GitLab

BUG=

Review URL: https://codereview.chromium.org/959303002

Cr-Commit-Position: refs/heads/master@{#318668}

BUG=456779
TBR=keybuk@chromium.org

Review URL: https://codereview.chromium.org/913773002

Cr-Commit-Position: refs/heads/master@{#318667}

BUG=461282

Review URL: https://codereview.chromium.org/952853002

Cr-Commit-Position: refs/heads/master@{#318666}

As a result of introducing templated traceImpl() function, the callee part of
"Base::trace()" call (i.e. Base::trace) may not be fully resolved at the time
of the plugin execution. This has caused another issue in cases where trace()
is intentionally omitted because the class has no fields to trace.

Consider the following example:

    class A : public GarbageCollected<A> { trace() { ... } };
    class B : public A { /* No trace() */ };
    class C : public B { trace() { B::trace(visitor); } };

The "B::trace()" call in C is actually A::trace(). If the callee part "B::trace"
is unresolved, we only know that "there is a function call which will be
resolved to A::trace()". We still want to mark B as traced when we see such
a function call.

This patch fixes this trace call detection logic so we can handle this case.
The new logic keeps going up the class hierarchy while the class is not
required to have a trace method, and if we find a class that matches the class
of the (unresolved) callee part, we mark the direct base class of the original
class as traced.

BUG=462511
R=kouhei@chromium.org
CC=zerny@chromium.org, oilpan-reviews@chromium.org

Review URL: https://codereview.chromium.org/964273002

Cr-Commit-Position: refs/heads/master@{#318665}

Cr-Commit-Position: refs/heads/master@{#318664}

Implement IdAllocator by storing complete id ranges of (min,max) in the
map instead of just ids. This should make allocating big amount of ids
scale a bit better.

Store a range by creating a mapping first id -> last id. Upon query of
id existence, find the lowest_bound with the id interested. The hit will
be found at the returned iterator position (range starts at the id) or
the position before (range starts before the id and expands to cover
it).

Changes the allocation results so that even though client has requested
AllocateIdAtOrAbove(high_id_number), the next AllocateId() will allocate
lowest available id (instead of lowest available after
high_id_number). This is why GLES2ImplementationTest::kBuffersStartId
changes.

BUG=344330

Review URL: https://codereview.chromium.org/692073002

Cr-Commit-Position: refs/heads/master@{#318663}

BUG=

Review URL: https://codereview.chromium.org/918913002

Cr-Commit-Position: refs/heads/master@{#318662}

BUG=462521
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/964343002

Cr-Commit-Position: refs/heads/master@{#318661}

Additional unit tests that caught this case are forthcoming. This is a
simple and safe patch for merging.

BUG=452607

Review URL: https://codereview.chromium.org/966833002

Cr-Commit-Position: refs/heads/master@{#318660}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/b675a73..db873d8

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

Review URL: https://codereview.chromium.org/972583003

Cr-Commit-Position: refs/heads/master@{#318659}

Check-select mode spec is here: https://docs.google.com/document/d/1kCG0AFtZSlxlkuipNHlKLY2nvjPdDCViG6RKhMI5-dc/edit

Mode transition rules:
 + Entering check-select mode
   - When two or more items are selected.
   - When checkmark area is clicked.
 + Leaving check-select mode
   - When all items are unselected.
   - When non-checkmark area on an item is clicked.
   - When direction key(UP,LEFT,etc...) results in single selection.

BUG=460220
TEST=manually tested

Review URL: https://codereview.chromium.org/962103002

Cr-Commit-Position: refs/heads/master@{#318658}

Separate out Touch Selection Orientation so that it can
be used both on Java side and native side easily.

BUG=NONE

Review URL: https://codereview.chromium.org/701823002

Cr-Commit-Position: refs/heads/master@{#318657}

HotwordServiceFactory::IsServiceAvailable() causes media device
enumeration to start as a side effect. On some platforms, this can block
resulting in a user-visible delay when opening the app list. Although
this probably doesn't affect ChromeOS, it will affect OSX if voice
search and hotwording are enabled on that platform in the future.

This was discovered as a result of investigating crbug.com/458397

Review URL: https://codereview.chromium.org/966323002

Cr-Commit-Position: refs/heads/master@{#318656}

Newly added MIDI input ports were detected and enumerated correctly,
but was not registered to listen. As a result, incoming MIDI messages
are not received at all.

BUG=n/a
TEST=manual

Review URL: https://codereview.chromium.org/960213006

Cr-Commit-Position: refs/heads/master@{#318655}

BUG=462521
TEST=out/Release/browser_tests --gtest_filter=FileManagerJsTest.ExifParser

Review URL: https://codereview.chromium.org/968793002

Cr-Commit-Position: refs/heads/master@{#318654}

MIDIPacketList can contain multiple MIDIPacket, but MIDIPacket objects
have different sizes depending on packet contents. So we should not
refer to MIDIPackets via array index, but should use MIDIPacketNext
that Core MIDI API provides to obtain the next object.

As a result, previous code may refer invalid memory region if large
MIDI messages are received at the same time.

BUG=456780, 374560

Review URL: https://codereview.chromium.org/965793002

Cr-Commit-Position: refs/heads/master@{#318653}

The previous logic was causing a memory corruption when aborting was not
invoked with the callback, but due to unmounting the file system.

As there are multiple paths for aborting, this CL adds a method IsAborted on
the Queue class so it's possible to check if a task has been aborted before.

TEST=unit_tests: *FileSystemProviderQueue*
BUG=454310

Review URL: https://codereview.chromium.org/968763002

Cr-Commit-Position: refs/heads/master@{#318652}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/54f1d4e..ed228f8

TBR=

Review URL: https://codereview.chromium.org/965293002

Cr-Commit-Position: refs/heads/master@{#318651}

This CL fixes an issue where the custom launcher page would still be
shown when the app list was opened even when setEnabled(false) was called.

BUG=462444

Review URL: https://codereview.chromium.org/958133005

Cr-Commit-Position: refs/heads/master@{#318650}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/73165bd..b675a73

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

Review URL: https://codereview.chromium.org/965273002

Cr-Commit-Position: refs/heads/master@{#318649}

This CL adds a method to create process using LowBox token on Windows. LowBox will help us tackle some of the escapes from Sandbox.

R=cpu,jschuh,rvargas,wfh,forshaw
BUG=455496

Review URL: https://codereview.chromium.org/937353002

Cr-Commit-Position: refs/heads/master@{#318648}

Fix for 'Rotate clockwise/counter clockwise' should not present in context menu in crashed PDF page.

When we are in the crashed PDf page then rotate does not have any use and it does
nothing as only sad_plugin bitmap is painted.
So rotate menu items should not present in the context menu.
Changes done to check when plugin is crashed then rotation is not allowed.
and rotate menu items are not added in context menu.

BUG=462525

Review URL: https://codereview.chromium.org/962193002

Cr-Commit-Position: refs/heads/master@{#318647}

Currently the toolbar won't get hidden until after there is some user input.

Review URL: https://codereview.chromium.org/966293002

Cr-Commit-Position: refs/heads/master@{#318646}

This fixed the remaining erros found by 'gn check', so we can enable it
on all bots.

Checked with the following command lines:

Linux
$ gn gen out-gn
$ gn check out-gn //google_apis/*

$ Android
$ gn gen out-gn-android --args='target_os="android"'
$ gn check out-android //google_apis/*

$ ChromeOS
$ gn gen out-gn-chromeos --args='host_os="chromeos"'
$ gn check out-gn-chromeos //google_apis/*

BUG=None
TEST=see above
R=dpranke@chromium.org
TBR=rogerta@chromium.org,zea@chromium.org

CQ_EXTRA_TRYBOTS=tryserver.chromium.linux:android_chromium_gn_compile_dbg,android_chromium_gn_compile_rel;tryserver.chromium.win:win8_chromium_gn_rel,win8_chromium_gn_dbg;tryserver.chromium.mac:mac_chromium_gn_rel,mac_chromium_gn_dbg

Review URL: https://codereview.chromium.org/970613002

Cr-Commit-Position: refs/heads/master@{#318645}

This requirement was an error. Code behaves correctly without nodes for
these layers.

BUG=386810

Review URL: https://codereview.chromium.org/970713003

Cr-Commit-Position: refs/heads/master@{#318644}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/2d29181..54f1d4e

TBR=

Review URL: https://codereview.chromium.org/966303002

Cr-Commit-Position: refs/heads/master@{#318643}

This CL fixes an issue with the app list start page sort order for
unlaunched apps. Folder items were being weighted too highly, causing
them to appear first. The items that are not in a folder will now
appear first in the recommendations, in app list order.

BUG=455338

Review URL: https://codereview.chromium.org/954363002

Cr-Commit-Position: refs/heads/master@{#318642}

R=dbeam@chromium.org
BUG=

Review URL: https://codereview.chromium.org/939533002

Cr-Commit-Position: refs/heads/master@{#318641}

Even though fixed position layers are constrained to keep a constant
position wrt to their "fixed container," their initial position must
account for the position of layers between them and this container.

BUG=386810

Review URL: https://codereview.chromium.org/967163002

Cr-Commit-Position: refs/heads/master@{#318640}

Example of warnings/errors/advices:
tools/gn/exec_process.h:5: Missing include guard. If you are ABSOLUTELY
sure that you don't want an include guard, then include a comment
containing override-include-guard in lieu of an include guard.
tools/gn/ninja_binary_target_writer.cc:50: Single-argument constructor
'IncludeWriter(PathOutput& path_output)' may inadvertently be used as a
type conversion constructor. Prefix the function with the 'explicit'
keyword to avoid this, or add an /* implicit */ comment to suppress this
warning.
tools/gn/standard_out.cc:98: Advice: Prefer `nullptr' to `NULL' in new
C++ code.  Unlike `NULL', `nullptr' can't accidentally be used in
arithmetic or as an integer. See
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2431.pdf for
details.
tools/gn/tutorial/hello.cc:7: The associated header file of .cpp files
should be included before any other includes.
tools/gn/target_generator.h:16: Warning: Base class TargetGenerator has
virtual functions but a public non-virtual destructor.

The errors were found with the following command line:

$ flint --recusive tools/gn

And cleaned up until silenced with:
$ gn gen out-gn
$ ninja -C out-gn gn gn_unittests

BUG=None
TEST=see above
R=scottmg@chromium.org,brettw@chromium.org

Review URL: https://codereview.chromium.org/964203002

Cr-Commit-Position: refs/heads/master@{#318639}

This re-lands #318523 with a fix for the mac build.

TBR=brettw@chromium.org
BUG=461019
CQ_EXTRA_TRYBOTS=tryserver.chromium.linux:android_chromium_gn_compile_dbg,android_chromium_gn_compile_rel;tryserver.chromium.win:win8_chromium_gn_rel,win8_chromium_gn_dbg;tryserver.chromium.mac:mac_chromium_gn_rel,mac_chromium_gn_dbg

Review URL: https://codereview.chromium.org/967683003

Cr-Commit-Position: refs/heads/master@{#318638}

These resources are used only by web_contents_delegate_android.

BUG=458328

Review URL: https://codereview.chromium.org/950223002

Cr-Commit-Position: refs/heads/master@{#318637}

It looks like these patches might be confusing the
analyze step on the ios bots. I don't understand
the errors I'm seeing on them, but they seem to
have started when the first of these patches landed,
so I'm going to try reverting them and see what happens.

This reverts:

  "Rework how the gn migration targets are included from all.gyp."
  commit 91cb1747 (#318599).

  "Fix a build failing in the x86 linux builds ..."
  commit 425ba073. (#318626)

  "Fix nacl-only targets in gn_migration.gypi"
  commit ad333257. (#318631)

TBR=brettw@chromium.org
BUG=461019
NOTRY=true

Review URL: https://codereview.chromium.org/971633002

Cr-Commit-Position: refs/heads/master@{#318636}

The rebaseline for SK_SUPPORT_LEGACY_ARCTO_QUADS has landed.

BUG=458898
TBR=reed@google.com,robertphillips@google.com
NOTRY=true

Review URL: https://codereview.chromium.org/968713002

Cr-Commit-Position: refs/heads/master@{#318635}

I think the analyze generator is getting confused by
a mismatch between bitmaptools and bitmaptools#host,
but am not sure, so this is a speculative fix.

TBR=brettw@chromium.org

Review URL: https://codereview.chromium.org/970713002

Cr-Commit-Position: refs/heads/master@{#318634}

Now valid tests:
  - WorkerTest except ones using SharedWorkers
  - MediaSourceTest.ConfigChangeVideo
  - CrossPlatformAccessibilityBrowserTest
  - ResourceDispatcherHostBrowserTest.SyncXMLHttpRequest_DuringUnload
  - ResourceDispatcherHostBrowserTest.CrossSiteAfterCrash
  - RenderFrameHostManagerTest.ProcessExitWithSwappedOutViews
  - ScreenOrientationBrowserTest
  - RenderViewHostTest.BaseURLParam
  - FileSystemBrowserTestWithLowQuota.QuotaTest
  - ChildProcessSecurityPolicyInProcessBrowserTest.NoLeak
  - DomSerializerTests
  - BookmarkletTest.NonEmptyResult
  - SignalTest
  - TracingControllerTest.EnableCaptureAndDisableMonitoring
  - TracingControllerTest.EnableCaptureAndDisableMonitoringWithFilePath
  - FrameTreeBrowserTest.NavigateWithLeftoverFrames

And unavailable tests:
  - WebGLConformanceTest
  - MessagePortTest
  - OffTheRecordClipboardTest
  - GpuPixelBrowserTest
  - RendererAccessibilityTest.EditableTextModeFocusEvents
  - RendererAccessibilityTest.EditableTextModeFocusNotifications
  - RenderWidgetTest.OnMsgPaintAtSize
  - PasswordFormConversionUtilsTest.ValidWebFormElementToPasswordForm
  - PasswordFormConversionUtilsTest.InvalidWebFormElementToPasswordForm
  - RenderWidgetHostBrowserTest.GetSnapshotFromRendererTest
  - BrowserGpuChannelHostFactoryTest

BUG=461624, 256238, 338408, 338411, 341995

Review URL: https://codereview.chromium.org/952293004

Cr-Commit-Position: refs/heads/master@{#318633}

This pulls in the following Native Client changes:

e13eda3: (stichnot@chromium.org) Update Subzero revision in pnacl/COMPONENT_REVISIONS.
20aea6a: (jvoung@chromium.org) Remove some unused PNaCl emulation handlers from sel_universal.
73c8589: (dschuff@chromium.org) Roll newlib nacl-gcc
da28a2e: (dschuff@chromium.org) Stop building biased-bitcode libstdc++
ab8a5c6: (jvoung@chromium.org) Update PNaCl to LLVM/Clang/Libcxx 223109 (nacl-change)
f84cb18: (dschuff@chromium.org) Increase coverage of memset in libc's memcpy/memset/memmove test.
45959d1: (jvoung@chromium.org) Update revision for PNaCl
f411324: (phosek@chromium.org) Add a SCons target for generating validator tries

BUG=none
TEST=browser_tests and nacl_integration
CQ_EXTRA_TRYBOTS=tryserver.chromium.linux:linux_chromium_dbg_32_ng,linux_arm_compile,linux_nacl_sdk_build

Review URL: https://codereview.chromium.org/964913004

Cr-Commit-Position: refs/heads/master@{#318632}

The 'nacl_loader_unittests' and 'remoting_key_tester'
targets require nacl and should only be built when
nacl is enabled.

TBR=dongseong.hwang@intel.com, tfarina@chromium.org
BUG=462945, 461019

Review URL: https://codereview.chromium.org/972563002

Cr-Commit-Position: refs/heads/master@{#318631}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/cc9ca14..2d29181

TBR=

Review URL: https://codereview.chromium.org/970703002

Cr-Commit-Position: refs/heads/master@{#318630}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/394ce76..73165bd

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

Review URL: https://codereview.chromium.org/968703002

Cr-Commit-Position: refs/heads/master@{#318629}