Commits · 31d95a457f774834e2cd7b9188446fe13154ad46 · App_Technologies / webkit

10 Nov, 2007 7 commits

* platform/win/fast/dom/Window/window-properties-expected.txt: Update for · 31d95a45

darin@apple.com authored Nov 11, 2007

        fix for http://bugs.webkit.org/show_bug.cgi?id=15922.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

31d95a45

Reviewed by Eric. · d5ab4fa9

zimmermann@webkit.org authored Nov 11, 2007

Fixes: http://bugs.webkit.org/show_bug.cgi?id=15937 (Add glyph-orientation-horizontal/vertical support)
Fixes: http://bugs.webkit.org/show_bug.cgi?id=13971 (text-anchor support on vertical text seems slightly off)

Add glyph-orientation-* support for text & textPath. (90/180/270 and 'auto')
Also fix all vertical text handling - for text & textPath! (especially latin1 characters auto orientation in vertical writing modes)

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

d5ab4fa9

Rubberstamped by Eric. · 436422be

zimmermann@webkit.org authored Nov 11, 2007

Update some SVG pixel test results - which show changes lately -> repainting fixes.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

436422be

JavaScriptCore: · 42667586

darin@apple.com authored Nov 11, 2007

        Reviewed by Sam.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15927
          REGRESSION(r27487): delete a.c followed by __defineGetter__("c", ...) incorrectly deletes another property
          and <rdar://problem/5586384> REGRESSION (r27487): Can't switch out of Edit HTML Source mode on Leopard Wiki

        Test: fast/js/delete-then-put.html

        * kjs/property_map.cpp:
        (KJS::PropertyMap::put): Added a missing "- 1"; code to find an empty slot was not working.
        (KJS::PropertyMap::checkConsistency): Added a missing range check that would have caught this
        problem before.

        - roll out a last-minute change to my evaluateToBoolean patch that was incorrect.

        * kjs/nodes.h: (KJS::ExprStatementNode::ExprStatementNode): Take out call to
        optimizeForUnnecessaryResult, since the result is used in some cases.

LayoutTests:

        Reviewed by Sam.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15927
          delete a.c followed by __defineGetter__("c", ...) incorrectly deletes another property

        * fast/js/delete-then-put-expected.txt: Added.
        * fast/js/delete-then-put.html: Added.
        * fast/js/resources/delete-then-put.js: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

42667586

WebKitTools: · 0be8d5fe

weinig@apple.com authored Nov 11, 2007

        Reviewed by Tim Hatcher.

        Follow up to <rdar://problem/5394877> Safari should not log unsafe JavaScript 
        attempts when in private browsing mode (only an issue if Log JavaScript Exceptions 
        is turned on)

        - Add LayoutTestController.setPrivateBrowsingEnabled(bool) (stub out implementation for windows)

        Added test: http/tests/security/cross-frame-access-private-browsing.html

        * DumpRenderTree/LayoutTestController.cpp:
        (setPrivateBrowsingEnabledCallback):
        (LayoutTestController::staticFunctions):
        * DumpRenderTree/LayoutTestController.h:
        * DumpRenderTree/mac/DumpRenderTree.mm:
        (runTest): Default to private browsing disabled.
        * DumpRenderTree/mac/LayoutTestControllerMac.mm:
        (LayoutTestController::setPrivateBrowsingEnabled):
        * DumpRenderTree/win/LayoutTestControllerWin.cpp:
        (LayoutTestController::setPrivateBrowsingEnabled):

LayoutTests:

        Reviewed by Tim Hatcher.

        Follow up to <rdar://problem/5394877> Safari should not log unsafe JavaScript 
        attempts when in private browsing mode (only an issue if Log JavaScript Exceptions 
        is turned on)

        Test using the new LayoutTestController.setPrivateBrowsingEnabled(bool)

        * http/tests/security/cross-frame-access-private-browsing-expected.txt: Added.
        * http/tests/security/cross-frame-access-private-browsing.html: Added.
        * platform/win/Skipped: Add new test to windows skip list until we have an
        implementation of LayoutTestController.setPrivateBrowsingEnabled(bool) for it.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27676 268f45cc-cd09-0410-ab3c-d52691b4dbfc

0be8d5fe

Reviewed by Darin. · e962e996

ap@webkit.org authored Nov 10, 2007

        http://bugs.webkit.org/show_bug.cgi?id=15922
        Implement more of Mozilla Selection API

        Tests: editing/selection/containsNode.html
               editing/selection/deleteFromDocument.html
               editing/selection/extend.html
               editing/selection/selectAllChildren.html

        * editing/SelectionController.cpp:
        (WebCore::SelectionController::deleteFromDocument):
        (WebCore::SelectionController::containsNode):
        (WebCore::SelectionController::selectAllChildren):
        (WebCore::SelectionController::extend):
        * editing/SelectionController.h:
        Added deleteFromDocument(), containsNode(), and selectAllChildren(). Reimplemented extend(),
        which existed, but didn't match Firefox behavior and wasn't exposed via bindings.
        Removed a comment mentioning removeRange(), as this method makes no sense without multiple
        selection range support.

        * page/DOMSelection.cpp:
        (WebCore::DOMSelection::extend):
        (WebCore::DOMSelection::deleteFromDocument):
        (WebCore::DOMSelection::containsNode):
        (WebCore::DOMSelection::selectAllChildren):
        * page/DOMSelection.h:
        * page/DOMSelection.idl:
        Exposed the new methods.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27666 268f45cc-cd09-0410-ab3c-d52691b4dbfc

e962e996

Reviewed by Darin. · fa296d6a

ap@webkit.org authored Nov 10, 2007

        http://bugs.webkit.org/show_bug.cgi?id=15892
        DOM Range operations are not implemented for ProcessingInstruction nodes

        Test: fast/dom/Range/range-processing-instructions.html

        * dom/Range.cpp:
        (WebCore::Range::processContents): Implemented ProcessingInstruction cases.
        (WebCore::Range::checkNodeWOffset): Removed a FIXME - yes, I think that we are supposed
        to use ProcessingInstruction.data.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

fa296d6a

09 Nov, 2007 4 commits

Reviewed by Sam Weinig. · 5decd396

tristan authored Nov 09, 2007

        * fast/dom/Window/window-resize-nan.html: Removed.
        * fast/dom/Window/window-resize-nan.html-disabled: Copied from fast/dom/Window/window-resize-nan.html.
        
        Disabled nan test until window resize issues can be resolved.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27657 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5decd396

Fix <rdar://5483519 > Pressing Enter on selected buttons should fire onclick · 745cedc5

aroben@apple.com authored Nov 09, 2007

 LayoutTests:

         Add a test for <rdar://5483519> Pressing Enter on selected buttons should fire onclick

         Reviewed by Adele.

         * fast/forms/enter-clicks-buttons-expected.txt: Added.
         * fast/forms/enter-clicks-buttons.html: Added.

 WebCore:

         Fix <rdar://5483519> Pressing Enter on selected buttons should fire onclick

         We now match the behavior of Firefox and IE, which is to always just
         send a click event to the focused button when the Enter key is pressed
         (previously we were submitting forms directly in some cases).

         Reviewed by Adele.

         Test: fast/forms/enter-clicks-buttons.html

         * html/HTMLButtonElement.cpp:
         (WebCore::HTMLButtonElement::defaultEventHandler): Don't do anything
         fancy when Enter is pressed on a <button type=button> -- just send a
         click event like we do for other button types.
         * html/HTMLInputElement.cpp:
         (WebCore::HTMLInputElement::defaultEventHandler): Treat type=button
         the same way we treat type=submit and type=reset: just send a click
         event when Enter is pressed.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

745cedc5

WebCore: · 5a4bdf3e

mitz@apple.com authored Nov 09, 2007

        Reviewed by Antti Koivisto.

        - fix a bug in invisible layer culling: dynamically changing a 
          descendant of an invisible layer to be visible did not work

        Test: fast/layers/layer-content-visibility-change.html

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::setHasVisibleContent): If we got visible content,
        make sure that our stacking context rebuilds its z-order lists to
        include us.

LayoutTests:

        Reviewed by Antti Koivisto.

        - test that dynamically changing a descendant of an invisible layer to
          be visible works

        * fast/layers/layer-content-visibility-change.html: Added.
        * platform/mac/fast/layers/layer-content-visibility-change-expected.checksum: Added.
        * platform/mac/fast/layers/layer-content-visibility-change-expected.png: Added.
        * platform/mac/fast/layers/layer-content-visibility-change-expected.txt: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5a4bdf3e

WebCore: · 58730d7c

bdakin@apple.com authored Nov 09, 2007

        Reviewed by Oliver.

        Fix for <rdar://problem/5586370> CSS Transform - incorrect matrix 
        math leads to crazy problems

        Transform matrices accept the first four parameters as CSS lengths. 
        CSS lengths get mapped into WebCore::Lengths as percents by 
        WebCore::convertToLength(). Percent lengths cannot call value(). It 
        does not yield a correct result and it asserts on Debug builds. 

        * rendering/RenderStyle.h:
        (WebCore::MatrixTransformOperation::apply): Instead of calling 
        value() on the lengths, call calcValue. This fixes the assert and 
        the bad rendering. 

LayoutTests:

        Reviewed by Oliver.

        Test for <rdar://problem/5586370> CSS Transform - incorrect matrix 
        math leads to crazy problems

        * fast/transforms/identity-matrix.html: Added.
        * platform/mac/fast/transforms/identity-matrix-expected.checksum: Added.
        * platform/mac/fast/transforms/identity-matrix-expected.png: Added.
        * platform/mac/fast/transforms/identity-matrix-expected.txt: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

58730d7c

08 Nov, 2007 3 commits

Fix regression caused by earlier bitwise and optimisation. 1 & undefined != 1. · e7f10a33

oliver authored Nov 09, 2007

Reviewed by Maciej.

The implementation of JSImmediate::areBothImmediateNumbers relies on
(JSImmediate::getTag(immediate1) & JSImmediate::getTag(immediate2)) having
a unique result when both immediate values are numbers.

The regression was due to UndefinedType & NumberType returning NumberType (3 & 1).
By swapping the value of NumberType and UndefinedType this ceases to be a problem.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27630 268f45cc-cd09-0410-ab3c-d52691b4dbfc

e7f10a33

WebCore: · 4a05631a

mitz@apple.com authored Nov 08, 2007

        Reviewed by Beth Dakin.

        - fix <rdar://problem/5491922> REGRESSION (Safari 2-3): Flash-based "Cash Optimizer" on etrade.com does not draw completely

        Test: fast/dom/length-attribute-mapping.html

        * dom/StyledElement.cpp:
        (WebCore::StyledElement::addCSSLength): Changed the garbage-stripping
        logic to stop after the first "%" or "*" in the string. This allows for
        "100%25" to be mapped to "100%" like it is in Firefox and WinIE.

LayoutTests:

        Reviewed by Beth Dakin.

        - test for <rdar://problem/5491922> REGRESSION (Safari 2-3): Flash-based "Cash Optimizer" on etrade.com does not draw completely

        * fast/dom/length-attribute-mapping-expected.txt: Added.
        * fast/dom/length-attribute-mapping.html: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27617 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4a05631a

WebCore: · 568b8655

darin@apple.com authored Nov 08, 2007

        Reviewed by Adam.

        - fix <rdar://problem/5552943> accesskey does not focus <button> elements

        Test: fast/forms/access-key.html

        * html/HTMLButtonElement.cpp: (WebCore::HTMLButtonElement::accessKeyAction):
        Added a call to focus.

LayoutTests:

        Reviewed by Adam.

        - test changes for <rdar://problem/5552943> accesskey does not focus <button> elements

        * fast/forms/access-key.html: Check for both focus and click events on all elements, but
        resisted the urge to switch to the American spelling for "focused".

        * fast/forms/access-key-expected.txt: Updated to reflect the bug fix ("1 button focussed"),
        the fact that <input type=button>, <input type=checkbox>, <input type=submit>, and
        <input type=reset> are all both focused and clicked ("2 input type button clicked",
        "3 input type checkbox clicked", "5 input type submit focussed", and
        "6 input type reset focussed"). Unfortunately this now also demonstrates that we don't
        send a click to <input type=text> and Gecko does. And there are still quite a few types
        that are not covered.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27599 268f45cc-cd09-0410-ab3c-d52691b4dbfc

568b8655

07 Nov, 2007 13 commits

WebCore: · 8b549c4c

ddkilzer authored Nov 07, 2007

WebKit confuses width/height for Media Queries device-aspect-ratio evaluation
<http://bugs.webkit.org/show_bug.cgi?id=14893>
<rdar://problem/5380295>

Reviewed by Darin.

Tests: fast/css/device-aspect-ratio.html
fast/css/max-device-aspect-ratio.html
fast/css/min-device-aspect-ratio.html

* css/MediaQueryEvaluator.cpp:
(WebCore::parseAspectRatio): Renamed method parameters from a/b to h/v.
(WebCore::device_aspect_ratioMediaFeatureEval): Renamed local variables from
a/b to h/v. Switched first two arguments of the call to cmpvalue() to fix
the bug.

LayoutTests:

WebKit confuses width/height for Media Queries device-aspect-ratio evaluation
<http://bugs.webkit.org/show_bug.cgi?id=14893>
<rdar://problem/5380295>

Reviewed by Darin.

Each of the following tests creates a <link> element and dynamically sets its
media query based on the aspect ratio (width/height) of the current screen.
When the <link> element is attached to the <head> node, the test expects the
stylesheet to be loaded for the test to pass.

* fast/css/device-aspect-ratio-expected.txt: Added.
* fast/css/device-aspect-ratio.html: Added.

This test sets the device-aspect-ratio to the current size of the screen.
NOTE: This test passed before the fix.

* fast/css/max-device-aspect-ratio-expected.txt: Added.
* fast/css/max-device-aspect-ratio.html: Added.

This test sets the max-device-aspect ratio to 100/1 for landscape monitors or
1/1 for portrait (or square) monitors.
NOTE: This test failed before the fix.

* fast/css/min-device-aspect-ratio-expected.txt: Added.
* fast/css/min-device-aspect-ratio.html: Added.

This test sets the min-device-aspect ratio to 1/1 for landscape monitors or
1/100 for portrait (or square) monitors.
NOTE: This test failed before the fix.

* fast/css/resources/device-aspect-ratio.css: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27581 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8b549c4c

WebCore: · 6201fc44

mitz@apple.com authored Nov 07, 2007

        Reviewed by Dave Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15887
          REGRESSION (r27576): Crash in RenderStyle::affectedByHoverRules clicking link on Digg

        Test: fast/css/display-none-inline-style-change-crash.html

        * dom/Element.cpp:
        (WebCore::Element::recalcStyle): Fixed the crash by null-checking
        the current style and removed other checks that are not strictly
        necessary.

LayoutTests:

        Reviewed by Dave Hyatt.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15887
          REGRESSION (r27576): Crash in RenderStyle::affectedByHoverRules clicking link on Digg

        * fast/css/display-none-inline-style-change-crash-expected.txt: Added.
        * fast/css/display-none-inline-style-change-crash.html: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27579 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6201fc44

WebCore: · 9e8a8657

mitz@apple.com authored Nov 07, 2007

        Reviewed by Darin Adler.

        - fix a bug where CSS rules with :hover in the ancestor chain stopped
          working after changing the inline style declaration of the ancestor

        Test: fast/css/affected-by-hover-after-style-change.html

        * dom/Element.cpp:
        (WebCore::Element::recalcStyle): If we are not forcing style
        recalculation for all descendants, preserve any "affected by
        {hover|active|drag} bits that we may have acquired from them.
        Also renamed _style to currentStyle.

WebKitSite:

        Reviewed by Darin Adler.

        - removed a workaround for the bug fixed in the WebCore part of this
          patch

        * misc/DatabaseExample.html:

LayoutTests:

        Reviewed by Darin Adler.

        - test that CSS rules with :hover in the ancestor chain continue to work
          after changing the inline style declaration of the ancestor

        * fast/css/affected-by-hover-after-style-change.html: Added.
        * platform/mac/fast/css/affected-by-hover-after-style-change-expected.checksum: Added.
        * platform/mac/fast/css/affected-by-hover-after-style-change-expected.png: Added.
        * platform/mac/fast/css/affected-by-hover-after-style-change-expected.txt: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27576 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9e8a8657

JavaScriptCore: · 8a50ec54

ggaren@apple.com authored Nov 07, 2007

        Reviewed by Darin Adler.
        
        Fixed part of http://bugs.webkit.org/show_bug.cgi?id=15861 
        15% of string-validate-input.js is spent compiling the same regular expression.

        Put RegExpImp properties into a static hashtable to avoid a slew of
        PropertyMap churn when creating a RegExpImp.
        
        Factored important bits of regular expression implementation out of
        RegExpImp (the JS object) and into RegExp (the PCRE wrapper class), 
        making RegExp a ref-counted class. (This will help later.)

        Removed PCRE_POSIX support because I didn't quite know how to test it 
        and keep it working with these changes.
        
        1.1% SunSpider speedup. 5.8% speedup on string-validate-input.js.

        * kjs/regexp.h: A few interface changes:
        1. Renamed "subpatterns()" => "numSubpatterns()"
        2. Made flag enumeration private and replaced it with public getters for
        specific flags.
        3. Made RegExp ref-counted so RegExps can be shared by RegExpImps.
        4. Made RegExp take a string of flags instead of an int, eliminating 
        duplicated flag parsing code elsewhere.

        * kjs/regexp_object.cpp:
        (KJS::RegExpProtoFunc::callAsFunction): For RegExp.compile: 
        - Fixed a bug where compile(undefined) would throw an exception. 
        - Removed some now-redundant code.
        - Used RegExp sharing to eliminate an allocation and a bunch of 
        PropertyMap thrash. (Not a big win since compile is a deprecated 
        function. I mainly did this to test the plubming.)

LayoutTests:

        Reviewed by Darin Adler.
        
        Beefed up the RegExp.compile testcase to cover a mistake in the 
        original check-in and a mistake I made while developing my new patch.

        * fast/js/resources/regexp-compile.js:



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@27571 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8a50ec54