GitLab

Prior to this commit, the IDL compiler was failing to resolve typedefs in
unions before generating new union code. In practice, this meant an IDL
construct such as:

    typedef boolean Foo;
    void func((Foo or long) arg);

Generated FooOrLong.cpp and FooOrLong.h, and "Foo" was considered an IDL
interface instead of a simple bool. Since Foo.cpp and Foo.h do not exist,
and neither does the Foo class, this caused build failures.

We now resolve the typedefs earlier, and in the example above we would now
generate BooleanOrLong.cpp and BooleanOrLong.h, which works as expected.

Fixing this requires a few separate, intertwined changes:
* IdlUnionType.resolve_typedefs() was broken: the |typedefs| argument it
  receives is a dictionary whose keys are typedef names, not
  IdlTypeBase-derived objects, so |typedefs.get()| was always failing and
  returning the original, unresolved member objects.
* CodeGeneratorUnionType actually needs to call resolve_typedefs(). It was
  the only generator class not doing so. The mechanism needs to be different
  from the other classes though, as IdlUnionType is not a type defined in
  idl_definitions.py as is the case for the types handled by the other
  generators. Instead of using TypedefResolver, we simply call the relevant
  parts of it manually.
* The code manually adding union header includes to interfaces and partial
  interfaces has been dropped: there isn't enough information to resolve the
  typedefs at that point, which meant interface .cpp files would include
  FooOrLong.h, while interface .h files include BooleanOrLong.h. The test
  expectations IDL files show a reduction in unneeded includes, and the
  build seems to work fine.

BUG=701411
R=bashi@chromium.org,haraken@chromium.org,yukishiino@chromium.org

Review-Url: https://codereview.chromium.org/2749253002
Cr-Commit-Position: refs/heads/master@{#457377}

This CL avoids VisibleSelectionInFlatTree creation if VSInDOMTree is none
 in SelectionEditor.cpp.
This reduces redundant canonicalization because
 VSInFlatTree availability == VSInDOMTree availability

BUG=668122
TEST=
 1. Open the page on comment#21 described at the bug.
 2. Open also chrome://tracing/ and record Input Latency.
 3. Confirm Messaseloop::RunTask blocking time gets faster 3.7 times(from 64ms to 17ms).

Review-Url: https://codereview.chromium.org/2741173002
Cr-Commit-Position: refs/heads/master@{#457376}

This CL has no behavior changes.

Review-Url: https://codereview.chromium.org/2755713003
Cr-Commit-Position: refs/heads/master@{#457375}

Add ExtensionUnpacker mojo, used to (unzip and) unpack a chrome
extension, and expose it to the browser via the utility process
policy file. Make callers use a utility process mojo client.

Add [Native] IPC enum param traits via a mojo typemap to handle
extensions::Manifest::Location. Set the wire limit for the enum
value to one less than Manifest::NUM_LOCATIONS (for compat with
the pre-existing custom traits removed in this patch since this
patch auto-generates them: manifest_location_param_traits.cc).

In the extensions/utility/utility_handler.cc, retain the CHECKs
on the Manifest::Location enum limits while bug 692069 is still
being investigated.

Move DecodeImages declarations from the extension messages-file
into its own file (extension_utility_types.h) to decouple those
declarations from this messages-file. A follow-up patch deletes
the messages-file so DecodeImages needs a new home.

Update all BUILD.gn for using clients to directly depend on the
//extensions/common:common target (and hence, its :mojo) to fix
post-compile step failures (warning about under-specified BUILD
files via "targets is_dirty" errors, see review comment #169).

Add TestContentBrowserClient to the extension test harness, and
also an overlay file in test/data (needed to expose the mojo in
the extension test harness to the unit tests listed below).

Covered by existing unit tests:
  {sandboxed_unpacker, zipfile_installer}_unittest.cc

BUG=691410

Review-Url: https://codereview.chromium.org/2697463002
Cr-Commit-Position: refs/heads/master@{#457374}

Primarily to address an issue we're seeing in Settings where initial clicks
on certain row items aren't being registered.

BUG=697353
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2751523005
Cr-Commit-Position: refs/heads/master@{#457373}

This CL does
- replaces m_scriptState->isolate() with an isolate pointer
- uses executionContext pointer
- folds not-parametarized long lines

BUG=660281

Review-Url: https://codereview.chromium.org/2758463002
Cr-Commit-Position: refs/heads/master@{#457372}

base::ScopedVector is deprecated, see bug.

BUG=554289
TBR=tsepez@chromium.org

Review-Url: https://codereview.chromium.org/2749203002
Cr-Commit-Position: refs/heads/master@{#457371}

(Otherwise everyone gets a presubmit failure if they change this file).

TBR=lgarron@chromium.org
NOTRY=true

Review-Url: https://codereview.chromium.org/2745253009
Cr-Commit-Position: refs/heads/master@{#457370}

Pass a Java long to C++ as a jlong type, not a C++ long which can have a
different size to avoid broken / negative event timestamp values.
ImeAdapter's Java side uses "long" in SendKeyEvent, so the C++ side must
use a jlong or int64_t, and not a C++ long. Otherwise things don't work
well when system uptime is over 2^31ms (~25 days).

Additionally, do not do an extra divide-by-1000 when the used helper
function will do the milliseconds to seconds conversion already, so the
timestamps are correctly measured in milliseconds.

The resulting keyboard event timestamps end up nicely sane and positive,
and no longer clamped to 0 in PerformanceBase.cpp.

BUG=701726
R=aelias@chromium.org

Review-Url: https://codereview.chromium.org/2755453004
Cr-Commit-Position: refs/heads/master@{#457369}

To enable tracing from chrome://tracing, we create a Mojo interface
for host side to trigger tracing in Android.

The trace is written into debugfs/tracefs in dev-mode, and is read back
by debugd.

Related changes:
- http://crrev.com/2749283003
- http://ag/1915548

BUG=653795
TEST=Run chrome://tracing and see events from Android.

Review-Url: https://codereview.chromium.org/2749283003
Cr-Commit-Position: refs/heads/master@{#457368}

Currently |TextControlElement::selection()| is called
only by Editor::selectionForCommand()
and returned |Range| is used for constructing |EphemeralRange|,
which again is used for constructing |SelectionInDOMTree|.

This CL makes |TextControlElement::selection()|
to return |SelectionInDOMTree| and avoid creation of Range.

BUG=691193

Review-Url: https://codereview.chromium.org/2750833002
Cr-Commit-Position: refs/heads/master@{#457367}

Revert of Initialize document m_URL, m_baseURL to blankURL. (patchset #4 id:60001 of https://codereview.chromium.org/2749803003/ )

Reason for revert:
This CL seems to break interactive_ui_tests on multiple bots:

A sample of error log was:
https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Ozone%20Tests%20%281%29/builds/44135

../../chrome/test/base/extension_js_browser_test.cc:64: Failure
Failed
Failed: RUN_TEST_F("AutomationTreeWalkerTest","Backward")
ReferenceError: DesktopAutomationHandler is not defined
    at AutomationTreeWalkerTest.ChromeVoxNextE2ETest (tree_walker_test.extjs:8556:3)
    at AutomationTreeWalkerTest (tree_walker_test.extjs:8663:24)
    at createTestCase (tree_walker_test.extjs:7607:19)
    at RUN_TEST_F (tree_walker_test.extjs:7721:25)
    at tree_walker_test.extjs:7523:20
    at runTestFunction (tree_walker_test.extjs:7593:37)
    at runTest (tree_walker_test.extjs:7567:18)
    at tree_walker_test.extjs:8851:1
gen/chrome/browser/resources/chromeos/chromevox/cvox2/background/tree_walker_test-gen.cc:69: Failure
Value of: RunJavascriptTestF( true, "AutomationTreeWalkerTest", "Backward")
  Actual: false
Expected: true
[  FAILED  ] AutomationTreeWalkerTest.Backward, where TypeParam =  and GetParam() =  (1196 ms)

Original issue's description:
> Initialize document m_URL & m_baseURL to blankURL.
>
> This patch will ensure that m_URL and m_baseURL are initialized to
> blankURL (about:blank) as is specified in [1]. This patch fixes the
> tests in [2] and [3].
>
> [1] https://dom.spec.whatwg.org/#interface-document
> [2] http://w3c-test.org/dom/nodes/DOMImplementation-createDocument.html
> [3] http://w3c-test.org/dom/nodes/DOMImplementation-createHTMLDocument.html
>
> BUG=563986
>
> Review-Url: https://codereview.chromium.org/2749803003
> Cr-Commit-Position: refs/heads/master@{#457347}
> Committed: https://chromium.googlesource.com/chromium/src/+/9168f165fe92e37664a5e39fef987f0e3ad72a62

TBR=tkent@chromium.org,martin@martinrogalla.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=563986

Review-Url: https://codereview.chromium.org/2757573002
Cr-Commit-Position: refs/heads/master@{#457366}

Reland of Enable FasterLocationReload feature by default (patchset #1 id:1 of https://codereview.chromium.org/2749513009/ )

Reason for revert:
The original change seems to be innocent.

Original issue's description:
> Revert of Enable FasterLocationReload feature by default (patchset #1 id:1 of https://codereview.chromium.org/2747413002/ )
>
> Reason for revert:
> This CL seems to break interactive_ui_tests on multiple bots:
>
> https://build.chromium.org/p/chromium.mac/builders/Mac10.10%20Tests/builds/15140
> https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Ozone%20Tests%20%281%29/builds/44135
>
> Original issue's description:
> > Enable FasterLocationReload feature by default
> >
> > As a third step of experiment-controlled rollouts, enable
> > FasterLocationReload feature by default for trunk.
> >
> > BUG=591245
> >
> > Review-Url: https://codereview.chromium.org/2747413002
> > Cr-Commit-Position: refs/heads/master@{#457346}
> > Committed: https://chromium.googlesource.com/chromium/src/+/bc7bbd667335209110948cd2f07b4e3b7f44b56c
>
> TBR=kinuko@chromium.org,toyoshim@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=591245
>
> Review-Url: https://codereview.chromium.org/2749513009
> Cr-Commit-Position: refs/heads/master@{#457358}
> Committed: https://chromium.googlesource.com/chromium/src/+/2a7d3504083f59f10b9150d58a00a863056750a3

TBR=kinuko@chromium.org,tzik@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=591245

Review-Url: https://codereview.chromium.org/2746293008
Cr-Commit-Position: refs/heads/master@{#457365}

Does the exact same thing, but imo more readable.

BUG=689758

Review-Url: https://codereview.chromium.org/2750123002
Cr-Commit-Position: refs/heads/master@{#457364}

BUG=694382,698622

Review-Url: https://codereview.chromium.org/2742853003
Cr-Commit-Position: refs/heads/master@{#457363}

Revert of Fix the condensed NTP layout for tablets. (patchset #2 id:20001 of https://codereview.chromium.org/2750503008/ )

Reason for revert:
This is causing KitKat Tablet Tester to fail the chrome_public_test_apk step.

Original issue's description:
> Fix the condensed NTP layout for tablets.
>
> BUG=700947
>
> Review-Url: https://codereview.chromium.org/2750503008
> Cr-Commit-Position: refs/heads/master@{#457068}
> Committed: https://chromium.googlesource.com/chromium/src/+/aa26b1a2d7c80e970051ee002dd072f37f9a3852

TBR=bauerb@chromium.org,mvanouwerkerk@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=700947

Review-Url: https://codereview.chromium.org/2755883002
Cr-Commit-Position: refs/heads/master@{#457362}

Revert of [Android:WebAPK] Don't add webapp to homescreen if WebAPK install times out part 2/3 (patchset #2 id:20001 of https://codereview.chromium.org/2728053002/ )

Reason for revert:
This is causing build failures on clang-clankium-tot-builder
See:
https://crbug.com/702086
https://uberchromegw.corp.google.com/i/internal.client.clank_tot/builders/clang-clankium-tot-builder/builds/43251

Original issue's description:
> [Android:WebAPK] Don't add webapp to homescreen if WebAPK install times out part 2/3
>
> This CL:
> - Changes GooglePlayWebApkInstallDelegate#installAsync() to return a
>   WebApkInstallResult in the callback.
> - Deletes unused GooglePlayWebApkInstallDelegate#canInstallWebApk().
>
> BUG=696132
>
> Review-Url: https://codereview.chromium.org/2728053002
> Cr-Commit-Position: refs/heads/master@{#457300}
> Committed: https://chromium.googlesource.com/chromium/src/+/19d929ae378491116d7683764baa6365d1ddd50d

TBR=dominickn@chromium.org,pkotwicz@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=696132

Review-Url: https://codereview.chromium.org/2757563002
Cr-Commit-Position: refs/heads/master@{#457361}

This issue was introduced by
https://codereview.chromium.org/2748383002
which renames InspectorInstrumentation.idl to InstrumentingProbes.idl.

I wonder why presubmit check did not fail in the CL.

BUG=None

Review-Url: https://codereview.chromium.org/2754763002
Cr-Commit-Position: refs/heads/master@{#457360}

In the current LoadNativeLibrary implementation, LoadLibraryW Windows
API is used to load a DLL. To be able to find dependencies in the same
folder, SetCurrentDirectory() is needed to search for the DLL directory,
and sets it back after the DLL is loaded. This is required because on
Windows, it'll search for dependencies in a search list, which includes
the system "current directory", but not the DLL directory.

However, SetCurrentDirectory() can be potentially problematic. It is not
recommended in a multithreaded application, and could pose a security
risk as "If an attacker gains control of one of the directories on the
DLL search path, it can place a malicious copy of the DLL in that
directory. This is sometimes called a DLL preloading attack or a binary
planting attack."

The right thing to do is to use LoadLibraryExW, where we can specify
additional flags like LOAD_LIBRARY_SEARCH_DLL_LOAD_DIR to enable
searching in the DLL directory. With this, we can eliminate the need of
doing SetCurrentDirectory().

Using these additional flags requires KB2533623 to be installed and the
method is "To determine whether the flags are available, use
GetProcAddress to get the address of the AddDllDirectory,
RemoveDllDirectory, or SetDefaultDllDirectories function. If
GetProcAddress succeeds, the LOAD_LIBRARY_SEARCH_* flags can be used
with LoadLibraryEx."

Therefore, we can dynamically call LoadLibraryExW if the API and the
flags are available. If not or its call fails, we should use the
LoadLibraryW API.

This CL also adds UMA histogram to record the calling status of both
LoadLibraryExW and LoadLibraryW APIs. Besides, this CL removes the
LoadNativeLibraryDynamically method as it is not used anywhere.

Running Chromium built with this CL locally shows that LoadLibraryExW
call were successful for kernel32.dll and widevinecdm.dll (which caused
crbug.com/700208), but failed when loading MDMRegistration.dll.
LoadLibraryW succeeds in loading MDMRegistration.dll though.

BUG=700503,700208

Review-Url: https://codereview.chromium.org/2744043003
Cr-Commit-Position: refs/heads/master@{#457359}

Revert of Enable FasterLocationReload feature by default (patchset #1 id:1 of https://codereview.chromium.org/2747413002/ )

Reason for revert:
This CL seems to break interactive_ui_tests on multiple bots:

https://build.chromium.org/p/chromium.mac/builders/Mac10.10%20Tests/builds/15140
https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Ozone%20Tests%20%281%29/builds/44135

Original issue's description:
> Enable FasterLocationReload feature by default
>
> As a third step of experiment-controlled rollouts, enable
> FasterLocationReload feature by default for trunk.
>
> BUG=591245
>
> Review-Url: https://codereview.chromium.org/2747413002
> Cr-Commit-Position: refs/heads/master@{#457346}
> Committed: https://chromium.googlesource.com/chromium/src/+/bc7bbd667335209110948cd2f07b4e3b7f44b56c

TBR=kinuko@chromium.org,toyoshim@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=591245

Review-Url: https://codereview.chromium.org/2749513009
Cr-Commit-Position: refs/heads/master@{#457358}

This is a follow-up CL for https://crrev.com/2735823006/

This CL...
- merges isolatedWorldMap() into worldMap() for simplification,
- fixes ArrayBuffer accumulation to correctly handle all worlds on non-main
  thread, and
- adds unit tests for DOMWrapperWorld.

BUG=697622, 697629

Review-Url: https://codereview.chromium.org/2747163002
Cr-Commit-Position: refs/heads/master@{#457357}

Rename the values of the EDisplay enum with the k prefix to match
the style generated by make_computed_style_base.py, which is in line with
the Blink rename to match Chromium style. This is prework for generating
EDisplay.

BUG=684966

Review-Url: https://codereview.chromium.org/2670643004
Cr-Commit-Position: refs/heads/master@{#457356}

We should have removed editing/deleting/password-delete-performance.html
 not password-delete-contents.html at
https://codereview.chromium.org/2730433004/

BUG=698626

Review-Url: https://codereview.chromium.org/2753743004
Cr-Commit-Position: refs/heads/master@{#457355}

This effectively merges ConsoleModel and MultitargetConsoleModel.

This is in preparation to extracting ConsoleModel out of SDK and
turning it into ConsoleView's presentation model.

BUG=none

Review-Url: https://codereview.chromium.org/2751173003
Cr-Commit-Position: refs/heads/master@{#457354}

Re-landing this change:

https://codereview.chromium.org/2730343003/

... along with the missing test expectations.

TBR=tzik@chromium.org
BUG=697751
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2757533002
Cr-Commit-Position: refs/heads/master@{#457353}

This eliminates the unnecessary distinction between 'initialState' and
'nextState'.

BUG=697706

Review-Url: https://codereview.chromium.org/2751543005
Cr-Commit-Position: refs/heads/master@{#457352}

It looks like this code was added for the old hotwording extension which
was removed. It's hard to tell how
MediaStreamDevicesController::ReqestPermission would be working from
this context if it was being used.

BUG=596786

Review-Url: https://codereview.chromium.org/2747943003
Cr-Commit-Position: refs/heads/master@{#457351}

Change log:
https://github.com/google/closure-compiler/compare/72c88ff7e9e85b4c2413274539897b6d43024d48...b2f2d2f3309c7ee61816e068050651af27bccdfa
chrome_extensions.js: 9f5d27bb1a9ad197c67f9eab61ea511d99aa4e8c -> a3f41820d1260b0d7ed13acac526b803ac6c1870

TBR=fukino@chromium.org
BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2753443007 .
Cr-Commit-Position: refs/heads/master@{#457350}

BUG=700581
TBR=jochen@chromium.org

Review-Url: https://codereview.chromium.org/2748443005
Cr-Commit-Position: refs/heads/master@{#457349}

To verify end to end reload behaviors against pages that contains
iframes, update ReloadCacheControlBrowserTestt to check cache
control flags against a page containing a simple frame in addition
to an image resource.

BUG=602900

Review-Url: https://codereview.chromium.org/2756443002
Cr-Commit-Position: refs/heads/master@{#457348}

This patch will ensure that m_URL and m_baseURL are initialized to
blankURL (about:blank) as is specified in [1]. This patch fixes the
tests in [2] and [3].

[1] https://dom.spec.whatwg.org/#interface-document
[2] http://w3c-test.org/dom/nodes/DOMImplementation-createDocument.html
[3] http://w3c-test.org/dom/nodes/DOMImplementation-createHTMLDocument.html

BUG=563986

Review-Url: https://codereview.chromium.org/2749803003
Cr-Commit-Position: refs/heads/master@{#457347}

As a third step of experiment-controlled rollouts, enable
FasterLocationReload feature by default for trunk.

BUG=591245

Review-Url: https://codereview.chromium.org/2747413002
Cr-Commit-Position: refs/heads/master@{#457346}

The only thing they provided is already inside of compiler.jar's
externs.zip/browser/html5.js at this point (HTMLImportElement#import).

R=tbreisacher@chromium.org
BUG=697353
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2754733002
Cr-Commit-Position: refs/heads/master@{#457345}

This CL is a follow-up of https://codereview.chromium.org/2745443002.

blink::WebScreenOrientationLockType traits is defined in
screen_orientation_messages.h but has only one user: manifest_manager_messages.h,
so we just move it into manifest_manager_messages.h, thus
screen_orientation_messages.h becomes empty and we can remove it completely.

BUG=678545

Review-Url: https://codereview.chromium.org/2749863004
Cr-Commit-Position: refs/heads/master@{#457344}

Revert of Null check for IME descriptor in GetKeyboardLayoutsForResovledLocale (patchset #3 id:40001 of https://codereview.chromium.org/2756523003/ )

Reason for revert:
landed a better workaround

Original issue's description:
> Null check for IME descriptor in GetKeyboardLayoutsForResovledLocale
>
> This is workaround for layouts returned by GetHardwareInputMethodIds
> not having a valid IME descriptor.
>
> BUG=700625
>
> Review-Url: https://codereview.chromium.org/2756523003
> Cr-Commit-Position: refs/heads/master@{#457228}
> Committed: https://chromium.googlesource.com/chromium/src/+/a6ba05eb3482758392c1e440f853f4e612140122

TBR=alemate@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=700625

Review-Url: https://codereview.chromium.org/2746363005
Cr-Commit-Position: refs/heads/master@{#457343}

Revert of Mojo EDK: Introduce MojoQueryHandleSignalsState API (patchset #9 id:160001 of https://codereview.chromium.org/2741033003/ )

Reason for revert:
This CL seems to break mojo_system_unittests on multiple bots:

https://build.chromium.org/p/chromium.linux/builders/Linux%20Tests/builds/53232
https://build.chromium.org/p/chromium.win/builders/Win7%20Tests%20%281%29/builds/64801
https://build.chromium.org/p/chromium.mac/builders/Mac10.11%20Tests/builds/9606

Here are samples of the failing tests. Most of them are failed by time out:
MessagePipeTest.DiscardMode
WatcherTest.WatchDataPipeConsumerReadable
WatcherTest.WatchMessagePipeReadable
MessagePipeTest.Basic
DataPipeTest.PeerClosedProducerWaiting

Original issue's description:
> Mojo EDK: Introduce MojoQueryHandleSignalsState API
>
> The only reliable way to inquire about handle signals now
> is to MojoWait (for e.g. 0 deadline). As a precursor to
> removing the wait APIs in favor of watchers, we need to
> retain the ability to efficiently query a handle's signals
> state.
>
> Rather than trying to retrofit the watcher APIs to support
> this use case in similar fashion to the wait APIs, this adds
> an API explicitly for the purpose of querying signals state.
>
> Adds a corresponding method to the C++ mojo::Handle
> and moves the EDK's internal HandleSignalsState helper class
> to mojo/public/cpp/system, adding some convenient accessors.
>
> Also introduces the API to the JS and Java libraries, and
> replaces any 0-deadline waits in those languages with usage of
> this new API.
>
> Because waitMany is not used in these languages (except for
> tests which test waitMany...) it has been removed. wait() is
> unused in Java after this change, so it has also been removed.
>
> Finally, this moves several tests away from calling MojoWait
> directly, instead using a simplified Watcher-based wait
> implementation in MojoTestBase.
>
> BUG=700171
> TBR=jam@chromium.org
>
> Review-Url: https://codereview.chromium.org/2741033003
> Cr-Commit-Position: refs/heads/master@{#457315}
> Committed: https://chromium.googlesource.com/chromium/src/+/853496a78ae997c2d8b80f3cd8fabf9423fb3361

TBR=yzshen@chromium.org,rockot@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=700171

Review-Url: https://codereview.chromium.org/2750273002
Cr-Commit-Position: refs/heads/master@{#457342}

When modifying selection programmatically by the method
FrameSelection::setSelectedRange preserve selection handles visibility.

BUG=685420

Review-Url: https://codereview.chromium.org/2701083002
Cr-Commit-Position: refs/heads/master@{#457341}

 to document.

In the condition, Range::expand calls create VisiblePosition and
 DCHECK(node.isConnected()) failed.
This CL lets Range::expand not do nothing in the condition.

BUG=698123

Review-Url: https://codereview.chromium.org/2747113003
Cr-Commit-Position: refs/heads/master@{#457340}

This patch fixes to handle a text node that contains only a collapsible
newline.

Before this change, collapsible newlines were lazily appended, when
following text is appended. When a text node contains only collapsible
newlines, NGLayoutInlineItemsBuilder does not keep the node and thus
unable to append lazily.

This patch changes to append pending collapsible newlines first, then
remove later if it needed to be collapsed.

BUG=636993

Review-Url: https://codereview.chromium.org/2749013003
Cr-Commit-Position: refs/heads/master@{#457339}

We need type info when copying video texture to another texture. 0 is
not a valid texture type. 0 type causes DoCopySubTextureCHROMIUM to
fail.

Format/internalformat info was added in the following two CLs:
https://codereview.chromium.org/775863005
https://codereview.chromium.org/759573002

BUG=612542
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2711903002
Cr-Commit-Position: refs/heads/master@{#457338}