is nothing running. Looking at the source I see the port set to "0" which according to the docs this is default. Any ideas?

@sp1d3rmxn thanks. Patch added.

New release new bugs.. @figue if you please could add this patch:

reported: https://bugzilla.mozilla.org/show_bug.cgi?id=1692363

patch just uplifted 2days ago to ESR but available for 2 months for their other versions: https://bugzilla.mozilla.org/show_bug.cgi?id=1684261

Yep... Updated. Thanks.

It's there: https://archive.mozilla.org/pub/firefox/releases/78.6.1esr/

@bittin 78.6.1esr is out? Can't found in official FTP.

@sp1d3rmxn Thanks for the advice. See ea05890161bc

@figue

The same patch for icecat https://aur.archlinux.org/packages/icecat/#comment-779101 is going to be needed for this PKGBUILD to please. The only diff between them being "patch -Np1 -i ../../../rust_1.48.patch" needs to be "patch -Np1 -i ../rust_1.48.patch" as you may already know.

@AndyRTR I'm building in a laptop with 16GB RAM with a W10 virtual machine running and it works for me. All patches are always welcome. I've cloned this package to my personal github, so pull requests will be accepted.

https://github.com/figue/firefox-esr

I guess somewhere around 78.3.0 the memory requirement while linking toolkit/library/build/libxul.so had a huge jump. Now my 16gb are no more sufficient and the system starts heavy swapping. Maybe we can reduce this. heftig made some suggestions to turn off pgo, lto and use bfd linker with its memory saving flags but with paying in code quality. maybe via an optional switch in the PKGBUILD.

@bittin why you flagged as out-of-date?

@bittin Probably we'll wait until final realease is released.

There will be a new update to ESR with 83 RC2 tomorrow evening

Thanks @heftig. All suggestions implemented.

You should set StartupWMClass=firefoxesr in the desktop file to match the remoting name set in .mozconfig, otherwise the desktop might not be able to assign the windows to the app correctly.

You should disable the crashreporter, as it's only useful if you also submit the built symbol archive to Mozilla. Since each AUR user builds the browser themselves, this isn't happening.

For the same reason, you can remove buildsymbols and the handling of the '*crashreporter-symbols-full.zip'. This will also remove a good chunk of the build time.

To continue using old 68 branch, you can use this package https://aur.archlinux.org/packages/firefox-esr68

@TheGoliath Added you as a co-maintainer.

@GI_Jack thanks. Seems good. Done.

Please change

provides=(firefox)

to

provides=(firefox=${pkgver})

This way it works with packages that version check FX

New prebuilt binaries in my own openSUSE Build Service account:

[home_ffigue_Arch]
Server = http://download.opensuse.org/repositories/home:/ffigue/Arch/$arch

To verify signature, you can do like this:

$ wget -O- http://download.opensuse.org/repositories/home:/ffigue/Arch/x86_64/home_ffigue_Arch.key | sudo gpg --homedir /etc/pacman.d/gnupg --import -
$ sudo pacman-key --lsign-key F5AAC5A5424ED5CA

@sp1d3rmxn thanks. Compilation with disable-elf-hack seems ok. Please test.

@figue

Ok the PKGBUILD change I added per the bug report has proven to also be successful with this.

Again like Icecat but, slightly different (no modifications to march/mtune either just like the Icecat build):

In the "prepare()" section under the heading "#Features" add this line:

ac_add_options --disable-elf-hack

It looks like this for me:

Features

ac_add_options --enable-alsa

ac_add_options --enable-jack

ac_add_options --enable-startup-notification

ac_add_options --enable-crashreporter

ac_add_options --disable-gconf

ac_add_options --disable-updater

ac_add_options --disable-elf-hack <---- ADD THIS LINE END

Again no further changes necessary and able to use CPU tuning as expected (no generics at all).

Hope this helps anyone going through the same headache.

@sp1d3rmxn thank you, I reply you in IceCat AUR package. I have to compile FF ESR the same way. Seems that something is wrong when you tune CFLAGS...

This bullshit has been going on for awhile. It seems as though specifying your CPU arch is now an issue. Guess the new amateur hour devs are to <fill in the blank> to realize "x86-64" for generic purposes is weak sauce if you don't know that your CPU has specific instruction sets that can get you more bang for the buck. Anyways here's the report: https://bugs.llvm.org/show_bug.cgi?id=43659

@figue - Yes and no answer to your question, yes for the last 4 years we have run linux-hardened, been through many of these upgrades and no because these are stock and installed as such. The only compilation is due to packages being in AUR. We use distcc occasionally if there are a ton to upgrade but, that compiler was eliminated in these test runs.

The issues when new versions of libraries get released seems to be a decades old problem for linux devs to keep their shit straight. The comment section here is proof of the constant issues that shouldn't exist with something called "extended support ...". I did find several others who have had to downgrade to clang/llvm/rust to previous versions or others have changed the compilation back to GCC instead of the shitty clang like here: https://www.linuxquestions.org/questions/slackware-14/llvm-9-0-0-clang-segfault-firefox-build-4175661799/ or this about waterfox (another mozilla based product experiencing similar bullshit to this): https://aur.archlinux.org/packages/waterfox-classic-git/#comment-716172

I have even more links but, I think you get the idea. Doesn't seem to matter to much about the customization as much as it seems in tracking down patches due to devs not being thorough because others are experiencing identical issues and they aren't even using Arch.

I suspect others may not have gone through this because having 500 machines to manage as I do tends to show issues like this quickly.

@sp1d3rmxn that's weird... Do you use stock kernel? Any other custom configuration?

This occurs in a fresh install with a fresh chroot environment (tried on 3 different machines that are brand new):

48:29.54 clang-9: error: unable to execute command: Segmentation fault (core dumped)

48:29.54 clang-9: error: clang frontend command failed due to signal (use -v to see invocation)

48:29.54 clang version 9.0.0 (tags/RELEASE_900/final)

48:29.54 Target: x86_64-pc-linux-gnu

48:29.54 Thread model: posix

48:29.54 InstalledDir: /usr/bin

48:29.54 clang-9: note: diagnostic msg: PLEASE submit a bug report to and include the crash backtrace, preprocessed source, and associated run script.

48:29.54 clang-9: error: unable to execute command: Segmentation fault (core dumped)

48:29.54 clang-9: error: clang frontend command failed due to signal (use -v to see invocation)

48:29.54 clang version 9.0.0 (tags/RELEASE_900/final)

48:29.54 Target: x86_64-pc-linux-gnu

48:29.54 Thread model: posix

48:29.54 InstalledDir: /usr/bin

48:29.54 clang-9: note: diagnostic msg: PLEASE submit a bug report to and include the crash backtrace, preprocessed source, and associated run script.

48:29.58 clang-9: note: diagnostic msg:

48:29.58 ****

48:29.58 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:

48:29.58 Preprocessed source(s) and associated run script(s) are located at:

48:29.58 clang-9: note: diagnostic msg: /tmp/test-ctors-36d1f3.c

48:29.58 clang-9: note: diagnostic msg: /tmp/test-ctors-36d1f3.sh

48:29.58 clang-9: note: diagnostic msg:

48:29.58 ****

48:29.58 make[4]: *** [/build/firefox-esr/src/firefox-68.2.0/config/rules.mk:755: test-ctors.o] Error 254

48:29.58 make[4]: *** Waiting for unfinished jobs....

48:29.58 clang-9: note: diagnostic msg:

48:29.58 ****

48:29.58 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:

48:29.58 Preprocessed source(s) and associated run script(s) are located at:

48:29.58 clang-9: note: diagnostic msg: /tmp/test-array-aae6dc.c

48:29.58 clang-9: note: diagnostic msg: /tmp/test-array-aae6dc.sh

48:29.58 clang-9: note: diagnostic msg:

48:29.58 ****

48:29.59 make[4]: *** [/build/firefox-esr/src/firefox-68.2.0/config/rules.mk:755: test-array.o] Error 254

48:29.59 make[3]: *** [/build/firefox-esr/src/firefox-68.2.0/config/recurse.mk:101: build/unix/elfhack/libs] Error 2

48:29.59 make[2]: *** [/build/firefox-esr/src/firefox-68.2.0/config/recurse.mk:34: libs] Error 2

48:29.59 make[1]: *** [/build/firefox-esr/src/firefox-68.2.0/config/rules.mk:413: default] Error 2

48:29.59 make: *** [client.mk:125: build] Error 2

48:29.61 159 compiler warnings present.

48:29.66 Notification center failed: Install notify-send (usually part of the libnotify package) to get a notification when the build finishes.

==> ERROR: A failure occurred in build().

Aborting...

Same exact thing outside of chroot:

52:44.67 clang-9clang-9: : error: unable to execute command: Segmentation fault (core dumped)error 52:44.67 : unable to execute command: Segmentation fault (core dumped) 52:44.67 clang-9clang-9: : error: error: clang frontend command failed due to signal (use -v to see invocation) 52:44.67 clang frontend command failed due to signal (use -v to see invocation) 52:44.67 clang version 9.0.0 (/home/sp1d3y/.cache/asp 424e7455481f941ec2259cc4c491238007cdf3d3) 52:44.67 Target: clang version 9.0.0 (/home/sp1d3y/.cache/asp 424e7455481f941ec2259cc4c491238007cdf3d3)x86_64-pc-linux-gnu 52:44.67 Thread model: Target: posixx86_64-pc-linux-gnu 52:44.67 InstalledDir: 52:44.67 /usr/binThread model: 52:44.67 posix 52:44.67 clang-9: InstalledDir: /usr/binnote 52:44.67 : diagnostic msg: PLEASE submit a bug report to and include the crash backtrace, preprocessed source, and associated run script.clang-9: 52:44.67 note: diagnostic msg: PLEASE submit a bug report to and include the crash backtrace, preprocessed source, and associated run script. 52:44.71 clang-9: note: diagnostic msg: 52:44.72 **** 52:44.72 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT: 52:44.72 Preprocessed source(s) and associated run script(s) are located at: 52:44.72 clang-9: note: diagnostic msg: /tmp/test-array-7c38e4.c 52:44.72 clang-9: note: diagnostic msg: /tmp/test-array-7c38e4.sh 52:44.72 clang-9: note: diagnostic msg: 52:44.72 *** 52:44.72 clang-9: note: diagnostic msg: 52:44.72 ** 52:44.72 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT: 52:44.72 Preprocessed source(s) and associated run script(s) are located at: 52:44.72 clang-9: note: diagnostic msg: /tmp/test-ctors-2a586c.c 52:44.72 clang-9: note: diagnostic msg: /tmp/test-ctors-2a586c.sh 52:44.72 clang-9: note: diagnostic msg: 52:44.72 ** 52:44.72 make[4]: [/home/sp1d3y/software/testing/firefox-esr/src/firefox-68.2.0/config/rules.mk:755: test-array.o] Error 254 52:44.72 make[4]: Waiting for unfinished jobs.... 52:44.72 make[4]: [/home/sp1d3y/software/testing/firefox-esr/src/firefox-68.2.0/config/rules.mk:755: test-ctors.o] Error 254 52:44.72 make[3]: [/home/sp1d3y/software/testing/firefox-esr/src/firefox-68.2.0/config/recurse.mk:101: build/unix/elfhack/libs] Error 2 52:44.72 make[2]: [/home/sp1d3y/software/testing/firefox-esr/src/firefox-68.2.0/config/recurse.mk:34: libs] Error 2 52:44.72 make[1]: [/home/sp1d3y/software/testing/firefox-esr/src/firefox-68.2.0/config/rules.mk:413: default] Error 2 52:44.72 make: [client.mk:125: build] Error 2 52:44.74 159 compiler warnings present. 52:44.83 /usr/bin/notify-send --app-name=Mozilla Build System Mozilla Build System Build failed ==> ERROR: A failure occurred in build(). Aborting...

clang has been recompiled and stock version used in both tests with both producing exactly the same results.

@AndyRTR indeed. Thank you!

FF-esr builds well with rust 1.39 and this patch applied. So there's no need to build another AUR (rust-bin) pkg:

https://svnweb.freebsd.org/ports/head/www/firefox-esr/files/patch-bindgen-rust1390?revision=516995&view=markup&pathrev=516995

I had to push a new release, because Rust version 1.39 breaks ESR build. Dependency rust-bin=1.37.0 it's on AUR.

If anybody wants to build in a clean chroot, I managed to build using clean-chroot-manager by graysky adding rust-bin to local repo then build it.

In classic moronic style the build just fails:

14:56.22 error: build failed

14:56.22 make[4]: *** [/home/user/.cache/pikaur/build/firefox-esr/src/firefox-68.2.0/config/makefiles/rust.mk:240: force-cargo-library-build] Error 101

14:56.22 make[3]: *** [/home/user/.cache/pikaur/build/firefox-esr/src/firefox-68.2.0/config/recurse.mk:74: toolkit/library/rust/target] Error 2

14:56.22 make[3]: *** Waiting for unfinished jobs....

16:36.62 make[2]: *** [/home/user/.cache/pikaur/build/firefox-esr/src/firefox-68.2.0/config/recurse.mk:34: compile] Error 2

16:36.62 make[1]: *** [/home/user/.cache/pikaur/build/firefox-esr/src/firefox-68.2.0/config/rules.mk:413: default] Error 2

16:36.62 make: *** [client.mk:125: build] Error 2

Don't you just love those explanations? Error 2? WTF is that? Google is not your friend for that lame shit. This is 2019 the lame ass errors and stupid compile errors should be well beyond done. What the hell guys.

@NekoBuster firefox.desktop updated. Please refresh your local repo. Cheers

Can you add StartupWMClass to the desktop app so we don't need to mannually add it every time after update?

@AndyRTR thanks for reporting... I forgot to check in a clean chroot. Can you check if all is OK now? I'll be online later to check it myself at home.

Package won't build due to missing python v3 dependency. Please compare your deps/makedeps and further build with last FF67 from extra repo or FF68 once it the repos. Please make sure it builds well in a clean chroot.

@dctxmei yeah, i686 can be dropped... I don't use for years. Any objections?

I use the command:

$ extra-x86_64-build -c

Result:

Checking PKGBUILD
PKGBUILD (firefox-esr) E: File referenced in $startdir
Checking firefox-esr-60.7.0-1-x86_64.pkg.tar.xz
firefox-esr W: Referenced library 'libmozgtk.so' is an uninstalled dependency
firefox-esr W: Referenced library 'libxul.so' is an uninstalled dependency
firefox-esr W: Referenced library 'libmozsqlite3.so' is an uninstalled dependency
firefox-esr W: Referenced library 'libmozsandbox.so' is an uninstalled dependency
firefox-esr W: Referenced library 'liblgpllibs.so' is an uninstalled dependency
firefox-esr W: Referenced library 'libmozavutil.so' is an uninstalled dependency
firefox-esr E: Symlink (usr/lib/firefox/dictionaries) points to non-existing /usr/share/hunspell
firefox-esr E: Symlink (usr/lib/firefox/hyphenation) points to non-existing /usr/share/hyphen
firefox-esr W: Unused shared library '/usr/lib/libpthread.so.0' by file ('usr/lib/firefox/gtk2/libmozgtk.so')
firefox-esr W: Unused shared library '/usr/lib/libgtk-x11-2.0.so.0' by file ('usr/lib/firefox/gtk2/libmozgtk.so')
firefox-esr W: Unused shared library '/usr/lib/libgdk-x11-2.0.so.0' by file ('usr/lib/firefox/gtk2/libmozgtk.so')
firefox-esr W: Unused shared library '/usr/lib/libpthread.so.0' by file ('usr/lib/firefox/liblgpllibs.so')
firefox-esr W: Unused shared library '/usr/lib/libstdc++.so.6' by file ('usr/lib/firefox/liblgpllibs.so')
firefox-esr W: Unused shared library '/usr/lib/libm.so.6' by file ('usr/lib/firefox/libmozavcodec.so')
firefox-esr W: Unused shared library '/usr/lib/libpthread.so.0' by file ('usr/lib/firefox/libmozgtk.so')
firefox-esr W: Unused shared library '/usr/lib/libgtk-3.so.0' by file ('usr/lib/firefox/libmozgtk.so')
firefox-esr W: Unused shared library '/usr/lib/libgdk-3.so.0' by file ('usr/lib/firefox/libmozgtk.so')
firefox-esr E: Dependency gtk2 detected and not included (libraries ['usr/lib/libgtk-x11-2.0.so.0', 'usr/lib/libgdk-x11-2.0.so.0'] needed in files ['usr/lib/firefox/gtk2/libmozgtk.so'])
firefox-esr W: Dependency included and not needed ('mozilla-common')
firefox-esr W: Dependency included and not needed ('ffmpeg')
firefox-esr W: Dependency included and not needed ('hunspell')
firefox-esr W: Dependency included and not needed ('libpulse')

It is recommended to give up i686 support :)

@GI_Jack done. Thanks

fix Icon= in the .destop file. pls point it to Icon=firefox-esr

change provides=(firefox)

to

provides=(firefox=${pkgver})

This way it doesn't break plugins. Also, borked with Arch32. No ideas why.

Had to add rust 1.33 patches... Please review if everything is fine.

Thanks @AndyRTR for the out-of-date notification. Since 60.6.0 was public I try to workaround the build process, because rust 1.33 is not compatible with 60.6.0. There is a patch available and it seems that it works, but there are other errors with "macro" that throws errors and build fails. I'll try to continue later...

17:51.06 libpacing_gn.a.desc
17:51.24    Compiling cubeb-pulse v0.1.1 (/DATA2/packages/figue/firefox-esr/src/firefox-60.6.0/media/libcubeb/cubeb-pulse-rs)
17:53.58    Compiling style_traits v0.0.1 (/DATA2/packages/figue/firefox-esr/src/firefox-60.6.0/servo/components/style_traits)
17:53.74 warning: use of deprecated item 'std::ascii::AsciiExt': use inherent methods instead
17:53.74   --> servo/components/style_traits/viewport.rs:10:30
17:53.74    |
17:53.74 10 | #[allow(unused_imports)] use std::ascii::AsciiExt;
17:53.74    |                              ^^^^^^^^^^^^^^^^^^^^
17:53.74    |
17:53.74    = note: #[warn(deprecated)] on by default
17:53.74
17:54.08 error: missing documentation for macro
17:54.08    --> servo/components/style_traits/values.rs:139:1
17:54.08     |
17:54.08 139 | macro_rules! serialize_function {
17:54.08     | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
17:54.08     |
17:54.08 note: lint level defined here
17:54.08    --> servo/components/style_traits/lib.rs:12:22
17:54.08     |
17:54.08 12  | #![deny(unsafe_code, missing_docs)]
17:54.08     |                      ^^^^^^^^^^^^
17:54.08
17:54.08 error: missing documentation for macro
17:54.08    --> servo/components/style_traits/values.rs:408:1
17:54.08     |
17:54.08 408 | macro_rules! define_css_keyword_enum {
17:54.08     | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
17:54.08
17:54.10 error: aborting due to 2 previous errors
17:54.10
17:54.12 error: Could not compile `style_traits`.
17:54.12
17:54.12 To learn more, run the command again with --verbose.
17:54.13 make[4]: *** [/DATA2/packages/figue/firefox-esr/src/firefox-60.6.0/config/rules.mk:979: force-cargo-library-build] Error 101
17:54.13 make[3]: *** [/DATA2/packages/figue/firefox-esr/src/firefox-60.6.0/config/recurse.mk:73: toolkit/library/rust/target] Error 2
17:54.13 make[3]: *** Se espera a que terminen otras tareas....
17:55.40 libremote_bitrate_estimator_gn.a.desc
17:58.06 libmedia_webrtc_signaling_gtest.a.desc
18:00.03 libaudio_processing_gn.a.desc
18:00.19 make[2]: *** [/DATA2/packages/figue/firefox-esr/src/firefox-60.6.0/config/recurse.mk:33: compile] Error 2
18:00.19 make[1]: *** [/DATA2/packages/figue/firefox-esr/src/firefox-60.6.0/config/rules.mk:442: default] Error 2
18:00.19 make: *** [client.mk:169: build] Error 2
18:00.20 92 compiler warnings present.
18:00.24 Failed to parse ccache stats output: stats updated                       Fri Jan 18 15:44:47 2019
18:00.24 /usr/bin/notify-send --app-name=Mozilla Build System Mozilla Build System Build failed
==> ERROR: Se produjo un fallo en build().
    Cancelando...

The pkg needs a rebuild for libvpx 1.8.x - maybe you follow firefox pkg from the repos that moved to more internal libs.

@BrLi are you sure that you have the community rust/cargo package? Did you clean your previous build with -C?

% LANG=C pacman -Qi rust
Name            : rust
Version         : 1:1.32.0-1
Description     : Systems programming language focused on safety, speed and concurrency
Architecture    : x86_64
URL             : https://www.rust-lang.org/
Licenses        : MIT  Apache
Groups          : None
Provides        : cargo  rustfmt
Depends On      : gcc-libs  llvm-libs  curl  libssh2
Optional Deps   : None
Required By     : None
Optional For    : None
Conflicts With  : cargo  rustfmt
Replaces        : cargo  rustfmt
Installed Size  : 192.62 MiB
Packager        : Johannes L
Build Date      : Thu Jan 17 23:32:42 2019
Install Date    : Fri Jan 18 21:50:31 2019
Install Reason  : Installed as a dependency for another package
Install Script  : No
Validated By    : Signature

You can try editing PKGBUILD and remove "ac_add_options --enable-rust-simd"

@figue I'm afraid that doesn't help, I turn to GCC then.

@BrLi can you try to workaround the build with clang?

env _CLANG=1 makepkg -si

Hi, I'm having issue when building under clean chroot... The chroot system is up-to-date

10:34.60 error: /home/user/firefox-esr/src/firefox-60.4.0/obj-ff/toolkit/library/release/deps/libcssparser_macros-bb0b1a733c43c15c.so: undefined symbol: atexit
10:34.60   --> /home/user/firefox-esr/src/firefox-60.4.0/third_party/rust/cssparser/src/lib.rs:73:14
10:34.60    |
10:34.60 73 | #[macro_use] extern crate cssparser_macros;
10:34.60    |              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
10:34.60 
10:34.60 error: aborting due to previous error
10:34.60 
10:34.61 error: Could not compile `cssparser`.
10:34.61 warning: build failed, waiting for other jobs to finish...
10:34.68 error: /home/user/irefox-esr/src/firefox-60.4.0/obj-ff/toolkit/library/release/deps/libxpcom_macros-39504dc553d554c1.so: undefined symbol: atexit
10:34.68   --> xpcom/rust/xpcom/src/lib.rs:21:1
10:34.68    |
10:34.68 21 | extern crate xpcom_macros;
10:34.69    | ^^^^^^^^^^^^^^^^^^^^^^^^^^
10:34.69 
10:34.69 error: aborting due to previous error
10:34.69 
10:34.69 error: Could not compile `xpcom`.
10:34.69 warning: build failed, waiting for other jobs to finish...

can you help me?

@leewdch pushed new release with that patch. It fixes compilation in IceCat (based off FF-ESR)... I haven't tested though...

Also, clang is now used only if you export the variable _CLANG before building or if you set inside PKGBUILD.

@leewdch Try the patch from https://bugzilla.mozilla.org/show_bug.cgi?id=1479540

Can you tell what's wrong? I'm trying to compile on a clean install but somehow it fails saying I need more than 2 value to extract: https://i.imgur.com/grw1vie.jpg

@AndyRTR thank you. It's really weird, but I tried even in a virtual machine on a Windows host, and it fails too. So, if I maintain this package I need to validate the build. Feel free to delete CC and CXX variables from my PKGBUILD if you want continue compiling with gcc

I see no need to switch to Clang compiler. Using gcc worked well for me building in a clean chroot.

I had to switch to clang instead of gcc to build FF-ESR on my machines... Will test deeply tomorrow, but it should be OK. Is this a problem for someone?

Edit: tests were ok, will push change tomorrow.

PLease test 60.2.0... I can't build it :''''(

I've successfully build FF-ESR 60.0.2 in other machine.

Please test and report if something is wrong.

One note: I've removed python path workaround, it builds fine without it.

Can somebody bump PKGVER and try to see if it builds for you? Mine fails in something related to skia or GL...

50:02.88 during GIMPLE pass: wrestrict
50:02.88 En la función miembro static ‘static void GrGLPath::InitPathObjectPathData(GrGLGpu*, GrGLuint, const SkPath&)’:
50:02.88 en pp_format, en pretty-print.c:1378
50:02.88  void GrGLPath::InitPathObjectPathData(GrGLGpu* gpu,
50:02.88       ^~~~~~~~
50:02.89 Por favor, envíe un informe completo de errores,
50:02.89 con el código preprocesado si es apropiado.
50:02.89 Véase <https://bugs.archlinux.org/> para instrucciones.
50:02.91 make[4]: *** [/DATA2/packages/figue/firefox-esr/src/firefox-60.0.2/config/rules.mk:1049: GrGLPath.o] Error 1

https://aur.archlinux.org/packages/firefox-esr52/

Hi, where can I find the firefox-esr v52.x package ? I want to downgrade it to v52.x.

@bogdanbiv just build it manually in another place:

cd your_desired_folder
git clone https://aur.archlinux.org/firefox-esr.git
cd firefox-esr
makepkg -sic

Or if you are using yaourt or pacaur, just change BUILDIR.

Cheers

Please help - for me Building Firefox ESR fails because it overfills the /tmp partition.

Building Firefox ESR fills my tmp partition and fails: Filesystem Size Used Avail Use% Mounted on tmpfs 7.7G 7.7G 4.0K 100% /tmp

Here is the last of the build log: 0:11.99 make: Leaving directory '/tmp/yaourt-tmp-bogdanbiv/aur-firefox-esr/src/firefox-52.5.2esr/obj-x86_64-pc-linux-gnu' Traceback (most recent call last): File "/usr/lib/python2.7/logging/init.py", line 885, in emit self.flush() File "/usr/lib/python2.7/logging/init.py", line 845, in flush self.stream.flush() IOError: [Errno 28] No space left on device Logged from file logging.py, line 54 ==> ERROR: A failure occurred in package(). Aborting... ==> ERROR: Makepkg was unable to build firefox-esr. ==> Restart building firefox-esr ? [y/N]

I read here that ESR 52.5.1 is still a release candidate and was pushed by mistake to releases folder in Mozilla's public FTP.

So I maintain my last comment or you can download tar.bz2 here.

Cheers

PD: I'm running the candidate just fine