GitLab

Revert of Revert of Never send a referrer header if the referrer is not a web url (https://codereview.chromium.org/133863003/)

Reason for revert:
Fixed the browser test meanwhile.

Original issue's description:
> Revert of Never send a referrer header if the referrer is not a web url (https://codereview.chromium.org/136583002/)
> 
> Reason for revert:
> This is the probable cause for the failing browser test CorrectReferrer. See:
> 
> http://build.chromium.org/p/chromium.webkit/builders/Linux%20Tests%20%28dbg%29/builds/1302/steps/browser_tests/logs/CorrectReferrer
> 
> http://build.chromium.org/p/chromium.webkit/builders/Mac10.8%20Tests/builds/5398/steps/browser_tests/logs/CorrectReferrer
> 
> http://build.chromium.org/p/chromium.webkit/builders/Win7%20Tests/builds/543/steps/browser_tests/logs/CorrectReferrer
> 
> Original issue's description:
> > Never send a referrer header if the referrer is not a web url
> > 
> > BUG=none
> > R=abarth@chromium.org
> > 
> > Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=165085
> 
> TBR=abarth@chromium.org,jochen@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> BUG=none
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=165090

TBR=abarth@chromium.org,tdanderson@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=none

Review URL: https://codereview.chromium.org/130173004

git-svn-id: svn://svn.chromium.org/blink/trunk@165140 bbb929c8-8fbe-4397-9dbb-9b2b20218538

(1) StyleEngine has two hashmaps, one is a mapping from css text to StyleSheetContents and the other is a mapping from StyleSheetContents to css text. When destroying StyleSheetContents, using the above mapping to remove css text.
The hashmaps are only for non-quirks document. Because non-quirks document cannot share styles with quirks document.
 
(2) made StyleSheetContents::checkLoaded to invoke new clients' sheetLoaded. To do so,
  (2-1) moved m_loadCompleted flag from StyleSheetContents to CSSStyleSheet,
  (2-2) added more "protect" to StyleSheetContents::checkLoaded to avoid CSSStyleSheet (i.e. client) and ownerNode being deleted by scripts, and
  (2-3) modified StyleSheetContents::loadCompleted() to see whether all clients completed loading.

(3) LinkStyle::setCSSStyleSheet should use StyleSheetContents' checkLoaded. Because now checkLoaded invokes only new clients' sheetLoaded. We don't need to directly invoke LinkStyle::sheetLoaded (this breaks m_loadCompleted flag and causes assertion failure). 

BUG=308781
TEST=all existing tests covers, so the tests should pass.

Review URL: https://codereview.chromium.org/28553005

git-svn-id: svn://svn.chromium.org/blink/trunk@165139 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=224786
TEST=none

Review URL: https://codereview.chromium.org/132823005

git-svn-id: svn://svn.chromium.org/blink/trunk@165137 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Mock temp file implementation is used in all inspector tests instead of the real implementation based on temporary filesystem which (unlike temp fs in regular DevTools window) has limited quota which is often exceeded in the tests.

This should fix flakiness in the following tests:
inspector/profiler/heap-snapshot-loader.html
inspector/profiler/cpu-profiler-save-load.html

Also a bug in BufferedTempFileWriter.close was fixed.

TempFile.getFile was replaced with writeToOutputSteam which operates on types defined in DevTools and simplifies mocks.

BUG=330617,333860
R=alph@chromium.org

Review URL: https://codereview.chromium.org/132993009

git-svn-id: svn://svn.chromium.org/blink/trunk@165136 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Now that DevTools caches CPU and heap profiles in temporary files we need to be careful to remove these files. The files are deleted in case a profile is removed as a result of user action and on DevTools shutdown.

To make sure we don't leave any garbage the temproary storage is also cleared on first attempt to create a new temp file. A SharedWorker is used to synchronize between DevTools front-ends when clearing temp files to avoid removing temp files that might be used by another front-end.

BUG=327298,324769
R=alph@chromium.org, pfeldman@chromium.org

Review URL: https://codereview.chromium.org/136793008

git-svn-id: svn://svn.chromium.org/blink/trunk@165135 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch moves inflate() to be a private internal detail of
FastTextAutosizer and relies on beginLayout to call inflate. This
patch also has a minor cleanup removing an untrue FIXME in inflate().

No new tests as this is just a refactoring.

BUG=302005

Review URL: https://codereview.chromium.org/138143004

git-svn-id: svn://svn.chromium.org/blink/trunk@165134 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also remove instances of String::append where StringBuilder can not be used.

Visited top ten sites and no more hits to String::append(WTF::String&).

Review URL: https://codereview.chromium.org/137293005

git-svn-id: svn://svn.chromium.org/blink/trunk@165133 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/139003002

git-svn-id: svn://svn.chromium.org/blink/trunk@165132 bbb929c8-8fbe-4397-9dbb-9b2b20218538

chromium side: https://codereview.chromium.org/133493003/

BUG=334263

Review URL: https://codereview.chromium.org/138303002

git-svn-id: svn://svn.chromium.org/blink/trunk@165131 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Our compositor knows how to do unions and deal with repaints for us, we
don't need to be deferring the repaints until after recalc style or
layout like this code was trying to do.

R=jamesr@chromium.org,eseidel@chromium.org

Review URL: https://codereview.chromium.org/134643008

git-svn-id: svn://svn.chromium.org/blink/trunk@165130 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No one calls maxLength for HTMLTextFormControlElement though
HTMLInputElement::maxLength and HTMLTextAreaElement::maxLength are used. We can
remove HTMLTextFormControlElement::maxLength, and stop overriding it in
HTMLInputElement and HTMLTextAreaElement.

BUG=none
TEST=none; no behavior changes.

Review URL: https://codereview.chromium.org/137783010

git-svn-id: svn://svn.chromium.org/blink/trunk@165129 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts commit b5c27b516609da2e9fa13187f6eb2a31e50f529c.

BUG=
TBR=pfeldman@chromium.org

Review URL: https://codereview.chromium.org/130483006

git-svn-id: svn://svn.chromium.org/blink/trunk@165128 bbb929c8-8fbe-4397-9dbb-9b2b20218538

r151702 added limit for pruning in ImageDecoded::clearCacheExceptFrame,
which would disallow pruning single-frame/static images altogether.

Allow such images to be pruned in their entirety by dropping the
condition for <= 1 frames - the for-loop will handle the kNotFound case.

(Also related is r156651, which re-added the destroyAll flag.)

Modify the ImageFrameGeneratorTest.frameHasAlpha unittest to use frame 0
instead of frame 1, to avoid clearing frame 1 (MockImageDecoder
maps all indices to frame 0.)

R=hclam@chromium.org,pkasting@chromium.org,senorblanco@chromium.org
BUG=302851

Review URL: https://codereview.chromium.org/136503002

git-svn-id: svn://svn.chromium.org/blink/trunk@165127 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=305105

Review URL: https://codereview.chromium.org/139033002

git-svn-id: svn://svn.chromium.org/blink/trunk@165126 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Change RenderText::computePreferredLogicalWidths to split "words" (i.e.
strings without spaces) at text direction boundaries to ensure accurate
measurements as text direction can effect text metrics.

BUG=333004
R=leviw@chromium.org
TEST=fast/text/international/inline-block-with-mixed-direction-words.html

Review URL: https://codereview.chromium.org/137723002

git-svn-id: svn://svn.chromium.org/blink/trunk@165125 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none
TEST=none; no behavior changes.

Review URL: https://codereview.chromium.org/136193008

git-svn-id: svn://svn.chromium.org/blink/trunk@165124 bbb929c8-8fbe-4397-9dbb-9b2b20218538

They are unused.
Note: We just make HTMLAnchorElement::target non-virtual because it is used
twice in HTMLAnchorElement.cpp.

BUG=none
TEST=none; no behavior changes.

Review URL: https://codereview.chromium.org/136033003

git-svn-id: svn://svn.chromium.org/blink/trunk@165123 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL supplies the necessary changes to Blink to support
the inner/outer viewport model for pinch-zoom and fixed-
position elements. The specification for these changes is
contained in the document "Layer-based Solution for Pinch
Zoom / Fixed Position".

Specifically, it provides the Blink-side infrastructure to
move scroll layers & scrollbars to using a clip-layer to
specify the max scroll offset and scrolbar positioning, as
opposed to pushing a max_scroll_offset from Blink. This
entails a number of changes to the Web* interface classes.

This CL supports the CC-side cl 
https://codereview.chromium.org/23983047/, and must land 
before that CL.

This code has previous review history at 
https://codereview.chromium.org/24302010/.

Review URL: https://codereview.chromium.org/138453004

git-svn-id: svn://svn.chromium.org/blink/trunk@165122 bbb929c8-8fbe-4397-9dbb-9b2b20218538

move recordsView construction code into a private method

Drive by fix: remove _popoverHelper and related code. We don't use it any more.

BUG=
R=aandrey@chromium.org, yurys@chromium.org

Review URL: https://codereview.chromium.org/137853005

git-svn-id: svn://svn.chromium.org/blink/trunk@165121 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Copy WebFrameImpl.{cpp,h} to TextFinder.{cpp,h} (https://codereview.chromium.org/119113003/)

Reason for revert:
The follow-up patch was reverted.

I guess it would be easier to do the copy and refactoring in one CL, so we don't end up with a copy of WebFrameImpl.* that gets stale before the refactoring is finished.

Original issue's description:
> Copy WebFrameImpl.{cpp,h} to TextFinder.{cpp,h}
> 
> This patch copies Source/web/WebFrameImpl.{cpp,h} into Source/web/TextFinder.{cpp,h} for decoupling WebFrameImpl and text finder.
> 
> Another patch will complete decoupling.
> 
> BUG=
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=164490

TBR=tkent@chromium.org,iceman@yandex-team.ru
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/139383002

git-svn-id: svn://svn.chromium.org/blink/trunk@165120 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Neither of these methods appear to be getting called.

BUG=260898

Review URL: https://codereview.chromium.org/138713002

git-svn-id: svn://svn.chromium.org/blink/trunk@165119 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove dead code from DocumentOrderedMap:
- clear() method is never used
- checkConsistency() is declared but unimplemented
- "wtf/HashCountedSet.h" include is no longer needed

R=esprehn, adamk

Review URL: https://codereview.chromium.org/138573003

git-svn-id: svn://svn.chromium.org/blink/trunk@165118 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It causes performance problems. See
https://code.google.com/p/chromium/issues/detail?id=331076

Micro-benchmarks like http://jsperf.com/setting-document-title regress
about one magnitude.

We need to figure out whether the spec should change or if the CL
could be written in a less performance-intrusive way, or if it's a
performance penalty we accept. So revert for now.

The follow-up "Inhibit title updates when removing children before
setting new value" is reverted too since it has no value without the
first patch.

Reverted CLs:

https://codereview.chromium.org/128603002
https://codereview.chromium.org/107513013

BUG=331076

Review URL: https://codereview.chromium.org/138213005

git-svn-id: svn://svn.chromium.org/blink/trunk@165117 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The same functionality is already in SVGGraphicsElement.

Review URL: https://codereview.chromium.org/138923003

git-svn-id: svn://svn.chromium.org/blink/trunk@165116 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Break out the handling of GC_M_MASK (multiple uchar category) into
separate inline function as a part of the effort to clean up the
logic to allow proper handling of script_common.

BUG=311372
R=leviw@chromium.org

Review URL: https://codereview.chromium.org/138553003

git-svn-id: svn://svn.chromium.org/blink/trunk@165115 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also devirtualize it along with dispatchInputEvent.

Based on http://trac.webkit.org/changeset/150805 by Andreas Kling.

Review URL: https://codereview.chromium.org/138003005

git-svn-id: svn://svn.chromium.org/blink/trunk@165114 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL removes the followings:
 - AXObject::selectedText
 - AXRenderObject::selectedText, which is an override of AXObject::selectedText
 - AXRenderObject::stringForRange and HTMLTextFormControlElement::selectedText,
   which are used only by AXRenderObject::selectedText.

BUG=none
TEST=none; no behavior changes

Review URL: https://codereview.chromium.org/137783006

git-svn-id: svn://svn.chromium.org/blink/trunk@165113 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We hold a small ring-list cache of recently-emptied pages, and only free the
oldest entry in the ring list when the cache is full. This makes us hold on to
a little bit of unused memory, but in return we do now eventually free the last
slot span in any given bucket, which would have been held indefinitely before.

This patch is a sufficient generic replacement for the bandaid added in
https://code.google.com/p/chromium/issues/detail?id=330798; accordingly, the
bandaid is removed. In fact, we now seem to be +4% faster than _before_ the
original regression.

BUG=332282

Review URL: https://codereview.chromium.org/133863006

git-svn-id: svn://svn.chromium.org/blink/trunk@165112 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Make generated Callback bindings classes FINAL and add OVERRIDE to their virtual
methods.

R=haraken
BUG=333316

Review URL: https://codereview.chromium.org/138613002

git-svn-id: svn://svn.chromium.org/blink/trunk@165111 bbb929c8-8fbe-4397-9dbb-9b2b20218538

period simpler.

We will need to have Members and Persistents instead of RefPtrs.
On RefPtrs the methods that correspond to raw and clear are called
get and release. Unify on the RefPtr methods to make the transition
period code simpler.

Fix comparison operators that were unused and did not work because
they accessed private members.

R=haraken@chromium.org, wibling@chromium.org, zerny@chromium.org, erik.corry@gmail.com
BUG=

Review URL: https://codereview.chromium.org/138203002

git-svn-id: svn://svn.chromium.org/blink/trunk@165110 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The TestGCScope only needs to stop and resume threads. It should not
change the state of the heap. The old code made the heap consistent
for scanning which cleared all freelists and never rebuild them.

R=haraken@chromium.org, erik.corry@gmail.com, wibling@chromium.org
BUG=

Review URL: https://codereview.chromium.org/138273003

git-svn-id: svn://svn.chromium.org/blink/trunk@165109 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=239771
R=haraken
NOTRY=true

Review URL: https://codereview.chromium.org/131183002

git-svn-id: svn://svn.chromium.org/blink/trunk@165107 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I'm seeing this taking up a big portion of FrameView::layout in my trace.

Review URL: https://codereview.chromium.org/138823003

git-svn-id: svn://svn.chromium.org/blink/trunk@165106 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As a part of RenderBlockLineLayout.cpp refactoring, I'm moving out LineLayoutState into a separate
file. I'm doing the refactoring step by step, the progress is tracked under 316347 tracking issue.

BUG=334363

Review URL: https://codereview.chromium.org/138773002

git-svn-id: svn://svn.chromium.org/blink/trunk@165105 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Levi is already an OWNER and owns a lot of the code that has now been
moved to platform.

R=eseidel@chromium.org
BUG=

Review URL: https://codereview.chromium.org/139043002

git-svn-id: svn://svn.chromium.org/blink/trunk@165104 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts 164797

R=abarth@chromium.org,eseidel@chromium.org
BUG=329603

Review URL: https://codereview.chromium.org/133273007

git-svn-id: svn://svn.chromium.org/blink/trunk@165102 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The checkForRepaintDuringLayout is needed for both the non-repaintAfterLayout
and the repaintAfterLayout case. So, I've split the method into checkForRepaint
and checkForRepaintAfterLayout which are used as needed.

This change pointed out a few missing LayoutRectRecorder entries that have
been added to fixup LayoutTests.

BUG=310398

Review URL: https://codereview.chromium.org/129853005

git-svn-id: svn://svn.chromium.org/blink/trunk@165101 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Adds layout test to verify that attempt to set
SourceBuffer's append mode fails when enough data has been
appended for the buffer to be in the middle of parsing a
media segment.
Also, removes default WebSourceBuffer::setMode()
implementation now that Chromium WSBI::setMode() has landed.

R=acolwell@chromium.org,jamesr@chromium.org
BUG=249422
TEST=All mediasource layout tests pass locally on Linux

Review URL: https://codereview.chromium.org/136523004

git-svn-id: svn://svn.chromium.org/blink/trunk@165100 bbb929c8-8fbe-4397-9dbb-9b2b20218538

|isSelfCollapsingBlock| should only be called on an object after it has had layout. In order to enforce this with an ASSERT I had to account for a couple of things I hadn't previously considered:

- In a couple of places we look at the previous child sibling and check whether it's selfCollapsing. When we do that we need to be sure that it's a normal flow block we're looking at.

- Placeholder elements are not laid out until the dimensions of their parent text control are known, so they don't get layout until their parent has had layout - this is unique in the layout tree and unfortunately when we call isSelfCollapsingBlock on them we find that they still need layout. Since placeholder elements only exist when they have content we can enforce the rule that they are never self-collapsing.

- When we are in partial layout and have entered layoutBlockChildren to layout one or more children we should return early when we are done. The current behaviour, which ignores the rest of the children and then adjusts the height of the parent block with the current margin information, is unnecessary and likely to be wrong.


BUG=

Review URL: https://codereview.chromium.org/129873004

git-svn-id: svn://svn.chromium.org/blink/trunk@165099 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No change in behavior, as ShadowRoot::baseURL still delegates to its Document.

BUG=311682
R=abarth@chromium.org

Review URL: https://codereview.chromium.org/126443005

git-svn-id: svn://svn.chromium.org/blink/trunk@165098 bbb929c8-8fbe-4397-9dbb-9b2b20218538