GitLab

Instead of using the locale zip in dep graph, alternative locale zip
could be used when building APK.

- Using dedicated variables in all_dependencies_settings for locale zip
- Not using locale zip from all_dependences_settings if alternative file
  specified.
- Make resources_zip_path in locale_pak_resources.gypi setable, so the
  file could be used as alternative lcoale zip for APK.

BUG=535379

Review URL: https://codereview.chromium.org/1357423011

Cr-Commit-Position: refs/heads/master@{#351190}

Previously, if a user navigated to a subpage directly via URL, i.e.
chrome://settings/searchEngines, and then tried to close that subpage,
the animation would not work correctly.

This patch fixes that issue.

BUG=535013

Review URL: https://codereview.chromium.org/1372713002

Cr-Commit-Position: refs/heads/master@{#351189}

$ git log 2eb98d819..772999bc2 --date=short --format='%ad %ae %s'
2015-09-28 andybons fix comment typo in BUILD.gn
2015-09-28 andybons Update build file to use asmflags instead of cflags.

BUG=none
R=brettw@chromium.org
CC=ajm@google.com

Review URL: https://codereview.chromium.org/1374003002

Cr-Commit-Position: refs/heads/master@{#351188}

This change implements FilteringNetworkManager which wraps IpcNetworkManager from PeerConnectionDependencyFactory and exposes rtc::NetworkManager to WebRTC. P2PNetworkManager will check mic/camera permissions and only if at least one of them is granted, the full network list from IpcNetworkManager is provided to WebRTC calls. Otherwise, WebRTC can only bind to the "any" address.

Currently, this is disabled by default and can be enabled with a finch experiment "WebRTC-LocalIPPermissionCheck". UMAs are added to track how many calls are impacted (either denied due to lack of permissions, or call set up time delayed) before fully turned on.

Test cases added for FilteringNetworkManager to test its handling of api StartUpdating() and asynchronous signals such as permission status update and IpcNetworkManager's SignalNetworksChanged in different order.

When multiple_routes option is not requested, EmptyNetworkManager is used instead which simply just returns BLOCKED as permission status which will cause WebRTC to use the "any" address networks.

This depends on the ongoing CL for MediaPermissionDispatcher https://codereview.chromium.org/1351443005/

In my test with my dev machine, I don't see the setup time has been increased by this call change and we have UMA to check that too.

BUG=520101

Review URL: https://codereview.chromium.org/1349823004

Cr-Commit-Position: refs/heads/master@{#351187}

BUG=None

Review URL: https://codereview.chromium.org/1374533003

Cr-Commit-Position: refs/heads/master@{#351186}

The existing PropertySet::Get function is asynchronous.
This CL adds the synchronous version of PropertySet::Get.
It is defined as PropertySet::GetAndBlock.

Also fix some typos and indent problems in previous
introduced SetAndBlock().

BUG=https://b.corp.google.com/u/0/issues/24131409
TEST=manually tested

Review URL: https://codereview.chromium.org/1368713002

Cr-Commit-Position: refs/heads/master@{#351185}

This changes adds a new entry point for the Media Router dialog by adding a "Cast..." menu item into the overflow/wrench menu as well as page/audio/video contextual menus.

This new menu item is currently only surfaced if Media Router is enabled (--enable-media-router) and if the user is not in Incognito mode. Once Media Router is re-enabled for Incognito, that check should be removed.

BUG=534841

Review URL: https://codereview.chromium.org/1360323002

Cr-Commit-Position: refs/heads/master@{#351184}

On OS X Snow Leopard and Lion (currently supported by Chrome) the
kIOHIDReportDescriptorKey is not available as it is in later versions of
the operating system. Since we are also unable to get the device's
report descriptor on Windows it seems okay to just ignore the failure.

BUG=479210

Review URL: https://codereview.chromium.org/1373923003

Cr-Commit-Position: refs/heads/master@{#351183}

Relanding https://codereview.chromium.org/1364043003/ after fixing
Svelte crash and CustomTabs test failure.

- Svelte was crashing due to getActivityTab being null when
  onOverviewModeFinishedHiding is called.
- CustomTabsActivity sets the status bar color explicitly and
  does not need to inherit setStatusBarColor calls from ChromeActivity.

BUG=535235,536698

Review URL: https://codereview.chromium.org/1375693002

Cr-Commit-Position: refs/heads/master@{#351182}

This patch creates a solid color tiling that pretends that it's not
solid, so that tiles can be created. This ensures that we don't try
to allocate resources/staging buffers for the tiles, and avoid a crash.

BUG=534911
R=danakj
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1365793003

Cr-Commit-Position: refs/heads/master@{#351181}

One of the design considerations in https://codereview.chromium.org/1358173005/ was that using base.copyWithoutNullFields made the code future-proof if new options are added. This unit test ensures that that holds true and also improves some of the existing unit tests to catch some obvious errors.

Review URL: https://codereview.chromium.org/1366093003

Cr-Commit-Position: refs/heads/master@{#351180}

BUG=490488

Review URL: https://codereview.chromium.org/1376533002

Cr-Commit-Position: refs/heads/master@{#351179}

Xcode 7.0 has been publicly released and iOS downstream requires it to build
so use the iOS 9.0 SDK with Xcode 7.0 for the bots upstream.

Run the tests on all the currently supported iOS version 7.1, 8.3 and 9.0.

BUG=516606

Review URL: https://codereview.chromium.org/1369243002

Cr-Commit-Position: refs/heads/master@{#351178}

This change is needed for testing recoverable SSL interstitials for
WKWebView.

BUG=462427

Review URL: https://codereview.chromium.org/1358033002

Cr-Commit-Position: refs/heads/master@{#351177}

The memory dump provider for OutputSurface was registered on the
client thread, but could be unregistered on the main thread in cases when
the client was cleaned up before the destructor was called (very common).
Move the unregistration code to the parallel of the registration code, so that
both only when appropriate, on the client thread.

BUG=536824
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1371163002

Cr-Commit-Position: refs/heads/master@{#351176}

Tints all action-bar menu icons so they are the same color as the
system icons (home/up button and three-dots overflow icon).

Review URL: https://codereview.chromium.org/1369653002

Cr-Commit-Position: refs/heads/master@{#351175}

These are types that we've seen on page_cycler.typical_25.

BUG=

Review URL: https://codereview.chromium.org/1372273002

Cr-Commit-Position: refs/heads/master@{#351174}

The "N cookies from this site" link gets updated in real time (to update the
value of N). That causes the link to lose click events and its tab index.
Solve the problem by only updating the link text, rather than destroying and
rebuilding the entire layout.

BUG=535587

Review URL: https://codereview.chromium.org/1372453002

Cr-Commit-Position: refs/heads/master@{#351173}

This allows users and integration tests to use the Data Saver API, but
override the list of Data Reduction Proxies to use for HTTP proxies with
a custom list.

BUG=526827

Review URL: https://codereview.chromium.org/1345523005

Cr-Commit-Position: refs/heads/master@{#351172}

Revert of scheduler: Add a base directory (patchset #4 id:60001 of https://codereview.chromium.org/1374653003/ )

Reason for revert:
Change causes compile failure:
http://build.chromium.org/p/chromium.linux/builders/Linux%20GN%20Clobber/builds/6135

Original issue's description:
> scheduler: Add a base directory
>
> This patch moves the more fundamental scheduling primitives (e.g.,
> TaskQueueManager) to a separate base/ directory. This makes the layered
> architecture of the scheduler more obvious and reduces the likelihood of
> accidentally introducing layering violations.
>
> With this refactoring, the overall dependencies are:
>
>        +------+
>        | base |
>        +------+
>            ^
>            |
>        +-------+     +------+
>        | child |<----| test |
>        +-------+     +------+
>           ^ ^          ^ ^
>           | |          | |
>           | '----------|-|---.
>           | .----------' '-. |
>           | |              | |
>       +----------+      +-------+
>       | renderer |      | ppapi |
>       +----------+      +-------+
>
> Committed: https://crrev.com/516bee16e9437521cdd20a78d683f9dbaf35df3e
> Cr-Commit-Position: refs/heads/master@{#351109}

TBR=jochen@chromium.org,alexclarke@chromium.org,skyostil@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1370343002

Cr-Commit-Position: refs/heads/master@{#351171}

Revert of Small fix for URLRequestJobs when entering suspending mode. (patchset #3 id:40001 of https://codereview.chromium.org/1366203004/ )

Reason for revert:
Seeing odd failures that make no sense on android bots that may somehow be due to this CL...Trying a revert.  :(

Original issue's description:
> Small fix for URLRequestJobs when entering suspending mode.
>
> When entering suspend mode, there was a window of time between
> when the request was canceled, and when the NetworkDelegate was
> informed it was canceled.  This didn't exist for other cancellation
> paths, and if an embedder invoked certain callbacks in this period,
> thinking the request was still live, crashes could result.
>
> BUG=289715
>
> Committed: https://crrev.com/05608da225e05b4a20da3f4ca0cb9a5276fcf3d6
> Cr-Commit-Position: refs/heads/master@{#350938}

TBR=davidben@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=289715

Review URL: https://codereview.chromium.org/1372213003

Cr-Commit-Position: refs/heads/master@{#351170}

R=prasadv@chromium.org
BUG=432959

Review URL: https://codereview.chromium.org/1375693003

Cr-Commit-Position: refs/heads/master@{#351169}

BUG=520266

Review URL: https://codereview.chromium.org/1364913005

Cr-Commit-Position: refs/heads/master@{#351168}

Revert of Enable subsequence caching (patchset #3 id:40001 of https://codereview.chromium.org/1366163002/ )

Reason for revert:
There are still some regressions.

BUG=410097,536764,536674,533717

Original issue's description:
> Enable subsequence caching
>
> BUG=410097
>
> Committed: https://crrev.com/80bdce2c101c3c6a8a2383b8d7035df7bdd5c0c4
> Cr-Commit-Position: refs/heads/master@{#350965}

TBR=chrishtr@chromium.org,pdr@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=410097

Review URL: https://codereview.chromium.org/1377643002

Cr-Commit-Position: refs/heads/master@{#351167}

V8DebuggerAgent was split into interface and implementation. All other agents should depend on the public part only.

Once InjectedScript, InjectedScriptHost and InjectedScriptManager are also moved to .../inspector/v8 some of the methods on V8DebuggerAgent will be removed and called only by the implementation classes.

BUG=439376

Review URL: https://codereview.chromium.org/1367623006

Cr-Commit-Position: refs/heads/master@{#351166}

This CL adds and implements NetworkDelegate::OnNetworkBytesSent(), which
is used to notify the network delegate when bytes have been sent over
the network.

BUG=518051

Review URL: https://codereview.chromium.org/1362793003

Cr-Commit-Position: refs/heads/master@{#351165}

If a user quits Chrome with two windows, both docked, on restore Chrome
will notice that neither window is fully visible and select the first
window in its restoration list to make visible. However, the profile
also has a notion of an active window, which session restore auto-
matically makes fully visible. The result is that depending on the order
in which the windows were placed in the dock, both windows can end up
restored fully visible instead of only one of them (which ideally should
be the last active window).

This change causes session restore to restore the last active window to
fully visible if all windows are hidden and the last active window
appears in the window restoration list.

BUG=536201

Review URL: https://codereview.chromium.org/1366313003

Cr-Commit-Position: refs/heads/master@{#351164}

At rockot@'s suggestion, I tested this with a small Chrome app that I
wrote (https://github.com/zeptonaut/serial-device-enumerator-app) that
opens a window showing all information that Chrome knows about serial
devices. We chose this route over unit tests because we didn't think
it was wise to write unit tests that assumed specific devices were
plugged into our test servers.

As far as devices, I used an FTDI Chipset High Speed USB 2.0 to Serial
RS-232 DB-9 Converter. This was useful because it's one of not-so-many
devices that has both serial properties (COM port) and USB properties
(vendor ID, device ID, and display name).

BUG=522217

Review URL: https://codereview.chromium.org/1357993002

Cr-Commit-Position: refs/heads/master@{#351163}

BUG=529471

Review URL: https://codereview.chromium.org/1365443004

Cr-Commit-Position: refs/heads/master@{#351162}

Per bug, we would send duplicate queries upon parsing previously-known forms on a page.

BUG=512173

Review URL: https://codereview.chromium.org/1371713002

Cr-Commit-Position: refs/heads/master@{#351161}

BUG=521646

Review URL: https://codereview.chromium.org/1364103003

Cr-Commit-Position: refs/heads/master@{#351160}

BUG=528847

Review URL: https://codereview.chromium.org/1365813003

Cr-Commit-Position: refs/heads/master@{#351159}

BUG=536092
TEST=None

Review URL: https://codereview.chromium.org/1373873003

Cr-Commit-Position: refs/heads/master@{#351158}

BUG=536832

Review URL: https://codereview.chromium.org/1369363002

Cr-Commit-Position: refs/heads/master@{#351157}

- download item click target takes up the entire vertical space
- move progress indicator padding from implicit (in DownloadShelf) to explicit (in relevant view)
- MD for shelf controls (show all, close buttons)

Unfortunately I did have to touch the existing shelf code, and therefore risk breaking it, but my visual inspection shows no changes to the appearance in LTR or RTL, warning state or normal state.

BUG=505804

Review URL: https://codereview.chromium.org/1367083002

Cr-Commit-Position: refs/heads/master@{#351156}

Previously the LayerTreeImpl's CurrentScrollingLayer could be either
the inner viewport or the outer viewport. It's clearer if it only ever
takes one of these values.

BUG=526462
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1365343002

Cr-Commit-Position: refs/heads/master@{#351155}

Also added a test to prevent a regression.

Essentially, ScreenLocker::HandleLockScreenRequest calls UserAddingScreen::Cancel and uses a callback to be notified for when the user-adding screen is completed. The crash happened because this callback was invoked too early - before the adding screen has been fully deinitialized.

BUG=467111

Review URL: https://codereview.chromium.org/1356713004

Cr-Commit-Position: refs/heads/master@{#351154}

* This houses previously static methods of NetworkPortalDetector, such as Get (now GetInstance), IsInitialized, InitializeForTesting, etc.
* InitializeNetworkPortalDetector moved to cbmc
* Number of renames of NetworkPortalDetector:: to network_portal_detector::

Files with non-trivial changes:
network_portal_detector_impl.{cc|h}
network_portal_detector.{cc|h}
chrome_browser_main_chromeos.cc

Everything else is a mechanical search/replace

BUG=530485
TEST=unit tests and browser tests.
TBR=kalman@chromium.org

Review URL: https://codereview.chromium.org/1358343003

Cr-Commit-Position: refs/heads/master@{#351153}

This patch adds memory tracing for GpuMemoryBuffers and graphics resources
used by media in GpuMemoryBufferVideoFramePool.

BUG=

Review URL: https://codereview.chromium.org/1355373003

Cr-Commit-Position: refs/heads/master@{#351152}

https://codereview.chromium.org/1376483002 broke trigger-only AMP calls
to build/android/test_runner.py. This fixes that workflow.

BUG=
TBR=mikecase@chromium.org

Review URL: https://codereview.chromium.org/1371243002

Cr-Commit-Position: refs/heads/master@{#351151}