WebKit-https.git
13 years ago Finished moving editing/pasteboard/5583362.html to platform/mac (this test...
ap@webkit.org [Wed, 7 Nov 2007 09:52:20 +0000 (09:52 +0000)]
    Finished moving editing/pasteboard/5583362.html to platform/mac (this test was crashing in ObjC test plugin,
        because it couldn't find its resources.

        * editing/resources/plaintext-pasteboard-data.dat: Removed.
        * platform/mac/editing/resources: Added.
        * platform/mac/editing/resources/plaintext-pasteboard-data.dat: Copied from editing/resources/plaintext-pasteboard-data.dat.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27505 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
mitz@apple.com [Wed, 7 Nov 2007 08:11:58 +0000 (08:11 +0000)]
WebCore:

        Reviewed by Brady Eidson.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15877
          REGRESSION: r27486 caused a layout regression at my bank's website

        Test: fast/block/float/overhanging-after-height-decrease-offsets.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::layoutBlock):

LayoutTests:

        Reviewed by Brady Eidson.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15877
          REGRESSION: r27486 caused a layout regression at my bank's website

        * fast/block/float/overhanging-after-height-decrease-offsets.html: Added.
        * platform/mac-leopard/fast/block/float/overhanging-after-height-decrease-offsets-expected.checksum: Added.
        * platform/mac-leopard/fast/block/float/overhanging-after-height-decrease-offsets-expected.png: Added.
        * platform/mac/fast/block/float/overhanging-after-height-decrease-offsets-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27504 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Eric.
mjs [Wed, 7 Nov 2007 07:23:25 +0000 (07:23 +0000)]
    Reviewed by Eric.

        - only collect when the heap is full, unless we have lots of extra cost garbage

        1.1% SunSpider speedup.

        This shouldn't hit memory use much since the extra space in those
        blocks hangs around either way.

        * kjs/collector.cpp:
        (KJS::Collector::heapAllocate):
        (KJS::Collector::collect): Fix logic error that reversed the sense of collect's
        return value.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27503 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Move a Mac-specific test to platform/mac
aroben [Wed, 7 Nov 2007 07:00:14 +0000 (07:00 +0000)]
    Move a Mac-specific test to platform/mac

        * editing/pasteboard/5583362.html: Removed.
        * platform/mac/editing/pasteboard/5583362.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27502 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoAvoid unnecessarily boxing the result from post inc/decrement for 0.3% gain in sunspider
oliver [Wed, 7 Nov 2007 06:03:20 +0000 (06:03 +0000)]
Avoid unnecessarily boxing the result from post inc/decrement for 0.3% gain in sunspider

Reviewed by Maciej

We now convert the common 'for (...; ...; <var>++) ...' to the semantically identical
'for (...; ...; ++<var>) ...'.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27501 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Eric Seidel <eric@webkit.org>
eric@webkit.org [Wed, 7 Nov 2007 05:41:21 +0000 (05:41 +0000)]
2007-11-06  Eric Seidel  <eric@webkit.org>

        * Scripts/build-testkjs: build fix... too many $$

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27500 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
bdakin [Wed, 7 Nov 2007 05:31:10 +0000 (05:31 +0000)]
WebCore:

        Reviewed by Sam.

        <rdar://problem/5575812> REGRESSION:When using absolute positioning
        with overflow:auto div, WebKit seems to add an additional 15px

        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::containingBlockWidthForPositioned): We need to
        subtract off the vertical scrollbar width too.

LayoutTests:

        Reviewed by Sam.

        New test for <rdar://problem/5575812> REGRESSION:When using
        absolute positioning with overflow:auto div, WebKit seems to add an
        additional 15px
        * fast/overflow/overflow-auto-position-absolute.html: Added.
        * platform/mac/fast/overflow/overflow-auto-position-absolute-expected.checksum: Added.
        * platform/mac/fast/overflow/overflow-auto-position-absolute-expected.png: Added.
        * platform/mac/fast/overflow/overflow-auto-position-absolute-expected.txt: Added.

        Test with changed results. The measurements in the rtl example now
        more closely mirror the ltr example.
        * platform/mac/fast/overflow/unreachable-overflow-rtl-bug-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27499 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Change WebLocalizableStrings to take UTF-8 C strings
aroben [Wed, 7 Nov 2007 05:29:02 +0000 (05:29 +0000)]
     Change WebLocalizableStrings to take UTF-8 C strings

         This matches the way things work on the Mac, and will allow source
         files containing localizable strings to be shared between Mac and
         Windows. The old functions have not been removed for compatibility
         reasons, but are now just wrappers around the new UTF-8 functions.

         Reviewed by Ada.

         * WebKit.vcproj/WebKit.def: Added new functions.
         * WebKit.vcproj/WebKit_debug.def: Ditto.
         * WebLocalizableStrings.cpp:
         (copyLocalizedStringFromBundle): Changed to take a WebCore::String
         representing the key.
         (localizedString): Refactored from WebLocalizedString. Takes a
         WebCore::String representing the key.
         (localizedLPCTSTR): Ditto for WebLocalizedLPCTSTR.
         (WebLocalizedStringUTF8): Added. Takes a UTF-8 C string as the key.
         (WebLocalizedLPCTSTRUTF8): Ditto.
         (WebLocalizedString): Changed to call localizedString.
         (WebLocalizedLPCTSTR): Changed to call localizedLPCTSTR.
         * WebLocalizableStrings.h: Changed macros to use the new UTF-8
         functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27498 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Eric Seidel <eric@webkit.org>
eric@webkit.org [Wed, 7 Nov 2007 05:24:17 +0000 (05:24 +0000)]
2007-11-06  Eric Seidel  <eric@webkit.org>

        Reviewed by Mark Rowe.

        * Scripts/build-testkjs: return xcodebuild's exit status, instead of grep's

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27497 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoMove Mac files from WebKit into WebKit/mac.
mrowe@apple.com [Wed, 7 Nov 2007 03:01:16 +0000 (03:01 +0000)]
Move Mac files from WebKit into WebKit/mac.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27496 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
justing [Wed, 7 Nov 2007 01:13:47 +0000 (01:13 +0000)]
WebCore:

        Reviewed by Dan Bernstein.

        <rdar://problem/5583387> ASSERTION FAILED: !refChild->hasTagName(bodyTag) when pasting newline in plain text into rich text Mail
        <rdar://problem/5583362> REGRESSION (5523.10.3-TOT): Newlines stripped when pasting plain text in Mail

        * editing/markup.cpp:
        (WebCore::createFragmentFromText): Put paragraphs of text into clones of the block
        that encloses the input context, unless that block is the body, which shouldn't
        be cloned.  In that case, use regular divs, as we did before r27369.

LayoutTests:

        Reviewed by Dan Bernstein.

        <rdar://problem/5583387> ASSERTION FAILED: !refChild->hasTagName(bodyTag) when pasting newline in plain text into rich text Mail
        <rdar://problem/5583362> REGRESSION (5523.10.3-TOT): Newlines stripped when pasting plain text in Mail

        * editing/pasteboard/5583362.html: Added.
        * editing/resources/plaintext-pasteboard-data.dat: Added.
        * platform/mac/editing/pasteboard/5583362-expected.checksum: Added.
        * platform/mac/editing/pasteboard/5583362-expected.png: Added.
        * platform/mac/editing/pasteboard/5583362-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27495 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Eric Seidel <eric@webkit.org>
eseidel [Wed, 7 Nov 2007 01:03:26 +0000 (01:03 +0000)]
2007-11-06  Eric Seidel  <eric@webkit.org>

        Reviewed by darin.

        This fixes a regressed layout test for string + object

        SunSpider claims this was an overall 0.3% speedup, although some individual tests were slower.

        * kjs/nodes.cpp:
        (KJS::add): remove erroneous "fast path" for string + *

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27494 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Eric Seidel.
ggaren@apple.com [Wed, 7 Nov 2007 00:06:43 +0000 (00:06 +0000)]
    Reviewed by Eric Seidel.

        Added toJSNumber, a fast path for converting a JSValue to a JS number,
        and deployed it in postfix expressions. In the fast case this
        eliminates a call to jsNumber.

        0.4% speedup on SunSpider.

        * ChangeLog:
        * kjs/nodes.cpp:
        (KJS::PostIncResolveNode::evaluate):
        (KJS::PostIncLocalVarNode::evaluate):
        (KJS::PostDecResolveNode::evaluate):
        (KJS::PostDecLocalVarNode::evaluate):
        (KJS::PostIncBracketNode::evaluate):
        (KJS::PostDecBracketNode::evaluate):
        (KJS::PostIncDotNode::evaluate):
        (KJS::PostDecDotNode::evaluate):
        (KJS::UnaryPlusNode::evaluate):
        * kjs/value.h:
        (KJS::JSValue::toJSNumber):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27493 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Christian Dywan <christian@twotoasts.de>
mrowe@apple.com [Wed, 7 Nov 2007 00:04:38 +0000 (00:04 +0000)]
2007-11-06  Christian Dywan  <christian@twotoasts.de>

        Reviewed by Darin.

        Fix http://bugs.webkit.org/show_bug.cgi?id=15828
        Bug 15828: WebKit GTK include and lib directory is installed in qt4-named directory

        * WebCore.pro: Use sane default install paths for the gtk port.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27492 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Rodney Dawes <dobey@wayofthemonkey.com>
mrowe@apple.com [Wed, 7 Nov 2007 00:02:04 +0000 (00:02 +0000)]
2007-11-06  Rodney Dawes  <dobey@wayofthemonkey.com>

        Fix http://bugs.webkit.org/attachment.cgi?id=17043&action=view
        Bug 15766: [GTK] WebKit sometimes spews binary data as text/plain into iframes

        FrameLoaderClient::objectContentType needs to check with the MIMETypeRegistry
        to determine whether the given MIME type is displayable as an image or non-image.

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (FrameLoaderClient::objectContentType): Change logic to match that in
        the Windows and Mac ports.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27491 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Correct test results.
antti [Tue, 6 Nov 2007 23:19:02 +0000 (23:19 +0000)]
    Correct test results.

        * platform/mac/fast/layers/layer-visibility-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27490 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Hyatt.
antti [Tue, 6 Nov 2007 23:00:27 +0000 (23:00 +0000)]
    Reviewed by Hyatt.

        Update test to cover
        <rdar://problem/5521068>
        Visibility not propagated correctly for children of a layer with z-index

        The fix for this was (accidentally) already checked in with r27277,
        with retroactive r=hyatt.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27489 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
justing [Tue, 6 Nov 2007 22:28:59 +0000 (22:28 +0000)]
WebCore:

        Reviewed by Darin Adler.

        <rdar://problem/5576619>
        REGRESSION: Caret disappears after deleting the last character in inline hole (15714)

        * editing/TypingCommand.cpp:
        (WebCore::TypingCommand::deleteSelection): Like the other TypingCommands, including
        both of the other deletion commands, call typingAddedToOpenCommand(), which takes
        the command's endingSelection and sets it as selection.

LayoutTests:

        Reviewed by Darin Adler.

        <rdar://problem/5576619> REGRESSION: Caret disappears after deleting the last character in inline hole (15714)

        * platform/mac/editing/input/5576619-expected.checksum: Added.
        * platform/mac/editing/input/5576619-expected.png: Added.
        * platform/mac/editing/input/5576619-expected.txt: Added.
        * platform/mac/editing/input/5576619.html: Added.
        * platform/mac/editing/input/text-input-controller-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27488 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
darin@apple.com [Tue, 6 Nov 2007 21:37:40 +0000 (21:37 +0000)]
    Reviewed by Maciej.

        - http://bugs.webkit.org/show_bug.cgi?id=15846
          REGRESSION (r27387): Memory corruption when running fast/js/kde/delete.html

        There was a mistake in the algorithm used to find an empty slot in the property
        map entries vector; when we were putting in a new property value and not overwriting
        an existing deleted sentinel, we would enlarge the entries vector, but would not
        overwrite the stale data that's in the new part. It was easy to pin this down by
        turning on property map consistency checks -- I never would have landed with this
        bug if I had run the regression tests once with consistency checks on!

        * kjs/property_map.cpp: (KJS::PropertyMap::put): Changed logic for the case where
        foundDeletedElement is false to always use the item at the end of the entries vector.
        Also allowed me to merge with the logic for the "no deleted sentinels at all" case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27487 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
mitz@apple.com [Tue, 6 Nov 2007 21:03:35 +0000 (21:03 +0000)]
WebCore:

        Reviewed by Antti Koivisto and Dave Hyatt.

        - fix <rdar://problem/5582961> Incorrect layout and floating object list corruption when CSS decreases a block's height

        Test: fast/block/float/overhanging-after-height-decrease.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::layoutBlock): If after calculating the height
        it turns out that there are overhanging floats that were not overhanging
        before, rescan children with overhanging floats and add them.
        (WebCore::RenderBlock::layoutBlockChildren): Added a parameter that
        returns the lowest float bottom of any of the children.
        (WebCore::RenderBlock::addOverhangingFloats): Changed to return the
        lowest float bottom.
        * rendering/RenderBlock.h:

LayoutTests:

        Reviewed by Antti Koivisto and Dave Hyatt.

        - test for <rdar://problem/5582961> Incorrect layout and floating object list corruption when CSS decreases a block's height

        * fast/block/float/overhanging-after-height-decrease.html: Added.
        * platform/mac-leopard/fast/block: Added.
        * platform/mac-leopard/fast/block/float: Added.
        * platform/mac-leopard/fast/block/float/overhanging-after-height-decrease-expected.checksum: Added.
        * platform/mac-leopard/fast/block/float/overhanging-after-height-decrease-expected.png: Added.
        * platform/mac/fast/block/float/overhanging-after-height-decrease-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27486 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
adele [Tue, 6 Nov 2007 19:53:18 +0000 (19:53 +0000)]
WebCore:

        Reviewed by Darin.

        Switched all uses of HTMLImageLoader to use OwnPtrs.

        * html/HTMLInputElement.h:
        * html/HTMLInputElement.cpp:
        (WebCore::HTMLInputElement::init):
        (WebCore::HTMLInputElement::~HTMLInputElement):
        (WebCore::HTMLInputElement::setInputType):
        (WebCore::HTMLInputElement::parseMappedAttribute):
        (WebCore::HTMLInputElement::attach):
        * html/HTMLObjectElement.h:
        * html/HTMLObjectElement.cpp:
        (WebCore::HTMLObjectElement::HTMLObjectElement):
        (WebCore::HTMLObjectElement::~HTMLObjectElement):
        (WebCore::HTMLObjectElement::parseMappedAttribute):
        (WebCore::HTMLObjectElement::attach):
        * html/HTMLVideoElement.h:
        * html/HTMLVideoElement.cpp:
        (WebCore::HTMLVideoElement::HTMLVideoElement):
        (WebCore::HTMLVideoElement::attach):
        (WebCore::HTMLVideoElement::detach):
        (WebCore::HTMLVideoElement::parseMappedAttribute):

LayoutTests:

        Reviewed by Darin.

        Added test for poster attribute.

        * media/content/abe.png: Added.
        * media/content/greenbox.png: Added.
        * media/video-poster-expected.txt: Added.
        * media/video-poster.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27485 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix previous patch to use a 3 bit shift, a 16 bit shift causes a regression in sunspider.
oliver [Tue, 6 Nov 2007 19:49:39 +0000 (19:49 +0000)]
Fix previous patch to use a 3 bit shift, a 16 bit shift causes a regression in sunspider.

RS=Darin

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27484 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoReplace boolean comparisons in AddNode with mask comparisons for a 0.2% improvement...
oliver [Tue, 6 Nov 2007 19:34:41 +0000 (19:34 +0000)]
Replace boolean comparisons in AddNode with mask comparisons for a 0.2% improvement in sunspider.

Reviewed by Darin

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27483 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Eric Seidel <eric@webkit.org>
eseidel [Tue, 6 Nov 2007 18:56:05 +0000 (18:56 +0000)]
2007-11-06  Eric Seidel  <eric@webkit.org>

        Reviewed by darin.

        SunSpider claims this is a 1.1% speedup.

        * kjs/nodes.cpp:
        (KJS::throwOutOfMemoryError): Added, non inline.
        (KJS::addSlowCase): renamed from add(), non inline.
        (KJS::add): add fast path for String + String, Number + Number and String + *

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27482 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam and Steve.
darin@apple.com [Tue, 6 Nov 2007 18:43:38 +0000 (18:43 +0000)]
    Reviewed by Adam and Steve.

        - added hooks needed to implement showModalDialog on Windows

        * Interfaces/IWebUIDelegate.idl: Added canRunModal, createModalDialog,
        and runModal functions to the end of IWebUIDelegate3.

        * WebChromeClient.h: Added uiDelegate and uiDelegate2 helper functions,
        so it's easier to write client functions.
        * WebChromeClient.cpp:
        (WebChromeClient::createWindow): Implemented dialog case. Calls IWebUIDelegate3.
        (WebChromeClient::canRunModal): Implemented. Calls IWebUIDelegate3.
        (WebChromeClient::runModal): Ditto.
        (WebChromeClient::uiDelegate): Added.
        (WebChromeClient::uiDelegate2): Added.
        (WebChromeClient::uiDelegate3): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27481 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
antti [Tue, 6 Nov 2007 18:11:06 +0000 (18:11 +0000)]
WebCore:

        Reviewed by Darin.

        Trigger media load on on src attribute changes as specified in new HTML5 draft.

        Tests: media/video-src-change.html
               media/video-src-remove.html
               media/video-src-set.html

        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::attributeChanged):
        * html/HTMLMediaElement.h:

LayoutTests:

        Reviewed by Darin.

        Tests for media src attribute changes.

        * media/video-src-change-expected.txt: Added.
        * media/video-src-change.html: Added.
        * media/video-src-remove-expected.txt: Added.
        * media/video-src-remove.html: Added.
        * media/video-src-set-expected.txt: Added.
        * media/video-src-set.html: Added.
        * media/video-test.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27480 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rubber-stamped by Adam Roben.
ap [Tue, 6 Nov 2007 18:10:22 +0000 (18:10 +0000)]
    Rubber-stamped by Adam Roben.

        Windows build fix.

        * WebFrame.cpp:
        (WebFrame::string): plainText() returns a String now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27479 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
mitz@apple.com [Tue, 6 Nov 2007 17:48:56 +0000 (17:48 +0000)]
WebCore:

        Reviewed by Darin Adler.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15838
          Incomplete repaint toggling "How you know this person" on Facebook

        Test: fast/repaint/make-children-non-inline.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::makeChildrenNonInline): Repaint the block. This
        is needed because the inline children may be repositioned as they move
        into new anonymous blocks, but those blocks have no knowledge of where
        their children used to be, so they cannot invalidate those areas.

LayoutTests:

        Reviewed by Darin Adler.

        - repaint test for http://bugs.webkit.org/show_bug.cgi?id=15838
          Incomplete repaint toggling "How you know this person" on Facebook

        * fast/repaint/make-children-non-inline.html: Added.
        * platform/mac/fast/repaint/make-children-non-inline-expected.txt: Added.
        * platform/mac-leopard/fast/repaint: Added.
        * platform/mac-leopard/fast/repaint/make-children-non-inline-expected.checksum: Added.
        * platform/mac-leopard/fast/repaint/make-children-non-inline-expected.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27478 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
ap [Tue, 6 Nov 2007 16:07:38 +0000 (16:07 +0000)]
    Reviewed by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=15847
        Some editing cleanup

        No change in functionality.

WebCore:
        * editing/TextIterator.cpp:
        (WebCore::plainText):
        * editing/TextIterator.h:
        Made WebCore::plainText() return String instead of DeprecatedString.

        * bridge/mac/WebCoreAXObject.mm:
        (-[WebCoreAXObject textUnderElement]):
        (-[WebCoreAXObject value]):
        (-[WebCoreAXObject doAXBoundsForTextMarkerRange:]):
        (-[WebCoreAXObject doAXNextSentenceEndTextMarkerForTextMarker:]):
        (-[WebCoreAXObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
        Updated for the above change. There is no need to explicitly convert to NSString now.

        * editing/EditCommand.cpp:
        (WebCore::EditCommand::EditCommand): m_startingSelection and m_endingSelection are actually
        initialized in constructor body, so the work done in initializer list was wasted.

        (WebCore::EditCommand::apply): Moved some stars.
        (WebCore::EditCommand::unapply): Ditto.
        (WebCore::EditCommand::reapply): Ditto.
        (WebCore::EditCommand::setStartingSelection): The loop exit condition was evaluated twice,
        removed one of the checks.

        * editing/SelectionController.cpp:
        (WebCore::SelectionController::toString): plainText() result type now matches what we need here.

        * page/mac/WebCoreFrameBridge.h:
        * page/mac/WebCoreFrameBridge.mm:
        Removed unused -[WebCoreFrameBridge deleteKeyPressedWithSmartDelete:granularity:] and
        -[WebCoreFrameBridge forwardDeleteKeyPressedWithSmartDelete:granularity:].

WebKit:
        * WebView/WebHTMLView.mm:
        (-[WebHTMLView deleteToEndOfLine:]):
        (-[WebHTMLView deleteToEndOfParagraph:]):
        WebCore had a duplicate of the same logic already. We are passing a boundary value to
        a function that expects granularity, this may need to be straightened out in the future.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27477 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-06 Eric Seidel <eric@webkit.org>
eseidel [Tue, 6 Nov 2007 09:08:56 +0000 (09:08 +0000)]
2007-11-06  Eric Seidel  <eric@webkit.org>

        Reviewed by mjs.

        Avoid more UString creation.

        SunSpider claims this is a 0.4% speedup.

        * kjs/regexp_object.cpp:
        (KJS::RegExpObjectImp::construct): use UString::find(UChar)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27476 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Tue, 6 Nov 2007 07:47:29 +0000 (07:47 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Mac build fix.

        * bindings/js/JSHTMLDocumentCustom.cpp:
        (WebCore::writeHelper):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27475 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Tue, 6 Nov 2007 07:30:22 +0000 (07:30 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Mac build fix.

        * kjs/array_object.cpp:
        (KJS::ArrayProtoFunc::callAsFunction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27474 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Don't crash if SafariTheme can't be loaded
aroben [Tue, 6 Nov 2007 07:20:24 +0000 (07:20 +0000)]
     Don't crash if SafariTheme can't be loaded

         PlatformScrollBarSafari and RenderThemeSafari were not checking
         whether SafariTheme was successfully loaded. All other uses of
         SafariTheme already check this.

         Reviewed by Steve.

         * platform/win/PlatformScrollBarSafari.cpp:
         (WebCore::PlatformScrollbar::paintButton):
         (WebCore::PlatformScrollbar::paintTrack):
         (WebCore::PlatformScrollbar::paintThumb):
         * rendering/RenderThemeSafari.cpp:
         (WebCore::RenderThemeSafari::isControlStyled):
         (WebCore::RenderThemeSafari::paintCheckbox):
         (WebCore::RenderThemeSafari::paintRadio):
         (WebCore::RenderThemeSafari::paintButton):
         (WebCore::RenderThemeSafari::paintTextField):
         (WebCore::RenderThemeSafari::paintCapsLockIndicator):
         (WebCore::RenderThemeSafari::paintTextArea):
         (WebCore::RenderThemeSafari::paintMenuList):
         (WebCore::RenderThemeSafari::paintSliderThumb):
         (WebCore::RenderThemeSafari::paintSearchField):
         (WebCore::RenderThemeSafari::paintSearchFieldCancelButton):
         (WebCore::RenderThemeSafari::paintSearchFieldResultsDecoration):
         (WebCore::RenderThemeSafari::paintSearchFieldResultsButton):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27473 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Windows build fix
aroben [Tue, 6 Nov 2007 07:20:06 +0000 (07:20 +0000)]
    Windows build fix

        * kjs/list.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27472 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Tue, 6 Nov 2007 07:09:25 +0000 (07:09 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Build fix.  Add missing #include.

        * kjs/operations.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27471 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Eric Seidel <eric@webkit.org>
eseidel [Tue, 6 Nov 2007 06:48:10 +0000 (06:48 +0000)]
2007-11-05  Eric Seidel  <eric@webkit.org>

        Reviewed by mjs.

        Remove another call to toString(exec)

        SunSpider claims this is a 0.5% speedup.

        * kjs/operations.cpp:
        (KJS::equal): remove another toString

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27470 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoAdding the rest of wx port's graphics impl.
kevino [Tue, 6 Nov 2007 06:39:22 +0000 (06:39 +0000)]
Adding the rest of wx port's graphics impl.

Rubber-stamped by Maciej Stachowiak.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27469 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Eric Seidel <eric@webkit.org>
eseidel [Tue, 6 Nov 2007 06:24:03 +0000 (06:24 +0000)]
2007-11-05  Eric Seidel  <eric@webkit.org>

        * kjs/operations.cpp:
        (KJS::equal): correct broken change.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27468 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Eric Seidel <eric@webkit.org>
eseidel [Tue, 6 Nov 2007 06:17:22 +0000 (06:17 +0000)]
2007-11-05  Eric Seidel  <eric@webkit.org>

        Reviewed by mjs.

        Remove one more call to toString(exec).

        SunSpider claims this is a 0.7% speedup.

        * kjs/operations.cpp:
        (KJS::equal): remove a call to toString()

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27467 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agowx port bug fixes and new files to get platform/wx building on trunk.
kevino [Tue, 6 Nov 2007 05:52:06 +0000 (05:52 +0000)]
wx port bug fixes and new files to get platform/wx building on trunk.

Rubber-stamped by Maciej Stachowiak.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27466 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
adele [Tue, 6 Nov 2007 05:43:16 +0000 (05:43 +0000)]
    Reviewed by Maciej.

        Fix for <rdar://problem/5579999> Add poster attribute for video element

        * html/HTMLAttributeNames.in: Added poster attribute.

        * html/HTMLImageLoader.cpp: (WebCore::HTMLImageLoader::updateFromElement):
          Calls imageSourceAttributeName instead of having special cases for the different kinds of elements.
        * dom/Element.cpp: (WebCore::Element::imageSourceAttributeName): Added.
        * dom/Element.h:
        * html/HTMLObjectElement.cpp: (WebCore::HTMLObjectElement::imageSourceAttributeName): Added.
        * html/HTMLObjectElement.h:

        * html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::movieNetworkStateChanged):
          Calls updatePosterImage when the network state is empty and when the first frame has been loaded.
        * html/HTMLMediaElement.h: (WebCore::HTMLMediaElement::rendererIsNeeded): Made this inline.  HTMLVideoElement now also implements this.

        * html/HTMLVideoElement.idl: Added case for poster attribute.
        * html/HTMLVideoElement.h: Added image loader and flag to keep track of whether or not the poster image should be shown.
        * html/HTMLVideoElement.cpp:
        (WebCore::HTMLVideoElement::HTMLVideoElement): Initialize m_imageLoader and m_shouldShowPosterImage.
        (WebCore::HTMLVideoElement::rendererIsNeeded): Calls HTMLElement::rendererIsNeeded since HTMLMediaElements have renderer by default.
        (WebCore::HTMLVideoElement::createRenderer): Create a RenderImage or RenderVideo depending on whether or not the poster image should be shown.
        (WebCore::HTMLVideoElement::attach): Set up image loader and RenderImage if necessary.
        (WebCore::HTMLVideoElement::detach): Delete image loader if its no longer needed.
        (WebCore::HTMLVideoElement::parseMappedAttribute): Added case to process poster attribute.
        (WebCore::HTMLVideoElement::poster): Added getter for poster attribute.
        (WebCore::HTMLVideoElement::setPoster): Added setter for poster attribute.
        (WebCore::HTMLVideoElement::isURLAttribute): Added.
        (WebCore::HTMLVideoElement::imageSourceAttributeName): Added.
        (WebCore::HTMLVideoElement::updatePosterImage): Added.  Updates m_shouldShowPosterImage and if its changed, detaches and attaches so the renderer is correct.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27465 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix <rdar://5563572> SVG image support is turned off
aroben [Tue, 6 Nov 2007 05:25:37 +0000 (05:25 +0000)]
Fix <rdar://5563572> SVG image support is turned off

WebCore:

        Fix <rdar://5563572> SVG image support is turned off

        Turned on SVG images for all platforms.

        Reviewed by Eric.

        * WebCore.vcproj/WebCore.vcproj: Added SVGImage.{cpp,h}
        * loader/CachedImage.cpp:
        (WebCore::CachedImage::createImage): Removed platform checks for
        SVGImage.

LayoutTests:

        Remove a now-passing test from the Windows Skipped file

        Reviewed by Eric.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27464 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix spelling errors in ChangeLogs, no review
eseidel [Tue, 6 Nov 2007 04:55:25 +0000 (04:55 +0000)]
Fix spelling errors in ChangeLogs, no review

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27463 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
antti [Tue, 6 Nov 2007 04:43:08 +0000 (04:43 +0000)]
    Reviewed by Darin.

        QTMovieView can generate callbacks during paint. This can lead to crashes.

        Delay callbacks so they get handled after painting is completed. No test case,
        I don't know how to reliably reproduce this.

        * platform/graphics/mac/MoviePrivateQTKit.mm:
        (WebCore::MoviePrivate::MoviePrivate):
        (WebCore::MoviePrivate::~MoviePrivate):
        (WebCore::MoviePrivate::paint):
        (-[WebCoreMovieObserver initWithCallback:WebCore::]):
        (-[WebCoreMovieObserver disconnect]):
        (-[WebCoreMovieObserver loadStateChanged:]):
        (-[WebCoreMovieObserver rateChanged:]):
        (-[WebCoreMovieObserver sizeChanged:]):
        (-[WebCoreMovieObserver timeChanged:]):
        (-[WebCoreMovieObserver volumeChanged:]):
        (-[WebCoreMovieObserver didEnd:]):
        (-[WebCoreMovieObserver setDelayCallbacks:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27462 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam.
antti [Tue, 6 Nov 2007 04:12:23 +0000 (04:12 +0000)]
    Reviewed by Adam.

        Add some missing WebCore* prefixes to ObjC classes

        * platform/mac/FileChooserMac.mm:
        (WebCore::FileChooser::FileChooser):
        * platform/mac/SharedBufferMac.mm:
        (WebCore::SharedBuffer::createNSData):
        (WebCore::SharedBuffer::createCFData):
        * platform/mac/SharedTimerMac.cpp:
        (WebCore::setSharedTimerFireTime):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27461 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Christian Dywan <christian@twotoasts.de>
alp [Tue, 6 Nov 2007 02:53:16 +0000 (02:53 +0000)]
2007-11-05  Christian Dywan  <christian@twotoasts.de>

        Reviewed by Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=15409
        FrameLoaderClientGtk hardcodes data, including platform to Linux i686

        Compute a proper user agent string.

        Patch includes fixes by Alp.

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (WebKit::agentPlatform):
        (WebKit::agentOS):
        (WebKit::composeUserAgent):
        (WebKit::FrameLoaderClient::userAgent):
        * WebCoreSupport/FrameLoaderClientGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27460 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago <rdar://problem/5579772> Regression: AltGr does not work
adachan [Tue, 6 Nov 2007 02:10:05 +0000 (02:10 +0000)]
    <rdar://problem/5579772> Regression: AltGr does not work
        Testing whether alt-key is down is not the right test for system key event.
        Added a m_isSystemKey flag in PlatformKeyboardEvent to keep track of whether
        this is a system key event, check that flag instead in handleEditingKeyboardEvent().

        Reviewed by Oliver.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27459 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Tue, 6 Nov 2007 00:35:47 +0000 (00:35 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Gtk build fix.

        * pcre/pcre.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27458 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build fix.
adachan [Tue, 6 Nov 2007 00:02:50 +0000 (00:02 +0000)]
    Build fix.

        Reviewed by Tristan.

        * WebChromeClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27457 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam.
antti [Mon, 5 Nov 2007 23:56:01 +0000 (23:56 +0000)]
    Reviewed by Sam.

        Update WKQTMovieViewSetDrawSynchronously.

        * libWebKitSystemInterfaceLeopard.a:
        * libWebKitSystemInterfaceTiger.a:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27456 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * WebView/WebView.mm:
sullivan@apple.com [Mon, 5 Nov 2007 23:21:10 +0000 (23:21 +0000)]
    * WebView/WebView.mm:
        (-[WebView _searchWithSpotlightFromMenu:]):

        Teeny style tweak to test svn access on other machine

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27455 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * WebView/WebView.mm:
sullivan@apple.com [Mon, 5 Nov 2007 23:12:58 +0000 (23:12 +0000)]
    * WebView/WebView.mm:
        (-[WebView computedStyleForElement:pseudoElement:]):

        Teeny style tweak to test svn access

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27454 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Mon, 5 Nov 2007 23:05:03 +0000 (23:05 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Gtk build fix.

        * kjs/list.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27453 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoqt:
tristan [Mon, 5 Nov 2007 22:43:45 +0000 (22:43 +0000)]
qt:

        Reviewed by Darin Adler.

        * WebCoreSupport/ChromeClientQt.cpp:
        (WebCore::ChromeClientQt::createWindow):
        * WebCoreSupport/ChromeClientQt.h:
        Revised to use new WebCore ChromeClient createWindow API.

WebCore:

        Reviewed by Darin Adler.

        * WebCore.xcodeproj/project.pbxproj:
        Set WindowFeatures.h as a Private header.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::createWindow):
        Revised to use a single createWindow function instead of
        two createWindow functions and one createModalDialog function.
        The logic for this is now addressed in WebKit in an effort
        to make this easier to follow.

        * page/Chrome.cpp:
        (WebCore::Chrome::createWindow):
        * page/Chrome.h:
        * page/ChromeClient.h:
        * page/ContextMenuController.cpp:
        (WebCore::openNewWindow):
        * platform/graphics/svg/SVGImageEmptyClients.h:
        (WebCore::SVGEmptyChromeClient::createWindow):
        Revised to take new additional windowFeatures parameter.

WebKit:

        Reviewed by Darin Adler.

        * DefaultDelegates/WebDefaultUIDelegate.m:
        (-[WebDefaultUIDelegate webView:createWebViewWithRequest:windowFeatures:]):
        Forward the UI delegate to call webView:createWebViewWithRequest: if
        this method doesn't exist.

        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::dispatchCreatePage):
        * Plugins/WebBaseNetscapePluginView.mm:
        (-[WebBaseNetscapePluginView loadPluginRequest:]):
        * WebView/WebView.mm:
        (-[WebView _openNewWindowWithRequest:]):
        Revised to use new webView:createWebViewWithRequest:windowFeatures: callback.

        * WebCoreSupport/WebChromeClient.h:
        * WebCoreSupport/WebChromeClient.mm:
        (WebChromeClient::createWindow):
        Added a new createWindow that accepts 3 parameters, so we can pass up windowFeatures to the chrome.
        Removed createModalDialog to use new createWindow function.

        * WebView/WebUIDelegatePrivate.h:
        Added new webView:createWebViewWithRequest:windowFeatures: method.

win:

        Reviewed by Darin Adler.

        Part of the WebKit/WebCore API changes for
        <rdar://problem/5368188>

        * WebChromeClient.cpp:
        (WebChromeClient::createWindow):
        Removed usage of createModalDialog and revised to use new createWindow
        ChromeClient API.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27452 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Touched a file to test my new HTTP access.
ggaren@apple.com [Mon, 5 Nov 2007 22:02:04 +0000 (22:02 +0000)]
    Touched a file to test my new HTTP access.

        * kjs/scope_chain.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27451 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build fix.
ggaren [Mon, 5 Nov 2007 21:56:09 +0000 (21:56 +0000)]
    Build fix.

        * ForwardingHeaders/kjs/value.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27450 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Alp Toker <alp@atoker.com>
alp [Mon, 5 Nov 2007 21:35:03 +0000 (21:35 +0000)]
2007-11-05  Alp Toker  <alp@atoker.com>

        Unreviewed build fix for qmake-based ports.

        Someone with a better understanding of qmake still needs to sort out
        the INCLUDEPATH/DEPENDPATH mess.

        * JavaScriptCore.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27449 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
ggaren [Mon, 5 Nov 2007 21:27:15 +0000 (21:27 +0000)]
JavaScriptCore:

        Reviewed by Darin Adler.

        http://bugs.webkit.org/show_bug.cgi?id=15835

        Switched List implementation from a custom heap allocator to an inline
        Vector, for a disappointing .5% SunSpider speedup.

        Also renamed List::slice to List::getSlice because "get" is the
        conventional prefix for functions returning a value through an out
        parameter.

        * kjs/array_object.cpp:
        (KJS::ArrayProtoFunc::callAsFunction): Removed some redundant function
        calls and memory accesses.

        * kjs/bool_object.cpp:
        (BooleanObjectImp::construct): Removed questionable use of iterator.

        * kjs/list.cpp:
        * kjs/list.h: New List class, implemented in terms of Vector. Two
        interesting differences:
            1. The inline capacity is 8, not 5. Many of the Lists constructed
            during a SunSpider run are larger than 5; almost none are larger
            than 8.

            2. The growth factor is 4, not 2. Since we can guarantee that Lists
            aren't long-lived, we can grow them more aggressively, to avoid
            excessive copying.

        * kjs/regexp_object.cpp:
        (RegExpObjectImp::construct): Removed redundant function calls.

        * kjs/string_object.cpp:
        (KJS::StringObjectImp::construct): Removed questionable use of iterator.

        * wtf/Vector.h:
        (WTF::::uncheckedAppend): Added a fast, unchecked version of append.

WebCore:

        Reviewed by Darin Adler.

        http://bugs.webkit.org/show_bug.cgi?id=15835

        Small adaptations to new KJS::List class.

        * bindings/js/kjs_window.cpp:
        (KJS::WindowFunc::callAsFunction):
        (KJS::ScheduledAction::ScheduledAction):

WebKit:

        Reviewed by Darin Adler.

        http://bugs.webkit.org/show_bug.cgi?id=15835

        Small adaptations to new KJS::List class.

        * ForwardingHeaders/kjs/value.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27448 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Mon, 5 Nov 2007 21:09:12 +0000 (21:09 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Reviewed by Alp Toker.

        Add DEPENDPATH to JavaScriptCore and pcre to help qmake with dependencies.

        * JavaScriptCore.pri:
        * pcre/pcre.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27447 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add support on Windows for WEBKIT_TESTFONTS
aroben [Mon, 5 Nov 2007 18:57:49 +0000 (18:57 +0000)]
     Add support on Windows for WEBKIT_TESTFONTS

         This environment variable lets you specify where the fonts to be used
         by DumpRenderTree reside. The Qt port is already using this, so I'm
         just following their lead.

         Reviewed by Darin.

         * DumpRenderTree/win/DumpRenderTree.cpp:
         (exePath): Refactored code out of initialize().
         (fontsPath): Returns either $WEBKIT_TESTFONTS or
         DumpRenderTree.resources.
         (initialize): Use the new fontsPath function.
         (main): Use the new exePath function.
         * Scripts/run-webkit-tests: Propagate the WEBKIT_TESTFONTS environment
         variable to DRT, like Qt does.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27446 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add IWebCache::disabled
aroben [Mon, 5 Nov 2007 18:57:30 +0000 (18:57 +0000)]
     Add IWebCache::disabled

         This matches the Mac WebCache class.

         Reviewed by Sam.

         * Interfaces/IWebCache.idl:
         * WebCache.cpp:
         (WebCache::disabled):
         * WebCache.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27445 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add IWebInspector and a way to get one from a WebView
aroben [Mon, 5 Nov 2007 18:57:10 +0000 (18:57 +0000)]
 Add IWebInspector and a way to get one from a WebView

 WebCore:

         Allow passing a base class pointer to COMPtr::copyRefTo

         Reviewed by Sam.

         * platform/win/COMPtr.h:

 WebKit/win:

         Add IWebInspector and a way to get one from a WebView

         This API matches the Mac one added in r27266.

         Reviewed by Sam.

         * Interfaces/IWebInspector.idl: Added.
         * Interfaces/IWebViewPrivate.idl: Added a new inspector method to
         match Mac.
         * WebInspector.cpp: Added.
         (WebInspector::createInstance):
         (WebInspector::WebInspector):
         (WebInspector::~WebInspector):
         (WebInspector::webViewClosed):
         (WebInspector::QueryInterface):
         (WebInspector::AddRef):
         (WebInspector::Release):
         (WebInspector::show):
         (WebInspector::showConsole):
         (WebInspector::showTimeline):
         (WebInspector::close):
         (WebInspector::attach):
         (WebInspector::detach):
         * WebInspector.h: Added.
         * WebKit.vcproj/Interfaces.vcproj: Added IWebInspector.idl.
         * WebKit.vcproj/WebKit.vcproj: Added WebInspector.{cpp,h}.
         * WebKit.vcproj/WebKitGUID.vcproj: Added IWebInspector_i.c.
         * WebView.cpp:
         (WebView::close): Notify the WebInspector that we're closing.
         (WebView::inspector): Added.
         * WebView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27444 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Cache strings that come directly from WebLocalizedString
aroben [Mon, 5 Nov 2007 18:56:49 +0000 (18:56 +0000)]
     Cache strings that come directly from WebLocalizedString

         Reviewed by Steve.

         * WebLocalizableStrings.cpp:
         (copyLocalizedStringFromBundle): This used to be WebLocalizedString,
         but is now just a static helper function.
         (findCachedStringInMap): Added static helper.
         (findCachedString): Refactored and cleaned up code from
         WebLocalizedLPCTSTR.
         (cacheString): Ditto.
         (WebLocalizedString): Rewrote to use the new helper functions.
         (WebLocalizedLPCTSTR): Changed to use the new helper functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27443 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Only override the default user agent string if we're actually given a custom one
aroben [Mon, 5 Nov 2007 18:55:46 +0000 (18:55 +0000)]
     Only override the default user agent string if we're actually given a custom one

         Reviewed by Sam.

         * WebView.cpp:
         (WebView::setCustomUserAgent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27442 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
mitz@apple.com [Mon, 5 Nov 2007 18:52:49 +0000 (18:52 +0000)]
WebCore:

        Reviewed by Oliver Hunt.

        - fix ASSERTION FAILED: !HashTranslator::equal(KeyTraits::emptyValue(), key) when a class attribute is all-whitespace

        Test: fast/dom/class-all-whitespace.html

        * dom/StyledElement.cpp:
        (WebCore::StyledElement::parseMappedAttribute): Check if there is any
        non-whitespace character in the class attribute.

LayoutTests:

        Reviewed by Oliver Hunt.

        - test that an all-whitespace class attribute does not cause an assertion failure

        * fast/dom/class-all-whitespace-expected.txt: Added.
        * fast/dom/class-all-whitespace.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27441 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam
beidson@apple.com [Mon, 5 Nov 2007 18:23:19 +0000 (18:23 +0000)]
    Reviewed by Sam

        Add transaction blocking to the DatabaseAuthorizer in preparation for the new version
        of the sql storage API

        * storage/DatabaseAuthorizer.cpp:
        (WebCore::DatabaseAuthorizer::allowTransaction):
        * storage/DatabaseAuthorizer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27440 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Mon, 5 Nov 2007 16:34:53 +0000 (16:34 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Reviewed by Alp Toker.

        Have getMIMETypeForExtension return a null string when no MIME type is known
        for the extension rather than returning "text/plain".  This prevents plugin data
        being dumped into object elements when plugins are disabled.

        * platform/gtk/MIMETypeRegistryGtk.cpp:
        (WebCore::MIMETypeRegistry::getMIMETypeForExtension):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27439 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Mon, 5 Nov 2007 15:01:02 +0000 (15:01 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Rubber-stamped by Alp Toker.

        Remove buggy "autocomplete" from GtkLauncher as it causes more problems than it solves.

        * GtkLauncher/main.cpp:
        (goToURLBarText):
        (main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27438 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Mon, 5 Nov 2007 14:39:42 +0000 (14:39 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Rubber-stamped by Alp Toker.

        Remove unused m_firstData member from FrameLoaderClientGtk.

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        * WebCoreSupport/FrameLoaderClientGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27437 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-05 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Mon, 5 Nov 2007 14:39:25 +0000 (14:39 +0000)]
2007-11-05  Mark Rowe  <mrowe@apple.com>

        Reviewed by Alp Toker.

        Fix http://bugs.webkit.org/show_bug.cgi?id=15842
        Bug 15842: [Gtk] about:blank doesn't work

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (WebKit::FrameLoaderClient::finishedLoading): Set the encoding on the frame loader to
        get work done that is normally done when the first bit of data is received, even in the
        case of a document with no data (like about:blank).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27436 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
ap [Mon, 5 Nov 2007 14:30:59 +0000 (14:30 +0000)]
    Reviewed by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=15841
        fast/dom/Document/early-document-access.html crashes under GuardMalloc

        * fast/dom/Document/early-document-access.html: Don't check for open() timeout;
        wait for popup window to close as long as it takes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27435 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago bisect-builds doesn't work with nightly build r19992 or newer on Leopard
ddkilzer [Mon, 5 Nov 2007 06:38:23 +0000 (06:38 +0000)]
    bisect-builds doesn't work with nightly build r19992 or newer on Leopard
        <http://bugs.webkit.org/show_bug.cgi?id=15830>

        Reviewed by Timothy.

        Restrict the range of nightly builds used by the bisect-builds script
        based on the version of Safari and the version of Mac OS X being used.

        Mac OS X 10.4: Safari 2.0: r11976 or newer
        Mac OS X 10.4: Safari 3.0: r19992 or newer

        Mac OS X 10.5: Safari 2.0: r19594 or newer
        Mac OS X 10.5: Safari 3.0: r25124 or newer

        * Scripts/bisect-builds:
        (findMacOSXVersion): Added.
        (makeNightlyList): Added argument to provide version of Mac OS X.
        Restrict range of nightly builds based on Safari and Mac OS X versions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27434 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rubber-stamped by Adam Roben.
weinig@apple.com [Mon, 5 Nov 2007 01:37:59 +0000 (01:37 +0000)]
    Rubber-stamped by Adam Roben.

        Rename SecurityOrigin::allowsAccessFrom to SecurityOrigin::canAccess to
        clear up ambiguity.

        * bindings/js/kjs_window.cpp:
        (KJS::Window::isSafeScript):
        * platform/SecurityOrigin.cpp:
        (WebCore::SecurityOrigin::canAccess):
        * platform/SecurityOrigin.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27433 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Land the HTML Editing Toolbar demo from WWDC 2007.
timothy@apple.com [Mon, 5 Nov 2007 00:20:55 +0000 (00:20 +0000)]
    Land the HTML Editing Toolbar demo from WWDC 2007.

        * demos: Added.
        * demos/editingToolbar: Added.
        * demos/editingToolbar/FancyToolbar.css: Added.
        * demos/editingToolbar/FancyToolbar.js: Added.
        * demos/editingToolbar/FancyToolbarImages: Added.
        * demos/editingToolbar/FancyToolbarImages/button.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonLeft.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonMiddle.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonPressed.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonPressedLeft.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonPressedMiddle.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonPressedRight.png: Added.
        * demos/editingToolbar/FancyToolbarImages/buttonRight.png: Added.
        * demos/editingToolbar/FancyToolbarImages/camera.png: Added.
        * demos/editingToolbar/FancyToolbarImages/link.png: Added.
        * demos/editingToolbar/FancyToolbarImages/textAlign.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutline.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineBottom.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineBottomLeft.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineBottomRight.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineLeft.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineRight.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineTop.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineTopLeft.png: Added.
        * demos/editingToolbar/FancyToolbarImages/toolbarOutlineTopRight.png: Added.
        * demos/editingToolbar/content.html: Added.
        * demos/editingToolbar/images: Added.
        * demos/editingToolbar/images/safari.png: Added.
        * demos/editingToolbar/index.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27432 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam.
timothy@apple.com [Mon, 5 Nov 2007 00:09:54 +0000 (00:09 +0000)]
    Reviewed by Adam.

        Bug 15834: There are many subtle bugs in the Styles pane of the Web Inspector
        http://bugs.webkit.org/show_bug.cgi?id=15834

        - Broke up DocumentPanel and added three SidebarPane sub-classes.
        - Fixed many Style pane bugs, including:
          * Poor handling of duplicate properties in the same rule. Some of this can't be
            fixed since we can't only get the "winning" value for duplicate properties.
            So we should only show one entry per unique property name.
          * Computed style does not show font shorthand sub-properties if 'font' was used.
          * Property priority was broken, the wrong properties were crossed out.
          * The 'border' shorthand shows null for the shorthand value.
          * Shorthands didn't show their priority (e.g. !important).
          * HSL and HTML hex colors didn't have preview swatch blocks.
          * Code refactoring, making it easier to reuse for console.log later.

        * page/inspector/DocumentPanel.js: Move sidebar pane code to three seperate
          classes in new files.
        * page/inspector/MetricsSidebarPane.js: Added.
        * page/inspector/Panel.js: Remove an InspectorController.log() call.
        * page/inspector/PropertiesSection.js: Add the section to the TreeOutline.
          So TreeElements can access properties on their section.
        * page/inspector/PropertiesSidebarPane.js: Added.
        * page/inspector/SidebarPane.js: Remove the explicit asignment of the
          onexpand and oncollapse to null. These were hiding prototypes.
        * page/inspector/StylesSidebarPane.js: Added.
        * page/inspector/inspector.html: Include the new script files.
        * page/inspector/treeoutline.js: If a null representedObject is passed
          in just use a empty object.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27431 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
kmccullo [Sun, 4 Nov 2007 22:36:32 +0000 (22:36 +0000)]
    Reviewed by Maciej.

        - This patch involves several changes, all of them were noticed that
        they were needed by the work being done in WebKit to get Drosera and
        WebKit working together on Windows.
        - The changes are:
        1) Added a debugger console for output messages.
        2) Drosera now listens for the WebKit server (before, the server would
        have to be running before Drosera was started.)
        3) Fixed a bug where the WebView started out as 0x0 pixels.
        4) Fixed a bug when there is no scope.
        5) Added the HTML, JS, and CSS to the project file to make them easy to
        find.
        6) Made the ServerConnection functions virtual.

        * Drosera/win/DebuggerClient.cpp: This is part of how Drosera listens
        for the WebKit server.
        (DebuggerClient::DebuggerClient):
        (DebuggerClient::~DebuggerClient):
        (DebuggerClient::didFinishLoadForFrame):
        (DebuggerClient::serverConnected):
        (DebuggerClient::attemptToCreateServerConnection):
        * Drosera/win/DebuggerClient.h: Ditto.
        * Drosera/win/DebuggerDocumentPlatform.cpp: Fixed a bug when there is no
        scope.
        (DebuggerDocument::getPlatformLocalScopeVariableNamesForCallFrame):
        (DebuggerDocument::platformValueForScopeVariableNamed):
        * Drosera/win/Drosera.cpp:
        (_tWinMain): Added a console in debug for output messages.
        (Drosera::Drosera): Listen for server.
        (Drosera::initUI): The server now Initializes COM.
        (Drosera::serverConnected): Part of the listening for the server.
        (Drosera::attemptToCreateServerConnection): Ditto.
        * Drosera/win/Drosera.h: New interface for listening for the server.
        * Drosera/win/Drosera.vcproj/Drosera.vcproj: Added HTML, JS, and CSS
        files to the VS project.
        * Drosera/win/ServerConnection.cpp: Part of listening for the server
        connection.
        (ServerConnection::ServerConnection):
        (ServerConnection::attemptToCreateServerConnection):
        * Drosera/win/ServerConnection.h: Ditto and virtualized the Interface
        methods.
        (ServerConnection::serverConnected):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27430 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoAdd svn:ignore for vcproj temp files under FindSafari.
weinig [Sun, 4 Nov 2007 22:30:20 +0000 (22:30 +0000)]
Add svn:ignore for vcproj temp files under FindSafari.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27429 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
ap [Sun, 4 Nov 2007 20:27:16 +0000 (20:27 +0000)]
    Reviewed by Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=15832
        fast/dom/gc-10.html crashes when run alone

        Check for !done before using objects that can be already deallocated.

        * DumpRenderTree/mac/EditingDelegate.mm:
        (-[EditingDelegate webView:shouldBeginEditingInDOMRange:]):
        (-[EditingDelegate webView:shouldEndEditingInDOMRange:]):
        (-[EditingDelegate webView:shouldInsertNode:replacingDOMRange:givenAction:]):
        (-[EditingDelegate webView:shouldInsertText:replacingDOMRange:givenAction:]):
        (-[EditingDelegate webView:shouldDeleteDOMRange:]):
        (-[EditingDelegate webView:shouldChangeSelectedDOMRange:toDOMRange:affinity:stillSelecting:]):
        (-[EditingDelegate webView:shouldApplyStyle:toElementsInDOMRange:]):
        (-[EditingDelegate webView:shouldChangeTypingStyle:toStyle:]):
        (-[EditingDelegate webViewDidBeginEditing:]):
        (-[EditingDelegate webViewDidChange:]):
        (-[EditingDelegate webViewDidEndEditing:]):
        (-[EditingDelegate webViewDidChangeTypingStyle:]):
        (-[EditingDelegate webViewDidChangeSelection:]):
        * DumpRenderTree/mac/FrameLoadDelegate.mm:
        (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]):
        (-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
        (-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]):
        (-[FrameLoadDelegate webView:didFinishLoadForFrame:]):
        (-[FrameLoadDelegate webView:didFailLoadWithError:forFrame:]):
        (-[FrameLoadDelegate webView:windowScriptObjectAvailable:]):
        (-[FrameLoadDelegate webView:didClearWindowObject:forFrame:]):
        (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
        (-[FrameLoadDelegate webView:didReceiveServerRedirectForProvisionalLoadForFrame:]):
        (-[FrameLoadDelegate webView:didReceiveIcon:forFrame:]):
        (-[FrameLoadDelegate webView:didChangeLocationWithinPageForFrame:]):
        (-[FrameLoadDelegate webView:willPerformClientRedirectToURL:delay:fireDate:forFrame:]):
        (-[FrameLoadDelegate webView:didCancelClientRedirectForFrame:]):
        (-[FrameLoadDelegate webView:willCloseFrame:]):
        (-[FrameLoadDelegate webView:didFinishDocumentLoadForFrame:]):
        (-[FrameLoadDelegate webView:didHandleOnloadEventsForFrame:]):
        * DumpRenderTree/mac/ResourceLoadDelegate.mm:
        (-[ResourceLoadDelegate webView:identifierForInitialRequest:fromDataSource:]):
        (-[ResourceLoadDelegate webView:resource:willSendRequest:redirectResponse:fromDataSource:]):
        (-[ResourceLoadDelegate webView:resource:didReceiveResponse:fromDataSource:]):
        (-[ResourceLoadDelegate webView:resource:didFinishLoadingFromDataSource:]):
        (-[ResourceLoadDelegate webView:resource:didFailLoadingWithError:fromDataSource:]):
        (-[ResourceLoadDelegate webView:resource:willCacheResponse:fromDataSource:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27428 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix build when spaces appear in the WebKit source path.
ddkilzer [Sun, 4 Nov 2007 16:46:50 +0000 (16:46 +0000)]
    Fix build when spaces appear in the WebKit source path.

        Reviewed by Mark Rowe.

        * bindings/scripts/IDLParser.pm: Use safer open() method which lists
        arguments individually and prevents the need to work around spaces
        in the path.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27427 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago CIA has changed its hostname from cia.navi.cx to cia.vc
ddkilzer [Sun, 4 Nov 2007 16:13:24 +0000 (16:13 +0000)]
    CIA has changed its hostname from cia.navi.cx to cia.vc
        <http://bugs.webkit.org/show_bug.cgi?id=15829>

        Reviewed by Mitz.

        * contact.html: Updated links to CIA.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27426 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-04 Alp Toker <alp@atoker.com>
alp [Sun, 4 Nov 2007 15:03:13 +0000 (15:03 +0000)]
2007-11-04  Alp Toker  <alp@atoker.com>

        Reviewed by Alexey Proskuryakov.

        Fix a crash when no clipboard text is available

        * platform/gtk/PasteboardGtk.cpp:
        (WebCore::Pasteboard::plainText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27425 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-04 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Sun, 4 Nov 2007 09:48:02 +0000 (09:48 +0000)]
2007-11-04  Mark Rowe  <mrowe@apple.com>

        Build fix.  Don't use Carbon.h as the prefix header as it triggers
        warnings that would otherwise be suppressed due to it being a system header.

        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27424 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * hosted/sunspider.html: Tweaked formatting a bit.
darin@apple.com [Sun, 4 Nov 2007 08:58:15 +0000 (08:58 +0000)]
    * hosted/sunspider.html: Tweaked formatting a bit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27423 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
darin@apple.com [Sun, 4 Nov 2007 08:28:22 +0000 (08:28 +0000)]
    Reviewed by Maciej.

        - http://bugs.webkit.org/show_bug.cgi?id=15826
          optimize opcode loop and case insensitive ASCII compares for a 30% speedup

        SunSpider says it's 2.6% faster overall, 32.5% in the regular expression tests.

        * pcre/pcre_internal.h: Added OP_ASCII_CHAR and OP_ASCII_LETTER_NC.

        * pcre/pcre_compile.c:
        (find_fixedlength): Added cases for OP_ASCII_CHAR and OP_ASCII_LETTER_NC. Also
        added OP_NOT since there was no reason it should not be in here.
        (could_be_empty_branch): Ditto.
        (compile_branch): Streamlined all the single-character cases; there was a bit of
        duplicate code. Added cases for OP_ASCII_CHAR and OP_ASCII_LETTER_NC as needed.
        But in particular, compile to those opcodes when the single character match is
        ASCII.
        (find_firstassertedchar): Added cases for OP_ASCII_CHAR and OP_ASCII_LETTER_NC.

        * pcre/pcre_exec.c: (match): Removed the "min", "minimize", and "op" fields from
        the matchframe, after I discovered that none of them needed to be saved and restored
        across recursive match calls. Also eliminated the ignored result field from the
        matchframe, since I discovered that rrc ("recursive result code") was already the
        exact same thing. Moved the handling of opcodes higher than OP_BRA into the default
        statement of the switch instead of doing them before the switch. This removes a
        branch from each iteration of the opcode interpreter, just as removal of "op"
        removed at least one store from each iteration. Last, but not least, add the
        OP_ASCII_CHAR and OP_ASCII_LETTER_NC functions. Neither can ever match a
        surrogate pair and the letter case can be handled efficiently.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27422 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * pcre/pcre_exec.c: (match): Try to fix the Windows build by removing unreach...
darin@apple.com [Sun, 4 Nov 2007 07:54:56 +0000 (07:54 +0000)]
    * pcre/pcre_exec.c: (match): Try to fix the Windows build by removing unreachable code.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27421 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - fix non-Mac builds; remove some more unused PCRE stuff
darin@apple.com [Sun, 4 Nov 2007 06:18:31 +0000 (06:18 +0000)]
    - fix non-Mac builds; remove some more unused PCRE stuff

        * pcre/pcre_compile.c:
        (compile_branch): Removed branch chain and some unused ESC values.
        (compile_regex): Ditto.
        (jsRegExpCompile): Ditto.
        * pcre/pcre_exec.c:
        (match): Removed unused branch targets. Don't use macros any more.
        (jsRegExpExecute): More of the same.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Update for removed files.
        * JavaScriptCore.xcodeproj/project.pbxproj: Ditto.
        * pcre/pcre.pri: Ditto.

        * pcre/MERGING: Removed.
        * pcre/pcre_fullinfo.c: Removed.
        * pcre/pcre_get.c: Removed.
        * pcre/pcre_internal.h:
        * pcre/ucp.h: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27420 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
darin@apple.com [Sun, 4 Nov 2007 05:22:44 +0000 (05:22 +0000)]
JavaScriptCore:

        Reviewed by Maciej.

        - http://bugs.webkit.org/show_bug.cgi?id=15821
          remove unused PCRE features for speed

        A first step toward removing the PCRE features we don't use.
        This gives a 0.8% speedup on SunSpider, and a 6.5% speedup on
        the SunSpider regular expression test.

        Replaced the public interface with one that doesn't use the
        name PCRE. Removed code we don't need for JavaScript and various
        configurations we don't use. This is in preparation for still
        more changes in the future. We'll probably switch to C++ and
        make some even more significant changes to the regexp engine
        to get some additional speed.

        There's probably additional unused stuff that I haven't
        deleted yet.

        This does mean that our PCRE is now a fork, but I think that's
        not really a big deal.

        * JavaScriptCore.exp: Remove the 5 old entry points and add
        the 3 new entry points for WebCore's direct use of the regular
        expression engine.

        * kjs/config.h: Remove the USE(PCRE16) define. I decided to flip
        its sense and now there's a USE(POSIX_REGEX) instead, which should
        probably not be set by anyone. Maybe later we'll just get rid of it
        altogether.

        * kjs/regexp.h:
        * kjs/regexp.cpp:
        (KJS::RegExp::RegExp): Switch to new jsRegExp function names and
        defines. Cut down on the number of functions used.
        (KJS::RegExp::~RegExp): Ditto.
        (KJS::RegExp::match): Ditto.

        * pcre/dftables.c: (main): Get rid of ctype_letter and ctype_meta,
        which are unused.

        * pcre/pcre-config.h: Get rid of EBCIDIC, PCRE_DATA_SCOPE, const,
        size_t, HAVE_STRERROR, HAVE_MEMMOVE, HAVE_BCOPY, NEWLINE,
        POSIX_MALLOC_THRESHOLD, NO_RECURSE, SUPPORT_UCP, SUPPORT_UTF8,
        and JAVASCRIPT. These are all no longer configurable in our copy
        of the library.

        * pcre/pcre.h: Remove the macro-based kjs prefix hack, the PCRE
        version macros, PCRE_UTF16, the code to set up PCRE_DATA_SCOPE,
        the include of <stdlib.h>, and most of the constants and
        functions defined in this header. Changed the naming scheme to
        use a JSRegExp prefix rather than a pcre prefix. In the future,
        we'll probably change this to be a C++ header.

        * pcre/pcre_compile.c: Removed all unused code branches,
        including many whole functions and various byte codes.
        Kept changes outside of removal to a minimum.
        (check_escape):
        (first_significant_code):
        (find_fixedlength):
        (find_recurse):
        (could_be_empty_branch):
        (compile_branch):
        (compile_regex):
        (is_anchored):
        (is_startline):
        (find_firstassertedchar):
        (jsRegExpCompile): Renamed from pcre_compile2 and changed the
        parameters around a bit.
        (jsRegExpFree): Added.

        * pcre/pcre_exec.c: Removed many unused opcodes and variables.
        Also started tearing down the NO_RECURSE mechanism since it's
        now the default. In some cases there were things in the explicit
        frame that could be turned into plain old local variables and
        other small like optimizations.
        (pchars):
        (match_ref):
        (match): Changed parameters quite a bit since it's now not used
        recursively.
        (jsRegExpExecute): Renamed from pcre_exec.

        * pcre/pcre_internal.h: Get rid of PCRE_DEFINITION, PCRE_SPTR,
        PCRE_IMS, PCRE_ICHANGED, PCRE_NOPARTIAL, PCRE_STUDY_MAPPED,
        PUBLIC_OPTIONS, PUBLIC_EXEC_OPTIONS, PUBLIC_DFA_EXEC_OPTIONS,
        PUBLIC_STUDY_OPTIONS, MAGIC_NUMBER, 16 of the opcodes,
        _pcre_utt, _pcre_utt_size, _pcre_try_flipped, _pcre_ucp_findprop,
        and _pcre_valid_utf8. Also moved pcre_malloc and pcre_free here.

        * pcre/pcre_maketables.c: Changed to only compile in dftables.
        Also got rid of many of the tables that we don't use.

        * pcre/pcre_tables.c: Removed the unused Unicode property tables.

        * pcre/pcre_ucp_searchfuncs.c: Removed everything except for
        _pcre_ucp_othercase.

        * pcre/pcre_xclass.c: (_pcre_xclass): Removed uneeded support
        for classes based on Unicode properties.

        * wtf/FastMallocPCRE.cpp: Removed unused bits. It would be good
        to eliminate this completely, but we need the regular expression
        code to be C++ first.

        * pcre/pcre_fullinfo.c:
        * pcre/pcre_get.c:
        * pcre/ucp.h:
        Files that are no longer needed. I didn't remove them with this
        check-in, because I didn't want to modify all the project files.

WebCore:

        Reviewed by Maciej.

        - WebCore part of http://bugs.webkit.org/show_bug.cgi?id=15821
          remove unused PCRE features for speed

        * page/Frame.cpp: (WebCore::Frame::matchLabelsAgainstElement):
        * page/mac/FrameMac.mm: (WebCore::Frame::matchLabelsAgainstElement):
        Remove use of "[[:digit:]]" syntax. This hasn't worked for some time.
        Use "\d" instead.

        * platform/RegularExpression.h: Remove the unused cap function. We can
        add it back later if we find we need it.
        * platform/RegularExpression.cpp:
        (WebCore::RegularExpression::Private::compile): Update for JavaScriptCore
        regular expression entry point changes.
        (WebCore::RegularExpression::Private::~Private): Ditto.
        (WebCore::RegularExpression::match): Remove the code to set PCRE_NOTBOL.
        This means that regular expressions with metacharactesr like ^ in them
        won't work any more with non-whole-string searches, but we don't use
        any regular expressions like that.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27419 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoUpdate the link stubs to match the current build, and fix coding style issues.
kevino [Sun, 4 Nov 2007 05:15:26 +0000 (05:15 +0000)]
Update the link stubs to match the current build, and fix coding style issues.

Reviewed by Mark Rowe.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27418 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build-fix.
weinig@apple.com [Sun, 4 Nov 2007 04:17:58 +0000 (04:17 +0000)]
    Build-fix.

        * UserObjectImp.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27417 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam.
mjs [Sun, 4 Nov 2007 03:37:21 +0000 (03:37 +0000)]
    Reviewed by Sam.

        - remove NaN check from JSImmediate::fromDouble for 0.5% SunSpider speedup

        It turns out that doing this check costs more than it saves.

        * kjs/JSImmediate.h:
        (KJS::JSImmediate::fromDouble):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27416 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agowx <-> WebKit conversions for IntPoint, IntRect and FloatRect
kevino [Sun, 4 Nov 2007 02:21:09 +0000 (02:21 +0000)]
wx <-> WebKit conversions for IntPoint, IntRect and FloatRect

Reviewed by Mark Rowe.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27415 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agowx <-> WebKit data type conversions for Pen and Color.
kevino [Sun, 4 Nov 2007 02:05:51 +0000 (02:05 +0000)]
wx <-> WebKit data type conversions for Pen and Color.

Reviewed by Darin Adler.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27414 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
weinig@apple.com [Sun, 4 Nov 2007 01:08:04 +0000 (01:08 +0000)]
JavaScriptCore:

        Reviewed by Oliver.

        Remove dummy variable from ClassInfo reducing the size of the struct by 1 word.
        The variable had been kept around for binary compatibility, but since nothing
        else is there is no point in continuing to keep it around.

        * API/JSCallbackConstructor.cpp:
        (KJS::):
        * API/JSCallbackFunction.cpp:
        (KJS::):
        * API/JSCallbackObject.cpp:
        (KJS::):
        * bindings/objc/objc_runtime.mm:
        * bindings/runtime_array.cpp:
        * bindings/runtime_object.cpp:
        * kjs/array_instance.cpp:
        (KJS::):
        * kjs/array_object.cpp:
        (KJS::):
        * kjs/bool_object.cpp:
        * kjs/date_object.cpp:
        (KJS::):
        * kjs/error_object.cpp:
        * kjs/function.cpp:
        (KJS::):
        * kjs/internal.cpp:
        (KJS::):
        * kjs/lookup.h:
        * kjs/math_object.cpp:
        * kjs/number_object.cpp:
        * kjs/object.h:
        * kjs/regexp_object.cpp:
        * kjs/string_object.cpp:
        (KJS::):

WebCore:

        Reviewed by Oliver.

        Remove dummy variable from ClassInfo reducing the size of the struct by 1 word.
        The variable had been kept around for binary compatibility, but since nothing
        else is there is no point in continuing to keep it around.

        * bindings/js/JSDOMExceptionConstructor.cpp:
        (WebCore::):
        * bindings/js/JSHTMLInputElementBase.cpp:
        (WebCore::):
        * bindings/js/JSNamedNodesCollection.cpp:
        (WebCore::):
        * bindings/js/JSXMLHttpRequest.cpp:
        (KJS::):
        * bindings/js/JSXSLTProcessor.cpp:
        (KJS::):
        * bindings/js/kjs_css.cpp:
        (WebCore::):
        * bindings/js/kjs_events.cpp:
        (WebCore::):
        * bindings/js/kjs_navigator.cpp:
        (KJS::):
        * bindings/js/kjs_window.cpp:
        (KJS::):
        * bindings/scripts/CodeGeneratorJS.pm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27413 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - Updated testkjs results to make the build bots green until we
kmccullo [Sat, 3 Nov 2007 23:52:57 +0000 (23:52 +0000)]
    - Updated testkjs results to make the build bots green until we
        can fix the tests that are failing.  The new failures are in DST.

        * tests/mozilla/expected.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27412 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
mjs [Sat, 3 Nov 2007 22:58:41 +0000 (22:58 +0000)]
    Reviewed by Darin.

        - Add compare mode and a description of the benchmark to the hosted version.

        * hosted/sunspider-results.html: Add compare mode - you can just paste in another results URL.
        * hosted/sunspider.html: Added intro text.
        * make-hosted: Copy new file.
        * resources/sunspider-compare-results.js: Made more reusable by putting the implementation in
        a function.
        * resources/sunspider-standalone-compare.js: Added. Calls the function in the way standalone mode
        expects.
        * sunspider-compare-results: Use sunspider-standalone-compare.js.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27411 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam.
mjs [Sat, 3 Nov 2007 21:13:36 +0000 (21:13 +0000)]
    Reviewed by Adam.

        - don't print the var twice for ForInNodes with a var declaration

        * kjs/nodes2string.cpp:
        (KJS::ForInNode::streamTo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27410 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rubber stamped by Adam.
mjs [Sat, 3 Nov 2007 20:54:57 +0000 (20:54 +0000)]
    Rubber stamped by Adam.

        - Turn off deprecated function warnings for TestNetscapePlugin because Carbon.h triggers them

        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27409 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * pcre/pcre_compile.c: (check_escape): Windows build fix. Get rid of
darin@apple.com [Sat, 3 Nov 2007 16:44:00 +0000 (16:44 +0000)]
    * pcre/pcre_compile.c: (check_escape): Windows build fix. Get rid of
        C-incompatible declaration.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27408 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-11-03 Mark Rowe <mrowe@apple.com>
mrowe@apple.com [Sat, 3 Nov 2007 16:42:26 +0000 (16:42 +0000)]
2007-11-03  Mark Rowe  <mrowe@apple.com>

        Gtk build fix.

        * kjs/nodes.cpp:  Add missing include.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27407 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
darin@apple.com [Sat, 3 Nov 2007 16:40:32 +0000 (16:40 +0000)]
JavaScriptCore:

        Reviewed by Maciej.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15814
          <rdar://problem/5536644> fast/js/kde/encode_decode_uri.html fails

        These changes cause us to match the JavaScript specification and pass the
        fast/js/kde/encode_decode_uri.html test.

        * kjs/function.cpp: (KJS::encode): Call the UTF-8 string conversion in its
        new strict mode, throwing an exception if there are malformed UTF-16 surrogate
        pairs in the text.

        * kjs/ustring.h: Added a strict version of the UTF-8 string conversion.
        * kjs/ustring.cpp:
        (KJS::decodeUTF8Sequence): Removed code to disallow U+FFFE and U+FFFF; while
        those might be illegal in some sense, they aren't supposed to get any special
        handling in the place where this function is currently used.
        (KJS::UString::UTF8String): Added the strictness.

LayoutTests:

        Reviewed by Maciej.

        - updates for http://bugs.webkit.org/show_bug.cgi?id=15814
          <rdar://problem/5536644> fast/js/kde/encode_decode_uri.html fails

        * fast/js/kde/resources/encode_decode_uri.js: Rewrote the test to cover edges better,
        and use the should functions in a way that makes failures easier to understand.
        * fast/js/kde/encode_decode_uri-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27406 268f45cc-cd09-0410-ab3c-d52691b4dbfc