GitLab

if the tracing category blink.invalidation is on.
BUG=

Review URL: https://codereview.chromium.org/320553004

git-svn-id: svn://svn.chromium.org/blink/trunk@175831 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Moves some logic from the JS code in to the code that generates
builders.jsonp for steps with test in the name that don't actually
run tests. Also expands this list a bit.

Also, creates a new list for test steps that do run tests, but
don't upload to the dashboard. In a future patch, I'll expand this
list and give it a nice UI in the flakiness dashboard that points
people to a way to get the test step to upload results.

Review URL: https://codereview.chromium.org/309783002

git-svn-id: svn://svn.chromium.org/blink/trunk@175829 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It didn't really make sense to be part of auto-rebaseline except
for the special noneedsrebaseline entries, which I've decided are
not really useful anyways. More importantly, move it up means
that we actually log everything to the server, including the outer
git pull call and it means that we log asynchronously to the
auto-rebaseline command, which should make it a lot faster.

Also, make the git cl upload logging log stderr as well.

BUG=382248

Review URL: https://codereview.chromium.org/327443002

git-svn-id: svn://svn.chromium.org/blink/trunk@175828 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=abarth
BUG=372837

Review URL: https://codereview.chromium.org/311273008

git-svn-id: svn://svn.chromium.org/blink/trunk@175827 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously, StringHasher contained methods for calculating the hash of a pure C
character array without having to pass in the length of the array. This was only
used for AtomicStrings constructed via a string literal. The impl of the specific
hashing function currently used in WTF already checks for null terminated character
array and when the final StringImpl* for the AtomicString is constructed we
do so again.

This patch just computes the length up front and passes that information along
to the hashing function.  Eventually this length is used to construct the
StringImpl without computing again.

BUG=none

Review URL: https://codereview.chromium.org/328453003

git-svn-id: svn://svn.chromium.org/blink/trunk@175826 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=abarth
BUG=382129

Review URL: https://codereview.chromium.org/323853007

git-svn-id: svn://svn.chromium.org/blink/trunk@175825 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We're adding the GN bots to the CQ, so they should be here as well.
Also, update the list to point at the correct set of android compile-only bots.

R=jamesr@chromium.org
BUG=380464

Review URL: https://codereview.chromium.org/319873002

git-svn-id: svn://svn.chromium.org/blink/trunk@175824 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Everything in the code assumes that the interval size will always fit in
an int. However, to make sure there are no negatives, the
RasterShapeInterval constructor takes an unsigned. This clamps that
value to make sure we can't overflow an int later.

Review URL: https://codereview.chromium.org/313393004

git-svn-id: svn://svn.chromium.org/blink/trunk@175823 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When acceleratedCompositingForFixedPositionEnabled is not set
FrameView::scrollContentsFastPath() performs invalidate for every sticky or
fixed layer. Page http://people.mozilla.org/~jorendorff/es6-draft.html#sec-terms-and-definitions-symbol-type 
causes Region::unite a lot. Vector buffer reallocations are hot in profiler.
This patch pre-allocates vector buffers in Region::Shape::shapeOperation() and
moves Vector allocation out of inner loop. This significantly reduces number of
vector buffer reallocs.

BUG=335306

Review URL: https://codereview.chromium.org/183663030

git-svn-id: svn://svn.chromium.org/blink/trunk@175822 bbb929c8-8fbe-4397-9dbb-9b2b20218538

-Pass -f instead of setting EDITOR.
-Put a message at the top of FlakyTests saying it's autogenerated.
-Add a test that goes through the upload codepath.
-Have ExecutiveMock log calls to run_and_throw_if_fail.

Review URL: https://codereview.chromium.org/312103005

git-svn-id: svn://svn.chromium.org/blink/trunk@175821 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Because font is inherited we had bugs due to the text autosizing
multiplier and the font getting out of sync. This patch changes
the text autosizing multiplier to inherit like the font properties.

TEST=display-type-change.html which failed before this patch.
BUG=378959

Review URL: https://codereview.chromium.org/319283002

git-svn-id: svn://svn.chromium.org/blink/trunk@175820 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently Clank could only delete (with the backspace key) nodes that are
visible to the android IME, and since a replaced element, e.g. <img>, is not
visible to the IME, thus it could not be easily deleted using the backspace key.
Another side effect is that if two text nodes are separated by an image element
alone, then they appear as a single composition to the IME.

To fix these issues, this CL exposes a replaced element as a single object
replacement char 0xFFFC to InputMethodController. For more details please refer
to the atatched bug.

BUG=311448

Review URL: https://codereview.chromium.org/297003003

git-svn-id: svn://svn.chromium.org/blink/trunk@175819 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=382589

Review URL: https://codereview.chromium.org/313393005

git-svn-id: svn://svn.chromium.org/blink/trunk@175818 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This test checks that the composited layer tree is updated immediately
after the inner iframe becomes composited due to an accelerated
overflow scrolling layer.

I have confirmed that this test fails without Ojan's change to give
all RenderParts render layers (this CL marks composited iframes as self
painting immediately rather than after a style update allowing the
GraphicsLayerUpdater to stitch in the subframe's graphics layer tree
right away) -- https://codereview.chromium.org/302083002

R=ojan@chromium.org
BUG=379097

Review URL: https://codereview.chromium.org/324733002

git-svn-id: svn://svn.chromium.org/blink/trunk@175817 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We're not actively working on CSSStickyPosition. It's not ready for web
developers to experiment with yet.

R=jamesr@chromium.org

Review URL: https://codereview.chromium.org/320323002

git-svn-id: svn://svn.chromium.org/blink/trunk@175816 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Return early from preload request for main resource
to avoid spinning up ResourceLoader unnecessarily.

Review URL: https://codereview.chromium.org/314423004

git-svn-id: svn://svn.chromium.org/blink/trunk@175815 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/319383002

git-svn-id: svn://svn.chromium.org/blink/trunk@175814 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Automatically generated using:
webkit-patch update-flaky-tests

R=dpranke@chromium.org

Review URL: https://codereview.chromium.org/319383004

git-svn-id: svn://svn.chromium.org/blink/trunk@175813 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Nothing calls this any more after
https://codereview.chromium.org/323633002/

BUG=none

Review URL: https://codereview.chromium.org/322553006

git-svn-id: svn://svn.chromium.org/blink/trunk@175812 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We were just looping over port names before, which will default
us to only use release builds. Also, that's wrong anyways because
there's no guarantee that a port actually runs on a builder yet.
Instead, loop over the builders. We assume throughout the codebase
that these builders upload results to the flakiness dashboard,
so it's correct to use the builders list here.

Review URL: https://codereview.chromium.org/326483002

git-svn-id: svn://svn.chromium.org/blink/trunk@175811 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The way we perform reset (i.e. clear records) was
badly wrong in a couple of places:
 - we generally should only reset the objects that we own,
so TimelineTracingView should have not tried to reset TracingModel;
 - TracingTimelineModel should reset before the record is
started,not after.

BUG=361045

Review URL: https://codereview.chromium.org/326693002

git-svn-id: svn://svn.chromium.org/blink/trunk@175810 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We still won't draw them as events, as we check for non-zero duration, but we'd
be able to process them in the TracingTimelineModel, e.g. to associate with other
events and frames.

BUG=361045

Review URL: https://codereview.chromium.org/323873002

git-svn-id: svn://svn.chromium.org/blink/trunk@175809 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Originally landed as r175749, this had to revert because of a memory
issue; the file arrays were not null terminated. Now using the
explicit length WTFString constructor.

R=jsbell@chromium.org,dcheng@chromium.org
TBR=jochen@chromium.org
BUG=374822

Review URL: https://codereview.chromium.org/325683003

git-svn-id: svn://svn.chromium.org/blink/trunk@175808 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In other words, don't recurse into composited descendants. In all cases, the callers
 of repaintRectIncludingDescendants only want to update repaint rects for child layers that
paint into the same backing.

Review URL: https://codereview.chromium.org/319203003

git-svn-id: svn://svn.chromium.org/blink/trunk@175807 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The failures on Windows Debug is probably due to a timing issue where
the FFT samples were grabbed at an unexpected time.  If the grabbed
time sample isn't a multiple of a full cycle and doesn't start at 0,
the FFT result is affected which can change the peak position and value.
This is aggravated by the Blackman window that is applied which smears the peak out anyway.

This is partly due to the use of one context for all of the
different-sized analyzers.  This CL replaces that with a separate
offline context for each analyzer test that should guarantee that
right time samples are used.

BUG=349523

Review URL: https://codereview.chromium.org/316293002

git-svn-id: svn://svn.chromium.org/blink/trunk@175806 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Should be safe to reland after this build fix:
https://codereview.chromium.org/325483002/

BUG=350394
TBR=peter@chromium.org,dominicc@chromium.org,jochen@chromium.org,michaeln@chromium.org,kinuko@chromium.org

Review URL: https://codereview.chromium.org/318413002

git-svn-id: svn://svn.chromium.org/blink/trunk@175804 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=340522

Review URL: https://codereview.chromium.org/317413003

git-svn-id: svn://svn.chromium.org/blink/trunk@175803 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is the master CL that contains the whole remaining Blink side.  It turns on Blob support in IDB, so it has to wait until the Chromium side has completely landed.

See https://codereview.chromium.org/18023022/ for the Chromium side.
See https://docs.google.com/a/chromium.org/document/d/1Kdr4pcFt4QBDLLQn-fY4kZgw6ptmK23lthGZdQMVh2Y/edit for the big picture document [chromium.org account required].
Read-only view at https://docs.google.com/document/d/1Kdr4pcFt4QBDLLQn-fY4kZgw6ptmK23lthGZdQMVh2Y/pub [no account required].

BUG=108012

Review URL: https://codereview.chromium.org/18590006

git-svn-id: svn://svn.chromium.org/blink/trunk@175802 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This simplifies many boundary time computations all around the code.

BUG=361045

Review URL: https://codereview.chromium.org/322783002

git-svn-id: svn://svn.chromium.org/blink/trunk@175801 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175793

BUG=380686
TBR=fmalita@chromium.org

Review URL: https://codereview.chromium.org/320923002

git-svn-id: svn://svn.chromium.org/blink/trunk@175799 bbb929c8-8fbe-4397-9dbb-9b2b20218538

With the three sides of the patch needed for 343327 landed, these temporary
casts are unnecessary. Removed.

BUG=343327

Review URL: https://codereview.chromium.org/313263005

git-svn-id: svn://svn.chromium.org/blink/trunk@175797 bbb929c8-8fbe-4397-9dbb-9b2b20218538

WebInspector.TimelineModel.Events.RecordingStopped is sufficient.

BUG=361045

Review URL: https://codereview.chromium.org/324683002

git-svn-id: svn://svn.chromium.org/blink/trunk@175796 bbb929c8-8fbe-4397-9dbb-9b2b20218538

NotificationController has maintained NotificationClient with OwnPtr since r173468. And, NotificationClient should not be a null. We can return a reference on NotificationController::clientFrom().

BUG=none

Review URL: https://codereview.chromium.org/268353004

git-svn-id: svn://svn.chromium.org/blink/trunk@175795 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Add Debugger.ExceptionDetails to protocol. Use ExceptionDetails for pass error when exception in snippet (run or compile).

Console show the position of exception when exception happens in snippet.

BUG=346013

Review URL: https://codereview.chromium.org/290633009

git-svn-id: svn://svn.chromium.org/blink/trunk@175794 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=380686
TBR=humper@google.com

Review URL: https://codereview.chromium.org/327513003

git-svn-id: svn://svn.chromium.org/blink/trunk@175793 bbb929c8-8fbe-4397-9dbb-9b2b20218538

multi-event type targets.

This patch performs a significant refactoring of the
DeviceSensorEvent{Controller,Dispatcher} and related classes.
It provides new Base classes to allow simple implementation of
APIs involving multi-event type non-window targets (e.g. Battery
Status API). Additionally it makes the implementation of existing
event-based APIs (device_orientation, device_light) simpler and
more compact.

In particular the following new classes have been added:
* DeviceEventControllerBase and DeviceEventDispatcherBase
(derived from by the Battery Status API).
* DeviceSingleWindowEventController is more specific and
contains functionality necessary for single-event type window
target APIs (used by Device Motion/Orientation and Device Light
APIs).

BUG=122593, 360068

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175661

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175686

Review URL: https://codereview.chromium.org/315573002

git-svn-id: svn://svn.chromium.org/blink/trunk@175792 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/322763002

git-svn-id: svn://svn.chromium.org/blink/trunk@175791 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=361045

Review URL: https://codereview.chromium.org/318343003

git-svn-id: svn://svn.chromium.org/blink/trunk@175790 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=380312
R=yurys, pfeldman@chromium.org, amikhaylova@google.com

Review URL: https://codereview.chromium.org/317703003

git-svn-id: svn://svn.chromium.org/blink/trunk@175789 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Using VisiblePosition has some not required performance implications.

BUG=None

Review URL: https://codereview.chromium.org/307353002

git-svn-id: svn://svn.chromium.org/blink/trunk@175788 bbb929c8-8fbe-4397-9dbb-9b2b20218538