GitLab

Since scaling a zero-sized image can result in dealing with NaNs (in case
that we're scaling it using an inf, which is fairly possible), we should
avoid doing that. This CL makes sure that we do.

As to why we're dealing with zero sized images to begin with, I've opened
crbug.com/522637, but it seems like something that can happen regardless,
when ImageResource doesn't have an Image.

BUG=521954

Review URL: https://codereview.chromium.org/1299493005

git-svn-id: svn://svn.chromium.org/blink/trunk@200879 bbb929c8-8fbe-4397-9dbb-9b2b20218538

At one point the caret was placed at an incorrect position for complex
scripts. This was caused by advances and/or bounds calculation was
incorrect when white-space:pre is applied to a text run of complex
scripts.

It was fixed in later patches. This CL adds a test for the failure to
prevent it happening again.

BUG=513994

Review URL: https://codereview.chromium.org/1303673005

git-svn-id: svn://svn.chromium.org/blink/trunk@200878 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch gets rid of used functions |kit(TextAffinity)| and
|core(TextAffinity)| in "TextAffinity.h" to cleanup code for improving code
health.

This is follow-up of a review comment in http://crrev.com/1293593004.

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

Review URL: https://codereview.chromium.org/1291283008

git-svn-id: svn://svn.chromium.org/blink/trunk@200877 bbb929c8-8fbe-4397-9dbb-9b2b20218538

https://codereview.chromium.org/1287283002/ unintentionally removed code
to consider estimatedRemovalRate when judging a threshold for page navigation GCs.
This CL recovers the code.

BUG=474470

Review URL: https://codereview.chromium.org/1301893003

git-svn-id: svn://svn.chromium.org/blink/trunk@200876 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Adds a layout test to ensure that client ordering is done properly.

Spec: https://slightlyoff.github.io/ServiceWorker/spec/service_worker/#clients-matchall-method

1. (Chromium) https://codereview.chromium.org/1285373002/
2. (Blink) This CL.

BUG=461411

Review URL: https://codereview.chromium.org/1286123004

git-svn-id: svn://svn.chromium.org/blink/trunk@200875 bbb929c8-8fbe-4397-9dbb-9b2b20218538

JavaScript Symbols are a new (in ES2015) primitive type. This caused
ScriptValueSerializer to go off the rails when it had exhausted its known
list of primitives and then assumed anything else might be an object.

Now the code checks for Object-ness before assuming it. Also refactored
the if/else-if cascade to group all Object handling together, and prune
dead branches.

BUG=522227

Review URL: https://codereview.chromium.org/1297223004

git-svn-id: svn://svn.chromium.org/blink/trunk@200874 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=501896

Review URL: https://codereview.chromium.org/1293793008

git-svn-id: svn://svn.chromium.org/blink/trunk@200872 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=
BUG=498539

Review URL: https://codereview.chromium.org/1306523002

git-svn-id: svn://svn.chromium.org/blink/trunk@200871 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=522674

Review URL: https://codereview.chromium.org/1286383007

git-svn-id: svn://svn.chromium.org/blink/trunk@200870 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=522274

Review URL: https://codereview.chromium.org/1300093002

git-svn-id: svn://svn.chromium.org/blink/trunk@200869 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=502118, 503170

Review URL: https://codereview.chromium.org/1284413004

git-svn-id: svn://svn.chromium.org/blink/trunk@200868 bbb929c8-8fbe-4397-9dbb-9b2b20218538

NOTRY=true
BUG=501896

Review URL: https://codereview.chromium.org/1290883006

git-svn-id: svn://svn.chromium.org/blink/trunk@200867 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=514595

Review URL: https://codereview.chromium.org/1297323002

git-svn-id: svn://svn.chromium.org/blink/trunk@200866 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This should make creating a subclass in a test easier, and hopefully
combined with FrameTestHelpers::WebViewHelper and some cleverness will
allow writing FOUC tests.

BUG=521692

Review URL: https://codereview.chromium.org/1287403005

git-svn-id: svn://svn.chromium.org/blink/trunk@200865 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=
NOTRY=

Review URL: https://codereview.chromium.org/1300253002

git-svn-id: svn://svn.chromium.org/blink/trunk@200864 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/1297903005

git-svn-id: svn://svn.chromium.org/blink/trunk@200863 bbb929c8-8fbe-4397-9dbb-9b2b20218538

filterNeedsPaintInvalidation does the same thing and makes the code more
clear. This also centralizes the callers of scheduleSVGFilterLayerUpdateHack.

Review URL: https://codereview.chromium.org/1292213004

git-svn-id: svn://svn.chromium.org/blink/trunk@200862 bbb929c8-8fbe-4397-9dbb-9b2b20218538

HTMLPluginElement::loadPlugin was using scheduleSVGFilterLayerUpdateHack()
to update the compositing state of the LayoutEmbeddedObject. Instead we
can use the setNeedsCompositingUpdate() API which will cause only a
compositing update and skip the wasteful style recalc on the <embed> or
<object>.

I also moved the call to updateSelfPaintingLayer into
setNeedsCompositingUpdate since all callers of it actually want to do
that too.

After this the only user of scheduleSVGFilterLayerUpdateHack is actually
SVG filters.

R=ojan@chromium.org

Review URL: https://codereview.chromium.org/1293533005

git-svn-id: svn://svn.chromium.org/blink/trunk@200861 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Gardening.

NOTRY=true
TBR=ojan@chromium.org,pfeldman@chromium.org,joelo@chromium.org
BUG=521764,521730

Review URL: https://codereview.chromium.org/1296223004

git-svn-id: svn://svn.chromium.org/blink/trunk@200860 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I manually verified this giant pile of tests are okay.
Mostly, the results change from blending border colors at the
corners and from a skia change to how images are scaled. Those
changes happened during the dark times when the 10.10 bot
wasn't processing rebaselines.

BUG=508724,509025
TBR=joelo

Review URL: https://codereview.chromium.org/1286043005

git-svn-id: svn://svn.chromium.org/blink/trunk@200859 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Follow-up to codereview.chromium.org/1286043005

BUG=509025
TBR=joelo

Review URL: https://codereview.chromium.org/1294113006

git-svn-id: svn://svn.chromium.org/blink/trunk@200858 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=510345

Review URL: https://codereview.chromium.org/1291733003

git-svn-id: svn://svn.chromium.org/blink/trunk@200857 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Verified all differences are expected.

BUG=509025
TBR=joelo

Review URL: https://codereview.chromium.org/1287423006

git-svn-id: svn://svn.chromium.org/blink/trunk@200856 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The mechanism that Blink was using to determine OSX version number was not
reliable. This has since been fixed.

> I landed a CL (https://codereview.chromium.org/1276003003/) which fixes system
> font problems on OSX 10.9 and OSX 10.10. The CL also fixed MatchNSFontFamily()
> to correctly return the system font on 10.11, but this causes glyph rendering
> problems. In the short term, I'm adding logic to MatchNSFontFamily() to return
> nil on 10.11, which was the original behavior before I landed that CL.
>
> BUG=521034
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=200618

BUG=521034
TBR=tkent@chromium.org, keishi@chromium.org

Review URL: https://codereview.chromium.org/1299713004

git-svn-id: svn://svn.chromium.org/blink/trunk@200855 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=522674

Review URL: https://codereview.chromium.org/1305433004

git-svn-id: svn://svn.chromium.org/blink/trunk@200854 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Like BeginTransform3DDisplayItem, these can also create transforms that
we need to know about. The logic is fairly straightforward and mostly shared
with 3D transforms.

BUG=460986

Review URL: https://codereview.chromium.org/1295403003

git-svn-id: svn://svn.chromium.org/blink/trunk@200853 bbb929c8-8fbe-4397-9dbb-9b2b20218538

UMA metrics indicate we have great validity. Enable the hit test cache
in release by default. Keep the validity checks around in ASSERTs
are enabled.

BUG=398920

Review URL: https://codereview.chromium.org/1294913004

git-svn-id: svn://svn.chromium.org/blink/trunk@200852 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This code is not implemented nor called. Discovered while digging into a
selection issue.

Review URL: https://codereview.chromium.org/1291253003

git-svn-id: svn://svn.chromium.org/blink/trunk@200851 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Extract BoxPainter::calculateBackgroundImageGeometry implementation
and related functions to the BackgroundImageGeometry class. This
enables shrinking the API surface of BackgroundImageGeometry and eases
further code health improvements.

BUG=521481

Review URL: https://codereview.chromium.org/1300103003

git-svn-id: svn://svn.chromium.org/blink/trunk@200850 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Repaint test expecations changed due to SlimmingPaint work, but
it happened during the window where Yosemite results weren't
being updated.

BUG=509025
TBR=joelo

Review URL: https://codereview.chromium.org/1296253004

git-svn-id: svn://svn.chromium.org/blink/trunk@200849 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It's too inaccurate.

BUG=502407
R=jchaffraix@chromium.org,leviw@chromium.org

Review URL: https://codereview.chromium.org/1285273004

git-svn-id: svn://svn.chromium.org/blink/trunk@200848 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Test relied on a timeout to detect iframe load failures, but on slow
bots (debug, msan, etc) the timeout might be hit. Remove the
timeout and disable the test cases that rely on it for now.

Also tweak the test structure slightly to use promise_test to ease
debugging.

BUG=522587
TBR=horo@chromium.org

Review URL: https://codereview.chromium.org/1288553009

git-svn-id: svn://svn.chromium.org/blink/trunk@200847 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This does not yet unprefix fill-available (aka fill),
because the CSSWG is not ready for that yet:
https://lists.w3.org/Archives/Public/www-style/2015Aug/0127.html

Intent to ship: https://groups.google.com/a/chromium.org/d/topic/blink-dev/2ljVa4n_ZX0/discussion

BUG=245157

Review URL: https://codereview.chromium.org/1293793007

git-svn-id: svn://svn.chromium.org/blink/trunk@200846 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Indirections--

BUG=521334

Review URL: https://codereview.chromium.org/1296353004

git-svn-id: svn://svn.chromium.org/blink/trunk@200845 bbb929c8-8fbe-4397-9dbb-9b2b20218538

They flaked awhile back but have since ceased to flake.

BUG=521852
TBR=ojan

Review URL: https://codereview.chromium.org/1301903007

git-svn-id: svn://svn.chromium.org/blink/trunk@200844 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As per webkit-patch print-flaky-tests.

TBR=leviw@chromium.org
BUG=522640, 522641, 522642, 522645, 522646, 522647, 522648, 522649, 522651, 522652, 521091

Review URL: https://codereview.chromium.org/1294933007 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200843 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No intended behavior change.

BUG=488634
TBR=rune@opera.com

Review URL: https://codereview.chromium.org/1294793004 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200842 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The class has m_lineBreakObj, which has already been converted to the new API.
This CL updates the childRemoved method to use box->lineLayoutItem() for
comparisons to m_lineBreakObj and for iterating.

BUG=499321

Review URL: https://codereview.chromium.org/1296803004

git-svn-id: svn://svn.chromium.org/blink/trunk@200841 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When an SVG resource was destroyed, it would notify the
SVGResourcesCache, which would walk the entire cache, notify the layout
object and add a "pending" entry for the id referring to the
corresponding element.
This would mean that every layout object which had any kind of resource
would get a "pending" reference from every id to itself - regardless of
if it ever referred to a resource with the given id. In the particular
test, this resulted in a fairly large (ever-growing) "pending" element
sets because there was persistent resource references in the document.

Fix by only adding "pending" entries for the current clients of the
resource that's being destroyed.
SVGResourcesCache::resourceDestroyed is removed in favor of new method
detachAllClients() in LayoutSVGResourceContainer. The part that
unregistered the resource itself as a client is removed in favor of the
pre-existing call to clientDestroyed() already existing in
LayoutSVGModelObject::willBeDestroyed (delegated to from the resource.)
SVGResources::resourceDestroyed is changed to not call
removeAllClientsFromCache() on the resource being passed - this is
instead done once after having cleared the references in all the
clients.

With this change, the "cycle time" of the test in the bug changes from
linearly increasing to constant.

BUG=521334

Review URL: https://codereview.chromium.org/1302713003

git-svn-id: svn://svn.chromium.org/blink/trunk@200840 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The newly added test VersionUtilMac.AppKitVersions had a logic error. It tried
to assert that IsOSMavericksOrEarlier() evalutes to false on 10.8 and 10.9.

Original issue's description:
> Fix the mechanism by which Blink determines OSX version number.
>
> Previously, Blink was checking NSAppKitVersionNumber against a compile time
> constant. This mechanism doesn't work (see the bug for more details). This CL
> converts the logic to use the same mechanism that Chromium uses.
>
> BUG=521034
>
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=200708

TBR=thakis@chromium.org
BUG=521034

Review URL: https://codereview.chromium.org/1288823004

git-svn-id: svn://svn.chromium.org/blink/trunk@200839 bbb929c8-8fbe-4397-9dbb-9b2b20218538