GitLab

Now that [1] has been fixed and libgtk-3-0 has been installed in
clusterfuzz's 32-bit configuration, it's safe to enable gtk3 for all
platforms.

[1] https://crbug.com/699669

BUG=79722
TBR=dpranke@chromium.org
CC=erg@chromium.org,inferno@chromium.org

Review-Url: https://codereview.chromium.org/2847863002
Cr-Commit-Position: refs/heads/master@{#468110}

NOPRESUBMIT=true
BUG=713190

Review-Url: https://codereview.chromium.org/2844593002
Cr-Commit-Position: refs/heads/master@{#468109}

This ensures proper capability check.

BUG=none

Review-Url: https://codereview.chromium.org/2845813003
Cr-Commit-Position: refs/heads/master@{#468108}

Specifically, add menu option displays and track site breakdown displays,
sorting, and expansion.

BUG=716111

Review-Url: https://codereview.chromium.org/2846953002
Cr-Commit-Position: refs/heads/master@{#468107}

Using a local proxy instance with the associated bug, the failed
response gave a 404 status. This check differentiates between the bug
and not.

BUG=715644

Review-Url: https://codereview.chromium.org/2844403002
Cr-Commit-Position: refs/heads/master@{#468106}

As suggested in a previous CL, the status for each keyId should have it's own
status value rather than relying on all the KeyIds to have the same status.

No longer passing |unexpected|, as the size check verifies that the correct
number of expected KeyIds matches.

BUG=616166
TEST=modified tests pass

Review-Url: https://codereview.chromium.org/2847803005
Cr-Commit-Position: refs/heads/master@{#468105}

This in preparation for USER_EVENTS, which is a commit only type and
some fields like version and not useful. We want to save as much
bandwidth as possible, and switching to optional saves us ~2 bytes each
before compression.

BUG=714828

Review-Url: https://codereview.chromium.org/2837993002
Cr-Commit-Position: refs/heads/master@{#468104}

BUG=714191

Review-Url: https://codereview.chromium.org/2829293002
Cr-Commit-Position: refs/heads/master@{#468103}

Directory::InsertEntry takes pointer to EntryKernel and inserts owning pointer
into methandles_mapo with WrapUnique. The object is still owned by unique_ptr in
ModelNeutralMutableEntry ctor. If one if the steps inside InsertEntry fails
ModelNeutralMutableEntry will not release unique_ptr which will cause object to
be freed while metahandles map still has entry pointing to it.

I changed InsertEntry to pass ownint pointer to EntryKernel. Caller is free to
stash non-owning pointer, but has to reset it to nullptr if InsertEntry fails.

I refactored couple of functions in Directory to be more strict with pointers to
EntryKernel. Particularly DeleteEntry should use entry found in metahandles_map,
not the entry passed as an argument to remove entry from different indices. It
shouldn't matter in terms of correctness, but makes it easier to reason about
the logic.

BUG=705704
R=skym@chromium.org

Review-Url: https://codereview.chromium.org/2844333003
Cr-Commit-Position: refs/heads/master@{#468102}

WindowManagerApplication's normal destruction order involves tearing down
everything is has created.

However on some fast tests it is possible for WindowManagerApplication to be
destroyed before OnStart is ever called. When this occurs we attempt to destroy
services which we never created, and various DCHECKS are triggered.

PhishingDOMFeatureExtractorTest.SubframeRemoval has a >50% repro rate of this
locally where the crash occurs after the test successfully passes.

TEST=PhishingDOMFeatureExtractorTest.SubframeRemoval
BUG=678687

Review-Url: https://codereview.chromium.org/2851763002
Cr-Commit-Position: refs/heads/master@{#468101}

BUG=None

Review-Url: https://codereview.chromium.org/2849903003
Cr-Commit-Position: refs/heads/master@{#468100}

When in tab fullscreen, replace the ESC button with "Exit Fullscreen"
and display the origin URL of the tab content in the Touch Bar.

A test for this is added in BrowserTouchBarUnitTest

BUG=690611

Review-Url: https://codereview.chromium.org/2816743006
Cr-Commit-Position: refs/heads/master@{#468099}

I think original issue was caused by ChromeApplication.onCreate calling
AccountManagerHelper.initializeAccountManagerHelper and discovering instance of
AccountManagerHelper that was set up by test. This issue is addressed by
http://crrev.com/2836373003.

Another issue was that waiting for syncability propagation to ContentResolver
got broken by refactoring http://crrev.com/2778443002. This issue was addressed
by change http://crrev.com/2847663003.

Seems like after these changes test should be passing. I ran it in a loop for
1000 iterations and it passed so I'm enabling the test.

BUG=605567
R=pnoland@chromium.org

Review-Url: https://codereview.chromium.org/2845173003
Cr-Commit-Position: refs/heads/master@{#468098}

Tests re-enabled are:
UseCounterTest.RecordingFeatures, UseCounterTest.SVGImageContextFeatures,
UseCounterTest.InspectorDisablesMeasurement, UseCounterTest.MutedDocuments

BUG=667913,654502

Review-Url: https://codereview.chromium.org/2847173002
Cr-Commit-Position: refs/heads/master@{#468097}

The security state decoration looks like this:

  Secure |

The divider is included in the bounds, and then the location bar text is placed
just after this decoration; because of that, the focus ring on the security
state decoration (which is exterior to the decoration, like Cocoa focus rings
usually are) draws over the top of the start of the location bar text. This
change:

1) Introduces LocationBarDecoration::GetRealFocusRingBounds() to allow each
   decoration to control its focus ring bounds if it needs to
2) Overrides that method in SecurityStateBubbleDecoration to inset the focus
   ring on the right side

BUG=715909

Review-Url: https://codereview.chromium.org/2847903003
Cr-Commit-Position: refs/heads/master@{#468096}

Webview does not have new tab page.

TBR=nednguyen@google.com,ssid@chromium.org
BUG=716041

Review-Url: https://codereview.chromium.org/2848893002
Cr-Commit-Position: refs/heads/master@{#468095}

With mask tiling, it is possible that a tiled mask gets a filter after
tiles are created. In order to make picture layer tilings throw away
tiles, we need to change the mask type from MULTI_TEXTURE_MASK to
SINGLE_TEXTURE_MASK, but mask type was set to be immutable as we do not
allow NOT_MASK to MASK conversion and vice versa.

This CL attempts to fix the demoting problem by making mask type a non-
const value and prevent NOT_MASK to MASK conversion by DCHECK.

BUG=567293
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2845143002
Cr-Commit-Position: refs/heads/master@{#468094}

BUG=707642
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2845773005
Cr-Commit-Position: refs/heads/master@{#468093}

In the (rare) case of a recording being started and the FilterData
structure being yanked away (and destroyed) from under
SVGFilterPainter's feet, we need to put the PaintController in a
consistent state before destroying it.
Add a new method SVGFilterRecordingContext::Abort() and call that when
SVGFilterPainter::FinishEffect encounters a null FilterData for the
LayoutObject.

BUG=713024
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2847133002
Cr-Commit-Position: refs/heads/master@{#468092}

BUG=none

Review-Url: https://codereview.chromium.org/2840413002
Cr-Commit-Position: refs/heads/master@{#468091}

This applies to Linux and Chrome OS (and Mac Views) until
713359 is fixed.

BUG=602756,713359

Review-Url: https://codereview.chromium.org/2839953005
Cr-Commit-Position: refs/heads/master@{#468090}

BUG=682420

Review-Url: https://codereview.chromium.org/2852543002
Cr-Commit-Position: refs/heads/master@{#468089}

std::less requires #include <functional>,
see e.g. http://en.cppreference.com/w/cpp/utility/functional/less

See the bug for more context.

BUG=716504

Review-Url: https://codereview.chromium.org/2849913002
Cr-Commit-Position: refs/heads/master@{#468088}

This CL
- Enables inline module scripts in ScriptLoader.
- Adds ModuleTreeLinker::FetchDescendantsForInlineScript() and
  related plumbing to ModuleTreeLinkerRegistry and Modulator.
  This is not yet spec'ed and we might need revisit once the spec is
  updated. https://github.com/whatwg/html/issues/2544
- Makes ModuleScript to store its |source_text| for CSP of inline scripts.

BUG=715369, 594639

Review-Url: https://codereview.chromium.org/2842923002
Cr-Commit-Position: refs/heads/master@{#468087}

Adds a 3 min delay before starting zero-touch enrollment during OOBE.

BUG=chromium:710716
TEST=browser test

Review-Url: https://codereview.chromium.org/2824283003
Cr-Commit-Position: refs/heads/master@{#468086}

Make FilterData::state_ only be checked and updated within
SVGFilterPainter, and not by SVGFilterRecordingContext or the local
painting helper.
Instead SVGFilterRecordingContext only manages the recording state, and
gets passed bounds while returning a paint record.
This simplifies some of the corner-cases with regards to how "aborted"
filters are handled, getting rid of some "FilterData is null" in many
cases.

BUG=713024
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2851753002
Cr-Commit-Position: refs/heads/master@{#468085}

Reland of SharedArrayBuffer throws serializing to indexeddb (patchset #1 id:1 of https://codereview.chromium.org/2845623002/ )

Reason for revert:
This CL was speculatively suspected as the cause of http://crbug.com/715567, but it turned out was unrelated.

Original issue's description:
> Revert of SharedArrayBuffer throws serializing to indexeddb (patchset #4 id:60001 of https://codereview.chromium.org/2838783002/ )
>
> Reason for revert:
> Suspect for http://crbug.com/715567 (crash when typing in hangouts chat window).
>
> Original issue's description:
> > SharedArrayBuffer throws serializing to indexeddb
> >
> > The StructuredDeserialize spec now has a special case for serializing
> > for storage: it doesn't allow SharedArrayBuffers to be serialized, and
> > instead throws a DataCloneError.
> >
> > BUG=chromium:714820
> >
> > Review-Url: https://codereview.chromium.org/2838783002
> > Cr-Commit-Position: refs/heads/master@{#467093}
> > Committed: https://chromium.googlesource.com/chromium/src/+/08d35f601c7cfbdc66f9a7f8836e32028d344559
>
> TBR=jbroman@chromium.org,haraken@chromium.org,jsbell@chromium.org,binji@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=chromium:714820
>
> Review-Url: https://codereview.chromium.org/2845623002
> Cr-Commit-Position: refs/heads/master@{#467325}
> Committed: https://chromium.googlesource.com/chromium/src/+/d86073fac8a47abe754598992e1b8424d7534b9f

TBR=jbroman@chromium.org,haraken@chromium.org,jsbell@chromium.org,gab@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=chromium:714820

Review-Url: https://codereview.chromium.org/2848933002
Cr-Commit-Position: refs/heads/master@{#468084}

Monthly update to the newest production version.

BUG=

Review-Url: https://codereview.chromium.org/2844243004
Cr-Commit-Position: refs/heads/master@{#468083}

https://skia.googlesource.com/skia.git/+log/0122af08f6af..2fcd4a480dee

$ git log 0122af08f..2fcd4a480 --date=short --no-merges --format='%ad %ae %s'
2017-04-28 benjaminwagner Revert "Only store width and height on SkPixelRef"
2017-04-28 msarett Revert "Revert "Delete copyTo(Allocator), hide copyTo() behind flag""
2017-04-28 reed add helper makeShader for the common Clamp case
2017-04-28 jvanverth Fix fast blur for shadow rrects.
2017-04-28 fmalita Scrub DeviceCM
2017-04-28 rmistry Testing permissions
2017-04-27 benjaminwagner Add non-Valgrind PreAbandonGpuContext bot.
2017-04-28 bsalomon Remove srgb flags from GrProcessorSet
2017-04-27 ethannicholas additional skslc vector optimizations
2017-04-28 reed add GM to compare atlas and vertices
2017-04-26 freiling handle GrVkTransferBuffer::Create failure
2017-04-27 benjaminwagner [infra/bots] Treat extra_config as a list.
2017-04-27 robertphillips Disable deferred proxies in Chrome
2017-04-27 msarett Only store width and height on SkPixelRef

Created with:
  roll-dep src/third_party/skia


Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls


CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_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;master.tryserver.chromium.android:android_optional_gpu_tests_rel
TBR=benjaminwagner@chromium.org

Change-Id: I288f95ee249db9b974f75f41c78b062996ce97f4
Reviewed-on: https://chromium-review.googlesource.com/490686Reviewed-by: Skia Deps Roller <skia-deps-roller@chromium.org>
Commit-Queue: Skia Deps Roller <skia-deps-roller@chromium.org>
Cr-Commit-Position: refs/heads/master@{#468082}

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2821323002
Cr-Commit-Position: refs/heads/master@{#468081}

This patch makes |InsertTextCommand::DoApply()| not to call |ApplyStyle()| for
applying typing style when selection after inserting text is empty since
|ApplyStyle()| doesn't work with empty selection.

The issue 714311 and the attached test case insert text into OPTION element to
get empty selection after insertion, since we can't place selection inside
OPTION element.

BUG=714311
TEST=run_webkit_unit_tests --gtest_filter=InsertTextCommandTest.WithTypingStyle

Review-Url: https://codereview.chromium.org/2847763004
Cr-Commit-Position: refs/heads/master@{#468080}

Since GestureScrolls can be ended by either GestureScrollEnd or
GestureFlingStart, it seems wrong that OOPIF scroll bubbling, in
RenderWidgetHostViewChildFrame/CrossProcessFrameConnector doesn't
bubble GestureFlingStart.

This CL modifies the bubbling code to include bubbling of
GestureFlingStart, and re-enables a related DHECK in
RenderWidgetHostImpl.

BUG=695187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2842203002
Cr-Commit-Position: refs/heads/master@{#468079}

Kernel 4.11 merges the commit that added linux/dma-buf.h
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=ef96152e6a36e0510387cb174178b7982c1ae879

This CL increases the required kernel to include this file to 4.11.

BUG=707604
R=danakj@chromium.org

Review-Url: https://codereview.chromium.org/2851803002
Cr-Commit-Position: refs/heads/master@{#468078}

This patch add DCHECK into |ApplyStyleCommand::DoApply()| to verify that we
don't apply |ApplyStyleCommand| to empty range to detect redundant command
execution for improving code health.

Actually, |ApplyStyleCommand::DoApply()| doesn't work with empty range, it
causes DCHECK for |Position.IsNull()| in |ComparePositions()|.

BUG=714311
TEST=n/a; no behavior changes

Review-Url: https://codereview.chromium.org/2848823002
Cr-Commit-Position: refs/heads/master@{#468077}

exo::Surface wasn't setting the device_scale_factor in CompositorFrames.
This was triggering a DCHECK when the SurfaceInfo for new cc::Surfaces
have device_scale_factor == 0.

BUG=b/37654593

Review-Url: https://codereview.chromium.org/2843723003
Cr-Commit-Position: refs/heads/master@{#468076}

Reland of [WebAPKs] Make WebAPK update flow more sensible (patchset #2 id:90001 of https://codereview.chromium.org/2845943006/ )

Reason for revert:
Reverting the revert now that https://chrome-internal-review.googlesource.com/c/363648/ has landed. That CL should prevent official builder from breaking

Original issue's description:
> Revert of [WebAPKs] Make WebAPK update flow more sensible (patchset #4 id:100001 of https://codereview.chromium.org/2827233006/ )
>
> Reason for revert:
> broke official-* bots, see crbug.com/716330
>
> Original issue's description:
> > [WebAPKs] Make WebAPK update flow more sensible
> >
> > Previously, whenever the shell_apk in shell_apk_version.gni there was a period
> > of time during which
> > - Chrome tries to update the WebAPK because the "shell_apk" version with which
> >   Chrome is compiled with was updated
> > - The WebAPK server is unable to fulfill the update request because the new
> >   ShellAPK has not yet been uploaded to the WebAPK server
> >
> > This CL splits shell_apk_version in shell_apk_version.gni into two variables:
> > |template_shell_apk_version| and |expected_shell_apk_version|
> >
> > The new flow for updating android/webapk/shell_apk will be:
> > 1) Write CL which:
> >    - Makes changes to code in android/webapk/shell_apk
> >    - Increments |template_shell_apk_version|
> >    This will:
> >    - Update the template ShellAPK which is built on the official build bot
> >    - Trigger an alert that the ShellAPK on the WebAPK Minting Server needs to
> >      be updated
> >    - Does not affect the "WebAPK update logic".
> > 2) Upload new ShellAPK template to the WebAPK Minting Server
> >    At this point:
> >    - new WebAPK installs will have the updated shell.
> >    - Chrome will still not request updates to the new ShellAPK version
> > 3) Write CL which:
> >    - Increments |expected_shell_apk_version|
> >    This will:
> >    - cause Chrome to request updates to the new ShellAPK version
> >
> > BUG=708053
> >
> > Review-Url: https://codereview.chromium.org/2827233006
> > Cr-Commit-Position: refs/heads/master@{#467708}
> > Committed: https://chromium.googlesource.com/chromium/src/+/d71abed75b07b4b9fc50591cbeb2b6c8874bf448
>
> TBR=hartmanng@chromium.org,hanxi@chromium.org,zpeng@chromium.org,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=708053
>
> Review-Url: https://codereview.chromium.org/2845943006
> Cr-Commit-Position: refs/heads/master@{#467966}
> Committed: https://chromium.googlesource.com/chromium/src/+/db08d35eab5d639a6ed751dd4e42dcd5b4f683ec

TBR=hartmanng@chromium.org,hanxi@chromium.org,zpeng@chromium.org,dominickn@chromium.org,timvolodine@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=708053

Review-Url: https://codereview.chromium.org/2847263002
Cr-Commit-Position: refs/heads/master@{#468075}

The current extension permission update bubble puts both old and new
permissions together in one list. This CL changes the bubble to only show the
new permissions.

BUG=443216

Review-Url: https://codereview.chromium.org/2821123002
Cr-Commit-Position: refs/heads/master@{#468074}

The "learn more" link has moved to the main content instead of the
details. Tests should check for it before clicking the details, since
mobile layouts will hide the main content after "details" is clicked.

BUG=674732

Review-Url: https://codereview.chromium.org/2845363003
Cr-Commit-Position: refs/heads/master@{#468073}

The lifetime of GuestViewBase doesn't match the lifetime of GuestViewMessageFilter, so it's possible that the GuestViewBase is already destroyed by the time this callback is called.

BUG=716223

Review-Url: https://codereview.chromium.org/2848723002
Cr-Commit-Position: refs/heads/master@{#468072}

BUG=

Review-Url: https://codereview.chromium.org/2848003002
Cr-Commit-Position: refs/heads/master@{#468071}