GitLab

In release mode a DCHECK was compiled out and caused a crash; it was
correctly asserting in the Debug builds. Change the DCHECK into a
runtime check that throws an exception. This addresses the cluster fuzz
crash.

BUG=472058
TEST=out/Release/content_shell --dump-render-tree --use-gl=any --disable-gl-drawing-for-test file:///work/issues/472058/test.html

Review URL: https://codereview.chromium.org/1140813007

Cr-Commit-Position: refs/heads/master@{#330659}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/9db912c..a65358c

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

Review URL: https://codereview.chromium.org/1146783002

Cr-Commit-Position: refs/heads/master@{#330658}

This is an incremental fix for 438313 that addresses the two issues mentioned here:
https://code.google.com/p/chromium/issues/detail?id=438313#c24
This issues are reproducible only with the alpha server with
implicit permanent folders enabled. I added a test case for
one of them - the out of order deletion which results in
deleting the permanent folder for a datatype before some of
the items of the same datatype. That occurs only in one
particular corner case when a failed datatype need to be
purged and re-downloaded.

BUG=438313

Review URL: https://codereview.chromium.org/1139883007

Cr-Commit-Position: refs/heads/master@{#330657}

Revert of do not hardcode -I/usr/include32 when cross-compiling (patchset #1 id:1 of https://codereview.chromium.org/1132853010/)

Reason for revert:
Looks like this broke some builds, where 'use_sysroot' isn't defined. See http://crbug.com/489783.

Sounds like sbc has an alternative solution here: https://codereview.chromium.org/1142793003/

Original issue's description:
> do not hardcode -I/usr/include32 when cross-compiling
>
> When we cross-compile, we do not want any /usr/include paths as that will
> conflict with our own system headers.  Disable the hardcoded path when we
> have an active sysroot.
>
> BUG=chromium:488360
>
> Committed: https://crrev.com/58b42a0c48127c60baad7ea79ea0e5576c6e1d0b
> Cr-Commit-Position: refs/heads/master@{#330514}

TBR=jochen@chromium.org,cjhopman@chromium.org,dpranke@chromium.org,phajdan.jr@chromium.org,scottmg@chromium.org,thakis@chromium.org,vapier@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=488360,489783

Review URL: https://codereview.chromium.org/1147793003

Cr-Commit-Position: refs/heads/master@{#330656}

Make this work independent for MsgLoop vs. ChildThread destruction
order and simplify the code a bit.

BUG=465397

Review URL: https://codereview.chromium.org/1140633005

Cr-Commit-Position: refs/heads/master@{#330655}

Before this change InvalidationService CHECKs that object id is only
registered with one handler. This causes browser crash when object id is
received from network already registered.

Solution is to let each component that uses InvalidationService decide
how to react to duplicate registration.
In this change:
- InvalidationService::UpdateRegisteredInvalidationIds doesn't CHECK
  inside, but returns boolean that indicates if update was successful.
- All places that call this function do CHECK to preserve existing
  behavior.
- Internal implementations and tests are updated accordingly.

BUG=475941
R=maniscalco@chromium.org

Review URL: https://codereview.chromium.org/1146533005

Cr-Commit-Position: refs/heads/master@{#330654}

Revert of bluetooth: Move testing IPC from BluetoothDispatcher to BlinkTestRunner (patchset #2 id:250001 of https://codereview.chromium.org/1125133005/)

Reason for revert:
Broke a bunch of Layout Tests. Here's an example: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&tests=virtual%2Fstable%2Fanimations-unprefixed%2Fanimation-events-prefixed-04.html

STDERR: ==28151==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000050 (pc 0x000006a00bd0 bp 0x7fff171be890 sp 0x7fff171be880 T0)
STDERR:     #0 0x6a00bcf in content::BluetoothDispatcherHost::SetBluetoothAdapterForTesting(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) content/browser/bluetooth/bluetooth_dispatcher_host.cc:63:43
STDERR:     #1 0x4e5d9d in OnSetBluetoothAdapter content/shell/browser/layout_test/layout_test_message_filter.cc:171:3
STDERR:     #2 0x4e5d9d in DispatchToMethodImpl<content::LayoutTestMessageFilter, void (content::LayoutTestMessageFilter::*)(const std::__1::basic_string<char> &), std::__1::basic_string<char> , 0> base/tuple.h:252:0

Original issue's description:
> bluetooth: Move testing IPC from BluetoothDispatcher to BlinkTestRunner
>
> This patch removes the testing IPC from BluetoothDispatcher and
> BluetoothDispatcherHost.
>
> This patch also changes the way the mock BluetoothAdapter was set in
> BluetoothDispatcherHost. Instead of receiving an IPC to set the adapter,
> BluetoothDispatcherHost exposes a function to directly set the adapter.
> This function is used by LayoutTestSupport to set the adapter.
>
> Mock adapter flow before:
> BlinkTestRunner -> LayoutTestSupport -> BluetoothDispatcher --IPC-->
> BluetoothDispatcherHost (Mock constructed here)
>
> After:
> BlinkTestRunner --IPC--> LayoutTestMessages(Mock constructed here) ->
> LayoutTestSupport -> BluetoothDispatcherHost
>
> This is the first of two patches to remove testing from BluetoothDispatcher
> and BluetoothDispatcherHost:
>
>  [1] This patch.
>  [2] http://crrev.com/1132943002
>
> BUG=436284
>
> Committed: https://crrev.com/3f7142d0acf5e930743cbe5d754084c464ac3c85
> Cr-Commit-Position: refs/heads/master@{#330647}

TBR=jam@chromium.org,scheib@chromium.org,tsepez@chromium.org,ortuno@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=436284

Review URL: https://codereview.chromium.org/1142303002

Cr-Commit-Position: refs/heads/master@{#330653}

Revert of Temporarily disable a webgl conformance test on D3D9 only. (patchset #1 id:1 of https://codereview.chromium.org/1135333004/)

Reason for revert:
Bug fixed.  Re-enabling the test.

Original issue's description:
> Temporarily disable a webgl conformance test on D3D9 only.
>
> BUG=488552
> TEST=webgl conformacne with D3D9 backend
> R=kbr@chromium.org
> NOTRY=true
>
> Committed: https://crrev.com/4fbdae9877d2b4775efa0dffdaa9c16ed6689c75
> Cr-Commit-Position: refs/heads/master@{#330138}

TBR=kbr@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=488552

Review URL: https://codereview.chromium.org/1124083006

Cr-Commit-Position: refs/heads/master@{#330652}

BUG=488146

Review URL: https://codereview.chromium.org/1130773004

Cr-Commit-Position: refs/heads/master@{#330651}

Revert of Fix missing GN dependencies. (patchset #4 id:60001 of https://codereview.chromium.org/1128163007/)

Reason for revert:
Broke Windows:
http://build.chromium.org/p/chromium.win/builders/Win%20x64%20GN/builds/1568/steps/compile/logs/stdio

Original issue's description:
> Fix missing GN dependencies.
>
> Fixes some cases where a dependency between a target's inputs and outputs exists without an explicit dependency.
>
> json_schema_api.gni was generating wrong names, this was fixed.
>
> mojo_application_package.gni was fixed to generate unique names (just a bug I noticed in passing) and support the testonly flag, which is necessary when I added the correct dependency.
>
> The rest of the cases are just adding deps or making existing deps public so that the ability to depend on the target's outputs is passed to dependents.
>
> BUG=487897
>
> Committed: https://crrev.com/eadea21096be21b0d74c93f5b1a0ae6a9cc057d7
> Cr-Commit-Position: refs/heads/master@{#330636}

TBR=dpranke@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=487897

Review URL: https://codereview.chromium.org/1137693006

Cr-Commit-Position: refs/heads/master@{#330650}

Add a layout and constants for supporting a menu
item with three icons in it.

BUG=478984

Review URL: https://codereview.chromium.org/1138993011

Cr-Commit-Position: refs/heads/master@{#330649}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/b1dbd0d..90421e2

TBR=hiroshige@chromium.org,szager@chromium.org,leviw@chromium.org

Review URL: https://codereview.chromium.org/1138823007

Cr-Commit-Position: refs/heads/master@{#330648}

This patch removes the testing IPC from BluetoothDispatcher and
BluetoothDispatcherHost.

This patch also changes the way the mock BluetoothAdapter was set in
BluetoothDispatcherHost. Instead of receiving an IPC to set the adapter,
BluetoothDispatcherHost exposes a function to directly set the adapter.
This function is used by LayoutTestSupport to set the adapter.

Mock adapter flow before:
BlinkTestRunner -> LayoutTestSupport -> BluetoothDispatcher --IPC-->
BluetoothDispatcherHost (Mock constructed here)

After:
BlinkTestRunner --IPC--> LayoutTestMessages(Mock constructed here) ->
LayoutTestSupport -> BluetoothDispatcherHost

This is the first of two patches to remove testing from BluetoothDispatcher
and BluetoothDispatcherHost:

 [1] This patch.
 [2] http://crrev.com/1132943002

BUG=436284

Review URL: https://codereview.chromium.org/1125133005

Cr-Commit-Position: refs/heads/master@{#330647}

- Update comment about "cenc" and "webm" support.
- Support keyids.

BUG=488759

Review URL: https://codereview.chromium.org/1128143005

Cr-Commit-Position: refs/heads/master@{#330646}

Cr-Commit-Position: refs/heads/master@{#330645}

The exp_directive test verifies the experiment flag is passed to the
page, by bypassing any request to http://aws1.mdw.la/exptest/... when
the flag --data-reduction-proxy-experiment=test is set.  Which works
as expected.

However the favicon request from http://aws1.mdw.la/favicon.ico won't
be bypassed, causing the test to fail.  To fix the test, I'm adding an
optional url_pattern parameter to the bypass verification, that will
verify only /exptest/ urls are checked.

BUG=488232

Review URL: https://codereview.chromium.org/1141923004

Cr-Commit-Position: refs/heads/master@{#330644}

Review URL: https://codereview.chromium.org/1130153009

Cr-Commit-Position: refs/heads/master@{#330643}

This restores the ability for the prefixed EME API to use secure codecs when the renderer preference is set.

BUG=489512

Review URL: https://codereview.chromium.org/1132323004

Cr-Commit-Position: refs/heads/master@{#330642}

WebRTC 9217:9227
Changes: https://chromium.googlesource.com/external/webrtc/trunk/webrtc.git/+log/5deb53c..6859ada

Libjingle 9216:9227
Changes: https://chromium.googlesource.com/external/webrtc/trunk/talk.git/+log/44d6a85..88eeffa

R=mcasas@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1137273003

Cr-Commit-Position: refs/heads/master@{#330641}

Composited selection coordinates are in floating pointer DIPs. However,
current APIs for manipulating the selection are all in integral
coordinates. This leads to cases where an adjusted selection can
chaotically hop across selection region boundaries.

Until all platforms adopt unified selection and we transition the APIs
to use floating point coordinates, use rounded coordinates when
manipulating the selection from the browser.

BUG=489852

Review URL: https://codereview.chromium.org/1139043004

Cr-Commit-Position: refs/heads/master@{#330640}

Depends on https://codereview.chromium.org/1134753005 to turn off
verify_property_trees everywhere so that this doesn't break the world.

This verification is turned off by default, and can be enabled
locally everywhere by reverting the above patch.

R=ajuma@chromium.org,vollick@chromium.org
BUG=481585

Review URL: https://codereview.chromium.org/1135393007

Cr-Commit-Position: refs/heads/master@{#330639}

The crash was probably due to accessing a dangling pointer to the plugin
container during a small time frame between PepperWebPluginImpl::destroy()
and the destructor being called. (Speculated from source since no reliable
repro is found.)

This CL clears eveything in the destroy() function as if the destructor has
been called, only delaying memory release.

In additional to the original CL https://codereview.chromium.org/1137663006/
This CL also switched the destruction order of pepper instance and the
throttler. Engaging throttle during destruction cause crash and doesn't make
sense anyway.

R=tommycli,raymes
BUG=483068,487607

Review URL: https://codereview.chromium.org/1141793002

Cr-Commit-Position: refs/heads/master@{#330638}

This is a pretty small optimization, but it saves a few struct
copies, which seems to make a small difference (~1% improvement) in a
targeted perftest that runs this function.

R=enne
BUG=488636

Review URL: https://codereview.chromium.org/1130343005

Cr-Commit-Position: refs/heads/master@{#330637}

Fixes some cases where a dependency between a target's inputs and outputs exists without an explicit dependency.

json_schema_api.gni was generating wrong names, this was fixed.

mojo_application_package.gni was fixed to generate unique names (just a bug I noticed in passing) and support the testonly flag, which is necessary when I added the correct dependency.

The rest of the cases are just adding deps or making existing deps public so that the ability to depend on the target's outputs is passed to dependents.

BUG=487897

Review URL: https://codereview.chromium.org/1128163007

Cr-Commit-Position: refs/heads/master@{#330636}

This behavior is behind the "DataReductionProxyUseDataSaverOnVPN"
Finch trial, so Data Saver will not be disabled only for users who are part of this trial.

BUG=462497

Review URL: https://codereview.chromium.org/1123173004

Cr-Commit-Position: refs/heads/master@{#330635}

The problem is that |response| was passed in the Bind call before the body was taken out. I saw this because of null dereference on Windows. The order of evaluation is undefined, which is why it worked on Linux.

Review URL: https://codereview.chromium.org/1138443004

Cr-Commit-Position: refs/heads/master@{#330634}

Move the remaining notifications from ExtensionSettingsHandler to the
DeveloperPrivate API.
Also adjust the WarningService observer method to include a set of modified
extensions.

BUG=461039

Review URL: https://codereview.chromium.org/1146693004

Cr-Commit-Position: refs/heads/master@{#330633}

This command will find an annotate dependencies between two targets in
the tree. I wanted this when tracking down some dependency issues so
wrote a tool for it.

Review URL: https://codereview.chromium.org/1130423006

Cr-Commit-Position: refs/heads/master@{#330632}

This is the experimental IDBIndex.getAll implementation for
retrieving multiple values in a single request as proposed in
https://www.w3.org/2008/webapps/wiki/IndexedDatabaseFeatures

Mozilla's documentation for IDBIndex.getAll is at:
https://developer.mozilla.org/en-US/docs/Web/API/IDBIndex/getAll

In order to use run Chrome with: --enable-experimental-web-platform-features

BUG=457450

Review URL: https://codereview.chromium.org/1147433002

Cr-Commit-Position: refs/heads/master@{#330631}

This test has been failing on cros trunk builder for several days.

BUG=480491

Review URL: https://codereview.chromium.org/1128153007

Cr-Commit-Position: refs/heads/master@{#330630}

This is so that we can use the telemetry/examples folder for putting an examples
end to end benchmark.

BUG=482227

Review URL: https://codereview.chromium.org/1111103002

Cr-Commit-Position: refs/heads/master@{#330629}

Instances of this class are created in an upcoming patch.

BUG=471928

Review URL: https://codereview.chromium.org/1123153002

Cr-Commit-Position: refs/heads/master@{#330628}

BUG=487849
TEST=manually

Review URL: https://codereview.chromium.org/1142253002

Cr-Commit-Position: refs/heads/master@{#330627}

Component build requires EXPORT macro when code from content
utilizes this class, as used in upcoming
https://codereview.chromium.org/1125133005

BUG=489330

Review URL: https://codereview.chromium.org/1140813004

Cr-Commit-Position: refs/heads/master@{#330626}

Revert of Remember previous keyboard type and use that when switching to "none" type. (patchset #2 id:20001 of https://codereview.chromium.org/1138103003/)

Reason for revert:
Patch landed despite rejection from code reviewers.

Original issue's description:
> Remember previous keyboard type and use that when switching to "none" type.
>
> BUG=484139
>
> Committed: https://crrev.com/e8e5d60be5f3b36634a1b012ec21874a4abe0ef0
> Cr-Commit-Position: refs/heads/master@{#330551}

TBR=aurimas@chromium.org,bcwhite@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=484139

Review URL: https://codereview.chromium.org/1148503003

Cr-Commit-Position: refs/heads/master@{#330625}

BUG=487456

Review URL: https://codereview.chromium.org/1129003003

Cr-Commit-Position: refs/heads/master@{#330624}

BUG=455391, 470147

Review URL: https://codereview.chromium.org/1135173007

Cr-Commit-Position: refs/heads/master@{#330623}

Add a new method to allow a DialogDelegate to provide a custom
amount to use as the minimum padding between the extra view and the
button strip in the dialog.
Use this to adjust the padding in the extension uninstall dialog.

BUG=441377

Review URL: https://codereview.chromium.org/1141253002

Cr-Commit-Position: refs/heads/master@{#330622}

This was a missing part of putting property trees on the compositor
thread. Because these temporary property trees are being used instead of
the ones on LayerTreeImpl, it makes future calls that touch property
trees through LayerImpls have bogus indices because the trees on the
LayerTreeImpl are not the same as the temporary property trees.

R=ajuma@chromium.org,vollick@chromium.org
BUG=481585

Review URL: https://codereview.chromium.org/1146523005

Cr-Commit-Position: refs/heads/master@{#330621}

These weren't being used.

Review URL: https://codereview.chromium.org/1143793003

Cr-Commit-Position: refs/heads/master@{#330620}