GitLab

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/465217f..21b204b

TBR=dpranke@chromium.org,caseq@chromium.org
INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_chromium_dbg_32_ng

Review URL: https://codereview.chromium.org/909983002

Cr-Commit-Position: refs/heads/master@{#315399}

BUG=456378

Review URL: https://codereview.chromium.org/908633002

Cr-Commit-Position: refs/heads/master@{#315398}

This is the last stage of the trace_event directory restructuring.
This is part of a set of 3 CLs which is moving tracing clients to use
the new base::trace_debug namespace.
See crrev.com/837303004 and the related bug for motivations.

BUG=451032
TBR=jam@chromium.org

Review URL: https://codereview.chromium.org/868603007

Cr-Commit-Position: refs/heads/master@{#315397}

This changelist:

Adding a basic UI for displaying host-offline-reason:
- When no host-offline-reason is provided, old behavior is preserved

- When a host-offline-reason is provided, the text used to present a host
  says:        "<hostname> (reported an error <time>)"
  rather than: "<hostname> (last online 11/11/2014)"

- When a known host-offline-reason is provided, a tooltip is added to the host
  giving localized description of the specific host-offline-reason.  Example
  tooltip: "Invalid host configuration."

- When an unknown host-offline-reason is provided (one without a localized
  description text), a generic tooltip is added saying:
  "Unrecognized error: INVALID_HOST_CONFIGURATION."

Context:

After crrev.com/719983002 Chromoting host sends an host-offline-reason string
before exiting.  The host-offline-reason value is currently always based on
ExitCodeToString function / HostExitCodes enum, but in principle
host-offline-reason is not necessarily tied to an exit code (i.e. in the future
the host should send "INVALID_POLICY" host-offline-reason but rather than
exiting should hang around monitoring policy for changes).  The
host-offline-reason sent by the host is stored in the Chromoting directory.
Since cr/79856832the directory sends host-offline-reason to the client as
hostOfflineReason field inside a Host object.

BUG=410050,247245

TEST=Manually tried the UI in the 3 cases: hor missing, known hor, unknown hor.

Review URL: https://codereview.chromium.org/882463003

Cr-Commit-Position: refs/heads/master@{#315396}

BUG=396634

Review URL: https://codereview.chromium.org/883403002

Cr-Commit-Position: refs/heads/master@{#315395}

When failing to connect to a previously-paired device, the profile would
not be cleared.  This sets adapter_ in RegisterProfile() so that it will
be set correctly for profile release later.

Also includes unit tests, adding a paired but non-connectable device.

BUG=455868

Review URL: https://codereview.chromium.org/910433002

Cr-Commit-Position: refs/heads/master@{#315394}

This patch changes a few behaviors.

1. Make implicitly sized (e.g. width="100%") plugins work with Plugin Power Saver. Previously sizes were read from plugin parameters, which doesn't work for implicitly sized plugins. Now, we look at the actual blink::WebRect view bounds.

2. Timeout keyframe extraction after 150 frames rather than a fixed 5 seconds. The previous behavior was broken for plugins that are loaded off-screen. (The countdown begins even though no frames are being generated until the plugin goes on-screen).

3. When "Detect and run..." is on, always respect the 'poster' parameter. This is a tweak to the existing behavior, which is to use the 'poster' param only for peripheral plugins. This should have no user impact, since no one actually uses the 'poster' param right now.

4. Clean up metric collection in LoadablePluginPlaceholder. Previous behavior was overcounting in some circumstances.

5. Reduces RenderFrame public interface and total SLOC.

BUG=443431,456217, 403800

Review URL: https://codereview.chromium.org/904913003

Cr-Commit-Position: refs/heads/master@{#315393}

Revert of Disable chromevox_tests BackgroundTest.InitialFeedback (patchset #2 id:20001 of https://codereview.chromium.org/908913002/)

Reason for revert:
Cl which caused this flake was reverted.

Original issue's description:
> Disable chromevox_tests BackgroundTest.InitialFeedback
>
> The test attempts to monitor speech immediately after startup. Disabling for now until we can figure out why it flakes on bots.
>
> NOTRY=true
> TBR=dtseng@chromium.org
>
> Committed: https://crrev.com/94007abc03dff9620d5fd64237b1fdb3c0c19957
> Cr-Commit-Position: refs/heads/master@{#315374}

TBR=
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/895533003

Cr-Commit-Position: refs/heads/master@{#315392}

FILE_ATTRIBUTE_NOT_CONTENT_INDEXED may or may not be set on a temp
directory depending on the OS version and location of the directory.

BUG=455470

Review URL: https://codereview.chromium.org/906173005

Cr-Commit-Position: refs/heads/master@{#315391}

    a)
    The swap accounting is a bit unreliable when the GPU process
    crashes or the context is lost. We might end up thinking that
    too many swap acks are pending (which will never arrive) and
    bail out of CompositorImpl::Composite() early.
    Ignore the swap limit when deciding to composite if we know
    the context was lost. Then calling Composite() will cause the
    output surface to get recreated.

    b)
    Fix infinite retry loops of which there are two:
    One is that we could indefinitely early-out of Composite()
    and repost if the GPU process keeps crashing at startup before
    LTH ever gets a chance to try create a context.

    Fix regression with https://codereview.chromium.org/738983002/
    which both removed the LOG(FATAL) for too many successive retry
    attempts (used to happen inside LTH), and added a similar retry
    loop where we might never call SetOutputSurface and keep
    relaunching the GPU process.

    Also essentially revert https://codereview.chromium.org/750643003/
    to simplify things. We don't need to early-out in Composite()
    anymore if the GPU channel is lost. This is handled better in
    cc now and we can just attempt to Composite(). It will schedule
    OutputSurface creation. Even if we call Composite() again in the
    meantime, that should be harmless. And after we eventually call
    SetOutputSurface(), this will trigger ScheduleComposite().

BUG=453511,453671,453819,453649,448549
NOTRY=True

Review URL: https://codereview.chromium.org/903933005

Cr-Commit-Position: refs/heads/master@{#315390}

This pulls in the following Native Client changes:

6c11549: (mseaborn@chromium.org) Add IRT interface for PNaCl's sandboxed linker to use to talk to Chromium
9fff96b: (teravest@chromium.org) NaCl: Remove use of NACL_BUILD_* in common.gypi.
76d6d4b: (mseaborn@chromium.org) PNaCl: Update Binutils revision in pnacl/COMPONENT_REVISIONS: use new IRT interface
a3e1e94: (jfb@chromium.org) Delete Windows time code
ea0a853: (mseaborn@chromium.org) Cleanup: Move declaration of NaClDebugExceptionHandlerRun() to src/public/
3fa058c: (mseaborn@chromium.org) Cleanup: Make nacl_desc required rather than optional in chrome_main.h
66f4b73: (mseaborn@chromium.org) Cleanup: Remove unused NaClBootstrapChannelErrorReporter and GetCrashLogOutput()

BUG=none
TEST=browser_tests and nacl_integration
CQ_EXTRA_TRYBOTS=tryserver.chromium.linux:linux_rel_precise32,linux_arm_compile,linux_nacl_sdk_build

Review URL: https://codereview.chromium.org/913553002

Cr-Commit-Position: refs/heads/master@{#315389}

TBR=kbr@chromium.org
NOTRY=true
BUG=456928

Review URL: https://codereview.chromium.org/905243005

Cr-Commit-Position: refs/heads/master@{#315388}

Instead of having separate actions for each test suite in components, provide a single target that can be used by both test targets.

BUG=451827
R=blundell@chromium.org,jam@chromium.org
TEST=components_unittests && components_browsertests

Review URL: https://codereview.chromium.org/900223003

Cr-Commit-Position: refs/heads/master@{#315387}

BUG=451286

Review URL: https://codereview.chromium.org/896583003

Cr-Commit-Position: refs/heads/master@{#315386}

The telemetry tests stopped using the real OSX keychain because the 10.9
machines were incorrectly configured. The 10.9 machines have had their
keychains fixed (Issue 397385). I've also added checks so that key Telemetry
tests log error messages if the OSX keychain is incorrectly configured
(http://code.google.com/p/chromium/issues/detail?id=443340#c3).

BUG=443340

Review URL: https://codereview.chromium.org/904163002

Cr-Commit-Position: refs/heads/master@{#315385}

BUG=

Review URL: https://codereview.chromium.org/905423002

Cr-Commit-Position: refs/heads/master@{#315384}

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/899453004

Cr-Commit-Position: refs/heads/master@{#315383}

Also, try enabling swarming for base_unittests on CrOS MSan.

The previous attempt broke every browser_test with GLib errors
(http://crbug.com/456205). I've committed a speculative fix for those.
Attempting to land again.

BUG=454828, 456205
TBR=phajdan.jr@chromium.org

Review URL: https://codereview.chromium.org/913583002

Cr-Commit-Position: refs/heads/master@{#315382}

It also add a check to verify that a media::VideoCaptureDevice is not deleted on the wrong thread.

BUG=450382

Review URL: https://codereview.chromium.org/899943004

Cr-Commit-Position: refs/heads/master@{#315381}

BUG=456902
TEST=none

Review URL: https://codereview.chromium.org/893343003

Cr-Commit-Position: refs/heads/master@{#315380}

BUG=none
TEST=none

Review URL: https://codereview.chromium.org/906793002

Cr-Commit-Position: refs/heads/master@{#315379}

We will start using these parameters in more callbacks, so it makes sense
to rename them into something more generic.

BUG=456782

Review URL: https://codereview.chromium.org/876413007

Cr-Commit-Position: refs/heads/master@{#315378}

WebsiteSettingsPopupLegacy was replaced with WebsiteSettingsPopup.
This change removes the old dialog and it's dependencies.

BUG=456167

Review URL: https://codereview.chromium.org/878413008

Cr-Commit-Position: refs/heads/master@{#315377}

An IRT layer is being added to Mojo NaCl support. The change to the Mojo
repository for this will temporarily break ppapi mojo support until a
corresponding chrome change is landed to make the Mojo IRT interface available
in Chrome.

BUG=454466
R=dmichael@chromium.org

Review URL: https://codereview.chromium.org/907913002

Cr-Commit-Position: refs/heads/master@{#315376}

Rolls in two CLs on Android tools:

1) Add aarch64 gold linker:
  https://chromium-review.googlesource.com/247060
2) Remove proguarding from GMS core
  https://chromium-review.googlesource.com/247611

BUG=None

Review URL: https://codereview.chromium.org/906363002

Cr-Commit-Position: refs/heads/master@{#315375}

The test attempts to monitor speech immediately after startup. Disabling for now until we can figure out why it flakes on bots.

NOTRY=true
TBR=dtseng@chromium.org

Review URL: https://codereview.chromium.org/908913002

Cr-Commit-Position: refs/heads/master@{#315374}

Revert of Reland #2: Ensure WebView notifies desktop automation on creation, destruction, and change  Original (patchset #7 id:120001 of https://codereview.chromium.org/890013006/)

Reason for revert:
Suspected to cause chromeos vox test failure:
http://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%281%29/builds/56046

Original issue's description:
> Reland #2: Ensure WebView notifies desktop automation on creation, destruction, and change
>
> Original issue: https://codereview.chromium.org/880063002
> Previous reland attempt: https://codereview.chromium.org/895623003/
>
> This cl changes the way in which ChromeVox initially verbalizes a page which led to flakeyness in
>   chromevox_tests --gtest_filter=BackgroundTest.*
>
> To attempt and address the flakeyness,
> - chromevox_e2e_test_base.js and chromevox_next_e2e_test_base.js now explicitly separates listening to page load events from the chrome.automation tree and the chrome.tabs APIs.
> They are named: runWithLoadedTree, runWithLoadedTab, and runWithTab. The last method allows a caller to create a tab, but not wait for it to load (based on the tab status being 'complete').
> - each test in background_test now no longer waits for ChromeVox to speak 'start' (which formerly could have been triggered by a loadComplete automation event).
> - tab creation is still done via the tabs API, but listening for load is done via chrome.automation. (i.e. add a listener for an automation tree to load completely, then create a tab).
>
> It is not known at this time if this cl will continue to cause flakes in the above tests. Try bots and local runs yielded no flakeyness. The next step may be to introduce logging to identify the source of the flakeyness.
>
> TBR=tfarina
>
> Committed: https://crrev.com/9b006b56ca402da3774708f08e062c89018d57a6
> Cr-Commit-Position: refs/heads/master@{#315362}

TBR=dmazzoni@chromium.org,tfarina@chromium.org,dtseng@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/910013002

Cr-Commit-Position: refs/heads/master@{#315373}

ObserverListThreadSafe::Notify is the 4th most called function in a
profile generated on my host. This is because all tasks posted from it
are stamped with this callsite, providing very little information.

This patch forces callers to pass in their own tracking data so that
the profile will differentiate between each observer list holder.

TBR=garykac@chromium.org,rdsmith@chromium.org,stuartmorgan@chromium.org,gbillock@chromium.org

Review URL: https://codereview.chromium.org/877993003

Cr-Commit-Position: refs/heads/master@{#315372}

BUG=413783

Review URL: https://codereview.chromium.org/887843003

Cr-Commit-Position: refs/heads/master@{#315371}

This is required to re-land CL 26009004 in webrtc (was reverted
due to compiler errors while rolling webrtc into chromium):
https://webrtc-codereview.appspot.com/26009004/

BUG=https://code.google.com/p/webrtc/issues/detail?id=3976

Review URL: https://codereview.chromium.org/896393005

Cr-Commit-Position: refs/heads/master@{#315370}

1. This allows other constraints about prefetched resources to be applied.
2. Together with another CL https://codereview.chromium.org/793823002/, the prefetched resources can temporarily bypass revalidation.

BUG=405690, 406209

Review URL: https://codereview.chromium.org/891343007

Cr-Commit-Position: refs/heads/master@{#315369}

This patch rolls trace-viewer to:
https://github.com/google/trace-viewer/commit/a9802a1384185f9c7ee250ce67d3d24d07b11141

Review URL: https://codereview.chromium.org/910773003

Cr-Commit-Position: refs/heads/master@{#315368}

Original: http://crrev.com/889723002
Reverted: http://crrev.com/895373002

This version restores a missing @VisibleForTests annotation
to GoogleServicesNotificationController.getInstance(). Its
removal caused the method to be pruned on build, breaking
downstream tests that depended on it.

GoogleServicesNotificationController is modified to work
the way most other singleton classes work, so that it is
not necessary to pass into the SyncNotificationController
constructor (and thus, into SyncController in the future).

SyncController gets a temporary setter method so that SNC
can be constructed downstream and passed in until all its
dependencies are upstream.

BUG=428882

Review URL: https://codereview.chromium.org/895283003

Cr-Commit-Position: refs/heads/master@{#315367}

BUG=none
TEST=compiles

Review URL: https://codereview.chromium.org/907523003

Cr-Commit-Position: refs/heads/master@{#315366}

TBR=thestig@chromium.org
NOTRY=true
BUG=442785

Review URL: https://codereview.chromium.org/904443004

Cr-Commit-Position: refs/heads/master@{#315365}

This regressed in https://codereview.chromium.org/897563002/.

I also fixed the ChromeContentUtilityClient::PreSandboxStartup() code as well since it was incorrect.

BUG=456623

Review URL: https://codereview.chromium.org/906923003

Cr-Commit-Position: refs/heads/master@{#315364}

BUG=453880

Review URL: https://codereview.chromium.org/905723004

Cr-Commit-Position: refs/heads/master@{#315363}

Original issue: https://codereview.chromium.org/880063002
Previous reland attempt: https://codereview.chromium.org/895623003/

This cl changes the way in which ChromeVox initially verbalizes a page which led to flakeyness in
  chromevox_tests --gtest_filter=BackgroundTest.*

To attempt and address the flakeyness,
- chromevox_e2e_test_base.js and chromevox_next_e2e_test_base.js now explicitly separates listening to page load events from the chrome.automation tree and the chrome.tabs APIs.
They are named: runWithLoadedTree, runWithLoadedTab, and runWithTab. The last method allows a caller to create a tab, but not wait for it to load (based on the tab status being 'complete').
- each test in background_test now no longer waits for ChromeVox to speak 'start' (which formerly could have been triggered by a loadComplete automation event).
- tab creation is still done via the tabs API, but listening for load is done via chrome.automation. (i.e. add a listener for an automation tree to load completely, then create a tab).

It is not known at this time if this cl will continue to cause flakes in the above tests. Try bots and local runs yielded no flakeyness. The next step may be to introduce logging to identify the source of the flakeyness.

TBR=tfarina

Review URL: https://codereview.chromium.org/890013006

Cr-Commit-Position: refs/heads/master@{#315362}

Chrome doesn't use overlay scrollbars. The only thing that we do with this
library is attempt to initialize it (as part of gtk2+ init) which occasionally
tries to read a missing settings schema and crashes.

BUG=456205
TBR=glider@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/905703003

Cr-Commit-Position: refs/heads/master@{#315361}

Unless we reconstruct an already sanitized referrer from the referrer
url and the referrer policy, we should always sanitize it before
constructing a request from it.

BUG=454621,422871
R=mkwst@chromium.org,jam@chromium.org

Review URL: https://codereview.chromium.org/898613004

Cr-Commit-Position: refs/heads/master@{#315360}