GitLab

Changed TabState.saveState to be able to safely write
mapped byte buffers by opening FileOutputStream after
retrieving buffer contents. Removed unnecessary call to
FileChannel.write which is not guaranteed to write all
bytes.

BUG=626815

Review-Url: https://codereview.chromium.org/2138503002
Cr-Commit-Position: refs/heads/master@{#405299}

This is part of the touch hud app for mustash (https://codereview.chromium.org/2092343002/).

BUG=588311
TEST=mus_ws_unittests

Review-Url: https://codereview.chromium.org/2118383002
Cr-Commit-Position: refs/heads/master@{#405298}

This is part of the touch hud app for mustash (https://codereview.chromium.org/2092343002/).

BUG=588311
TEST=views_mus_unittests

Review-Url: https://codereview.chromium.org/2125663002
Cr-Commit-Position: refs/heads/master@{#405297}

This is part of the touch hud app for mustash (https://codereview.chromium.org/2092343002/).

BUG=588311
TEST=ash_unittests

Review-Url: https://codereview.chromium.org/2118223004
Cr-Commit-Position: refs/heads/master@{#405296}

This patch periodically re-enables GPU rasterization after it is
disabled due to a content veto.

R=enne
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_blink_rel

Review-Url: https://codereview.chromium.org/2108033004
Cr-Commit-Position: refs/heads/master@{#405295}

Revert of RELAND: ImageCapture: Implement takePhoto() for Mac AVFoundation (patchset #2 id:20001 of https://codereview.chromium.org/2146973002/ )

Reason for revert:
It caused timeout in
WebRtcWebcamBrowserTests/WebRtcWebcamBrowserTest.MANUAL_TestAcquiringAndReacquiringWebcam/0

see
https://build.chromium.org/p/chromium.webrtc/builders/Mac%20Tester/builds/56929

Original issue's description:
> RELAND: ImageCapture: Implement takePhoto() for Mac AVFoundation
>
> Got reverted due to breaking
> WebRtcWebcamBrowserTests/WebRtcWebcamBrowserTest.MANUAL_TestAcquiringAndReacquiringWebcam/0
> in some WebRtc Mac bot, possibly because it has no real WebCam
> or it has one without  AVCaptureStillImageOutputClass support.
>
> This CL adds a guard  [captureSession_ canAddOutput:...]
> before [captureSession_ addOutput:...]
>
> Original description ----------------------------------------------
> ImageCapture: Implement takePhoto() for Mac AVFoundation
>
> By pulling the necessary symbols out of AVFoundation
> and using them; most notably (Cr)AVCaptureStillImageOutput.
>
> TakePhoto is already implemented for Android (both APIs).
>
> Also adding VideoCaptureDeviceTest::MAYBE_TakePhoto
> test case, enabled only for Mac.
>
> BUG=518807
> TEST=Run build with flag --enable-blink-features=ImageCapture,
> navigate to [1] and push buttons
> - Open Camera ...
> - Create ImageCapturer
> - takePhoto()  (N times!) --> profit
>
> [1] https://rawgit.com/Miguelao/demos/master/imagecapture.html
>
> TBR=rsesek@chromium.org since the change is trivial.
>
> Committed: https://crrev.com/d4179235b928634aeadb0c3379d385caf0ad3a07
> Cr-Commit-Position: refs/heads/master@{#405243}

TBR=
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=518807

Review-Url: https://codereview.chromium.org/2151443003
Cr-Commit-Position: refs/heads/master@{#405294}

R=tandrii@chromium.org
BUG=617627
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_blink_rel

Review-Url: https://codereview.chromium.org/2147793003
Cr-Commit-Position: refs/heads/master@{#405293}

This patch updates the code to prevent auto deducing to a raw pointer.

R=brettw@chromium.org, danakj, dcheng
BUG=554600

Review-Url: https://codereview.chromium.org/2105553005
Cr-Commit-Position: refs/heads/master@{#405292}

This regressed in https://codereview.chromium.org/2092623002 which caused a revert of the GN roll. The particular issue is that a build argument in a BUILD.gn file was used in only one of the toolchains, so the other toolchains were giving unused variable errors on it.

Build variables kept at their default values already are marked implicitly used, and anything in a .gni file is the same. The remaining cases are pretty confusing so it seems better to not warn in this case instead of requiring the author to wrap the build variable in a condition based on the usage.

TBR=dpranke

Review-Url: https://codereview.chromium.org/2151633002
Cr-Commit-Position: refs/heads/master@{#405291}

BUG=605198
TEST=add unittest in audio_manager_unittest; unittest doesn't include beamformingdevices case yet as I am still not quite clear about its usage.

tested with https://webrtc.github.io/samples/src/content/devices/input-output/, showing enumerating audio devices in a format of (default) plus other chromeos audio devices.

Review-Url: https://codereview.chromium.org/2079843003
Cr-Commit-Position: refs/heads/master@{#405290}

Also convert !container.size() to container.empty().

Review-Url: https://codereview.chromium.org/2126433002
Cr-Commit-Position: refs/heads/master@{#405289}

Move nearer use in session_state_delegate_chromeos.cc.

A temporary/punt measure to avoid some minor issues.
It'll need to communicate w/ash for share/capture stuff.
Avoids SystemTray access and dialog conversion tasks.

BUG=619636
TEST=compiles; unit tests; no behavior changes.
R=jamescook@chromium.org
TBR=sky@chromium.org

Review-Url: https://codereview.chromium.org/2144083002
Cr-Commit-Position: refs/heads/master@{#405288}

m_forCrossProcessAncestor was intended to mean that fullscreen was
entered for an iframe container which is an ancestor of the actual
fullscreen element, which also mirrors the
:-webkit-full-screen-ancestor style that the container will gain.  But
on the flipside, this naming can also be confusing: if an ancestor of
a cross-process iframe goes fullscreen, that iframe doesn't need to
know about this at all, and instead it only needs to know when an
element in a descendant frame goes fullscreen.  It might make more
sense to name this flag to reflect where the actual fullscreen element
is, rather than how m_fullscreenElement relates to it, so this CL
renames the flag to m_forCrossProcessDescendant and updates comments.

BUG=550497

Review-Url: https://codereview.chromium.org/2147803002
Cr-Commit-Position: refs/heads/master@{#405287}

These events are 100x more common than other errors. User input events
can arrive when the user is e.g. looking at an image, which isn't a
tracked page but is something a user can interact with, so it's valid
to receive input events without a committed load. Thus, it doesn't make
sense to log this error.

Review-Url: https://codereview.chromium.org/2142743002
Cr-Commit-Position: refs/heads/master@{#405286}

Channel code used to be in //chrome, so it couldn't be used in
//extensions.  Now that it's moved to components/, we can safely
include it and can simplify Features code significantly.

BUG=626774

Review-Url: https://codereview.chromium.org/2137463003
Cr-Commit-Position: refs/heads/master@{#405285}

To avoid using the BrowserThread we instantiate a ThreadChecker instance on the NetLogObserver::Attach method and use this to validate that calls to the NetLogObserver instance occur on the same thread.

To avoid using ContentBrowserClient, we pass the NetLog pointer in the NetLogObserver::Attach method.

BUG=598073

Review-Url: https://codereview.chromium.org/2115823004
Cr-Commit-Position: refs/heads/master@{#405284}

Revert of Revive experiment to isolate shell operations. (patchset #5 id:100001 of https://codereview.chromium.org/2122303002/ )

Reason for revert:
Possibly broke compile on"Win8 GYP (dbg)"

Original issue's description:
> Isolate shell operations to the utility process.
>
> Shell operations can cause 3rd-party shell extensions to be
> loaded into the caller's process. The utility process protects the browser process from potential instability.
>
> Also convert the IPC mechanism to Mojo.
>
> BUG=73098
>
> Committed: https://crrev.com/b7faec8d94ef286d2b92144c2836b170c2cdea95
> Cr-Commit-Position: refs/heads/master@{#405270}

TBR=sky@chromium.org,dcheng@chromium.org,wfh@chromium.org,pmonette@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=73098

Review-Url: https://codereview.chromium.org/2141093005
Cr-Commit-Position: refs/heads/master@{#405283}

BUG=627392

R=kbr@chromium.org

Review-Url: https://codereview.chromium.org/2143913002
Cr-Commit-Position: refs/heads/master@{#405282}

ProtoZero plugin is a custom C++ stubs generator for protobuf compiler.
These stubs are Tracing V2 specific and meet perfomance and memory
requirements being less functional than conventional ones.
This CL has no complete implementation (see crrev.com/2083373002) but
contains all invasive changes to build files.
See crrev.com/2082693002 for background about protobuf plugin support.

BUG=608721

Review-Url: https://codereview.chromium.org/2147613002
Cr-Commit-Position: refs/heads/master@{#405281}

Revert of Improve ios/web/public/test/web_js_test.h (patchset #2 id:20001 of https://codereview.chromium.org/2148723003/ )

Reason for revert:
I'm afraid DCHECK_NSEQ is not defined, just EXPECT and ASSERT. It seems to break compilation downstream. Let me revert and fix.

Vaclav

Original issue's description:
> Improve ios/web/public/test/web_js_test.h
>
> This CL does some minor changes to web_js_test.h based on eugenebut's
> recommendations in https://codereview.chromium.org/2150513002/#msg21.
>
> R=eugenebut@chromium.org
> BUG=627775
>
> Committed: https://crrev.com/2de879b6567bdd5383db462d0d3ee194bb4d998b
> Cr-Commit-Position: refs/heads/master@{#405234}

TBR=eugenebut@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=627775

Review-Url: https://codereview.chromium.org/2146793003
Cr-Commit-Position: refs/heads/master@{#405280}

To match the TimeDeltas in page_load_metrics, the
DataReductionProxyPageLoadTiming TimeDeltas are changed to
base::Optional TimeDeltas. Functionally, this will only change the case
where the TimeDelta is zero.

The constructor for DataReductionProxyPageLoadTiming is moved to a new
cc file due to the templated variables creating a complex constructor.

BUG=626040

Review-Url: https://codereview.chromium.org/2145743002
Cr-Commit-Position: refs/heads/master@{#405279}

This CL implements the OpenGL canvas, rendering layer, and desktop. This is part
of the project of implementing OpenGL rendering component for Android and iOS.
Please see design doc in crbug for more information.

BUG=385924

Review-Url: https://codereview.chromium.org/2045963004
Cr-Commit-Position: refs/heads/master@{#405278}

Given std::unique_ptr<T> t, t.get()-> is the same as t->.
Similarly, *t.get() is the same as *t. Simplify extensions
code to conform.

BUG=None

Review-Url: https://codereview.chromium.org/2137013005
Cr-Commit-Position: refs/heads/master@{#405277}

The mock GL bindings are useful, but they are very complex to setup and use
(gmock), and too slow for some uses (e.g. a fuzzer), and painful to maintain as
the decoder code evolves.

This CL introduces a GLStubApi as an alternative:
- GLStubApi hooks at the GLApi level rather than the DriverGL level, that is, it
implements a full interface rather than individual functions looked up through
GetProcAddress. It is simpler to override.
- the CL implements an auto-generated GLStubApiBase which only has trivial
functions, as well as a more concrete GLStubApi that implements enough to
initialize and run a GLES2Decoder.
- GpuChannelTest was changed to use the GLStubApi

BUG=None
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2136223002
Cr-Commit-Position: refs/heads/master@{#405276}

The current Dockerfile uploaded with the testing package is just a NOOP stub.  This change creates a functional stub from the shared base used by test and engine and handles additional filesystem setup steps for testing.

The updated Dockerfile will be used to validate service support.

BUG=616945, 608487

Review-Url: https://codereview.chromium.org/2136363002
Cr-Commit-Position: refs/heads/master@{#405275}

VideoCaptureDeviceClient buffer warnings show that capture is
stopped. In most of the capture related errors, capture stops
and we print this warning for each following frame. As a result,
it is much harder to read through the error logs, such as below
as we print this line hundreds of times.

This CL addresses this issue by removing the logs and replacing with
a DCHECK that comes in after consistent errors.

https://build.chromium.org/p/chromium.webrtc/builders/Win8%20Tester/builds/26646/steps/browser_tests/logs/stdio

Review-Url: https://codereview.chromium.org/2124073006
Cr-Commit-Position: refs/heads/master@{#405274}

Users can add "use_new_wrapper_types = true" to their mojom targets to make the generator map mojom array/map/string to STL types (for chromium variant) or WTF types (for blink variant).

This CL also switches bindings tests to use the new mode

BUG=624136

Review-Url: https://codereview.chromium.org/2136733002
Cr-Commit-Position: refs/heads/master@{#405273}

This removes trace endpoints from the content interface and moves them
to TracingControllerImpl. We don't want to expose both TraceDataSinks
and TraceEndpoints currently to avoid confusion.

Review-Url: https://codereview.chromium.org/2052773002
Cr-Commit-Position: refs/heads/master@{#405272}

TBR=dgozman@chromium.org

Review-Url: https://codereview.chromium.org/2145123002
Cr-Commit-Position: refs/heads/master@{#405271}

Shell operations can cause 3rd-party shell extensions to be
loaded into the caller's process. The utility process protects the browser process from potential instability.

Also convert the IPC mechanism to Mojo.

BUG=73098

Review-Url: https://codereview.chromium.org/2122303002
Cr-Commit-Position: refs/heads/master@{#405270}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/8b7bf3c29a56..04bbb644880f

$ git log 8b7bf3c29..04bbb6448 --date=short --no-merges --format='%ad %ae %s'

BUG=577224

TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2140173005
Cr-Commit-Position: refs/heads/master@{#405269}

flaky on Windows 7 and fix TestExpectations for picker appearance tests.

BUG=627782,593567
TBR=engedy

Review-Url: https://codereview.chromium.org/2150523003
Cr-Commit-Position: refs/heads/master@{#405268}

SurfaceId is 128-bits and so it cannot fit in a single register.
Thus passing it around by const reference is potentially more
efficient, and more consistent with how we pass around other non-trivial
structs such as gfx::Rect.

BUG=none
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review-Url: https://codereview.chromium.org/2150633002
Cr-Commit-Position: refs/heads/master@{#405267}

BUG=622707
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2133833002
Cr-Commit-Position: refs/heads/master@{#405266}

Increase allowed percentage of pixels that differ in ImageBackgroundFilter.BackgroundFilterRotated_GL test

The old value was 3.1% and on Windows a Skia change causes slightly more pixels to be different. It is increased to 4%.
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review-Url: https://codereview.chromium.org/2145043002
Cr-Commit-Position: refs/heads/master@{#405265}

This reverts r403440 (commit c825d655),
as it caused a security regression.

BUG=626790
TBR=shanmuga.m@samsung.com

Review-Url: https://codereview.chromium.org/2147003002
Cr-Commit-Position: refs/heads/master@{#405264}

Now that cpplint.py doesn't trigger a false positive with
base::hash_map, we don't need the NOLINT tag anymore.

BUG=627202
TEST=git cl lint  # clean output

Review-Url: https://codereview.chromium.org/2146833003
Cr-Commit-Position: refs/heads/master@{#405263}

Implements migration of a QUIC connection to a different destination address if specified by the server in an ASAD tag in the Server Hello. Quietly falls back to original socket if creation and connecting a new socket fails.

R=rch@chromium.org
BUG=

Review-Url: https://codereview.chromium.org/2124753005
Cr-Commit-Position: refs/heads/master@{#405262}

Use the cursor theme that GTK tells us is the default,
instead of always using the default white cursor theme.

BUG=377525

Review-Url: https://codereview.chromium.org/2135353002
Cr-Commit-Position: refs/heads/master@{#405261}

All clients have transitioned to using the respective devil modules
directly, so these are no longer needed.

BUG=617761

Review-Url: https://codereview.chromium.org/2139133004
Cr-Commit-Position: refs/heads/master@{#405260}