• https://commondatastorage.googleapis.com/chromium-browser-official/ (Download)

• component
• debug
• lto

• (2022-08-03) Updated to version: chromium-104.0.5112.79
• (2022-07-20) Updated to version: chromium-103.0.5060.134
• (2022-07-07) Updated to version: chromium-103.0.5060.114
• (2022-06-23) Updated to version: chromium-103.0.5060.53
• (2022-06-10) Updated to version: chromium-102.0.5005.115
• (2022-05-25) Updated to version: chromium-102.0.5005.61
• (2022-05-14) Updated to version: chromium-101.0.4951.67
• (2022-05-11) Updated to version: chromium-101.0.4951.64
• (2022-05-03) Updated to version: chromium-101.0.4951.54
• (2022-04-27) Updated to version: chromium-101.0.4951.41

Log message:
update file path watcher code to match the kqueue implementation
and disable some FD passing code that is not used on OpenBSD

Log message:
start taming chromium in low FD situations; starting with disk cache
cap the file descriptor limit at 85% of the size of the
file descriptor table and also substract the amount of
currently used file descriptors as this should give us
enough reserve to avoid hitting the limit

Log message:
udpate to 96.0.4664.45

Log message:
chromium/iridium: allow parallel build on i386 again, it is currently
working without going up in flames on the build cluster (avail mem =
2205MB thanks to the annoying memory layout on newer intels) and
seems to save some time.

Log message:
remove python3 symlink and append to PATH instead of overwriting it
so that /usr/local/bin/python3 is picked up

Log message:
use the use_allocator_shim build flag instead of a patch

Log message:
Revert previous python3 symlink removal
The build process runs "env python3" with PATH=${WRKDIR}/bin in some places
so /usr/local/bin/python3 won't be found.
Spotted by naddy and phessler

Log message:
Do not symlink python3 anymore, it is provided by the default package
Symlinking the unversioned `python` executable is fine.
These two are the only exceptions in the ports tree handling a versioned
symlink.
OK robert

Log message:
add a missing ifdef to have DocumentAnalysis support

Log message:
update to 95.0.4638.69

Log message:
unbreak vulkan support by adding the required pledges and unveil paths
in the gpu process

Log message:
remove last python2 tentacles from chromium
robert@ switched chromium from python2 to python3 in rev 1.578. This diff
also switches to using the python3 flavor of py-xcbgen to get rid of
chromium's last dependency on python2.
While here, also drop MODPY_VERSION since it's already set to python3 by
default.
ok robert@ (MAINTAINER)

Log message:
update to 95.0.4638.54

Log message:
udpate to 94.0.4606.81

Log message:
update to 94.0.4606.71

Log message:
update to 93.0.4577.82

Log message:
Add unveil for ~/.config/fcitx to make the input method work when
using fcitx.
Reported by Yifei Zhan.
ok sthen@

Log message:
regen ffmpeg config on arm64

Log message:
unbreak on i386

Log message:
include two missing functions in libvulkan.so on OpenBSD

Log message:
update to 92.0.4515.131

Log message:
update to 92.0.4515.107

Log message:
update to 91.0.4472.164

Log message:
update to 91.0.4472.124

Log message:
update to 91.0.4472.114

Log message:
update to 91.0.4472.106

Log message:
update to 91.0.4472.101

Log message:
remove ports libexecinfo, we now have a libunwind-based libexecinfo in base

Log message:
chromium, iridium: removes RTLD_NODELETE removal patches
ok robert@

Log message:
chromium: fix managed prefs (/etc/chromium/policies/managed/prefs.json)

Log message:
chromium: fix managed prefs (/etc/chromium/policies/managed/prefs.json)
ok robert

Log message:
update to 91.0.4472.77

Log message:
disable on i386: mozilla-related ports appear to run out of memory in
rust compilation; chromium has a "reference to type XX could not bind
to an lvalue of type YY" error

Log message:
do not set an mmap(2) hint on OpenBSD; this also unbreaks chromium on arm64
where the addresses were bigger than VM_MAXUSER_ADDRESS

Log message:
Fix missing LIB_DEPEND and update some WANTLIBS. ok robert@

Log message:
update to 90.0.4430.212

Log message:
mark these as broken on powerpc64. while there, fix the COMMENT for electron.
ok robert@

Log message:
implement GetMacAddresses() based on the FreeBSD implementation and
add libvulkan.so to the PLIST

Log message:
unbreak build with llvm11

Log message:
update to 90.0.4430.93

Log message:
update to 90.0.4430.85 and add an flock pledge to the utility process

Log message:
update to 90.0.4430.72; ok naddy@

Log message:
security update to 89.0.4389.128; ok sthen@

Log message:
remove obsolete patch

Log message:
unbreak multicast and getnetworklist

Log message:
mark BROKEN-i386

Log message:
update to 89.0.4389.114

Log message:
update to 89.0.4389.90

Log message:
chromium: put back "DPB_PROPERTIES=lonesome" and get rid of "parallel" for i386.
I was trying to avoid this as only some parts of the build use large
amounts of memory and the build is so slow that it would really help to
have 2 build jobs, but it's causing the bulk build machines to swap too
much, which causes clocks to go out of sync (11 minutes in my current
build), which breaks all the KF5 things built on that machine afterwards
because cmake does stupid time checks on files installed in /usr/local
not just in the build directory.

Log message:
reenable chromium on i386 following MAXTSIZ changes in base

Log message:
revert accidental reintroduction of carriage returns in patch file

Log message:
Allow OpenBSD's mmap to select a random address. Fixes aarch64 at least.
ok robert@ phessler@ sthen@

Log message:
Remove google_default_client_id and google_default_client_secret from
GN_ARGS.
From 2013 until early 2021, Google permitted distribution builds of
Chromium to access Google APIs that added significant features to
Chromium including, but not limited to, Sync and geolocation.
As of March 15, 2021, any Chromium builds which pass client_id and/or
client_secret during build will prevent end-users from signing into their
Google account.
Hint:
If you wish to use sync, you have to build your own chromium package
and use the publicly available client_id and client_secret of the Chrome
browser.

Log message:
remove carriage returns ('\r') from patch files and set FIX_CRLF_FILES

Log message:
update to 88.0.4324.190

Log message:
Reverse the polarity of MODPY_VERSION; default is now 3.x,
if a port needs 2.x then set MODPY_VERSION=${MODPY_DEFAULT_VERSION_2}.
This commit doesn't change any versions currently used; it may be that
some ports have MODPY_DEFAULT_VERSION_2 but don't require it, those
should be cleaned up in the course of updating ports where possible.
Python module ports providing py3-* packages should still use
FLAVOR=python3 so that we don't have a mixture of dependencies some
using ${MODPY_FLAVOR} and others not.

Log message:
Back-out USB data toggle fix for HID devices, since we received multiple
reports about broken devices, e.g. for ukbd(4) and fido(4).
ok ajacoutot@

Log message:
update to 88.0.4324.150

Log message:
www/chromium remove terrible_ping_kludge since kernel is fixed
ok mglocker@

Log message:
unbreak usrsctp by taking the proper codepath on openbsd
also makes google stadia work

Log message:
reinstate BROKEN-i386 again, it will not run so it is useless to waste
30 hours building it

Log message:
update to 88.0.4324.96

Log message:
now that we have struct ip_mreqn, start using that to avoid multiple
multicast errors

Log message:
update to 87.0.4280.141

Log message:
unbreak policies once again by re-introducing a patch that was removed
in a previous commit

Log message:
put back BROKEN-i386 in chromium and add it to iridium, the binaries
cannot be loaded, they immediately exit:
: sthen@i386.p; chrome
/usr/local/bin/chrome[60]: /usr/local/chrome/chrome: Cannot allocate memory
: sthen@i386.p; size /usr/local/chrome/chrome
text	data	bss	dec	hex
180104608	3805736	1949616	185859960	b13ff78
exiting happens very early, before ld.so loads libraries:
: sthen@i386.p; LD_DEBUG=1 /usr/local/chrome/chrome
ksh: /usr/local/chrome/chrome: Cannot allocate memory
no point wasting 2x24h in build for these two.

Log message:
update to 87.0.4280.88

Log message:
update to 87.0.4280.66

Log message:
make it possible for gn to determine its own executable path so that
regenerating rules with ninja is possible

Log message:
unbreak xcursor theme handling by using XcursorLibraryPath() to determine
our own path for the icons instead of using the hardcoded linux paths
tested and okidoki aja@

Log message:
update to 86.0.4240.198

Log message:
update to 86.0.4240.193

Log message:
update to 86.0.4240.185

Log message:
chromium: disable build on i386, the binary cannot be execve()'d, it
immediately returns ENOMEM so there's no point wasting 26h in a build
for it
2254 sh       NAMI  "/usr/local/chrome/chrome"
2254 sh       RET   execve -1 errno 12 Cannot allocate memory

Log message:
change a linux ifdef to include bsd as well for consistency, no functional change

Log message:
update to 86.0.4240.111

Log message:
update to 85.0.4183.121

Log message:
update to 84.0.4147.135

Log message:
update to 84.0.4147.125

Log message:
check if the ENABLE_WASM environment variable is set and do not disable
webassembly if it is; this makes zoom.us (evil) work

Log message:
install missing unveil and pledge files for the video process

Log message:
unbreak video recording now that it uses its own utility process by moving
all the unveil and pledge magic to the proper place

Log message:
update to 84.0.4147.105

Log message:
unbreak build again by not including my own pobj path

Log message:
update to 84.0.4147.89

Log message:
Fix README --disable-unveil flag
ok robert@

Log message:
update to 83.0.4103.116

Log message:
unbreak sync to google and add some missing OS_BSD ifdefs to the viz and
autofill code
sync was broken because we had the a function neutered that pulled hardware
information (manufacturer and model) which was required for sync to work.
this commit implements this function but sends dummy information about the
hardware: manufacturer is OpenBSD and the model is an empty string

Log message:
Migrate off base::ThreadPool()-as-a-trait in /services/device
https://source.chromium.org/chromium/chromium/src/+/6ddec2fbc4e0f917b5dc257184089fdc402d849c
Fixes SEGV when using FIDO keys.
ok robert@

Log message:
update to 83.0.4103.106 and unbreak the debug flavor which was broken
because of an accidentally removed patch to to build/config/compiler/BUILD.gn

Log message:
bump after libxslt dependency changed

Log message:
update to 83.0.4103.97

Log message:
disable the memory dump manager

Log message:
update to 83.0.4103.61

Log message:
last minute update to 81.0.4044.138 which includes 3 security fixes:
https://chromereleases.googleblog.com/2020/05/stable-channel-update-for-desktop.html
ok naddy@

Log message:
re-introduce a debug PSEUDO_FLAVOR so that the regular chromium build can
be used to build DEBUG_PACKAGES as well

Log message:
Remove left-over from the electron split
"go ahead" robert

Log message:
regen plist, spotted by aja@

Log message:
chromium: tweaks to the readme, ok espie robert

Log message:
document some specificities of chromium on openbsd
okay robert@

Log message:
update to 81.0.4044.129

Log message:
remove the audio pledge from the audio process as direct audio device access it not possible anymore

Log message:
update to 81.0.4044.122

Log message:
add another missing internal dependency for a mojom rule

Log message:
update to 81.0.4044.113

Log message:
add a missing internal dependency to avoid random build order failures during bulks

Log message:
remove the openbsd specific storage monitor dummy implementation because
chromium ships with its own dummy implementation that we can use

Log message:
switch back to using the internal freetype library as it fixes some website
font rendering issues
switching back is now possible because a bug has been fixed in freetype:
commit 5b1a37af563349172230842fcfdb59b5f22afcd0
Author: Dominik R_ttsches <drott@chromium.org>
Date:   Tue Feb 25 11:08:11 2020 +0200
Fix for CFF space glyph regression (#57541).

Log message:
update to 81.0.4044.92;
this commit also gets rid of an rpath pledge in the renderer
process by initializing fontconfig fontconfig

Log message:
work around an shmget(2) called and fall back to the XImage API
this should make screen sharing work with most of the online chat systems

Log message:
get rid of more setpriority(2) to avoid pledge violations

Log message:
update to 80.0.3987.163

Log message:
update to chromium-80.0.3987.162;
enable widevine support while here so that Netflix can be watched

Log message:
neuter dwz preventively since it doesn't work with those 4 ports
(it's likely the debug info is bogus)

Log message:
update to 80.0.3987.149 and apply the following changes:
- re-enable pointer compression (it is not the cause of crashes)
- use system harfbuzz and freetype
- add a debug package for the component flavor so that backtraces
are useful
- remove the dependency on java by disabling the closure compiler
and other UI "optimizations" using javascript

Log message:
revert "re-enabling pointer compression", all hell broke loose.
okay robert@

Log message:
zap electron cruft

Log message:
re-enable pointer compression in v8 now that the new RLIMIT_DATA accounting
has been commited

Log message:
update to 80.0.3987.132 and fix the component flavor while here

Log message:
electron still installs natives_blob.bin, so let's just revive it

Log message:
electron can still use the jumbo build which was removed from chromium

Log message:
update to 80.0.3987.116;
disable pointer compression in v8 and jit performance diagnostics until
some things in base are ironed out

Log message:
merge two upstream commits that should fix the issue where a long directory
listing eats up the CPU while the tab is sluggish or stuck
https://chromium.googlesource.com/chromium/src.git/+/37078f9468a5a3ca90a29bcf2bd714824ae90a3f
https://chromium.googlesource.com/chromium/src/+/cc007f73137623795e01102fe6175a39e4d18b81

Log message:
update to 79.0.3945.130

Log message:
MFC chromium updates:
- unbreak chrome://system by not trying to get the amount of used file
descriptors through the metrics code so that a pledge violation is
avoided if the sandbox is enabled (which is the default)
- update to 79.0.3945.117
(skipping U2F because it needs libfido2 which wasn't in base in 6.6)

Log message:
update to 79.0.3945.117

Log message:
avoid a crash when a fido device is being unplugged in the middle of
being accessed by handling close() the same way it is done on other
posix systems
bump the electron package as well to unbreak package builds
from Greg Steuck

Log message:
add support for fido(4) devices; work done by Greg Steuck

Log message:
unbreak chrome://system by not trying to get the amount of used file
descriptors through the metrics code so that a pledge violation is
avoided if the sandbox is enabled (which is the default)

Log message:
MFC updates to chromium-79.0.3945.88 / electron-6.0.3p3 (based on 76.0.3809.132)

Log message:
update to 79.0.3945.88

Log message:
Missed bump on electron.
WANTLIB += iconv lzma

Log message:
merge upstream commit d3a31888bd978cb1f6da614fa25f668d2cc360b8 which
allows us to disable MIT-SHM extension for certain environments and
because pledge(2) does not allow shmget(2) for now we completely
disable this by default

Log message:
use KERN_SHMINFO_SHMMAX to get the value of shmmax because IPC_INFO cmd for
shmctl(2) is not available on OpenBSD

Log message:
update to 79.0.3945.79

Log message:
update to 78.0.3904.106

Log message:
update to 78.0.3904.97

Log message:
merge upstream commit to silence missing font warnings:
commit 2b580e805103141bb1d9da1e120db17911a14c49
Author: Dominik R_ttsches <drott@chromium.org>
Date:   Mon Sep 30 19:42:18 2019 +0000
Remove noisy error log for failing to find font fallback

Log message:
update to 78.0.3904.87

Log message:
oops, unbreak - I put the FLAVOR:Melectron REVISION block too high

Log message:
bump REVISION for electron, WANTLIB was changed (lost uuid).

Log message:
update to 77.0.3865.120 and make sure to use HW_NCPUONLINE instead of HW_NCPU

Log message:
swiftshader is x86 only. Add PFRAG for electron too.
okay naddy@ robert@

Log message:
the network process needs a unix pledge for inter process communication
as well

Log message:
make sure that we support the /etc/chromium/policies directory by actually
enabling the code and disable the usage of the internal fontconfig which
was periodically reloaded causing fontconfig related pledge violations

Log message:
fixup component flavor's plist

Log message:
allow the audio process to create the sndio cookie

Log message:
remove the id pledge from the renderer process and disable the code in
chromium that changes the process priority on the fly because as per POSIX,
lowering the priorities can only be done by the superuser anyway

Log message:
for now add a missing id pledge for the renderer process to be able to call
getpriority() and setpriority()

Log message:
add .zip to the ELECTRON_ARCH string which is only used to substitute in
PLISTs (electron/releases/electron-v6.0.3-openbsd-${ELECTRON_ARCH} etc),
this will prevent share/icons/hicolor/64x64 from being accidentally written
as share/icons/hicolor/64${ELECTRON_ARCH}. looks ok to robert@

Log message:
fix bad subst; share/icons/hicolor/64${ELECTRON_ARCH}/apps/chrome.png

Log message:
fixup patches where WRKSRC was substituted

Log message:
update to 77.0.3865.90;
This update enables two new process types, the network and audio processes
which are also pledge'd and unveil'd.
These processes come with their own pledge and unveil config files for
the time being so that we can shrink or extend their rights as needed.

Log message:
rename the electorn distfile to include the chromium version as well

Log message:
update to 76.0.3809.132 and regen electron tarball to reflect the change

Log message:
fix packaging by not hardcoding the architecture of the machine in the plist

Log message:
add build dependency on libnotify

Log message:
unbreak build by not searching for libcares with pkg-config

Log message:
unbreak by not calling kvm becuase pledge does not allow it

Log message:
fix sqlports

Log message:
add an electronjs (https://electronjs.org/) port;
the electronjs port is going to live inside the chromium port,
as a separate flavor, which is called electron (obviously)
because it requires all of our chromium patches as well and
this way we can avoid having to maintain another subdirectory
with hundreds of patches; this is mostly unconventional but
for me this is the best solution (maintainence-wise)
a couple of patches for the actual electron source are shamelessly
stolen from freebsd
this commit also adds a --no-sandbox flag because electron
cannot use unveil and pledge just yet, i will work on that
in the future
basic stuff works, but don't expect anything fancy yet
--
initial versoin of electron-6.0.3:
Build cross platform desktop apps with JavaScript, HTML, and CSS.
It's easier than you think.
If you can build a website, you can build a desktop app. Electron is a
framework for creating native applications with web technologies like
JavaScript, HTML, and CSS. It takes care of the hard parts so you can
focus on the core of your application.

Log message:
update to 76.0.3809.100

Log message:
get rid of "parallel" in DPB_PROPERTIES on i386, just keep the "lonesome"

Log message:
update to 76.0.3809.87

Log message:
Fix confusion in time units breaking audio/video synchronization.
ok robert

Log message:
Fix audio-video synchronisation bug revealed by last commit.
A/V synchronization was caused by confusion between bytes and samples
in the hardware delay calculation. Bug reported and fix tested by
Tobias Kortkamp and robert@, thanks.
ok robert@

Log message:
remove whitespace after revision number

Log message:
disable the network change notifier because it seems to be broken,
so this way the browser will report an online status all the time
and web applications checking the status will work again until the
actual issue with the notifier is fixed

Log message:
update to 75.0.3770.142

Log message:
install vulkan libraries as well

Log message:
Switch to PERMIT_PACKAGE.  CDROM restrictions are no longer applicable.

Log message:
unbreak on arm64

Log message:
unbreak video(4) support by using the proper type for ioctl(2)'s request
argument and add the necessary pledge to the main process and unveil /dev/video

Log message:
Add support for audio recording using sndio.
ok and (a lot) of help from robert.

Log message:
bump REVISION in a bunch of ports known or likely to use struct kinfo_proc,
there may be some missing as my unpacked ports source is a little out of date
but this should catch the main things people might run into
the struct was reordered a second time in sysctl.h r1.192 to improve
compatibility but amd64 snapshot packages made it out before that happened
so the bumps are still needed

Log message:
update to 75.0.3770.100

Log message:
update to 75.0.3770.90

Log message:
sync WANTLIB

Log message:
unbreak build by setting the PATH properly to ${WRKDIR}/bin so that
python can be found by the build and regen patches while here

Log message:
update to 75.0.3770.80;
this brings us back to the stable channel again

Log message:
update to 75.0.3770.75

Log message:
unbreak the component flavor once again

Log message:
update to 75.0.3770.66

Log message:
fix functions to show the proper number of open fds

Log message:
add support for setting the thread and process priorities on bsd

Log message:
update to 75.0.3770.52

Log message:
set png16 WANTLIB entries back to png following the change in soname
part done by portbump, part by hand (it is easily confused)

Log message:
fix the sysctl(2) call for testing if we are running inside gdb
and make the best effort for figuring out our own exectuable name
and fall back to a hardcoded value if that fails

Log message:
just the regular round of OS_BSD define additions for linux code

Log message:
remove the check for a wxallowed filesystem

Log message:
update to 75.0.3770.27;
75.0.3770.27 is still a beta, but it fixes a really bad bug with notifications

Log message:
remove some noisy notimplemented notifications

Log message:
add an upstream fix for https://bugs.chromium.org/p/chromium/issues/detail?id=956061
Issue 956061: Context menu and on top elements draw a black rectangle

Log message:
call XGetErrorDatabaseText() before pledge(2) to have the XErrorDB cached
for future XGetErrorDatabaseText() calls

Log message:
unbreak policies by actually setting the correct policy dir

Log message:
update to 74.0.3729.131 and fix component flavor

Log message:
fix LD_LIBRARY_PATH

Log message:
add some missing OS_BSD defines

Log message:
update to 74.0.3729.108

Log message:
disable wasm (WebAssembly) by default because it is a disaster

Log message:
update to 73.0.3683.103

Log message:
unbreak i386, from robert

Log message:
the main process needs a vminfo pledge so that the internal
resource coordinator can get live data about the amount of
free available memory in high workloads

Log message:
unbreak build on arm64

Log message:
fix missing patches from previous commit

Log message:
upate to 73.0.3683.86

Log message:
minimize the v8 javascript engine memory footprint as much as possible
to not eat up all the memory with sites using lots of javascript

Log message:
add getpw pledge to the main process because when using cups, there will
be getpnam(3) calls by the cups library

Log message:
ignore ENOENT up until the last element of the path when doing mkdir(2)
on unveil(2)'d directory trees ... essentially the same modification that
went into gtk+3
this allows us to remove some hacks from unveil(2)

Log message:
update to 72.0.3626.121

Log message:
do not try to raise limits on-the-fly to avoid breaking the pledge(2)

Log message:
update to 72.0.3626.119

Log message:
be nice and create ~/Downloads if its not there

Log message:
Make sure to always default to the Downloads directory in case a file
dialog is opened and unveil(2) is used.
This fixes the issue when another program not using unveil(2) changes
the last used path of the file dialog and then the file dialog opened
by chrome will get stuck inside a non-existent directory.
To make sure that the ~/Downloads XDG directory is always available
for glib to pick up, the chrome wrapper script will check if the
~/.config/user-dirs.dirs does not exist and create it with one entry
configured: XDG_DOWNLOAD_DIR="$HOME/Downloads"

Log message:
update to 72.0.3626.109

Log message:
update to 72.0.3626.96

Log message:
update to 72.0.3626.81
now chromium uses f'in java to use the closure compiler which
creates "better javascript" by parsing and converting all js
code into unreadable garbage. this feature is only enable on
amd64 because on i386 java randomly fails to allocate the needed
memory and on arm64 we do not have jdk yet. on these architectures,
we use the best method to create "better javascript", which is cat(1)

Log message:
set DPB_PROPERTIES+=lonesome on chromium/iridium for i386 (like is done
for firefox) in the hope that this will reduce build problems. these
ports often end up building together on a machine and they aren't coping
too well.
compiling some individual files is now bumping into my 75 minute DPB
stuck_timeout (possibly the llvm update took it over the edge) - raising
the timeout means that the other ports which occasionally hang during
compilation tie up a build slot for too long.

Log message:
pre-create the .cache/chromium and .local/share/applications directories
so that unveil will not fail on them if they do not exist yet

Log message:
remove error pledge because it is not needed and sync the sandbox code
with linux

Log message:
lock down the unveils even more to avoid having access to the complete
~/.local and ~/.config directories and initialize NSS before unveil so
that the ~/.pki and the subdirectories can be created just in time

Log message:
update to 71.0.3578.98

Log message:
update to 71.0.3578.80

Log message:
fix theme handling by actually showing the necessary buttons in the settings
page

Log message:
add unveil(2) for /usr/bin/lpr to make the system print dialog work when
using lpr for printing

Log message:
Allow text relocations on i386, caused by optimized non-PIC assembly in
the embedded copy of FFmpeg.  Chromium build test by sthen@,
looks good to robert@

Log message:
update to 70.0.3538.110

Log message:
enable unveil(2) by default in chromium;
if you wish to disable it, please run chromium with the --disable-unveil flag

Log message:
update to 70.0.3538.102

Log message:
unbreak i386 build by defining explicit WebVector(size_t size) : data_(size) {}
for non 64-bit arches on OpenBSD as well

Log message:
update to 70.0.3538.77

Log message:
add some missing openbsd ifdefs after the overhaul of the UI

Log message:
disable USE_LLD on i386, it still doesn't work there for these ports:
can't create dynamic relocation R_386_32 against local symbol in
readonly segment; recompile object files with -fPIC

Log message:
set USE_LLD to Yes for chrome to shave at least 30 minutes of build time

Log message:
update to 69.0.3497.100 and enable jumbo build
To improve compilation times it is possible to use Jumbo builds,
in Chromium.
The idea is to merge many source files and compile them together.
Since a large portion of Chromium's code is in shared header files
that dramatically reduces the total amount of work needed.
fix an internal build dependency while here

Log message:
annotate the monster with parallel2.
dpb now has a special feature where parallel2 may be set to use even
more cores than parallel.

Log message:
enable the new chromium UI on openbsd as well

Log message:
fix the swiftshader PKG_ARG on aarch64

Log message:
update to 69.0.3497.92 and fix a missing dependency on e2fsprogs

Log message:
add support for arm64;
This is now possible due to the libcxx update and pmap changes in the kernel.
Since swiftshader is not supported on aarch64 (amd64), create a PFRAG to be
included on anything else besides arm64.

Log message:
update to 69.0.3497.81

Log message:
update-patches strikes again .. do not include my pobj path

Log message:
remove more /sys and /proc access and add required unveil for /etc/machine-id
and some missing pledges for the main process

Log message:
update to 68.0.3440.106 and include all the unveil configuration files
by default now, so there is no need to clone the git repository for that

Log message:
update to 68.0.3440.84

Log message:
update to 68.0.3440.75

Log message:
remove bogous patch

Log message:
Introduce unveil(2) support for chromium;
I am commiting this so that work can continue in tree because the diff
is starting to get really big. This commit includes several changes
required for unveil(2) described below:
- overhaul of the sandboxing code for pledge(2) to match what is being
done on other platforms
- avoid using /dev/urandom and switch to arc4random(3) or arc4random_buf(3)
- start using pledge(2) for the main browser process as well, disabled by
default for now as the list changes rapidly. The list of pledges are read
from the /etc/chromium/pledge.main file if available. When this is complete
the list of pledges will be hardcoded just as it is for the other process
types.
- add the --enable-unveil flag to use unveil(2) in all of the browser processes
by reading the list of unveil'd paths from files located in /etc/chromium,
e.g.: /etc/chromium/unveil.main or /etc/chromium/unveil.gpu.
These files are not included in the package by default as they are work in
progress. If you want to help please visit: https://github.com/rnagy/chromium
- if unveil(2) is used, chromium will not be able to access most of the filesystem
so for example if you would like to download or upload something, only the unveil'd
paths are going to be available, which is by default include ~/Downloads.

Log message:
@tag gtk-update-icon-cache

Log message:
first tag: update-desktop-database

Log message:
67.0.3396.87

Log message:
add tmppath pledge for the gpu process for DRI3

Log message:
unbreak build by not hardcoding my pobj

Log message:
do not link to libatomic as we use clang

Log message:
update to 67.0.3396.62

Log message:
update to 66.0.3359.181

Log message:
update to 66.0.3359.170

Log message:
update to 66.0.3359.139 and remove unnecessary patch

Log message:
fix syntax erron in python script

Log message:
add a flavor called component to allow chromium to be built as a non-static
binary to ease development

Log message:
avoid hitting pledge aborts while running the task manager and
fix a multiple symbol definition

Log message:
update to 66.0.3359.117

Log message:
the utility process needs a sendfd pledge as well in case chrome is running
in complete process separation mode to be able to send back data to the
renderer, for printing to pdf files for example

Log message:
update to 65.0.3325.181

Log message:
update to 65.0.3325.162

Log message:
update to 65.0.3325.146

Log message:
update to 64.0.3282.186

Log message:
update to 64.0.3282.167

Log message:
update to 64.0.3282.140

Log message:
unbreak build by not hardcoding my pobj path

Log message:
bump the v8 max code range size to 512M from 256M

Log message:
update to 64.0.3282.119

Log message:
update to 63.0.3239.132

Log message:
Make Chromecast devices and casting work by doing several different things:
- the media router extension which handles chromecast is only enabled
on official Chrome builds so enable it on chromium as well
- our multicast implementation is a legacy one so patch it around
to make it work and use SO_REUSEADDR and SO_REUSEPORT on the socket
to make other clients able to bind to the same port and address so
that everything will work smoothly together and chromium will not
"block" the multicast ports
- enable a good amount of features that are used in the javascript extension
code that were only enabled for linux and used by the media router
extension
Everything works from youtube casting, desktop casting to browser tab casting.
The only thing that does not work is the actual initial setup of the chromecast
device because that requires connecting to the device's wifi network and the browser
cannot do that on OpenBSD so another device has to be used for that, but I guess
everyone has a smartphone nowadays :)

Log message:
Add an i686 string as well on i386 now that chromium is re-enabled on i386 builds

Log message:
update to 63.0.3239.108 and re-enable on i386 for now

Log message:
unbreak build by adding missing patch for malloc.h

Log message:
use -Wl,--no-keep-memory on i386, linking fails too often without it. ok robert@

Log message:
update to 61.0.3163.79

Log message:
update to 60.0.3112.113

Log message:
update to 60.0.3112.101

Log message:
update to 60.0.3112.90 and add some patching for chromium remote desktop

Log message:
update to 60.0.3112.78

Log message:
bump LIBCXX/LIBECXX/COMPILER_LIBCXX ports.

Log message:
Extend PATCH_LIST with clang-* for base clang builds; this was missed out from
the previous commit.

Log message:
make chromium build with base clang by applying the following changes:
- use the proper compiler and linker in ${WRKDIR}/bin
- casting to nullptr where required (diff from espie@)
- link the character_data_generator with c++abi twice until a linker
bug is figured out

Log message:
use COMPILER_LIBCXX where applicable

Log message:
Backport a fix from chromium git so that DPI settings are honored.
https://bugs.chromium.org/p/chromium/issues/detail?id=732440
https://chromium.googlesource.com/chromium/src/+/b50cab37a165f1f11d9a046d7f3367e4a563a73f

Log message:
update to 59.0.3071.115;
use a hand-crafted tarball this time because google failed to create a
a working one :)

Log message:
do not call malloc and sysctl twice that was introduced by a bad copy & paste
diff from Caspar Schutijser

Log message:
update to 59.0.3071.109

Log message:
allows base clang to pass the bootstrap step, even though there are lots
of issues later

Log message:
switch everything to new COMPILER idiom, even stuff that won't build with clang
yet, but at least that part is done.

Log message:
use the MODGCC4* variables instead of hardcoding dependencies.
convert to WANT_CXX even though it might not work yet.
bump the ports that didn't even the full spec

Log message:
update to 58.0.3029.110

Log message:
update to 58.0.3029.96

Log message:
update to 58.0.3029.81

Log message:
update to 57.0.2987.133

Log message:
update to 57.0.2987.110

Log message:
update to 57.0.2987.98

Log message:
use more linux specific code for font handling and gpu

Log message:
update to 56.0.2924.87

Log message:
update to 56.0.2924.76

Log message:
update to 55.0.2883.87

Log message:
unbreak build

Log message:
update to 55.0.2883.75; add lpr support to the system dialog while here

Log message:
update to 54.0.2840.100

Log message:
update to 54.0.2840.87

Log message:
update to 54.0.2840.71

Log message:
update to 53.0.2785.143

Log message:
update to 53.0.2785.116

Log message:
Unbreak with cups>=2.2.0.
Hopefully catched all breakage since a few ports aren't building currently
(ligtop2...).

Log message:
update to 53.0.2785.101

Log message:
update to 53.0.2785.89

Log message:
use USE_WXNEEDED for chromium/iridium now that it works with gcc4

Log message:
remove unused patch

Log message:
remove some unnecessary patches

Log message:
Move chromium to the new build system (gn - generate ninja)
This commit moves the chromium port to use gn to generate ninja files
which will hopefully resolve the random build errors. Gyp is dead so
this move has to be done anyway.
Also change the way how we handle openbsd. Let's assume we are running
on linux and only handle !linux cases where needed. This also ensures
that a change on linux will for sure end up in our builds as well.

Log message:
add another pre-build target

Log message:
add ax_gen into prebuild

Log message:
update to 52.0.2743.116

Log message:
update to 52.0.2743.82

Log message:
add the required recvfd pledges

Log message:
add another pre-build target

Log message:
update to 51.0.2704.106

Log message:
update to 51.0.2704.103

Log message:
update to 51.0.2704.84

Log message:
since the chrome wrapper script is already doing some checks and warning
the user if they fail, also add a check for wxallowed on the mountpoint,
a message giving a clue is better than "syntax error: `)' unexpected"
ok robert

Log message:
update to 51.0.2704.79

Log message:
Link chromium with -Wl,-z,wxneeded. This has two effects:
1. it will disable the kernel logs about W^X violations.
2. the binary will not run at all unless the filesystem holding
/usr/local has the "wxallowed" flag set (in /etc/fstab), rather
than starting but having tabs crash all the time.
ok robert@

Log message:
add more mojom prebuild targets

Log message:
update to 51.0.2704.63

Log message:
pre-build the autofill_server_proto target as well

Log message:
update to 50.0.2661.102

Log message:
enable webrtc support and fix video recording. the only thing
that's missing from working video conferencing is the sndio
input support, which is being worked on

Log message:
cache the amount of physical memory in the UtilityProcess as well before
the pledge(2) call; issue found by sthen@

Log message:
update to 50.0.2661.86

Log message:
update to 50.0.2661.75

Log message:
update to 49.0.2623.112

Log message:
update to 49.0.2623.110

Log message:
update to 49.0.2623.108

Log message:
oops

Log message:
droid-fonts -> noto-fonts

Log message:
add common_constants to the pre-build list

Log message:
update to 49.0.2623.75

Log message:
Bump REVISION for lang/clang MODULE changes.

Log message:
security update to 48.0.2564.116; ok sthen@

Log message:
security update to 48.0.2564.109; ok sthen@, naddy@

Log message:
update to 48.0.2564.103

Log message:
since we are not using the Zygote model, make sure to initialize the
sandbox ipc channel for the renderer in order to get proper font
handling outside of the sandbox if necessary

Log message:
unbreak build by adding a missing patch

Log message:
Initialize the FcInit library early just like it is being done on Linux

Log message:
update to 48.0.2564.97

Log message:
prebuild a missing dep that I ran into with chromium on a bulk build; ok robert@

Log message:
move to libjpeg_turbo from using the system libjpeg

Log message:
use the ipc channel to replace fonts inside the sandbox that are not
available on the system, this fixes a lot of pdf viewer issues where
the text was missing

Log message:
update to 48.0.2564.82

Log message:
sync WANTLIB

Log message:
add an easily-greppable marker for ports known to use pledge().
it's not so important while they're only in patches in the ports tree,
but for any which get upstreamed it saves a ~100GB grep to find them.

Log message:
update to 47.0.2526.111

Log message:
remove rpath from the GPU process's pledge call now that mesa has been taken care of

Log message:
pledge(2) the chromium processes;
The renderer, gpu, plugin and utility processes are now using pledge(2)
Unfortunately the GPU process only requires an rpath pledge because of
Mesa trying to parse two configuration files, /etc/drirc and ${HOME}/.drirc
So currently the GPU process will use an rpath pledge in the next
week or so so that people can test, but this situation has to be
resolved because it is not acceptable that a mostly unused configuration
file is being parsed from a library and that stops us from using less
pledges and thus disallowing the GPU process to have read access
to the filsystem ... like your ssh keys.

Log message:
add mojo bindings to the pre-build targets

Log message:
update to 47.0.2526.106

Log message:
fix build by adding a missing patch

Log message:
Disable upstream's attempt to use a 64-bit linker for i386 ("ia32") builds
ok robert@

Log message:
remove unused patches up and enable some more functionalty like HiDPI support

Log message:
re-enable pre-build hack for resources

Log message:
update to 47.0.2526.80

Log message:
update to 47.0.2526.73

Log message:
remove devel/flex dep, these are ok with base flex

Log message:
pre-build gfx_vector_icons in chromium+iridium, ok robert@

Log message:
update to 46.0.2490.86

Log message:
update to 46.0.2490.80

Log message:
add fonts/droid-fonts as a run dependency to ease displaying
some special Japanese, Russian and other characters by default

Log message:
update to 46.0.2490.71

Log message:
re-enable the pre-build targets which i accidentally disabled in the previous
commit

Log message:
make v8 pre-allocate less memory on OpenBSD/amd64 to avoid v8 running out
of memory

Log message:
update to 45.0.2454.101

Log message:
update to 45.0.2454.99

Log message:
sync the heap handling code of v8 with linux and regen patches

Log message:
fix typo

Log message:
add cacheinvalidation and cacheinvalidation_proto_cpp to the pre-build targets

Log message:
update to 45.0.2454.93

Log message:
fix build by applying a patch from upstream

Log message:
update to 45.0.2454.85

Log message:
detect *BSD in the javascript code of webui

Log message:
update to 44.0.2403.157

Log message:
update to 44.0.2403.156

Log message:
drop the dependency on libusb in favor of the internal libusb due to
extra functionality implemented in the bundled version

Log message:
update to 44.0.2403.155

Log message:
add net_interfaces to the list of pre-built targets

Log message:
add the suggestions target to the list of pre-built targets

Log message:
update to 44.0.2403.125;
add a hack to build the generated resources to avoid random build failures

Log message:
add a missing patch and remove some bad practice shell code from the Makefile

Log message:
register another missing internal dependency (there are still more, though)

Log message:
add a missing dependency on internal resource

Log message:
update to 44.0.2403.107

Log message:
fix a patch for llvm

Log message:
update to 44.0.2403.89

Log message:
update to 43.0.2357.134

Log message:
allow chromium to restart itself when necessary after asking the user
for confirmation

Log message:
update to 43.0.2357.132

Log message:
update to 43.0.2357.130

Log message:
dependency fix, okay robert@ (testing same with iridium)

Log message:
update to 43.0.2357.125

Log message:
update to 43.0.2357.124

Log message:
add a missing build dependency to avoid ninja failing

Log message:
remove unused MASTER_SITE

Log message:
update to 43.0.2357.81

Log message:
update to 43.0.2357.65

Log message:
Zap bogus '${SUBST_CMD}' line

Log message:
Our nm doesn't support neither '-D' nor '-f', hence some non-critical
errors occur during chromium and iridium builds:
---8<---
[2581/10537] CXX obj/components/policy/core/common/cloud/policy_component_common.cloud_external_data_manager.o
[2582/10537] CXX obj/components/policy/core/common/cloud/policy_component_common.external_policy_data_updater.o
[2583/10537] CXX obj/components/policy/core/common/cloud/policy_component_common.component_cloud_policy_service.o
[2584/10537] SOLINK(module) libffmpegsumo.so
nm: unknown option -- D
usage: nm [-aCegnoprsuw] [file ...]
[8618/16362] STAMP obj/extensions/extensions_browser.actions_depends.stamp
[8619/16362] AR obj/extensions/libextensions_browser.a
nm: unknown option -- D
usage: nm [-aCegnoprsuw] [file ...]
---8<---
Fix these errors by using llvm-nm instead.
While here, also fix a typo into iridium's pkg/DESCR (missing verb).
OK robert@ (maintainer)

Log message:
unbreak i386 and make the debug flavor closer to build

Log message:
update to 42.0.2311.135

Log message:
Mark BROKEN-i386, binutils-2.15 can no longer link this in available ram.

Log message:
update to 42.0.2311.90

Log message:
Bump after the recent changes in packaging dbus and dconf.

Log message:
update to 41.0.2272.118

Log message:
Move fontpath from /usr/local/lib/X11/fonts to /usr/local/share/fonts to match
XDG_DATA_DIR (where Desktop tools will look for by default).
discussed with sthen@ and matthieu@
ok sthen@

Log message:
x11/gtk+2,-guic -> x11/gtk+3,-guic && bump

Log message:
update to 41.0.2272.101 and add missing patches from last commit

Log message:
update to 41.0.2272.89

Log message:
Rework how the GPU vendor and device id is detected by using direct calls
to Mesa. The current solution is based on a diff from jsg@.
Also fix font redering issues with WebKit, with 3rd party TTFs with a
diff from Yoshihisa Matsushita.

Log message:
update to 40.0.2214.115

Log message:
use mesa to get the gpu vendor and device id and pass it to chromium because
we can't use libpci due to permission restrictions and use some more UI code
on BSD as well

Log message:
update to 40.0.2214.114

Log message:
restore PLIST, unbreaks packaging on i386 (insufficient memory to link with
libpdf enabled).

Log message:
update to 40.0.2214.111

Log message:
update to 40.0.2214.93

Log message:
Repair chromium builds on i386, we are bumping into the hard 2GB MAXDSIZ
limit during linking:
- use --no-keep-memory
- disable pdf support on i386
(conditionalize the libpdf.so PLIST entry based on arch)
We are still very close to the limit and I suspect it may still intermittently
fail.
ok robert@

Log message:
add a missing dependency to fix ninja random build

Log message:
include the internal pdf viewer/generator

Log message:
update to 40.0.2214.91

Log message:
update to 39.0.2171.99

Log message:
update to 39.0.2171.95

Log message:
update to 39.0.2171.71

Log message:
update to 39.0.2171.65

Log message:
update to 38.0.2125.122

Log message:
Fix another dependency issue
From upstream git commit 46c567aecd2b80239fee6fc4006474db9c281229
ok robert@ (maintainer)

Log message:
Fix dependency error
ok robert@ (maintainer)

Log message:
update to 38.0.2125.111

Log message:
add a dependency on theme_resources, hopefully this will fix an intermittent
build failure some of us have been seeing with ninja random builds.  ok espie

Log message:
fix build on i386

Log message:
update to 38.0.2125.104

Log message:
update to the next stable release (13.0.782.112)

Log message:
update to 12.0.742.122

Log message:
fix build on 5.0

Log message:
Fix build with png-1.5.
All fixes by yours truly.

Log message:
fix the chrome wrapper script a bit
- skip the tests if you already have a running chrome instance
- default to 'no' (meaning you must not use the mouse for doing the
sensible thing)
- do not start chrome if you just close the window (alt+f4 etc)
ok robert@

Log message:
update to 12.0.742.112

Log message:
Add gnome-icon-theme to RUN_DEPENDS, it's needed to render some icons.
ok robert@ (maintainer)

Log message:
Try to set the number of open files to the desired value before warning
about openfiles-max. ok robert@

Log message:
Use MODPY_VERSION instead of hardcoding 2.6 which is not the case anymore
since we default to python 2.7

Log message:
update to chromium-12.0.742.100

Log message:
fix AtomicWord definition on i386, no bump needed

Log message:
cvs missed this, tsk tsk

Log message:
Update to the new stable branch: 12.0.742.91
http://googlechromereleases.blogspot.com/2011/06/chrome-stable-release.html

Log message:
bump after default python version switch

Log message:
update to chromium-11.0.696.77

Log message:
Instead of reserving 32MB of virtual memory let's reserve the quarter of
RLIMIT_DATA (ulimit -d) if RLIMIT_DATA is not 0 (unlimited).
If RLIMIT_DATA is 0 then the default 512MB reservation will be used.
Using 32MB made v8 use less memory for sure but it was a huge performance
loss too.

Log message:
Make v8 pre-allocate at least 32M of virtual memory to avoid
crashing with vmmap. This is needed because of their JIT is
not fully 64-bit.

Log message:
- re-enable flock in the build
- add support for detecting if an sqlite3 db is on an NFS mount
- fix printing: Chrome can only print to PDF and that is why you
cannot print to LPD directly. It requres CUPS.

Log message:
update to 11.0.696.71

Log message:
Add WIP progress sndio support to chrome, developed by jakemsr@ and
with modifications by me.
It is not working yet so you can only enable it by setting the
CHROME_SNDIO environment variable, so by default nothing changes.
I commit this so that other people can work on it too.

Log message:
Add missing RUN_DEPENDS.
Remove trailing blank line.

Log message:
Add a .desktop file for the menus and install the icons too.
Remove the message about bumping shmall and datasize too.

Log message:
sync transport dib code

Log message:
- use the same memory allocation values on amd64 and i386 too.
this makes chrome use way less memory and since it does not hit
datasize limits anymore, it is way more stable too
- remove the warning of datasize bumping from the wrapper script
tested by sthen@, jasper@ and me

Log message:
deve/gconf2 is about to loose its dependency on x11/gtk+2 so explicitely
register x11/gtk+2 for ports depending on gconf2 that don't have another
dependency that would bring it.
discussed with jasper@

Log message:
update to chromium-11.0.696.68

Log message:
xdg-open is used to start helpers, so add devel/xdg-utils to RUN_DEPENDS.
ok robert@ (maintainer)

Log message:
update to 11.0.696.65

Log message:
update to the latest stable version which is now 11.0.696.60
+ switch from using system sqlite to the bundled one because they have
many modifications that are needed
+ add a new mirror for the distfile because the google one is utterly slow

Log message:
update to 10.0.648.205 and use python 2.6

Log message:
update to 10.0.648.204

Log message:
merge patches from http://codereview.chromium.org/6682026

Log message:
sync with the freebsd transport dib code

Log message:
make sure we depend on kvm>=10.0 otherwise it will fail to build
from Pascal Stumpf <Pascal.Stumpf@cubes.de>

Log message:
update to 10.0.648.151

Log message:
Add some missing patches that were missed by the previous commit
and move back to KERN_PROC from KERN_PROC2 now that we can.

Log message:
update to 10.0.648.133

Log message:
install the manpage, this also fixes chrome --help
ok robert@

Log message:
WANTLIB+=kvm and bump, ok robert@

Log message:
turns out that disable_sse2 is only checked on i386 so remove the check
for amd64

Log message:
- remove LINK=/usr/bin/g++ from make env
- fix an ifdef
- enable SSE2 support on amd64
diff from Matthew Dempsky

Log message:
update to 9.0.597.107

Log message:
no need to use gcc from ports anymore, the pragma patch has been
commited to base

Log message:
Use KERN_PROC2 everywhere instead of trying to use /proc and use
kvm(3) to help chrome identify it's own processes.

Log message:
- SECURITY UPDATE of chromium to 9.0.597.94, fixes about 13 CVE's.
tested by myself on amd64, and espie@ on i386.
"sure, make sure it works" robert@ (MAINTAINER), ok espie@

Log message:
fix MESSAGE, bump

Log message:
tweak wrapper script; use larger default limits for FD and datasize
(you may still need to raise them if you open many tabs, but this
should be enough to avoid problems in normal use), only warn if
softlimits can't be raised, and adjust messages to mention login.conf
(and login.conf.db if detected).
ok espie@ "feel free to modify it if you don't like it" robert@

Log message:
.. and ditch older patches too

Log message:
late update to chromium 9, robert@ did the work, but he doesn't have the
time right now.
Tree is still locked!
Reasons for the update:
- good work that can be used now
- only builds on i386/amd64, and they're fast
- multiple security problems in old chrome
- old chrome half working, can't break it more than it was.
This one requires resources (see MESSAGE), but it's ways more stable.
Again, discussed with sthen@, jasper@, naddy@, miod@

Log message:
Disable on amd64, because it does not work.
After unlock a chrome update will come...
ok ajacoutot@, sthen@ and requested by many people

Log message:
zap stray

Log message:
cut the "monster patches" into small pieces. pval can fix it to his liking
if he comes back.
In the mean time, fix build of chromium by adding the required
-I/usr/local/includes to gyp glue where needed.

Log message:
new depends

Log message:
Use the python MODULE instead of harcoding python version.
This port is still broken though, it missed a -I/usr/local/include somewhere...

Log message:
Remove USE_X11.

Log message:
- fix WANTLIB after cairo update.
(please note chromium is still broken, but will be dealt with soon)

Log message:
- new style depends
- tweak master sites
ok pval@ (MAINTAINER)

Log message:
now that this builds again, bump for the gcc3->gcc4 switch

Log message:
fix build. make sure build includes come before systeme includes, thus
avoiding confusion between built-in libicu and textproc/icu4c

Log message:
minor update to chromium-5.0.359.1

Log message:
being too chummy with modules... fix gcc4 build.

Log message:
Update to chromium-5.0.539.0; tested by many on ports@
- still some issues on some amd64 machines, etc. but a step forward
- see http://sightly.net/peter/openbsd/chromium/ for more

Log message:
While I'm still unsure why gamin is needed at all, at least use a
correct run_depends: we mean gamin, not libgamin.
"go ahead" jasper@

Log message:
gamin RUN_DEPENDS, from speedvin@archlinux.us; ok jasper espie

Log message:
Mention in MESSAGE that this package currently requires SSE2,
and show how to check. This should change sometime, but the
fix is fiddly and for now we should let people know. Committing
now to make sure something is in pre-4.7.
Discussed with pval, this diff ok jasper@ landry@

Log message:
remove extra under-specified stdc++ libdepends.

Log message:
Move some LIB_DEPENDS to WANTLIB.
Regen PLIST.

Log message:
fix build