Package Details: chromium-dev 53.0.2783.2-1

Git Clone URL: https://aur.archlinux.org/chromium-dev.git (read-only)
Package Base: chromium-dev
Description: The open-source project behind Google Chrome (Dev Channel)
Upstream URL: http://www.chromium.org
Licenses: BSD
Submitter: None
Maintainer: sl1pkn07
Last Packager: sl1pkn07
Votes: 147
Popularity: 0.467346
First Submitted: 2010-05-17 09:04
Last Updated: 2016-07-02 08:49

Dependencies (45)

Required by (3)

Sources (10)

Latest Comments

sl1pkn07 commented on 2016-06-26 15:29

@misc thanks for the patch. but without patch build for me ok

greetings

misc commented on 2016-06-23 15:45

Dunno if 53.0.2774.3 is already affected by…

../../content/public/common/common_param_traits_macros.h:25:92: fatal error: third_party/WebKit/public/platform/modules/permissions/permission_status.mojom.h: file not found

… but anyway, solved for me with this patch: http://pastebin.com/raw/xz5ghExY

ogarcia commented on 2016-06-17 09:44

@sl1pkn07 The error is here http://pastebin.com/CtSr29jS

The full log can be donwloaded from here https://db.tt/FX5GtMOR

sl1pkn07 commented on 2016-06-16 15:02

@ogracia: please paste in a pastebin the error

greetings

ogarcia commented on 2016-06-16 13:55

@sl1pkn07 I don't know why, but if I try to make package without gtk2 installed gives me missing gtk2 error :(

You can make test with a clean build environment using chroot: https://wiki.archlinux.org/index.php/DeveloperWiki:Building_in_a_Clean_Chroot

TingPing commented on 2016-06-15 23:00

Just wanted to note that:

If _use_bundled_clang is disabled system clang will fail to build. Probably requires newer development version.

If _use_clang is disabled the resulting binary just constantly crashes with memory access errors.

sl1pkn07 commented on 2016-06-14 18:48

gtk2 for what?

_use_gtk3=1 # If set 1, then build with GTK3 support, if set 0, then build with GTK2
....
...
..
# Build with GTK3?
if [ "${_use_gtk3}" = "1" ]; then
depends+=('gtk3')
_launcher_gtk='GTK=3'
optdepends+=('libappindicator-gtk3: Needed for show systray icon in the panel in plasma-next (KF5)')
elif [ "${_use_gtk3}" = "0" ]; then
depends+=('gtk2')
optdepends+=('libappindicator-gtk2: Needed for show systray icon in the panel in plasma-next (KF5)')
fi

ogarcia commented on 2016-06-14 18:40

You forget gtk2 as build dependency

bryanparadis commented on 2016-06-01 00:30

When building this package make sure you have at least 9GB of space on /tmp. Build finished but makepkg didn't have enough room to generate the packages with a 7.8GB /tmp. Managed to recover from it without rebuilding. /tmp will be by default 50% of your ram. You can override this by adding an fstab entry with the size option.

neeks commented on 2016-05-19 01:56

I've been getting errors the past two weeks in `third_party/WebKit/Source/build/scripts/make_css_value_keywords.py` and `third_party/WebKit/Source/build/scripts/make_css_property_names.py`.

Turns out that the gperf package - 3.0.4.7 - that was pushed to `extra` 16 days ago is to blame. Downgrading gperf to 3.0.4.6 solved those specific build errors.

j1simon commented on 2016-05-18 18:01

Yes, I have disabled the [multilib] repository.

$ pacaur -y chromium-dev
:: resolving dependencies...
:: no results found for lib32-gcc-libs
:: no results found for lib32-zlib

chrome://nacl -> http://pix.toile-libre.org/upload/original/1463594452.png

sl1pkn07 commented on 2016-05-18 17:55

in a clean environment?

j1simon commented on 2016-05-18 17:54

@sl1pkn07 I've compiled it with the line ("makedepends_x86_64=('lib32-gcc-libs' 'lib32-zlib')") commented without problems and it compiles NaCl plugin.

sl1pkn07 commented on 2016-05-18 13:22

because is need for build the NaCL stack

j1simon commented on 2016-05-18 10:21

Why "makedepends_x86_64=('lib32-gcc-libs' 'lib32-zlib')"?

LinguinePenguiny commented on 2016-04-23 08:09

Compile proxy/plugin_message_filter.cc
[FLAGS I REMOVED]
Traceback (most recent call last):
File "../native_client/build/build_nexe.py", line 1136, in <module>
sys.exit(Main(sys.argv))
File "../native_client/build/build_nexe.py", line 1108, in Main
out = build.Compile(filename)
File "../native_client/build/build_nexe.py", line 574, in Compile
self.CleanOutput(outd)
File "/tmp/yaourt-tmp-aaron/aur-chromium-dev/src/chromium-51.0.2704.19/native_client/build/build_nexe_tools.py", line 81, in CleanOutput
if IsFile(out):
File "/tmp/yaourt-tmp-aaron/aur-chromium-dev/src/chromium-51.0.2704.19/native_client/build/build_nexe_tools.py", line 41, in IsFile
return os.path.isfile(FixPath(path))
File "/tmp/yaourt-tmp-aaron/aur-chromium-dev/src/chromium-51.0.2704.19/native_client/build/build_nexe_tools.py", line 32, in FixPath
path = os.path.normpath(os.path.join(os.getcwd(), path))
OSError: [Errno 2] No such file or directory
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().

It's not compiling, still.

sl1pkn07 commented on 2016-04-22 13:23

@juston_li done

tnx

juston_li commented on 2016-04-22 02:17

install: cannot stat 'chromium-51.0.2704.19/chrome/installer/linux/common/default-app.tempate': No such file or directory

Looks like just a typo on line 627 tempate -> template

wicast commented on 2016-04-21 13:09

MimeType is missing in .desktop file. xdg-open is not working well.

axfelix commented on 2016-04-19 16:33

h264 VAAPI hardware decoding is working again with the most recent build (51.0.2704)! Thanks, folks.

axfelix commented on 2016-04-15 16:51

Just confirmed it's back to working for flash video containers, at least, just not HTML5. I think this was the case up until Chromium 44-45 or so, when HTML5 h264 would actually work with the vaapi patches, but it seems that lasted a fairly short time.

axfelix commented on 2016-04-15 16:38

Build works with the minizip patch and it's back to showing Video Decode as hardware accelerated in chrome://gpu with the new vaapi patch (it went back to showing software only in 50) but GPU decoding still isn't kicking in after that change, in 51.

misc commented on 2016-04-10 17:08

Patch for the minizip update: http://pastebin.com/raw/QCqSDam5

If I understand correctly, the old version interpreted a 0 as sign to pick strcmp, the new one directly takes a function pointer and falls to strcmp when it's NULL.

guiodic commented on 2016-04-06 19:34

"Seems like GPU video decoding patches may have stopped working between 49 and 50..."

Yes, it's true. But when I run chromium from the cli I get the va-api message. Nevertheless chromium takes a lot of cpu.

Maybe an upstream bug related to the switch to GTK3.

axfelix commented on 2016-03-24 22:16

Seems like GPU video decoding patches may have stopped working between 49 and 50...

sl1pkn07 commented on 2016-03-06 22:39

yeah!

misc commented on 2016-03-06 21:51

I don't have the line '-Dicu_use_data_file_flag=0', just try after killing it entirely.

sl1pkn07 commented on 2016-03-06 21:32

obj/mojo/shell/runner/mojo_runner_host_lib.init.o: In function `mojo::shell::CallLibraryEarlyInitialization(void*)':
init.cc:(.text._ZN4mojo5shell30CallLibraryEarlyInitializationEPv+0x16): undefined reference to `base::i18n::GetRawIcuMemory()'


mmmm

moscar commented on 2016-03-06 21:11

This currently fails to build if gtk2 is not installed:

Updating projects from gyp files...
Package gdk-2.0 was not found in the pkg-config search path.
Perhaps you should add the directory containing `gdk-2.0.pc'
to the PKG_CONFIG_PATH environment variable
No package 'gdk-2.0' found
gyp: Call to 'pkg-config --cflags gdk-2.0' returned exit status 1 while in /sources/chromium-dev/src/chromium-50.0.2633.3/build/linux/system.gyp.
==> ERROR: A failure occurred in build().

I'm guessing the `-Duse_gtk3` flag is not enough, or something else is missing.

misc commented on 2016-03-06 15:39

One revert and a workaround I need to compile ATM. Dunno if the reverted commit already made it into this tarball.

http://pastebin.com/raw/2FKGi944 ← revert
http://pastebin.com/raw/nz3KrTbc ← system_icu=1 issue

moscar commented on 2016-01-30 06:05

imagemagick should be in makedepends since the convert command is used here: https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=chromium-dev#n453

juston_li commented on 2016-01-27 17:40

The constant was removed in a cleanup patch that sl1pkn07 linked in the issue tracker
https://chromium.googlesource.com/chromium/src/+/8921a7faaf42c6e216ad3de32093a0c3a121e696.

Looks to be a rather silly and trivial error, not sure why they haven't fixed or how it got though review.
The commit simply removed the declaration of "kDefaultTintFrameInactive" but left a reference to it on line 1226.
Not sure if the constant was accidentally removed or they forgot to remove the reference, I'll try and look into it but safer way to fix is just to add the constant again.
Here's the simple patch tested working:
https://drive.google.com/file/d/0BxTqqkmPl8mTODFLM29nM3lFc3c/view?usp=sharing

prazola commented on 2016-01-27 16:30

"kDefaultTintFrameInactive" should be a constant indicating a color. Have you checked if the constant is missing in the some theme source inside the project?

sl1pkn07 commented on 2016-01-25 01:04

@corepoint

before, the patch is applied with this:
https://aur.archlinux.org/cgit/aur.git/commit/?h=chromium-dev&id=cf179b8e640c8ed6cc0600fae730d6e60f7d1515 (the sed line)

now, using the chromium [extra] patch. (more elaborated than my "patch")

for widevine(netflix)support in this package you need install this package: https://aur.archlinux.org/packages/chromium-widevine-dev/

I not have netflix account, but working with this example page
http://www.dash-player.com/demo/drm-test-area/

greetings

CorePoint commented on 2016-01-24 17:55

@sl1pkn07

Hey, can you elaborate? They don't seem to have widevine in their repository, so i'm out of luck?

sl1pkn07 commented on 2016-01-22 05:54

congrats!

greetings

juston_li commented on 2016-01-22 05:44

Built fine this time
Thanks

juston_li commented on 2016-01-22 04:12

32GB ram
Though it might have been full as I had just built 2623.0 then it got updated to 2623.13. If yaourt didn't clear it then having both packages might have been a bit too much =p

I'll try again with /tmp clear
Thanks

sl1pkn07 commented on 2016-01-22 03:56

maybe a insufficient space? :S

http://wstaw.org/m/2016/01/22/Screenshot_20160122_045937.png

juston_li commented on 2016-01-22 02:45

2623.13 build fails at the end for me [19398/19407] =/
http://pastebin.com/pW14MEcK

sl1pkn07 commented on 2016-01-15 17:51

@Prazola:

in this moment, no

https://code.google.com/p/chromium/issues/detail?id=575038

@CorePoint:

that patch is applied, but in another way

prazola commented on 2016-01-15 17:50

No more GTK3 default in last build? :D

CorePoint commented on 2016-01-13 10:37

Hey,

i would really like to have a binary for this in the aur too.
The package in the archlinuxcn fails too use widevine.
When i install chromium-widevine-dev from the aur, it shows up in chrome://plugins but not in chrome://components. So netflix shows its usual error "missing widevine, make sure to enable it blah".

It seems they didn't inculde this patch.
https://code.google.com/p/chromium/issues/detail?id=429452#c16

Just telling, because maybe nobody saw a reason to make a chromiun-dev binary aur, and now has a reason.

Have a nice day

sl1pkn07 commented on 2015-12-19 10:52

@misc with pnacl?

misc commented on 2015-12-18 18:40

I've been able to use the -lite.tar.xz tarballs without issue, which are still around 450 MiB.

sl1pkn07 commented on 2015-12-16 16:51

yes and no

https://groups.google.com/a/chromium.org/forum/#!topic/chromium-packagers/c6lVrV8IuwA/discussion

prazola commented on 2015-12-16 16:11

The source tar.xz is about 1.4GB instead of the usal 380~ MB. Is it ok? :/

axfelix commented on 2015-12-06 20:47

@skzzorz

thanks for finding that link -- looks like it is indeed explicitly disabled on mesa.

veluria commented on 2015-12-01 17:15

Does anybody know if there's a binary build with GTK3 support?

sl1pkn07 commented on 2015-11-12 17:50

nope, chromium-dev still need bundled libvpx (it needs headers not included in libvpx 1.5)

'vpx/vpx_codec_impl_bottom.h',
'vpx/vpx_codec_impl_top.h',

greetings

prazola commented on 2015-11-12 14:57

Does 48.0.2560.0-1 build support libvpx 1.5?

prazola commented on 2015-11-10 16:21

Wow, latest nightly builds are a bit to heavy :D
https://download-chromium.appspot.com/dl/Linux_x64
The executable reach 2gb when uncompressed. Do they have some build problem?

skzzorz commented on 2015-11-08 01:36

@axfelix
a qucik search at cs.chromium.org leads to: https://code.google.com/p/chromium/codesearch#chromium/src/content/browser/gpu/browser_gpu_memory_buffer_manager.cc&sq=package:chromium&l=102&rcl=1446928296
it doesn't look like it's disabled by OS-related macro.

axfelix commented on 2015-11-05 22:04

I'm noticing that the --enable-native-gpu-memory-buffers flag doesn't seem to be working on Linux in newest chromium-dev despite the fact that this functionality *should* now be there; is it possible that this is blacklisted at compile time on non-ChromeOS platforms the way that vaapi was? Might be fun to have a look into that...

axfelix commented on 2015-11-03 23:52

Thanks!

skzzorz commented on 2015-10-29 09:59

@quartz55 @axfelix
Pre-compiled builds for x86-64 with VA-API patch can be found at archlinuxcn's unofficial community repo at repo.archlinuxcn.org

ethail commented on 2015-10-27 11:46

Looking at the post on the chrome releases blog [0] for this release, shouldn't we be using 48.0.2541.0 instead of 48.0.2540.0?

Not that it matters that much, as a release should happen shortly, I guess, but maybe for a future we could pay a little more attention when two versions targeted at diferent OSes are released.

[0] http://googlechromereleases.blogspot.com.es/2015/10/dev-channel-update_20.html

EDIT (a few hours later): I realized my tone could get misinterpreted, by "we" I mean everyone who uses this package and cares even a little about it, not just the maintainer.

prazola commented on 2015-10-21 12:10

I have a script for precompiled snapshot. I prefer to compile this package myself.

danbruegge commented on 2015-10-20 07:41

@axfelix +1 for Precompiled builds

quartz55 commented on 2015-10-17 16:52

@axfelix Precompiled builds would be awesome, it does take quite a while for the builds to finish :p

axfelix commented on 2015-09-08 00:36

Any interest in doing precompiled builds? I used this because I wanted the vaapi patch (super thrilled to someone maintaining it in the AUR, think you're the only one), and it works great, but man, a Chromium build is not for the faint of heart :)

sl1pkn07 commented on 2015-08-31 20:17

You can't wait?

mortan commented on 2015-08-31 19:06

@sl1pkn07: Any plans to update this package?

kdex commented on 2015-08-11 21:50

@sl1pkn07: There's probably a typo in the optional dependencies; it's "libexif", not "llibexif".

sl1pkn07 commented on 2015-07-18 23:30

yes. i know the situation. for the moment, just remove python2-ply, build chromium, and then reinstall. the patch for fix the issue still no merge

https://github.com/dabeaz/ply/issues/66

https://bugs.archlinux.org/task/45625

emerino commented on 2015-07-18 23:22

Just to give you a heads up, current python-ply 3.6 WON'T work, it'll cause an error at compile time. You have to either downgrade to 3.4 or use python-ply's git version.

https://github.com/dabeaz/ply/issues/63

andre.vmatos commented on 2015-07-06 23:37

Anyone has a repository or could post somewhere x86_64 version of this package, please? Thank you.

prazola commented on 2015-06-06 16:16

I was missing some gcc 32bit libs, I'm recompiling right now.

prazola commented on 2015-06-06 14:15

Wow, no, not installed. Idk why, I'm upgrading from a previous version of you package, really strange. Thx

Core2 stop on "Compile posix/nacl_timestamp.c" without informations about the error.

sl1pkn07 commented on 2015-06-06 13:59

you have installed bison? (strange, because is part of base-devel (mandatory to build any AUR package)

prazola commented on 2015-06-06 13:23

i7 error
http://pastebin.com/T5F4Z1iV

sl1pkn07 commented on 2015-06-06 12:39

log?

prazola commented on 2015-06-06 12:10

Build problems on two different platforms (native, o3) ninja subcommand failed.

Build dir is in my home.

Config 1: core2 p8600 - 4GB ddr2 - ssd 240gb
Config 2: i7 4790k - 16GB ddr3 - ssd 120gb
Config 3: i5 450m - 8GB ddr3 - hdd 640gb -> build OK!

Python2-ply not installed, so what should I do?

vdemin commented on 2015-06-03 06:51

@foutrelis, you are right. I changed location of the build folder from tmpfs to hard drive and decreased the tmpfs size, and compilation finished successfully. Thank you for the help!

foutrelis commented on 2015-06-02 20:44

> c++: internal compiler error: Killed (program cc1plus)

That indicates that you're running out of memory during compilation.

sl1pkn07 commented on 2015-06-02 20:34

build ok here (x86_64)

snip the clang build log

http://sl1pkn07.wtf/paste/view/561d7015

sl1pkn07 commented on 2015-06-02 20:32

build ok here (x86_64)

vdemin commented on 2015-06-02 20:03

I can't build the package on x86_64:
http://pastebin.com/CZJsQh60

hotaru commented on 2015-05-06 16:13

yeah, if you remove '--insecure' it should be ok.

sl1pkn07 commented on 2015-05-06 15:59

oh. you mean the '--insecure' flag?. is removed in the next version (now in build/test)

hotaru commented on 2015-05-06 15:01

> sha1sums=("$(curl -sL "https://commondatastorage.googleapis.com/chromium-browser-official/chromium-${pkgver}.tar.xz.hashes" --insecure | grep sha1 | cut -d " " -f3)"
doesn't that completely negate the point of using sha1sums?

sl1pkn07 commented on 2015-05-03 12:29

yes. that is

but for the moment, the plugin don't work in a linux platform

juston_li commented on 2015-05-03 03:34

Hmm seems like widevine build included again, got this error:
error: failed to commit transaction (conflicting files)
chromium-dev: /usr/lib/chromium-dev/libwidevinecdm.so exists in filesystem

Removed'chromium-widevine-dev' package then I could upgrade. chromium-dev shows widevine in chrome://plugins

eggze commented on 2015-04-29 11:44

It helped, thank you.

sl1pkn07 commented on 2015-04-27 13:23

uninstall python2-ply first

eggze commented on 2015-04-27 12:31

Build failed on my machine with similar error as katta's one:
FAILED: cd ../../ui/accessibility; python ../../tools/json_schema_compiler/compiler.py "ax_enums.idl" "--root=../.." "--destdir=../../out/Release/gen" "--namespace=ui" "--generator=cpp" "--include-rules="
Traceback (most recent call last):
File "../../tools/json_schema_compiler/compiler.py", line 191, in <module>
opts.namespace, opts.impl_dir, include_rules)
File "../../tools/json_schema_compiler/compiler.py", line 53, in GenerateSchema
api_def = schema_loader.LoadSchema(schema)
File "/home/jprokop/Downloads/builds/chromium-dev/src/chromium-44.0.2376.0/tools/json_schema_compiler/schema_loader.py", line 90, in LoadSchema
api_defs = idl_schema.Load(schema_path)
File "/home/jprokop/Downloads/builds/chromium-dev/src/chromium-44.0.2376.0/tools/json_schema_compiler/idl_schema.py", line 525, in Load
return Process(contents, filename)
File "/home/jprokop/Downloads/builds/chromium-dev/src/chromium-44.0.2376.0/tools/json_schema_compiler/idl_schema.py", line 535, in Process
idl = idl_parser.IDLParser().ParseData(contents, filename)
File "/home/jprokop/Downloads/builds/chromium-dev/src/chromium-44.0.2376.0/tools/json_schema_compiler/../../ppapi/generators/idl_parser.py", line 906, in __init__
IDLLexer.__init__(self)
File "/home/jprokop/Downloads/builds/chromium-dev/src/chromium-44.0.2376.0/tools/json_schema_compiler/../../ppapi/generators/idl_lexer.py", line 215, in __init__
self.lexobj = lex.lex(object=self, lextab=None, optimize=0)
File "/usr/lib/python2.7/site-packages/ply/lex.py", line 893, in lex
if '.' not in lextab:
TypeError: argument of type 'NoneType' is not iterable
ninja: build stopped: subcommand failed.

My version of python2-ply is 3.6-1

sl1pkn07 commented on 2015-04-26 23:20

ok, is because chromium build system don't like the new python2-ply 1.5.0

try now

katta commented on 2015-04-26 15:48

build failed:

[1226/17650] cd ../../third_party/zlib; python ../../tools/generate_shim_headers/generate_shim_headers.py --headers-root . --output-directory ../../out/Release/gen/shim_headers/zlib/target zlib.h --generate
FAILED: cd ../../third_party/WebKit/Source/bindings/scripts; python blink_idl_parser.py ../../../../../out/Release/gen/blink/bindings/scripts
Traceback (most recent call last):
File "blink_idl_parser.py", line 456, in <module>
sys.exit(main(sys.argv))
File "blink_idl_parser.py", line 452, in main
parser = BlinkIDLParser(outputdir=outputdir, rewrite_tables=True)
File "blink_idl_parser.py", line 428, in __init__
picklefile=picklefile)
File "/usr/lib/python2.7/site-packages/ply/yacc.py", line 3232, in yacc
read_signature = lr.read_pickle(picklefile)
File "/usr/lib/python2.7/site-packages/ply/yacc.py", line 1990, in read_pickle
in_f = open(filename, 'rb')
IOError: [Errno 2] No such file or directory: '../../../../../out/Release/gen/blink/bindings/scripts/parsetab.pickle'
ninja: build stopped: subcommand failed.

c00w commented on 2015-04-24 03:05

I don't think HIDPI support is working anymore.

sl1pkn07 commented on 2015-04-19 10:06

@juston_li

fixed

also now use a chromium-launcher tarball instead of git and fixed minnor issues

greetings

juston_li commented on 2015-04-19 00:27

Got this error trying to run chromium-dev, compiled with GCC:

Can't locate File/BaseDir.pm in @INC (you may need to install the File::BaseDir module) (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at /usr/bin/chromium-dev line 16.
BEGIN failed--compilation aborted at /usr/bin/chromium-dev line 16.

Fixed by installing perl-file-basedir

juston_li commented on 2015-04-19 00:26

Got this error after compiling with GCC:

Can't locate File/BaseDir.pm in @INC (you may need to install the File::BaseDir module) (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at /usr/bin/chromium-dev line 16.
BEGIN failed--compilation aborted at /usr/bin/chromium-dev line 16.

Fixed by installing perl-file-basedir

sl1pkn07 commented on 2015-04-18 21:12

update to 44 series

- now uses a launcher like a chromium in [extra]. please check the instruction after installation
- fix the NaCL installation
- for the moment now use a system clang instead of build (fail build)
- build widevine lib not need anymore. please install a chromium-widevine/chromium-widevine-dev instead

greetings

sl1pkn07 commented on 2015-04-18 21:10

update to 44 series

- now uses a launcher like a chromium in [extra]. please check the instruction after installation
- fix the NaCL installation
- for the moment now use a system clang instead of build (fail build)

greetings

walkindude commented on 2015-04-10 19:11

@Valeth
no need, looks like the latest build solved this issue

Valeth commented on 2015-04-09 19:41

@walkindude

try to launch chromium-dev with the MESA_GL_VERSION_OVERRIDE environment variable set to 4.0

sl1pkn07 commented on 2015-04-08 19:58

ok. in the next release

greetings

@walkindude. sorry. try add the issue in crbug.com (chromium issue tracker). i can't help you :(

juston_li commented on 2015-04-08 19:41

Should probably make ccache usage optional by making it a build option.
It takes up about 1GB to storage compiling with GCC so I feel it should require explicit action to enable.

That aside, I've been using ccache for awhile now and if you don't mind giving up some disk space, ccache really speeds things up.
Compile times from 40-50m down to 5m with ccache on an SSD

walkindude commented on 2015-04-08 14:45

...
[8293:8293:0408/174044:ERROR:logger.cc(46)] Too many GL errors, not reporting any more for this context. use --disable-gl-error-limit to see all errors.
[8293:8293:0408/174044:ERROR:gles2_cmd_decoder.cc(11546)] [GroupMarkerNotSet(crbug.com/242999)!:D0D63D32FF3B0000]GL ERROR :GL_INVALID_OPERATION : glTexStorage2DEXT: <- error from previous GL command
[8293:8293:0408/174044:ERROR:shader_manager.cc(128)] Shader translator allowed/produced an invalid shader unless the driver is buggy:
--Log from shader translator--

--original-shader--
#define TexCoordPrecision highp
attribute TexCoordPrecision vec4 a_position; attribute TexCoordPrecision vec2 a_texCoord; attribute float a_index; uniform mat4 matrix; uniform TexCoordPrecision vec2 quad[4]; uniform TexCoordPrecision vec4 vertexTexTransform; varying TexCoordPrecision vec2 v_texCoord; void main() { vec2 pos = quad[int(a_index)]; gl_Position = matrix * vec4(pos, a_position.z, a_position.w); v_texCoord = a_texCoord * vertexTexTransform.zw + vertexTexTransform.xy; }
--translated-shader--
#version 150
in vec4 a_position;
in vec2 a_texCoord;
in float a_index;
uniform mat4 matrix;
uniform vec2 quad[4];
uniform vec4 vertexTexTransform;
out vec2 v_texCoord;
void main(){
vec2 pos = quad[int(clamp(float(int(a_index)), 0.0, float(3)))];
(gl_Position = (matrix * vec4(pos, a_position.z, a_position.w)));
(v_texCoord = ((a_texCoord * vertexTexTransform.zw) + vertexTexTransform.xy));
}

--info-log--
0:1(10): error: GLSL 1.50 is not supported. Supported versions are: 1.10, 1.20, 1.30, 1.00 ES, and 3.00 ES

[8293:8293:0408/174044:ERROR:shader_manager.cc(128)] Shader translator allowed/produced an invalid shader unless the driver is buggy:
--Log from shader translator--

--original-shader--
#define TexCoordPrecision mediump
#define SamplerType sampler2D
#define TextureLookup texture2D
#define ApplyBlendMode(X, Y) (X)
precision mediump float; varying TexCoordPrecision vec2 v_texCoord; uniform SamplerType s_texture; uniform float alpha; void main() { vec4 texColor = TextureLookup(s_texture, v_texCoord); gl_FragColor = ApplyBlendMode(texColor * alpha, 0.0); }
--translated-shader--
#version 150
out vec4 webgl_FragColor;
in vec2 v_texCoord;
uniform sampler2D s_texture;
uniform float alpha;
void main(){
vec4 texColor = texture(s_texture, v_texCoord);
(webgl_FragColor = (texColor * alpha));
}

--info-log--
0:1(10): error: GLSL 1.50 is not supported. Supported versions are: 1.10, 1.20, 1.30, 1.00 ES, and 3.00 ES

[8251:8251:0408/174054:ERROR:extension_downloader.cc(695)] Invalid URL: '' for extension dgpdioedihjhncjafcpgbbjdpbbkikmi
[8251:8488:0408/174055:ERROR:get_updates_processor.cc(243)] PostClientToServerMessage() failed during GetUpdates

from the latest build

I have:
walkindude@Desktop> glxinfo | grep version ~
server glx version string: 1.4
client glx version string: 1.4
GLX version: 1.4
OpenGL core profile version string: 3.3 (Core Profile) Mesa 10.6.0-devel (git-eb51c6d)
OpenGL core profile shading language version string: 3.30
OpenGL version string: 3.0 Mesa 10.6.0-devel (git-eb51c6d)
OpenGL shading language version string: 1.30
OpenGL ES profile version string: OpenGL ES 3.0 Mesa 10.6.0-devel (git-eb51c6d)
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.00

no more chromium for me?

harryharryharry commented on 2015-04-06 10:25

@sl1pkn07 Yes I already changed the makeflags. But it's okay, at least I now know I should take a beer in the meantime :)

sl1pkn07 commented on 2015-04-05 22:27

@harryharryharry

yes. is normal. but can try edit the /etc/makepkg.conf and edit/comment out the MAKEFLAGS= variable with -jyour_number_of_cores (-j4 for quad-core or -j8 if have quad-core with HT) (if you had it before)

the pkgbuild has been fixed

greething and sorry

harryharryharry commented on 2015-04-05 21:35

@sl1pkn07 I had the same error as rekoil, gokcen & daurnimator. Adding ccache as a dependency solved it for me. Thanks!
The build took over 6 hours (with quad-core, 4GB ram) is that normal ?

sl1pkn07 commented on 2015-04-05 12:02

oh.

try to add 'ccache' to a makedepends()

daurnimator commented on 2015-04-05 03:50

I get same error:

-> Starting building Chromium...
Updating projects from gyp files...
Traceback (most recent call last):
File "../../tools/compile_test/compile_test.py", line 65, in <module>
sys.exit(DoMain(sys.argv[1:]))
File "../../tools/compile_test/compile_test.py", line 52, in DoMain
stderr=subprocess.PIPE)
File "/usr/lib/python2.7/subprocess.py", line 710, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1335, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
gyp: Call to 'python ../../tools/compile_test/compile_test.py --code "int main() { return 0; }" --run-linker --on-success "-lwebp -lwebpdemux" --on-failure "-lwebp" -- -lwebpdemux' returned exit status 1.
==> ERROR: A failure occurred in build().
Aborting...
==> ERROR: Makepkg was unable to build .

illuser commented on 2015-04-05 01:31

Getting the same error, even with python2-lxml added to dependencies.

sl1pkn07 commented on 2015-04-04 11:11

@gokcen @rekoil

add 'python2-lxml' to dependencies

greetings

sl1pkn07 commented on 2015-04-04 11:01

@gokcen @rekoil

try to change:

'python2-beautifulsoup3'

with

'python2-beautifulsoup4' 'python2-lxml'

greetings

sl1pkn07 commented on 2015-04-04 11:00

try to change:

'python2-beautifulsoup3'

with

'python2-beautifulsoup4' 'python2-lxml'

greetings

gokcen commented on 2015-04-04 10:21

Getting the same error here. FWIW, _use_system_clang=1 does not fix the error.

rekoil commented on 2015-04-03 18:07

I'm having the following issue when attempting to compile the latest chromium-dev. Not sure exactly what is going on, perhaps there's some dependency missing? Can't really test because it keeps starting from scratch every time I run makepkg.

-> Starting building Chromium...
Updating projects from gyp files...
Traceback (most recent call last):
File "../../tools/compile_test/compile_test.py", line 65, in <module>
sys.exit(DoMain(sys.argv[1:]))
File "../../tools/compile_test/compile_test.py", line 52, in DoMain
stderr=subprocess.PIPE)
File "/usr/lib/python2.7/subprocess.py", line 710, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1335, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
gyp: Call to 'python ../../tools/compile_test/compile_test.py --code "int main() { return 0; }" --run-linker --on-success "-lwebp -lwebpdemux" --on-failure "-lwebp" -- -lwebpdemux' returned exit status 1.
==> ERROR: A failure occurred in build().
Aborting...

gokcen commented on 2015-03-21 07:38

I hope hidpi support will get better with the latest fix released: https://code.google.com/p/chromium/issues/detail?id=143619#c71

sl1pkn07 commented on 2015-03-15 22:55

@jebarb please fill bugreport in http://crbug.com

greetings

jebarb commented on 2015-03-15 22:29

@sl1pkn07

Package finally built with hidpi flag enabled. Everything works, but the menus are nearly unusable because the text size is about 3x everything else. Right click has the same issue, and the boxes for both are popping up in the wrong place. The menu pops up in the middle of the screen, and the right click menu is all over the place.

jebarb commented on 2015-03-15 20:11

@sl1pkn07

Scratch that. Build eventually failed at the same place for the same reason as the system clang and gcc, though it annoyingly happened when almost complete instead of just starting. I'm running 64 bit with multilib enabled.

jebarb commented on 2015-03-15 19:58

Both systems are 64 with multilib enabled.

jebarb commented on 2015-03-15 19:57

@sl1pkn07

Looks like I was somehow running out of memory, causing segfaults. Nothing useful in any logs. Computer was becoming 99% unresponsive. Strange, because it was failing on a system with 12G memory with 6G tmpfs. Bumped it up to 8G and added a swapfile, now the build is moving along fine.

sl1pkn07 commented on 2015-03-14 08:13

@jbarb

For first error. Please paste more log (search 'FAILED' string).
For second error, i'll investigare.

One cuestión. You use 32 or 64 bits?

Greetings

jebarb commented on 2015-03-13 23:55

Package is failing to build with packaged clang, system clang, and gcc on two devices.

Error with packaged clang:

[1837/2229] Building CXX object tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o
FAILED: /usr/bin/c++ -DCLANG_ENABLE_ARCMT -DCLANG_ENABLE_OBJC_REWRITER -DCLANG_ENABLE_STATIC_ANALYZER -DGTEST_HAS_RTTI=0 -D_DEBUG -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -fPIC -fvisibility-inlines-hidden -Wall -W -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wno-missing-field-initializers -pedantic -Wno-long-long -Wno-maybe-uninitialized -Wno-comment -std=c++11 -ffunction-sections -fdata-sections -fno-common -Woverloaded-virtual -fno-strict-aliasing -O3 -Itools/clang/lib/ASTMatchers/Dynamic -I/tmp/pacaurtmp-jebarb/chromium-dev/src/chromium-42.0.2311.22/third_party/llvm/tools/clang/lib/ASTMatchers/Dynamic -I/tmp/pacaurtmp-jebarb/chromium-dev/src/chromium-42.0.2311.22/third_party/llvm/tools/clang/include -Itools/clang/include -Iinclude -I/tmp/pacaurtmp-jebarb/chromium-dev/src/chromium-42.0.2311.22/third_party/llvm/include -UNDEBUG -fno-exceptions -fno-rtti -MMD -MT tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o -MF tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o.d -o tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o -c /tmp/pacaurtmp-jebarb/chromium-dev/src/chromium-42.0.2311.22/third_party/llvm/tools/clang/lib/ASTMatchers/Dynamic/Registry.cpp
c++: internal compiler error: Killed (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://bugs.archlinux.org/> for instructions.
[1837/2229] Building CXX object tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaDeclCXX.cpp.o
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().


Error with system clang (gcc error is similar):

-> Starting building Chromium...
File "build/gyp_chromium", line 118
except SyntaxError, e:
^
SyntaxError: invalid syntax
==> ERROR: A failure occurred in build().

sl1pkn07 commented on 2015-03-11 16:28

@misc

you can help me with your patch? in the 43 series has been changed. i'm not sure if my changes is a best option to handle it

I used (revert) this patchset https://codereview.chromium.org/980603003, and apply (little modified) your patch

this is ok?

@ethail

added (in the next release). thanks :)

greetings

sl1pkn07 commented on 2015-03-11 16:27

@misc

you can help me with your patch? in the 43 series has been changed. i'm not sure if my changes is a best option to handle it

I used (revert) this patchset https://codereview.chromium.org/980603003, and apply (little modified) your patch

this is ok?

@ethail

added (in the next release). thanks and greetings :)

greetings

ethail commented on 2015-03-09 22:54

Would it be too much trouble to add some kind of config option that results on using ccache when the system clang is enabled?

It would have to change lines 364 onwards for something like this:

msg2 "Setup system clang"
export CC="ccache clang -Qunused-arguments"
export CXX="ccache clang++ -Qunused-arguments"
export CCACHE_CPP2=yes
export CCACHE_SLOPPINESS=time_macros
export CXXFLAGS="${CXXFLAGS} -Wno-unknown-warning-option"

For those that have set-up ccache, the change could result in a speed-up of subsequent rebuilds or updates.

I know anyone could make that changes manually every time or something like that, but having it as a config option in the PKGBUILD seemed a better way to handle it.

Thanks in advance

sl1pkn07 commented on 2015-02-27 10:38

Ops!

I whill fix when go to home

Greerings

ledti commented on 2015-02-27 04:51

I noticed that the chromium-dev wrapper looks for /etc/chromium/default rather than /etc/chromium-dev/default, which makes chromium-dev incompatible with all the pepper-flash packages unless you also have chromium installed (or you edit the wrapper).

Is that intentional?

sl1pkn07 commented on 2015-02-25 05:42

ouch!

misc commented on 2015-02-25 01:15

Just change the COMPLETE to SUCCESS in the hunk.

Not sure if there weren't any other changes between 2311 and 2314, though.

sl1pkn07 commented on 2015-02-24 21:42

Thanks @misc but your patch, but don't work in 42.0.2311.4 (with some changes like remove this http://sl1pkn07.wtf/paste/view/670f462a)

Hunk #11 FAILED at 465.
Hunk #12 succeeded at 477 (offset -6 lines).
1 out of 12 hunks FAILED -- saving rejects to file ui/file_manager/file_manager/background/js/media_import_handler.js.rej

http://sl1pkn07.wtf/paste/view/29ac699d

and please don't use pastebin, that breaks the EOL

greetings

misc commented on 2015-02-24 16:43

For those who don't want the Google Analytics crap built into their Chromium:
http://pastebin.com/raw.php?i=ZUJS8ecT

sl1pkn07 commented on 2015-02-20 15:36

how you update chromium dev? (which helper i mean)

because if '_use_widevine' is set to '1', then set makedepend directly on 'chromium-widevine-dev'

see line 90:

# Use Widevide?
[ "${_use_widevine}" = "1" ] && makedepends+=('chromium-widevine-dev')

greetings

sl1pkn07 commented on 2015-02-20 15:35

how you update chromium dev? (which helper i mean)

mecause if _use_widevine is set to 1 then make a makedepend directly on 'chromium-widevine-dev'

see line 90:

# Use Widevide?
[ "${_use_widevine}" = "1" ] && makedepends+=('chromium-widevine-dev')

drizz commented on 2015-02-20 15:03

When building:

==> Starting prepare()...
-> Fix to save configuration in ~/.config/chromium-dev
-> Add Widevine Support
/tmp/pacaurtmp-/chromium-dev/PKGBUILD: line 322: /usr/lib/chromium-dev/widevine.version: No such file or directory
==> ERROR: A failure occurred in prepare().
Aborting...

sl1pkn07 commented on 2015-02-19 20:15

do'h, jesuchrist

now really fixed ><

sl1pkn07 commented on 2015-02-19 05:27

OW men!! http://upload.wikimedia.org/wikipedia/commons/f/f4/The_Scream.jpg

yes. the pkgbuild have a typo in prepare() (widevide instead of widevine)

fixed

very sorry :S

juston_li commented on 2015-02-19 00:54

After it finishes compiling I get:
install: cannot stat ‘libwidevinecdmadapter.so’: No such file or directory

sl1pkn07 commented on 2015-02-18 21:11

- now support Google Widevine plugin (active by default. to disable set '_use_widevine' to 0)
- now can handle more configuration files in /etc/chromium-dev. useful for plugins. for use their own conf files instead edit the general conf file. see https://bugs.archlinux.org/task/43804
- libpdf now integrated into a binary

greetings

sl1pkn07 commented on 2015-02-10 14:55

32 or 64 bits?

shackra commented on 2015-02-10 07:10

FAILED with 1: ../../../../out/Release/gen/sdk/linux_x86/pnacl_newlib/bin/pnacl-clang -c posix/nacl_timestamp.c -o ../../../../out/Release/obj/native
_client/src/shared/platform/platform_lib.gen/pnacl_newlib-pnacl/platform_lib/nacl_timestamp_362f7b12.o -MD -MF ../../../../out/Release/obj/native_cli
ent/src/shared/platform/platform_lib.gen/pnacl_newlib-pnacl/platform_lib/nacl_timestamp_362f7b12.o.d -O2 -g -Wall -fdiagnostics-show-option -gline-ta
bles-only -Wno-unused-function -Wno-char-subscripts -Wno-c++11-extensions -Wno-unnamed-type-template-args -Wno-extra-semi -Wno-unused-private-field -
Wno-char-subscripts -Wno-unused-function -D__STDC_LIMIT_MACROS=1 -D__STDC_FORMAT_MACROS=1 -D_GNU_SOURCE=1 -D_DEFAULT_SOURCE=1 -D_BSD_SOURCE=1 -D_POSI
X_C_SOURCE=199506 -D_XOPEN_SOURCE=600 -DDYNAMIC_ANNOTATIONS_ENABLED=1 -DDYNAMIC_ANNOTATIONS_PREFIX=NACL_ -DV8_DEPRECATION_WARNINGS -D_FILE_OFFSET_BIT
S=64 -D_DEFAULT_SOURCE=1 -D_BSD_SOURCE=1 -D_POSIX_C_SOURCE=199506 -D_XOPEN_SOURCE=600 -D_GNU_SOURCE=1 -D__STDC_LIMIT_MACROS=1 -DCHROMIUM_BUILD -DCR_C
LANG_REVISION=223108 -DTOOLKIT_VIEWS=1 -DUI_COMPOSITOR_IMAGE_TRANSPORT -DUSE_AURA=1 -DUSE_ASH=1 -DUSE_PANGO=1 -DUSE_CAIRO=1 -DUSE_DEFAULT_RENDER_THEM
E=1 -DUSE_LIBJPEG_TURBO=1 -DUSE_X11=1 -DUSE_CLIPBOARD_AURAX11=1 -DENABLE_ONE_CLICK_SIGNIN -DENABLE_PRE_SYNC_BACKUP -DENABLE_REMOTING=1 -DENABLE_WEBRT
C=1 -DUSE_PROPRIETARY_CODECS -DENABLE_PEPPER_CDMS -DENABLE_CONFIGURATION_POLICY -DENABLE_NOTIFICATIONS -DUSE_UDEV -DDONT_EMBED_BUILD_METADATA -DENABL
E_TASK_MANAGER=1 -DENABLE_EXTENSIONS=1 -DENABLE_PLUGINS=1 -DENABLE_SESSION_SERVICE=1 -DENABLE_THEMES=1 -DENABLE_AUTOFILL_DIALOG=1 -DENABLE_BACKGROUND
=1 -DENABLE_GOOGLE_NOW=1 -DCLD_VERSION=2 -DENABLE_PRINTING=1 -DENABLE_BASIC_PRINTING=1 -DENABLE_PRINT_PREVIEW=1 -DENABLE_SPELLCHECK=1 -DENABLE_CAPTIV
E_PORTAL_DETECTION=1 -DENABLE_APP_LIST=1 -DENABLE_SETTINGS_APP=1 -DENABLE_SUPERVISED_USERS=1 -DENABLE_MDNS=1 -DENABLE_SERVICE_DISCOVERY=1 -DV8_USE_EX
TERNAL_STARTUP_DATA -DUSE_LIBPCI=1 -DUSE_GLIB=1 -DUSE_NSS=1 -DNACL_WINDOWS=0 -DNACL_OSX=0 -DNACL_LINUX=0 -DNACL_ANDROID=0 -DNACL_BUILD_ARCH=pnacl -I.
./../../../out/Release/gen/tc_pnacl_newlib/include -I../../../.. -I../../../../out/Release/gen -I../../../src/third_party -I../../../.. -DNDEBUG -std
=gnu99
Unknown clang plugin argument: check-weak-ptr-factory-order
ninja: build stopped: subcommand failed.

sl1pkn07 commented on 2015-02-06 21:39

no problem ;) thanks for you I found various errors in the pkgbuild ;)

greetings

Lynthium commented on 2015-02-06 20:38

@sl1pkn07
Ye, fully up to date 64bit. The build was with the bundled clang. I couldn't find the "virdual" in the file the error mentioned though. Guess it's just a gremlin poking its head up in my system (or in this insanely complicated build system).

Building it with system clang right now. It's past 14.5k/17.2k mark so it should be fine.

My apologies for wasting your time with this :(

sl1pkn07 commented on 2015-02-06 20:31

build ok with GCC

sl1pkn07 commented on 2015-02-06 19:57

@Lynden build ok with system clang. trying with GCC....

sl1pkn07 commented on 2015-02-06 19:37

@chrisdunder

with your option spawn a lot of

warning: unknown warning option '-Wno-inconsistent-missing-override' [-Wunknown-warning-option]

XD

sl1pkn07 commented on 2015-02-06 19:19

update pkgbuild to fix build bundled ffmpeg with system clang

sl1pkn07 commented on 2015-02-06 18:58

you system is up to date? 32 or 64 bits?

build fine here (with bundled clang). build with system clang in progress..

greetings

Lynthium commented on 2015-02-06 16:27

It fails at 10553 / 17202 with

In file included from ../../third_party/WebKit/Source/core/fetch/ResourcePtr.h:29:
../../third_party/WebKit/Source/core/fetch/Resource.h:191:5: error: unknown type name 'virdual'
virdual void willFollowRedirect(ResourceRequest&, const ResourceResponse&);
^
../../third_party/WebKit/Source/core/fetch/Resource.h:191:13: error: expected member name or ';' after declaration specifiers
virdual void willFollowRedirect(ResourceRequest&, const ResourceResponse&);
~~~~~~~ ^
2 errors generated.

Lynthium commented on 2015-02-06 15:49

@sl1pkn07

Thanks, it builds fine now :)

sl1pkn07 commented on 2015-02-06 15:15

@Lynden
try now

@misc
thanks for the info!

@chrisdunder
added cmake to makedepends if use bundled clang (fixed typo (c&p code xd)
and added -Wno-inconsistent-missing-override if use system clang

python path ('export PATH') still in prepare()

greetings

Lynthium commented on 2015-02-06 14:58

Addition: the previous comment fixes the ffmpeg build, but the chromium build shoots itself in the foot still later on.

-> Starting building Chromium...
Updating projects from gyp files...
ninja: Entering directory `out/Release'
ninja: warning: multiple rules generate obj/third_party/zlib/google/zip.zip.o. builds involving this target will not be correct; continuing anyway
ninja: warning: multiple rules generate obj/third_party/zlib/google/zip.zip_internal.o. builds involving this target will not be correct; continuing anyway
ninja: warning: multiple rules generate obj/third_party/zlib/google/zip.zip_reader.o. builds involving this target will not be correct; continuing anyway
ninja: error: '/data/chromium-dev/src/flapper_version.h', needed by 'gen/flapper_version.h', missing and no known rule to make it

"/data/chromium-dev/src" doesn't exist, it should be "/tmp/makepkg/chromium-dev/src".

Lynthium commented on 2015-02-06 14:50

"$(pwd)/src" in _build_ffmpeg_args (line 84) breaks the ffmpeg build if you have a custom BUILDDIR set in makepkg.conf, since the src is no longer created in the current pwd.

Changing it to "${srcdir}" and moving said statement into prepare() works, but it's prob not the cleanest solution.

misc commented on 2015-02-06 09:41

Just a heads-up, libpdf will be integrated directly into the chromium binary in the next release, leading to space savings of some 7 MiB.

That and they just enabled multithreading for the gold linker, which consistently crashes on my machine. Doubt that's gonna make it into the release, though: https://codereview.chromium.org/894463003

sl1pkn07 commented on 2015-02-06 04:20

@chrisdunder

thanks for the suggestions.

one question: why need move 'export PATH' to to build()?

greetings

chrisdunder commented on 2015-02-05 23:01

Can you move the `export PATH` line from prepare() to build()?

Also cmake is a builddep if use_system_clang==0.

Also can you add -Wno-unknown-warning-option to CXXFLAGS if use_system_clang==1? Otherwise it spams "warning: unknown warning option '-Wno-inconsistent-missing-override'" with the latest (system) clang.

From one to another, Thanks for maintaining this!

chrisdunder commented on 2015-02-05 22:41

cmake is a builddep if use_system_clang=0

Also can you add -Wno-unknown-warning-option to flags? Otherwise it spams "warning: unknown warning option '-Wno-inconsistent-missing-override'" with the latest (system) clang.

TNC commented on 2015-01-28 12:18

Worked like a charm, thank you!

greetings

TNC commented on 2015-01-27 15:25

I'll try that and keep you updated.

greetings

sl1pkn07 commented on 2015-01-27 15:16

Update (no need rebuild) pkgbuild to add support to use system instead of bundled clang

@TNC
try to set "'use_system_clang=' to '1'. or if you want use GCC, set '_use_clang' to '0'

greetings

TNC commented on 2015-01-27 09:55

the newest version failed to build twice:

/backward -internal-isystem /usr/local/include -internal-isystem /usr/bin/../lib/clang/3.5.1/include -internal-externc-isystem /include -internal-externc-isystem /usr/include -O3 -Wall -Wno-unused-parameter -Wno-missing-field-initializers -Wno-reserved-user-defined-literal -Wheader-hygiene -Wno-char-subscripts -Wno-unneeded-internal-declaration -Wno-covered-switch-default -Wstring-conversion -Wno-c++11-narrowing -Wno-deprecated-register -Wno-inconsistent-missing-override -Wglobal-constructors -Wsign-compare -std=gnu++11 -fdeprecated-macro -fdebug-compilation-dir /tmp/yaourt-tmp-mpd/aur-chromium-dev/src/chromium-41.0.2272.12/out/Release -ferror-limit 19 -fmessage-length 0 -fvisibility hidden -fvisibility-inlines-hidden -pthread -stack-protector 2 -stack-protector-buffer-size 4 -stack-protector-buffer-size 4 -mstackrealign -fno-rtti -fno-threadsafe-statics -fobjc-runtime=gcc -fdiagnostics-show-option -fcolor-diagnostics -vectorize-loops -vectorize-slp -load /tmp/yaourt-tmp-mpd/aur-chromium-dev/src/chromium-41.0.2272.12/third_party/llvm-build/Release+Asserts/lib/libFindBadConstructs.so -add-plugin find-bad-constructs -plugin-arg-find-bad-constructs strict-virtual-specifiers -load /tmp/yaourt-tmp-mpd/aur-chromium-dev/src/chromium-41.0.2272.12/third_party/llvm-build/Release+Asserts/lib/libBlinkGCPlugin_14.so -add-plugin blink-gc-plugin -o obj/third_party/WebKit/Source/wtf/wtf.DataLog.o -x c++ ../../third_party/WebKit/Source/wtf/DataLog.cpp
clang: error: unable to execute command: Segmentation fault (core dumped)
clang: error: clang frontend command failed due to signal (use -v to see invocation)
clang version 3.5.1 (tags/RELEASE_351/final)
Target: x86_64-unknown-linux-gnu
Thread model: posix
clang: note: diagnostic msg: PLEASE submit a bug report to http://llvm.org/bugs/ and include the crash backtrace, preprocessed source, and associated run script.
clang: note: diagnostic msg:
********************

PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:
Preprocessed source(s) and associated run script(s) are located at:
clang: note: diagnostic msg: /tmp/DataLog-8e4e9f.cpp
clang: note: diagnostic msg: /tmp/DataLog-8e4e9f.sh
clang: note: diagnostic msg:

********************
warning: unknown warning option '-Wno-inconsistent-missing-override' [-Wunknown-warning-option]
1 warning generated.
warning: unknown warning option '-Wno-inconsistent-missing-override' [-Wunknown-warning-option]
1 warning generated.
ninja: build stopped: subcommand failed.
==> FEHLER: Ein Fehler geschah in build().
Breche ab...

I can provide the 2 mentioned files if needed. I tried to built the package with clang.

sl1pkn07 commented on 2015-01-25 21:11

update for add patch like a [extra] chromium. now fix open pages like https://en.wikipedia.org/wiki/List_of_Unicode_characters

greetings and sorry the inconvenience

sl1pkn07 commented on 2015-01-22 14:48

update to 41.0.2272.12

now uses bundled clang

greetings

scadu commented on 2015-01-17 04:51

Builds fails for me with following error message:

[3581/17051] cd ../../third_party/WebKit/Source/core; python ../build/scripts/make_token_matcher.py ../core/css/parser/BisonCSSParser-in.cpp ../../../../out/Release/gen/blink/core/BisonCSSParser.cpp
FAILED: cd ../../third_party/WebKit/Source/core; python ../build/scripts/rule_bison.py "css/parser/CSSGrammar.y" ../../../../out/Release/gen/blink/core bison
Traceback (most recent call last):
File "../build/scripts/rule_bison.py", line 82, in <module>
returnCode = subprocess.call([bisonExe, '-d', '-p', prefix, inputFile, '-o', outputCpp])
File "/usr/lib/python2.7/subprocess.py", line 522, in call
return Popen(*popenargs, **kwargs).wait()
File "/usr/lib/python2.7/subprocess.py", line 710, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1335, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
Aborting...

Tried twice, same output.

scadu commented on 2015-01-17 04:48

Build fails for me with following error message:
[3581/17051] cd ../../third_party/WebKit/Source/core; python ../build/scripts/make_token_matcher.py ../core/css/parser/BisonCSSParser-in.cpp ../../../../out/Release/gen/blink/core/BisonCSSParser.cpp
FAILED: cd ../../third_party/WebKit/Source/core; python ../build/scripts/rule_bison.py "css/parser/CSSGrammar.y" ../../../../out/Release/gen/blink/core bison
Traceback (most recent call last):
File "../build/scripts/rule_bison.py", line 82, in <module>
returnCode = subprocess.call([bisonExe, '-d', '-p', prefix, inputFile, '-o', outputCpp])
File "/usr/lib/python2.7/subprocess.py", line 522, in call
return Popen(*popenargs, **kwargs).wait()
File "/usr/lib/python2.7/subprocess.py", line 710, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1335, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
Aborting...

scadu commented on 2015-01-17 04:47

Build fails for me with following error message:

atcher.py ../core/css/parser/BisonCSSParser-in.cpp ../../../../out/Release/gen/blink/core/BisonCSSParser.cpp
FAILED: cd ../../third_party/WebKit/Source/core; python ../build/scripts/rule_bison.py "css/parser/CSSGrammar.y" ../../../../out/Release/gen/blink/core bison
Traceback (most recent call last):
File "../build/scripts/rule_bison.py", line 82, in <module>
returnCode = subprocess.call([bisonExe, '-d', '-p', prefix, inputFile, '-o', outputCpp])
File "/usr/lib/python2.7/subprocess.py", line 522, in call
return Popen(*popenargs, **kwargs).wait()
File "/usr/lib/python2.7/subprocess.py", line 710, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1335, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
Aborting...

sl1pkn07 commented on 2014-12-31 17:46

not free space. to build chromium-dev need 2Gb or more

tajidinabd commented on 2014-12-31 17:36

When trying to build this package in the prepare field all kinds of errors
complaining about write failed as you can see. Is there a permissions problem?


chromium-41.0.2251.0/rlz/lib/rlz_lib.h: Write failed
chromium-41.0.2251.0/rlz/lib/lib_values.h: Write failed
chromium-41.0.2251.0/rlz/lib/financial_ping.h: Write failed
chromium-41.0.2251.0/rlz/lib/machine_id_unittest.cc: Write failed
chromium-41.0.2251.0/rlz/lib/rlz_lib.cc: Write failed
chromium-41.0.2251.0/rlz/test/rlz_test_helpers.cc: Write failed
chromium-41.0.2251.0/rlz/test/rlz_unittest_main.cc: Write failed
chromium-41.0.2251.0/rlz/test/rlz_test_helpers.h: Write failed
bsdtar: Error exit delayed from previous errors.
==> FEHLER: Ein Fehler geschah in prepare().


Excuse me for the german basically it is saying it failed in the prepare().
any idea the situation

TNC commented on 2014-12-18 10:21

For some reason the build was fine, but I didn't install or update any packages...

greetings

TNC commented on 2014-12-16 21:53

As I mentioned, I forgot to write the missing files down. I will recompile the package with gcc, so you have to be patient. :-)

Greetings

sl1pkn07 commented on 2014-12-16 10:32

Can you tell me what files is missing?

Greetings

TNC commented on 2014-12-16 09:13

Build did fail with gcc on my amd c60, but with clang it went through.

With gcc the build failed because of some missing python files. I can't tell you which, because I forgot to write it down.

A failed build is really infuriating, because on this slow machine, the build takes over 12h. :-)

greetings

sl1pkn07 commented on 2014-11-21 01:40

update to 41 branch

- switch back to GCC (with clang fail build)

greeetings

sparse commented on 2014-11-20 14:06

Is there any repo for this package?

sl1pkn07 commented on 2014-10-29 12:09

git download over 3Gb in first time. and need a lot of work to "mount" all sources to build same like a tar.xz sources

Nefelim4ag commented on 2014-10-29 11:15

Sorry but, may be you can create a git version of this package?
(Downloading of 250 Mb not good, and i use DEVELSRCDIR="/var/cache/src-csv-cache" in yaourt for cache VCS, and i think, not only i, its useful)

sl1pkn07 commented on 2014-09-29 06:29

@skelet nope, both is from [core] (base & base-devel group)

skelet commented on 2014-09-29 06:03

It seems it needs some additional dependencies:
* patch
* diffutils (for cmp command)

sl1pkn07 commented on 2014-09-26 01:37

please insert this after line 105

pwd $srcdir
ls $srcdir
exit

try now and paste here the output

sl1pkn07 commented on 2014-09-26 01:36

please insert this after 105 line

pwd $srcdir
ls $srcdir
exit

and paste here the output

sl1pkn07 commented on 2014-09-26 01:30

@roukoswarf what version of pacman you use? git?

roukoswarf commented on 2014-09-26 01:28

Package fails prepare on latest versions, path needs ../ removed on decompress line, as it will fail to find the archive.

abandonedaccount commented on 2014-09-21 17:17

You're right. I understand it now:
remove_bundled_libraries.py <preserved-directories>
Thank you.

sl1pkn07 commented on 2014-09-20 19:10

that array in "Remove unnecessary components to save space" is for exceptions (only apply in all thrid_party folders)

sl1pkn07 commented on 2014-09-20 19:07

thats a test I did. nothing happens

fixed

greetings

abandonedaccount commented on 2014-09-20 19:05

chromium-system-jinja-r7.patch seems to be patching only files in third_party/WebKit folder which is later removed via msg2 "Remove unnecessary components to save space"

abandonedaccount commented on 2014-09-20 18:59

in this line:
find . -not \( -path ./git -prune \) -type f -exec sed -i -r \
./git should be ./.git or else sed will dive into it
- nobody was affected by this unless they manually git checkout the repo instead of the archive

sl1pkn07 commented on 2014-09-18 05:22

fixed. sorry

elric_ed commented on 2014-09-17 18:43

Got the following error while running makepkg

==> Validating source files with sha1sums...
chromium-39.0.2159.4.tar.xz ... Passed
chromium-dev.desktop ... Passed
chromium-dev.sh ... Passed
chromium-dev.default ... Passed
chromium-dev.xml ... Passed
chromium-system-jinja-r7.patch ... FAILED
==> ERROR: One or more files did not pass the validity check!

updated sha1sum for chromium-system-jinja-r7.patch on line 54 of PKGBUILD with c24d14029714d2295f3220a7173a5a7362f578a2

It is building now. Will report back if it fails!

juston_li commented on 2014-09-17 18:29

chromium-system-jinja-r7.patch ... FAILED
==> ERROR: One or more files did not pass the validity check!

abandonedaccount commented on 2014-09-07 19:47

that pacman-git bug is tracked here: https://bugs.archlinux.org/task/41862

abandonedaccount commented on 2014-09-07 17:28

I'll try to find out why it isn't softlinked for me(likely pacman-git), I'll let you know what I find.
I've tested so far with untouched makepkg.conf and pacman.conf but the error persists, pwd is: /tmp/yaourt-tmp-emacs/aur-chromium-dev/src and all other files are softlinked.

Btw, thank you for maintaining this package! Much appreciated!

sl1pkn07 commented on 2014-09-07 17:20

maybe is a pacman-git problem

all my test is with pacman 4.1.2-6 and yaourt 1.5-1

sl1pkn07 commented on 2014-09-07 17:20

maybe is a problem by pacman-git

all my test is with pacman 4.1.2-6 and yaourt 1.5-1

sl1pkn07 commented on 2014-09-07 17:15

is softlinked. add that option versions ago. always work

the softlink is added if the files exist in sources array

abandonedaccount commented on 2014-09-07 17:01

I think the reason the tar.gz isn't softlinked is because of this line:
noextract=("chromium-${pkgver}.tar.xz")
haven't tested yet

sl1pkn07 commented on 2014-09-07 16:51

from yaourt (same test)

http://sl1pkn07.wtf/paste/view/9a74f85f

sl1pkn07 commented on 2014-09-07 16:44

makepkg and yaourt softlink the sources from SRCDEST to srcdir

http://sl1pkn07.wtf/paste/view/9dfbc763


tested with edit the pkgbuild and add

pwd
ls -la
exit

in lines 102, 103 and 104 (http://wstaw.org/m/2014/09/07/plasma-desktopil2384.png)

all with untouch /etc/makepkg.conf

abandonedaccount commented on 2014-09-07 16:32

I was afraid that was the case which means it may be my (newer)pacman version:
Pacman v4.1.2-406-g3e19-dirty - libalpm v8.0.2
putting an echo `pwd` before the bsdtar command shows current folder to be:
/tmp/makepkg/chromium-dev/src
while the .tar.gz resides in /tmp/sources/
I've the following /etc/makepkg.conf settings:
BUILDDIR=/tmp/makepkg
PKGDEST=/tmp/packages
SRCDEST=/tmp/sources
SRCPKGDEST=/tmp/srcpackages

testing with SRCDEST unset in /etc/makepkg.conf
leaves the downloaded tar.xz file in /tmp/yaourt-tmp-emacs/aur-chromium-dev
and still fails.
Rebuilding pacman yields the same version. makepkg (pacman) 4.1.2-406-g3e19-dirty (August 7 commit was last)
was using yaourt 1.5 but now(next) using yaourt-git: yaourt 1.4-1-gb6b8
nothing changed.
bsdtar: Error opening archive: Failed to open 'chromium-39.0.2145.4.tar.xz'
pwd is: /tmp/makepkg/chromium-dev/src
archive is actually in /tmp/sources/
or when SRCDEST above isn't set(and running yaourt in a new terminal) then archive is in: /tmp/yaourt-tmp-emacs/aur-chromium-dev/ where PKGBUILD &co. are also downloaded

sl1pkn07 commented on 2014-09-07 15:14

working for me (makepkg(local) and yaourt(online))

abandonedaccount commented on 2014-09-07 14:49

had to replace this:
bsdtar -xf "chromium-${pkgver}.tar.xz"
with this:
bsdtar -xf "${SRCDEST}/chromium-${pkgver}.tar.xz"
or else it would fail to find that file and fail.

sl1pkn07 commented on 2014-08-24 22:24

im not sure. but libwidevine is include directly in chrome binary

└───╼ strings chrome |grep widevine
libwidevinecdmadapter.so
application/x-ppapi-widevine-cdm

jeagoss commented on 2014-08-24 21:25

No. I would like to copy the libwidevinecdm*.so components from the binary build to the opensource build and have them register properly. Kind of like how the flash plugin is copied currently. I've been doing minimal searching and not turning up much on how to enable recognition for the components.

sl1pkn07 commented on 2014-08-24 21:23

@jeagos you mean this? http://www.chromium.org/administrators/pre-installed-extensions (docs.crx, drive.crx youtube.crx, etc..)

sl1pkn07 commented on 2014-08-23 21:02

Nope. Sorry. But i can have a look when back from my little flash holiday

Regards

jeagoss commented on 2014-08-23 20:29

Would you happen to know if there is a way to register the widevine components from the binary build with the open source build?

sl1pkn07 commented on 2014-08-22 03:24

38.0.2125.8 release today.... bah!!

sl1pkn07 commented on 2014-08-22 03:04

If this package build/compact in one minute i don't make this effusive apology. but this take an a hour and more.. I can not allow such a silly error like this. I do not want waste the time to other users.

greetings :)

jeagoss commented on 2014-08-22 02:08

Heh. Everyone forgets stuff sometimes. No need to apologize. Thanks for fixing it :)

sl1pkn07 commented on 2014-08-21 19:51

fixed in pkgbuild

i'm so sorry :S

NO NEED REBUILD ALL. only just copy the file with the right name and before next update remove that file

sorry the inconvenience

sl1pkn07 commented on 2014-08-21 19:43

fixed in pkgbuild

i'm so sorry :S

sl1pkn07 commented on 2014-08-21 19:41

BIG DO'H! my faut!

please rename

/usr/lib/chromium-dev/chrome_100_percent

to

/usr/lib/chromium-dev/chrome_100_percent.pak


BIG SORRY :S

jeagoss commented on 2014-08-21 19:08

This last build has resulted in an incredibly unusable package. When loading chromium, Almost all UI elements are Red. Plugins are very crashy.

I've done a bit of searching and nothing seems to be being reported. Is this a known issue? Has anyone else run into it?

confusedfla commented on 2014-06-28 14:09

FAILED: Command '['../../out/Release/gen/sdk/linux_x86/pnacl_newlib/bin/pnacl-clang++', '-o', '../../out/Release/nacl_irt_x86_64.nexe.debug.raw', '-Wl,--as-needed', '--target=x86_64-unknown-nacl', '-stdlib=libstdc++', '-arch', 'x86-64', '--pnacl-allow-translate', '--pnacl-allow-native', '-B../../out/Release/gen/tc_irt/lib64', '-Wt,-ffunction-sections', '-Wt,-fdata-sections', '-Wn,--gc-sections', '-Wl,--start-group', '-lirt_browser', '-llatency_info_nacl', '-lpnacl_irt_shim_for_irt', '-lppapi_proxy_nacl', '-lppapi_ipc_nacl', '-lppapi_shared_nacl', '-lgles2_implementation_nacl', '-lgles2_cmd_helper_nacl', '-lgles2_utils_nacl', '-lcommand_buffer_client_nacl', '-lcommand_buffer_common_nacl', '-ltracing_nacl', '-lgpu_ipc_nacl', '-lipc_nacl', '-lbase_nacl', '-lshared_memory_support_nacl', '-lsrpc', '-limc_syscalls', '-lplatform', '-lgio', '-Wl,--end-group', '-lm', '-Wl,-Ttext-segment=0xfce0000', '-Wl,-Trodata-segment=0x3ef40000']' returned non-zero exit status 247
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().

anyone else experiencing this build error?

sl1pkn07 commented on 2014-06-19 13:30

@matchulinux

try now (disable ffmpeg patch)

matchulinux commented on 2014-06-19 12:45

It seems like the sha1sum for the file chromium-ffmpeg-r1.patch is wrong.
==> Validating source files with sha1sums...
[...]
chromium-ffmpeg-r1.patch ... FAILED
[...]
==> ERROR: One or more files did not pass the validity check!

PS: Thanks for maintaining this package :)

sl1pkn07 commented on 2014-06-06 09:47

@lordmocha

that error is problem from google side.

and for re2 related... i think is because makepkg (makepkg -s) cannot handle packages from AUR directly

greetings

lordmocha commented on 2014-06-06 04:03

IOError: Could not download URL (https://storage.googleapis.com/nativeclient-once/object/llvm_i686_linux_3c0f38277820277b0ede45b2b22fb54fcd9ded5c.tgz): The read operation timed out
==> ERROR: A failure occurred in prepare().
Aborting...

and i also had to install re2 from aur first, it didn't autoget it as a dependency

sl1pkn07 commented on 2014-06-05 22:27

try add

-Drelease_extra_cflags='march=native'

into _flags= array

greetings

sudsbud commented on 2014-06-05 21:57

I was wondering how I might specify some more compile options - for instance I want to specify march=native. I tried exporting C and CXX FLAGS and editing the PKGBUILD along with editing my makepkg.conf but it didn't seem to work. How should I specify flags for this build?

sl1pkn07 commented on 2014-06-05 10:09

"warning" is a internal command from makepkg, like "msg" and "msg2"

nothing to do

sl1pkn07 commented on 2014-06-05 10:09

"warning" is a internal command from makepkg, like "msg" and "msg2"

nothig to do

sl1pkn07 commented on 2014-06-05 10:05

warning is a internal command from makepkg

pfrenssen commented on 2014-06-05 08:26

I'm getting the following error:

/tmp/pacaurtmp-pieter/chromium-dev/PKGBUILD: line 28: warning: command not found

It tries to execute the command called "warning" which is not present on my system. You can either list the package which supplies "warning" as a dependency, or just use echo instead of warning.

sl1pkn07 commented on 2014-05-30 21:27

my bad. I upload new pkgbuild with disable unittest

please, you can safe rebuild with "makepkg -fR" (only compress package, no rebuild everything)

greetings and sorry :S

DarkMind commented on 2014-05-30 19:20

[1444/1444] ThreadCollisionTest.MTSynchedScopedRecursiveBookCriticalSectionTest (4000 ms)
Retrying 4 tests (retry #1)
[1445/1448] TraceEventTestFixture.TraceContinuousSampling (0 ms)
[ RUN ] FileUtilTest.ChangeFilePermissionsAndWrite
../../base/file_util_unittest.cc:772: Failure
Value of: WriteFile(file_name, kData.data(), kData.length())
Actual: 5
Expected: -1
../../base/file_util_unittest.cc:773: Failure
Value of: PathIsWritable(file_name)
Actual: true
Expected: false
[ FAILED ] FileUtilTest.ChangeFilePermissionsAndWrite (0 ms)
[1446/1448] FileUtilTest.ChangeFilePermissionsAndWrite (0 ms)
[ RUN ] FileUtilTest.ChangeFilePermissionsAndRead
../../base/file_util_unittest.cc:732: Failure
Value of: ReadFile(file_name, buffer, buffer_size)
Actual: 5
Expected: -1
[ FAILED ] FileUtilTest.ChangeFilePermissionsAndRead (0 ms)
[1447/1448] FileUtilTest.ChangeFilePermissionsAndRead (0 ms)
[ RUN ] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate
../../base/file_util_unittest.cc:819: Failure
Value of: c1.size()
Actual: 1
Expected: 0
../../base/file_util_unittest.cc:820: Failure
Value of: GetPosixFilePermissions(file_name, &mode)
Actual: true
Expected: false
[ FAILED ] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate (0 ms)
[1448/1448] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate (0 ms)
Retrying 3 tests (retry #2)
[ RUN ] FileUtilTest.ChangeFilePermissionsAndWrite
../../base/file_util_unittest.cc:772: Failure
Value of: WriteFile(file_name, kData.data(), kData.length())
Actual: 5
Expected: -1
../../base/file_util_unittest.cc:773: Failure
Value of: PathIsWritable(file_name)
Actual: true
Expected: false
[ FAILED ] FileUtilTest.ChangeFilePermissionsAndWrite (0 ms)
[1449/1451] FileUtilTest.ChangeFilePermissionsAndWrite (0 ms)
[ RUN ] FileUtilTest.ChangeFilePermissionsAndRead
../../base/file_util_unittest.cc:732: Failure
Value of: ReadFile(file_name, buffer, buffer_size)
Actual: 5
Expected: -1
[ FAILED ] FileUtilTest.ChangeFilePermissionsAndRead (1 ms)
[1450/1451] FileUtilTest.ChangeFilePermissionsAndRead (0 ms)
[ RUN ] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate
../../base/file_util_unittest.cc:819: Failure
Value of: c1.size()
Actual: 1
Expected: 0
../../base/file_util_unittest.cc:820: Failure
Value of: GetPosixFilePermissions(file_name, &mode)
Actual: true
Expected: false
[ FAILED ] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate (0 ms)
[1451/1451] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate (0 ms)
Retrying 3 tests (retry #3)
[ RUN ] FileUtilTest.ChangeFilePermissionsAndWrite
../../base/file_util_unittest.cc:772: Failure
Value of: WriteFile(file_name, kData.data(), kData.length())
Actual: 5
Expected: -1
../../base/file_util_unittest.cc:773: Failure
Value of: PathIsWritable(file_name)
Actual: true
Expected: false
[ FAILED ] FileUtilTest.ChangeFilePermissionsAndWrite (1 ms)
[1452/1454] FileUtilTest.ChangeFilePermissionsAndWrite (0 ms)
[ RUN ] FileUtilTest.ChangeFilePermissionsAndRead
../../base/file_util_unittest.cc:732: Failure
Value of: ReadFile(file_name, buffer, buffer_size)
Actual: 5
Expected: -1
[ FAILED ] FileUtilTest.ChangeFilePermissionsAndRead (0 ms)
[1453/1454] FileUtilTest.ChangeFilePermissionsAndRead (0 ms)
[ RUN ] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate
../../base/file_util_unittest.cc:819: Failure
Value of: c1.size()
Actual: 1
Expected: 0
../../base/file_util_unittest.cc:820: Failure
Value of: GetPosixFilePermissions(file_name, &mode)
Actual: true
Expected: false
[ FAILED ] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate (0 ms)
[1454/1454] FileUtilTest.ChangeDirectoryPermissionsAndEnumerate (0 ms)
3 tests failed:
FileUtilTest.ChangeDirectoryPermissionsAndEnumerate
FileUtilTest.ChangeFilePermissionsAndRead
FileUtilTest.ChangeFilePermissionsAndWrite
Tests took 23 seconds.
==> ERROR: Se produjo un error en check().
Cancelando...

sl1pkn07 commented on 2014-05-23 14:10

update to 37.0.2008.2

- new 37 series
- include own libpdf builded from sources (see "pdfium" in http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/trunk/src&range=269467:272007&mode=html). no need chromium-libpdf(-dev) anymore

enjoy!

sl1pkn07 commented on 2014-05-23 14:03

update to 37.0.2008.2

- new 37 series
- include own libpdf builded from sources (see "pdfium" in http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/trunk/src&range=269467:272007&mode=html)

enjoy!

sl1pkn07 commented on 2014-05-22 12:03

@dz0ny: what happens??

sl1pkn07 commented on 2014-05-21 21:21

OW!

fixed ><

thanks

trizen commented on 2014-05-21 21:05

In chromium-dev.sh, at line 6, should be '/etc/chromium-dev/default', instead of /etc/chromium/default', otherwise chromium-dev will not start.

sl1pkn07 commented on 2014-05-21 12:18

@Vi0L0

i'll test 8K (with you images in crbug) images with this new version. the scroll and resize back to smooth

sl1pkn07 commented on 2014-05-21 12:13

update to 36.0.1985.18

- Back to use clang (more general stablity, specially with nvidia drivers 337.19))
- Remove PepperFlash and libpdf: if want use this plugins install chromium-pepper-flash(-dev) and chromium-libpdf(-dev)
- adjust dependency list
- Back to use system ICU

greetings

sl1pkn07 commented on 2014-05-14 21:23

update to 36.0.1985.2

- use bundled jsoncpp
- back to use GCC(4.9) instead of Clang (if set "_use_clang" to "1", then use arch clang instead of rebuild bundled clang (this use "-Dclang_use_chrome_plugins=0" to work)
- set different exclude unittest for clang/GCC (ones work with clang, others not work with GCC)

greetings

Vi0L0 commented on 2014-05-12 20:39

Sorry for delay. Bug report is here [1], feel free to add your two cents ;)

[1] https://code.google.com/p/chromium/issues/detail?id=372538

sl1pkn07 commented on 2014-05-09 07:51

@vi0l0, same behavior with Nvidia (beta) here :S

Vi0L0 commented on 2014-05-08 17:27

Guys, maybe you noticed performance problem when resizing big image to its actual size? Take this 8k image for example:
http://www.glif.is/publications/maps/GLIF_5-11_World_8k.jpg
Resize it, and try to move around it.
I'm catching almost a freeze, but only on this one images tab.
I'm thinking about posting bug report, but decided to ask here at first.
BTW: compiled while fglrx(catalyst) was installed, it's lagging on both fglrx and oss ati(-git).

sl1pkn07 commented on 2014-05-05 19:44

reger. then update now

trizen commented on 2014-05-05 17:58

After I've added the "--force-local-build --without-android" arguments to "tools/clang/scripts/update.sh", I couldn't compile clang with clang-3.4 (doh), so I uninstalled clang and tried again so it builds using gcc-3.9.0. At this moment, clang is still compiling...

Update: the process completed successfully -- chromium is working now. So the solution on an x86 arch, is to compile clang locally with gcc, then compile chromium with the locally built clang.

trizen commented on 2014-05-05 12:07

After added the "--force-local-build --without-android" arguments to "tools/clang/scripts/update.sh", I couldn't compile clang with clang-3.4 (doh), so I uninstalled clang and tried again so it builds using gcc-3.9.0. At this moment, clang is still compiling...

sl1pkn07 commented on 2014-05-05 12:05

@trizen: nope, need disable clang plugins (-Dclang_use_chrome_plugins=0")

but fail build http://sl1pkn07.no-ip.com/paste/view/ef1c1eca

related: https://bbs.archlinux.org/viewtopic.php?id=181094

:S

about --force-local-build. i think is not a solution. but i'll make test

greetings

sl1pkn07 commented on 2014-05-05 12:03

@trizen: nope, need disable clang plugins (-Dclang_use_chrome_plugins=0")

but fail build http://sl1pkn07.no-ip.com/paste/view/ef1c1eca

related: https://bbs.archlinux.org/viewtopic.php?id=181094

:S

trizen commented on 2014-05-05 10:25

No, I don't use any of that. By looking at "src/build/tools/clang/scripts/update.sh", you can see at line 207 that it actually downloads a x64 version of clang:

if [ "${OS}" = "Linux" ]; then
CDS_FULL_URL="${CDS_URL}/Linux_x64/${CDS_FILE}"

Where ${CDS_URL} is: https://commondatastorage.googleapis.com/chromium-browser-clang

I also tried to link /usr/bin/clang, /usr/bin/clang++ and /usr/bin/llvm-symbolizer into src/build/third_party/llvm-build/Release+Asserts/bin, but got another error:

error: unable to load plugin '/home/swampyx/Others/Build/chromium-dev/src/build/tools/clang/scripts/../../../third_party/llvm-build/Release+Asserts/lib/libFindBadConstructs.so': '/home/swampyx/Others/Build/chromium-dev/src/build/tools/clang/scripts/../../../third_party/llvm-build/Release+Asserts/lib/libFindBadConstructs.so: cannot open shared object file: No such file or directory'
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
Aborting...

So, it needs clang-3.5, I guess. I'm thinking about passing the "--force-local-build" argument to "tools/clang/scripts/update.sh" to see if this works...

sl1pkn07 commented on 2014-05-05 10:12

no, is for both. you use selinux/apparmor/something?

trizen commented on 2014-05-05 09:47

Thanks. It works now, but I made a mistake compiling it with gcc-4.9.0. Trying again now with clang.

Update: It doesn't work for me (maybe clang is for x64?):
/bin/sh: ../../third_party/llvm-build/Release+Asserts/bin/clang: cannot execute binary file: Exec format error
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
Aborting...

trizen commented on 2014-05-05 09:40

Thanks. It works now, but I made a mistake compiling it with gcc-4.9.0. Trying again now with clang.

sl1pkn07 commented on 2014-05-04 22:08

@tizen

try now

greetings and sorry

trizen commented on 2014-05-04 20:38

On x86, "https://dl.google.com/linux/direct/google-chrome-unstable_current_i386.rpm" returns 404 - Not found.

Using line 83 seems to fix the problem.

sl1pkn07 commented on 2014-05-04 16:36

updated to 36.0.1964.2

- now build with clang (provide by chromium) (Turn back to GCC when fix run with GCC 4.9)
- now NaCl/PNaCl toolchain download by chromium internal tool
- pepperflash/libpdf provide from official common download site instead Fedora repository (still use RPM sources)

greetings and sorry the delay

sl1pkn07 commented on 2014-05-03 22:25

i've tried build with clang, the error are gone (36.0.1964.2)

pnacl/nacl build ok for me

in the new pkgbuild i've made have several changes respect this (use internal tool provide by chromium sources instead download manually) and add one flag:

-Dlinux_use_bundled_binutils=0

about .rpm (flash/pdf plugin), 36.0.1964.2 x86_64 not found (i386 ok) in google official fedora repository. this time use the google-chrome-unstable_current_${_rpm_arch}.rpm from official download site

when finish test if all work ok, i'll upload the update

if any want test the update: https://dl.dropboxusercontent.com/u/6596386/chromium-dev-36.0.1964.2-1.src.tar.gz

greetings

sl1pkn07 commented on 2014-05-03 22:19

i've tried build with clang, the error are gone (36.0.1964.2)

pnacl/nacl build ok for me

in the new pkgbuild i've made have several changes respect this (use internal tool provide by chromium sources instead download manually) and add one flag:

-Dlinux_use_bundled_binutils=0

about .rpm (flash/pdf plugin), 36.0.1964.2 x86_64 not found (i386 ok) in google official fedora repository. this time use the chromium-browser-unstable_current_${_rpm_arch}.rpm from official download site

when finish test if all work ok, i'll upload the update

if any want test the update: https://dl.dropboxusercontent.com/u/6596386/chromium-dev-36.0.1964.2-1.src.tar.gz

greetings

misc commented on 2014-05-03 20:38

There's 36.0.1964.5 now, maybe try that?

misc commented on 2014-05-03 20:32

It's gone with 1974.0 (r267764, actually), so it's either somebody doing a bisect to find the fix, or waiting for the next dev release.

Also, I can no longer get [p]nacl to build and had to disable it (-Ddisable_nacl=1 and remove the pnacl line + adjust package(), obviously). Luckily Flash nowadays works without it, I wasn't aware of that.

sl1pkn07 commented on 2014-05-03 20:14

have problems with GCC 4.9. build OK. but...

http://wstaw.org/m/2014/05/03/plasma-desktopF11541.png

all plugins/extensions stop working and all pages (include internal) gets like screenshot

console log: http://sl1pkn07.no-ip.com/paste/view/64746be0
gdb: http://sl1pkn07.no-ip.com/paste/view/914bee2a
dmesg: tons of http://sl1pkn07.no-ip.com/paste/view/83a83e4b

any ideas?

sl1pkn07 commented on 2014-05-03 20:13

have problems with GCC 4.9. build OK. but...

http://wstaw.org/m/2014/05/03/plasma-desktopF11541.png

all plugins/extensions stop working.

console log: http://sl1pkn07.no-ip.com/paste/view/64746be0
gdb: http://sl1pkn07.no-ip.com/paste/view/914bee2a
dmesg: tons of http://sl1pkn07.no-ip.com/paste/view/83a83e4b

any ideas?

misc commented on 2014-05-03 09:12

If that's still the case, use the deb files instead. For x64:
https://dl.google.com/linux/chrome/deb/pool/main/g/google-chrome-unstable/google-chrome-unstable_36.0.1964.2-1_amd64.deb

Needs to be extracted, and the two bsdtar lines adjusted to
bsdtar -xf data.tar.lzma opt/google/chrome-unstable/[…]

fluxboxer commented on 2014-04-25 13:52

curl: (22) The requested URL returned error: 404 Not Found

fluxboxer commented on 2014-04-25 13:37

All right, will try

tmpfs /tmp tmpfs nodev,nosuid,size=3G 0 0

fluxboxer commented on 2014-04-25 13:25

I guess so. Sorry for botherin you but should I chage my fstab entry:

tmpfs /tmp tmpfs nodev,nosuid

to anything else? (Have SSD)

sl1pkn07 commented on 2014-04-25 13:08

@qanjaro are you sure have a sufficient free space?, uncompress tar need +/- 2Gb

fluxboxer commented on 2014-04-25 10:23

Extracting chromium-36.0.1941.0.tar.xz with bsdtar
chromium-36.0.1941.0/tools/deps2git/PRESUBMIT.py: Write failed
chromium-36.0.1941.0/tools/deps2git/deps2git.py: Write failed
chromium-36.0.1941.0/tools/deps2git/git_tools.py: Write failed
chromium-36.0.1941.0/tools/deps2git/svn_to_git_public.py: Write failed
chromium-36.0.1941.0/tools/deps2git/WATCHLISTS: Write failed
chromium-36.0.1941.0/tools/deps2git/buildspec_to_git.py: Write failed
chromium-36.0.1941.0/tools/symsrc/README.chromium: Write failed
chromium-36.0.1941.0/tools/symsrc/COPYING-pefile: Write failed
chromium-36.0.1941.0/tools/symsrc/img_fingerprint.py: Write failed
chromium-36.0.1941.0/tools/symsrc/pefile.py: Write failed
chromium-36.0.1941.0/tools/symsrc/pdb_fingerprint_from_img.py: Write failed
chromium-36.0.1941.0/tools/symsrc/source_index.py: Write failed
chromium-36.0.1941.0/tools/set_default_handler/DEPS: Write failed
chromium-36.0.1941.0/tools/set_default_handler/set_default_handler.gyp: Write failed
chromium-36.0.1941.0/tools/set_default_handler/set_default_handler_main.cc: Write failed
chromium-36.0.1941.0/tools/linux/dump-static-initializers.py: Write failed
chromium-36.0.1941.0/tools/linux/procfs.py: Write failed
chromium-36.0.1941.0/tools/linux/PRESUBMIT.py: Write failed
chromium-36.0.1941.0/tools/linux/tests/procfs_tests.py: Write failed
chromium-36.0.1941.0/tools/memory_watcher/DEPS: Write failed
chromium-36.0.1941.0/tools/memory_watcher/mini_disassembler.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/call_stack.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/memory_watcher.gyp: Write failed
chromium-36.0.1941.0/tools/memory_watcher/hotkey.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/scripts/memtrace.pl: Write failed
chromium-36.0.1941.0/tools/memory_watcher/scripts/memprof.pl: Write failed
chromium-36.0.1941.0/tools/memory_watcher/scripts/summary.pl: Write failed
chromium-36.0.1941.0/tools/memory_watcher/scripts/finditem.pl: Write failed
chromium-36.0.1941.0/tools/memory_watcher/memory_hook.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/preamble_patcher.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/README: Write failed
chromium-36.0.1941.0/tools/memory_watcher/mini_disassembler_types.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/mini_disassembler.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/ia32_modrm_map.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/memory_watcher.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/preamble_patcher_with_stub.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/memory_watcher.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/ia32_opcode_map.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/dllmain.cc: Write failed
chromium-36.0.1941.0/tools/memory_watcher/preamble_patcher.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/call_stack.h: Write failed
chromium-36.0.1941.0/tools/memory_watcher/memory_hook.cc: Write failed
chromium-36.0.1941.0/rlz/DEPS: Write failed
chromium-36.0.1941.0/rlz/rlz.gyp: Write failed
chromium-36.0.1941.0/rlz/OWNERS: Write failed
chromium-36.0.1941.0/rlz/mac/lib/machine_id_mac.cc: Write failed
chromium-36.0.1941.0/rlz/mac/lib/rlz_value_store_mac.h: Write failed
chromium-36.0.1941.0/rlz/mac/lib/rlz_value_store_mac.mm: Write failed
chromium-36.0.1941.0/rlz/chromeos/lib/machine_id_chromeos.cc: Write failed
chromium-36.0.1941.0/rlz/chromeos/lib/rlz_value_store_chromeos.cc: Write failed
chromium-36.0.1941.0/rlz/chromeos/lib/rlz_value_store_chromeos.h: Write failed
chromium-36.0.1941.0/rlz/examples/rlz_id.cc: Write failed
chromium-36.0.1941.0/rlz/ios/lib/machine_id_ios.cc: Write failed
chromium-36.0.1941.0/rlz/win/dll/dll_main.cc: Write failed
chromium-36.0.1941.0/rlz/win/dll/exports.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/lib_mutex.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/rlz_value_store_registry.h: Write failed
chromium-36.0.1941.0/rlz/win/lib/machine_deal.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/rlz_value_store_registry.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/process_info.h: Write failed
chromium-36.0.1941.0/rlz/win/lib/machine_id_win.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/lib_mutex.h: Write failed
chromium-36.0.1941.0/rlz/win/lib/registry_util.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/machine_deal_test.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/registry_util.h: Write failed
chromium-36.0.1941.0/rlz/win/lib/process_info.cc: Write failed
chromium-36.0.1941.0/rlz/win/lib/machine_deal.h: Write failed
chromium-36.0.1941.0/rlz/win/lib/rlz_lib_win.cc: Write failed
chromium-36.0.1941.0/rlz/lib/string_utils.h: Write failed
chromium-36.0.1941.0/rlz/lib/assert.cc: Write failed
chromium-36.0.1941.0/rlz/lib/crc32.h: Write failed
chromium-36.0.1941.0/rlz/lib/string_utils_unittest.cc: Write failed
chromium-36.0.1941.0/rlz/lib/machine_id.cc: Write failed
chromium-36.0.1941.0/rlz/lib/financial_ping.cc: Write failed
chromium-36.0.1941.0/rlz/lib/lib_values_unittest.cc: Write failed
chromium-36.0.1941.0/rlz/lib/machine_id.h: Write failed
chromium-36.0.1941.0/rlz/lib/assert.h: Write failed
chromium-36.0.1941.0/rlz/lib/lib_values.cc: Write failed
chromium-36.0.1941.0/rlz/lib/rlz_enums.h: Write failed
chromium-36.0.1941.0/rlz/lib/crc32_wrapper.cc: Write failed
chromium-36.0.1941.0/rlz/lib/crc8_unittest.cc: Write failed
chromium-36.0.1941.0/rlz/lib/rlz_lib_clear.cc: Write failed
chromium-36.0.1941.0/rlz/lib/financial_ping_test.cc: Write failed
chromium-36.0.1941.0/rlz/lib/rlz_lib_test.cc: Write failed
chromium-36.0.1941.0/rlz/lib/crc32_unittest.cc: Write failed
chromium-36.0.1941.0/rlz/lib/string_utils.cc: Write failed
chromium-36.0.1941.0/rlz/lib/recursive_cross_process_lock_posix.h: Write failed
chromium-36.0.1941.0/rlz/lib/rlz_value_store.h: Write failed
chromium-36.0.1941.0/rlz/lib/crc8.h: Write failed
chromium-36.0.1941.0/rlz/lib/crc8.cc: Write failed
chromium-36.0.1941.0/rlz/lib/recursive_cross_process_lock_posix.cc: Write failed
chromium-36.0.1941.0/rlz/lib/rlz_lib.h: Write failed
chromium-36.0.1941.0/rlz/lib/lib_values.h: Write failed
chromium-36.0.1941.0/rlz/lib/financial_ping.h: Write failed
chromium-36.0.1941.0/rlz/lib/machine_id_unittest.cc: Write failed
chromium-36.0.1941.0/rlz/lib/rlz_lib.cc: Write failed
chromium-36.0.1941.0/rlz/test/rlz_test_helpers.cc: Write failed
chromium-36.0.1941.0/rlz/test/rlz_unittest_main.cc: Write failed
chromium-36.0.1941.0/rlz/test/rlz_test_helpers.h: Write failed
bsdtar: Error exit delayed from previous errors.

misc commented on 2014-04-22 19:11

In package()'s "Extract and install PepperFlash" section, makepkg cd's into $srcdir (where pepper-flash resides), yet wants to copy that file from the parent directory (ie, ../), and thus fails.

crondog commented on 2014-04-18 11:29

@sl1pkn07 Just tried. That only changes the font on webpages etc

sl1pkn07 commented on 2014-04-18 10:52

@trizen. Thanks. added

@crondog. try with https://chrome.google.com/webstore/detail/advanced-font-settings/caclkomlalccbpcdllchkeecicepbmbm

crondog commented on 2014-04-18 03:24

So i finally got around to updating from 34->36. Anyone know how to reduce the size of the omnibox font? It's twice the size of the bookmark fonts and setting the font size with lxappearance does not change the omnibox font size

trizen commented on 2014-04-17 23:40

Minor fix: 'pnacl_public_x86_64_libpnacl_irt_shim_a' should be 'pnacl_public_x86_32_libpnacl_irt_shim_a' on a 32-bit machine.

PKGBUILD: http://pastebin.com/XDUfHtku

sl1pkn07 commented on 2014-04-17 11:12

try now

walkindude commented on 2014-04-17 10:45

-> 1337's g0 7h3 p4r7y!
Updating projects from gyp files...
Running build/landmines.py...
ninja: Entering directory `out/Release'
ninja: warning: multiple rules generate obj/third_party/zlib/google/zip.zip.o. builds involving this target will not be correct; continuing anyway
ninja: warning: multiple rules generate obj/third_party/zlib/google/zip.zip_internal.o. builds involving this target will not be correct; continuing anyway
ninja: warning: multiple rules generate obj/third_party/zlib/google/zip.zip_reader.o. builds involving this target will not be correct; continuing anyway
ninja: error: 'gen/sdk/toolchain/linux_x86_newlib/stamp.untar', needed by 'gen/sdk/toolchain/linux_x86_newlib/stamp.prep', missing and no known rule to make it
==> ERROR: A failure occurred in build().
Aborting...
==> ERROR: Makepkg was unable to build chromium-dev.

sl1pkn07 commented on 2014-04-16 07:20

NOTICE

in the 36.0.1941.0, GTK port has been complety remove in sources

from now, AURA is now set by default

see version log:

http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/trunk/src&range=263744:262849&mode=html

sl1pkn07 commented on 2014-04-16 07:18

NOTICE

in the 36.0.1941.0, GTK port has been complety remove in sources

from now, AURA is now set by default

ledti commented on 2014-04-06 09:33

I managed to get around the error by building with clang.

ledti commented on 2014-04-06 00:57

For the last two updates, trying to build an aura enabled build results in the following error near the very end:

libremoting client plugin.an: error adding symbols: Archive has no index; run ranlib to add one

Any tips?

sl1pkn07 commented on 2014-04-04 00:57

https://code.google.com/p/chromium/issues/detail?id=359177

:(

WhoNeedszzz commented on 2014-03-28 14:54

Error occurs at this point: " -> L3ts g0 th3 p4rTy!"

WhoNeedszzz commented on 2014-03-28 14:52

Since the build is out of date I tried updating it myself, but I'm running into this error:
Exception: Failed to open ../third_party/dom_distiller_js/js/domdistiller.js while trying to flatten ../third_party/dom_distiller_js/js/domdistiller.js. (No such file or directory)

Searching google for this returns only one useless result.

ledti commented on 2014-03-19 09:01

Here's a patch for the PKGBUILD to build a more recent version: http://ix.io/bad

+ updated to latest development version.
+ updated to latest nacl toolchain.
+ changed to use internal zlib (-Duse_system_zlib=0), otherwise you'll get 'log_serializer.cc' related errors at the end of compilation.
- removed cups-r0 patch, it's already applied.
- removed gn-r2.patch, it's no longer needed as far as I can tell. The patched/deleted lines are now commented by default.
- removed system-jinja-r5.patch, jinja related errors.
- removed 'remove_bundled_libraries.py' related lines, errors related to jinja, ply, and markupsafe. Too lazy to figure out which specific libraries are required/safe to remove.

sl1pkn07 commented on 2014-02-14 19:22

I've tried, but have a lots of problems with for example D&D from the chromium to the dolphin (the desktop get unresponsive/slow when switch between aplications (alt+tab))

jeagoss commented on 2014-02-14 19:06

I've been using Aura with a clang build for a while now. (The default on my repo.) I haven't run into any issues recently. Even with themes.

Nordlicht commented on 2014-02-14 15:12

Here is an interesting discussion about the "Aura on Linux" Topic:

https://code.google.com/p/chromium/issues/detail?id=317859#c71

sl1pkn07 commented on 2014-02-14 08:09

DAMN!!!!!!

https://wiki.archlinux.org/index.php/Creating_Packages#The_check.28.29_function

option=('!check') wrong!!! (fixed now in wiki)

please, if use yaourt to update package:

DON'T CANCEL THE INSTALLATION IF FAILS!!!!!

just go to yaourt temporal directory (/tmp/yaourt-tmp->user<) (our your tmp path defined in yaourt.rc file) and type:

makepkg -fR

to avoid the rebuild everiting and only start repackage

if you build without AUR helper only need run the same command

i've upload new PKGBUILD por solve this


sorry and grettings :S

sl1pkn07 commented on 2014-02-14 08:08

DAMN!!!!!!

https://wiki.archlinux.org/index.php/Creating_Packages#The_check.28.29_function

option=('!check') wrong!!! (fixed now in wiki)

please, if use yaourt to update package:

DON'T CANCEL THE INSTALLATION IF FAILS!!!!!

just go to yaourt temporal directory (/tmp/yaourt-tmp->user<) (our your tmp path defined in yaourt.rc file) and type:

makepkg -fR

to avoid the rebuild everiting and only start repackage

if you build without AUR helper only need the same comand

i've upload new PKGBUILD por solve this


sorry and grettings :S

sl1pkn07 commented on 2014-02-14 08:03

fixed wiki

sl1pkn07 commented on 2014-02-14 07:44

DAMN!!!!!!

https://wiki.archlinux.org/index.php/Creating_Packages#The_check.28.29_function is broken!! option=('!check') don't exist!!!

please, if use yaourt to update package:

DON'T CANCEL THE INSTALLATION IF FAILS!!!!!

just go to yaourt temporal directory (/tmp/yaourt-tmp->user<) (our your tmp path defined in yaourt.rc file) and type:

makepkg -fR

to avoid the rebuild everiting and only start repackage

if you build without AUR helper only need the same comand

i've upload new PKGBUILD por solve this


sorry and grettings :S

sl1pkn07 commented on 2014-02-14 06:15

update to 34.0.1838.2

- The pepperflash flags now store in own file (/etc/chromium-dev/pepper-flash) instead in /etc/chromium-dev/default. I've changed the wrapper launcher for handle this (/usr/bin/chromium-dev).
- Unittest is now "stable", but disable by default. only for debugging


@ All users with PaX environment:

_use_pax=1 works?

greetings

jeagoss commented on 2014-02-08 03:08

I don't think aura decoration should be turned on by default. However, I think the PKGEXT variable should be removed. People need to control that in their /etc/makepkg.conf. Not in a PKGBUILD file.

Nordlicht commented on 2014-02-07 17:42

"POLL:

active Aura decorations by default in next release?

- Yes
- No"

Devinitly no. There are a lot of showstoppers atm (Dropdown fields, Theme bugs, etc.). I would wait a while, until the bigger bugs are solved.

sl1pkn07 commented on 2014-02-07 03:24

@Lucy

new found: https://code.google.com/p/chromium/wiki/NinjaBuild#Parallelism

in next release remove all ${makeflags}

greetings

sl1pkn07 commented on 2014-02-07 03:18

updated to 34.0.1825.4

@Lucy

according of this: http://martine.github.io/ninja/manual.html#_running_ninja

ninja by default no need parse -j. anyway, i've added method to use "-j N" instead "-jN" from $MAKEFLAGS (remember: to use this variable is need comment out MAKEFLAGS in /etc/makepkg.conf)


for @ALL

POLL:

active Aura decorations by default in next release?

- Yes
- No

greetings

sl1pkn07 commented on 2014-02-07 03:15

updated to 34.0.1825.4

@Lucy

according of this: http://martine.github.io/ninja/manual.html#_running_ninja

ninja by default no need parse -j. anyway, i add method to add "-j N" instead "-jN" from $MAKEFLAGS (remember: to use this variable is need comment out MAKEFLAGS in /etc/makepkg.conf)


for @ALL

POLL:

active Aura decorations by default in next release?

- Yes
- No

greetings

Anonymous comment on 2014-02-04 15:52

MAKEFLAGS shouldn't be passed to ninja (lines 359 and 364) as it doesn't support the same flags as make, so it fails with what's normally a valid MAKEFLAGS.

roukoswarf commented on 2014-02-02 04:45

build is currently broken for me.

Hunk #1 succeeded at 420 (offset 50 lines).
patching file chrome/browser/ui/gtk/autofill/autofill_popup_view_gtk.cc
patching file chrome/browser/ui/gtk/browser_actions_toolbar_gtk.cc
patching file chrome/browser/ui/gtk/location_bar_view_gtk.cc
-> Fix to really use python2.

just hangs there, launches a find process, but it doesnt seem to be doing much.

sl1pkn07 commented on 2014-02-01 21:07

works for me

http://privatepaste.com/a55def467d

lockheed commented on 2014-02-01 20:03

There is a problem with downloading the package:
-> Downloading chromium-34.0.1809.0.tar.xz...
--2014-02-01 21:01:38-- https://commondatastorage.googleapis.com/chromium-browser-official/chromium-34.0.1809.0.tar.xz
Resolving commondatastorage.googleapis.com (commondatastorage.googleapis.com)... 127.0.0.1, 2a00:1450:4001:802::100b
Connecting to commondatastorage.googleapis.com (commondatastorage.googleapis.com)|127.0.0.1|:443... failed: Connection refused.
Connecting to commondatastorage.googleapis.com (commondatastorage.googleapis.com)|2a00:1450:4001:802::100b|:443... failed: Address family not supported by protocol.
==> ERROR: Failure while downloading chromium-34.0.1809.0.tar.xz
Aborting...


I found a misplaced " inside the file, but correcting it did not help.

sl1pkn07 commented on 2014-01-30 23:45

thanks @jeagoss!

udate to 34.0.1809.0

jeagoss commented on 2014-01-30 21:19

And here is the full modification set.

http://src.chromium.org/viewvc/chrome?view=revision&revision=247501

jeagoss commented on 2014-01-30 21:19

Here is the fix.

http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/ui/gtk/autofill/autofill_popup_view_gtk.cc?r1=242491&r2=247501

misc commented on 2014-01-30 15:50

The GTK issue was fixed after 1809 already, but dunno by which commit.

sl1pkn07 commented on 2014-01-30 03:05

https://dl.dropboxusercontent.com/u/6596386/chromium-dev-34.0.1809.0-1.src.tar.gz

sl1pkn07 commented on 2014-01-30 03:04

file:///home/sl1pkn07/Dropbox/Public/chromium-dev-34.0.1809.0-1.src.tar.gz

jeagoss commented on 2014-01-30 02:14

Do you host copies of the PKGBUILD anywhere? I'd like to give it a go and see if I can figure anything out.

sl1pkn07 commented on 2014-01-30 02:07

fail build with GTK decorations, but build OK with AURA decorations

https://code.google.com/p/chromium/issues/detail?id=339158

update package anyway?

sl1pkn07 commented on 2014-01-17 14:59

if the hares waiting...

update to 34.x series

sl1pkn07 commented on 2014-01-09 22:53

waiting for libwebp update....

misc commented on 2013-12-11 13:48

New version needs python_ver=2.7 added to GYP_DEFINES.

dimaqq commented on 2013-12-11 11:17

If anyone wants a prebuilt 64-bit arch package with aura, I can offer this:
chromium-dev-33.0.1729.3-1-x86_64.pkg.tar.xz (45MB, md5 183b4b03ee45354ac132f483ab02b1cc)
built-in: aura, ncl, pepper-flash, pdf
can be installed and ran side-by-side with regular chromium, profiles are not shared.
however, requires that you uninstall chromium's separately packaged pepper-flash and pdf. feel free to reinstall these later and disable in chromium-dev.

https://docs.google.com/file/d/0B3d3mgl1sAwLOUdwOEhDNU12alk/edit?pli=1

dimaqq commented on 2013-12-11 08:53

D'oh! I'm so stupid... ran makepkg and I see these two files:

-rw-r--r-- 1 root root 185371392 Dec 10 14:53 chromium-33.0.1729.3.tar.xz
-rw-r--r-- 1 root root 45142436 Dec 10 19:39 chromium-dev-33.0.1729.3-1-x86_64.pkg.tar.xz

and I was trying to install the first instead of the second...

sl1pkn07 commented on 2013-12-02 01:29

because ninja and python2-ply is stored in AUR

govno.ebanoe commented on 2013-12-02 00:45

Please add to dependencies "ninja" and "python2-ply" packages. On clean system, makepkg says "Not found in resitories".

sl1pkn07 commented on 2013-11-23 17:32

finaly update to 33 series

add _use_aura= if you want use new Aura Decoration


greetings

sl1pkn07 commented on 2013-11-22 05:28

update to 33.0.1711.3

- Add option to enable Aura Toolkit (disable by default)

greetings

misc commented on 2013-11-18 16:59

The gnome-keyring issue appears fixed in the next release.

misc commented on 2013-11-15 18:05

No, I meant the behavior that is controlled by the "Use System Title Bar and Borders" option, which is gone/disabled in 1707.0. It's back/fixed in the version I'm currently running (1711.0); dunno if the keyring issue was fixed too, though.

sl1pkn07 commented on 2013-11-14 17:58

@ Misc

"native window decorations" means Aura?

-Duse_aura=1
-Dtoolkit_uses_gtk=0

?

misc commented on 2013-11-13 17:11

BTW the native window decorations are entirely broken, thus theming/Linux integration messed up, and for whatever reason compilation took a bunch of GB more, so you might want to skip this 1707.

misc commented on 2013-11-13 16:42

Apparently somebody killed an option check, because gnome-keyring was mandatory for my compilation of 1707 — meaning I had to enable Duse_gconf, Dlinux_link_gnome_keyring and Duse_gnome_keyring, which also implies installing libgnome-keyring.

Moreover, I had to hardcode the include to <gnome-keyring.h> in chrome/browser/password_manager/native_backend_gnome_x.{h,cc} to "/usr/include/gnome-keyring-1/gnome-keyring.h" or it wouldn't find it.

sl1pkn07 commented on 2013-11-01 18:02

try now.

update to 32.0.1687.2

fixed missing "$" in _use_pax variables

test0 commented on 2013-11-01 17:55

I circumvented the problem by setting -Dv8_use_snapshot=0.

sl1pkn07 commented on 2013-11-01 15:51

ow. missing "$"

fixed in next release (in 2 or 3 hours)

test0 commented on 2013-11-01 15:42

Does not compile when grsecurity/pax mprotect is active even with _use_pax set. Either mksnapshot is never marked with paxctl, or (which is more probable) it is needlessly recompiled and overwritten.

sl1pkn07 commented on 2013-10-29 15:12

why? not works with ninja-git?

grimboy commented on 2013-10-29 00:15

Relies on too specific a version of ninja.

sl1pkn07 commented on 2013-10-17 19:04

Need rebuild again

juston_li commented on 2013-10-17 18:57

icu update today seems to breaks chromium
/usr/lib/chromium-dev/chromium: error while loading shared libraries: libicui18n.so.51: cannot open shared object file: No such file or directory

sl1pkn07 commented on 2013-10-16 18:21

ninja and python2-ply set in makedepends

Nordlicht commented on 2013-10-16 17:31

New dependencies: ninja, python2-ply. Sadly, it doesn't build. I only get the error: "ninja: build stopped: subcommand failed."

juston_li commented on 2013-10-02 06:19

chromium-system-icu-r0.patch has a sha1sum mismatch, doesn't matter since PKGBUILD doesn't use it, commented out.

On a separate note, Do you know how I would go about includeing https://code.google.com/p/chromium/wiki/ChromotingBuildInstructions in the build.

Thanks!

juston_li commented on 2013-10-02 06:17

chromium-system-icu-r0.patch has a sha1sum mismatch, PKGBUILD doesn't seem to use it anyways though commented out.

On a separate note, How would I go about include https://code.google.com/p/chromium/wiki/ChromotingBuildInstructions into the build

Thanks!

sl1pkn07 commented on 2013-09-28 14:01

update 31.0.1650.0

- sync with gentoo ebuild
- back chromedriver
- now use bundle V8

sl1pkn07 commented on 2013-09-10 20:14

update 31.0.1622.0

- fix build with libwebp (disable for the moment)
- disable build with system ffmpeg (solve gmail, google+ and other fail render)
- change method to remove unused components (now use python tool include in sources)

greetings

crondog commented on 2013-09-08 01:27

Just to let others know (if you didn't know already) I managed to get it working with use_system_ffmpeg=0. You need to remove the chromium-system-ffmpeg-r7.patch and uncomment install -Dm755 libffmpegsumo.so "${pkgdir}/usr/lib/chromium-dev/libffmpegsumo.so"

I found this out from http://packages.gentoo.org/package/www-client/chromium

Anonymous comment on 2013-09-04 23:21

related issue on chromium bug tracker:
https://code.google.com/p/chromium/issues/detail?can=2&start=0&num=100&q=&colspec=ID%20Pri%20M%20Iteration%20ReleaseBlock%20Cr%20Status%20Owner%20Summary%20OS%20Modified&groupby=&sort=&id=278876

t4c commented on 2013-09-04 12:34

media-video/ffmpeg-1.2.2

sl1pkn07 commented on 2013-09-04 10:26

What version of ffmpeg use gentoo?.

t4c commented on 2013-09-04 10:08

Same here, Chromium crashed on google.*/ig, gmail and many more pages.
Only on Archlinux, no problems with devchannel on Gentoo Linux, so it seems to be AUR problem.

Anonymous comment on 2013-09-02 18:07

Crashes on G+, Gmail, Youtube (with HTML5 active) and many other web-site.
Test on 30.x.x.x and 31.x.x.x
I have investigating the issue. This problem with ffmpeg. I build with use_system_ffmpeg=1 and also with use_system_ffmpeg=0. Every time crash. I have create test page here: http://legendar.info/tmp/html5video/index.html (click on any button crash the page). Nothing interesting in traceback, but only one in dmesg:

[18062.849091] Media[2262]: segfault at 0 ip 00007f6ed98181c0 sp 00007f6eb7ffe7f8 error 4 in libavutil.so.52.38.100[7f6ed980a000+35000]

I also try build ffmpeg-full from aur. Nothing helps =(

juston_li commented on 2013-09-01 23:12

Crashes on Google+ for me too, otherwise fine.

crondog commented on 2013-09-01 08:50

Just wondering if anyone else is having crashes on gmail and google plus while using version >=30.x.x.x. I am currently running 29.0.1547.0 with no problems.

misc commented on 2013-08-12 01:39

Seems like they're gonna drop the custom path and filename for the sandbox in the next release. Its install line will need to be changed from chromium-sandbox to chrome-sandbox.

shaunren commented on 2013-07-30 14:41

@t4c same here.

t4c commented on 2013-07-29 09:04

@xavier114fch update you v8-dev package.

I get a segfault with aur/chromium-dev 30.0.1573.2-0:

(gdb) run
Starting program: /usr/lib/chromium-dev/chromium
warning: no loadable sections found in added symbol-file system-supplied DSO at 0x7ffff7ffa000
warning: Could not load shared library symbols for linux-vdso.so.1.
Do you need "set solib-search-path" or "set sysroot"?
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/libthread_db.so.1".
[New Thread 0x7fffdf247700 (LWP 12248)]
[New Thread 0x7fffdea46700 (LWP 12249)]
[New Thread 0x7fffde245700 (LWP 12250)]
[New Thread 0x7ffff7ff9700 (LWP 12251)]
[New Thread 0x7ffff7f7d700 (LWP 12252)]
[New Thread 0x7fffdda44700 (LWP 12253)]
[New Thread 0x7ffff7fd8700 (LWP 12254)]
[New Thread 0x7fffdc8db700 (LWP 12255)]
[New Thread 0x7fffbffff700 (LWP 12256)]
[New Thread 0x7fffbf7fe700 (LWP 12257)]
[New Thread 0x7fffbeffd700 (LWP 12258)]
[New Thread 0x7fffbe7fc700 (LWP 12259)]
[New Thread 0x7fffbdffb700 (LWP 12260)]
[New Thread 0x7fffbd7fa700 (LWP 12261)]
[New Thread 0x7fffbc88b700 (LWP 12262)]

Program received signal SIGSEGV, Segmentation fault.
0x00007fffed0edbe5 in __memcpy_ssse3_back () from /usr/lib/libc.so.6

xavier114fch commented on 2013-07-27 04:28

Build error: http://pastebin.com/UR12nZ2h

xavier114fch commented on 2013-07-26 03:13

30.0.1573.2 is available now.
https://gsdview.appspot.com/chromium-browser-official/chromium-30.0.1573.2.tar.xz

V8 3.20.7 is also there.
https://gsdview.appspot.com/chromium-browser-official/v8-3.20.7.tar.bz2

sl1pkn07 commented on 2013-07-25 00:01

chromium-30.0.1573.2.tar.xz no aviable on https://gsdview.appspot.com/chromium-browser-official/ or https://commondatastorage.googleapis.com/chromium-browser-official/

no sources = no party :(


PS: send comment from Euskal Encounter 21

misc commented on 2013-07-24 09:16

dev's been updated to 1573.2 just now.

Some of the things I needed to do for compilation until now. Not sure if all are still required:

1. ICU includes patch (for reversal if its use_system is set): http://pastebin.com/1k64p5ht
2. V8-git patch (for reversal only *after* replace_gyp_files.py): http://pastebin.com/bAC2tktn
3. add libusb to exclude list
4. add libwebp to exclude list (and disable its use_system line)
5. add ply to exclude list

xavier114fch commented on 2013-07-24 06:22

Latest version (30.0.1568.0) is available now:
https://gsdview.appspot.com/chromium-browser-official/chromium-30.0.1568.0.tar.xz

v8-dev also needs to update to 3.20.6.0.

sl1pkn07 commented on 2013-07-23 13:28

https://code.google.com/p/chromium/issues/detail?id=262123

misc commented on 2013-07-23 11:06

Looks like they're generated at about one tarball per hour. Since minor version bumps are included, this could take another day or two.

xavier114fch commented on 2013-07-22 02:59

Seems they are coming back, but the version lags behind the latest one shown on their release calendar.
http://www.chromium.org/developers/calendar

sl1pkn07 commented on 2013-07-16 21:28

again. no sources (what fu*cking happen?)

sl1pkn07 commented on 2013-07-08 22:39

no sources again ¬¬

sl1pkn07 commented on 2013-06-27 10:14

fixed

xavier114fch commented on 2013-06-27 02:27

Checksum failed for the main xz package. The correct one is: f868be3b5939dc1aa83fe64fa1ca24b625dfd8e8

Det commented on 2013-06-26 09:31

Does ninja change so harshly upon every bugfix release that 'ninja>=1.3.4' is overdoing it?

n1f7 commented on 2013-06-20 12:26

Thanks for that

sl1pkn07 commented on 2013-06-20 12:20

ok, update pkgbuild

n1f7 commented on 2013-06-20 12:13

I had old version of ninja. I was able to build chromium only after I updated to recent ninja-1.3.4-1, but build should warn about outdated ninja instead of throwing error. I might misunderstood something, sorry for inconvenience

sl1pkn07 commented on 2013-06-20 12:04

mmm why you need mark the version in ninja array?

n1f7 commented on 2013-06-20 11:23

It should be 'ninja=1.3.4' then

sl1pkn07 commented on 2013-06-20 11:16

it's makedepend

makedepends=('elfutils' 'gperf' 'mesa' 'ninja' 'perl-json' 'python2-jinja' 'python2-ply' 'python2-simplejson' 'subversion' 'xdg-utils' 'yasm') # 'libsrtp'

n1f7 commented on 2013-06-20 11:13

Nope, it's working with ninja-1.3.4-1 but it's not listed as a dependancy

sl1pkn07 commented on 2013-06-20 10:56

what version of ninja? not working with ninja-1.3.4-1?

n1f7 commented on 2013-06-20 10:45

Please also add versioned dependency to ninja, this should fix the problem

n1f7 commented on 2013-06-20 10:34

I got this ninja error again
Updating projects from gyp files...
ninja: Entering directory `out/Release'
ninja: error: build.ninja:17: unexpected variable 'deps'
deps = gcc
^ near here
==> ERROR: A failure occurred in build().
Aborting...

sl1pkn07 commented on 2013-06-20 10:08

roger

test0 commented on 2013-06-20 08:37

(Which actually means: Please add a versioned dependency.)

test0 commented on 2013-06-20 08:33

Funny, I got a build error somewhere else. (v8-related)

n1f7 commented on 2013-06-20 05:03

I have this error:
-> Fix to really use python2.
==> Prepare NaCl SDK
==> Starting build()...
-> Building Chromium...
Updating projects from gyp files...
ninja: Entering directory `out/Release'
ninja: error: build.ninja:17: unexpected variable 'deps'
deps = gcc
^ near here
==> ERROR: A failure occurred in build().
Aborting...

sl1pkn07 commented on 2013-06-16 14:03

roger!

misc commented on 2013-06-16 13:43

With 29.0.1539.0 they've *finally* resumed to provide new tarballs — without filling in the missing ones, however.

Anyway, in order to build three changes to the PKGBUILD are required:
1. Exclude third_party/usrsctp/*
2. Exclude third_party/lzma_sdk/*
3. Add perl-json to makedepends.

test0 commented on 2013-06-08 17:04

The dropbox-hosted tarball works for me.

sl1pkn07 commented on 2013-06-08 08:00

https://dl.dropboxusercontent.com/u/6596386/chromium-dev-29.0.1521.3-0.src.tar.gz

try with this pkgbuild

sl1pkn07 commented on 2013-06-08 07:57

install harfbuzz-icu

sl1pkn07 commented on 2013-06-08 07:57

install harzbuzz-icu

juston_li commented on 2013-06-08 04:58

Failed to open Chormium after upgrading harfbuzz too.
/usr/lib/chromium-dev/chromium: symbol lookup error: /usr/lib/chromium-dev/chromium: undefined symbol: hb_icu_get_unicode_funcs

Downgraded harfbuzz

xavier114fch commented on 2013-06-08 03:47

Fail to build after harfbuzz upgrade.

extra/harfbuzz 0.9.18-1 [installed]
OpenType text shaping engine

http://pastie.org/8021767

sl1pkn07 commented on 2013-06-07 11:46

no sources = no party :(

misc commented on 2013-06-06 20:38

Crash reporting is not available for Chromium. From their FAQ page:

> Chrome handles its own crashes by linking in Breakpad, whereas Chromium does not link in Breakpad. Therefore, when a Chromium crash occurs, it is simply ignored.

You can still get a classic trace, see the ArchWiki "Debug - Getting Traces" for that. (Warning — I did that one, the compilation took ages and the package grew to little less than 900 MiB.)

juston_li commented on 2013-05-30 14:18

Perhaps a dumb question but is crash reporting disabled?
I can't seem to find a way to retrieve crash ID. Say's "Crash reporting is disabled. Crash reporting is not available in Chromium." And the report crash statistics is not in settings.

Thanks for the packages by the way

juston_li commented on 2013-05-30 14:17

Perhaps a dumb question but is crash reporting disabled?
I can't seem to find a way to retrieve crash ID. Say's "Crash reporting is disabled. Crash reporting is not available in Chromium." And the report crash statistics is not in settings.

Thanks for the packages by the way

sl1pkn07 commented on 2013-05-10 01:02

update to 28.0.1500.5

- toolchain_linux -> no need anymore, thanks @misc for the tip

note for:
* Instead of patching libvpx, adding "-Dlibvpx_source=/usr/include" to _flags does the same.

I prefer patch, and none of the two ways is necessary en next version https://code.google.com/p/chromium/issues/detail?id=226860#c18

greetings

misc commented on 2013-05-09 18:01

Two notes:
* toolchain_linux_x86.tar.bz2 isn't necessary, at least I've never downloaded it.
* Instead of patching libvpx, adding "-Dlibvpx_source=/usr/include" to _flags does the same.

sl1pkn07 commented on 2013-05-08 23:12

update to 28.0.1500.3

- add initial support to unit test (disable for default, need huge test (for me get out of memory (?¿))
- add some patch from gentoo build
- add part of code from @misc pkgbuild (pnacl related) (tnx bro)
- back need lib32-gcc-libsa and lib32-zlib. need active multilib repository in pacman.conf (finaly make more test (without trolling me) with chrooted installation, yes, only need these two packages, no need more.) and add warning message if detect x64 systems
- now use own v8 packgae called v8-dev, this package is updated when released dev channel

thanks all for the patient

greetings and enjoy! http://www.youtube.com/watch?v=EsMNqUVv3mQ

sl1pkn07 commented on 2013-05-08 16:05

jummm. I think I tested the pkgbuild only with lib32-zlib and lib32-gcc and fail with missing files http://paste.rufian.eu/6YXA/raw/, only build with rebuild chroot with gcc-multilib (include denpendencies)

well, I try pkgbuild again later when resolve build new version (ICU related?)

thanks for your pkgbuild @misc. i take some parts to include in my pkgbuild (pnacl related)

greetings.

sl1pkn07 commented on 2013-05-08 15:40

jummm. I think I tested the pkgbuild only with lib32-zlib and lib32-gcc and fail with missing files http://paste.rufian.eu/6YXA/raw/, only build with rebuild chroot with gcc-multilib (include denpendencies, without any lib32 packages)

well, I try pkgbuild again later when resolve build new version (ICU related?)

thanks for your pkgbuild @misc. i take some parts to include in my pkgbuild (pnacl related)

greetings.

misc commented on 2013-05-08 15:06

Here's my PKGBUILD; I can compile with nacl/pnacl enabled using just the "pure" x64 gcc (though lib32-zlib and lib32-gcc-libs are installed):

http://pastie.org/7818390

It's a bit of a frankenbuild with some changes, not all are annotated.

sl1pkn07 commented on 2013-05-08 13:24

https://code.google.com/p/chromium/issues/detail?id=238984

:/

sl1pkn07 commented on 2013-05-08 08:49

lib32-zlib/lib32-gcc-libs is part of [multilib],

I test build in a chrooted environment [1], these libraries not need, only need gcc-multilib (or multilib-devel)[2] (in new instalation)

[1] https://wiki.archlinux.org/index.php/DeveloperWiki:Building_in_a_Clean_Chroot
[2] https://www.archlinux.org/groups/x86_64/multilib-devel/

TheWaffleGuy commented on 2013-05-08 03:56

I've always compiled chromium with just lib32-zlib and lib32-gcc-libs, like the old PKGBUILD and it worked fine on my x86_64 system.

sl1pkn07 commented on 2013-05-07 22:30

roger

misc commented on 2013-05-07 22:25

The gold linker is now stable enough to be used for Chromium; it's much faster and requires much less memory. (It's also possible to rm -rf the chromium-build/third_party folder after make if one needs the space for package(), like I used to do with my 3G tmpfs.)

To use it, simply add gold/* to the third_party exclusions and toggle the two use_gold flags.

sl1pkn07 commented on 2013-05-05 15:14

@macau

if is possible fill bug in chromium bug tracker

http://crbug.com

greetings

sl1pkn07 commented on 2013-05-05 13:02

build with clang is broken for the moment. use gcc instead

sl1pkn07 commented on 2013-05-05 12:57

build with clang is broken at the moment. use gcc instead

Anonymous comment on 2013-05-05 11:19

Trying compilation with clang. with clean pkgbuild i catch a bug at start:
/usr/lib64/gcc/x86_64-unknown-linux-gnu/4.8.0/../../../../include/c++/4.8.0/type_traits:268:39: error: use of undeclared
identifier '__float128'
struct __is_floating_point_helper<__float128>

winth -std=c++11 flag compilation go ok, but then i catch other bug:
In file included from sync/engine/apply_control_data_updates.cc:11:
In file included from ./sync/syncable/directory.h:12:
./base/file_util.h:416:11: error: use of undeclared identifier 'typeof'
if (HANDLE_EINTR(close(*x)) < 0)
^
./base/posix/eintr_wrapper.h:24:3: note: expanded from macro 'HANDLE_EINTR'
typeof(x) eintr_wrapper_result; \
^
In file included from sync/engine/apply_control_data_updates.cc:11:
In file included from ./sync/syncable/directory.h:12:
./base/file_util.h:416:11: error: use of undeclared identifier 'eintr_wrapper_result'
./base/posix/eintr_wrapper.h:26:5: note: expanded from macro 'HANDLE_EINTR'
eintr_wrapper_result = (x); \
^
In file included from sync/engine/apply_control_data_updates.cc:11:
In file included from ./sync/syncable/directory.h:12:
./base/file_util.h:416:11: error: use of undeclared identifier 'eintr_wrapper_result'
./base/posix/eintr_wrapper.h:27:12: note: expanded from macro 'HANDLE_EINTR'
} while (eintr_wrapper_result == -1 && errno == EINTR); \
^
In file included from sync/engine/apply_control_data_updates.cc:11:
In file included from ./sync/syncable/directory.h:12:
./base/file_util.h:416:11: error: use of undeclared identifier 'eintr_wrapper_result'
./base/posix/eintr_wrapper.h:28:3: note: expanded from macro 'HANDLE_EINTR'
eintr_wrapper_result; \
^
In file included from sync/engine/apply_control_data_updates.cc:11:
In file included from ./sync/syncable/directory.h:12:
./base/file_util.h:416:35: error: invalid operands to binary expression ('void' and 'int')
if (HANDLE_EINTR(close(*x)) < 0)
~~~~~~~~~~~~~~~~~~~~~~~ ^ ~
5 errors generated.

sl1pkn07 commented on 2013-05-05 05:23

@jeagos, tested the package in chroot build

yes, need multilib-devel (tested only with gcc-multilib) in makedepends

but this have a problem, this is not a normal makedepend, because afected all system.

Little explain (my english is limited): other normal makedepends, after build package only need remove that with pacman and go.
Install multilib-devel implies active [multilib] and alternate play with [core] gcc (and depends) before make the pkgbuild (for resolve cycle dependency) if want have the system clean (more people want pure 64bits systems)

I only see two ways:

1) add warning in pkgbuild to need active [multilib] before build package (with all that implies)

2) change method to build Nacl components.
this method is back to use https://www.archlinux.org/packages/extra/x86_64/nacl-toolchain-newlib/.

I don't know if this package is out of date, and build in separate package with recent version of source stored in http://gsdview.appspot.com/nativeclient-archive2/x86_toolchain/ take a LOOOOOONNNGGGGG time (i remember need bit more time than chomium)

then:

build Nacl with actual method:

pros:
- all code up to date
- no need huge size dependency, nacl-toolchain-newlib have 102Mb (uncompress) only for build (no need for run). need all time when need build chromium. Toolchain include in the actual method is cleaned after build by helpers, like yaourt (if use remote install).

cons:
- need enable multilib repos and install multilib-devel group before build chromium
- if install multilib-devel is complicated (but not impossible) go back

older method:

pros:
- no need multilib (for pure 64bits system)
- the nacl-toolchain-newlib package is maintained by distribution

cons:
- code outdated
- need install this package when need build chromium

ladies and gentlemen, open poll :)


Notice in next version: the actual v8 package in extra is "mess". don't build chromium with that (now build with a static lib provide by chromium bundle).

I make a new package, calling "v8-dev" with some changes than the [extra] v8 package (use the real method to build this lib, like a Gentoo package) (and yes, if any ask, this package is a little bit clone like gentoo chromium-dev package).

this package update when dev channel is released

if all people have any question, commentary or insult xd. please say it :)

gretings and sorry for the huge post (and of course, for my "burn brain" english)

jeagoss commented on 2013-05-04 07:05

multilib-devel only does the following

binutils-multilib - replaces binutils
gcc-multilib - replaces gcc
gcc-libs-multilib - replaces gcc-libs
libtool-multilib - replaces libtool
lib32-gcc-libs - addition
lib32-glibc - addition

lib32-zlib is not needed

As always, it is a pleasure working with you :)

sl1pkn07 commented on 2013-05-04 06:53

no, all my systems have multilib actived. yes, is possible for build nacl components need add lib32 makedepends or multilib-devel like you say,

in older pkgbuilds have in makedepends array this packages: ('lib32-zlib' 'lib32-gcc-libs')

can you test if only need that packages instead all multilib-devel? (for save donwnloads tons of dependencies)

i need mount a chroot, but don't have space and time.

i try to do something in the next few days

thanks for the (regression?) report

greetings

and sorry my bad english :S

jeagoss commented on 2013-05-04 06:39

I have no idea if this has been mentioned before, but I ran into a problem compiling this on a vanilla x64 install. The build kept erroring out with a 'No such file or directory' error with x86_64-nacl-gcc. To fix the problem, I installed multilib-devel. Once the multilib tools were installed, it compiled fine. I've reproduced this on 3 different systems now and fixed it with multilib-devel.

@sl1pn07 - Have you built this on a system without multilib-devel? If not, should it be considered a builddep?

jeagoss commented on 2013-05-04 06:29

I have no idea if this has been mentioned before, but I ran into a problem compiling this on a vanilla x64 install. The build kept erroring out with a 'No such file or directory' error with x86_64-nacl-gcc. To fix the problem, I installed multilib-devel. Once the multilib tools were installed, it compiled fine. I've reproduced this on 3 different systems now and fixed it with multilib-devel.

@sl1pn07 - Have you built this on a system without multilib-devel? If not, should it be considered a builddep?

sl1pkn07 commented on 2013-05-02 02:57

update to 28.1490.2

adapt pkgbuild to new system to set use_system_foo libraries

greetings

sl1pkn07 commented on 2013-05-02 01:44

the new version need pkgbuild rework, please have patience
greetings

n1f7 commented on 2013-04-18 14:11

Ok. Got it working with gcc and without -flto flag

n1f7 commented on 2013-04-18 10:24

Thanks for update, but still segfaulting on both gcc and clang builds

sl1pkn07 commented on 2013-04-17 12:01

sorry the delay, yesterday don't have good day :(

finaly update to 28.0.1478.0

- fix build with GCC 4.8
- clean dependencies array
- rework several PKGBUILD structure for pacman 4.1
- "fix" in pkgbuild build with clang (make segfault)
- disable build with internal sqlite
- disable pnacl support
- other forgotten changes...

greetings

sl1pkn07 commented on 2013-04-14 04:13

libpdf is is include directly in the pkgbuild by option (enable), no need external package

emmerkar commented on 2013-04-14 02:42

Compiled without problems, but segfaulting at launch.
Removing the package gives me this message:
chromium-libpdf optionally requires chromium-dev: development version of default browser, requires modifying the PKGBUILD

n1f7 commented on 2013-04-09 08:19

Thanks. Also segfaulting

sl1pkn07 commented on 2013-04-09 07:23

http://sl1pkn07.no-ip.com/paste/view/8b0255a0

n1f7 commented on 2013-04-09 07:15

Please provide updated pkgbuild, I get this with clang http://sl1pkn07.no-ip.com/paste/view/ee1d38ab

sl1pkn07 commented on 2013-04-09 06:57

fail build 28.0.1469.0 with gcc4.8

build with Clang OK. but get Segment Fault :S

emmerkar commented on 2013-04-09 06:32

thanks for the fix :)

emmerkar commented on 2013-04-09 06:30

thanks for the fix :)

emmerkar commented on 2013-04-09 06:30

Well it is a intel core duo P8600 2.40GHz nmot veri new but with enough power to compile. Anyway i'll retry it with longer time to compile.

Thanks

emmerkar commented on 2013-04-09 06:29

Well it is a intel core duo P8600 2.40GHz nmot veri new but with enough power to compile. Anyway i'll retry it with longer time to compile.

Thanks

sl1pkn07 commented on 2013-04-08 12:23

thanks for the little pkgbuild fix

n1f7 commented on 2013-04-08 11:52

quickfix for PKGBUILD
http://sl1pkn07.no-ip.com/paste/view/f7aace3c

n1f7 commented on 2013-04-08 11:45

quickfix for PKGBUILD
http://sl1pkn07.no-ip.com/paste/view/f7aace3c

sl1pkn07 commented on 2013-04-08 10:30

https://code.google.com/p/chromium/issues/detail?id=227506

any help?

n1f7 commented on 2013-04-08 06:59

gcc 4.8 is fine.
Do you have enough processor power? On my atom N570(4 threads) compilation takes around 14-15 hours (without linkage).

sl1pkn07 commented on 2013-04-08 06:37

in next build clean dependencies

i don't test with gcc4.8

please wait

emmerkar commented on 2013-04-08 06:23

I forgot to say that for ttf-font the package should be ttf-ms-fonts

Emmerkar

emmerkar commented on 2013-04-08 06:12

I cannot com,pile cjhromium-dev because of a strange behaviour. it begins building the package and then it acts like it entered a compiliign loop. I launched the building process yesterday and this moring it was still there and i had the im,presison it was repeting some routine. After 10 hrs is too much i guess :)

here is my configuration:

3.8.5-1-ARCH x86_64
gcc (GCC) 4.8.0

alsa-lib 1.0.26-1
desktop-file-utils 0.21-1
expat 2.1.0-1
ffmpeg 1:1.2-2
flac 1.2.1-3
gpsd 3.7-1
gtk2 2.24.17-1
harfbuzz 0.9.15-1
hicolor-icon-theme 0.12-2
icu 51.1-1
jsoncpp 0.5.0-3
libevent 2.0.21-2
errore: impossibile trovare il pacchetto 'libjpeg'
libpng 1.5.14-1
errore: impossibile trovare il pacchetto 'libusb'
libwebp 0.3.0-1
libxslt 1.1.26-3
libxss 1.2.2-1
minizip 1.2.7-1
nspr 4.9.5-1
nss 3.14.3-2
opus 1.0.2-1
protobuf 2.5.0-3
speech-dispatcher 0.8-1
speex 1.2rc1-3
sqlite 3.7.16.1-1
errore: impossibile trovare il pacchetto 'ttf-font'
errore: impossibile trovare il pacchetto 'udev'

Libjpeg missing becuase is installed
libusb missing because libusb compact is isntalled instead
ttf-fon missing at all in official and AUR repositories
udev missing at all in official and AUR repositories

Maybe the prob is in the gcc version?


Emmerkar

sl1pkn07 commented on 2013-03-28 18:55

updated:

"fixed" by gentoo patch and disable pnacl support

greetings

sl1pkn07 commented on 2013-03-27 03:53

waiting to fix http://crbug.com/224169

greetings

misc commented on 2013-03-09 12:54

Patchie for speech-dispatcher 0.8:
http://pastie.org/6429359

sl1pkn07 commented on 2013-03-07 06:15

oks :)

thanks for testing, now wait any 32bit build status to launch release

greetings

SirWuffleton commented on 2013-03-07 04:42

Looks like something under chrome://flags was affecting it for me.
Tested with a clean profile and it worked, and resetting all of those to default helped as well.

sl1pkn07 commented on 2013-03-07 04:01

works for me

sl1pkn07 commented on 2013-03-07 03:58

works for me (see screenshot)

SirWuffleton commented on 2013-03-07 03:36

Yeah, that type of dropdown menu.
It appears as a couple of white pixels for me.

sl1pkn07 commented on 2013-03-07 03:35

scriptsafe: http://wstaw.org/m/2013/03/07/plasma-desktopC18744.png

sl1pkn07 commented on 2013-03-07 03:34

dropdown menu lite this? http://wstaw.org/m/2013/03/07/plasma-desktopa18744.png

SirWuffleton commented on 2013-03-07 03:16

Testing the prerelease PKGBUILD, I've encountered a pretty nasty bug: Dropdown menus no longer appear on extension buttons. Renders things like ScriptSafe pretty useless since they rely on those dropdown menus.

Other than that, everything else seems to build fine and work properly on x86_64.

sl1pkn07 commented on 2013-03-06 14:40

https://dl.dropbox.com/u/6596386/chromium-dev/chromium-dev-27.0.1430.0-1.src.tar.gz

please test, new enable build with system mesa (in special, 32 bits users)

@det, You could have waited a little XD

Det commented on 2013-03-06 14:25

:D. I'm not saying you're incompetent if I'm flagging your package.

I'm just letting people know there's a new one available. Like now.

sl1pkn07 commented on 2013-03-05 20:01

it's simple, if don't update package is for i found problems in build or "all" other fail (for example, sourcecode don't ship in google "ftp")

and i repeat, i build the package by myself before upload any update (of course, i'm human, make fails, like latest revision with jsoncpp), install and test some time.

you like lost 40min in build time to fail in stupid typo or bad apply change?

and yes, i following dev-channel RSS

besides, the date of launch new version almost always grabs me sleeping (time difference)

You can rest assured. this package don't dead

greetings and sorry for my enlish

Det commented on 2013-03-05 11:09

And that has anything to do with pkg-config or is somehow a bad thing because?

Forgive me a little bit here but if you're following the Dev Channel why is it a bad thing that you're actually doing so?

sl1pkn07 commented on 2013-03-04 19:32

all my email out-of-date reported is only by you

Det commented on 2013-03-04 19:19

@sl1pkn07, why should I care about a pkg-config bug not reported by me, not commented by me, not linked by me and not ever, once, mentioned by me?

sl1pkn07 commented on 2013-03-02 11:29

yes, i see

http://wstaw.org/m/2013/03/02/plasma-desktopj22940.png

for my it's not bad, but could be better

SirWuffleton commented on 2013-03-02 03:01

Is it just me, or does the new version ignore system font settings and/or uses a horrid font for tab titles?
I've noticed that the tab title text is much blurrier and narrower in this build.

sl1pkn07 commented on 2013-03-02 00:24

are?? not include??

:S

sorry guys. i don't know in my local copy have include it :S

updating

jeagoss commented on 2013-03-02 00:14

I experienced the same issue as SirWuffleton. After adding jsoncpp to the depends array and adding jsoncpp to my system from AUR, everything worked just fine.

sl1pkn07, you might update the depends array to include jsoncpp

sl1pkn07 commented on 2013-03-01 22:41

jsoncpp is as dependency

sl1pkn07 commented on 2013-03-01 22:40

jsoncpp is dependency

sl1pkn07 commented on 2013-03-01 22:38

OOOOOOO

SirWuffleton commented on 2013-03-01 22:34

In that case, could it be that jsoncpp is a dependency?
There exists a jsoncpp package in the AUR, and the presence of such a package is usually implied by use_system_X=1

I'm currently testing this theory to see if that's the issue.

sl1pkn07 commented on 2013-03-01 22:29

change use_system_jsoncpp=1 to use_system_jsoncpp=0

but build without any probelm for me :S

(always build, install and test chromium by myseft before update, is some cases several times, if have error don't upload)

SirWuffleton commented on 2013-03-01 22:28

@sl1pkno7: It's only happening for me on the new version (27.0.1425.2-1)
The previous version didn't have this issue and compiled without a problem.

n1f7 commented on 2013-03-01 22:22

@sl1pkn07

I have the same error after last update.

sl1pkn07 commented on 2013-03-01 22:20

before or after new version?

SirWuffleton commented on 2013-03-01 22:03

Current version will not build on my x86_64 box. It fails with lines in reference to libjingle's json component: http://pastie.org/6362611

sl1pkn07 commented on 2013-02-27 11:23

@Det:

https://bugs.archlinux.org/task/34046

sl1pkn07 commented on 2013-02-25 14:38

yea, i've upload pkgbuild with the change days ago

shosca commented on 2013-02-25 14:24

It builds with use_system_mesa=0

sl1pkn07 commented on 2013-02-23 19:24

http://paste.ubuntu.com/5559153

quick fix build with mesa 9.1 in 64bits (?)

use_system_mesa=${_use_mesa}

to

use_system_mesa=0

(need test, don't set out-of-date package)

sl1pkn07 commented on 2013-02-23 19:18

http://paste.ubuntu.com/5559153

quick fix build with meda 9.1 in 64bits(?)

use_system_mesa=${_use_mesa}

to

use_system_mesa=0

(need test, don't set out-of-date package)

sl1pkn07 commented on 2013-02-23 19:16

quick fix build with meda 9.1 (?)

use_system_mesa=${_use_mesa}

to

use_system_mesa=0

(need test, don't set out-of-date package)

sl1pkn07 commented on 2013-02-23 14:09

aw, sorry fail upload src.tar.gx

sorry the inconvenience

thanks Det, for once your "out-of-date" for something worth

/me runs

Anonymous comment on 2013-02-21 21:20

@sl1pkn07 works for me

sl1pkn07 commented on 2013-02-20 09:56

https://dl.dropbox.com/u/6596386/chromium-dev/chromium-dev-26.0.1410.10-1.src.tar.gz

please test

Anonymous comment on 2013-02-20 09:49

The gl_bindings_skia_in_process.o fail occurs on 32 bit

sl1pkn07 commented on 2013-02-20 04:09

oh, i have this isuue! (close cromium keep process running)

back linux_use_tcmalloc = 0

thanks for the report!

ipha commented on 2013-02-20 03:16

Seems like the tcmalloc bug is back(still there?) I have to compile it with "linux_use_tcmalloc=0" or it will leave a few processes running when it exits.

ipha commented on 2013-02-20 03:16

Seems like the tcmalloc is back(still there?) I have to compile it with "linux_use_tcmalloc=0" or it will leave a few processes running when it exits.

sl1pkn07 commented on 2013-02-20 02:48

https://bugs.gentoo.org/show_bug.cgi?id=457130

@all "gl_bindings_skia_in_process.o" fail users

32 or 64 bits?

sl1pkn07 commented on 2013-02-20 02:45

https://bugs.gentoo.org/show_bug.cgi?id=457130

ok, chage to bundle mesa in next release

Anonymous comment on 2013-02-19 17:33

I think I'm gettin the same problem Architeckt had.
Here it is: http://pastie.org/6227128

sl1pkn07 commented on 2013-02-17 21:05

if you're interested

https://mailman.archlinux.org/pipermail/arch-dev-public/2013-February/024459.html

shosca commented on 2013-02-17 20:15

No its my own package https://github.com/shosca/pkgbuild-xorg-git/blob/master/mesa-git/PKGBUILD

sl1pkn07 commented on 2013-02-17 19:36

with libgles from extra/testing?

shosca commented on 2013-02-17 19:13

I use mesa-git package: http://pastie.org/6205353

Just did a clean build with your changes and still get the same error.

sl1pkn07 commented on 2013-02-17 18:50

@shosca

how you built the package?

GLES2/gl2ext.h is part of libgles, and this is in makedepends

shosca commented on 2013-02-17 18:43

With those changes:
==> Prepare NaCl SDK
==> Building Chromium...
-> Building build project...
Updating projects from gyp files...
Traceback (most recent call last):
File "generate_bindings.py", line 1827, in <module>
sys.exit(main(sys.argv[1:]))
File "generate_bindings.py", line 1778, in main
print ResolveHeader(header, options.header_paths)
File "generate_bindings.py", line 1763, in ResolveHeader
raise Exception('Header %s not found.' % header)
Exception: Header GLES2/gl2ext.h not found.
gyp: Call to 'python generate_bindings.py --header-paths=../../third_party/mesa/MesaLib/include:../../third_party/khronos --inputs' returned exit status 1. while loading dependencies of gpu/gpu.gyp while loading dependencies of third_party/WebKit/Source/WebCore/WebCore.gyp/WebCore.gyp while loading dependencies of third_party/WebKit/Source/WebKit/chromium/WebKit.gyp while loading dependencies of content/browser/devtools/devtools_resources.gyp while loading dependencies of chrome/chrome_resources.gyp while loading dependencies of chrome/chrome.gyp while loading dependencies of build/all.gyp while trying to load build/all.gyp
==> ERROR: A failure occurred in build().
Aborting...

sl1pkn07 commented on 2013-02-17 07:55

uhmmmmmmm...

thi is weird... i build chromium without any problem

try this:

in: msg "Remove unnecesary components"

add: \! -path 'third_party/mesa/*' \

and in GYP_DEFINES="\

change "use_system_mesa=1" to "use_system_mesa=0"

greetings

sl1pkn07 commented on 2013-02-17 07:54

uhmmmmmmm...

thi is wreid... i build chromium without any problem

try this:

in: msg "Remove unnecesary components"

add: \! -path 'third_party/mesa/*' \

and in GYP_DEFINES="\

change "use_system_mesa=1" to "use_system_mesa=0"

greetings

shosca commented on 2013-02-17 00:20

Build fails with http://pastie.org/6079040

I've fixed it with http://pastie.org/6199586 but now fails with http://pastie.org/6199590

thorbert commented on 2013-02-16 18:36

I'm seeing the issue Architeckt reported as well. I'm thinking it's something missing in the gcc-4.7 patch but I haven't looked yet.

sl1pkn07 commented on 2013-02-16 18:20

added

thanks for the report

@architeckt, please, test if this fix your problem

greetings

thorbert commented on 2013-02-16 18:13

makedepends seems to be missing libegl as well.

Anonymous comment on 2013-02-10 11:13

build fails: http://pastie.org/6108981

sl1pkn07 commented on 2013-02-06 15:31

strange, before post this update i test the build some times, build without error in all times

http://paste.ubuntu.com/1616902/ (build log, finished a few minutes ago)

shosca commented on 2013-02-06 12:59

build fails with this > http://pastie.org/6079040

sl1pkn07 commented on 2013-02-06 02:36

try now

sl1pkn07 commented on 2013-02-06 02:34

ow, instal libgles

fixed soon

SirWuffleton commented on 2013-02-06 01:31

The new version doesn't compile on my box, complaining about a missing header: http://pastie.org/pastes/6076223/text

sl1pkn07 commented on 2013-02-06 00:23

finaly updated \*O*/

- back to use internal ffmpeg (wait to update system ffmpeg 1.1.1)
- back to use internal libvpx (need update to support new VP9 codec)
- back to use internal V8 engine (need downgrade ?¿?¿)
- remove and add some patches
- remove old

sorry for this long wait, but 26.0.1397.2 (dev channel for build) not allowed in google servers

greetings

sl1pkn07 commented on 2013-02-05 19:03

yes, is typo

nasedo commented on 2013-02-05 18:33

A typo?

use_sistem_harfbuzz=1 \

Anonymous comment on 2013-01-28 08:00

Any idea why chromium always fails to build with the same exact error? Here's the log: http://dpaste.com/895001/

eduardosm commented on 2013-01-20 12:21

I started with a new profile and now works, but it still stucks on some pages when I use the back button.

moonman commented on 2013-01-19 19:23

For those who are trying to build this for ARM, there are extra patches required. The topic is here: http://archlinuxarm.org/forum/viewtopic.php?f=18&t=4878

moonman commented on 2013-01-19 19:23

For those who are trying to build this for ARM, there are extra patches required. The topic is here: http://archlinuxarm.org/forum/viewtopic.php?f=18&t=4878

sl1pkn07 commented on 2013-01-19 14:44

@eduardosm

http://wstaw.org/m/2013/01/19/works_1.png

sl1pkn07 commented on 2013-01-19 14:42

@eduardosm

http://wstaw.org/m/2013/01/19/works.png

sl1pkn07 commented on 2013-01-19 14:39

ned more info to build in ARM architecture

please use http://paste.ubuntu.com/1549160 and paste the output here

sl1pkn07 commented on 2013-01-19 14:38

ned more info to ARM architecture

eduardosm commented on 2013-01-19 14:30

You cannot access the settings either, it stucks like this: https://docs.google.com/file/d/0B-CglhwBY0JqLUxWejFGOV85Smc/edit

eduardosm commented on 2013-01-19 14:19

This last version (26.0.1386.0-1) is quite defective. Some pages freeze before loading. I do not recommend updating.

eduardosm commented on 2013-01-19 14:18

This last version (26.0.1386.0-1) is quite defective. Some pages before before loading. I do not recommend updating.

rpodgorny commented on 2013-01-19 13:20

...for me, it fails to build on armv6h (raspberry pi):

==> Building Chromium...
-> Building build project...
Updating projects from gyp files...
gyp: Undefined variable asm_sources in third_party/ffmpeg/ffmpeg.gyp while loading dependencies of third_party/WebKit/Source/WebCore/WebCore.gyp/WebCore.gyp while loading dependencies of third_party/WebKit/Source/WebKit/chromium/WebKit.gyp while loading dependencies of content/browser/devtools/devtools_resources.gyp while loading dependencies of chrome/chrome_resources.gyp while loading dependencies of chrome/chrome.gyp while loading dependencies of build/all.gyp while trying to load build/all.gyp
==> ERROR: A failure occurred in build().
Aborting...

moonman commented on 2013-01-18 04:48

lol no worries. Thanks for keeping it up to date. Trying to build your version for ARM as the official one doesn't have the patch to use system ffmpeg and included one just won't build. They really messed it up...

sl1pkn07 commented on 2013-01-17 15:07

fuck

fixed & sorry

moonman commented on 2013-01-17 14:55

use_system_protobuf=1 --> no slash so something doesn't get used, probably the line after...

misc commented on 2013-01-16 13:12

tcmalloc has been working fine for me for months now.

jleclanche commented on 2013-01-16 04:30

"""
# Fix compilation with glib-2.31.6 (http://crbug.com/109527)
sed 's|glib/gutils.h|glib.h|' -i "${srcdir}/chromium-build/ui/base/l10n/l10n_util.cc"
"""

This was fixed a while back, glib/gutils.h no longer appears in the file.

sl1pkn07 commented on 2013-01-11 03:11

update to 25.0.1364.29

- finaly, build with system ffmpeg support
- fix build with system libpng 1.5x
- build with opus support (?) (html5test say no)
- add API keys from Chromium package in [extra]
- add/remove some depend & makedepend
- "clean" Pkgbuild


Poll: enable by default NaCL support in next release: yes? no?

greetings

sl1pkn07 commented on 2013-01-11 02:59

update to 25.0.1364.29

- finaly, build with system ffmpeg support
- fix build with system libpng 1.5x
- build with opus support (?) (html5test say no)
- add API keys from Chromium package in [extra]
- add/remove almost depend & makedepend
- "clean" Pkgbuild


Poll: enable by default NaCL support in next release: yes? no?

greetings

sl1pkn07 commented on 2012-12-25 13:36

fixed

jeagoss commented on 2012-12-25 09:32

In your nacl install section, there is an error on the 32bit build. The following line:

[ "${CARCH}" = "i686" ] && install -Dm755 out/Release/nacl_irt_{,_srpc}x86_32.nexe "${pkgdir}/usr/lib/${pkgname}"

nacl_irt_{,_srpc}x86_32.nexe

needs to be changed to

nacl_irt{,_srpc}_x86_32.nexe

eherranzr commented on 2012-12-13 21:27

Yesterday it didn't build, today it did. Maybe some upgrade. Thanks

sl1pkn07 commented on 2012-12-12 19:16

need more info. post more lines before error message

eherranzr commented on 2012-12-12 19:10

Get the following error while building (in /home/emilio/Chorradas):
FAILED with 1: /home/emilio/Chorradas/yaourt-tmp-emilio/aur-chromium-dev/src/chromium-build/out/Release/obj/gen/sdk/toolchain/linux_x86_newlib/bin/x86_64-nacl-g++ -c proxy/ppapi_messages.cc -o /home/emilio/Chorradas/yaourt-tmp-emilio/aur-chromium-dev/src/chromium-build/out/Release/obj.target/ppapi_ipc_untrusted/geni/newlib-x86-64/ppapi_ipc_untrusted/ppapi_messages.o -MD -MF /home/emilio/Chorradas/yaourt-tmp-emilio/aur-chromium-dev/src/chromium-build/out/Release/obj.target/ppapi_ipc_untrusted/geni/newlib-x86-64/ppapi_ipc_untrusted/ppapi_messages.o.d -m64 -mtls-use-call -fomit-frame-pointer -fasynchronous-unwind-tables -O2 -g -Wall -Werror -fdiagnostics-show-option -D__linux__ -D__STDC_LIMIT_MACROS=1 -D__STDC_FORMAT_MACROS=1 -D_GNU_SOURCE=1 -D_BSD_SOURCE=1 -D_POSIX_C_SOURCE=199506 -D_XOPEN_SOURCE=600 -DDYNAMIC_ANNOTATIONS_ENABLED=1 -DDYNAMIC_ANNOTATIONS_PREFIX=NACL_ -DNACL_BUILD_ARCH=x86 -DNACL_PPAPI_IPC_PROXY -DENABLE_PEPPER_THREADING -D_FORTIFY_SOURCE=2 -D_FILE_OFFSET_BITS=64 -DNO_TCMALLOC -DCHROMIUM_BUILD -DUSE_DEFAULT_RENDER_THEME=1 -DUSE_LIBJPEG_TURBO=1 -DUSE_NSS=1 -DENABLE_ONE_CLICK_SIGNIN -DGTK_DISABLE_SINGLE_INCLUDES=1 -DENABLE_REMOTING=1 -DENABLE_WEBRTC=1 -DUSE_PROPRIETARY_CODECS -DENABLE_PEPPER_THREADING -DENABLE_CONFIGURATION_POLICY -DENABLE_INPUT_SPEECH -DENABLE_NOTIFICATIONS -DENABLE_GPU=1 -DENABLE_EGLIMAGE=1 -DUSE_SKIA=1 -DENABLE_TASK_MANAGER=1 -DENABLE_WEB_INTENTS=1 -DENABLE_EXTENSIONS=1 -DENABLE_PLUGIN_INSTALLATION=1 -DENABLE_PLUGINS=1 -DENABLE_PROTECTOR_SERVICE=1 -DENABLE_SESSION_SERVICE=1 -DENABLE_THEMES=1 -DENABLE_BACKGROUND=1 -DENABLE_AUTOMATION=1 -DENABLE_GOOGLE_NOW=1 -DENABLE_LANGUAGE_DETECTION=1 -DENABLE_PRINTING=1 -DENABLE_CAPTIVE_PORTAL_DETECTION=1 -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -DNACL_WINDOWS=0 -DNACL_OSX=0 -DNACL_LINUX=0 -I/home/emilio/Chorradas/yaourt-tmp-emilio/aur-chromium-dev/src/chromium-build/out/Release/obj/gen/tc_newlib/include -I.. -I../ppapi -I.. -I../third_party/khronos


make: *** [out/Release/obj/gen/tc_newlib/lib64/libppapi_ipc_untrusted.a] Error 1

sl1pkn07 commented on 2012-12-08 16:15

ow :S

matchulinux commented on 2012-12-08 16:09

I think you should remove wget from makedepends since it's not needed anymore

Det commented on 2012-11-27 18:45

@lessdevrandom, that is the stupidest thing I've heard in a long time.

E: Nor frankly do I understand what's up with the whole "next version" syndrome. You don't have to raise the pkgver/rel every time you do something.

Det commented on 2012-11-27 14:50

@lessdevrandom, that is the stupidest thing I've heard in a long time.

sl1pkn07 commented on 2012-11-18 02:07

oks, in next release add curl instead wget

SirWuffleton commented on 2012-11-18 00:09

@lessdevrandom: Actually, it's still pretty easy to just use curl (took all of a couple minutes to replace the wget commands with curl)

@sl1pkn07: Consider using this PKGBUILD instead? (Replaced wget with curl and eliminated a couple unnecessary uses of cat that I saw) http://pastie.org/5394424

Anonymous comment on 2012-11-17 22:55

That is because wget is scripted into the PKGBUILD on lines 66, 67, 68 and 90. It was easier to add wget as a build dependency than to edit the PKGBUILD to use curl.

Det commented on 2012-11-17 14:43

Why wouldn't you just use curl?

sl1pkn07 commented on 2012-11-16 02:08

thanks for the report

Anonymous comment on 2012-11-16 01:52

I needed to add wget to makedepends in the PKGBUILD, but it otherwise builds and runs perfectly.

sl1pkn07 commented on 2012-11-06 12:37

python-simplejson move to community whit name python2-simplejson

change this in new release

SirWuffleton commented on 2012-11-06 09:37

Is it just me or did the makedepend 'python-simplejson' vanish from the AUR with the updated interface? Packer keeps spitting errors about it and I can't even find a trace of it in the aur-mirror-git.

sl1pkn07 commented on 2012-11-04 10:00

in the pkgbuild have this (with little changes):

rm -f *.sha1hash*
wget -q "${_nacl_sdk_path}"/naclsdk_linux_x86.tgz.sha1hash
wget -q "${_nacl_sdk_path}"/naclsdk_pnacl_linux_x86.tgz.sha1hash
wget -q "${_toolchain_path}"/toolchain_linux_x86.tar.bz2.sha1hash
_nacl_sdk_sha1sum="$(cat naclsdk_linux_x86.tgz.sha1hash)"
_pnacl_sdk_sha1sum="$(cat naclsdk_pnacl_linux_x86.tgz.sha1hash)"
_toolchain_sha1sum="$(cat toolchain_linux_x86.tar.bz2.sha1hash)"
rm -f *.sha1hash*

[ -e "${srcdir}/../naclsdk_linux_x86.tgz" ] && [ "$(sha1sum "${srcdir}/../naclsdk_linux_x86.tgz" | cut -d " " -f1)" != "${_nacl_sdk_sha1sum}" ] && rm -fr "${srcdir}/../naclsdk_linux_x86.tgz"
[ -e "${srcdir}/../naclsdk_pnacl_linux_x86.tgz" ] && [ "$(sha1sum "${srcdir}/../naclsdk_pnacl_linux_x86.tgz" | cut -d " " -f1)" != "${_pnacl_sdk_sha1sum}" ] && rm -fr "${srcdir}/../naclsdk_pnacl_linux_x86.tgz"
[ -e "${srcdir}/../toolchain_linux_x86.tar.bz2" ] && [ "$(sha1sum "${srcdir}/../toolchain_linux_x86.tar.bz2" | cut -d " " -f1)" != "${_toolchain_sha1sum}" ] && rm -fr "${srcdir}/../toolchain_linux_x86.tar.bz2"

source+=("${_nacl_sdk_path}/naclsdk_linux_x86.tgz"
"${_nacl_sdk_path}/naclsdk_pnacl_linux_x86.tgz"
"${_toolchain_path}/toolchain_linux_x86.tar.bz2")
sha1sums+=("${_nacl_sdk_sha1sum}"
"${_pnacl_sdk_sha1sum}"
"${_toolchain_sha1sum}")

steps for this:

- download the sha1file (from google) for these files and make variable with file contains. and add in sha1sums pkgbuild array
- search if exists file
--- if yes -> get the sha1 hash from existing file and compare with sha1has variable (get from sha1file download from google) if not same hash
------- if yes (not same hash) -> remove existing file and jump to next file
------- if no (same hash) -> jump to next file
--- if no -> jump to next file



make this for prevent you say

i test this: https://wiki.archlinux.org/index.php/PKGBUILD#source

Tip: You can specify a different name for the downloaded file - if the downloaded file has a different name for some reason like the URL had a GET parameter - using the following syntax: filename::fileuri, for example $pkgname-$pkgver.zip::http://199.91.152.193/7pd0l2tpkidg/jg2e1cynwii/Warez_collection_16.4.exe

if working (the build need symlink nacl/pnacl/toolchain zips with the specific names (like download files)), add in next release

greetings


pumbur commented on 2012-11-03 21:43

....

==> Validating source files with sha1sums...
....
pulse_audio_fix.patch ... Passed
naclsdk_linux_x86.tgz ... FAILED
naclsdk_pnacl_linux_x86.tgz ... FAILED
toolchain_linux_x86.tar.bz2 ... Passed


https://bugs.archlinux.org/index.php?do=details&task_id=29106

sl1pkn07 commented on 2012-10-31 14:42

@Det: what report?

jeagoss commented on 2012-10-31 01:28

@Det I apologize for not clarifying....

Det commented on 2012-10-31 00:00

@jeagoss, not as much out-of-date as meant for the _stable_ channel chromium.

@sl1pkn07, jeagoss's patch (http://archrepo.jeago.com/sources/chromium-dev/pulse_audio_fix.patch) fixes http://crbug.com/157876? Well, shouldn't this be mentioned in the report?

crondog commented on 2012-10-30 06:32

Hey sl1pkn07

I have been building chromium-dev for a while now and have always had to fiddle with the PKGBUILD manually to get it working with BUILDDIR=/tmp/makepkg. I have finally made a PKGBUILD which keeps your functionality but is able to cope with the builddir being in a different location.

--- PKGBUILD 2012-10-28 14:28:27.000000000 +1100
+++ PKGBUILD.new 2012-10-30 17:26:09.405889398 +1100
@@ -92,6 +92,7 @@
source+=("http://dl.google.com/linux/chrome/rpm/stable/"${_rpm_arch}"/google-chrome-unstable-"${pkgver}"-"${_rpm_build}"."${_rpm_arch}".rpm")
sha1sums+=("${_rpm_sha1}")
noextract+=("google-chrome-unstable-"${pkgver}"-"${_rpm_build}"."${_rpm_arch}".rpm")
+ _before_srcdir="$(pwd)"
fi

# Are we in Gnome?
@@ -372,7 +373,7 @@
install -Dm755 "${srcdir}"/"${pkgname}".sh "${pkgdir}"/usr/bin/"${pkgname}"

install -Dm644 LICENSE "${pkgdir}"/usr/share/licenses/"${pkgname}"/LICENSE
- install -Dm644 "${srcdir}"/../default "${pkgdir}"/etc/chromium-dev/default
+ install -Dm644 "${srcdir}"/default "${pkgdir}"/etc/chromium-dev/default

if [ "${_use_nacl}" = 1 ]; then
msg2 "Adding NaCl components"
@@ -406,7 +407,7 @@
for i in "${srcdir}"/opt/google/chrome/PepperFlash/*; do install -m644 "$i" "${_chromium_home}"/PepperFlash; done
chmod 775 "${_chromium_home}"/PepperFlash/libpepflashplayer.so
_flash_version="$(cat "${_chromium_home}"/PepperFlash/manifest.json | grep version | sed 's|[a-z,": ]*||g')"
- sed -e "s|use_pepperflash=0|use_pepperflash=1|" -e "s|version=0|version=${_flash_version}|" -i "${srcdir}"/../"${pkgname}".install
+ sed -e "s|use_pepperflash=0|use_pepperflash=1|" -e "s|version=0|version=${_flash_version}|" -i $_before_srcdir/"${pkgname}".install
rm -fr "${srcdir}"/opt
fi

sl1pkn07 commented on 2012-10-28 11:29

update to 24.0.1305.3

- fix http://crbug.com/157876 (thanks to @jeagoss)
- back to use internal copy of V8

sorry for delay

jeagoss commented on 2012-10-27 18:55

I have updated to 24.0.1305.3. In order to get it to compile on Arch, you'll need the following patch. http://archrepo.jeago.com/sources/chromium-dev/pulse_audio_fix.patch

Also, for all of you experiencing crashes on Facebook, Google Reader, etc... the V8 library included in Arch is out of date. Try installing the v8-svn package from aur. Doing so has solved my problems with crashing. (You might need to recompile chromium after updating v8.)

sl1pkn07 commented on 2012-10-12 09:08

@hard

try install v8 from community

sl1pkn07 commented on 2012-10-12 09:05

@cpatrick

try now

HarD commented on 2012-10-12 06:46

In file included from ./webkit/plugins/webview_plugin.h:15:0,
from webkit/plugins/webview_plugin.cc:5:
./third_party/WebKit/Source/WebKit/chromium/public/WebFrameClient.h:46:16: fatal error: v8.h: No such file or directory compilation aborted.
make: *** [out/Release/obj.target/glue/webkit/plugins/webview_plugin.o] Error 1

cpatrick08 commented on 2012-10-12 01:16

error installing curl: (22) The requested URL returned error: 404 Not Found
==> ERROR: Failure while downloading google-chrome-unstable-24.0.1290.1-159713.x86_64.rpm

sl1pkn07 commented on 2012-09-21 13:26

the _toochain_rev is determine by hand because i search the revision by date release. the 23.0.1271.1 released 20 september 2012. then search toolchain revision release same day

make this for possible incompatibility in buld

misc commented on 2012-09-21 12:49

The latest _toochain_rev can be dynamically determined eg. with:

curl -s 'https://gsdview.appspot.com/nativeclient-archive2/x86_toolchain/?marker=x86_toolchain%2Fr9801%40' | grep toolchain | tail -n2 | head -n1 | cut -d '>' -f 3 | cut -c 2-5

(Note: For some weird reason that Google server acts up every now and then, requiring one to add/remove the "%40".)

Anonymous comment on 2012-09-08 05:32

crashes some sites: facebook, orkut and some albums picasa

sl1pkn07 commented on 2012-09-05 08:38

adblock and autopagerize fails for me

SirWuffleton commented on 2012-09-05 08:33

Is anyone else experiencing constant crashes of some of their extensions with the latest version of chromium? Adblock and Flag for Chrome won't load at all for me and simply crash every time I try to reload them.

sl1pkn07 commented on 2012-08-31 10:09

good fix -> if "sudo pacman -Qo /usr/lib/libudev.so.0" result is don't provide any package. remove libudev.so.0 (rm /usr/lib/libudev.so) and rebuild chromium-dev. if other program fail to use libudev.so.0, rebuild these program (all packages in core/extra/community/testing use libudev.so.1 instead libudev.so.0) (probably AUR package)

ugly fix -> sudo rm /usr/lib/libudev.so.0 && ln -s /usr/lib/libudev.so.1.1.0 /usr/lib/libudev.so.0

sl1pkn07 commented on 2012-08-31 10:08

good fix -> if "sudo pacman -Qo /usr/lib/libudev.so.0" result is don't provide any package. remove libudev.so.0 (rm /usr/lib/libudev.so) and rebuild chromium-dev. if other program fail to use libudev.so.0, rebuild these program (all packages in core/extra/community/testing use libudev.so.1 instead libudev.so.0) (probably AUR package)

ugly fix -> rm /usr/lib/libudev.so.0 && ln -s /usr/lib/libudev.so.1.1.0 /usr/lib/libudev.so.0

sl1pkn07 commented on 2012-08-31 10:07

good fix -> if "sudo pacman -Qo /usr/lib/libudev.so.0" result is = none (nothing). remove libudev.so.0 (rm /usr/lib/libudev.so) and rebuild chromium-dev. if other program fail to use libudev.so.0, rebuild these program (all packages in core/extra/community/testing use libudev.so.1 instead libudev.so.0) (probably AUR package)

ugly fix -> rm /usr/lib/libudev.so.0 && ln -s /usr/lib/libudev.so.1.1.0 /usr/lib/libudev.so.0

lockheed commented on 2012-08-31 08:56

$ chromium-browser
/opt/chromium-browser/chrome: error while loading shared libraries: libudev.so.0: cannot open shared object file: No such file or directory

And I already have all this in /usr/lib/
libudev.a libudev.so.0 libudev.so.1.1.0
libudev.so libudev.so.1

sl1pkn07 commented on 2012-08-28 19:14

finaly update to 23.0.1246.0

- use V8 from extra (finaly update to latest version)
- use nacl-toolchain-newlib from extra instead external build (keep old code in PKGBUILD)
- add new makedependencies
- Remove unnecesary files (resources folder)

greetings and sorry the inconveniencies (i'm back to the holidays)

misc commented on 2012-08-25 20:28

Wrote a few lines that move the nacl_sdk download to pacman, thus requiring only one when it does not change over various Chromium releases. The relevant parts:

1. At the top of the PKGBUILD:
_nacl_sdk=$(curl -s 'https://commondatastorage.googleapis.com/nativeclient-mirror/nacl/nacl_sdk/naclsdk_manifest2.json' | grep linux -B 4 | tail -n 5 | head -n 1 | cut -d '/' -f 7)

(Note: The json file does change its structure occasionally.

You get the most recent Chromium tarball with eg.: "curl -s 'https://gsdview.appspot.com/chromium-browser-official/?marker=chromium-23.0.1245.0.tar.bz' | grep chromium-2 | tail -1 | cut -d '>' -f 2 | cut -d '/' -f 3 | cut -c 10-20"

To get then the version for the Chrome rpm with the Pepper Flash plugin, use "_rpm_ver="$(cat filelists.xml | grep -e "unstable" | cut -d \" -f10)"" as replacement variable.)

2. In source, add:
naclsdk_linux-$_nacl_sdk.tar.bz2::http://commondatastorage.googleapis.com/nativeclient-mirror/nacl/nacl_sdk/$_nacl_sdk/naclsdk_linux.tar.bz2
The first part adds the revision to the downloaded file's name, so pacman can know if any present one is outdated.

3. Within the "if [ "${_use_nacl}" = 1 ]; then" paragraph change:
i) only rm -fr sdk_cache
ii) add below that "mv $srcdir/pepper_2? $srcdir/nacl_sdk/"${_pepper_version}""

That should be it. However, I did not care about the pacman checksum, since I build with --skipchecksums.

misc commented on 2012-08-25 11:02

File structure has been altered with Chromium 23, now the revision line must read eg.:

curl -s 'https://commondatastorage.googleapis.com/nativeclient-mirror/nacl/nacl_sdk/naclsdk_manifest2.json' | grep 'revision' | tail -1 | cut -d ' ' -f 8

misc commented on 2012-08-22 21:35

Wrote a few lines that move the nacl_sdk download to pacman, thus requiring only one when it does not change over various Chromium releases. The relevant parts:

1. At the top of the PKGBUILD:
_nacl_sdk=$(curl -s 'https://commondatastorage.googleapis.com/nativeclient-mirror/nacl/nacl_sdk/naclsdk_manifest2.json' | grep 'trunk.*linux' | tail -1 | cut -d '/' -f 7)

(You get the most recent Chromium tarball with eg.: "curl -s 'https://gsdview.appspot.com/chromium-browser-official/?marker=chromium-23.0.1242.0.tar.bz' | grep chromium-2 | tail -1 | cut -d '>' -f 2 | cut -d '/' -f 3 | cut -c 10-20"

To get then the version for the Chrome rpm with the Pepper Flash plugin, use "_rpm_ver="$(cat filelists.xml | grep -e "unstable" | cut -d \" -f10)"" as replacement variable.)

2. In source, add:
naclsdk_linux-$_nacl_sdk.tar.bz2::http://commondatastorage.googleapis.com/nativeclient-mirror/nacl/nacl_sdk/$_nacl_sdk/naclsdk_linux.tar.bz2
The first part adds the revision to the downloaded file's name, so pacman can know if any present one is outdated.

3. Within the "if [ "${_use_nacl}" = 1 ]; then" paragraph change:
i) only rm -fr sdk_cache
ii) add below that "mv $srcdir/pepper_2? $srcdir/nacl_sdk/"${_pepper_version}""

That should be it. However, I did not care about the pacman checksum, since I build with --skipchecksums.

misc commented on 2012-08-09 15:01

Anybody else getting consistent site crashes (sad tab) on flash / HTML5 videos (apparently those using hardware acceleration)? Getting that both with the "Old" and Pepper Flash.

(Next version has the separate "ui_resources_100_percent,theme_resources_100_percent" become a single "chrome_100_percent".)

sl1pkn07 commented on 2012-07-24 19:57

finally! update new version

- Fix bison 2.6 build http://crbug.com/138243
- Fix Nacl build fail https://crbug.com/137628
- Fix build libyuv when use external libjpg-turbo build

sl1pkn07 commented on 2012-07-24 19:38

https://projects.archlinux.org/svntogit/packages.git/tree/trunk/chromium-20.0.1132.57-bison-2.6-remove-yyparse-decl.patch?h=packages/chromium

misc commented on 2012-07-24 19:21

Hacky bison 2.6 build fix in Chromium 20.0.1132.57-2

SirWuffleton commented on 2012-07-24 14:35

Just wondering, does anyone have Chromium working with the --use-pulseaudio flag? I downgraded bison and recompiled with PulseAudio present on my system, but it seems to segfault as soon as it tries to play something (I tested with a HTML5 Youtube Video).
It outputs this when run from the command line:
Assertion 's' failed at pulse/stream.c:2050, function pa_stream_set_write_callback(). Aborting.
zsh: abort chromium-dev --use-pulseaudio

sl1pkn07 commented on 2012-07-21 22:49

yes. i see in https://code.google.com/p/chromium/issues/detail?id=138243

but the latest version in core is 2.6. to fix build need downgrade bison out of repository or wait external patch (the attachment diff is generate code output from bison)

i prefeer wait external patch

misc commented on 2012-07-21 22:16

Had it happen now, too. Following the Gentoo people, a revert to bison 2.5.1 fixed it.

sl1pkn07 commented on 2012-07-20 13:48

shit. other fail build with disable NACL

https://code.google.com/p/chromium/issues/detail?id=138243

:S

misc commented on 2012-07-17 20:47

Not sure about that NaCL breakage as I don't use it. However, as for that tcmalloc-fix the one in extra's Chromium appears less tinker-ish than mine. Also, the recent addition libc_urandom_override.cc (note: it came with 1208.0 IIRC) lacks an include for GCC 4.7. Link points to both fixes in one patch: http://pastie.org/4273855
Moreover, 1210.0+ needs "\! -path 'third_party/hyphen/*' \".

sl1pkn07 commented on 2012-07-17 04:40

fail build

https://code.google.com/p/chromium/issues/detail?id=137628

sl1pkn07 commented on 2012-07-08 18:47

zankius!

misc commented on 2012-07-08 18:44

Here's the diff that I've been using to make tcmalloc work again:
http://pastie.org/4221566

misc commented on 2012-07-06 17:59

Appears that tcmalloc is again broken with glibc 2.16; compilation fails rather instantly (worked fine with 2.15).

No problem for this package though, as it's already disabled in the PKGBUILD.

SirWuffleton commented on 2012-06-22 08:34

If you're referring to a segfault when attempting to print, then I've been experiencing that for the past few versions. I'm also on x86_64.
Here's my journal output from after the crash: Jun 22 01:30:52 Night kernel: [181718.631052] chromium[2032] general protection ip:7f1d1baac40e sp:7fffeb168d70 error:0 in libgtk-x11-2.0.so.0.2400.10[7f1d1b85b000+422000]

sl1pkn07 commented on 2012-06-21 00:27

"oops" when print screen?. same issue here :/

jk121960 commented on 2012-06-20 18:57

Hi, I am using the latest build 21.0.1171.0-1 when I click to print it just closes all instances that are open. I am using Arch-64 Awesome desktop, GTK2 themeing using lxappearance. On Chromium I am using sync profiling. I hope I am supposed to post this here of not please direct me as to where. If there is anything else you need please let me know.

thanks

--jerry

misc commented on 2012-06-18 17:45

With r142750 (in 1180) the missing include for memory_details_linux has been added. IOW, amazingly no patches at all are needed — for now.

"* \! -path 'third_party/qcms/*' \" has become necessary, though.

sl1pkn07 commented on 2012-06-14 00:03

* \! -path 'third_party/libwebp/*' \ is include in this pkgbuild (+ use_system_libwebp=0)

thansk for the info

greetings

misc commented on 2012-06-13 22:36

I've been able to compile without setting use_system-libvpx at all. However as of 1173, the next version will need:

* use_system_libwebp=0
* \! -path 'third_party/trace-viewer/*' \
* \! -path 'third_party/libwebp/*' \
* No more sqlite patch as it has been integrated.

Also, there is no need for the remaining GCC 4.7 patch/additional cstddef includes anymore. Right now the memory_details_linux patch is the only one truly required.

sl1pkn07 commented on 2012-06-12 17:49

update to 21.0.1171.0

- remove disable_glibc (fail build)
- set use_system-libvpx to 0 (if set 1 fail build)
- add patch to fix memory_details_linux.cc build
- remove patch chromium-expat-r0.patch
- change patch gcc47v3.patch to gcc47v4.patch
- remove patch ipc-fix.diff

greetings

SirWuffleton commented on 2012-06-08 08:17

Looks like my ALSA issue was due to chromium not liking the IPC key I had set for dmixer. Changing it from 4096 to 3072 resolved the issue.

I'm currently compiling 21.0.1167.0 to see if the fix has been merged yet. Besides a few of the patches becoming unnecessary or requiring edits, it would seem that disable_glibc needs to be set to 0 in order to compile. The bug that required it to be set to 1 appears to have been resolved, and if set to 1 now, the build will fail. (Just a heads up for any other people wanting try the bleeding-edge tarballs for bug fixes)

jeagoss commented on 2012-06-07 22:39

The extensions problem will be fixed in the next revision.

http://src.chromium.org/viewvc/chrome?view=rev&revision=140954

SirWuffleton commented on 2012-06-05 18:24

Any other pure-ALSA users having issues with the new version and HTML5 Audio/Video?

It floods messages similar to this on the command line when attempting to do anything that makes noise:
[13373:13386:5014272109:ERROR:alsa_output.cc(670)] ALSA returned 4073 of 3763 frames available.
[13373:13386:5014457904:ERROR:alsa_output.cc(670)] ALSA returned 4058 of 3763 frames available.

sl1pkn07 commented on 2012-06-03 17:07

no crash for me, but if left click in plugins don't show menu of these plugins

sl1pkn07 commented on 2012-06-03 17:06

no crash for me, but if right click in plugins don't show menu of these plugins

misc commented on 2012-06-03 11:37

Me too. Also, Chromium crashes if I try to deactivate such extensions (eg. ScriptNo) via right click menu.

jeagoss commented on 2012-06-03 01:12

I've noticed the same behavior as DarkSniper. Trying to step through it in gdb.

SirWuffleton commented on 2012-06-03 01:04

Anyone else having weird behavior with extension menus in this version? I've found that the menus for extensions like adblock and ghostery (which have toolbar buttons next to the wrench menu) appear on the wrong side of the window if at all.

monoloco commented on 2012-05-26 02:04

i have narrowed it down to the Override software rendering list flag so if is chrome crashing for anyone disable that flag and it may help

monoloco commented on 2012-05-26 01:55

Disregard my issue i started looking further into it and it occurred to me that it might be a flag i set, i disabled Override software rendering list and GPU Accelerated SVG and CSS Filters and i've had no more crashes for at least a few hours should i delete my post?

sl1pkn07 commented on 2012-05-24 13:07

http://code.google.com/p/chromium/issues/detail?id=129221 ?

monoloco commented on 2012-05-24 08:46

I am getting many crashes when closing tabs it happens quite often on YouTube to me, also if chromium-dev is run from terminal and getting "ERROR:gles2_cmd_decoder.cc(4911)] glTexImage2D: texture is immutable" spam the console. I am not sure if they're related though.

Chromium just crashed for me and spat out "ERROR:gl_context_glx.cc(163)] Couldn't make context current with X drawable." and "ERROR:x11_util.cc(1242)] X Error detected: serial 4161, error_code 8 (BadMatch (invalid parameter attributes)), request_code 135, minor_code 5 (X_GLXMakeCurrent)"

monoloco commented on 2012-05-24 08:44

I am getting many crashes when closing tabs it happens quite often on YouTube to me, also if chromium-dev is run from terminal and getting "ERROR:gles2_cmd_decoder.cc(4911)] glTexImage2D: texture is immutable" spam the console. I am not sure if they're related though.

Chromium just crashed for me and spat out "ERROR:gl_context_glx.cc(163)] Couldn't make context current with X drawable." and "ERROR:x11_util.cc(1242)] X Error detected: serial 4161, error_code 8 (BadMatch (invalid parameter attributes)), request_code 135, minor_code 5 (X_GLXMakeCurrent)"

sl1pkn07 commented on 2012-05-23 21:48

mmm no for me (for the moment)

misc commented on 2012-05-23 20:32

Anybody else getting permanent crashes when closing a certain tab?

sl1pkn07 commented on 2012-05-23 07:10

at last!!

update to 21.0.1145.0

- remove pulseaudio patch
- mod gcc47.diff to gcc47v3.diff
- add symlink needed for build with Nacl
- add patch for http://crbug.com/129173 (include modifications in gcc47.diff)
- add changes sugsted by @misc

greetings and enjoy!

misc commented on 2012-05-22 20:59

Odd. All I can tell is that with issue10378057_5035.diff reverted, compilation for me ends successful, otherwise I get exactly the same error.

Maybe applying this patch instead (ie, without the revert) will solve it. Haven't got any time left to test it today:

https://pastebin.com/raw.php?i=VbFK7MjL

(Like last time, pastebin ate the last two lines again; one with two spaces, one blank.)

misc commented on 2012-05-22 20:27

Odd. All I can tell is that with issue10378057_5035.diff reverted, compilation for me ends successful, otherwise I get exactly the same error.

Maybe applying this patch instead (ie, without the revert) will solve it. Haven't got any time left to test it today:

https://pastebin.com/raw.php?i=2jTZiR4e

(Like last time, pastebin ate the last two lines again; one with two spaces, one blank.)

sl1pkn07 commented on 2012-05-22 19:52

yea. the fail appear with changes you say:

- "\! -path 'third_party/zlib/contrib/minizip/*' \" → "\! -path 'third_party/zlib/*' \" (change in new pkgbuild)

- "\! -path 'third_party/libxml/*' \" → find&delete or compilation fails. (done in 1132.11)

- Patch fragment for base/message_pump_libevent.cc needs to be removed as include is no longer missing. (change new pkgbuild, rejected gcc47.diff patch, new gcc47v2.diff: http://paste.kde.org/484514)

- Like Chromium in Extra, "CFLAGS+=' -Wno-c++0x-compat'" has to be added or else GCC 4.7 prints a couple of thousand useless warnings. (done in 1132.8 or 11. this add new tons of messages: cc1: warning: command line option '-Wno-c++0x-compat' is valid for C++/ObjC++ but not for C [enabled by default])

- This commit must be reverted (or perhaps a new include is missing) for compilation to not fail: http://codereview.chromium.org/10378057. (change in new pkgbuild)

changes by me:

add symbolic link from native_client/toolchain/linux_x86_newlib → native_client/toolchain/linux_x86 (fail build without this)


but fails (see post below)

need other change?

the new pkgbuild http://paste.kde.org/484568 (issue10378057_5035.diff: http://codereview.chromium.org/download/issue10378057_5035.diff)

misc commented on 2012-05-22 16:28

There's a reason that I wrote those comments, you know.

sl1pkn07 commented on 2012-05-22 15:05

g++ '-D_FILE_OFFSET_BITS=64' '-DCHROMIUM_BUILD' '-DUSE_NSS=1' '-DGTK_DISABLE_SINGLE_INCLUDES=1' '-DENABLE_REMOTING=1' '-DENABLE_WEBRTC=1' '-DENABLE_P2P_APIS=1' '-DUSE_PROPRIETARY_CODECS' '-DENABLE_CONFIGURATION_POLICY' '-DENABLE_INPUT_SPEECH' '-DENABLE_NOTIFICATIONS' '-DENABLE_GPU=1' '-DENABLE_EGLIMAGE=1' '-DUSE_SKIA=1' '-DENABLE_TASK_MANAGER=1' '-DENABLE_WEB_INTENTS=1' '-DENABLE_EXTENSIONS=1' '-DENABLE_PLUGIN_INSTALLATION=1' '-DENABLE_PROTECTOR_SERVICE=1' '-DENABLE_SESSION_SERVICE=1' '-DENABLE_THEMES=1' '-DENABLE_BACKGROUND=1' '-DENABLE_PROMO_RESOURCE_SERVICE=1' '-DENABLE_AUTOMATION=1' '-D__STDC_FORMAT_MACROS' '-DIPC_IMPLEMENTATION' '-DNDEBUG' '-DNVALGRIND' '-DDYNAMIC_ANNOTATIONS_ENABLED=0' -I. -pthread -fno-exceptions -fno-strict-aliasing -Wall -Wno-unused-parameter -Wno-missing-field-initializers -fvisibility=hidden -pipe -fPIC -pthread -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -pthread -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/gdk-pixbuf-2.0 -I/usr/include/pango-1.0 -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng15 -O2 -fno-ident -fdata-sections -ffunction-sections -march=x86-64 -mtune=native -O2 -pipe -fstack-protector --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2 -Wno-c++0x-compat -fno-rtti -fno-threadsafe-statics -fvisibility-inlines-hidden -Wsign-compare -MMD -MF out/Release/.deps/out/Release/obj.target/ipc/ipc/ipc_sync_message.o.d.raw -c -o out/Release/obj.target/ipc/ipc/ipc_sync_message.o ipc/ipc_sync_message.cc
ipc/ipc_platform_file.cc: In function 'IPC::PlatformFileForTransit IPC::GetFileHandleForProcess(base::PlatformFile, base::ProcessHandle, bool)':
ipc/ipc_platform_file.cc:39:43: error: '::dup' has not been declared
make: *** [out/Release/obj.target/ipc/ipc/ipc_platform_file.o] Error 1
make: *** Waiting for unfinished jobs....

sl1pkn07 commented on 2012-05-22 14:15

CXX(target) out/Release/obj.target/content_common/content/public/common/ssl_status.o
content/public/common/sandbox_init.cc: En la función ‘IPC::PlatformFileForTransit content::BrokerGetFileHandleForProcess(base::PlatformFile, base::ProcessId, bool)’:
content/public/common/sandbox_init.cc:35:43: error: ‘::dup’ no se ha declarado
make: *** [out/Release/obj.target/content_common/content/public/common/sandbox_init.o] Error 1
make: *** Se espera a que terminen otras tareas....

sl1pkn07 commented on 2012-05-22 12:27

little "fix" in nacl

msg "Copy NaCl SDK toolchain"
cd "${srcdir}"
_nacld=`bsdtar -tf naclsdk_linux.tgz | head -1`
bsdtar -xf naclsdk_linux.tgz "${_nacld}"toolchain/linux_x86
mv "${_nacld}"/toolchain/linux_x86 chromium-build/native_client/toolchain/linux_x86
rm -fr "${_nacld}"
cd "${srcdir}"/chromium-build

sl1pkn07 commented on 2012-05-22 12:19

oh. sorry. the gcc47v2.diff: http://paste.kde.org/484214

sl1pkn07 commented on 2012-05-22 12:18

i have problem with new build

the nacl support has change. i make update pkgbuild for solvent this

now use http://commondatastorage.googleapis.com/nativeclient-mirror/nacl/nacl_sdk/staging/naclsdk_linux.tgz instead nacl_sdk.zip method (update launcher and download sources from google)

the old method make:

../../../../native_client/toolchain/linux_x86/x86_64-nacl/bin/gcc -c invoke.c -o /home/sl1pkn07/aplicaciones/chromium-dev/src/chromium-build/out/Release/obj.target/srpc_lib/geni/glibc-x86-64/srpc_lib/invoke.o -std=gnu99 -m64 -O3 -fomit-frame-pointer -mtls-use-call -D__linux__ -D__STDC_LIMIT_MACROS=1 -D__STDC_FORMAT_MACROS=1 -D_GNU_SOURCE=1 -D_BSD_SOURCE=1 -D_POSIX_C_SOURCE=199506 -D_XOPEN_SOURCE=600 -DDYNAMIC_ANNOTATIONS_ENABLED=1 -DDYNAMIC_ANNOTATIONS_PREFIX=NACL_ -DNACL_BUILD_ARCH=x86 -DNACL_BUILD_SUBARCH=64 -I/home/sl1pkn07/aplicaciones/chromium-dev/src/chromium-build/out/Release/obj/gen/tc_glibc/include -I../../../.. -I../../../../ppapi -I../../../.. -I../../../../ppapi
FAILED: [Errno 2] No such file or directory

../../../../native_client/toolchain/linux_x86/x86_64-nacl/bin/gcc don't provide by nacl_sdk.zip method. the "new" sdk solvent this

but have other problem:

CXX(target) out/Release/obj.target/ipc/ipc/ipc_sync_channel.o
ipc/ipc_platform_file.cc: En la función ‘IPC::PlatformFileForTransit IPC::GetFileHandleForProcess(base::PlatformFile, base::ProcessHandle, bool)’:
ipc/ipc_platform_file.cc:39:43: error: ‘::dup’ no se ha declarado (don't declare)
CXX(target) out/Release/obj.target/ipc/ipc/ipc_sync_message.o
make: *** [out/Release/obj.target/ipc/ipc/ipc_platform_file.o] Error 1

is a bug? :/

need little help

this is the new pkgbuild http://paste.kde.org/484208

please test it

greetings


misc commented on 2012-05-21 16:33

Still need to revert that commit with 1045.0.

@softwayer: Can happen, but should not (tarballs are indicatory: 1084.46 184.0M vs. 1132.11 185.4M).

Softwayer commented on 2012-05-18 13:40

So the version of latest stable build is bigger than of the unstable one?

sl1pkn07 commented on 2012-05-18 13:30

in dev-channel? no

http://googlechromereleases.blogspot.com.es/2012/05/dev-channel-update_17.html

sl1pkn07 commented on 2012-05-18 13:25

@Softwayer

out of date?? why???

Softwayer commented on 2012-05-18 13:23

The latest version is 21.0.1141.0, isn't it?

sl1pkn07 commented on 2012-05-18 10:06

oh my bad!

oks. is for active clang option instead GCC

sorry. change now

FredBezies commented on 2012-05-18 09:26

A little better, but got a building problem later :

chrome/browser/policy/policy_path_parser_linux.cc:35:36: error: use of
undeclared identifier 'geteuid'
struct passwd* user = getpwuid(geteuid());
^
chrome/browser/policy/policy_path_parser_linux.cc:45:9: error: use of undeclared
identifier 'gethostname'
if (gethostname(machinename, 255) == 0) {
^
2 errors generated.
make: *** [out/Release/obj.target/browser/chrome/browser/policy/policy_path_parser_linux.o] Erreur 1

FredBezies commented on 2012-05-18 07:59

Didn't think about this problem. Bad letters :]

sl1pkn07 commented on 2012-05-18 02:46

@fredbezies

Téléchargements <-

try in other place (without accents)

misc commented on 2012-05-17 20:02

≥1140: "\! -path 'third_party/zlib/contrib/minizip/*' \" → "\! -path 'third_party/zlib/*' \"

FredBezies commented on 2012-05-16 17:11

Crash on building :

==> Patch, update and copy NaCl SDK
~/Téléchargements/chromium-dev/src/nacl_sdk ~/Téléchargements/chromium-dev/src/chromium-build
Traceback (most recent call last):
File "/home/fred/Téléchargements/chromium-dev/src/nacl_sdk/sdk_tools/sdk_update.py", line 1060, in <module>
sys.exit(main(sys.argv[1:]))
File "/home/fred/Téléchargements/chromium-dev/src/nacl_sdk/sdk_tools/sdk_update.py", line 1053, in main
InvokeCommand(args)
File "/home/fred/Téléchargements/chromium-dev/src/nacl_sdk/sdk_tools/sdk_update.py", line 1042, in InvokeCommand
command(options, args[1:])
File "/home/fred/Téléchargements/chromium-dev/src/nacl_sdk/sdk_tools/sdk_update.py", line 899, in Update
bundle_path = os.path.join(options.sdk_root_dir, bundle_name)
File "/usr/lib/python2.7/posixpath.py", line 71, in join
path += '/' + b
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 12: ordinal not in range(128)
==> ERREUR : Une erreur s'est produite dans build().
Abandon...

misc commented on 2012-05-16 14:18

≥1138:

4. "\! -path 'third_party/libxml/*' \" to find&delete or compilation fails.

misc commented on 2012-05-16 11:59

Applies to 1133 (or so) til at least 1137.5:

1. Patch fragment for base/message_pump_libevent.cc needs to be removed as include is no longer missing.

2. This commit must be reverted (or perhaps a new include is missing) for compilation to not fail: http://codereview.chromium.org/10378057

3. Like Chromium in Extra, "CFLAGS+=' -Wno-c++0x-compat'" has to be added or else GCC 4.7 prints a couple of thousand useless warnings.

sl1pkn07 commented on 2012-05-12 18:43

sl1pkn07@sL1pkn07 ~ $ locate localefwd.h
/home/sl1pkn07/aplicaciones/chromium-dev/src/nacl_sdk/pepper_20/toolchain/linux_x86_glibc/x86_64-nacl/include/c++/4.4.3/bits/localefwd.h
/home/sl1pkn07/aplicaciones/chromium-dev/src/nacl_sdk/pepper_20/toolchain/linux_x86_newlib/x86_64-nacl/include/c++/4.4.3/bits/localefwd.h
/usr/include/c++/4.7.0/bits/localefwd.h
sl1pkn07@sL1pkn07 ~ $ pkgfile localefwd.h
core/gcc
community/arm-wince-cegcc-gcc
community/mingw32-gcc
multilib/gcc-multilib
sl1pkn07@sL1pkn07 ~ $ yaourt -Qo /usr/include/c++/4.7.0/bits/localefwd.h
/usr/include/c++/4.7.0/bits/localefwd.h es propiedad de gcc-multilib 4.7.0-6
sl1pkn07@sL1pkn07 ~ $

in this case. localefwd.h is provided by pepper_20 download by chromium-dev (if nacl is active)

regarding .pacnew file..

in the old packages, the variable CHROMIUM_FLAGS=" " is empty . in new default only remove the space between " ". only have hard modifications the chromium-dev.install file

i'm tested this in last night with all possibility: uninstall all, fresh install with/out pepperfrash, update pachage with/out pepperflash when initial fresh install have pepperflash and not, edit by hand the version of pepperflah to view update version take effect, etc. all test sucessfull (include if have other flags before/afther pepperflash flags)


of course. nothig is perfect :)

gretings

SirWuffleton commented on 2012-05-12 17:37

20.0.1132.3-1's compile succeeded for me, but I noticed that the .pacnew file for /etc/default didn't work. When I removed the flags relating to pepper-flash (the new default is blank), it was not detected by chromium.

Regarding the compile issue: I do not have this file either, but it compiled just fine on my box.

imrehg commented on 2012-05-12 16:17

compilaton of 20.0.1132.3-1 fails with:
In file included from ./remoting/protocol/authentication_method.h:16:0,
from remoting/protocol/authentication_method.cc:5:
/usr/lib/gcc/x86_64-unknown-linux-gnu/4.7.0/../../../../include/c++/4.7.0/string:45:69: fatal error: /usr/include/cairo/bits/localefwd.h: Permission denied
compilation terminated.
make: *** [out/Release/obj.target/remoting_protocol/remoting/protocol/authentication_method.o] Error 1

Actually, the file /usr/include/cairo/bits/localefwd.h does not exists in my system, is that a missing dependency? What would it be?

sl1pkn07 commented on 2012-05-09 16:28

with new compilation (rev2) fails for me :S

sl1pkn07 commented on 2012-05-09 15:11

oh! i see. yes i'm use the raw patch (http://chromium.googlecode.com/issues/attachment?aid=1225250019000&name=sqlite-3.7.6.3-fix-out-of-scope-memory-reference.patch&token=jqQlLFZFmSnbOxt60rCB6H5mAIo%3A1336575820598) oks. update pkgbuild to v2

sorry the inconvenient :S

SirWuffleton commented on 2012-05-09 14:41

Looks like you're using the raw patch (which came directly from sqlite) you need to patch third_party/sqlite/amalgamation/sqlite3.c in order for it to be included in the build. It was mentioned on the bug report comments and the new chromium PKGBUILD does this. The edited patch I posted earlier is pointed at that file (or you could just do what the PKGBUILD in extra is doing). Things have been stable for me with this patch included.

sl1pkn07 commented on 2012-05-09 13:03

i think is my faul to apply sqlite patch. i'm test new pkgbuild with chages like chromium package from extra

please patient

sl1pkn07 commented on 2012-05-09 12:46

ag! still fails :S

SirWuffleton commented on 2012-05-09 02:01

Just a heads up, a patch that fixes the segfaults was posted on the bug report page ( http://crbug.com/122525 ).
I edited the patch since a later comment indicated that the patch needed to be used on a different file in order to be included in the build. You can get that patch here: http://dl.dropbox.com/u/6464618/sqlite-3.7.6.3-fix-out-of-scope-memory-reference.patch
Then just add this to the PKGBUILD and you're good to go: patch --silent -Np0 -i "${srcdir}"/sqlite-3.7.6.3-fix-out-of-scope-memory-reference.patch

Also, 20.0.1130.1 is out according to the RPM list.

cornholio commented on 2012-05-07 15:23

That's strange that I'm the only one having the problem. After searching /usr/lib for *expat* before and after installing pyxml, it looks like:
/usr/lib/python2.7/lib-dynload/pyexpat.so
is superseded by:
/usr/lib/python2.7/site-packages/_xmlplus/parsers/pyexpat.so

If I rename the new pyexpat.so, the same error returns. I have no idea how this arose, but reinstalling all of the relevant packages didn't change anything. Its not a big problem since its working now, but it is strange.

jeagoss commented on 2012-05-07 08:46

I definitely don't have pyxml installed, and chromium-dev compiles for me every time. It seems there might have been some kind of install problem that the installation of pyxml might have repaired.

SirWuffleton commented on 2012-05-07 08:22

I can confirm this. I have been able to to build chromium-dev successfully and have never needed pyxml. I'm running x86_64 if it makes a difference.

My output of the stuff sl1pkn07 ran:
[dark@Night ~]% python2 -c "import xml.parsers.expat"
[dark@Night ~]% packer -Qs xml
core/expat 2.1.0-1
An XML parser library
extra/libplist 1.8-2
A library to handle Apple Property List format whereas it's binary or XML
extra/libxml++ 2.34.2-1
C++ bindings to libxml2
extra/libxml2 2.7.8-2
XML parsing library, version 2
extra/libxslt 1.1.26-3
XML stylesheet transformation library
extra/perl-xml-parser 2.41-2
Expat-based XML parser module for perl
extra/perl-xml-simple 2.18-4
Simple XML parser for perl
extra/raptor 2.0.7-2
A C library that parses RDF/XML/N-Triples into RDF triples
extra/xcb-proto 1.7.1-1
XML-XCB protocol descriptions

sl1pkn07 commented on 2012-05-07 07:02

i don't have install pyxml. and build chromium-dev sucessfull in all version


sl1pkn07@sL1pkn07 ~ $ python2 -c "import xml.parsers.expat"
sl1pkn07@sL1pkn07 ~ $

my yaourt xml output http://paste.kde.org/470684

anyone can confirm this?

cornholio commented on 2012-05-07 00:02

@sl1pkn07

build.gyp would fail with a python error regarding libexpat.so (from expat.py)

Without pyxml installed, attempting to load expat would produce an error (even though expat was installed):

[name@host ~]$ python2 -c "import xml.parsers.expat"
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/usr/lib/python2.7/xml/parsers/expat.py", line 4, in <module>
from pyexpat import *
ImportError: /usr/lib/python2.7/lib-dynload/pyexpat.so: undefined symbol: XML_SetHashSalt

After installing pyxml, this command no longer produced an error and chromium-dev would build successfully.
My machine configuration isn't very exotic so it would be strange if I am the only one with this dependency.

sl1pkn07 commented on 2012-05-06 12:40

@cornholio

are you sure?

sl1pkn07@sL1pkn07 ~ $ yaourt pyxml
1 extra/pyxml 0.8.4-9
Python XML parsing library

misc commented on 2012-05-06 09:28

Haven't had any crashes anymore again with 1027. Might be worth updating to that (or 1028 that's out in one/two hours) for those who have issues with 1123.4.

cornholio commented on 2012-05-06 01:01

pyxml is a make dependency

misc commented on 2012-05-05 10:19

I'm getting some seemingly random crashes since 1024 or so, too. Doubt they're related to ICU though, for a few builds after the patch Chromium was entirely stable again.

sl1pkn07 commented on 2012-05-04 16:30

@imhumn

fixed. thanks for report!

SirWuffleton commented on 2012-05-04 14:44

Looks like that icu patch didn't solve the crashes after all. While my last build seemed to crash maybe once a day tops, the new version seems to be crashing just as much as before:
May 03 18:25:31 Night kernel: Chrome_HistoryT[11640] general protection ip:7f6a891cbaa7 sp:7f6a6eff2090 error:0 in chromium[7f6a87bae000+3b05000]
May 03 18:29:03 Night kernel: Chrome_HistoryT[15473] general protection ip:7fe03147baa7 sp:7fe0172a2090 error:0 in chromium[7fe02fe5e000+3b05000]
May 03 18:29:53 Night kernel: Chrome_HistoryT[16084] general protection ip:7fcd88f12aa7 sp:7fcd6ed39290 error:0 in chromium[7fcd878f5000+3b05000]

Anonymous comment on 2012-05-04 14:28

The build fails because one of the build steps calls the svnversion command, which provided by the subversion package. Please update buildtime dependencies accordingly!

sl1pkn07 commented on 2012-05-03 22:09

er..... ops

Anonymous comment on 2012-05-03 21:20

The PKGBUILD uploaded now has use_system_icu=1 but the "\! -path 'third_party/icu/*' \" is not removed. Did you manually do that before you built with the patched ICU? Or is it still broken despite?

sl1pkn07 commented on 2012-05-03 21:13

OPS! :S

for me still broken :S

sl1pkn07 commented on 2012-05-03 20:55

@darksnipper

SUCESS! don't crash!

thanks man!


update to 20.0.1123.2 and change pepper_19 to pepper_20

please. use ICU build by @darksniper (can you upload to AUR?)

greetings

SirWuffleton commented on 2012-05-03 19:27

@sl1pkn07: Yeah, I just downloaded icu from the ABS and put the patch above ./configure like misc said (it needed -Np4 not -Np1 though). Here's the makepkg source archive I used for icu if you need it: http://dl.dropbox.com/u/6464618/icu-49.1.1-1.src.tar.gz

For the chromium-dev pkgbuild, I set use_system_icu=1, added icu to the depends array and removed the line "\! -path 'third_party/icu/*' \" and it built just fine.

sl1pkn07 commented on 2012-05-03 19:15

hi. the patch provide by @misc not work in 20.0.1123.2 (don't apply patch) without system ICU (use_system_icu=0 and withouth \! -path 'third_party/icu/*' \)

i test download icu 4.9.1.1 from sources (http://download.icu-project.org/files/icu4c/49.1.1/icu4c-49_1_1-src.tgz) and put on third_party folder (with patch by @misc)...

don't build: http://paste.kde.org/468140

howto get fix @darksnipper?

download extra ICU from ABS (or pkgbuild) and build with patch? (and build chromium-dev with use_system_icu=1)


greetings:


sorry the delay, yesterday my pc exploded

sl1pkn07 commented on 2012-05-03 19:09

hi. the patch provide by @misc not work in 20.0.1123.2 (don't apply patch) without system ICU (use_system_icu=0 and withouth \! -path 'third_party/icu/*' \)

i test download icu 4.9.1.1 from sources (http://download.icu-project.org/files/icu4c/49.1.1/icu4c-49_1_1-src.tgz) and put on third_party folder (with patch by @misc)...

don't build: http://paste.kde.org/468140

howto get fix @darksnipper?

download extra ICU from ABS (or pkgbuild) and build with patch? (and build chromium-dev with use_system_icu=1)


greetings:


sorry the delay, yesterday have a personal "Jericho" with my PC

SirWuffleton commented on 2012-05-01 20:12

I can confirm that the icu patch fixes the crashes with icu 49.1.1 and chromium 20.0.1120.0. Even after heavily abusing the omnibox for a good 5 minutes, I could not get it to crash :D

SirWuffleton commented on 2012-05-01 20:11

I can confirm that that icu patch fixes the crashes with icu 49.1.1 and chromium 20.0.1120.0. Even after heavily abusing the omnibox for a good 5 minutes, I could not get it to crash :D

SirWuffleton commented on 2012-05-01 19:17

Yep, that's definitely the problem. My ldd output shows no links to libicu since I built chromium using a modified version of the current PKGBUILD. I've made the appropriate changes to the PKGBUILD and am recompiling chromium now. I'll test it and post back with my results when it's done :)

misc commented on 2012-05-01 19:09

Forgot that patch files can be used to revert them, too ("-R" instead of "-N"). :)

So I just tested 50-svn with said patch reverted, and indeed it crashed again.

Which also means that if 50-svn or a patched 49.1.1 still crashes for you, it's a different bug — in which case you should of course submit your own ticket.

misc commented on 2012-05-01 18:58

> Do you think that a recompile of chromium against the new ICU would help?

It should not; if Chromium links dynamically to ICU (check with eg "ldd /usr/lib/chromium-dev/chromium | grep libicu" — I get three links to /usr/lib/libicu[…].so.50, whereas you'd get *.49, of course) all that patch ought to have changed would be some of those files in /usr/lib.

> would this prevent the patched ICU from being used by chromium?

Indeed it does. I'm afraid that if you have not changed back "use_system_icu=0" to 1 and removed that above "\! -path 'third_party/icu/*' \" line, your Chromium compiles most likely linked to the ICU version shipped with the Chromium tarball (some past version, 4.5 or so I think). Just use ldd again to check — those three links would probably not show at all in that case, or perhaps be a self-reference to chromium itself.

SirWuffleton commented on 2012-05-01 17:33

@misc: That patch applied cleanly to 49.1.1's source and the compile was successful, however, I can still replicate the crash in the 20.0.1120.0 package I built earlier. Do you think that a recompile of chromium against the new ICU would help? (Also, I noticed that the PKGBUILD configures chromium to not use the system's ICU, would this prevent the patched ICU from being used by chromium?)

misc commented on 2012-05-01 16:48

pedberg suggests that the following fix is what solved the crashes for me with ICU 50-svn, and could readily be applied to 49.1.1 as well: http://pastebin.com/raw.php?i=FyTd1ea4

Just add it as eg "patch -Np1 -i <path>" before ./configure in build() of ICU's PKGBUILD. If Chromium has been compiled with 49.1.1 then there should be no need to recompile it — or the other packages — to test (just close the programs or better restart, of course).

SirWuffleton commented on 2012-04-30 08:02

Looks like it was just stubborn with its crashes before. I'm experiencing the exact same issues as before with the normal build.

SirWuffleton commented on 2012-04-30 03:32

Derp. I just realized that leaving Ccache enabled while bug testing is a terrible idea.
I built icu-49.1.1-1 and chromium-dev 20.0.1120.0 with debug symbols and ccache disabled with the intent of getting valgrind/gdb traces, however I can't reproduce the crash anymore.
I was unable to get GDB to start chromium, but the crash I was able to cause through valgrind doesn't look related. (I was using a script to spam the omnibox rather than just typing random stuff into it as I did before. Here's my trace if anyone else can make sense of it: http://pastebin.com/0REr7xRX )

I'm going to try clearing CCache and compiling the normal version of chromium-dev to confirm that it doesn't exhibit the bug under normal use.

misc commented on 2012-04-29 17:04

The crashes of ICU 49.1.1 are indeed gone for me with ICU 50-svn (C20.0.1122.0).

Note that if you install the latter you'll also have to recompile all packages that depend on it for those (and those that in turn depend on them, etc.) to work again.

Again, additional valgrind & gdb traces will help to locate and properly fix the bug(s?).

misc commented on 2012-04-29 17:01

The crashes of ICU 49.1.1 are indeed gone for me with ICU 50-svn (C20.0.1122.0).

misc commented on 2012-04-29 13:20

@DarkSniper: Thanks for testing. Maybe it's an ICU 49 bug then, after all — I'll test it again next time.

@all: There's a ticket for it on ICU's bugtracker; the dev has asked for valgrind traces with ICU compiled for debug (see eg the Arch Wiki page "Debug - Getting Traces" for instructions):
http://bugs.icu-project.org/trac/ticket/9276

SirWuffleton commented on 2012-04-29 04:39

@misc: I've compiled version 20.0.1120.0 against those package versions and it still experiences the same segfault as before on my end.

misc commented on 2012-04-28 14:43

Note for my previous comment: Preferably with extra's icu-49.1.1-1, gcc 4.7.0-5 and chromium-dev 20.0.1120.0.

Rationale: Given that it works now for me without any apparently relevant code change on ICU's side, I'm trying to figure out whether this was actually a bug actually in eg a prior GCC 4.7 snapshot or Chromium itself.

misc commented on 2012-04-28 13:46

Not getting anymore crashes with icu (svn). Can someone confirm, please?

sl1pkn07 commented on 2012-04-26 17:43

ow. sorry :S

very thanks

SirWuffleton commented on 2012-04-26 17:12

Just a note on the force flags sl1pkno7 mentioned: I'd recommend that you use only -Sf/-Uf if recompiling the package is that inconvenient for you. It will override the file ownership check that would prevent you from upgrading otherwise, though using -dd (skipping the dependency checks) is dangerous and unnecessary in this situation.

sl1pkn07 commented on 2012-04-26 10:30

oks. now fixed in pkgbuild

upload missing resources (not need rebuild again the package)

http://dl.dropbox.com/u/6596386/theme_resources_standard.pak
http://dl.dropbox.com/u/6596386/ui_resources_standard.pak

put in /usr/lib/chromium-dev

in the next update. force instalation by -Sfdd (or rebuild now with the changes)

greetings

misc commented on 2012-04-26 10:06

Hurr. I should have looked at the PKGBUILD first: Line 306 just doesn't copy those two new files. Once amended to

install -Dm644 out/Release/{chrome,resources,content_resources,ui_resources_standard,theme_resources_standard}.pak "${_chromium_home}"/

all's well again. (Except for that unrelated memory corruption in "third_party/tcmalloc/chromium/src/free_list.cc:117" on exit.)

misc commented on 2012-04-25 22:25

Made a post about it on Google's bugtracker. The more persons who star it, the more attention it gets:

https://code.google.com/p/chromium/issues/detail?id=125060

misc commented on 2012-04-25 20:27

Faulty themes started with 1115.0; there are a couple of commits concerned with them between it and 1114.0. Upon start (and somewhere before finishing initialization) Chromium now also complains about not finding some resource files in /usr/lib/chromium-dev.

sl1pkn07 commented on 2012-04-25 17:23

is bug. i think is related by our friend gcc 4.7

http://sl1pkn07.no-ip.com/looooooooooooooooooooooooonnnnnnnnnnnnnnngggggggggggglog.txt

yetAnotherZero commented on 2012-04-25 16:48

@imhumn I got the same problem.

Anonymous comment on 2012-04-25 16:15

Giant angry red bar at the top of the screen, lack of icons, and this error message:
(chromium:1528): GdkPixbuf-WARNING **: GdkPixbufLoader finalized without calling gdk_pixbuf_loader_close() - this is not allowed. You must explicitly end the data stream to the loader before dropping the last reference.

Feature? or bug? hahaha... Rebuilding now to see if this was just a bad build on my system, or if it's Chromium's fault. May wanna hold off on upgrading in the meantime.

yetAnotherZero commented on 2012-04-21 20:06

@DarkSniper Thanks for that. I recompiled it and it's awesome.
I just did a pacman -Rdd gcc gcc-libs libtool lib32-gcc-libs then pacman -U on all those pkg versions you mentioned below. After the compile finished I did a pacman -Syu to get my system back up to par

Thanks again

SirWuffleton commented on 2012-04-20 18:12

@speed: I compiled it in a chroot using GCC4.6 with the Clang option and the crashes are gone for me. People reported that it doesn't crash if compiled under GCC 4.6, but downgrading GCC on your working install could potentially cause catastrophic breakage. A chroot seems to be the best (and only) existing workaround at this point.

I followed this guide for the chroot creation: https://wiki.archlinux.org/index.php/DeveloperWiki:Building_in_a_Clean_Chroot (Beware: If you're using BTRFS, it makes a subvolume for the chroot, so you have to use btrfs sub delete if you wish to remove the chroot, as rm -rf won't work)

Here's my package list for the chroot, which has everything you need to compile chromium (replace the base base-devel sudo part of the mkarchroot command with this): http://pastebin.com/vj6NL75r
Get version 4.6.3-1 of core/gcc, core/gcc-libs, and multilib/lib32-gcc-libs (x86_64 only) and version 2.4.2-4 of core/libtool from http://schlunix.org/archlinux/ and install them into your chroot with "pacman -r /path/to/your/chroot/root -U <paths to downloaded packages here>"

Before running mkchrootpkg, you'll probably need this edited PKGBUILD (It's up to you whether to use clang or not): http://pastebin.com/D7W50u6R
The '"${srcdir}"/../' bits in package() don't work with mkchrootpkg because it doesn't store the source files one level above $srcdir. I've changed instances of this to '"${srcdir}"/', since install seems to treat the symlinks as normal files. I have not tested this yet (I copied the source files to $srcdir/../ manually), but it should work as intended.

Hope this helps at least a few other users until the chromium team fixes GCC 4.7 support!

yetAnotherZero commented on 2012-04-20 17:21

Anyone have feedback regarding the omnibar crashes? It's a little sad everytime I forget and try to use it...

SirWuffleton commented on 2012-04-12 08:53

Using those options didn't seem to fix the segfault. After putting the omnibox through some abuse, it still crashes. I'm getting something slightly different on the command line now, though the dmesg output seems to be very similar if not identical to the previous segfault.

From Command Line:
[2695:2695:120527473823:ERROR:content_settings_pref_provider.cc(493)] Invalid pattern strings: ,*
[2695:2695:120527475053:ERROR:content_settings_pref_provider.cc(363)] Invalid pattern strings: ,*
[2695:2730:120535207943:ERROR:diagnostic_error_delegate.h(31)] sqlite error 11, errno 0: database disk image is malformed (This one doesn't always occur)
Segmentation fault (core dumped)

From dmesg:
[120553.729454] Chrome_HistoryT[2730] general protection ip:7f1df11d6247 sp:7f1dd88722a0 error:0 in chromium[7f1defabb000+4524000]

sl1pkn07 commented on 2012-04-12 06:29

in my case. with ICU 4.9 from testing (with use_system_icu=1)

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff3f16354 in icu_49::RegexMatcher::reset(icu_49::UnicodeString const&) () from /usr/lib/libicui18n.so.49

sl1pkn07 commented on 2012-04-12 06:28

in my case. with ICU 4.9 from testing

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff3f16354 in icu_49::RegexMatcher::reset(icu_49::UnicodeString const&) () from /usr/lib/libicui18n.so.49

SirWuffleton commented on 2012-04-12 05:54

@ sl1pkn07: Awesome! Compiling now to see if it fixes the issue.

sl1pkn07 commented on 2012-04-12 04:52

i think found problem described by @ DarkSniper (segfaults in omnibox)

please test with:

use_system_icu=0 and add "\! -path 'third_party/icu/*' \ in" in remove unnecesary components

greetings

jeagoss commented on 2012-04-12 02:28

The patches still need testing to make sure I got everything. So please test away.

sl1pkn07 commented on 2012-04-12 02:13

@jeagoss

http://i.imgur.com/uIMfV.png

jeagoss commented on 2012-04-12 01:23

I have a working PKGBUILD and patches to make chromium compile with pulseaudio. I also commented on your bug thread with the patch.

You can download the patches and PKGBUILD from here.

http://archrepo.jeago.com/sources/chromium-dev/

Anonymous comment on 2012-04-11 14:52

Thanks, as always for the updates! One suggested change to the PKGBUILD: if building with clang and you use plugins, both from and not from the Web Store, clang_use_chrome_plugins=1 will also have to be added to the GYP_DEFINES variable. Was using the clang build earlier, and this was necessary to see my plugins in-browser. AFAIK, this shouldn't cause any problems if you don't use plugins. See [https://code.google.com/p/chromium/wiki/Clang] for more informations.

sl1pkn07 commented on 2012-04-11 00:33

update to 20.0.1096.1

now back 64bit pepperflash!!!
now use pkgbuild sugested by @nagisa
now pepperflash 32 and 64 bits use same paths /usr/lib/chromium-dev/Pepperflash/libpepflashplayer.so

notes for @nagisa

- In /etc/makepkg.conf

#-- Make Flags: change this for DistCC/SMP systems
MAKEFLAGS="-j8"

no "need" fast build option. only uncomment and edit makepkg.conf (accoirding you cores) to enable multicore compilation

and ${MAKEFLAGS} dont work. use $MAKEFLAGS instead

- The disable pulseaudio is for http://crbug.com/119903. i dont know the fail build is related by http://crbug.com/111392 patch. (still fail build with/out patches)

- All ffmpeg system libs/ffmpegsumo build is according by http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/www-client/chromium/ (make by phajdan.jr , Chromium developer)
i comment all build with ffmpeg system libs for use in the future. active now don't have changes in perfomance.

- all no common variables warned by namcap. change to: "_<variable>"

any other suggestion are welcome

greetings

and PD: really sorry my english :S

sl1pkn07 commented on 2012-04-11 00:30

update to 20.0.1096.1

now back 64bit pepperflash!!!
now use pkgbuild sugested by @nagisa

notes for @nagisa

- In /etc/makepkg.conf

#-- Make Flags: change this for DistCC/SMP systems
MAKEFLAGS="-j8"

no "need" fast build option. only uncomment and edit makepkg.conf (accoirding you cores) to enable multicore compilation

and ${MAKEFLAGS} dont work. use $MAKEFLAGS instead

- The disable pulseaudio is for http://crbug.com/119903. i dont know the fail build is related by http://crbug.com/111392 patch. (still fail build with/out patches)

- All ffmpeg system libs/ffmpegsumo build is according by http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/www-client/chromium/ (make by phajdan.jr , Chromium developer)
i comment all build with ffmpeg system libs for use in the future. active now don't have changes in perfomance.

- all no common variables warned by namcap. change to: "_<variable>"

any other suggestion are welcome

greetings

and PD: really sorry my english :S

nagisa commented on 2012-04-06 19:37

I made (didn't test it too much tough) new PKGBUILD based on your PKGBUILD. Maybe you want to take some parts from it or even take whole PKGBUILD.
https://gist.github.com/7df6bd01b6e5f5d8bc15

Anonymous comment on 2012-04-06 14:43

That switch didn't work for me when Pepper Flash first hit this PKGBUILD. I had to point Chromium directly to the file with another switch. It's possible that's a Windows switch, not sure.

misc commented on 2012-04-06 07:38

Well, the only change that commit actually requires (assuming, again, that it's really the only one responsible) is to start Chromium with --enable-bundled-ppapi-flash .

misc commented on 2012-04-06 07:23

Couldn't you just include a patch that reverts
http://codereview.chromium.org/9969131/diff/1/chrome_content_client.cc
?

Seems like that's the only change responsible for disabling it. Haven't tested it though since I don't use pepper yet. No idea why they did that (and only on 1084), either.

sl1pkn07 commented on 2012-04-06 01:23

http://dl.dropbox.com/u/6596386/google-chrome-unstable-19.0.1084.9-130290.x86_64.rpm

for older flash 64

sl1pkn07 commented on 2012-04-06 01:18

update to 19.0.1084.15

DONT HAVE PEPPERFLASH ON 64BITS BRRRRRRRRRRRRRRRRRR!!!!!!!!!!

òó


see http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/branches/1084/src&range=130290:130829&mode=html

jeagoss commented on 2012-04-05 22:25

It appears in the latest release that google has renamed the flash plugin for the 32bit version. So there is no longer a need to have a 32bit/64bit check. The flash plugin is named exactly the same and located in the same place as the 64bit version.

sl1pkn07 commented on 2012-04-05 20:53

yes, i have this issue. but i dont know is related :S

i think related by GCC 4.7 (with gcc 4.6.x don't appear) , but dont have idea fix it. in compialation don't show output error

SirWuffleton commented on 2012-04-05 20:26

This version has been segfaulting occasionally for me when typing something in the omnibox. Has anyone else been experiencing similar behavior? The output seems to indicate something related to history, so I tried clearing my history, but it didn't seem to have an effect, as it's still segfaulting.

Here's the relevant log output from my Systemd journal:
Apr 05 13:19:48 Night kernel: Chrome_HistoryT[7870] general protection ip:7fc64046e397 sp:7fc6269e42a0 error:0 in chromium[7fc63eea6000+39a5000]
Apr 05 13:19:48 Night[20497]: Process 7835 (Chrome_HistoryT) dumped core.

Anonymous comment on 2012-04-04 08:20

Tested again. Works fine. My fault, i guess. Sorry for inconvenience.

sl1pkn07 commented on 2012-03-29 14:38

testing. no fail for me (gcc 4.6.3 and disable NaCL)

http://dl.dropbox.com/u/6596386/log

greetings

sl1pkn07 commented on 2012-03-29 14:37

i test. no fail for me (gcc 4.6.3 and disable Nacl)

http://dl.dropbox.com/u/6596386/log

Anonymous comment on 2012-03-29 10:34

Latest build without NaCL fails with:

Updating projects from gyp files...
Traceback (most recent call last):
File "build/gyp_chromium", line 171, in <module>
sys.exit(gyp.main(args))
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/__init__.py", line 471, in main options.circular_check)
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/__init__.py", line 111, in Load depth, generator_input_info, check, circular_check)
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/input.py", line 2292, in Load depth, check)
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/input.py", line 433, in LoadTargetBuildFile includes, depth, check)
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/input.py", line 433, in LoadTargetBuildFile includes, depth, check)
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/input.py", line 433, in LoadTargetBuildFile includes, depth, check)
File "/tmp/yaourt-tmp-dm/aur-chromium-dev/src/chromium-build/tools/gyp/pylib/gyp/input.py", line 414, in LoadTargetBuildFile "Unable to find targets in build file %s" % build_file_path
Exception: Unable to find targets in build file ppapi/native_client/native_client.gyp while loading dependencies of chrome/chrome.gyp while loading dependencies of chrome/browser/sync/tools/sync_tools.gyp while loading dependencies of build/all.gyp while trying to load build/all.gyp

sl1pkn07 commented on 2012-03-27 11:55

oks. fix and update

misc commented on 2012-03-27 11:23

Looks like pastebin cut off the last two lines as they are "empty", which has patch complain about syntax (not an error though).

So what's missing after "class SkBitmap;": One line with one space, then another line that's entirely empty.

sl1pkn07 commented on 2012-03-27 11:08

@misc

(Stripping trailing CRs from patch.)
patching file ui/gfx/codec/jpeg_codec.h
patch unexpectedly ends in middle of line
Hunk #1 succeeded at 8 with fuzz 2 (offset 1 line).

sl1pkn07 commented on 2012-03-27 06:53

try now

SirWuffleton commented on 2012-03-27 01:42

nacl_sdk.zip seems to have a bad checksum. I'm repeatedly getting 6819b50db888bb4c3c215c3c96f32eff431b6024 but the PKGBUILD expects e4991439b4832a5a80134a452fb824992fc9818d. Changing the checksum doesn't seem to cause any problems.

sl1pkn07 commented on 2012-03-26 23:48

ains....

update

- disable (temporally, even fix build) pulseaudio support http://crbug.com/119903

- now use sha1sums instead md5sums. to work for script for download automatically official chrome RPM for libpdf/pepperflash plugins
(search official yum repository (http://dl.google.com/linux/chrome/rpm/stable/"${_arch}"/repodata/filelists.xml.gz) and download RPM according this)

- add support (not tested) to GCC 4.7.0 (testing/multilib repository) (patch provided by @misc). please test it. (only apply if detect GCC 4.7.0 has been installed)


please report more bugs, fix, patches, ideas, encouragement, spit, blablabla

greetings

Det commented on 2012-03-26 20:59

But so pulseaudio still needs to be disabled?

misc commented on 2012-03-25 15:07

Well that wasn't hard to fix. Here's my patch to have Chromium (tested with 1079.0 & 1080.0) compile with GCC 4.7 — just a bunch of missing includes (those adding cstddef are those that still apply from extra's chromium package to solve warnings for GCC 4.6; aren't necessary but don't hurt either):

http://pastebin.com/CQu4GwC9

misc commented on 2012-03-24 16:15

1078.0 built fine for me with pulseaudio disabled, with GCC 4.7 however I'm getting prototype and missing declaration errors (for 1079.0).

sl1pkn07 commented on 2012-03-24 02:56

fail build 19.0.1077.3

http://crbug.com/119903

Anonymous comment on 2012-03-20 15:55

Just wanted to put it out there that Arch's default CFLAGS appear to work just
fine, even when compiling Nacl, Pepper Flash, and the libpdf extras. There
should be no need to remove the stack-smashing flags at the moment, AFAIK.

There may be a few outstanding quirks regarding using the --enable-pulseaudio
switch for 64-bit users using PulseAudio and Pepper Flash together, but I'm
looking into it.

nagisa commented on 2012-03-16 22:14

Building without _system_ffmpeg_libs is OK, however with them (without patching, because it rejects) I get

out/Release/obj.target/ffmpeg/geni/ffmpeg_stubs.cc:19:10: fatal error: 'libavformat/url.h' file not found
#include <libavformat/url.h>

For now I will bear with builing ffmpeg_sumo, but it would be great to see system_ffmpeg_libs working again soon.

sl1pkn07 commented on 2012-03-16 18:41

@Nagisa

oks. now disable (with comment) all references to build/install system FFmpeg library.

nagisa commented on 2012-03-16 18:27

You should also add a note near "compile with system ffmpeg" option that compiling with system ffmpeg currently fails and you can remove patch_for_ffmpeg-git.patch as it doesn't work anyway.

nagisa commented on 2012-03-16 18:24

You should also add a note near that compiling with system ffmpeg currently fails and you can remove patch_for_ffmpeg-git.patch as it doesn't work anyway.

sl1pkn07 commented on 2012-03-16 18:23

:D

thanks for this report! (need more of this). updated and added lib32-gcc-libs makedepend (if nacl is acctived)

greetings

and little cosmetics

Anonymous comment on 2012-03-16 16:01

Great work on getting all of this PKGBUILD fun worked out! I've been until now unable to compile Nacl, even with safe CFLAGS (took out the stack-smashing protection flags manually). Was receiving fatal errors that some gcc library files were missing. lib32-gcc-libs needs to be a dependency for 64-bit users if building Nacl, it seems.

sl1pkn07 commented on 2012-03-16 15:23

do'h

fixed

jeagoss commented on 2012-03-16 12:56

On line 344 of the PKGBUILD, you need to change the first ${pkgdir} to ${srcdir}. This change allows for full compilation on i686.


OLD: install -m775 "${pkgdir}"/opt/google/chrome/libgcflashplayer.so "${pkgdir}"/usr/lib/chromium-dev/libgcflashplayer.so

NEW: install -m775 "${srcdir}"/opt/google/chrome/libgcflashplayer.so "${pkgdir}"/usr/lib/chromium-dev/libgcflashplayer.so

sl1pkn07 commented on 2012-03-16 12:19

@jeagoss

try now (update pkgbuild)

in 32bits, the PapperFlash library is named "libgcflashplayer.so". now install in /usr/lib/chromium-dev/libgcflashplayer.so (if detect 32bits) (like a RPM package)

@imhumn

edit /etc/chromium-dev/default and add "--ppapi-flash-path=/usr/lib/chromium-dev/PepperFlash/libpepflashplayer.so --ppapi-flash-version=11.2.31.114" in CHROMIUM_FLAGS (--enable-greasemonkey --enable-sync not need)

i think fixed now in chromium-dev.install script.


@all

new pkgbuild enable compile "chromiumdrive" (chromedrive). for new major updates include it for default

please test it

greetings

jeagoss commented on 2012-03-16 10:15

It appears that the i386 versions of chrome do not actually include the pepperflash directory opt/google/chrome/PepperFlash in the rpm archive. I also checked the i386 versions of the deb's that are available, and the same seems to be true. Due to i386 not including the plugin, it causes the build to fail.

I'm not sure if google is maintaining an i386 version of the plugin or not. I've been trying to figure this out. Until you can find a reliable source for the plugin, you might consider disabling the install on i686 architectures.

Anonymous comment on 2012-03-16 02:35

Nope. The CHROMIUM_FLAGS entry in /etc/chromium-dev/default is empty.

sl1pkn07 commented on 2012-03-16 01:28

#imhumn

please. in /etc/chromium-dev/default exist this line?:

CHROMIUM_FLAGS=" --ppapi-flash-path=/usr/lib/chromium-dev/PepperFlash/libpepflashplayer.so --ppapi-flash-version=11.2.31.114 --enable-greasemonkey --enable-sync"

greetings

Anonymous comment on 2012-03-16 01:07

@sl1pkn07: Thanks! Pepper Flash is working just fine using the --ppapi-flash-path="/path/to/plugin" command-line switch.

sl1pkn07 commented on 2012-03-16 00:43

#imhumm

fixed (but not affected. typo present in all scripts)

acfrazier commented on 2012-03-15 18:52

curl: (22) The requested URL returned error: 404
==> ERROR: Failure while downloading google-chrome-unstable-19.0.1068.0-126342.x86_64.rpm

I'm assuming this is due to the fact that we're now on 19.0.1068.1, so I'll wait for an update.

Anonymous comment on 2012-03-15 15:33

@sl1pkn07: Minor typo in your sed command around the end of the PKGBUILD, where the PKGBUILD-internal variable is written _use_papperflash instead of 'pepperflash'. Not sure if that's gonna cause issues in the installation of the Pepper plugin. I can't find it in the browser, despite the file existing, and the --enable-bundled-ppapi-flash switch being on. Does the Pepper Flash plugin require Nacl perhaps? I'd rebuild to test this, but I'm getting a 404 Not Found on the .rpm download after having successfully built your new PKGBUILD once. URL change?

sl1pkn07 commented on 2012-03-15 06:09

upadate to 19.0.1068.0

add new options:

- Add "_use_libpdf" option to install (enable by default) "libpdf.so" from Official Chrome-Browser unstable (RPM)
- Add "_use_pepperflash"option to install (enable by default) "PepperFlash" (libpepflashplayer.so) from Official Chrome-Browser unstable (RPM)
need new dependence: openssl098
- Add support custom flags in /etc/chromium-dev/default (like chromium PPA packages from Ubuntu .deb) (is need to install PepperFlash plugin)


please. test it

greetings

sl1pkn07 commented on 2012-03-11 18:19

fixed http://i.imgur.com/rn8Zh.png

Det commented on 2012-03-11 17:41

By the way, there's a couple of typos in the PKGBUILD. It says "Enviorement" instead of "Environment" in the "Automatic Build Options" and the Gnome/Pulseaudio (sub)sections.

sl1pkn07 commented on 2012-03-09 01:06

@nagisa, the ffmpeg patch is provide by @misc. please ask him

if you have working patch. please send me

greetings

nagisa commented on 2012-03-08 15:37

In addition to yet unfixed problem reported in 24th February there's another problem now which is related with patches.
There's no `third_party/ffmpeg/avcodec-53.sigs` anymore and the file is in `third_party/ffmpeg/chromium/avcodec-54.sigs` now. Also trying to patch this file simply fails with yet another reject.

As with previous case, failing patch is called `patch_for_ffmpeg-git.patch`

nagisa commented on 2012-02-24 15:53

By the way reject looks like this: https://gist.github.com/9db1661d4b1474dbb9dc

nagisa commented on 2012-02-24 15:37

I'm getting

patch unexpectedly ends in middle of line
patch unexpectedly ends in middle of line
1 out of 1 hunk FAILED -- saving rejects to file ui/base/l10n/l10n_util.cc.rej

when patching for ffmpeg-git.

Also, is there a way to get rid of that nasty lib32-zlib dependency and only use lib-zlib, when compiling with NaCl?

@sl1pkn07, I have been using this nickname way before I watched Clannad, so... not.
Thanks for update.

nagisa commented on 2012-02-24 15:34

Is there a way to get rid of that nasty lib32-zlib dependency and only use lib-zlib, when compiling with NaCl?

@sl1pkn07, I have been using this nickname way before I watched Clannad, so... not.
Thanks for update.

sl1pkn07 commented on 2012-02-24 02:55

update to 19.0.1049.3 (sorry for the delay :S)

- Add changes suggested by @Nagisa (Dango Daikazoku?)
- Add third_party/adobe/ in remove unnecesary components exception
- Delete chromium-revert-jpeg-swizzle-r2.patch. no need anymore (fixed by libjpeg-turbo 1.2.0)

greetings

sl1pkn07 commented on 2012-02-24 02:16

update to 19.0.1049.3 (sorry for the delay :S)

- Add changes suggestion by @Nagisa (Dango Daikazoku?)
- Add third_party/adobe/ in remove unnecesary components exception
- Delete chromium-revert-jpeg-swizzle-r2.patch. no need anymore (fixed by libjpeg-turbo 1.2.0)

greetings

nagisa commented on 2012-02-16 14:25

To better integrate with Gnome Shell and provided .desktop file you should change line
exec /usr/lib/chromium-dev/chromium "$@"
to
exec /usr/lib/chromium-dev/chromium --class=chromium-dev "$@"
in chromium-dev.sh

Here's new file https://gist.github.com/a6def8dfd86fc9ac4a6f and it's md5 is 48e11fa61360ec851c375e6deede11e2

sl1pkn07 commented on 2012-02-12 00:30

update to 19.0.1036.7

- remove some patches

- add GYP_DEFINES

linux_use_gold_binary=0
linux_use_gold_flags=0

greetings

misc commented on 2012-02-11 17:49

Finally — after "python2 build/gyp_chromium -f make --depth=. build/all.gyp":

sed -i 's/LINK.target ?= $(LINK)/LINK.target ?= $(LINK) -fuse-linker-plugin/g' Makefile
sed -i 's/LINK.host ?= g++/LINK.host ?= g++ -fuse-linker-plugin/g' Makefile

I'm currently compiling .1039; if it fails I'll add another comment.

misc commented on 2012-02-10 12:28

Gold accessible only as ld.gold, not vanilla ld, so one needs two plugin parameters for gcc. However, reading build/common.gypi makes it clear that the script searches for its own gold in third_party, so I excluded it with third_party/gold/* from removal — to no avail, linking still fails with unrecognized option. Adding "-B/usr/bin/ld.gold" to CFLAGS & CXXFLAGS before calling make didn't solve it, either.

misc commented on 2012-02-10 10:56

Damn, I really should wait a minute before I submit a comment — now I know why it fails:

Gold accessible only as ld.gold, not vanilla ld.

misc commented on 2012-02-10 10:48

@Miche: No idea what's included in that 10GB figure — perhaps all revisions of the source files that ever existed — but as for the tarballs, they only contain one. Thus the ball (~190MB) and its extracted contents (~900MB) take up ~ 1GB on my /tmp, which has up to 2GB allocated of my 4GB RAM. So far compilation has worked hassle-free — but only if no other memory intensive programs run.

@sl1pkn07: The glib2 patch appears superfluous as of 1036; as for the ffmpeg one, I doubt it's enough for the current git.

However, there's another issue now, at least on my machine: Compilation aborts with "/usr/bin/ld: unrecognized option '--icf=none'". Identical code folding is a feature added by (and afaik only available to) Google's gold linker, which ought to be accessible since Arch's binutils are compiled with the --enable-gold parameter — so no idea why it fails.

misc commented on 2012-02-10 10:47

@Miche: No idea what's included in that 10GB figure — perhaps all revisions of the source files that ever existed — but as for the tarballs, they only contain one. Thus the ball (~190MB) and its extracted contents (~900MB) take up ~ 1GB on my /tmp, which has up to 2GB allocated from my 4GB RAM. So far compilation has worked without hassle-free — but only if no other unnecessary, memory intensive programs run.

@sl1pkn07: The glib2 patch appears superfluous as of 1036; as for the ffmpeg one, I doubt it's enough for the current git.

However, there's another issue now, at least on my machine: Compilation aborts with "/usr/bin/ld: unrecognized option '--icf=none'". Identical code folding is a feature added by (and afaik only available to) Google's gold linker, which ought to be accessible since Arch's binutils are compiled with the --enable-gold parameter — so no idea why it fails.

Anonymous comment on 2012-02-07 11:06

Chromium wiki says "It takes about 10GB or so of disk space to check out and build the source tree. This number grows over time." I tried twice and the process end with a "no space left on the device" error. The second time i set the build dir in my home (yaourt -S chromium-dev --tmp ~/tmp) but didn't worked anyway.
Did you got this issue? Where do you build it?

acfrazier commented on 2012-02-04 02:57

Thank you, I appreciate your work on this package.

acfrazier commented on 2012-02-04 02:56

MD5 for chromium-18.0.1025.3.tar.bz2:
22e6b19ffdbc7b40da0fe65d6c38a539

NOTE: I don't use PulseAudio, but it built properly on my system.

sl1pkn07 commented on 2012-02-03 20:30

this morning release 1025.2 (yes, i sure)

now look new release

http://feedproxy.google.com/~r/GoogleChromeReleases/~3/kvVrvuQCnDQ/dev-channel-update.html

updating in progress

acfrazier commented on 2012-02-03 19:49

http://omahaproxy.appspot.com/viewer

Latest Linux dev release is 18.0.1025.3
I'm not sure if it builds with the current pkgbuild, testing now..

misc commented on 2012-01-30 22:04

1020< needs
\! -path 'third_party/libsrtp/*' \
in "Remove unnecesary components".

Btw, my patch is actually two in one, the last two changes are for glib2 >= 2.31 but should be downwards compatible.

sl1pkn07 commented on 2012-01-28 20:18

apply patch prvide in http://crbug.com/111392

now compile

upload and updane new version

add patch provided by @misc (only active if use system ffmpeg libs)

greetings

acfrazier commented on 2012-01-25 23:03

It built just fine on my machine, I don't know why you are seeing that error.

sl1pkn07 commented on 2012-01-25 22:30

nacl-g++ /home/sl1pkn07/aplicaciones/chromium-dev/src/chromium-build/out/Release/obj.target/ppruntime_lib/geni/newlib-x86-64/ppruntime_lib/plugin_ppb_audio.o
In file included from ./base/memory/weak_ptr.h:63:0,
from ./media/audio/pulse/pulse_output.h:25,
from media/audio/pulse/pulse_output.cc:5:
./base/logging.h: In function 'std::string* logging::CheckEQImpl(const t1&, const t2&, const char*) [with t1 = scoped_refptr<base::MessageLoopProxy>, t2 = MessageLoop*, std::string = std::basic_string<char>]':
media/audio/pulse/pulse_output.cc:129:3: instantiated from here
./base/logging.h:556:1: error: no match for 'operator==' in 'v1 == v2'
./base/logging.h:556:1: note: candidates are:
./base/logging.h:556:1: note: operator==(MessageLoop*, MessageLoop*) <built-in>
./base/logging.h:556:1: note: no known conversion for argument 1 from 'const scoped_refptr<base::MessageLoopProxy>' to 'MessageLoop*'
./base/logging.h:556:1: note: operator==(base::MessageLoopProxy*, base::MessageLoopProxy*) <built-in>
./base/logging.h:556:1: note: no known conversion for argument 2 from 'MessageLoop* const' to 'base::MessageLoopProxy*'
./base/memory/scoped_ptr.h:203:6: note: template<class C> bool operator==(C*, const scoped_ptr<C>&)
./base/memory/scoped_ptr.h:315:6: note: template<class C> bool operator==(C*, const scoped_array<C>&)
./base/memory/scoped_ptr.h:441:6: note: template<class C, class FP> bool operator==(C*, const scoped_ptr_malloc<C, FP>&)
CXX(target) out/Release/obj.target/media/media/base/byte_queue.o
./base/logging.h:556:1: warning: control reaches end of non-void function [-Wreturn-type]
make: *** [out/Release/obj.target/media/media/audio/pulse/pulse_output.o] Error 1

acfrazier commented on 2012-01-25 19:08

MD5 for chromium-18.0.1017.2.tar.bz2:
186593a3f943e3d9bdf89c7899f382a3

Builds fine only changing the package version/MD5, and leaving the rest of the things as-is.

misc commented on 2012-01-24 15:58

My bad, I misread an announcement.

jeagoss commented on 2012-01-24 08:25

Why is this flagged out of date? The latest official dev release is the version stated.

misc commented on 2012-01-22 10:09

Here's a little patch to make it compile with ffmpeg-git & glib2 >= 2.31:
http://pastebin.com/raw.php?i=ixheuye6

sl1pkn07 commented on 2012-01-19 12:43

oks. thanks :)

jeagoss commented on 2012-01-19 11:37

That looks to be the only dependency that wasn't there.

sl1pkn07 commented on 2012-01-19 11:05

@jeagoss

please. test package with namcap and post the results (for needed add other dependencies)

greetings

sl1pkn07 commented on 2012-01-19 11:01

roger. update pkgbuild

don't up pkgrel because is the first issue by this. i supposed all user have install dbus-glib package

thanks @jeagoss

greetings

jeagoss commented on 2012-01-19 10:42

dbus-glib should be added as a dependency. I was building from a clean system, and the build fails out in the gyp stage if it is missing.

sl1pkn07 commented on 2012-01-18 19:28

update to 18.0.1010.1

fixed:

http://crbug.com/103360
http://crbug.com/41887

add udev and icu to dependency

greetings

pfreire commented on 2012-01-15 20:41

Thanks Det, that was it. now i have another error, but I believe it's because of the CFLAGS i'm using just like in this problem http://code.google.com/p/chromium/issues/detail?id=104569. They advise to use safe CFLAGS.

FAILED with 1: ../native_client/toolchain/linux_x86_newlib/x86_64-nacl/bin/gcc -c src/untrusted/stubs/crti_x86_64.S -o /tmp/yaourt-tmp-root/aur-chromium-dev/src/chromium-build/out/Release/obj/gen/tc_newlib/lib64/crti.o -std=gnu99 -m64 -O3 -fomit-frame-pointer -mtls-use-call -DNACL_BUILD_ARCH=x86 -D__linux__ -D__STDC_LIMIT_MACROS=1 -D__STDC_FORMAT_MACROS=1 -D_GNU_SOURCE=1 -D_BSD_SOURCE=1 -D_POSIX_C_SOURCE=199506 -D_XOPEN_SOURCE=600 -DDYNAMIC_ANNOTATIONS_ENABLED=1 -DDYNAMIC_ANNOTATIONS_PREFIX=NACL_ -DNACL_BUILD_SUBARCH=64 -I/tmp/yaourt-tmp-root/aur-chromium-dev/src/chromium-build/out/Release/obj/gen/tc_newlib/include -I.. -I../ppapi -I.. -I../ppapi


make: *** [out/Release/obj/gen/tc_newlib/lib64/crti.o] Error 255

Det commented on 2012-01-15 15:30

It's in [multilib]. Dunno why the "Native Client" (NACL) needs it.

pfreire commented on 2012-01-15 15:27

Hello, I get this error (using yaourt):

==> Building and installing package
==> Install or build missing dependencies for chromium-dev:
error: target not found: lib32-zlib

Where is lib32-zlib, and why do I need it on a x64 system?

sl1pkn07 commented on 2012-01-12 16:18

http://crbug.com/106954 now fixed!

thanks phajdan.jr!

sl1pkn07 commented on 2012-01-11 20:18

compile chromium take several minutes, force compile all for a "little" revision (incrementing pkgrel number) is a "dirty trick" (sorry, my english not good)

only incrementing pkgrel number if a make several changes (not include update package)

imrehg commented on 2012-01-11 15:09

Just two more thoughts:
- chromium-revert-jpeg-swizzle-r0.patch seems to be completely commented out anyway, I guess shouldn't be included
- would be great if package fixes would be noted with incremented pkgrel number, it gets confusing that I downloaded the PKGBUILD 3 or 4 times today, all worked (or not-worked) differently but nominally they are all the "same".

Other then these things, thanks a lot for maintaining it! For a while I tried to keep track of all the changes that needed to be taken care of to get Chromium to build, but I had much much less patience with it. :)

sl1pkn07 commented on 2012-01-11 14:26

yesyes. fixed in pkgbuild

:S

imrehg commented on 2012-01-11 14:23

Had to except "third_party/libjpeg_turbo/" in this latest build, otherwise got compilation error.

sl1pkn07 commented on 2012-01-11 12:14

fuck!

http://crbug.com/106954 not fixed ><

back GYP_Define to "use_libjpeg=0"

sl1pkn07 commented on 2012-01-11 11:58

update to 18.0.1003.1

fixed:

http://crbug.com/104569 (no need patch anymore)
http://crbug.com/108228 (fixed in 18.) @darksniper
http://crbug.com/106954 (no need patch anymore)

back mv instead cp -R

add exception for "third_party/libyuv/" (thanks @imrehg)

remove Debug code

greetings

SirWuffleton commented on 2012-01-08 21:09

That was exactly the issue, built fine [without eating up loads of space] after changing that =D
However, it seems that this version still contains this bug, whereas the latest stable one fixed it: http://crbug.com/108228

sl1pkn07 commented on 2012-01-07 21:33

try this:

in build()

mv instead cp -R

only source code need 852MiB

SirWuffleton commented on 2012-01-07 16:15

Just wondering... Is it just me, or does building chromium now take an outrageous amount of disk space?
On earlier builds I was able to build the whole thing in my 2GB tmpfs, now I can't even extract the source without running out of space. For some odd reason, the source alone it's taking up about 700 megs more on /tmp than it does on my hard drive, and I'm pretty sure that BTRFS's compression isn't responsible for such a big difference in size. I bumped /tmp's size up to 2.5 gigs and it barely got to the end of the NACL SDK thing before running out of space again.
Any ideas on what might be causing this?

imrehg commented on 2012-01-02 11:22

For newer versions I had to add an exception for "third_party/libyuv/" in the "Remove unnecesary components" section, otherwise the build breaks. With this small change (and disabling nacl) I could successfully build 18.0.993.0

By the way, is that section really necessary? Can't the build system decide what it needs the build and what not? It took a while to track down while a build breaks with "missing header file", when it is originally there.

sl1pkn07 commented on 2011-12-28 14:50

install with yaourt. its more easy

yaourt -Sa nameofpackage

or with edit /etc/yaourtrc AURSEARCH=1 , install with simply yaourt -S nameofpackage

whist this install all necesary dependencies in all packages

example this:

yaourt -S chromium-dev

https://wiki.archlinux.org/index.php/Yaourt

darnir commented on 2011-12-28 11:27

Okay, my mistake. I had only made the package, not installed it. Still new to Arch, figuring my way out..
Once I ran makepkg -si, it installed libwebp, and no problem thence.
Though, shouldnt that be automated when I run makepkg -si for the chromium-dev package?

sl1pkn07 commented on 2011-12-28 10:43

are you sure install libwebp?. try install with yaourt (yaourt -Sa libwebp)

darnir commented on 2011-12-28 08:04

While installing, I ran makepkg -s, and got an error:
error: target not found: libwebp
=> Error: 'pacman' failed to install missing dependencies.

So, I followed the link to the libwebp package and manually installed it. Yet I keep getting this error and cannot install the chromium-dev package.

sl1pkn07 commented on 2011-12-16 19:56

@oraerk

in compile time. buid 64 and 32bits plugin in same time

Anonymous comment on 2011-12-16 14:51

According to Chromium docs NACL should be supported in 64bit mode. Why use lib32-zlib?

sl1pkn07 commented on 2011-12-14 12:50

New PKGBUILD Update

- Remove Build Options: _use_cups and _use_krb
now build by default without touch any config

- Now NACL suppot buid by default. disable with build option _use_nacl
this is for change system build NACL (see http://projects.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/chromium)

thanks to @Evangelos for guide me to the light XD

sorry for the any conveniences :S

sl1pkn07 commented on 2011-12-14 08:22

update to fix post below

and remember (now fix)

_use_nacl
_use_cups
_use_kerb

ALL set to 0, to enable set to 1 (to this edit PKGBUILD)

sl1pkn07 commented on 2011-12-14 08:14

edit /etc/makepkg.con and uncomment MAKEFLAGS entry

add -jX (X is a number of cored you use. the first core is 1 (example quad core = -j4, dual core = -j2, etc)

sl1pkn07 commented on 2011-12-14 08:09

edit /etc/makepkg.con and uncomment MAKEFLAGS entry

naelstrof commented on 2011-12-14 06:09

I get the following error while building chromium-dev:

http://j.mp/vrlCvv

sl1pkn07 commented on 2011-12-13 22:54

new PKGBUILD update

add in GYP_DEFINES

use_system_vpx : disable (0). prepare to use libvpx system libs (don't use system lib, waiting TODO)
use_system_hunspell : prepare to use hunspell system libs (upstream changes needed)
use_system_ssl : prepare to use openssl system libs (http://crbug.com/58087)

BUILD OPTIONS

_use_kerb (kerberos) : enable/disable support to build Kerberos environment


fix install NACL libs and components when _use_nacl=1
some cosmetic changes

greetings

sl1pkn07 commented on 2011-12-13 18:53

okok. update to 17.0.963.6

changes in pkgbuild:

add in GYP_DEFINES:

use_system_v8=0 : prepare to use V8 system libs (when update V8 extra package to +3.7.6)
use_system_sqlite=0 : prepare to use sqlite3 system libs (not working)
use_system_flac=1 : use system FLAC libs (and add FLAC dependency)
use_system_libwebp=1 : use system libwebp libs (and add libwebp dependency)
use_system_libjpeg=0 : back config (see http://code.google.com/p/chromium/issues/detail?id=106954)
ffmpeg_branding=Chrome


BUILD OPTIONS

_use_cups : enable support to build CUPS environment
_use_pulseaudio : Now is automatic when detect pulseaudio server installed
_use_nacl : Now WORKS when set to 1. thanks @jeagoss. for more info see http://code.google.com/p/chromium/issues/detail?id=104569

Back Don't use type 'settings' in gyp files: http://code.google.com/p/chromium/issues/detail?id=96629
Add Remove unnecesary components
Add More internal patches
Mod Force usage Python2

greetings

Det commented on 2011-12-13 07:47

Same thing: not yet mentioned in the blog but .6 is the latest Dev one.

sl1pkn07 commented on 2011-12-13 00:36

@big_bum

the latest DEV-CHANNEL is 17.0.963.2

http://googlechromereleases.blogspot.com/search/label/Dev%20updates

@jeagoss

O_O

i test with this 3 days ago and don't compile :S

http://code.google.com/p/chromium/issues/detail?id=104569#c19

that if, without CFLAGS edit and without "filter-sp.patch". only CXXFLAGS whitout -fastak-protector

don't test with CFLAGS, CXXFLAGS and patch (in test now)

waiting result...

EDIT: WORKS O_o

THANKS!!!!

sl1pkn07 commented on 2011-12-12 21:18

@big_bum

the latest DEV-CHANNEL is 17.0.963.2

http://googlechromereleases.blogspot.com/search/label/Dev%20updates

@jeagos

O_O

i test with this 3 days ago and dot compile :S

http://code.google.com/p/chromium/issues/detail?id=104569#c19

that if, without CFLAGS edit and without "filter-sp.patch". only CXXFLAGS whitout -fastak-protector

don't test with CFLAGS, CXXFLAGS and patch (in test now)

waiting result...

jeagoss commented on 2011-12-12 20:33

When building with NaCl, I found something that might work as a fix for the missing symbols error.

I had to add the following lines

export CXXFLAGS="`echo ${CXXFLAGS} | sed 's| -fstack-protector||'`"
export CFLAGS="`echo ${CFLAGS} | sed 's| -fstack-protector||'`"

I put them right under line 99 where you link the newlib folder into the srcdir

I tested the changes on x86_64 and i686. This was the only way I could get the build to complete with _use_nacl=1

big_bum commented on 2011-12-12 18:15

Isn't the latest version 18.0.969.0 ?

sl1pkn07 commented on 2011-12-12 16:36

try now (update)

TamCore commented on 2011-12-12 15:12

resolving dependencies...
warning: cannot resolve "gconf libgnome-keyring", a dependency of "chromium-dev"
:: The following package cannot be upgraded due to unresolvable dependencies:
chromium-dev

sl1pkn07 commented on 2011-12-11 22:12

update

libjpeg-turbo not longer need (http://code.google.com/p/chromium/issues/detail?id=106954)
perl-switch not longer need

add use_system_libexpat and use_system_speex GYP_defines (and speex dependency)

sorry for delay

sl1pkn07 commented on 2011-12-09 04:35

new version. render images still fails but..

http://code.google.com/p/chromium/issues/detail?id=106954

when use internal libjpeg works without render fails

upload new pkgbuild here before upload to AUR

http://dl.dropbox.com/u/6596386/chromium-dev-17.0.963.2-1.src.tar.gz

please tests with use_system_libjpg=0 and use_system_libjpeg=1

and post the result

greeting and sorry for all

PS: not upload directly in AUR because prevent massive fail update

sl1pkn07 commented on 2011-12-08 04:47

i have problem in new version. the render images is poorly

http://wstaw.org/m/2011/12/08/plasma-desktopR14002.png


http://dl.dropbox.com/u/6596386/chromium-dev-17.0.963.0-1.src.tar.gz

please, test it :S

sl1pkn07 commented on 2011-11-10 17:38

fixed

sl1pkn07 commented on 2011-11-10 17:38

fixed

(but strange... repack or corrupt?)

kfgz commented on 2011-11-10 12:45

Md5sum of chromium-17.0.932.0.tar.bz2 is not correct. It should be afe0e00c9745a145bdc910c12497780e

kfgz commented on 2011-11-10 12:42

Md5sum of chromium-17.0.932.0.tar.bz2 is not correct.

sl1pkn07 commented on 2011-11-10 10:38

ok. fixed

imrehg commented on 2011-11-10 02:05

New location to tarballs (according to the README in the previous directory), like http://commondatastorage.googleapis.com/chromium-browser-official/chromium-17.0.934.0.tar.bz2

sl1pkn07 commented on 2011-11-07 06:25

sorry for the dalay

this weekend broken one of my HDD. fortunately restore all of data


update to 17.0.928.0

sl1pkn07 commented on 2011-10-19 22:18

update to 16.0.912.0

fix build with pulseaudio

change use gyp flags to build correctly with pulseaudio support. disable by default, to enable edit "_use_pulseaudio" variable to 1

greetings

misc commented on 2011-10-08 13:49

Nevermind. I does so anyway, or my limit is just too low (350). Still strange however as that wasn't an issue with previous versions.

misc commented on 2011-10-08 13:11

So I thought :)

Oddly enough, on my setup it (or ffmpegsumo.so) appears to have caused Chromium to improperly spawn/kill child processes, causing my nproc limit to be reached — thus preventing the creation of other processes. Serves me right for neglecting the options…

misc commented on 2011-10-08 13:10

So I thought :)

Oddly enough, on my setup it (or ffmpegsumo.so) appears to have caused Chromium to improperly spawn/kill child processes, causing my nproc limit to be reached — thus preventing the creation of other processes. Serves me right for not neglecting the options…

sl1pkn07 commented on 2011-10-06 22:05

why i use it.

:)

misc commented on 2011-10-06 17:20

[Not a complaint:] Why is "Build With Pulseaudio" enabled by default?

terminalmage commented on 2011-09-26 12:16

http://googlechromereleases.blogspot.com/2011/09/dev-channel-update_22.html

"Due to known bug, Linux isn't updated."

http://code.google.com/p/chromium/issues/detail?id=97655

I would upgrade at your own risk. This isn't an "official" release for Linux.

sl1pkn07 commented on 2011-09-26 12:10

update to 16.0.889.0

sl1pkn07 commented on 2011-09-16 18:46

update to 15.0.874.15

sl1pkn07 commented on 2011-09-14 14:13

update to 15.0.874.12

sl1pkn07 commented on 2011-09-09 21:00

update to 15.0.874.5

sl1pkn07 commented on 2011-09-09 19:13

doh! the RSS dont mark the update

update later, previous compile by myselft to test build correctly

sl1pkn07 commented on 2011-09-09 19:10

doh! the RSS dont mark the update

update later

Det commented on 2011-09-09 19:10

Note: Chrome Dev Channel has been updated to 15.0.874.5 for Windows, Mac, Linux, and Chrome Frame with the same changes as below.

http://googlechromereleases.blogspot.com/2011/09/dev-channel-update.html

sl1pkn07 commented on 2011-09-09 19:09

The Dev channel has been updated to 15.0.874.5 (Platform version: 1011.0) [b] for Chromebooks (Acer AC700, Samsung Series 5, and Cr-48)[/b].

http://googlechromereleases.blogspot.com/2011/09/dev-channel-updates-for-chromebooks_08.html

sl1pkn07 commented on 2011-09-09 19:08

The Dev channel has been updated to 15.0.874.5 (Platform version: 1011.0) [b] for Chromebooks (Acer AC700, Samsung Series 5, and Cr-48)[/b].

Det commented on 2011-09-09 19:04

So does 15.0.874.5 build?

sl1pkn07 commented on 2011-09-09 19:00

oks, thanks

greetings

terminalmage commented on 2011-09-09 18:58

Yes, I was able to build on i686 with the updated patch.

sl1pkn07 commented on 2011-09-09 18:01

archtaku. whit the patch mod fix your problem?

terminalmage commented on 2011-09-09 16:33

Just a note to people who tried to build 15.0.874.0 before the patch was modified to include i686: If you're getting patch errors, remove/rename the "src" dir.

sl1pkn07 commented on 2011-09-09 15:50

if have fails post in comment. dot mark outdated please. im not soothsayer

greetings

sl1pkn07 commented on 2011-09-08 19:55

try now

i'm modiffied patch to make same fix on 32bits (before only patch 64 bits)

greetings and sorry

terminalmage commented on 2011-09-08 19:45

Updated PKGBUILD does not work. CARCH is correctly identified as i686 and patch is not applied, but build error is still the same.

sl1pkn07 commented on 2011-09-08 19:39

fix in pkgbuild(?)

only when detect x86_64 apply the patch

fixed?¿

greetings

terminalmage commented on 2011-09-08 19:36

Didn't work.

sl1pkn07 commented on 2011-09-08 19:25

mmm oks

comment

patch -p1 -i ../nacl.gypi.patch

line

fix later

thanks

terminalmage commented on 2011-09-08 19:07

Build works on x86_64 but fails on i686: http://pastebin.com/fYg5LUGY

sl1pkn07 commented on 2011-09-08 18:12

update dev channel 15.0.874.0

add patch for error compile (see http://code.google.com/p/chromium/issues/detail?id=94518)

more configs for gnome users (add "use_gnome_keyring" and "linux_link_gnome_keyring" gyps flags)

sorry for delay

sl1pkn07 commented on 2011-08-30 21:29

probe this patches. but not works (and anhother patches and script fixes)


for this need help :S

terminalmage commented on 2011-08-30 15:51

Looks like this has been submitted to their bug tracker. May want to follow this bug report:

http://code.google.com/p/chromium/issues/detail?id=94518

sl1pkn07 commented on 2011-08-30 14:28

please. need help!

http://paste.kde.org/116143/

sl1pkn07 commented on 2011-08-20 18:27

update to update to 15.0.854.0

fixed don't close tabs

sl1pkn07 commented on 2011-08-18 01:56

back to 849

http://code.google.com/p/chromium/issues/detail?id=93086

sl1pkn07 commented on 2011-08-18 01:45

update to 15.0.854.0

sl1pkn07 commented on 2011-08-12 16:04

update to 15.0.849.0

- gyp_define "use_system_libevent=0" (fail build with "1")
- don't install "product_logo_64.png" icon

sl1pkn07 commented on 2011-08-08 12:47

fixed.

add gyp_define "include_pulse_audio"

sorry for this :S

imrehg commented on 2011-08-08 09:23

Looking at the changes, it only adds the option to the PKGBUILD, but "_use_pulseaudio" is not actually used in the GYP_DEFINES, so nothing changes in the build, except the "depends" of the package. Did I miss something?

sl1pkn07 commented on 2011-08-07 20:21

update to 14.0.835.18

addep "_use_pulseaudio" option

by default is 1 (enable), if you NOT use pulseaudio, set to 0 (disable, don't add pulseaudio in depends)

Meow commented on 2011-08-06 11:32

pulseaudio also needs to be included in dependencies, please add it.

sl1pkn07 commented on 2011-07-29 04:40

update dev channel 14.0.835.8

sl1pkn07 commented on 2011-07-27 14:12

update dev channel 14.0.835.4

- Change method force use Python2

imrehg commented on 2011-07-22 07:43

sl1pkn07: Still need to add to the PKGBUILD because if the package does not use 'make ${MAKEFLAGS}', then the line in makepkg.conf does not matter at all. Once it is included, the parameter value can be added manually (like I did in the example), or automatically like using the conf you referenced.

chmielu commented on 2011-07-19 18:28

@piotrek91666

pacman -S base-devel

piotrek91666 commented on 2011-07-19 18:19

I have error:
| *** Starting to parse CSSPropertyNames.in...
|
| *** Starting to parse CSSValueKeywords.in...
|
ACTION webcore_bindings_sources_MathMLNames out/Release/obj/gen/webkit/MathMLNames.cpp
ACTION webcore_bindings_sources_UserAgentStyleSheets out/Release/obj/gen/webkit/UserAgentStyleSheets.h
ACTION webcore_bindings_sources_XLinkNames out/Release/obj/gen/webkit/XLinkNames.cpp
ACTION webcore_bindings_sources_XMLNSNames out/Release/obj/gen/webkit/XMLNSNames.cpp
ACTION webcore_bindings_sources_XMLNames out/Release/obj/gen/webkit/XMLNames.cpp
ACTION webcore_bindings_sources_tokenizer out/Release/obj/gen/webkit/tokenizer.cpp
RULE webcore_bindings_sources_bison_0 out/Release/obj/gen/webkit/CSSGrammar.cpp
Traceback (most recent call last):
File "scripts/action_maketokenizer.py", line 101, in <module>
sys.exit(main(sys.argv))
File "scripts/action_maketokenizer.py", line 89, in main
p1 = subprocess.Popen(['flex', '-t', flexInput], stdout=subprocess.PIPE)
File "/usr/lib/python2.7/subprocess.py", line 679, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1228, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
RULE webcore_bindings_sources_bison_1 out/Release/obj/gen/webkit/XPathGrammar.cpp
make: *** [out/Release/obj/gen/webkit/tokenizer.cpp] Błąd 1
make: *** Oczekiwanie na niezakończone zadania....
Traceback (most recent call last):
File "scripts/rule_bison.py", line 72, in <module>
returnCode = subprocess.call(['bison', '-d', '-p', prefix, inputFile, '-o', outputCpp])
File "/usr/lib/python2.7/subprocess.py", line 493, in call
return Popen(*popenargs, **kwargs).wait()
File "/usr/lib/python2.7/subprocess.py", line 679, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1228, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
make: *** [out/Release/obj/gen/webkit/CSSGrammar.cpp] Błąd 1
Traceback (most recent call last):
File "scripts/rule_bison.py", line 72, in <module>
returnCode = subprocess.call(['bison', '-d', '-p', prefix, inputFile, '-o', outputCpp])
File "/usr/lib/python2.7/subprocess.py", line 493, in call
return Popen(*popenargs, **kwargs).wait()
File "/usr/lib/python2.7/subprocess.py", line 679, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1228, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
make: *** [out/Release/obj/gen/webkit/XPathGrammar.cpp] Błąd 1

sl1pkn07 commented on 2011-07-19 14:20

update Dev channel. 14.0.825

- GCC 4.6 patch not need (commented in pkgbuild)
- build ffmpegsumo.so by default (to use system ffmpeg libs change "_use_ffmpeg_system_libs" to "1"

sl1pkn07 commented on 2011-07-17 18:57

yes. I'd added this option because i'd had the same problem. i dont know why occur this, i think this occur because ffmpeg what is installed in the system uses new API (chromium uses old API)

the difference between ffmpeg or internal ffmpeg is minimal: (ffmpeg uses for decoder, not for encode)

http://i.imgur.com/rLj8j.png (compile with ffmpegsumo.so)

greetings

sl1pkn07 commented on 2011-07-17 18:40

yes. for this i add this option. i dont know occur this, i'm supposed occur for newer version of ffmpeg/libav (chromium uses old API)

the difference on used between system ffmpeg or internal ffmpeg is minimal: (ffmpeg uses for decoder, not for encode)

http://i.imgur.com/rLj8j.png

greetings

blkqi commented on 2011-07-17 18:23

compile was OK with _use_ffmpeg_system_libs=0

with _use_ffmpeg_system_libs=1 I get the av_malloc() conflict posted below.

sl1pkn07 commented on 2011-07-16 19:14

this not depend of your /etc/makepkg.conf?


#-- Make Flags: change this for DistCC/SMP systems
MAKEFLAGS="-j9"

imrehg commented on 2011-07-16 10:44

Hi, could you add ${MAKEFLAGS} to the build line, like "make BUILDTYPE=${_buildtype} ${MAKEFLAGS} chrome chrome_sandbox"? It's the same as in core/kernel26, and lets me build things much faster on a multicore system (e.g. running "MAKEFLAGS=-j8 makepkg"). Cheers!

solstice commented on 2011-07-14 13:31

why "update-desktop-database -q" is not run in post_install in .install file as for chromium ?

sl1pkn07 commented on 2011-07-13 15:10

update. thanks huulivoide for the "fix" xd

Huulivoide commented on 2011-07-13 12:11

Turn the system_icu off
till it gets fixed, upstream, or someone
writes a patch, unless your working on a patch alredy ^^

sl1pkn07 commented on 2011-07-12 12:25

fail build new release Dev channel

CXX(target) out/Release/obj.target/i18n_api/v8/src/extensions/experimental/number-format.o
v8/src/extensions/experimental/number-format.cc: En la función ‘icu_46::DecimalFormat* v8::internal::CreateFormatterFromSkeleton(const icu_46::Locale&, const icu_46::UnicodeString&, UErrorCode*)’:
v8/src/extensions/experimental/number-format.cc:254:5: error: ‘EStyles’ no es un miembro de ‘icu_46::NumberFormat’
v8/src/extensions/experimental/number-format.cc:254:32: error: expected ‘;’ before ‘style’
v8/src/extensions/experimental/number-format.cc:257:9: error: ‘style’ no se declaró en este ámbito
v8/src/extensions/experimental/number-format.cc:257:17: error: ‘kCurrencyStyle’ no es un miembro de ‘icu_46::NumberFormat’
v8/src/extensions/experimental/number-format.cc:260:17: error: ‘kIsoCurrencyStyle’ no es un miembro de ‘icu_46::NumberFormat’
v8/src/extensions/experimental/number-format.cc:263:17: error: ‘kPluralCurrencyStyle’ no es un miembro de ‘icu_46::NumberFormat’
v8/src/extensions/experimental/number-format.cc:267:55: error: ‘style’ no se declaró en este ámbito
CXX(target) out/Release/obj.target/i18n_api/gen/i18n-js.o
AR(target) out/Release/obj.target/remoting/libremoting_base.a
make: *** [out/Release/obj.target/i18n_api/v8/src/extensions/experimental/number-format.o] Error 1

sl1pkn07 commented on 2011-07-12 04:24

new DEV channel release. update later

sl1pkn07 commented on 2011-07-11 12:47

media/ffmpeg/ffmpeg_common.cc: En la función ‘base::TimeDelta media::ConvertFromTimeBase(const AVRational&, int64)’:
media/ffmpeg/ffmpeg_common.cc:15:70: error: ‘av_rescale_q’ no se declaró en este ámbito
media/ffmpeg/ffmpeg_common.cc: En la función ‘int64 media::ConvertToTimeBase(const AVRational&, const base::TimeDelta&)’:
media/ffmpeg/ffmpeg_common.cc:21:73: error: ‘av_rescale_q’ no se declaró en este ámbito
media/ffmpeg/ffmpeg_common.cc:22:1: aviso: el control alcanza el final de una función que no es void [-Wreturn-type]
make: *** [out/Release/obj.target/media/media/ffmpeg/ffmpeg_common.o] Error 1

Huulivoide commented on 2011-07-11 09:49

Here this should include
all

http://paste.pocoo.org/show/435997/

Huulivoide commented on 2011-07-11 09:45

Oh yeh I forgot the stubs dir
I remember there were two places I did modifications in
but as it was few weeks ago I didn't rememeber that
let me update the patch

sl1pkn07 commented on 2011-07-10 16:39

fail for me

out/Release/obj.target/geni/ffmpeg_stubs.cc:41:64: error: la declaración de la función C ‘void* av_malloc(unsigned int)’ genera un conflicto con
/usr/include/libavutil/mem.h:73:7: error: declaración previa de ‘void* av_malloc(size_t)’ aquí
CXX(target) out/Release/obj.target/wtf/third_party/WebKit/Source/JavaScriptCore/wtf/DecimalNumber.o

test with libav-git and using system ffmpeg (_use_ffmpeg_system_libs=1)

Huulivoide commented on 2011-07-10 09:55

http://paste.pocoo.org/show/435053/

This patch should fix some of the symbols
but im not sure if this is all, please
install some svn or git liav pacxkage
and test it ^^

sl1pkn07 commented on 2011-06-28 13:35

update dev: 14.0.803 and patches

sl1pkn07 commented on 2011-06-27 19:59

what patch? link please

sl1pkn07 commented on 2011-06-27 12:22

how patch? link please

Huulivoide commented on 2011-06-22 09:18

could you add a patch for newer ffmpeg-revisions?
It only recuires few changes to change depricated
funktion names with new ones. Like CODEC_TYPE_VIDEO
becomes AVMEDIA_TYPE_VIDEO, these are alredy so old
that they hawe been completely removed from ffmpeg
for some time.

sl1pkn07 commented on 2011-06-21 13:52

update dev channel: 14.0.797.0

- update patch for gcc46 and glib 2.14

- add option: "_use_ffmpeg_system_libs="

0 = build "ffmpegsumo.so" library
1 = use system ffmpeg

greetings

Rogof commented on 2011-06-20 03:23

Will you keep following the Dev Channel?

joshdmiller commented on 2011-06-19 22:42

I am using glib 1.2.10 still, so I commented out the tmalloc patch line and everything's groovy now.

sl1pkn07 commented on 2011-06-19 18:59

update: when using Glib 2.14 [testing] apply the patch for compile with glib 2.14, if use glib 2.10 [core], skip this patch

sl1pkn07 commented on 2011-06-19 15:59

what glib version using you?

joshdmiller commented on 2011-06-19 13:58

I still get tcmalloc compilation errors:


CXX(target) out/Release/obj.target/allocator/third_party/tcmalloc/chromium/src/thread_cache.o
third_party/tcmalloc/chromium/src/tcmalloc.cc:379:36: error: conflicting declaration ‘void* (* volatile __malloc_hook)(size_t, const void*)’
/usr/include/malloc.h:165:16: error: ‘__malloc_hook’ has a previous declaration as ‘void* (* __malloc_hook)(size_t, const void*)’
third_party/tcmalloc/chromium/src/tcmalloc.cc:387:47: error: conflicting declaration ‘void* (* volatile __realloc_hook)(void*, size_t, const void*)’
/usr/include/malloc.h:167:16: error: ‘__realloc_hook’ has a previous declaration as ‘void* (* __realloc_hook)(void*, size_t, const void*)’
third_party/tcmalloc/chromium/src/tcmalloc.cc:393:60: error: conflicting declaration ‘void (* volatile __free_hook)(void*, const void*)’
/usr/include/malloc.h:163:15: error: ‘__free_hook’ has a previous declaration as ‘void (* __free_hook)(void*, const void*)’
third_party/tcmalloc/chromium/src/tcmalloc.cc:1587:64: error: conflicting declaration ‘void* (* volatile __memalign_hook)(size_t, size_t, const void*)’
/usr/include/malloc.h:169:16: error: ‘__memalign_hook’ has a previous declaration as ‘void* (* __memalign_hook)(size_t, size_t, const void*)’
make: *** [out/Release/obj.target/allocator/third_party/tcmalloc/chromium/src/tcmalloc.o] Error 1
make: *** Waiting for unfinished jobs....
WARNING - duplicate keys exist in shortcut group IDD_CHROME_FRAME_FIND_DIALOG

Rogof commented on 2011-06-18 22:56

ffmpeg-gyp-vpx and webkit-debug-build patches are not needed anymore.

sl1pkn07 commented on 2011-06-18 22:48

update and adopt this package

fixed with patch for Glib 2.14

greetings

Anonymous comment on 2011-06-17 12:03

These days Stable is good enough for me. Have fun.

sl1pkn07 commented on 2011-06-14 16:11

+1 to jim1960 error :S

KittyKatt commented on 2011-06-14 04:16

Getting the same error as jim1960. Waiting for a fix, perhaps.

Anonymous comment on 2011-06-09 10:58

@yvonney: Both gperf and libgnome-keyring are in makedepends, meaning that they are needed at build time but could be uninstalled afterwards.

jim1960 commented on 2011-06-08 00:49

I just upgrade glibc to 2.14-1 then compile error:
third_party/tcmalloc/chromium/src/tcmalloc.cc:379:36: 錯誤:相互衝突的宣告 「void* (* __malloc_hook)(size_t, const void*)」
/usr/include/malloc.h:176:39: 錯誤:「__malloc_hook」早先被宣告為 「void* (* volatile __malloc_hook)(size_t, const void*)」
third_party/tcmalloc/chromium/src/tcmalloc.cc:387:47: 錯誤:相互衝突的宣告 「void* (* __realloc_hook)(void*, size_t, const void*)」
/usr/include/malloc.h:179:39: 錯誤:「__realloc_hook」 早先被宣告為 「void* (* volatile __realloc_hook)(void*, size_t, const void*)」
third_party/tcmalloc/chromium/src/tcmalloc.cc:393:50: 錯誤:相互衝突的宣告 「void (* __free_hook)(void*, const void*)」
/usr/include/malloc.h:173:38: 錯誤:「__free_hook」 早先被宣告為 「void (* volatile __free_hook)(void*, const void*)」
third_party/tcmalloc/chromium/src/tcmalloc.cc:1587:54: 錯誤:相互衝突的宣告 「void* (* __memalign_hook)(size_t, size_t, const void*)」
/usr/include/malloc.h:183:39: 錯誤:「__memalign_hook」 早先被宣告為 「void* (* volatile __memalign_hook)(size_t, size_t, const void*)」
make: *** [out/Release/obj.target/allocator/third_party/tcmalloc/chromium/src/tcmalloc.o] Error 1

yvonney commented on 2011-06-07 04:00

Upon install the dev version wants to install: gperf libgnome-keyring

are they now essential to the dev build?
forum posting suggestion great. big thanks for all the help!
no luck yet though that's ok for now.

yvonney commented on 2011-06-07 02:33

kindthanks. and sorry to be messing up things here with my post size etc.
can delete then in a bit I'd think good.

SO, now upon install the dev version wan't to install:
the following (2) packages will be installed by pacman: gperf libgnome-keyring

So unless I've missed something these two gnomey packages which are not wanted.
are they now essential to the dev build?
many thanks for all efforts. I'm used to lots of stumbling. forum suggestion great.
will delete my posts below in a day or two to clean it up here if that's ok.

Vi0L0 commented on 2011-06-07 00:42

@yvonney: i just dunno why you are getting this:
install file (chromium-dev.install) does not exist.
error. chromium-dev.install surelly is there, so maybe there is some problem with your filesystem?
I dunno, i think that you should write new post about this issue on the forum.

Ps. ah sorry you were writing about aurget, not yourt... i'm kinda sleepy (-.-)Zzz
then i have no idea is there possibility of editing PKGBUILD before building when using aurget...

Vi0L0 commented on 2011-06-07 00:28

@yvonney: i'm not yaourt user, although i'm pretty sure that with it you can edit PKGBUILD before building package.
So edit PKGBUILD of perl-switch and simply remove "make test &&"
You will find it in:
make &&make test &&
line inside build() function. Then exit editor and it shall build package.

yvonney commented on 2011-06-06 22:14

THANK YOU! me still having probs.
if I use makepkg with tarball amd PKGBUILD i get:
install file (chromium-dev.install) does not exist.

If I use aurget -S I get:

Looks good
Writing Makefile for Switch
Writing MYMETA.yml
cp Switch.pm blib/lib/Switch.pm
Manifying blib/man3/Switch.3pm
PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/given.t ... Useless use of numeric gt (>) in void context at t/given.t line 19.
t/given.t ... Failed 2/293 subtests
t/nested.t .. ok
t/switch.t .. ok

Test Summary Report
-------------------
t/given.t (Wstat: 0 Tests: 293 Failed: 2)
Failed tests: 2-3
Files=3, Tests=590, 0 wallclock secs ( 0.06 usr 0.00 sys + 0.48 cusr 0.01 csys = 0.55 CPU)
Result: FAIL
Failed 1/3 test programs. 2/590 subtests failed.
make: *** [test_dynamic] Error 255
==> ERROR: A failure occurred in build().
Aborting...
notice: perl-switch failed while building

Anonymous comment on 2011-06-06 12:02

Fixed build with icu 4.8

yvonney commented on 2011-06-06 00:05

THANKS everyone! can you advise? WHICH 'test' line to comment?
Still getting: '==> ERROR: install file (chromium-dev.install) does not exist.'

# for n in avcodec avdevice avfilter avformat avutil postproc swscale; do
# if test -e /usr/lib/lib${n}.so.[0-9]; then
# f=`echo /usr/lib/lib${n}.so.[0-9]`
# else
# f=`echo /usr/lib/lib${n}.so.[0-9][0-9]`
# fi
# f=`basename "$f"`
# ln -s ../$f "${chromium_home}/${f}"
# done

yvonney commented on 2011-06-06 00:02

THANKS everyone!
ONE problem for me: WHICH 'make test' line ... ??

Vi0L0 commented on 2011-06-05 23:19

yes, with use_system_icu=0 it's compiling fine

Vi0L0 commented on 2011-06-05 22:45

edit: or to change from:
use_system_icu=1 \
to:
use_system_icu=0 \

Vi0L0 commented on 2011-06-05 22:42

@alub: it's caused by icu which were updated lately:
icu (4.6.1-2 -> 4.8-1)
when i tried re-compile with new icu i'm getting error :/
i don't have time to investigate so i just copied needed libs from older icu package.

If you don't have compiled on older icu chromium-dev package then i guess it's good idea to remove:
use_system_icu=1 \
line from chromium-dev's PKGBUILD

alub commented on 2011-06-05 22:33

With 13.0.782.1-2, i have the following error:
/usr/lib/chromium-dev/chromium: error while loading shared libraries: libicui18n.so.46: cannot open shared object file: No such file or directory

sl1pkn07 commented on 2011-06-05 19:19

comment the line "make test"

this is the "tune pkgbuild" need


yvonney commented on 2011-06-05 04:49

ERROR: install file (chromium-dev.install) does not exist.

yvonney commented on 2011-06-05 04:47

runnin testing repo also. is it that perl version is too new in testing repo?
could use a solution:
Test Summary Report
-------------------
t/given.t (Wstat: 0 Tests: 293 Failed: 2)
Failed tests: 2-3
Files=3, Tests=590, 0 wallclock secs ( 0.06 usr 0.01 sys + 0.49 cusr 0.01 csys = 0.57 CPU)
Result: FAIL
Failed 1/3 test programs. 2/590 subtests failed.
make: *** [test_dynamic] Error 255
==> ERROR: A failure occurred in build().
Aborting...
notice: perl-switch failed while building, remove source files

Vi0L0 commented on 2011-06-04 01:07

@sl1pkn07: same here (testing), thanks for solution

sl1pkn07 commented on 2011-06-03 19:03

hello. (again xd)

fail build

Can't locate Switch.pm in @INC (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at ../make-hash-tools.pl line 23.
BEGIN failed--compilation aborted at ../make-hash-tools.pl line 23.
Can't locate Switch.pm in @INC (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at ../make-hash-tools.pl line 23.
BEGIN failed--compilation aborted at ../make-hash-tools.pl line 23.
make: *** [out/Release/obj/gen/webkit/DocTypeStrings.cpp] Error 2
make: *** Se espera a que terminen otras tareas....
make: *** [out/Release/obj/gen/webkit/ColorData.cpp] Error 2
/tmp/yaourt-tmp-sl1pkn07/aur-chromium-dev/src/chromium-13.0.782.1/third_party/WebKit/Source/WebCore/css/CSSGrammar.y:946.5: aviso: a `;' might be needed at the end of action code
/tmp/yaourt-tmp-sl1pkn07/aur-chromium-dev/src/chromium-13.0.782.1/third_party/WebKit/Source/WebCore/css/CSSGrammar.y:946.5: aviso: future versions of Bison will not add the `;'
WARNING - duplicate keys exist in shortcut group IDD_CHROME_FRAME_FIND_DIALOG

need install perl-switch (with tune pkgbuild)

im using [testing]

sl1pkn07 commented on 2011-06-03 19:02

hello. (again xd)

fail build

Can't locate Switch.pm in @INC (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at ../make-hash-tools.pl line 23.
BEGIN failed--compilation aborted at ../make-hash-tools.pl line 23.
Can't locate Switch.pm in @INC (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at ../make-hash-tools.pl line 23.
BEGIN failed--compilation aborted at ../make-hash-tools.pl line 23.
make: *** [out/Release/obj/gen/webkit/DocTypeStrings.cpp] Error 2
make: *** Se espera a que terminen otras tareas....
make: *** [out/Release/obj/gen/webkit/ColorData.cpp] Error 2
/tmp/yaourt-tmp-sl1pkn07/aur-chromium-dev/src/chromium-13.0.782.1/third_party/WebKit/Source/WebCore/css/CSSGrammar.y:946.5: aviso: a `;' might be needed at the end of action code
/tmp/yaourt-tmp-sl1pkn07/aur-chromium-dev/src/chromium-13.0.782.1/third_party/WebKit/Source/WebCore/css/CSSGrammar.y:946.5: aviso: future versions of Bison will not add the `;'
WARNING - duplicate keys exist in shortcut group IDD_CHROME_FRAME_FIND_DIALOG

need install perl-switch (with tune pkgbuild)

sl1pkn07 commented on 2011-06-03 17:44

hello. (again xd)

fail build

Can't locate Switch.pm in @INC (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at ../make-hash-tools.pl line 23.
BEGIN failed--compilation aborted at ../make-hash-tools.pl line 23.
Can't locate Switch.pm in @INC (@INC contains: /usr/lib/perl5/site_perl /usr/share/perl5/site_perl /usr/lib/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr/lib/perl5/core_perl /usr/share/perl5/core_perl .) at ../make-hash-tools.pl line 23.
BEGIN failed--compilation aborted at ../make-hash-tools.pl line 23.
make: *** [out/Release/obj/gen/webkit/DocTypeStrings.cpp] Error 2
make: *** Se espera a que terminen otras tareas....
make: *** [out/Release/obj/gen/webkit/ColorData.cpp] Error 2
/tmp/yaourt-tmp-sl1pkn07/aur-chromium-dev/src/chromium-13.0.782.1/third_party/WebKit/Source/WebCore/css/CSSGrammar.y:946.5: aviso: a `;' might be needed at the end of action code
/tmp/yaourt-tmp-sl1pkn07/aur-chromium-dev/src/chromium-13.0.782.1/third_party/WebKit/Source/WebCore/css/CSSGrammar.y:946.5: aviso: future versions of Bison will not add the `;'
WARNING - duplicate keys exist in shortcut group IDD_CHROME_FRAME_FIND_DIALOG


i find this patch: http://www.freebsd.org/cgi/query-pr.cgi?pr=157237

i have [testing]

sl1pkn07 commented on 2011-05-15 08:58

i used FFmpeg-MT fork from Mplayer2 repository

http://git.mplayer2.org/mplayer2-build/ -> http://repo.or.cz/w/FFMpeg-mirror/mplayer-patches.git

Anonymous comment on 2011-05-14 14:34

@sl4pkn07: Cannot reproduce.
$ pacman -Q ffmpeg
ffmpeg 20110330-2

sl1pkn07 commented on 2011-05-14 13:31

Fail Build (ffmpeg)

http://paste.kde.org/67309

imrehg commented on 2011-05-14 07:18

Hi, current version needs to disable building the native client for the moment, as it has to download a precompiled executable, but it's not a simple thing just yet.
As e.g. Gentoo did, add the following to the gyp setting list: disable_nacl=1

Also, gcc46.patch needs a little update because one of the files (third_party/libjingle/source/talk/base/sigslot.h) changed within the patched region, here's a patched version https://github.com/imrehg/aur/blob/72233efe203ff3dae2fe4dbac57037a29e2f1069/usedpackages/chromium/gcc46.patch

Anonymous comment on 2011-05-01 17:44

Apparently the path has changed from build.chromium.org/buildbot/official/ to build.chromium.org/official/. Not a problem as the former is set to redirect to the new address, just a warning.

Anonymous comment on 2011-04-19 07:50

12.0.733.0-2: Pulled gcc 4.6 build fix from http://trac.webkit.org/changeset/84123

Vi0L0 commented on 2011-04-18 18:53

@mvorozhtsov: yes, you are right, i just compiled it on virtual machine with gcc 4.5 and it's fine.

You may download x86_64 package from: http://catalyst.apocalypsus.net/other_pkgs/chromium-dev-12.0.733.0-1-x86_64.pkg.tar.xz

Anonymous comment on 2011-04-18 12:19

It's http://code.google.com/p/chromium/issues/detail?id=46411
I can't even build 12.0.725.0 anymore.
Does it happen with gcc 4.5? I can't downgrade my machine because I use GNOME 3.

Vi0L0 commented on 2011-04-18 11:11

I am also getting failure, although mine looks kinda different: http://hpaste.org/45804
+ i'm using [testing]

sl1pkn07 commented on 2011-04-18 11:06

fail build:

http://paste.kde.org/13156

sl1pkn07 commented on 2011-04-18 10:30

fail build:

http://paste.kde.org/13156

Anonymous comment on 2011-04-08 07:56

Use Arch about 2 years and this is the first package that requires base-devel:)

Anonymous comment on 2011-04-08 05:19

@4ybaka
https://wiki.archlinux.org/index.php/AUR#Getting_started

Anonymous comment on 2011-04-07 16:04

@mvorozhtsov
Thanks, I haven't some of these packages. Why them not included in dependencies?

Anonymous comment on 2011-04-07 13:31

@4ybaka
$ pacman -S base-devel

Anonymous comment on 2011-04-07 13:12

Build failed on last update...

Traceback (most recent call last):
File "scripts/action_maketokenizer.py", line 101, in <module>
sys.exit(main(sys.argv))
File "scripts/action_maketokenizer.py", line 89, in main
p1 = subprocess.Popen(['flex', '-t', flexInput], stdout=subprocess.PIPE)
File "/usr/lib/python2.7/subprocess.py", line 672, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1202, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
make: *** [out/Release/obj/gen/webkit/tokenizer.cpp] Error 1

Anonymous comment on 2011-04-06 07:11

Yes, I'm on testing. I assume then that the issue lies with my toolchain.

Anonymous comment on 2011-04-05 07:45

@robert_paxon: Not me. Are you on [testing]?

Anonymous comment on 2011-04-04 21:35

This compilation error has been present since 12.0.712.0-1:
CXX(target) out/Release/obj.target/browser/chrome/browser/search_engines/template_url_prepopulate_data.o
In file included from chrome/browser/search_engines/template_url_prepopulate_data.cc:5:0:
./chrome/browser/search_engines/template_url_prepopulate_data.h:30:29: error: ‘size_t’ has not been declared
make: *** [out/Release/obj.target/browser/chrome/browser/search_engines/template_url_prepopulate_data.o] Error 1

-------------------------

I assume it's completely an upstream issue but I'm posting to see if others are getting the same error.

ydasilva commented on 2011-03-18 22:14

@mvorozhtsov my bad then :)

Anonymous comment on 2011-03-15 11:23

@dl.zerocool
This package tracks the Dev channel (http://www.chromium.org/getting-involved/dev-channel, announcements are published at http://googlechromereleases.blogspot.com/) and NOT supposed to provide the "last-published-tarball" version.

ydasilva commented on 2011-03-15 10:31

To get the new version :

pkgver=11.0.696.3
to
pkgver=12.0.703.0

You can always check the last version on : http://build.chromium.org/official/

Shall you do a small script to update the pkgbuild daily :)
(And change md5 signature of course)

ydasilva commented on 2011-03-15 10:27

To get the new version :

pkgver=11.0.696.3
to
pkgver=12.0.703.0

You can always check the last version on : http://build.chromium.org/official/

Shall you do a small script to update the pkgbuild daily :)

Anonymous comment on 2011-03-13 21:21

It appears not to work with the NaCl demos.
ex, http://nacl-gallery.appspot.com/life/life.html

Note, I have --no-sandbox and I have enabled NaCl in about:flags.


[6949:6949:110715291860:ERROR:gpu_info_collector_linux.cc(178)] Failed to locate libpci
[6972:6972:110715410744:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[6974:6974:110715477114:ERROR:renderer_main.cc(299)] Running without renderer sandbox
Moonlight: 3.99.0.1.99
Moonlight: Attempting to load libmoonloaderxpi
[6977:6977:110715706753:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[6981:6981:110715884523:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[6986:6986:110716197973:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[6997:6997:110721749765:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[7000:7000:110753067670:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[7003:7003:110753212924:ERROR:native_library_linux.cc(32)] dlopen failed when trying to open default_plugin: default_plugin: cannot open shared object file: No such file or directory
[7005:7005:110767985590:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[7019:7019:110913395047:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[7022:7022:110918882152:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[7025:7025:110932539981:ERROR:renderer_main.cc(299)] Running without renderer sandbox
[7029:7029:111125708219:ERROR:renderer_main.cc(299)] Running without renderer sandbox

Huulivoide commented on 2011-03-12 12:49

ffmpeg version must be lower than 20110215 as there was a bigger change in the
code base and chromium hasnt been updated to reflect it yet.

Anonymous comment on 2011-03-09 13:54

Beware of https://bugs.archlinux.org/task/23204

Anonymous comment on 2011-02-18 12:50

11.0.672.2: Disabled system ICU for now as it breaks linking for me.

Anonymous comment on 2011-02-11 23:33

Now, after nearly two hours of compiling, it works very smooth :)

Anonymous comment on 2011-02-01 14:42

There was no .11 tarball so I packaged .12.
@Det: "next update" = "when I have time".

Det commented on 2011-02-01 13:55

Out of curiosity, is there a specific reason you want to wait until the next release each time there's something to be changed in the PKGBUILD?

If you checked/verified that (10.0.648).11 was already released, then it's of course fine to just say that you'll be doing both changes on the same go, though.

Anonymous comment on 2011-02-01 13:21

lastchange.py uses git, I'll add it to makedepends in the next update.

terminalmage commented on 2011-01-31 19:36

I have the same error when building on i686, only with out/Release/obj.target/geni/webkit_version.h instead of out/Release/obj/gen/chrome/grit/component_extension_resources.h. Didn't have a problem building on x86_64 though.

jpaechnatz commented on 2011-01-31 06:59

...fails to compile...
RULE chrome_extra_resources_grit_0 out/Release/obj/gen/chrome/grit/component_extension_resources.h
Traceback (most recent call last):
File "./lastchange.py", line 149, in <module>
sys.exit(main())
File "./lastchange.py", line 134, in main
version_info = FetchVersionInfo(opts.default_lastchange)
File "./lastchange.py", line 84, in FetchVersionInfo
if IsGitSVN(directory):
File "./lastchange.py", line 31, in IsGitSVN
cwd=directory)
File "/usr/lib/python2.7/subprocess.py", line 486, in call
return Popen(*popenargs, **kwargs).wait()
File "/usr/lib/python2.7/subprocess.py", line 672, in __init__
errread, errwrite)
File "/usr/lib/python2.7/subprocess.py", line 1202, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
make: *** [out/Release/obj/gen/build/LASTCHANGE.always] Fehler 1
make: *** Warte auf noch nicht beendete Prozesse...
WARNING - duplicate keys exist in shortcut group IDD_CHROME_FRAME_FIND_DIALOG
sk duplicates: &N
fr duplicates: &R
ro duplicates: &G
nl duplicates: &O
no duplicates: &N
pt-PT duplicates: &P
sw duplicates: &F
ca duplicates: &A
de duplicates: &N
tr duplicates: &A
sv duplicates: &S
da duplicates: &O
lv duplicates: &R, &U
pl duplicates: &W
cs duplicates: &N
fi duplicates: &K
es-419 duplicates: &A
id duplicates: &T
es duplicates: &B

Anonymous comment on 2011-01-25 08:23

I've also tried to compile without 'disable_sse2=1' on x86_64 Intel Core 2 Duo, Intel Core i5, Intel Xeon. Everything seems to work fine.

Anonymous comment on 2011-01-15 07:06

OK, I'll add this in the next update.

Vi0L0 commented on 2011-01-14 17:54

i've compiled it on x86_64 with sse2 and everything is ok - so i could recommend it to other x86_64 users.

ofcourse you can leave it how it looks now - if somebody wants he can always remove 'disable_sse2=1'
but maybe a good consensus here is to copy solution from [1] extra's chromium:

$([ "${CARCH}" == 'i686' ] && echo '-Ddisable_sse2=1')


[1] http://projects.archlinux.org/svntogit/packages.git/tree/chromium/repos/extra-x86_64/PKGBUILD

Anonymous comment on 2011-01-14 16:08

It seems that enabling SSE2 results in wrong CFLAGS on some systems: http://bugs.gentoo.org/show_bug.cgi?id=336871
At least it is forcibly disabled in Gentoo for that reason: http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/www-client/chromium/chromium-10.0.634.0.ebuild?revision=1.1&view=markup

Vi0L0 commented on 2011-01-14 15:41

hmm... i can only see disable_sse2=1
and afaik sse2 is supported by all x86_64 cpus - so we don't need any check for this arch
ofcourse if im right, but i should be - sse2 is known from 2001's pentium 4 cpus, and amd added such support in their athlon64 cpus...

Det commented on 2011-01-14 13:31

@mvorozhtsov, can't you add a check of /proc/cpuinfo whether sse is supported? :p

Malix commented on 2011-01-14 10:07

@mvorozhtsov: ok, sounds reasonable. so if I have SSE capable cpu, I assume I should change that to zero. Does makepkg.conf override that if I use -march=native, for example.

Anonymous comment on 2011-01-14 09:59

@Malix: Not all i686 CPUs support SSE.

Malix commented on 2011-01-13 15:25

why do we disable_sse=1 in this PKGBUILD, is there some negative effect of using it? Also, is there some place where I could do some reading on these options, tried to search them with google but didn't find anything worthwhile.

Koshon commented on 2011-01-06 22:20

...
pkgver=10.0.628.0
...
md5sums=('2c50b326039453077c6b99197e0204ed'...

Det commented on 2011-01-06 07:36

@theislainman, you are free to push the *Flag Out-of-date* button too :).

As for mvorozhtsov, please check out the minor tweaks I made to your PKGBUILD on the previous comment (in case you missed it).

Anonymous comment on 2011-01-06 05:56

http://googlechromereleases.blogspot.com/2011/01/dev-channel-update.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+GoogleChromeReleases+%28Google+Chrome+Releases%29

Det commented on 2010-12-23 14:26

You are free to take whatever liberties you like with your _own_ packages. For good reasons is always even better.

Anonymous comment on 2010-12-23 12:55

@Det: I looked up the RegEx bug they mention on the blog and it seems to affect Linux machines too, so I took the liberty of upgrading to the newer version.

Det commented on 2010-12-23 11:15

That'd be nice. The only possible drawbacks with system libraries would be that they were either too new or buggy. I don't think sqlite is.

Malix commented on 2010-12-23 11:00

couldn't we use more system libraries etc, for example: Sqlite ("use_system_sqlite=1")? Seems to work, but is there any drawbacks to this? Afaik we're currently compiling stuff most of us have already installed.

Det commented on 2010-12-22 16:17

Another btw: here's some minor improvements to your PKGBUILD: http://aur.pastebin.com/CzMuGTTz

Det commented on 2010-12-22 16:16

Another btw: here's some minor improvements to your PKGBUILD: http://aur.pastebin.com/gYgvzGJR

Det commented on 2010-12-22 16:11

Another btw: here's some minor improvements to your PKGBUILD: http://aur.pastebin.com/PBspdtLu

Det commented on 2010-12-22 13:32

Btw. 10.0.612.3 is an 'incorrect' version (http://googlechromereleases.blogspot.com/2010/12/dev-channel-update_16.html):
"Update: The Dev channel has been updated to 10.0.612.3 for _Windows_ and _Mac_ to correct a RegEx regression."

Det commented on 2010-12-20 18:39

Please remove the svn makedependency on the same go (in case you missed c00kiemon5ter's comment).

terminalmage commented on 2010-12-20 18:30

@mvorozhtsov: Looks like they finally posted it

Anonymous comment on 2010-12-20 08:33

Waiting for the 10.0.612.1 tarball at http://build.chromium.org/buildbot/official/

Det commented on 2010-12-14 15:37

No.

c00kiemon5ter commented on 2010-12-14 14:19

also I can build this without svn installed, is it really a dep ?
makedepends=('svn' .. )

c00kiemon5ter commented on 2010-12-14 14:17

also I can build this without svn installed, is it really a dep ?

Det commented on 2010-12-13 19:13

@dolby, you thought correct: http://code.google.com/p/chromium/issues/detail?id=13322

Anonymous comment on 2010-12-13 19:05

I thought gconf wasnt required now: http://phajdan-jr.blogspot.com/2010/11/www-clientchromium-gconf-no-longer.html

alphazo commented on 2010-12-12 21:23

Never mind, Chromed Bird and Proxy Switchy are working back on 9.0.597.16.

alphazo commented on 2010-12-12 20:33

Since 9.0.596, "Chromed Bird" and "Proxy Switchy!" extensions have stopped working (icon is there but nothing happens when clicking on it). Does anyone have them running with the latest version (Dec. 12)?

Det commented on 2010-12-10 16:40

@c00kiemon5ter, I can't because there is nothing to correct.

c00kiemon5ter commented on 2010-12-10 01:59

@fettouchi
as ioni pointed in #archlinux -dev is different from 'latest snapshot'
ie -dev is this: http://googlechromereleases.blogspot.com/
"The Dev channel has been updated to 9.0.597.10 @ Monday, December 6, 2010"
while at the same time http://build.chromium.org/buildbot/official/ already included -10.0.60x

this is supposed to be following -dev so I dont think it'll be updated to -10.x soon
correct me if I'm wrong

fettouhi commented on 2010-12-08 11:03

I can see that 10.0.x is out. Are you going to switch to that branch soon? does the google webstore and ebook store work with chromium or do you need the official google chrome browser?

Anonymous comment on 2010-12-05 21:30

@Det: Thank you for the answer. I will wait ;)

Det commented on 2010-12-05 19:12

You're just gonna have to wait until they fix hardware acceleration again. It didn't work with Egl/Gles2 enabled mesa 7.9 for me and it doesn't work with mesa-git either.

With both I just get some Egl symbol lookup error about some undefined symbol or something. If I then enable accelerated 2d canvas the extension LastPass stops working - while the accelerated 2d canvas doesn't work either anyway.

If you really want to you can rebuild your mesa with "--enable-gles2" but I really don't think this issue is just on my system.

Anonymous comment on 2010-12-05 15:47

So, what is the solution of the problem about libGLESv2? Thanks ;)
Now, i'm using the package of chromium from official repository :(

Det commented on 2010-12-05 08:31

Ahhh, of course it was _Gles2_ not _Egl_. Ok then, my fault the whole thing then. I dunno how I got the idea that "libGLESv2.so" would suddenly be a Mesa Egl component but no harm done... except for archtaku I guess. Sorry about that ;).

Anonymous comment on 2010-12-05 08:19

alphazo problem is with missing libGLESv2.so, not libffmpegsumo.so. I don't know how he managed to pull this dependency though, looking at media/media.gyp, the default backend is x11, not GLES. I have neither of these shared libraries on my system and chromium is running fine (it complains about libffmpegfumo.so, but that's fine). Maybe it has something to do with enabling experimental 3D/acceleration features (I don't use them).

Det commented on 2010-12-05 08:07

So when alphazo said he cannot start chromium because of the two errors - Mesa Egl and libffmpegsumo.so - he actually could?

Anonymous comment on 2010-12-05 08:00

Just to clear things up. libffmpegsumo.so is entirely optional and it's absence can't cause any failure.

Koshon commented on 2010-12-04 23:15

to Det: it's not for update, only for who want.

Koshon commented on 2010-12-04 22:41

to Det: it's not for update, only for who want.

Det commented on 2010-12-04 19:28

Stay in your pants, it's not about being hard to understand. I'm just wondering how can you bypass the Mesa Egl requirement by copying libffmpegsumo.so (Ffmpeg component) to the install folder.

terminalmage commented on 2010-12-04 18:56

Det: Because it _was_ less work. I already had chromium installed, and even if I didn't, installing it and copying the .so _is_ less work than installing one package and recompiling another. I'm fully aware it's not the cleanest method of resolving the issue, which is why I added that it was a dirty hack.

Why is this so hard for you to understand?

Det commented on 2010-12-04 18:35

archtaku: Hemm, in that case I _really_ wonder why did you say "but less work than installing an extra package and rebuilding mesa."

Koshon: if you mean this package should be updated, you are wrong. Check here the latest stable/beta/_dev_ channel releases: http://googlechromereleases.blogspot.com/

Det commented on 2010-12-04 18:34

archtaku: Hemm, in that case I _really_ wonder why did you say "but less work than installing an extra package and rebuilding mesa."

terminalmage commented on 2010-12-04 18:25

Det: You say this as if you think it's all news to me.

Koshon commented on 2010-12-04 16:43

10.0.601.0 md5sum c84b48d9468eeeb71ddc85e4b4f74d4f
use of health..

Det commented on 2010-12-04 10:06

@archtaku, if you get an error about a missing mesa feature, your 'dirty hacks' with moving chromium files around aren't going to help. Additionally not everybody even has two chromiums installed at the time and the problem with manually doing stuff is that to keep up doing it you need to do it yourself - whilst packages are auto-updating.

@alphazo, that happens with me too. I just don't use accelerated compositing/2d canvas for now. Dunno about your extensions, though. At least with me LastPass started working again when I disabled accelerated 2d canvas (about:flags or remove the manual launcher switch).

alphazo commented on 2010-12-03 22:51

I went the recompilation way (mesa) and installed chromium-dev-codecs-ffmpeg-nonfree-bin. Chromium-dev starts again with only one error message though: /usr/lib/chromium-dev/chromium --type=gpu-process --channel=5542.0x58382b8.1442616187: symbol lookup error: /usr/lib/egl/egl_dri2.so: undefined symbol: _glapi_get_proc_address

Now I have two extensions that stopped working on this release (9.0.597): Chromed Bird and Proxy Switch!

alphazo commented on 2010-12-03 20:27

Still no luck even with having the .so file in /usr/lib/chromium-dev. I still have this last issue:

[5652:5652:3751961263:ERROR:base/native_library_linux.cc(32)] dlopen failed when trying to open libGLESv2.so: libGLESv2.so: cannot open shared object file: No such file or directory

Anonymous comment on 2010-12-03 20:11

bortol ~ $ chromium-dev
[14546:14546:29055732666:ERROR:base/native_library_linux.cc(32)] dlopen failed when trying to open libGLESv2.so: libGLESv2.so: cannot open shared object file: No such file or directory chromium-dev-codecs-ffmpeg-nonfree-bin
[14471:14471:29057246618:ERROR:chrome/browser/renderer_host/render_sandbox_host_linux.cc(625)] sendmsg: Connessione rifiutata
Segmentation fault

I have already installed chromium-dev-codecs-ffmpeg-nonfree-bin from aur.

terminalmage commented on 2010-12-03 17:39

Instead of doing that, I copied the .so from /usr/lib/chromium/ to /usr/lib/chromium-dev/. A dirty hack, yes, but less work than installing an extra package and rebuilding mesa.

Det commented on 2010-12-03 10:54

You need the ffmpeg codecs from the package chromium-dev-codecs-ffmpeg-nonfree-bin (should be marked as a dependency) and rebuild mesa with "--enable-gles2" and "--enable-egl".

Maybe [extra]'s mesa will be built with those flags in the end anyway... just have to wait.

alphazo commented on 2010-12-03 07:30

I have been using chromium-dev for quite some time now. However chromium-dev 9.0.597.0-1 compiles fine but I cannot start it:

[1:1:287546422:ERROR:base/native_library_linux.cc(32)] dlopen failed when trying to open /usr/lib/chromium-dev/libffmpegsumo.so: /usr/lib/chromium-dev/libffmpegsumo.so: cannot open shared object file: No such file or directory
[3:3:288278871:ERROR:base/native_library_linux.cc(32)] dlopen failed when trying to open /usr/lib/chromium-dev/libffmpegsumo.so: /usr/lib/chromium-dev/libffmpegsumo.so: cannot open shared object file: Permission denied
[6171:6171:288841360:ERROR:base/native_library_linux.cc(32)] dlopen failed when trying to open libGLESv2.so: libGLESv2.so: cannot open shared object file: No such file or directory

I looked inside /usr/lib/chromium-dev/ and there is no such libffmpegsumo.so.

Det commented on 2010-11-25 14:12

@bsergik, for the 999999th time, flex, bison, automake, autoconf, etc. is included in the group "base-devel" that consists of the most common stuff needed to build stuff from source.

Spread the world.

bsergik commented on 2010-11-25 12:13

Please add flex and bison as a makedepends. Otherwise build fails.


ACTION Executing flex on compiler/glslang.l out/Release/obj.target/geni/glslang.cpp
/bin/sh: flex: command not found
make: *** [out/Release/obj.target/geni/glslang.cpp] Error 127
Aborting...

ACTION Executing bison on compiler/glslang.y out/Release/obj.target/geni/glslang_tab.cpp
/bin/sh: bison: command not found
make: *** [out/Release/obj.target/geni/glslang_tab.cpp] Error 127
Aborting...

tomk commented on 2010-11-14 12:34

9.0.580.0 requires libxtst in depends.

Det commented on 2010-11-05 16:48

9.0.570.1 brought only new flash for Chrome users. Not really a reason for rebuild...: http://googlechromereleases.blogspot.com/2010/11/dev-channel-update_04.html

Det commented on 2010-10-30 07:40

@Dogmeat, he probably means this package. Unless Jesses uses accelerated 2d canvas there's no help with 8.0.552.23 either: http://googlechromereleases.blogspot.com/2010/10/dev-channel-update_29.html

It was removed from "about:flags" with .23 but it can, of course, be re-enabled with the command line switch: "--enable-accelerated-2d-canvas"

Det commented on 2010-10-30 07:39

@Dogmeat, he probably means this package. Unless Jesses uses accelerated 2d canvas there's no help with 8.0.552.23: http://googlechromereleases.blogspot.com/2010/10/dev-channel-update_29.html

It was removed from "about:flags" with .23 but it can - of course - be re-enabled with the command line switch: "--enable-accelerated-2d-canvas"

Det commented on 2010-10-30 07:38

@Dogmeat, he probably means this package. Unless Jesses uses "--enable-accelerated-2d-canvas" there's no help with 8.0.552.23: http://googlechromereleases.blogspot.com/2010/10/dev-channel-update_29.html

Accelerated 2D Canvas was removed from "about:flags" with .23 but it can - of course - be re-enabled with the command line switch: "--enable-accelerated-2d-canvas"

Anonymous comment on 2010-10-29 23:55

Jesses: sometimes with 9.0.563.0.

Huulivoide commented on 2010-10-29 17:43

Anyone elsa having lock ups whit youtube?

Vi0L0 commented on 2010-10-28 13:58

i686 package: http://www.mediafire.com/file/pw3mtcrgq84j71u/chromium-dev-8.0.552.18-1-i686.pkg.tar.xz

+ gpu acceleration can be enabled in about:flags tab

Vi0L0 commented on 2010-10-28 13:39

i686 package: http://www.mediafire.com/file/pw3mtcrgq84j71u/chromium-dev-8.0.552.18-1-i686.pkg.tar.xz

Det commented on 2010-10-22 13:02

Didn't the Ffmpeg stuff come with Google Chrome and for "chromium-browser-bin" users through the "chromium-codecs-ffmpeg-nonfree-bin" package?

Anonymous comment on 2010-10-22 05:55

@hokapoka
I may be wrong, but I think ffmpegsumo stuff is installed only when you build the ffmpeg bundled with Chromium, and this PKGBUILD uses the Arch version.

Det commented on 2010-10-21 17:00

@mvorozhtsov, of course, sorry about that.

Anonymous comment on 2010-10-21 16:30

Not sure if it's just my machines but when starting chromium-dev on all 4 of my boxes I was getting the following error :

[1:1:6238048324:ERROR:base/native_library_linux.cc(28)] dlopen failed when trying to open /usr/lib/chromium-dev/libffmpegsumo.so: /usr/lib/chromium-dev/libffmpegsum o.so: cannot open shared object file: No such file or directory

The browser appeared to function perfectly without, I assume it's to do with some video playback but haven't tried to see what it might be effecting.

I located the file from the binary install of chromium from the extra repos and copied it into place :

# sudo cp /usr/lib/chromium/libffmpegsumo.so /usr/lib/chromium-dev/

Restarted chromium-dev and the error hasn't re-appeared.

It's a crappy hack I know, but I thought that I would pass it on.

Thanks again mvorozhtsov for maintaing this package.



Anonymous comment on 2010-10-21 15:24

@Det
As you can see at http://googlechromereleases.blogspot.com/
>Update: The Dev channel has been updated to 8.0.552.5 for Windows, Mac, and ChromeFrame
But not for Linux.

Det commented on 2010-10-21 15:14

To whoever unflagged this as Out-of-date have a look here before doing it again (8.0.552.5 released): http://googlechromereleases.blogspot.com/

jackoneill commented on 2010-10-20 18:37

Commenté par: Det sur Fri, 15 Oct 2010 10:52:46 +0000

You get a binary package with google-chrome/-dev/-beta.

Kwpolska commented on 2010-10-20 18:36

Is there any binary package for chromium-dev?

Anonymous comment on 2010-10-20 12:33

Ouch, I didn't know that, thanks for replying so fast.

baghera commented on 2010-10-20 12:19

Also, packages already included in base-devel should not be included in makedepends. base-devel had been created to avoid cluttering of makedepends field.

baghera commented on 2010-10-20 12:18

@ TrashMaster

You should have done just

pacman -S base-devel

base-devel group is mandatory for those who want to build something on arch (it's written in wiki). I suggest to install the whole group.

Anonymous comment on 2010-10-20 12:14

It also depends on patch, bison, and another packages. I've got to restart the build of chromium tree times, installing missing dependencies.

Anonymous comment on 2010-10-19 07:41

8.0.552.0-2: Compatibility with the new python environment.

Det commented on 2010-10-15 10:52

You get a binary package with google-chrome/-dev/-beta.

dcelasun commented on 2010-10-14 20:51

Is there a place where I can find a binary package for this? It just takes too much time for every update.

alphazo commented on 2010-10-14 13:11

Oops.. My mistake. I had "-g" in my /etc/makepkg.conf. I enabled it to track down a bug on Thunderbird. Thanks for the tip!

alphazo commented on 2010-10-14 13:07

Why is chromium-dev-8.0.552.0-1 package so fat (155MB) and why does it require 787MB to install?

Malix commented on 2010-10-14 13:06

@alphazo
did you edit the PKGBUILD and disabled stripping the debug symbols?

atleast on my box the package is roughly 17MB

alphazo commented on 2010-10-14 13:04

Why does chromium-dev-8.0.552.0-1 package is so fat (155MB) and requires 787MB to install?

tomk commented on 2010-10-13 18:57

I guess I cheated then. :)

I did (briefly) consider wading through it myself for python stuff, but then I decided to rebuild testing/python2 with a /usr/bin/python symlink. After that, the build was successful.

It runs fine too.. mostly. That is, it seems to be working as expected so far, but I got a "no such file or directory" error about /usr/lib/chromium-dev/libffmpegsumo.so - and there is indeed no such file or directory in the package I created. extra/chromium has that lib btw.

demonicmaniac commented on 2010-10-05 12:53

Hi.
There's some issues with this PKGBUILD when you have testing enabled as testing to moved to python3.
The steps suggested by Allan in his blog don't suffice to get it built as some .py files explicitly call python instead of python2 and I can't be arsed to wade through this mess to fix it >.> Perhaps somebody with more patience might be able to.

Anonymous comment on 2010-10-05 07:17

7.0.536.2-2: Added fix for http://code.google.com/p/chromium/issues/detail?id=57167

mutantmonkey commented on 2010-10-02 20:35

My apologies; I must have neglected to install base-devel after my reinstall.

baghera commented on 2010-10-02 18:16

mutantmonkey, bison is in base-devel group, thus it should not be included in makedepends. base-devel is a requirement to compile anything on arch.

mutantmonkey commented on 2010-10-02 16:41

Please add bison to makedepends, I wasn't able to build without installing it.

Anonymous comment on 2010-09-29 14:43

@jpaechnatz: Looks like http://code.google.com/p/chromium/issues/detail?id=57275, except I use -Duse_system_ssl=0.
It seems you build somehow gets hooked up with the system nss in spite of that.

jpaechnatz commented on 2010-09-29 07:42

Since 7.0.517.8 I wasn't able to compile a newer version...caused by similar failures like:
CXX(target) out/Release/obj.target/net/net/spdy/spdy_http_stream.o
net/socket/ssl_client_socket_nss.cc: In Funktion »int net::<unnamed>::MapNSPRError(PRErrorCode)«:
net/socket/ssl_client_socket_nss.cc:212:10: Fehler: »SSL_ERROR_WEAK_SERVER_KEY« wurde in diesem Gültigkeitsbereich nicht definiert
make: *** [out/Release/obj.target/net/net/socket/ssl_client_socket_nss.o] Fehler 1
make: *** Warte auf noch nicht beendete Prozesse...
Breche ab ...

Do I miss a package? I already checked for nss/nspr, but they are current (maybe to new?)
Thanks.

terminalmage commented on 2010-09-25 01:57

New official release: 7.0.517.13

Det commented on 2010-09-23 18:05

Well, at least the "google devs is lazy" and "only one excuse for not updating" parts made my day :).

Det commented on 2010-09-23 18:04

Well, at least the "google devs is lazy for updating the blog" and "excuse for not updating" parts made my day :).

terminalmage commented on 2010-09-22 22:54

Again, sl1pkn07, if you want a newer version than the latest official dev release, download the tarball (http://aur.archlinux.org/packages/chromium-dev/chromium-dev.tar.gz), unpack, modify the release number in the PKGBUILD, and run makepkg. It's not that hard.

But, don't expect every tarball to build because, as baghera said, the tarballs are auto-generated and aren't guaranteed to build.

baghera commented on 2010-09-22 18:37

@ sl1pkn07
You should look a bit more closer at the link you provided. There are _at least_ two tarballs a day. Those are _not_ all official release, they're official tarballs generated by a buildbot. Sometimes those tarballs are broken and won't compile, while official releases tracked on the blog brings code in a consistent state.
Google devs are not lazy updating the blog.

sl1pkn07 commented on 2010-09-22 18:29

you no look the link address?

http://build.chromium.org/buildbot/>official</chromium-7.0.530.0.tar.bz2


the google devs is lazy for updating the blog (of course, working in the program)

last post in dev channel: "Thursday, September 16, 2010 | 17:52"

the #malix post is only one excuse for not updating this package

if i have want a new version, download from the SVN... o wait: http://build.chromium.org/buildbot/official/chromium-7.0.517.8.log

greetings

terminalmage commented on 2010-09-22 16:53

@sl1pkn07: Please read the comments before flagging out-of-date. The PKGBUILD is only updated for official dev releases. (see http://googlechromereleases.blogspot.com/)

If you want a newer version, download the PKGBUILD, update the release number, and build it.

sl1pkn07 commented on 2010-09-22 16:12

http://build.chromium.org/buildbot/official/chromium-7.0.530.0.tar.bz2


yaes, is out of date

Malix commented on 2010-09-16 06:23

@FredBezies
this package follows the dev-channel releases, last released is what the pkg offers (see: http://googlechromereleases.blogspot.com/2010/09/dev-channel-update.html).
so not out of date.

FredBezies commented on 2010-09-16 06:20

7.0.524.0 is the last one available. Please update.

Anonymous comment on 2010-09-14 19:55

yay, i have managed to build it but youtube html5 didn't work (BSOD: http://i.imgur.com/9iPhF.png)

@hokapoka: Thanks man, like you said it took +/- 1.2 GB hard disk and 3 hours to build (P4, 768 MB RAM)

Anonymous comment on 2010-09-14 11:16

Mikhail

Many thanks for this AUR, works a treat on x86 and x64.

I only had one issue, the md5sum wasn't correct. I tried 2 different downloads of the tarball from the buildbot repos and neither was the same as in the PKBUILD.

Had to use : c2b142a835eed3ed7fa382612f3d9b4a

It appears that the DateTime stamp on the Tarball is later than the last update of the package, maybe they rebuilt it after you created the package.

Again, many thanks.

Anonymous comment on 2010-09-14 11:12

bintang,

eh? 10Gb... where did you get that size from?

I've just checked my "chromium-dev" build folder and it was 1.2 Gb

If you want I can send you a copy of my "chromium-dev-7.0.517.0-1-i686.pkg.tar.xz", it just finished building on my old P4 took ~3 hours.

I've never tried to install AURs that have been built on other machines before, so I can't be sure it'll work, you're very welcome to try it. Let me know if you want a link to it.





Anonymous comment on 2010-09-14 09:34

@hokapoka, alright i'll try ^_^
But do i really need 10 GB hard disk space? I think this is my biggest concern, only 2 gigs left ^_^

Anonymous comment on 2010-09-14 09:19

bintang, just fire it off overnight regardless how slow your box is it should be complete by the morning, at least it does on one of my old (>5 years old) P4's.

That being said there is a another package that appears to have a binary version available, but I had issues some dependencies.

It was this one, http://aur.archlinux.org/packages.php?ID=27031, currently flagged out of date. Personally I'd prefer to use this (chromium-dev) from source over the that binary AUR.

HTH

Anonymous comment on 2010-09-13 22:30

Would anyone share their i686 build (or maybe set up a repo.. hehe)? My computer is too slow.

Malix commented on 2010-09-12 20:24

@dcelasun
I'd guess it took roughly 20 minutes or so on a phenom2 945, using MAKEFLAGS="-j4" in /etc/makepkg.conf

dcelasun commented on 2010-09-12 20:20

Quick question, how long did it take for you to compile this? On what hardware? It's been a while since I started it and I wonder if, somewhere, I did something wrong.

thermionix commented on 2010-09-11 00:42

Just modified the pkgbuild, don't know if anyones interested. This re-enables phantom tabs, a great feature which was removed recently

http://pastebin.com/AaWMx1qX -pkgbuild
http://pastebin.com/uiAAc5PV -patch

Anonymous comment on 2010-09-05 14:34

@joint007
Try to upgrade gcc to 4.5.1-6 and rebuild the package:
$ rm -rf src
$ makepkg -s -f
I also recommend to disable custom CFLAGS/CXXFLAGS if you are using any.

Anonymous comment on 2010-09-05 14:23

@mvorozhtsov
This is second time build for this version chromium.
gcc version is 4.5.0-6.

Anonymous comment on 2010-09-05 13:54

@joint007
Looks like http://code.google.com/p/chromium/issues/detail?id=52748 but that bug is already fixed in the current version. What version of gcc are you using?
Your log is rather short, did you build from scratch (i.e. no src dir)?

Anonymous comment on 2010-09-05 12:18

Hello, i have a problem: http://pastebin.com/jZwZYnsg
Ideas?(sorry for my very bad english)

jackoneill commented on 2010-09-01 08:26

@ViOLO:
I see, thanks.

Vi0L0 commented on 2010-09-01 01:10

@cantabile:
i tested only with .509, but Gillian00 said it's working fine with .503 : https://bbs.archlinux.org/viewtopic.php?pid=818805#p818805

jackoneill commented on 2010-08-31 20:13

@ViOLO:
In that bbs post you offer a source tarball for version 7.0.508; is it necessary to use _that_ version (or higher) or will this one (7.0.503) work as well?

Malix commented on 2010-08-30 21:37

@Vi0L0
thanks for the link, I tried experimenting with the commandline switches but managed to trigger some ridiculous memoryleaks. I probably was doing strong overdose with them.
anyway, now works like a charm. fluid youtube-video among other things <3

Vi0L0 commented on 2010-08-30 21:04

GPU acceleration in newest chromium works great. At least with catalyst driver, i guess it will also work great with nvidia one.
More here: https://bbs.archlinux.org/viewtopic.php?pid=818410#p818410

baghera commented on 2010-08-19 17:46

@JASN
If you look a few comments below, you can see that mvorozhtsov updates this package following official release of dev channel (or beta if newer).
You can see official release here http://googlechromereleases.blogspot.com/

Version 7.0.497.0 doesn't build because that tarball is automatically generated by buildbot, even if trunk is broken.

Anonymous comment on 2010-08-19 08:29

New version 7.0.497.0 is out, although it didn't build for me with that PKGBUILD

Anonymous comment on 2010-08-15 06:39

Updated, the issue with make should be fixed now.
See http://code.google.com/p/chromium/issues/detail?id=51288

stricte commented on 2010-08-14 22:08

the same with "chrome/common_constants.target.mk:6: *** brakujący separator. Stop." on x86_64

Anonymous comment on 2010-08-14 14:48

@gaunt
I'm on x86_64 myself, cannot reproduce it. Try to remove the src dir. Is anyone else having this issue?

Anonymous comment on 2010-08-14 14:36

As of this timestamp, (490.1) doesn't build on x86_64 with the error:


chrome/common_constants.target.mk:6: *** missing separator. Stop.
Aborting...

Anonymous comment on 2010-08-12 13:18

@Malix: I package only officially promoted (via one of the Stable/Beta/Dev channels) versions.

Malix commented on 2010-08-12 13:07

Just out of curiosity, there seems to be chromium-6.0.491.0 sources at http://build.chromium.org/buildbot/official/
why did you update this to .472.33 instead of .491.0 ?

baghera commented on 2010-08-12 10:33

Now beta channel have higher version than dev channel (6.0.472.33 vs 6.0.472.25), will you update this package or stick with this version?

jackoneill commented on 2010-08-12 09:23

Thank you mvorozhtsov.

Anonymous comment on 2010-08-12 09:20

@cantabile
1) yes
2) no, the folders in $HOME are different.

jackoneill commented on 2010-08-12 08:00

Hi!

Two questions:
1 - if my cpu has sse2, can I use disable_sse2=0 ?

2 - an earlier comment says "This package can be installed alongside with the chromium package from [extra]." I assume that means there are no file conflicts. But what about the user's profile? Do they both use the same profile? The same folder(s) in my home?

baghera commented on 2010-08-09 13:32

flex is part of base-devel group, so it should _not_ be listed as makedepend, that's the purpose of base-devel. If not, also gcc, make, fakeroot and others should be listed as makedepends.

markuman commented on 2010-08-09 13:29

flex is a make dependecie

baghera commented on 2010-07-29 11:58

Hi, I flagged out of date since 6.0.472.11 is out.

Can you also move libgnome-keyring to makedepends/optdepends please?

baghera commented on 2010-07-29 11:57

Hi, I flagged out of date since 6.0.472.11 is out.

Can you also move libgnome-keyring to optdepends please?

xenoterracide commented on 2010-07-26 10:15

yeah that worked down to 57 MB thanks

xenoterracide commented on 2010-07-26 10:02

oh didn't think about strip. I'd turned that off to try debugging some other less stable package. I'll try rebuilding with it on.

Anonymous comment on 2010-07-26 09:58

The problem is that the package doesn't get stripped.
Possible causes include:
- DEBUG=yes _at the time of building_
- nostrip in options=(...) (possibly via /etc/makepkg.conf)
- Failure of actual invocation of strip(1) (not enough disk space?)
- Some other magic

xenoterracide commented on 2010-07-26 09:42

-rwxr-xr-x 1 root root 743860615 Jul 24 09:54 /usr/lib/chromium-dev/chromium

xenoterracide commented on 2010-07-26 09:42

I don't have DEBUG enabled

Anonymous comment on 2010-07-26 09:35

What does 'ls -l /usr/lib/chromium-dev/chromium' print?

xenoterracide commented on 2010-07-26 09:28

I don't have DEBUG enabled

Anonymous comment on 2010-07-26 09:08

Check your environment. If the environment contains DEBUG=yes, then a debug version of the package will be built.

xenoterracide commented on 2010-07-26 08:58

well in my case... Installed Size : 738648.00 K wtf

Anonymous comment on 2010-07-26 06:18

@xenoterracide
$ pacman -Qi chromium-dev | grep Size
Installed Size : 59132.00 K

xenoterracide commented on 2010-07-24 14:13

it says 700MB install is that right? that doesn't seem right? also why install as -dev instead of just providing the regular chromium

OttoA commented on 2010-07-11 15:15

@baghera: it seems as if it is only bundled with google chrome at the moment. It is bundled with the google-chrome-dev package

baghera commented on 2010-07-04 21:57

In 6.0.453.1 log I see

* The PDF plugin is now available (turn it on in about:plugins).

but I can't find it in about:plugins. Missing file in package maybe?

baghera commented on 2010-07-02 19:08

@ Limoto
As you can see in the link you posted, there are at least 2 releases a day, but those tarballs are not intended to be all packaged (they are generated by buildbot). If you check http://googlechromereleases.blogspot.com/ you can see which of those tarballs are actually official releases, intended as tarball with code in a consistent state, and no breakage (apart from known issues).

Limoto commented on 2010-07-02 16:00

http://build.chromium.org/buildbot/official/ and scrool down ;-)

baghera commented on 2010-06-22 23:56

To the one who flagged this out of date: where is the new release? http://googlechromereleases.blogspot.com/ still reports 6.0.437.3 as the last release.

sl1pkn07 commented on 2010-06-20 15:50

wow. i dont touch nothing.

now compile ok! :S

(without change pkgbuild/makepkg.conf)
(with gcc from testing)

sl1pkn07 commented on 2010-06-20 13:41

@mvorozhtsov

CXXFLAGS="-march=x86-64 -mtune=generic -O2 -pipe"

not have -Werror in this config

Anonymous comment on 2010-06-20 12:56

@gaelic: Didn't check that - but also didn't change the PKGBUILD at all. So i'd say the PKGBUILD is missing the dependency for its configuration.

Anonymous comment on 2010-06-20 11:43

Good PKGBUILD!

polslinux commented on 2010-06-20 08:40

there's a new release

Anonymous comment on 2010-06-20 06:19

@sl1pkn07: You got -Werror in your CXXFLAGS. Check your /etc/makepkg.conf.

sl1pkn07 commented on 2010-06-19 20:36

CXX(target) out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o
cc1plus: warnings being treated as errors
webkit/glue/plugins/pepper_device_context_2d.cc: En la función ‘PP_Resource pepper::<sinnombre>::Create(PP_Module, int32_t, int32_t, bool)’:
webkit/glue/plugins/pepper_device_context_2d.cc:69:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
webkit/glue/plugins/pepper_device_context_2d.cc:73:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
make: *** [out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o] Error 1


on arch64bits

sl1pkn07 commented on 2010-06-19 20:34

CXX(target) out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o
cc1plus: warnings being treated as errors
webkit/glue/plugins/pepper_device_context_2d.cc: En la función ‘PP_Resource pepper::<sinnombre>::Create(PP_Module, int32_t, int32_t, bool)’:
webkit/glue/plugins/pepper_device_context_2d.cc:69:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
webkit/glue/plugins/pepper_device_context_2d.cc:73:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
make: *** [out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o] Error 1

sl1pkn07 commented on 2010-06-19 20:33

CXX(target) out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o
cc1plus: warnings being treated as errors
webkit/glue/plugins/pepper_device_context_2d.cc: En la función ‘PP_Resource pepper::<sinnombre>::Create(PP_Module, int32_t, int32_t, bool)’:
webkit/glue/plugins/pepper_device_context_2d.cc:69:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
webkit/glue/plugins/pepper_device_context_2d.cc:73:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
make: *** [out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o] Error 1

sl1pkn07 commented on 2010-06-19 20:31

CXX(target) out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o
cc1plus: warnings being treated as errors
webkit/glue/plugins/pepper_device_context_2d.cc: En la función ‘PP_Resource pepper::<sinnombre>::Create(PP_Module, int32_t, int32_t, bool)’:
webkit/glue/plugins/pepper_device_context_2d.cc:69:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
webkit/glue/plugins/pepper_device_context_2d.cc:73:12: error: se convierte al tipo ‘long int’ que no es puntero desde NULL
make: *** [out/Release/obj.target/glue/webkit/glue/plugins/pepper_device_context_2d.o] Error 1

gaelic commented on 2010-06-19 18:43

hmm. i think libgnome-keyring is only optional, just like kwallet.

Anonymous comment on 2010-06-19 13:38

The PKGBUILD is missing the dependency libgnome-keyring and fails in the buildprocess due to gnome-keyring-1.pc not being found.

Thanks for the packagebuild anyway!

Anonymous comment on 2010-06-19 08:46

6.0.437.1-1: Switched to the system ffmpeg.

gaelic commented on 2010-06-17 14:49

Anyone enabled kwallet support in chromium?

http://code.google.com/p/chromium/issues/detail?id=12351

gaelic commented on 2010-06-17 14:48

@mvorozhtsov:

thx. my /usr/lib/pkgconfig was missing a lot of pc files (and i don't know why)
I did a pacman -Ql gtk2 and saw the .pc files of the package, but as mentioned no .pc files in the config_path.

A reinstall did it, but not only gkt, also cairo, pango, glib, etc.

jinks commented on 2010-06-17 06:12

Another question: Since the new ffmpeg came out today with WebM support, can we use_system_ffmpeg=1 now?

jinks commented on 2010-06-17 06:11

If you want to play back h264, you have to add
"proprietary_codecs=1"
to the GYP_DEFINES variable.

Also, 6.0.437.1 is out.

Anonymous comment on 2010-06-17 04:37

@gaelic: There is no "devel" packages on Arch, gtk+-2.0.pc is a part of the gtk2 package (just do $ pacman -Ql gtk2 to see all the contents). Perhaps your environment variables are screwed (PKG_CONFIG_PATH and PKG_CONFIG_* in general).

gaelic commented on 2010-06-16 09:01

I have problems installing because pkg-config is unable to find gtk+-2.0.pc and gconf***.pc. But I don't see any devel packages for gtk+2? Any suggestions?

baghera commented on 2010-06-11 20:48

All packages in base-devel should NOT be listed as (make)depends, base-devel has been created for this reason, it's a requirement for those who want to compile something on arch.

daneel commented on 2010-06-11 20:30

Please add "flex" and "bison" (core/base-devel) to makedepends. Thanks.

daneel commented on 2010-06-11 20:28

Please add "flex" (core/base-devel) to makedepends. Thanks.

baghera commented on 2010-06-11 13:30

Just add

export CC=/usr/bin/gcc-4.4
export CXX=/usr/bin/g++-4.4

before GYP_DEFINES and it should work properly.

sl1pkn07 commented on 2010-06-11 13:28

oks. thanks.

but what option edit the on the pkgbuild for use gcc44?

sl1pkn07 commented on 2010-06-11 13:22

oks. thanks XD

i have install gcc44 from aur (before compiling this package). how to use it?

baghera commented on 2010-06-11 13:15

You can:
1. Downgrade gcc , gcc-libs , mpfr and libmpc to core version
or
2. install gcc44 from aur and build chromium-dev using that gcc version
or
3. Wait a few hours for gcc 4.5.0-6 to hit testing :) Allan is gonna fix this soon, commits for 4.5.0-6 are in trunk yet (see http://mailman.archlinux.org/pipermail/arch-commits/2010-June/091885.html )

baghera commented on 2010-06-11 13:14

You can:
1. Downgrade gcc , gcc-libs , mpfr and libmpc to core version
or
2. install gcc44 from aur and build chromium-dev using that gcc version
or
3. Wait a few hours for gcc 4.5.0-6 to hit testing :) Allan is gonna fix this soon, commits for 4.5.0-6 are in trunk yet.

sl1pkn07 commented on 2010-06-11 13:09

yes., i'm use testing

howto fix without uninstall testing?

thanks

sl1pkn07 commented on 2010-06-11 13:07

6.0.427 make me all "¡Oh, no!" pages

baghera commented on 2010-06-11 12:52

@ sl1pkn07
Are you using testing? gcc 4.5.0-5 in testing is broken atm , see http://mailman.archlinux.org/pipermail/arch-dev-public/2010-June/017066.html

sl1pkn07 commented on 2010-06-11 12:50

6.0.427 make me all "¡Oh, no!" pages

baghera commented on 2010-05-28 19:31

I also tried to build this package (and -beta) with gcc-4.5.0-3 from [testing], and I got the same results, on x86_64 as well. I didn't try on i686.

Anonymous comment on 2010-05-28 15:03

The last gcc snapshot that got into [core] (as 4.5.0-2) makes chromium go OOM on any page. At least on x86_64 with -O2 (Debug version works fine). Is it the case for i686? For now I added gcc44 to makedeps on x86_64.

baghera commented on 2010-05-28 11:15

Hi mvorozhtsov,
I just tried to build this package in a clean chroot, and during sources checkout I got an error about missing git executable. Adding git to makedeps and rebuilding fixed the issue.

Anonymous comment on 2010-05-23 04:58

why is it SO LARGE???, with my connection this is painful PAINFUL

soupault commented on 2010-05-17 14:07

when this version will be moved to [extra]?

Anonymous comment on 2010-05-17 09:04

This package can be installed alongside with the chromium package from [extra].