GitLab

Revert of Enable Surfaces on Android (try #4) (patchset #1 id:1 of https://codereview.chromium.org/1147193005/)

Reason for revert:
The following tests started failing on Android Debug Nexus 5, Nexus 6 and Nexus 9 bots:

pixel_tests: Pixel.Canvas2DRedBox, Pixel.CSS3DBlueBox, Pixel.WebGLGreenTriangle
gpu_rasterization_tests: GpuRasterization.BlueBox

Original issue's description:
> Enable Surfaces on Android (try #4)
>
> The browser compositor contents and renderer contents are each put into separate
> layers which are aggregated together by a Display.
>
> BUG=449319
>
> Committed: https://crrev.com/c95c7a17c51b905ef491fcea157852adebf85bfe
> Cr-Commit-Position: refs/heads/master@{#339793}

TBR=sievers@chromium.org,jbauman@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=512676

Review URL: https://codereview.chromium.org/1250993002

Cr-Commit-Position: refs/heads/master@{#339834}

This generates a proper version manifest for chrome.exe and also for the version assembly.

This adds a new way to get the chrome version info programatically at GN time. Although this adds another exec_script call, it removes four such calls in remoting so there is a net speed increase.

BUG=

Review URL: https://codereview.chromium.org/1250853007

Cr-Commit-Position: refs/heads/master@{#339833}

BUG=508746
TEST=gpu_unittests

Review URL: https://codereview.chromium.org/1246203003

Cr-Commit-Position: refs/heads/master@{#339832}

TBR=bartfab@chromium.org,mnissler@chromium.org
NOTRY=true
BUG=512670

Review URL: https://codereview.chromium.org/1241203004

Cr-Commit-Position: refs/heads/master@{#339831}

TBR=hirono@chromium.org
NOTRY=true
BUG=512669

Review URL: https://codereview.chromium.org/1250983002

Cr-Commit-Position: refs/heads/master@{#339830}

This adds the data and data_deps logic for ui_base_unittests
and cacheinvalidation_unittests for the GN build, so that we
can run those two binaries under swarming.

R=sky@chromium.org, brettw@chromium.org, maruel@chromium.org
BUG=504079

Review URL: https://codereview.chromium.org/1246103005

Cr-Commit-Position: refs/heads/master@{#339829}

XHR stops throwing NetworkError exceptions from
https://codereview.chromium.org/1060113004/

Prepare for it by removing chrome.test.fail() call but keeping the
expectation set for the error event.

After landing the Blink side patch, remove the try-catch.

BUG=479484
R=yhirano

Review URL: https://codereview.chromium.org/1238313003

Cr-Commit-Position: refs/heads/master@{#339828}

Cr-Commit-Position: refs/heads/master@{#339827}

NOPRESUBMIT=true
TEST=All current tests pass. Closure compiler shows no errors.
BUG=507210

Review URL: https://codereview.chromium.org/1233823003

Cr-Commit-Position: refs/heads/master@{#339826}

Add a copy of the existing SRPC-based implementation --
native_client/src/untrusted/irt/irt_pnacl_translator_{compile,link}.c
-- to the Chromium side.  This will allow the two implementations to
be independently changed to remove use of SRPC.

In copying this code, I have made some changes to follow the Chromium
style:

 * Switching to C++ and using C++ comments
 * Using anon namespaces instead of "static"
 * "*" spacing (using clang-format)
 * "FooBar" function naming instead of "foo_bar"
 * Using the Chromium copyright notice

BUG=302078
TEST=e.g. NaClBrowserTestPnacl.PPAPICore in browser_tests
  (also manually tested with NaCl-side interface disabled, to ensure the
  Chromium-side one is really being used)

Review URL: https://codereview.chromium.org/1240343004

Cr-Commit-Position: refs/heads/master@{#339825}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/c71239b..78f55fc

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
TBR=joshualitt@google.com

Commits in this roll:
78f55fc mtklein@google.com Revert of 565 support for SIMD xfermodes (patchset #4 id:60001 of https://codereview.chromium.org/1245673002/)
860dcaa mtklein@chromium.org 565 support for SIMD xfermodes

Review URL: https://codereview.chromium.org/1245763003

Cr-Commit-Position: refs/heads/master@{#339824}

BUG=398202
TEST=none
TBR=shess

Review URL: https://codereview.chromium.org/1247803007

Cr-Commit-Position: refs/heads/master@{#339823}

* Fix typo
* Introduced Scoped125DSFForUIScaling to temporarily disable it for testing.
* Make sure id isn't invalid when checking if it's internal.
 This problem exists only in unit tests, not in production.

BUG=None

Review URL: https://codereview.chromium.org/1251573006

Cr-Commit-Position: refs/heads/master@{#339822}

Currently we rely on a flag (--ozone-use-surfaceless) to trigger the
surfaceless code path in both GPU & browser processes.

This makes surfacelessness a context capability, so that we can create
the appropriate output surface in the browser without using the flag.

BUG=512631
TEST=Manually on link. UI is drawn.

Review URL: https://codereview.chromium.org/1243263002

Cr-Commit-Position: refs/heads/master@{#339821}

Fixes:
..\..\media\audio\win\audio_device_listener_win_unittest.cc(24,19) : error:
unused variable 'kNoDevice' [-Werror,-Wunused-const-variable]
static const char kNoDevice[] = "";
                  ^
No intended behavior change.

BUG=506712,82385
TBR=dalecurtis@chromium.org

Review URL: https://codereview.chromium.org/1250933002.

Cr-Commit-Position: refs/heads/master@{#339820}

R=kalman@chromium.org, toyoshim@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1208993011

Cr-Commit-Position: refs/heads/master@{#339819}

BUG=497258

Add ViewStructure class temporarily. This commit also rolls
=============================================
Cherry-pick libc++ bug fixes (7aa54b450fb756fecafb327faa03ae336fea3671).
================================================

Review URL: https://codereview.chromium.org/1246163003

Cr-Commit-Position: refs/heads/master@{#339818}

The |time_progressing_| variable should only be accessed via the
media thread, it can't be accessed under lock since it is set
during a call which may be reentrant to locked methods.

The fix is to use a proxy for |time_progressing_| when called from
other threads; luckily the inverse of |render_first_frame_and_stop_|
will suffice.

BUG=512371
TEST=media_unittests

Review URL: https://codereview.chromium.org/1247973002

Cr-Commit-Position: refs/heads/master@{#339817}

Currently there is logic in the renderer which checks whether a plugin can
handle a mime type and if it can the plugin is started. This is also how
MimeHandlerView plugins are started in the renderer. MimeHandlerViews also
rely on the original http request being intercepted as a stream which the
plugin can read when it chooses. This happens in the browser. Currently
the logic for determining whether the stream should be intercepted is different
to the logic for determining whether to start the plugin which can lead to
problems in some cases. This CL makes sure the logic is the same: on both
sides we check whether the plugin can handle the mime type.

MimeHandlerView plugins load a chrome extension inside a BrowserPlugin.
Once it has been determined that the plugin can handle the mime type inside
BufferedResourceHandler, the extension ID of the plugin is passed down through into
ChromeResourceDispatcherHostDelegate::ShouldInterceptResourceAsStream.
It is then checked that the extension is enabled and if so the stream is intercepted.
There is some complexity because we also need to support the legacy case
where the stream is intercepted for the streamsPrivate api in which case there is no
MimeHandlerView plugin.

This CL was originally written by deepak.m1@samsung.com (Deepak Mittal)
and landed here: https://codereview.chromium.org/953793003/. It was reverted
because of a bug it triggered with not downloading PDFs when the PDF plugin
was disabled. This bug has been addressed and a test added for that case.

BUG=443466

Review URL: https://codereview.chromium.org/1164073006

Cr-Commit-Position: refs/heads/master@{#339816}

It looks like some software changes internal flows instead of
changing the device id.  In place of this deduplication logic
I've added a rate limit of 1 change per 250ms, which is enough
to fix the original issue of multiple events.

Also fixes an issue where we were firing device changes four
times (!) --[eRender, eCapture] * [eConsole, eCommunications].
The AudioManager only supports output device notifications so
we should only be firing the listener callback for eRender.

With the rate limiting logic and only firing notifications for
eRender, this takes us down from four device changes to just
1 for unplug / replug of a device.

BUG=506712
TEST=device changes work, single stream created per multiple events.
TBR=henrika

Review URL: https://codereview.chromium.org/1248543002

Cr-Commit-Position: refs/heads/master@{#339815}

R=esprehn@chromium.org,thestig@chromium.org
BUG=511486

Review URL: https://codereview.chromium.org/1246253002

Cr-Commit-Position: refs/heads/master@{#339814}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/10b371c..c71239b

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
TBR=joshualitt@google.com

Commits in this roll:
c71239b herb@google.com Change the GlyphCache to use a hash table instead of doing its own ad-hoc hashing. This change appears to be performance neutral. BUG=skia:

Review URL: https://codereview.chromium.org/1250793006

Cr-Commit-Position: refs/heads/master@{#339813}

To prevent spurious coded frame processing discontinuity detection for
potentially incorrect streams with coded frames having decode time >
presentation time, this change updates DTS by the same delta as PTS
during partial append window trimming.
This change is option 2 from comment 9 in bug 511128.

BUG=511128
R=dalecurtis@chromium.org
TEST=*/FrameProcessorTest.PartialAppendWindowFilterNoDiscontinuity_DtsAfterPts

Review URL: https://codereview.chromium.org/1240323003

Cr-Commit-Position: refs/heads/master@{#339812}

Flaky on Linux ChromiumOS bots.

TBR=nkostylev@chromium.org
NOTRY=true
BUG=512648

Review URL: https://codereview.chromium.org/1249783003

Cr-Commit-Position: refs/heads/master@{#339811}

BUG=

Review URL: https://codereview.chromium.org/1243233002

Cr-Commit-Position: refs/heads/master@{#339810}

a browser window that are shown on the given |profile|'s desktop to prevent returning
a browser window on other users' desktop.

Also clean up and refactor codes related to TestBrowserWindowAura class. There were
multiple files defining TestBrowserWindowAura that were almost identical. Extracted this class to
test_browser_window.h .cc files and added helper function CreateBrowserWithAuraTestWindowForParams()
to handle its lifetime.

BUG=500027

Review URL: https://codereview.chromium.org/1198313003

Cr-Commit-Position: refs/heads/master@{#339809}

As previously discussed, DownloadResouceThrottle is the best place to handle file access requests.
This CL moves that logic from DownloadTargetDeterminer to DownloadResourceThrottle.
Unit test is also added for DownloadResourceThrottle.

BUG=501606

Review URL: https://codereview.chromium.org/1229933010

Cr-Commit-Position: refs/heads/master@{#339808}

Extension profile extender for profile generator and a new benchmark for
startup time on a profile that loads extensions. Also for this purpose, a
script to maintain a static set of extensions in cloud storage, and small
additions to python gsutil wrapper (cloud_storage).

BUG=444230
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_perf_bisect

Review URL: https://codereview.chromium.org/1240703003

Cr-Commit-Position: refs/heads/master@{#339807}

Adding a flag to enable the use of the newly rewritten implementation of
the task manager.

BUG=471006

Review URL: https://codereview.chromium.org/1248823004

Cr-Commit-Position: refs/heads/master@{#339806}

BUG=505040
TEST=None
TBR=aurimas
NOTRY=true

Review URL: https://codereview.chromium.org/1243273002

Cr-Commit-Position: refs/heads/master@{#339805}

This patch changes how MB will generate the .isolate files
for a GN build:

First, we rename 'ninja_to_gn.pyl' to 'gn_isolate_map.pyl',
and change it from a straight compile_target -> gn_label map to
a map of compile_target -> { "label": gn_label, "type": string,
"args": optional list<string> }.

Valid values for "type" are
"windowed_test_launcher"
: the test is a gtest-based test that uses the
  'brave-new-test-launcher' from //base/test:test_support and
  needs to run under Xvfb if run on an X11-based platform
  (i.e., if use_x11=true in GN).
"console_test_launcher"
: the test is a gtest-based test that uses the
  'brave-new-test-launcher' from //base/test:test_support but does
  not need Xvfb.
"raw"
: the test is just an executable and may take an optional list
  of command-line arguments (specified in the 'args' key),
  but does not need any extra files or other special processing.
"unknown"
: (the default), which indicates that we don't know what the command
  line needs to be (this is a fatal error).

Second, we read the new file in, and build a lookup map of test
types to command lines. We will probably need to keep iterating
on the list of test types and command lines, and perhaps provide
some escape hatches so that we can handle the full variety of tests.

TBR=maruel@chromium.org
BUG=503942, 504079

Review URL: https://codereview.chromium.org/1239343003

Cr-Commit-Position: refs/heads/master@{#339804}

Using ClipDrawable progress bar helps future changes for
http://crbug.com/461148 as we only need pass to color parameters.

Also it has minor performance benefit.

Note: second landing attempt as the previous one made Chrome shell tests
      flaky. https://codereview.chromium.org/1170843002/

TBR=tedchoc@chromium.org
BUG=466140, 455891, 461148

Review URL: https://codereview.chromium.org/1245333002

Cr-Commit-Position: refs/heads/master@{#339803}

Revert of gpu: Use shader cache on ES3 (patchset #2 id:20001 of https://codereview.chromium.org/1223393003/)

Reason for revert:
This codepath is actually flaky.

OnLoadedShader() races with dynamic GL binding initialization. Previously it meant that we just might not end up loading this shader (as long as b_GL_OES_get_program_binary etc. default to false), but the GetGLVersionInfo() might actually crash.

Original issue's description:
> gpu: Use shader cache on ES3
>
> The ES3 core specs supports gl(Get)ProgramBinary().
> BUG=510172
>
> Committed: https://crrev.com/809d9b979c0edd3976132eaaf77fc4740fcee019
> Cr-Commit-Position: refs/heads/master@{#339369}

TBR=dyen@chromium.org,zmo@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=510172,512464

Review URL: https://codereview.chromium.org/1247233002

Cr-Commit-Position: refs/heads/master@{#339802}

When a session is suspended by the UI from the media notifications, it
does not release the audio focus and might receive focus loss/gain
messages from the system. It shouldn't try to resume the session when
the focus is gained back in such case.

BUG=511921

Review URL: https://codereview.chromium.org/1243683003

Cr-Commit-Position: refs/heads/master@{#339801}

In L apps are able to extend itself under the navigation bar. Yet this
feature is not needed in enhanced bookmarks, and it caused a UI bug with
snackbar.

BUG=508873

Review URL: https://codereview.chromium.org/1250853005

Cr-Commit-Position: refs/heads/master@{#339800}

Introduce a queue in TabDelegate for creating multiple Tabs in a row.
This prevents triggering some issue in Android's ActivityManager that
causes it to throw away the tasks for Activities after it launches them,
which in turn causes Android to destroy our Activities.

BUG=498920

Review URL: https://codereview.chromium.org/1229853006

Cr-Commit-Position: refs/heads/master@{#339799}

Hooks up the allocator shim on Windows. The GN build uses the Windows heap like GYP does.

This makes tcmalloc actually compile on Windows if you manually request it, although it doesn't actually seem to work correctly. I added an assert noting this if you try to set the flag on Windows. Note that this is never run by default on GYP and we may never want it.

This sets the ALLOCATOR_SHIM for the one media file that uses it.

Review URL: https://codereview.chromium.org/1246083005

Cr-Commit-Position: refs/heads/master@{#339798}

direct_dependent_configs is the deprecated name for public_configs,
and public_configs is already handled.

gn-dev thread:
https://groups.google.com/a/chromium.org/d/topic/gn-dev/h9N85MIx_Bk/discussion

BUG=None
TEST=gn gen out-gn/Debug
R=brettw@chromium.org

Review URL: https://codereview.chromium.org/1246183002

Cr-Commit-Position: refs/heads/master@{#339797}

BUG=None
R=bengr@chromium.org

Review URL: https://codereview.chromium.org/1244923002

Cr-Commit-Position: refs/heads/master@{#339796}

Previously, this member was assigned ad-hoc during initialization of a
RenderWidget. Since there are different initialization paths for
different kinds of RenderWidgets, this led to some strange-looking code:
for example, the initialization of a frame widget would double-assign
CompositorDependencies.

BUG=none
TBR=mkwst@chromium.org

Review URL: https://codereview.chromium.org/1245863003

Cr-Commit-Position: refs/heads/master@{#339795}