GitLab

If a collection backing is reallocated during pre finalization, the backing containing stale pointers ends up uncollected for one cycle. If conservative GC ends up tracing this buffer by accident, the stale pointers gets traced.

This CL disallows collection backing reallocation during pre finalizers. The backing size will end up being adjusted later.

BUG=709201

Review-Url: https://codereview.chromium.org/2815663002
Cr-Commit-Position: refs/heads/master@{#464359}

 NodeAsRangePastLastNode() as Range::PastLastNode().

This was introduced once 2 years ago:
https://codereview.chromium.org/1000533004

but reverted of crash:
https://codereview.chromium.org/1121633004

Now original code won't crash because {start/end}Position() doesn't return
 Null Position since:
https://codereview.chromium.org/2812733003

TEST=LayoutTests/fast/dom/Range/range-toString-non-anchor-no-crash.html

Review-Url: https://codereview.chromium.org/2818643003
Cr-Commit-Position: refs/heads/master@{#464358}

f425c1c cryptohome: Add a new mount argument for migration

BUG=702526
TEST=build

Review-Url: https://codereview.chromium.org/2815163004
Cr-Commit-Position: refs/heads/master@{#464357}

This line was introduced before 2011 splitting DragCaretController from
 FrameSelection:
https://chromium.googlesource.com/chromium/src/+/fa14cd2b25316fb358f7f5ebc16f752ebcb4093a

Now DragCaret doesn't touch Selection Layout and this is unnecessary call.

BUG=708453

Review-Url: https://codereview.chromium.org/2813883005
Cr-Commit-Position: refs/heads/master@{#464356}

This is essentially the result of the following shell command followed by
git cl format:

    find . -type f -exec sed -i 's/"wtf\//"platform\/wtf\//g' {} \;

DEPS is added in order to prevent further introduction of "wtf/".

BUG=691465
R=haraken@chromium.org

Review-Url: https://codereview.chromium.org/2815193003
Cr-Commit-Position: refs/heads/master@{#464355}

This is essentially the result of the following shell command followed by
git cl format:

    find . -type f -exec sed -i 's/"wtf\//"platform\/wtf\//g' {} \;

DEPS is added in order to prevent further introduction of "wtf/".

BUG=691465
R=yosin@chromium.org

Review-Url: https://codereview.chromium.org/2817863002
Cr-Commit-Position: refs/heads/master@{#464354}

This is essentially the result of the following shell command followed by
git cl format:

    find . -type f -exec sed -i 's/"wtf\//"platform\/wtf\//g' {} \;

DEPS restriction is lifted to platform/ level. Duplicated entries in
the subdirectories are removed.

BUG=691465
R=haraken@chromium.org

Review-Url: https://codereview.chromium.org/2810413002
Cr-Commit-Position: refs/heads/master@{#464353}

LayoutView::ClearSelection was originally introduced at 2004 to assure no
 crash:
https://chromium.googlesource.com/chromium/src/+/10f7ac6ea6784e33161c7979e9a59c5e2cae14b5

Even now that code doesn't make sense because we update LayoutSelection after
layout in following sequence:
1. FrameView::PerformPostLayoutTasks() checks
 LayoutSelection::SetHasPendingSelection()
2. PaintLayerCompositor::UpdateIfNeededRecursiveInternal() calls
 LayoutSelection::Commit() and it updates layout selection.

BUG=708453

Review-Url: https://codereview.chromium.org/2811333003
Cr-Commit-Position: refs/heads/master@{#464352}

This allows observers that might outlive the service to unregister
themselves, and to clear out any raw pointers to the service they might
have.
In the same spirit, verify that there are no observers left during
TemplateURLService destruction.

BUG=704518

Review-Url: https://codereview.chromium.org/2815643007
Cr-Commit-Position: refs/heads/master@{#464351}

BUG=709964

Review-Url: https://codereview.chromium.org/2806273002
Cr-Commit-Position: refs/heads/master@{#464350}

Add the libjingle_xmpp_unittests test to the following bots:
Cast Audio Linux
Cast Linux
Android Tests
Android Tests (dbg)
Linux Tests
Linux Tests (dbg)(1)
Linux Tests (dbg)(1)(32)
Mac10.10 Tests
Mac10.11 Tests
Mac10.9 Tests
Mac10.9 Tests (dbg)
Win 7 Tests x64 (1)
Win10 Tests x64
Win7 Tests (1)
Win7 Tests (dbg)(1)

BUG=708983
TESTED=Local execution and a run in Swarming using:
tools/mb/mb.py isolate //out/Release libjingle_xmpp_unittests
tools/swarming_client/isolate.py archive -I https://isolateserver.appspot.com   -i out/Release/libjingle_xmpp_unittests.isolate -s out/Release/libjingle_xmpp_unittests.isolated
tools/swarming_client/swarming.py trigger -S https://chromium-swarm.appspot.com -I https://isolateserver.appspot.com -d pool Chrome -d os Linux -d cpu x86-64 out/Release/libjingle_xmpp_unittests.isolated

Review-Url: https://codereview.chromium.org/2803333002
Cr-Commit-Position: refs/heads/master@{#464349}

Revert of Don't set a default value for the NTPContentSuggestionsEnabled policy if ChromeHome is enabled (patchset #3 id:40001 of https://codereview.chromium.org/2806453002/ )

Reason for revert:
The hack isn't required anymore, let's get it removed before branch point :)

Original issue's description:
> Don't set a default value for the NTPContentSuggestionsEnabled policy if ChromeHome is enabled
>
> This is a temporary hack to allow dogfooding of ChromeHome.
>
> BUG=708191
>
> Review-Url: https://codereview.chromium.org/2806453002
> Cr-Commit-Position: refs/heads/master@{#462451}
> Committed: https://chromium.googlesource.com/chromium/src/+/d4b92153f875c211e8c6f01603ed510962200747

TBR=atwilson@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=708191

Review-Url: https://codereview.chromium.org/2813423002
Cr-Commit-Position: refs/heads/master@{#464348}

As per spec Pageshow event must be fired asynchonously all the time, but to be compatible with other browsers blink used to fire Pageshow synchronously.
And thus Pageshow event was able to modify nodes when it should not.

This CL makes Pageshow fired asynchronously when it is in EventQueueScope to solve the problem.
BUG=660269,543424

Review-Url: https://codereview.chromium.org/2815803003
Cr-Commit-Position: refs/heads/master@{#464347}

This is just moving:
https://codereview.chromium.org/2800813006/

TEST=No change in behavior.
BUG=708453

Review-Url: https://codereview.chromium.org/2813403002
Cr-Commit-Position: refs/heads/master@{#464346}

core/html/forms/OWNERS should be responsible for these files.
This CL has no behavior changes.

Review-Url: https://codereview.chromium.org/2815263002
Cr-Commit-Position: refs/heads/master@{#464345}

{EventHandler,MouseEventManager} is storing a {clicked,tapped} node as Node*,
instead of Element*, that would be a bad practice. Actually, that has been the
cause of several wrong behaviors in Blink, such as:

1. Blink does NOT fire a click event when a clicked text node has been removed
   in mouseup.  This is wrong because a click event should be fired on an
   element node.  See [1] for details. This bug was just tentatively fixed at
   another CL [2].

2. Blink DOES fire a touch event on a non-element node.  This is wrong because
   TouchEvent specification says all kinds of touch events' event target types
  are limited to Document and Element [3].

To prevent these wrong behaviors, this CL does:
- Have EventHandlingUtil::ParentElementIfNeeded and use it everywhere so that
  we surely adjust a node to the appropriate element, at which an event should be
  fired.
- Replaces the usage of Node* to Element* as much as possible so that it rejects
  the wrong code at type level check.

- [1]: topmost event target;
  https://www.w3.org/TR/uievents/#topmost-event-target
- [2] https://codereview.chromium.org/2812613004
- [3]: Trusted proximal event target types are: Document and Element;
  https://w3c.github.io/touch-events/#list-of-touchevent-types

BUG=708394,710425

Review-Url: https://codereview.chromium.org/2807123002
Cr-Commit-Position: refs/heads/master@{#464344}

This is essentially the result of the following shell command followed by
git cl format:

    find . -type f -exec sed -i 's/"wtf\//"platform\/wtf\//g' {} \;

DEPS is added in order to prevent further introduction of "wtf/".

BUG=691465
R=japhet@chromium.org

Review-Url: https://codereview.chromium.org/2818613003
Cr-Commit-Position: refs/heads/master@{#464343}

And rename files
gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.{cc,h} =>
gpu/ipc/client/gpu_memory_buffer_impl_native_pixmap.{cc,h}

Turn on GpuMemoryBufferImplNativePixmap on OS_LINUX instead of only USE_ZONE.

This a pre-step to enable glCreateImageCHROMIUM with linux dma buf on linux.

BUG=584248
R=reveman@chromium.org
TBR=kbr@chromium.org, piman@chromium.org
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/2710223002
Cr-Commit-Position: refs/heads/master@{#464342}

The clang tool had a bug in how it processes variables named
fooBARsBaz into foo_ba_rs_baz rather than foo_bars_baz.  Examples
include URLs -> ur_ls and XHRs -> xh_rs.

No intended behavior change.

BUG=675877
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/2814853002
Cr-Commit-Position: refs/heads/master@{#464341}

This CL adds and changes some files in ui/file_manager to build ZIP archiver.
We are going to add ZIP archiver to the list of component extensions
in another CL, so ZIP archiver still won't show up in the context
menu with this patch.

BUG=607078
TEST=manually tested
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2808993003
Cr-Commit-Position: refs/heads/master@{#464309}
Committed: https://chromium.googlesource.com/chromium/src/+/77065d9816b34a5cc7913894b8b756e7c3e6fb79

patch from issue 2808993003 at patchset 80001 (http://crrev.com/2808993003#ps80001)

Review-Url: https://codereview.chromium.org/2814313003
Cr-Commit-Position: refs/heads/master@{#464340}

Revert of Expand Mash Browser Test Coverage (patchset #11 id:220001 of https://codereview.chromium.org/2786583003/ )

Reason for revert:
Some tests are very flaky on ozone try bot.

https://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_chromeos_ozone_rel_ng?numbuilds=200

Original issue's description:
> Based on running of the tests locally, along with on the trybots, the filter for mash_browser_tests have been updated with more tests
>
> TEST=mash_browser_tests
> BUG=678687
>
> Review-Url: https://codereview.chromium.org/2786583003
> Cr-Commit-Position: refs/heads/master@{#464015}
> Committed: https://chromium.googlesource.com/chromium/src/+/0190af327299091a9bdd376da2093cd48fe1f993

TBR=sky@chromium.org,jonross@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=678687

Review-Url: https://codereview.chromium.org/2820483002
Cr-Commit-Position: refs/heads/master@{#464339}

This CL removes a cross thread malloc/free pair from simple disk cache backend.

After a PostTaskAndReply change in http://crrev.com/e5de8d551e80115b, it
destroys |task| part of the given tasks on the target thread, and that
introduced a number of cross thread malloc/free pairs around net::IOBuffer.
The cross thread malloc/free pair increased the apparent size of memory
usage on some Android perf bots by 2~3%, that is likely due to thread
specific free list caches.

BUG=708644

Review-Url: https://codereview.chromium.org/2815563002
Cr-Commit-Position: refs/heads/master@{#464338}

Revert of Move logic of recording main thread scrolling reasons from cc to blink::ScrollManager (patchset #17 id:320001 of https://codereview.chromium.org/2773593005/ )

Reason for revert:
Likely cause of webkit_unit_tests failing on chromium.webkit/WebKit Android (Nexus4).

BUG=711190

Original issue's description:
> Move logic of recording main thread scrolling reasons from cc to
> blink::ScrollManager
>
> Due to crbug.com/701355, all style related main thread scrolling reasons
> stop being recorded. This patch moves the logic of storing the style
> related reasons to PLSA and move the recording logic to ScrollManager
> because the information on the compositor side is insufficient. The
> disabled test due to the previous bug has been enabled again with minor
> modification.
>
> BUG=704805
> TEST=All/NonCompositedMainThreadScrollingReasonTest.*;
> EventHandlerTest.NonCompositedMainThreadScrollingReason*
>
> CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
> master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
> master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
>
> Review-Url: https://codereview.chromium.org/2773593005
> Cr-Commit-Position: refs/heads/master@{#464073}
> Committed: https://chromium.googlesource.com/chromium/src/+/76d0929033d6cd3ec36b96d1b4845166e70c9349

TBR=flackr@chromium.org,ajuma@chromium.org,bokan@chromium.org,tdresser@chromium.org,jwd@chromium.org,yigu@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=704805

Review-Url: https://codereview.chromium.org/2816973002
Cr-Commit-Position: refs/heads/master@{#464337}

To remove circular deps between file_table and file_table_list,
- Moved filelist namespace from file_table to file_table_list.
- Use observer pattern instead of calling
|updateHighPriorityRange| in FileTable from FileTableList.

BUG=636289
TEST=run_compiler, try
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2807033002
Cr-Commit-Position: refs/heads/master@{#464336}

If used in combination with a transparent setDefaultBackgroundColor,
the background color override applied by setDeviceMetricsOverride is
harmful. The background override does not seem to be beneficial
anymore, as the frontend already doesn't use it.

BUG=689349

Review-Url: https://codereview.chromium.org/2808123004
Cr-Commit-Position: refs/heads/master@{#464335}

BUG=

Review-Url: https://codereview.chromium.org/2811093002
Cr-Commit-Position: refs/heads/master@{#464334}

This CL is a reraise of http://crrev.com/2785943004 PS4.

A null callback posted to PostTask family eventually causes a null pointer
access crash. And its crash report doesn't indicate the culprit directly
without looking into the local variable of the minidump.

After this CL, the crash happens earlier and the crash report will
have meaningful trace.

BUG=711167, 709149

Review-Url: https://codereview.chromium.org/2815573002
Cr-Commit-Position: refs/heads/master@{#464333}

- Add CryptohomeClient::NeedsDircryptoMigration wrapper;
- Check whether a user needs dircryto migration when the relevant
  user pod get focus;
- Show a banner message when the user needs dircryto migration;

BUG=708015

Review-Url: https://codereview.chromium.org/2808423002
Cr-Commit-Position: refs/heads/master@{#464332}

Added null check for the case when arc_custom_notification_item is
already destroyed.

BUG=711149

Review-Url: https://codereview.chromium.org/2812363003
Cr-Commit-Position: refs/heads/master@{#464331}

This adds a new QuickOpenProvider '?' that can be be used to jump to
another provider or search for help.

BUG=662081

Review-Url: https://codereview.chromium.org/2814263002
Cr-Commit-Position: refs/heads/master@{#464330}

According to the spec, even if location header is not set, we should
treat the redirect response as "opaqueredirect" if the redirect mode
of the fetch request is "manual".

This behavior was changed by this commit on the spec.
https://github.com/whatwg/fetch/commit/3e501f29eceff41eb81c60fb9937e33e23cf5492

BUG=707185

Review-Url: https://codereview.chromium.org/2785123002
Cr-Commit-Position: refs/heads/master@{#464329}

BUG=537099

Review-Url: https://codereview.chromium.org/2806013002
Cr-Commit-Position: refs/heads/master@{#464328}

This lets a C++ embedder send messages to and from JS on a HeadlessWebContents.
Note for this to work the headless browser context must have been created with
AddTabSocketMojoBindings and the headless web contents must have been created
with CreateTabSocket(true).

This will not affect chrome.exe --headless because only a C++ embedder can set
the above options.

Note that C++ embedders can already do this and more with the current headless API.
At a later date we will remove the more general headless mojo interface in favor of
TabSockets, unless it turns out somebody is actually using them.

BUG=546953

Review-Url: https://codereview.chromium.org/2813953002
Cr-Commit-Position: refs/heads/master@{#464327}

This CL does two things:
1) Replace old pictures with new ones for powerwash page.
2) Rebuild everything using Polymer for material design.

BUG=653269
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2756863002
Cr-Commit-Position: refs/heads/master@{#464326}

GetChildFrameView returns FrameView associated with local Node
if node is HTMLFrameOwnerElement or nullptr.

The method name GetFrameView is already declared in
LayoutObject which returns Document().View(), so a different
name was required.

BUG=637460
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2810873007
Cr-Commit-Position: refs/heads/master@{#464325}

This CL just remove a line which depends on libarchive.

BUG=607078
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2815083003
Cr-Commit-Position: refs/heads/master@{#464324}

This CL splits UpdateData() function into UpdateResourceData() and
UpdateRedirectData() with less number of arguments.
ResourcePrefetchPredictor uses UpdateData() to update only one of the tables
at the same time so there is no need to pass extra arguments each time.

BUG=631966

Review-Url: https://codereview.chromium.org/2804283002
Cr-Commit-Position: refs/heads/master@{#464323}

Queries to this API have a sleep of 0--8ms, randomly chosen based on a
SHA-256 hash of the app ID, salted with a unique stable device ID. This
prevents websites from being able to determine whether a non-related app
is installed based on the time the query takes.

The website is unable to compare the timing of different apps because
they have different random delays, and unable to predict the random
delay for a given device, because the salt is unknown.

BUG=707071

Review-Url: https://codereview.chromium.org/2802603002
Cr-Commit-Position: refs/heads/master@{#464322}

CL https://crrev.com/2809843002 took  a stab at it but clashed with
the definition in [1]:
   using Position = PositionTemplate<EditingStrategy>;
which is, sadly, used in a bazillion places (see e.g. [2]).

This CL takes a least bad solution by renaming the idl-exposed
Geoposition to Position, but keeping the underlying c++ classes
unchanged.

BUG=708206

[1] https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/editing/Position.h?dr&sq=package:chromium&l=224
[2] https://cs.chromium.org/search/?q=%22+Position+%22+file:%5Esrc/third_party/WebKit/Source/core/+package:%5Ewebrtc$&type=cs

Review-Url: https://codereview.chromium.org/2816713004
Cr-Commit-Position: refs/heads/master@{#464321}

The MD Settings proxy implementation did not support
proxy configurations that did not include a HTTP proxy.
This is a bug, we should support, e.g., just setting a
FTP proxy.

BUG=708561
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2815923002
Cr-Commit-Position: refs/heads/master@{#464320}