GitLab

metrics_web_contents_observer.h/.cc have become quite complex.
This change factors the PageLoadTracker code into its own files,
in order to make each individual file more manageable.

Review-Url: https://codereview.chromium.org/2435233002
Cr-Commit-Position: refs/heads/master@{#427216}

This patch converts "editing/selection/5497643.html" to use |assert_selection()|
to simplify test case and renames to "after-removing-textarea.html" to provide
a hint what this test verifies for improving code health.

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

Review-Url: https://codereview.chromium.org/2441363002
Cr-Commit-Position: refs/heads/master@{#427215}

Changes include:
(1) Suspect content::RunAllPendingInMessageLoop is the reason for flaky, since BrowserWindow::IsActive is asynchronous on Mac and Linux.

(2) Construct a non-empty url as there is a DCHECK here: https://cs.chromium.org/chromium/src/chrome/browser/ui/browser_navigator.cc?type=cs&sq=package:chromium&rcl=1476314269&l=530

(3) function->set_extension as there is a DCHECK here:
https://cs.chromium.org/chromium/src/chrome/browser/extensions/browser_extension_window_controller.cc?type=cs&q=browser_extension_window_controller.cc&sq=package:chromium&l=87

BUG=136562
TEST=change is on test itself
If flaky is still observed, please disable the test again.

Review-Url: https://codereview.chromium.org/2419763002
Cr-Commit-Position: refs/heads/master@{#427214}

Now it returns PaintInvalidationBoundsChange instead of
PaintInvalidationIncremental when bounds change for non-LayoutBox.
Then we don't need to override PaintInvalidationIncremental with
PaintInvalidationFull in ObjectPaintInvalidatorWithContext::
invalidatePaintIfNeeded().

About layout test changes:
- Some test no longer produce PaintInvalidationSelection invalidations
  when the object is doing full paint invalidation;
- Some "full" invalidations are now "bounds change" invalidations.
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2442383002
Cr-Commit-Position: refs/heads/master@{#427213}

Observer lists now support range-based for loops.

BUG=655021

Review-Url: https://codereview.chromium.org/2440303002
Cr-Commit-Position: refs/heads/master@{#427212}

Previously we were only flushing queued messages in OnProcessLaunched.
This means that any messages queued during initialization would never
be delivered for in-process renderers.

Specifically this can negatively affect webview behavior in production.

BUG=656582
R=jam@chromium.org

Review-Url: https://codereview.chromium.org/2448663003
Cr-Commit-Position: refs/heads/master@{#427211}

Previously clank would not call View#startActionMode for a FloatingPastePopup menu
if the copy/paste clipboard was empty (because there would be nothing in the menu).
This is a problem because apps using WebView may want to add items to the
FloatingPastePopup menu (using startActionMode) even if the clipboard is empty.

This patch makes clank call View#startActionMode and make an empty menu when there
is nothing to paste.

BUG=652916

Review-Url: https://codereview.chromium.org/2443963003
Cr-Commit-Position: refs/heads/master@{#427210}

This patch addresses the following problems:
* Offline icon is not shown in omnibox on tablets
* Offline verbose status is sometimes shown with green padlock
* URL emphasis lags behind security icon update.

This fix does the following:
* Moves the offline icon from the navigation button to security button
* Ensures that navigation button is never shown on phones
* Calculates when to show: navigation/security/no button, to properly
  show or hide icon container.
* Moves verbose status visibility control to security button related
  code (as that is where we show offline stuff and would show verbose
  security states)

BUG=656088,648129
R=tedchoc@chromium.org

Review-Url: https://codereview.chromium.org/2438413002
Cr-Commit-Position: refs/heads/master@{#427209}

BUG=656572, 658668
CQ_INCLUDE_TRYBOTS=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/2443413002
Cr-Commit-Position: refs/heads/master@{#427208}

This is an automated CL created by the recipe roller. This CL rolls recipe
changes from upstream projects (e.g. depot_tools) into downstream projects
(e.g. tools/build).

More info is at https://goo.gl/zkKdpD. Use https://goo.gl/noib3a to file a bug
(or complain)

build:
  https://crrev.com/71945ff218759ff4121b5c3f8ddc4843bba423c7 Roll recipe dependencies (trivial). (recipe-roller@chromium.org)
  https://crrev.com/93c1c44e6424b382a801036419c975fd4f7063ec Reland "Support xctests" (smut@google.com)
  https://crrev.com/a3a2062832e5218037413c5495ffc093d4b1ff8e Roll recipe dependencies (trivial). (recipe-roller@chromium.org)
depot_tools:
  https://crrev.com/1a8439a7d9d0593a953a89b16cc154cd35f6a34e Clean up gclient sync's construction of the set of modified files (agable@chromium.org)
  https://crrev.com/c972d189fa495709f455b060a3c963e482d19a25 Remove SVN support from checkout.py (agable@chromium.org)

TBR=martiniss@chromium.org,phajdan.jr@chromium.org

Recipe-Tryjob-Bypass-Reason: Autoroller
Bugdroid-Send-Email: False
Review-Url: https://codereview.chromium.org/2449703002
Cr-Commit-Position: refs/heads/master@{#427207}

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2444643002
Cr-Commit-Position: refs/heads/master@{#427206}

When a download's MIME type is remapped, Chrome should use the new MIME
type.
However, when opening a download from download home, GUID is used to
find the download.
The download history stores the original MIME type, which is not
remapped.
This causes Chrome to incorrectly launch an intent even if the type is
supported.

BUG=657594

Review-Url: https://codereview.chromium.org/2437243002
Cr-Commit-Position: refs/heads/master@{#427205}

When clicking on download notification, Chrome checks if the download file
can be opened by itself.
This calls into a native function.
However, if user kills Chrome, and then clicks the notification, this
generates a Broadcast to Chrome.
When BroadcastReceiver runs, native library is not loaded.
As a result, this generates a crash.
This CL solves the problem by performing the mime type checking when
download completes.
So that we don't need to check it when BroadcastReceiver runs

BUG=658012

Review-Url: https://codereview.chromium.org/2446643004
Cr-Commit-Position: refs/heads/master@{#427204}

https://chromium.googlesource.com/external/github.com/googlei18n/sfntly/+log/1ef790a..e33ba7a

e33ba7a Merge pull request #60 from leizleiz/leizleiz-crbug
1bc53e1 Fix undefined shifts in ReadableFontData::ReadShort.
d651349 Check for negative size in NameTable::NameAsBytes.
8475d2f Avoid NULL derefs inside FontHeaderTable::Builder.
1fba3b3 Fix undefined shifts in ReadableFontData::ReadLong.
083b02b Fix NULL pointer derefs in sfntly::Font::Builder.
6d1efaa Fix out of bound access in subtly sample program.
cafc4c8 Merge pull request #59 from HalCanary/pronounciation
7d5169e README: pronounciation guide

BUG=638573,641452,646300,646347,654663,655914
TBR=behdad@chromium.org

Review-Url: https://codereview.chromium.org/2444123002
Cr-Commit-Position: refs/heads/master@{#427203}

This CL compiles PermissionInfoBarDelegate only on Android. It also
refactors the delegate subclasses to move their Create methods to
PermissionInfoBarDelegate, simplifying the PermissionsQueueController.

This refactor simplifies the work necessary to allow infobars to be
switched out for modal permission prompt dialogs by centralising the
delegate creation logic.

BUG=658125

Review-Url: https://codereview.chromium.org/2443463003
Cr-Commit-Position: refs/heads/master@{#427202}

Creating derived constraint spaces from the layout opportunity iterator wasn't being used for the float case, and isn't needed for the text case either. Removing!

BUG=635619

Review-Url: https://codereview.chromium.org/2438313003
Cr-Commit-Position: refs/heads/master@{#427201}

Also, refactor buffer access validation a bit to reduce
duplicated codes.

(Note that this is still part 1 of the work. We have more
buffer access paths that we need to put in buffer access
validation.)

BUG=654201
TEST=gpu_unittests, webgl_conformance
R=piman@chromium.org
CQ_INCLUDE_TRYBOTS=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/2435803004
Cr-Commit-Position: refs/heads/master@{#427200}

FLAG_ACTIVITY_RETAIN_IN_RECENTS is required on L+ or
the activities disappear on upgrade.

BUG=656196

Review-Url: https://codereview.chromium.org/2445133002
Cr-Commit-Position: refs/heads/master@{#427199}

They already work.

BUG=504347,575215

Review-Url: https://codereview.chromium.org/2448623002
Cr-Commit-Position: refs/heads/master@{#427198}

We're crashing due to a null m_stateRestorer in the fuzzer. It's unclear
how this is possible, so add CHECKs to indicate where this is going
wrong.

These CHECKS should be changed to DCHECKs when the issue is addressed.

BUG=658265

Review-Url: https://codereview.chromium.org/2439073003
Cr-Commit-Position: refs/heads/master@{#427197}

Before, switching devices in Device Mode would set the zoom scale to fit the
window. Now, devices will remember their last used mode.

BUG=656020

Review-Url: https://codereview.chromium.org/2431823002
Cr-Commit-Position: refs/heads/master@{#427196}

BUG=653836
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2428083002
Cr-Commit-Position: refs/heads/master@{#427195}

Most of these tests are testing conformance for TypedArrays, ArrayBuffer,
bufferData, and some basic WebGL functions. I believe all these tests
are upstreamed in Github, so there is no reason to keep it here.

Note that I didn't delete all tests under this dir, only the ones I
can quickly locate its duplicates on conformance test suite.

BUG=none

Review-Url: https://codereview.chromium.org/2447493002
Cr-Commit-Position: refs/heads/master@{#427194}

BUG=555865

Review-Url: https://codereview.chromium.org/2442953002
Cr-Commit-Position: refs/heads/master@{#427193}

Also, add GetAppIndex() to ArcNavigationThrottle and use it from
ArcExternalProtocolDialog to reduce code duplication.

BUG=None
TEST=try

Review-Url: https://codereview.chromium.org/2443313002
Cr-Commit-Position: refs/heads/master@{#427192}

BUG=658124
TEST=self sustained

Review-Url: https://codereview.chromium.org/2442933002
Cr-Commit-Position: refs/heads/master@{#427191}

Changes serviceConnectionError() to be a no-op if Platform::current()
is null. This can happen because MessageLoop destruction may invoke
the callback after Blink shutdown.

TBRing since this precise change was recommended by haraken@ and I'd
like to make sure the fix lands ASAP.

BUG=650637
TBR=haraken@chromium.org

Review-Url: https://codereview.chromium.org/2446723002
Cr-Commit-Position: refs/heads/master@{#427190}

Revert of Add threadsafe version of PermissionManager::GetPermissionStatus (patchset #5 id:80001 of https://codereview.chromium.org/2439673004/ )

Reason for revert:
Going to address a number of issues that Raymes pointed out including:

1) It's unclear the purpose the overloaded GetPermissionStatus function (see https://google.github.io/styleguide/cppguide.html#Function_Overloading). It could be good to name it something that reflects its usage from a background thread.

2) In PermissionManager, it's unclear that it is threadsafe while the rest of the class isn't - we should add a comment.

3) It's unclear why HostContentSettingsMap needs to be passed in - it's already available in the PermissionContextBase.

4) I feel like it would be better to leave the default implementation empty in the PermissionContextBase and have NOTREACHED. Then if someone really wants to support querying a permission safely from a background thread, they have to think about it.

Original issue's description:
> Add threadsafe version of PermissionManager::GetPermissionStatus
>
> Add a version of GetPermissionStatus that takes a host_content_settings_map.
> This method is guaranteed to be threadsafe.
>
> BUG=658020
>
> R=jochen, mlamouri
>
> Committed: https://crrev.com/d6b19d48cf91f89c1ec414485b781f2ef2434521
> Cr-Commit-Position: refs/heads/master@{#427078}

TBR=mlamouri@chromium.org,jochen@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=658020

Review-Url: https://codereview.chromium.org/2446863002
Cr-Commit-Position: refs/heads/master@{#427189}

The transferToImageBitmap is an API in OffscreenCanvas. This API could
throw exception in certain cases, which means that the
OffscreenCanvas::transferToImageBitmap() method has a ExceptionState&
parameter.

In our current implementation, we do all the exception check in the above
method. However, we still pass the ExceptionState& parameter to the
CanvasRenderingContext::transferToImageBitmap() method, and this method
is overriden by OffscreenCanvasRenderingContext2D, WebGLRenderingContext,
and WebGL2RenderingContext. So this method in all these sub-classes also
take a ExceptionState& parameter but do nothing with it. We should not
need to pass this parameter to CanvasRenderingContext::transferToImageBitmap.

TBR=kbr@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2440213003
Cr-Commit-Position: refs/heads/master@{#427188}

Added by mistake in https://chromiumcodereview.appspot.com/2436913003

BUG=webrtc:6376

Review-Url: https://codereview.chromium.org/2450553002
Cr-Commit-Position: refs/heads/master@{#427187}

BUG=656099

Review-Url: https://codereview.chromium.org/2440733002
Cr-Commit-Position: refs/heads/master@{#427186}

Memory coordinator should update renderers' memory state when their
visibility have changed. For a visible (foreground) renderer,
the state is set to NORMAL. For background renderers, the states
are set to the global state.

BUG=617492

Review-Url: https://codereview.chromium.org/2443093002
Cr-Commit-Position: refs/heads/master@{#427185}

When compiling against the 10.9 SDK, IOBluetoothAddServiceDict() and
IOBluetoothRemoveServiceWithRecordHandle() generate deprecation errors.
This cl removes the instances of this old API.

R=isherman@chromium.org
BUG=650836

Review-Url: https://codereview.chromium.org/2393723002
Cr-Commit-Position: refs/heads/master@{#427184}

The new ResourceHandler's OnWillStart was never being called, which
violates the ResourceHandler's API contract, and has caused at least one
crash previously (Though that fixed by another CL).

BUG=650717

Review-Url: https://codereview.chromium.org/2436163002
Cr-Commit-Position: refs/heads/master@{#427183}

Some targets (e.g. unit_tests_apk) are starting to hit the limit!

BUG=658782

Review-Url: https://codereview.chromium.org/2445883002
Cr-Commit-Position: refs/heads/master@{#427182}

BUG=600808

Review-Url: https://codereview.chromium.org/2444493002
Cr-Commit-Position: refs/heads/master@{#427181}

For most regular fetches, we apply the user agent twice:
- addAdditionalRequestHeaders
- prepareRequest

This patch changes FrameLoader::applyUserAgent to
FrameLoader::applyUserAgentIfNeeded which first checks if the user agent
is set. This avoids going to the embedder to calculate the override twice.

BUG=348655

Review-Url: https://codereview.chromium.org/2445653003
Cr-Commit-Position: refs/heads/master@{#427180}

This is an automated CL created by the recipe roller. This CL rolls recipe
changes from upstream projects (e.g. depot_tools) into downstream projects
(e.g. tools/build).

More info is at https://goo.gl/zkKdpD. Use https://goo.gl/noib3a to file a bug
(or complain)

build:
  https://crrev.com/9488f450ac9e1b5cf503cd1bb76a964dab60fc57 Add WebView Ui Test to chromium.fyi (yolandyan@chromium.org)
  https://crrev.com/3a8bfe4caf559c1126a5761154c778dd7ff887de Support xctests (smut@google.com)
  https://crrev.com/fc06b185bc76c25108800aa40525268d42da6deb Roll recipe dependencies (trivial). (recipe-roller@chromium.org)
  https://crrev.com/284d0f102f55bf7c629a297d89c04b8fde110020 Disable galaxy and new mac perf bots for SOM (martiniss@chromium.org)
  https://crrev.com/b9e5951f09ad926b5cb8e0563c5c56b3f4e678e6 Revert "Support xctests" (smut@google.com)
  https://crrev.com/997541305c3617899e17a76fe9c0e5d9a517bdd0 Roll recipe dependencies (trivial). (recipe-roller@chromium.org)
depot_tools:
  https://crrev.com/1fdfcb63166412adae9361acf85705c5290523c6 Refactoring: Extract helper functions from CMDtry in git_cl.py. (qyearsley@chromium.org)
  https://crrev.com/037edb823f045ca3e6a13217efd0fcbdcfb2a384 Remove SVN support from apply_issue (agable@chromium.org)
  https://crrev.com/83faed04862b6c6b7b2034bb256687afeb9d31c0 Remove unnecessary rev_str variable from gclient_scm (agable@chromium.org)

TBR=martiniss@chromium.org,phajdan.jr@chromium.org

Recipe-Tryjob-Bypass-Reason: Autoroller
Bugdroid-Send-Email: False
Review-Url: https://codereview.chromium.org/2447913002
Cr-Commit-Position: refs/heads/master@{#427179}

In Ash, there is logic to close menus then asynchronisly perform an accelerator command. However, the accelerator that opens the context menu itself can and does trigger the close path depending on how a person executes the command.

On Chrome OS, the shortcut to open a context menu is Search+Shift+Volume Up. Any of these keys should leave the menu as is otherwise, the context menu is effectively a no-op when triggered from the keyboard.

TEST=on device, hit search; verify apps list opens. Press Search+Shift+Volume up; verify context menu opens. Press Search+m with ChromeVox on, verify context menu opens.
BUG=621331,616130

Review-Url: https://codereview.chromium.org/2434483003
Cr-Commit-Position: refs/heads/master@{#427178}

When mapping from a fixed to an absolute ancestor, mapLocalToAncestor()
deals with the case that the ancestor is skipped by fixed's container,
but forgot to adjust fixed-position offset.

Also fix the similar issue of mapAncestorToLocal().

BUG=656172
TEST=MapCoordinatesTest.FixedPosIn(Fixed|Absolute)PosScrollView
TEST=paint/invalidation/fixed-with-border-under-composited-absolute-scrolled.html
NOTRY=true (to workaround OOM of android bot)

Review-Url: https://codereview.chromium.org/2436283002
Cr-Commit-Position: refs/heads/master@{#427177}