GitLab

Blink r187605 https://crrev.com/804873004 introduced DEFINE_TRACE macros,
which dedupes trace(Visitor*) and trace(InlinedGlobalMarkingVisitor) by
using template method:
template<typename VisitorDispatcher> void traceImpl(VisitorDispatcher);

This CL updates the GC plugin, so that it:
- recognizes traceImpl method, and mark trace(field) calls made inside.
- recognizes trace method that just delegate to traceImpl, and skip its processing.
  - e.g.) void trace(Visitor* visitor) { traceImpl(visitor); }

BUG=420515

Review URL: https://codereview.chromium.org/824853003

Cr-Commit-Position: refs/heads/master@{#309627}

In the case when an app is inside a framework, app_positions[0] should be
greater than framework_positions[0].

BUG=444343
TBR=inferno@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/819403003

Cr-Commit-Position: refs/heads/master@{#309626}

BUG=444510
TEST=GYP_GENERATORS=ninja tools/gyp/gyp --depth . ui/file_manager/file_manager/foreground/js/metadata/compiled_resources.gyp && ninja -C out/Default

Review URL: https://codereview.chromium.org/824003004

Cr-Commit-Position: refs/heads/master@{#309625}

When no search term are entered, remove the "result-headers" on mobile
as otherwise it looks incorrect. Use :empty CSS pseudo selector.

BUG=440368

Review URL: https://codereview.chromium.org/821703003

Cr-Commit-Position: refs/heads/master@{#309624}

BUG=444485
TEST=GYP_GENERATORS=ninja tools/gyp/gyp --depth . ui/file_manager/video_player/js/compiled_resources.gyp && ninja -C out/Default | grep "caster.js"

Review URL: https://codereview.chromium.org/823583005

Cr-Commit-Position: refs/heads/master@{#309623}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/50a87e2..c22662b

TBR=gavinp@chromium.org,johnme@chromium.org

Review URL: https://codereview.chromium.org/818363003

Cr-Commit-Position: refs/heads/master@{#309622}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/ce0377c..50a87e2

TBR=gavinp@chromium.org,johnme@chromium.org

Review URL: https://codereview.chromium.org/826623002

Cr-Commit-Position: refs/heads/master@{#309621}

The interactive_ui_tests for toolkit-views' custom menus (e.g. for combo
boxes) would work fine when running tests locally, but refuse pass when
run on a trybot.

Tracing narrowed the discrepancy down to the
"WindowAtCurrentMouseLocation()" function in ui_controls_mac.mm. This
was working for windows at some positions on the screen, but not others,
and attempting to add a [NSApp activateIgnoringOtherApps:YES] call at
the appropriate place (e.g. ShowAndFocusNativeWindow() in
interactive_test_utils_mac.mm) had no effect.

This CL makes WindowAtCurrentMouseLocation more robust by effectively
ignoring the influence of windows in other applications that may be on
top of windows created in the test. Instead, iterate through [NSApp
orderedWindows] and use NSPointInRect. This allows all the toolkit-views
menu interactive_ui_tests to pass (except drag & drop - still working on
that).

Note that, historically, Mac's SendMouseMoveNotifyWhenDone would just
use the keyWindow, but that doesn't work for menus (it's also "wrong"
for mouse events).

I suspect this will also allow me to de-flake a bunch of disabled, Cocoa
interactive_ui_tests that have been accumulating.

BUG=403679

Review URL: https://codereview.chromium.org/821883002

Cr-Commit-Position: refs/heads/master@{#309620}

TBR=jrummell@chromium.org
BUG=432998
TEST=Manually tested.
R=jrummell@chromium.org

Review URL: https://codereview.chromium.org/817143003

Cr-Commit-Position: refs/heads/master@{#309619}

We need it to supported encrypted media playback. Note that |is_encrypted| is
already in VideoDecoderConfig.

BUG=432998
TEST=Added unittest to cover this.
R=jrummell@chromium.org

Review URL: https://codereview.chromium.org/820103002

Cr-Commit-Position: refs/heads/master@{#309618}

Example to enable logging in both html_viewer and media app:
out/GN/mojo_shell http://<host>/bear-src-a.html
--args-for='mojo://html_viewer --enable-mojo-media-renderer
--vmodule=*mojo_render*=3' --args-for='mojo://media --vmodule=*service*=3'

TBR=jamesr@chromium.org
BUG=432998,434802
TEST=See above.
R=jrummell@chromium.org

Review URL: https://codereview.chromium.org/823853003

Cr-Commit-Position: refs/heads/master@{#309617}

Previously the TaskController calls the private version of executeDefaultTask
method that does not wait for initialization completion of FileTasks.  Thus it
tries to call default task before the default task is determined.

BUG=443942
TEST=manually

Review URL: https://codereview.chromium.org/821003002

Cr-Commit-Position: refs/heads/master@{#309616}

util.loadImage and util.cancelLoadImage seem to be called only by thumbnail_loader.js. Since common/js/util.js is included by many applications, unnecessarily putting codes in common/js/util.js brings unnecessary dependencies.

BUG=433728
TEST=GYP_GENERATORS=ninja tools/gyp/gyp --depth . ui/file_manager/gallery/js/compiled_resources.gyp && ninja -C out/Default

Review URL: https://codereview.chromium.org/823593006

Cr-Commit-Position: refs/heads/master@{#309615}

BUG=444510
TEST=GYP_GENERATORS=ninja tools/gyp/gyp --depth . ui/file_manager/file_manager/foreground/js/metadata/compiled_resources.gyp && ninja -C out/Default

Review URL: https://codereview.chromium.org/821653002

Cr-Commit-Position: refs/heads/master@{#309614}

BUG=425993

Review URL: https://codereview.chromium.org/795053003

Cr-Commit-Position: refs/heads/master@{#309613}

BUG=444485
TEST=GYP_GENERATORS=ninja tools/gyp/gyp --depth . ui/file_manager/video_player/js/compiled_resources.gyp && ninja -C out/Default | grep "media_controls.js"

Review URL: https://codereview.chromium.org/824973002

Cr-Commit-Position: refs/heads/master@{#309612}

This enforces thread-safety by design via transactions for any
implementation code accessing shared SafeBrowsingDatabase members.

Previously this was enforced by assuming everybody would make proper
usage of |lookup_lock_| but this was deemed insufficient as programming
mistakes proved easy to make twice.

One of the subtleties of this transactional model (which made the previous
ad-hoc locking even harder to prove (and keep) correct) is that only the
main database thread is allowed to modify this state, allowing for
unlocked reads on the main thread which are important to avoid
contention when flushing to disk.

BUG=440517

Review URL: https://codereview.chromium.org/794273002

Cr-Commit-Position: refs/heads/master@{#309611}

The issue was caused by a oversight of crrev.com/805713003.
As the menu was not treated as focus parent, 'focus' event was dispatched with the text-field as an target. This make the menu hidden just after the menu is shown.
This CL corrects the selector to make the menu focus parent.

BUG=444480
TEST=go through repro steps written in the issue

Review URL: https://codereview.chromium.org/822663003

Cr-Commit-Position: refs/heads/master@{#309610}

BUG=444902, 297672
TBR=thakis

Review URL: https://codereview.chromium.org/824963002

Cr-Commit-Position: refs/heads/master@{#309609}

This is the followup to https://codereview.chromium.org/822823002/ that fixes the crash for DETECT.

This is lower priority than the preceding bug, since it only affects 41/Canary, but it still addresses a crash.

BUG=444524

Review URL: https://codereview.chromium.org/819183002

Cr-Commit-Position: refs/heads/master@{#309608}

BUG=408189

Review URL: https://codereview.chromium.org/823923002

Cr-Commit-Position: refs/heads/master@{#309607}

MediaSource is no longer used in the webapp and there is
no plan to use it in the future.

Review URL: https://codereview.chromium.org/825793002

Cr-Commit-Position: refs/heads/master@{#309606}

BUG=444955

Review URL: https://codereview.chromium.org/823643003

Cr-Commit-Position: refs/heads/master@{#309605}

Review URL: https://codereview.chromium.org/826563003

Cr-Commit-Position: refs/heads/master@{#309604}

PixelBufferTileTaskWorkerPool::CheckForCompletedRasterTasks() has a DCHECK
that triggers due to TileManager::DidFinishRunningTileTasks recursing
back into PixelBufferTileTaskWorkerPool.

This patch removes the recursion.

BUG=442598

Review URL: https://codereview.chromium.org/807273005

Cr-Commit-Position: refs/heads/master@{#309603}

On platforms (i.e. ChromeOS using Ozone) where Chromium is responsible for
displaying the buffers, we want to delay SwapBufferAck till we know that
the buffer is displayed on screen. Currently, we block the GPU main thread till
that data is available. This patch introduces SwapBufferAsync apis, which are
similar to current SwapBuffer calls except that it adds a callback function
which can be used to delay sending SwapBufferAck as needed.

BUG=443543,chrome-os-partner:34292

Review URL: https://codereview.chromium.org/797843005

Cr-Commit-Position: refs/heads/master@{#309602}

Current string "Detect important content" in the WebUI exceptions dialog
is too long. Changing to "Detect content", which is... probably fine.
I mean, before PPS, it was "Ask", which was even more meaningless.

BUG=444520

Review URL: https://codereview.chromium.org/805643003

Cr-Commit-Position: refs/heads/master@{#309601}

Set the owner_extension_id in GuestViewBase::Init so that it's available in request.embedderExtensionId in onEmbedRequested in the app to be embedded. This also simplifies code in GuestViewBase a bit.

BUG=444749
TBR=rockot@chromium.org for guest_view_internal_api.cc

Review URL: https://codereview.chromium.org/823543003

Cr-Commit-Position: refs/heads/master@{#309600}

It's disabled, because we don't have a fix yet
(and I'm about to go on leave).

This required moving the WebContentsAddedObserver from
content_browser_test_utils.h->browser_test_utils.h. I took
the opportunity to RenderFrame-ify it slightly.

BUG=444945

Review URL: https://codereview.chromium.org/818263004

Cr-Commit-Position: refs/heads/master@{#309599}

Two regressions were introduced by https://crrev.com/30c6291b559f48df49.
The first was that the wValue and wIndex fields of control transfers
were passed as uint8s, truncating these normally 16-bit values. The
second was that the number of packets of an isochronous transfer was
no longer passed to libusb_alloc_transfer.

BUG=443107

Review URL: https://codereview.chromium.org/827433002

Cr-Commit-Position: refs/heads/master@{#309598}

To include recent updates to deploy_chrome script.

TEST=gclient sync
BUG=None
NOTRY=true
TBR=vapier@chromium.org

Review URL: https://codereview.chromium.org/825803002

Cr-Commit-Position: refs/heads/master@{#309597}

Use correct browser context when creating web_contents.

R=gunsch@chromium.org
BUG=

Review URL: https://codereview.chromium.org/818283002

Cr-Commit-Position: refs/heads/master@{#309596}

This adds a free list to the browser-wide discardable memory
manager in child processes. This reduces the number of open
file descriptors and improves performance significantly by
avoiding a lot of browser process round-trips.

Address-ordered best-fit policy is currently used as it is more
likely to result in fewer locked segments but the policy can
easily be adjusted if needed.

BUG=429415
TEST=content_unittests --gtest_filter=DiscardableSharedMemoryHeapTest.*

Review URL: https://codereview.chromium.org/807303002

Cr-Commit-Position: refs/heads/master@{#309595}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/0e9ab4e..0eafc9b

CQ_EXTRA_TRYBOTS=tryserver.blink:linux_blink_rel,linux_blink_dbg
TBR=bungeman@google.com

Review URL: https://codereview.chromium.org/818323002

Cr-Commit-Position: refs/heads/master@{#309594}

Pulseaudio uses default pipe buffer size to calculate pipe
latency, ignoring the actual buffer size set by the reading
side. AudioPipeReader was setting buffer size that's about 2
times bigger than default. As result it reports incorrect
latency and some clients did not work properly (particularly
AudioContext API implementation in chrome). Now
AudioPipeReader will be using default buffer size determined
by the OS, which matches pulseaudio's behavior.

BUG=442455
R=lambroslambrou@chromium.org

Review URL: https://codereview.chromium.org/824943002

Cr-Commit-Position: refs/heads/master@{#309593}

By default open() blocks when opening a fifo until the writer also
opens the fifo. In the past AudioPipeReader was using O_NONBLOCK
when opening the pipe for reading, but that was changed in r277315,
and may result in host hanging if pulseaudio doesn't open the pipe
for some reason. Fixed AudioPipeReader to always pass O_NONBLOCK
to open().

BUG=420090
R=lambroslambrou@chromium.org

Review URL: https://codereview.chromium.org/820123002

Cr-Commit-Position: refs/heads/master@{#309592}

android_chromium_gn_compile_dbg is red failing to compile a cc file:
In file included from ../../device/vibration/android/vibration_jni_registrar.cc:9:0:
../../device/vibration/vibration_manager_impl_android.h:9:54: fatal error: device/vibration/vibration_manager.mojom.h: No such file or directory
 #include "device/vibration/vibration_manager.mojom.h"
                                                      ^

This was added with incorrect dependencies in
https://codereview.chromium.org/583663003/ two weeks ago, but
apparently no clobber happened since then.  The problem can be
reproduced locally by running

  rm -rf out/Debug
  gn gen out/Debug --args=is_debug=true\ os=\"android\"\ cpu_arch=\"arm\"\ use_goma=true\ goma_dir=\"$HOME/goma\"
  ninja -C out/Debug obj/device/vibration/android/android.vibration_jni_registrar.o

BUG=none
R=dcheng@chromium.org

Review URL: https://codereview.chromium.org/826593003

Cr-Commit-Position: refs/heads/master@{#309591}

Last instrumentations showed that:
SSL_do_handshake call is 17 jph (I mean janks per startup-hour)
This excludes all callbacks that SSL_do_handshake can invoke.

SSLClientSocketOpenSSL::UpdateServerCert is 5.8 jph.

Adding instrumentations to differentiate the cert from no-cert cases.
Also instrumenting SSLClientSocketOpenSSL::UpdateServerCert.

BUG=424386

Review URL: https://codereview.chromium.org/787913003

Cr-Commit-Position: refs/heads/master@{#309590}

A lot of SideEffectWhiteListPrefixSet related code was mostly a duplicate
 of the other PrefixSet code in the same file.

Only logical side-effect this has is grouping histograms for all PrefixSets
in the same bucket. This will be addressed as a follow-up in
https://codereview.chromium.org/815863002/ where relevant histograms will be
split once again.

This cleanup makes it easier to progress on issue 440517.

BUG=440517
TBR=jwd

Review URL: https://codereview.chromium.org/790703003

Cr-Commit-Position: refs/heads/master@{#309589}

GN bots are having trouble picking up the changes to the resouce ID
range, which was changed in https://codereview.chromium.org/734243003/.

BUG=444902
R=eroman@chromium.org, thakis@chromium.org

Review URL: https://codereview.chromium.org/821303002

Cr-Commit-Position: refs/heads/master@{#309588}