GitLab

Revert of Eagerly dispose of ScheduledActions. (patchset #2 id:20001 of https://codereview.chromium.org/2552673002/ )

Reason for revert:
Speculative revert for reported perf decrease on system_health.memory_mobile, https://crbug.com/672098

Original issue's description:
> Eagerly dispose of ScheduledActions.
>
> The DOMTimer's ScheduledAction hold on to the script source and
> state needed to execute the timer action. Let go of ShceduledAction's
> resource early.
>
> Apart from reducing the lifetime of script source, this is a speculative
> fix for crashes reported in v8::PersistentValueVector::Clear() during
> lazy sweeping of ScheduledAction objects.
>
> R=
> BUG=
>
> Committed: https://crrev.com/11bd50343795ed1dc1977da91e9a1588687522fd
> Cr-Commit-Position: refs/heads/master@{#436298}

TBR=oilpan-reviews@chromium.org,haraken@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=

Review-Url: https://codereview.chromium.org/2568103002
Cr-Commit-Position: refs/heads/master@{#437889}

Revert of Delay Input.dispatchKeyEvent response until after key event ack. (patchset #16 id:300001 of https://codereview.chromium.org/2387353004/ )

Reason for revert:
Tentative revert for:

Failed steps failed chrome_public_test_apk on android failed content_browsertests on android

https://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20%28dbg%29?numbuilds=200

as of this CL (and 2 others which are mere renames + don't touch content so suspecting this one from the 3: https://chromium.googlesource.com/chromium/src/+log/b671a5234f41f1784bd261a7c1eb6010434cd2c2..92aad3c0c15d18263af3e5ce0dad525c0b063297?pretty=fuller&n=10000)

No logs from failures on Android bot unfortunately :-(... Don't understand why it would have passed CQ and not waterfall but still can't see what else in range could be at fault, sorry..

Original issue's description:
> Delay Input.dispatchKeyEvent response until after key event ack.
>
> BUG=chromedriver:1506
>
> Review-Url: https://codereview.chromium.org/2387353004

TBR=tdresser@chromium.org,dtapuska@chromium.org,dgozman@chromium.org,pfeldman@chromium.org,samuong@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=673218, chromedriver:1506

Review-Url: https://codereview.chromium.org/2565313002
Cr-Commit-Position: refs/heads/master@{#437888}

Originally, this was a virtual, but it meant that LocalFrame::domWindow
could not return a LocalDOMWindow, due to the circular dependencies
between the LocalFrame and LocalDOMWindow header. As a workaround,
LocalFrame exposed a special localDOMWindow getter. However, this leads
to code that reads very LocalFrame-centric. Since the ultimate goal of
OOPIF integration is to integrate seamlessly, it would be nice to
remove the "local" prefix from the getter.

Since this function doesn't need to be virtual, just de-virtualize it
and have the getters of derived classes return the appropriate type.

BUG=none

Review-Url: https://codereview.chromium.org/2562323002
Cr-Commit-Position: refs/heads/master@{#437887}

BUG=
TEST=

Review-Url: https://codereview.chromium.org/2562363002
Cr-Commit-Position: refs/heads/master@{#437886}

Before OnWebDataServiceRequestDone is called for a request, the request is marked as being complete by being removed from the WebDataRequestManager's pending_requests_ map.

When a query from AutofillCounter fails, it is handled inside OnWebDataServiceRequestDone. By that time, it has been removed from the pending_requests_ map and is no longer allowed to be canceled. Therefore, make sure that AutofillCounter marks it as being complete so that it doesn't try to cancel it.

BUG=672823

Review-Url: https://codereview.chromium.org/2555373007
Cr-Commit-Position: refs/heads/master@{#437885}

Review-Url: https://codereview.chromium.org/2501113002
Cr-Commit-Position: refs/heads/master@{#437884}

Revert of Mark org.chromium.chrome.browser.payments.PaymentRequestIncompleteContactDetailsTest#testEditIncomp… (patchset #1 id:1 of https://codereview.chromium.org/2566173002/ )

Reason for revert:
The tree is closed:
../../chrome/android/javatests/src/org/chromium/chrome/browser/payments/PaymentRequestIncompleteContactDetailsTest.java:47: error: cannot find symbol
    @FlakyTest
     ^
  symbol:   class FlakyTest
  location: class PaymentRequestIncompleteContactDetailsTest
Note: Some input files additionally use or override a deprecated API.

Original issue's description:
> Mark org.chromium.chrome.browser.payments.PaymentRequestIncompleteContactDetailsTest#testEditIncompleteContactAndCancel as flaky
>
> BUG=673299
> TBR=grt@chromium.org
>
> Committed: https://chromium.googlesource.com/chromium/src/+/257f431c44c930193c082bf608c18282a063785e

TBR=grt@chromium.org,vasilii@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=673299

Review-Url: https://codereview.chromium.org/2574433002
Cr-Commit-Position: refs/heads/master@{#437883}

Recently we discovered, that our dismissed IDs pruning may happen at a
wrong time, when DownloadManager does not have a complete list.
Unfortunately, there is no way to check whether DownloadManager
finished starting up.
Therefore, this CL removes pruning of asset downloads and instead
limits maximal number of stored IDs to 100 (the oldest are removed
once the threshold is reached).

BUG=672758

Review-Url: https://codereview.chromium.org/2562073002
Cr-Commit-Position: refs/heads/master@{#437882}

Mark org.chromium.chrome.browser.payments.PaymentRequestIncompleteContactDetailsTest#testEditIncompleteContactAndCancel as flaky

BUG=673299
TBR=grt@chromium.org

Review-Url: https://codereview.chromium.org/2566173002 .
Cr-Commit-Position: refs/heads/master@{#437881}

The comment specified the purpose and possible values of the --password-store
switch. It was outdated in both respects. Since the proper implementation of
os_crypt on Linux earlier this year, the storage was used not just for
passwords, but also for os_crypt's encryption key.

This is even more problematic because the comment is picked up to generate
http://peter.sh/experiments/chromium-command-line-switches/.

This CL updates the description based on the relevant code in
components/os_crypt/key_storage_util_linux.cc.

The CL does not change the name of the flag on purpose: it is likely contained
in a lot of testing and developers' set-ups, and changing it could have
unwanted side-effects of testing data suddenly piling up in people's
Keyrings/KWallets. Once the alternative backends for Linux will stop being
used for storing passwords (https://crbug.com/571003), and only for the single
encryption key of os_crypt, it will be much safer to rename the switch.

BUG=673121

Review-Url: https://codereview.chromium.org/2563253002
Cr-Commit-Position: refs/heads/master@{#437880}

I marked it as slow in https://codereview.chromium.org/2568793002/ but it didn't help.

BUG=673296
TBR=grt@chromium.org

Review-Url: https://codereview.chromium.org/2565423003 .
Cr-Commit-Position: refs/heads/master@{#437879}

Add a new target for GmsCore integration to let WebView code depend on
GmsCore without depending on //android_webview/android_webview_java.
Note that we don't remove PlatformServiceBridge from the
android_webview_java target yet as that would break downstream.

BUG=657314

Review-Url: https://codereview.chromium.org/2566903003
Cr-Commit-Position: refs/heads/master@{#437878}

```--
Allows for custom comparison of results and exotic navigation setup.

Details
```

--
- New browser tests will be needed to test edge cases. For example verifying behavior during cross tab navigation. Useful for regression with PlzNavigate.
- To be able to add the test cases we need to have some more customization
than what the current test framework provides. Specifically, partial
comparison and new navigation types. Implemented that.

BUG=650253

Review-Url: https://codereview.chromium.org/2561493003
Cr-Commit-Position: refs/heads/master@{#437877}

this adds the user agent to the dump. This allows identifying
the chrome version from the dump.

BUG=

Review-Url: https://codereview.chromium.org/2543493003
Cr-Commit-Position: refs/heads/master@{#437876}

Revert of Disable v8.browsing_desktop & v8.browsing_desktop_ignition on Win Perf bots (patchset #1 id:1 of https://codereview.chromium.org/2401253002/ )

Reason for revert:
Speculatively re-enabling to see if it still fails on the bots.

Original issue's description:
> Disable v8.browsing_desktop & v8.browsing_desktop_ignition on Win Perf bots
>
> TBR=nednguyen@google.com
> BUG=chromium:654209
>
> Committed: https://crrev.com/1a039dd675016be7597d84dcef408e1b786a56d4
> Cr-Commit-Position: refs/heads/master@{#424086}

TBR=zhenw@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=chromium:654209

Review-Url: https://codereview.chromium.org/2562373002
Cr-Commit-Position: refs/heads/master@{#437875}

Also fixes a typo.

This is a follow-up from https://codereview.chromium.org/2391883006

BUG=612339

Review-Url: https://codereview.chromium.org/2564953003
Cr-Commit-Position: refs/heads/master@{#437874}

Also changed the order of the attributes/methods under Shadow DOM to match the spec.

BUG=662005

Review-Url: https://codereview.chromium.org/2562343002
Cr-Commit-Position: refs/heads/master@{#437873}

Keep retries on 5xx errors only for user-triggered, interactive requests.
Other fetches have no immediate priority and their number will
be determined by a variation parameter.

This CL does not reschedule/retry non-interactive requests,
as we figured that this small change improves the current state on its own and rescheduling requires some more discussion.

BUG=668074

Review-Url: https://codereview.chromium.org/2548343002
Cr-Commit-Position: refs/heads/master@{#437872}

This will fix Windows python license check presubmit errors. It
will ensure the presubmit check always sees "LF" line ending.

BUG=669941

Review-Url: https://codereview.chromium.org/2558413002
Cr-Commit-Position: refs/heads/master@{#437871}

We don't require a write barrier here as TraceWrapperMember is used for the
following scenarios:
- Initial initialization: The write barrier will not fire as the parent is
   initially white.
- Wrapping when inserting into a container: The write barrier will fire upon
  establishing the move into the container.
- Assignment to a field: The regular assignment operator will fire the write
  barrier.

Note that support for black allocation would require a barrier here.

BUG=chromium:468240

Review-Url: https://codereview.chromium.org/2564153002
Cr-Commit-Position: refs/heads/master@{#437870}

Revert of perf: Disable browse:news:reddit on Mac (patchset #1 id:1 of https://codereview.chromium.org/2484783002/ )

Reason for revert:
The failure does not reproduce at TOT locally.
Re-enabling to see if it still fails on bots.

Original issue's description:
> perf: Disable browse:news:reddit on Mac
>
> BUG=662959
> TBR=perezju@chromium.org
>
> Committed: https://crrev.com/31caf2b1bb934e61b2b79fda4ebb8a8791546f54
> Cr-Commit-Position: refs/heads/master@{#430320}

TBR=perezju@chromium.org,skyostil@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=662959

Review-Url: https://codereview.chromium.org/2565393002
Cr-Commit-Position: refs/heads/master@{#437869}

Before this CL, we were creating a SupervisedUserResourceThrottle for
each top-level navigation. If the current user isn't supervised, this is
pointless. One side-effect was that we were recording
ManagerUsers.Filter* histograms for all navigations, including for
non-supervised users.

This CL adds an "enabled" field on SupervisedUserURLFilter, which is set
to true iff the user is supervised. This allows us to not create the
resource throttle for regular users.

BUG=671472

Review-Url: https://codereview.chromium.org/2564043002
Cr-Commit-Position: refs/heads/master@{#437868}

- A part of Image, as a kind of a wrapper of blink::Image and
  the entry point of ImageResourceObserver, and
- A part of Resource, because it is a subclass of Resource and highly
  involved into internals of Resource such as state control of Resource.
This complicates the ImageResource, causes dependencies and
layering violation between loading- and image-related things, and
complicates reloading logic.

This CL splits ImageResource into two parts:
ImageResource, the Resource-part:
- A subclass of Resource.
ImageResourceContent, the Image-part:
- Manages blink::Image and ImageResourceObserver, and inherits ImageObserver.
- Has limited access to Resource-part (if any) via ImageResourceInfo.

Most of current users of ImageResource, especially those in core/style, are
migrated to use ImageResourceContent and/or
to inherit ImageResourceObserver instead of ResourceClient,
as they want to access Image, not necessarily resource loading.

Design doc:
https://docs.google.com/document/d/1O-fB83mrE0B_V8gzXNqHgmRLCvstTB4MMi3RnVLr8bE/edit?usp=sharing

In this CL, the Resource-part and the Image-part has 1:1 correspondence that
doesn't change during their lifetime and therefore ImageResource::m_content
is always non-null.
Subsequent CLs will make them detachable.

Changes in most of the files are replacing ImageResource with
ImageResourceContent.
Notable changes are:

Main part of CL: split ImageResource into two parts:
- core/fetch/ImageResource.h
- core/fetch/ImageResource.cpp
- core/fetch/ImageResourceContent.h
- core/fetch/ImageResourceContent.cpp

emulateLoadStartedForInspector() hack is now called via ImageResourceContent:
- core/css/CSSImageValue.cpp

ImageLoader is mostly migrated to ImageResourceContent, except for
imageResourceForImageDocument() that returns ImageResource only for
ImageDocument.
- core/html/HTMLImageElement.h
- core/html/ImageDocument.cpp
- core/loader/ImageLoader.h
- core/loader/ImageLoader.cpp

A reference to Resource from FrameSerializer is replaced with |mimeType|
String and |hasCacheControlNoStoreHeader| boolean in order to migrate to
ImageResourceContent.
- web/WebFrameSerializer.cpp
- core/frame/FrameSerializer.h
- core/frame/FrameSerializer.cpp

BUG=667641

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2469873002
Cr-Commit-Position: refs/heads/master@{#437867}

Applies to:
  * chrome://snippets-internals/
  * chrome://popular-sites-internals/

It's nice to start the user off at zoom 1.0, but there's no good reason
to disallow zooming in further.

BUG=none
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2552883011
Cr-Commit-Position: refs/heads/master@{#437866}

BUG=667207

Review-Url: https://codereview.chromium.org/2562243002
Cr-Commit-Position: refs/heads/master@{#437865}

This is a regression by crbug.com/432795.  Scrollbar thickness is fixed regardless
of zoom level.  So popup width was too narrow in less-than-100% zoom level.

With this CL, popupInternalPaddingEnd() returns zoomed value as ever if zoom level
is 100%+, and returns a value based on actual scrollbar thickness otherwise.

Also, popupIntenalPaddingEnd() respects to the actual scrollbar thickness instead
of returning fixed '18' pixel.  The default scrollbar thickness on Windows is 17.
popupInternalPaddingEnd() returns 1 + <scrollbar thickness>.

We need to update ThemePainterDefault::setupMenuListArrow() so that it can support
variable width of scrollbars.

Summary of behavior changes:

All platforms except Mac:
  menulist box is wider in less-than-100% zoom level.

Windows Aura theme:
  menulist box is not changed in 100%+ zoom level.

Non-Windows Aura theme and Mock theme:
  menulist box  is narrower by 2px because scrollbar thickness is 15px.

BUG=667236
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2560733002
Cr-Commit-Position: refs/heads/master@{#437864}

Renames the MediaStream "label" used e.g. in addStream and
onaddstream to "id" to be more aligned with the specification.

BUG=none

Review-Url: https://codereview.chromium.org/2495983002
Cr-Commit-Position: refs/heads/master@{#437863}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/1ba626d6..35e713ce

Please follow these instructions for assigning/CC'ing issues:
https://github.com/v8/v8/wiki/Triaging%20issues

Please close rolling in case of a roll revert:
https://v8-roll.appspot.com/
This only works with a Google account.

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

TBR=hablich@chromium.org,machenbach@chromium.org,littledan@chromium.org,vogelheim@chromium.org

Review-Url: https://codereview.chromium.org/2569673002
Cr-Commit-Position: refs/heads/master@{#437862}

Revert of PaymentApp: Introduce PaymentAppDatabase class. (patchset #3 id:100001 of https://codereview.chromium.org/2560293002/ )

Reason for revert:
Broke content_unittests:
[ RUN      ] PaymentAppManagerTest.SetAndGetManifest

[libprotobuf ERROR c:\b\c\b\win\src\third_party\protobuf\src\google\protobuf\message_lite.cc:120] Can't parse message of type "content.PaymentAppManifestProto" because it is missing required fields: (cannot determine missing fields for lite message)

c:\b\c\b\win\src\content\browser\payments\payment_app_manager_unittest.cc(145): error: Value of: payments::mojom::PaymentAppManifestError::NONE

  Actual: PaymentAppManifestError::NONE

Expected: read_error

Which is: PaymentAppManifestError::MANIFEST_STORAGE_OPERATION_FAILED

[  FAILED  ] PaymentAppManagerTest.SetAndGetManifest (3 ms)

Original issue's description:
> PaymentApp: Introduce PaymentAppDatabase class.
>
> The class is providing APIs to read/write payment app related data(e.g. manifest
> and associated service worker). It can be shared between payment_app_manager.cc
> and payment_app_context.cc.
>
> The payment_app_manager.cc contains a actual implementation of the
> PaymentAppManager interface[1] used on the page or worker.
> Also, the payment_app_context.cc can be used to query all manifests data[2]
> in order to update payment request UI in Chrome layer.
>
> We can summarize this CL as follows.
>   1. Rename PaymentAppContext in content/browser/payments to
>      PaymentAppContextImpl and then add PaymentAppContext interface in
>      content/public to expose some APIs to Chrome layer.
>   2. Write PaymentAppDatabase class and then move most of logics to access
>      database from PaymentAppManager to the new class.
>   3. Add a interface to query all manifests data in Chrome layer but it's not
>      implemented yet in this CL. It will be implemented in follow-up CL[2].
>
> [1] https://w3c.github.io/webpayments-payment-apps-api/#payment-app-manager
> [2] https://codereview.chromium.org/2556433002
>
> BUG=661608
> TEST=existing unittests
>
> Review-Url: https://codereview.chromium.org/2560293002

TBR=rouslan@chromium.org,avi@chromium.org,tommyt@opera.com,jinho.bang@samsung.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=661608

Review-Url: https://codereview.chromium.org/2565423002
Cr-Commit-Position: refs/heads/master@{#437861}

BUG=673133
TBR=grt@chromium.org

Review-Url: https://codereview.chromium.org/2568803002 .
Cr-Commit-Position: refs/heads/master@{#437860}

No need to export PowerMonitorBroadcastSource.

BUG = 647247

Review-Url: https://codereview.chromium.org/2565293002
Cr-Commit-Position: refs/heads/master@{#437859}

The template buildflag_header allows generating a header file with
defines based on gn variables. This is better than using "defines"
on a target as only the files using the macro will depend on it
instead of all files in the target.

BUG=None

Review-Url: https://codereview.chromium.org/2569663002
Cr-Commit-Position: refs/heads/master@{#437858}

BUG=673296
TBR=grt@chromium.org

Review-Url: https://codereview.chromium.org/2568793002 .
Cr-Commit-Position: refs/heads/master@{#437857}

This is refactoring only, with the exception of VRDisplay.cpp, which
changes from unprefixed to prefixed.

As per the TODO, the idea is to make the default type depend on
RuntimeEnabledFeatures::fullscreenUnprefixedEnabled().

Also make RequestType an enum class while in the area.

BUG=402376

Review-Url: https://codereview.chromium.org/2565203002
Cr-Commit-Position: refs/heads/master@{#437856}

The class is providing APIs to read/write payment app related data(e.g. manifest
and associated service worker). It can be shared between payment_app_manager.cc
and payment_app_context.cc.

The payment_app_manager.cc contains a actual implementation of the
PaymentAppManager interface[1] used on the page or worker.
Also, the payment_app_context.cc can be used to query all manifests data[2]
in order to update payment request UI in Chrome layer.

We can summarize this CL as follows.
  1. Rename PaymentAppContext in content/browser/payments to
     PaymentAppContextImpl and then add PaymentAppContext interface in
     content/public to expose some APIs to Chrome layer.
  2. Write PaymentAppDatabase class and then move most of logics to access
     database from PaymentAppManager to the new class.
  3. Add a interface to query all manifests data in Chrome layer but it's not
     implemented yet in this CL. It will be implemented in follow-up CL[2].

[1] https://w3c.github.io/webpayments-payment-apps-api/#payment-app-manager
[2] https://codereview.chromium.org/2556433002

BUG=661608
TEST=existing unittests

Review-Url: https://codereview.chromium.org/2560293002
Cr-Commit-Position: refs/heads/master@{#437855}

This is mostly about mechanical changes such as:

- Move InputMethodController-related logic inside InputMethodController
  and call it first - such that we can expand the logic there.
- Rename method name for better readability.

Also, there are subtle logical changes:

1) Try finishComposingText() even when the current input type is NONE.
   This should not have any effect.

2) When finishComposingText() returns false, we no longer call
   UpdateTextInputState(). This should not have any effect.

3) When finishComposingText() returns false, we no long call
   UpdateCompositionInfo(). This is used by Android and Mac.
   Previously, there was a chance that we propagate InvalidRange() all the
   way up to the keyboard app. Now ShouldUpdateCompositionInfo() returns
   false to make it clear.

This is originally found in resolving crbug.com/664317, but is also
needed for crbug.com/630137 because if we no longer cancel composition
on selection change, then WebViewTest may fail because it uses
TestWebViewClient which does not implement resetInputMethod().

BUG=664317, 630137

Review-Url: https://codereview.chromium.org/2561543003
Cr-Commit-Position: refs/heads/master@{#437854}

Implemented analogously to font-feature-settings, as their definition in
the spec is quite similar as well [1].

Add an experimental web platform features flag for CSSVariableFonts.

[1] https://drafts.csswg.org/css-fonts-4/#low-level-font-variation-settings-control-the-font-variation-settings-property

TEST=fast/css/font-variation-settings-css-support.html
BUG=669455,669460

Review-Url: https://codereview.chromium.org/2558053002
Cr-Commit-Position: refs/heads/master@{#437853}

I'm replacing DOMWindowProperty with ContextLifecycleObserver.
(ContextLifecycleObserver::contextDestroyed() is called already almost at the same
timing as DOMWindowProperty::frameDestroyed().)

To make the replacement easier, this CL introduces ContextLifecycleObserver::frame().
Also this CL moves ApplicationCache to ContextLifecycleObserver.

BUG=610176

Review-Url: https://codereview.chromium.org/2562303002
Cr-Commit-Position: refs/heads/master@{#437852}

Add the android.permission.INTERNET permission to the standalone
WebView's apk, to allow Services belonging to WebView's package to
connect to the internet.

BUG=673280

Review-Url: https://codereview.chromium.org/2562353002
Cr-Commit-Position: refs/heads/master@{#437851}

Make ReadingListModelStorage inherit from syncer::ModelTypeSyncBridge.
ReadingList require the storage to implement USS so add a direct
inheritance to avoid multiple inheritance.

Mark ReadingListStoreDelegate constructor and destructor protected.
ReadingListStoreDelegate is an interface, mark constructor and destructor
protected.

BUG=673169

Review-Url: https://codereview.chromium.org/2568023002
Cr-Commit-Position: refs/heads/master@{#437850}