WebKit-https.git
17 years ago - updated pixel results for tests with results that changed recently
darin [Sat, 22 Oct 2005 22:55:33 +0000 (22:55 +0000)]
    - updated pixel results for tests with results that changed recently

        * editing/selection/3690703-2-expected.checksum:
        * editing/selection/3690703-2-expected.png:
        * editing/selection/3690703-expected.checksum:
        * editing/selection/3690703-expected.png:
        * editing/selection/3690719-expected.checksum:
        * editing/selection/3690719-expected.png:
        * fast/forms/001-expected.checksum:
        * fast/forms/001-expected.png:
        * fast/forms/form-element-geometry-expected.checksum:
        * fast/forms/form-element-geometry-expected.png:
        * fast/forms/formmove3-expected.checksum:
        * fast/forms/formmove3-expected.png:
        * fast/forms/input-value-expected.checksum:
        * fast/forms/input-value-expected.png:
        * fast/selectors/064-expected.checksum:
        * fast/selectors/064-expected.png:

        - added pixel results for a bunch of tests that, for whatever reason, didn't have them

        * dom/xhtml/level2/html/HTMLBaseElement01-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBaseElement01-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBaseElement02-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBaseElement02-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBodyElement01-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBodyElement01-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBodyElement02-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBodyElement02-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBodyElement03-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBodyElement03-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBodyElement04-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBodyElement04-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBodyElement05-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBodyElement05-expected.png: Added.
        * dom/xhtml/level2/html/HTMLBodyElement06-expected.checksum: Added.
        * dom/xhtml/level2/html/HTMLBodyElement06-expected.png: Added.
        * dom/xhtml/level2/html/body01-expected.checksum: Added.
        * dom/xhtml/level2/html/body01-expected.png: Added.
        * dom/xhtml/level2/html/frame-expected.checksum: Added.
        * dom/xhtml/level2/html/frame-expected.png: Added.
        * dom/xhtml/level2/html/iframe-expected.checksum: Added.
        * dom/xhtml/level2/html/iframe-expected.png: Added.
        * dom/xhtml/level3/core/canonicalform08-expected.checksum: Added.
        * dom/xhtml/level3/core/canonicalform08-expected.png: Added.
        * dom/xhtml/level3/core/canonicalform09-expected.checksum: Added.
        * dom/xhtml/level3/core/canonicalform09-expected.png: Added.
        * dom/xhtml/level3/core/documentgetinputencoding03-expected.checksum: Added.
        * dom/xhtml/level3/core/documentgetinputencoding03-expected.png: Added.
        * dom/xhtml/level3/core/documentgetxmlencoding03-expected.checksum: Added.
        * dom/xhtml/level3/core/documentgetxmlencoding03-expected.png: Added.
        * dom/xhtml/level3/core/entitygetinputencoding02-expected.checksum: Added.
        * dom/xhtml/level3/core/entitygetinputencoding02-expected.png: Added.
        * dom/xhtml/level3/core/entitygetxmlversion02-expected.checksum: Added.
        * dom/xhtml/level3/core/entitygetxmlversion02-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri05-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri05-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri07-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri07-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri09-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri09-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri10-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri10-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri11-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri11-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri15-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri15-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri17-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri17-expected.png: Added.
        * dom/xhtml/level3/core/nodegetbaseuri18-expected.checksum: Added.
        * dom/xhtml/level3/core/nodegetbaseuri18-expected.png: Added.
        * dom/xhtml/level3/core/nodelookupnamespaceuri01-expected.checksum: Added.
        * dom/xhtml/level3/core/nodelookupnamespaceuri01-expected.png: Added.
        * dom/xhtml/level3/core/nodelookupprefix19-expected.checksum: Added.
        * dom/xhtml/level3/core/nodelookupprefix19-expected.png: Added.
        * fast/forms/button-sizes-expected.checksum: Added.
        * fast/forms/button-sizes-expected.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10916 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agocheck in revised results for forms minwidth test
hyatt [Sat, 22 Oct 2005 21:32:27 +0000 (21:32 +0000)]
check in revised results for forms minwidth test

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLand updated editing layout tests for buttons.
hyatt [Sat, 22 Oct 2005 18:24:46 +0000 (18:24 +0000)]
Land updated editing layout tests for buttons.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by darin
hyatt [Sat, 22 Oct 2005 18:19:27 +0000 (18:19 +0000)]
    Reviewed by darin

        * khtml/css/cssstyleselector.cpp:
Add support for the GrayText CSS2 system color to represent
the disabled text color for controls.

        * khtml/css/html4.css:
Make sure that the default cursor is set for all types of buttons
and not just the HTML4 button.

Set the foreground color for buttons using the CSS2 system color.
Set the disabled text color using the CSS2 system color.

Move the default padding for buttons into the CSS file.  It will
get overridden by Aqua buttons but be honored by all other types
of buttons.  This allows the author to override it as well when
customizing.

        * khtml/rendering/render_block.h:
        Make paintChildren virtual so that RenderButton can push a clip
and pop a clip when painting children.

* khtml/rendering/render_box.cpp:
        (RenderBox::paintBoxDecorations):
        Let the theme now decide whether the border/background should
still be painted after the theme has painted.  This allows buttons
to paint custom borders but still use a pretty OS X background for
the button face.

* khtml/rendering/render_button.cpp:
        (khtml::RenderButton::paintChildren):
        Overridden to push/pop a clip so that the contents of a button
don't ever spill out.

* khtml/rendering/render_button.h:
        Added paintChildren method.

* khtml/rendering/render_style.h:
        (khtml::BorderData::hasBorder):
        Fixed a regression in hasBorder.  The border-image logic was
inverted, resulting in hasBorder being true for every element!

* khtml/rendering/render_theme.cpp:
        (khtml::RenderTheme::paint):
        Patched to return a boolean indicating whether or not the
border/background should be painted.

* khtml/rendering/render_theme.h:
        * khtml/rendering/render_theme_mac.h:
        * khtml/rendering/render_theme_mac.mm:
        (khtml::RenderThemeMac::adjustRepaintRect):
        Forgot to include radio and button in this method so that repaint
rects would be correct.

(khtml::RenderThemeMac::paintCheckbox):
        (khtml::RenderThemeMac::paintRadio):
        Adjusted to return the boolean to indicate that painting of
border/background should not occur.

(khtml::RenderThemeMac::adjustButtonStyle):
        Added support for border/background/colors.  Also locked
white-space to nowrap for Aqua buttons to avoid line wrapping.

(khtml::RenderThemeMac::setButtonCellState):
(khtml::RenderThemeMac::paintButton):
More support for custom border/background/color.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoPut back in the hasRef() check for jumping to anchors to fix the regression (bug...
hyatt [Sat, 22 Oct 2005 04:38:49 +0000 (04:38 +0000)]
Put back in the hasRef() check for jumping to anchors to fix the regression (bug 5457)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by darin.
ggaren [Sat, 22 Oct 2005 01:42:04 +0000 (01:42 +0000)]
    Reviewed by darin.

        WebKit side of the fix for <rdar://problem/4184719> window.print() followed by
        window.close() causes world leak

        No test case added because I have another reviewed patch that will include a test
        for this bug as well as many others.

        Under some conditions, [WebDataSource stopLoading] did not set [WebDataSource isLoading]
        to false, so the didFInishLoad delegates never fired.

        The reason isLoading didn't become false was that the tokenizer was still running.
        The fix here is to move the call to [WebCoreBridge stopLoading] above the early return
        inside [WebDataSource stopLoading] -- since the tokenizer may still be running even if the
        loader is finished loading -- and then to call [WebFrame _checkLoadComplete] to
        give the frame a chance to fire its delegates.

        * WebView.subproj/WebDataSource.m:
        (-[WebDataSource _stopLoadingInternal]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by darin.
ggaren [Sat, 22 Oct 2005 01:41:36 +0000 (01:41 +0000)]
    Reviewed by darin.

        WebCore side of fix for <rdar://problem/4184719> window.print() followed by window.close()
        causes world leak

        No test case added because I have another reviewed patch that will include a test
        for this bug as well as many others.

        Under some conditions, [WebDataSource stopLoading] did not set [WebDataSource isLoading]
        to false, so the didFInishLoad delegates never fired.

        The reason isLoading didn't become false was that the tokenizer was still running.
        The fix here is to add a call to Tokenizer::stopParsing inside HTMLPart::stopLoading.

        Since the WebKit side of the fix eliminates an early return that used to protect
        HTMLPart::stopLoading from multiple calls, I added an early return inside
        HTMLPart::stopLoading.

        I also merged stopped and stopLoading, and merged the way XMLTokenizer and HTMLTokenizer
        track whether they're in a stopped state, because anything less would be uncivilized.

        * khtml/html/htmltokenizer.cpp:
        (khtml::HTMLTokenizer::HTMLTokenizer):
        (khtml::HTMLTokenizer::write):
        (khtml::HTMLTokenizer::stopParsing):
        (khtml::HTMLTokenizer::processToken):
        * khtml/html/htmltokenizer.h:
        * khtml/khtml_part.cpp:
        (KHTMLPart::stopLoading):
        (KHTMLPart::endIfNotLoading):
        (KHTMLPart::stop):
        * khtml/xml/xml_tokenizer.cpp:
        (khtml::Tokenizer::Tokenizer):
        (khtml::Tokenizer::finishedParsing):
        (khtml::XMLTokenizer::XMLTokenizer):
        (khtml::XMLTokenizer::endElementNs):
        (khtml::XMLTokenizer::characters):
        (khtml::XMLTokenizer::error):
        (khtml::XMLTokenizer::processingInstruction):
        (khtml::XMLTokenizer::cdataBlock):
        (khtml::XMLTokenizer::comment):
        (khtml::XMLTokenizer::stopParsing):
        * khtml/xml/xml_tokenizer.h:
        (khtml::Tokenizer::stopParsing):
        * kwq/KWQSignalStubs.mm:
        (Loader::requestStarted):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago*** empty log message ***
bdakin [Sat, 22 Oct 2005 01:10:32 +0000 (01:10 +0000)]
*** empty log message ***

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago*** empty log message ***
bdakin [Sat, 22 Oct 2005 00:11:19 +0000 (00:11 +0000)]
*** empty log message ***

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Eric, landed by Maciej.
mjs [Sat, 22 Oct 2005 00:06:43 +0000 (00:06 +0000)]
    Reviewed by Eric, landed by Maciej.

- fixed xmlhttprequest test to not depend on property enumeration order

        * fast/js/xmlhttprequest-get-expected.txt:
        * fast/js/xmlhttprequest-get.xhtml:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - temporarily reverted test changes corresponding to reverted changes
mjs [Fri, 21 Oct 2005 23:35:27 +0000 (23:35 +0000)]
    - temporarily reverted test changes corresponding to reverted changes

        * dom/xhtml/level3/core/nodeissamenode01-expected.txt:
        * dom/xhtml/level3/core/nodeissamenode02-expected.txt:
        * dom/xhtml/level3/core/nodeissamenode03-expected.txt:
        * dom/xhtml/level3/core/nodeissamenode04-expected.txt:
        * dom/xhtml/level3/core/nodeissamenode05-expected.txt:
        * dom/xhtml/level3/core/nodeissamenode06-expected.txt:
        * dom/xhtml/level3/core/nodeissamenode10-expected.txt:
        * fast/js/for-in-expected.txt: Removed.
        * fast/js/for-in.html: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoforgot to add this file in my previous commit
vicki [Fri, 21 Oct 2005 23:13:31 +0000 (23:13 +0000)]
forgot to add this file in my previous commit

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by John.
vicki [Fri, 21 Oct 2005 23:11:04 +0000 (23:11 +0000)]
    Reviewed by John.

- fixed <rdar://problem/4304213> sometimes have to click twice to jump to top, on href=""

        * manual-tests/linkjump-1.html: Added.
        * khtml/khtml_part.cpp:
        (KHTMLPart::gotoAnchor): remove check for m_url.hasRef() - encodedHtmlRef and gotoAnchor handle empty
strings for scrolling to the top of the document

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #:
bdakin [Fri, 21 Oct 2005 22:55:27 +0000 (22:55 +0000)]
Bug #:
Submitted by:
Reviewed by:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10896 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #:
bdakin [Fri, 21 Oct 2005 22:49:21 +0000 (22:49 +0000)]
Bug #:
Submitted by:
Reviewed by:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago No test cases added.
sullivan [Fri, 21 Oct 2005 22:45:53 +0000 (22:45 +0000)]
    No test cases added.

        * khtml/rendering/render_theme_mac.mm:
        (khtml::RenderThemeMac::setButtonPaddingFromControlSize):
        fix deployment build by adding default: statement to button-size switch statement

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #:
bdakin [Fri, 21 Oct 2005 22:04:17 +0000 (22:04 +0000)]
Bug #:
Submitted by:
Reviewed by:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoadd new test of html4 button font sizes.
hyatt [Fri, 21 Oct 2005 21:43:33 +0000 (21:43 +0000)]
add new test of html4 button font sizes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10890 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoUpdate the fast layout tests to have the new buttons.
hyatt [Fri, 21 Oct 2005 21:41:54 +0000 (21:41 +0000)]
Update the fast layout tests to have the new buttons.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10889 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Land the new "NSView-less" buttons. With this landing both
hyatt [Fri, 21 Oct 2005 21:27:51 +0000 (21:27 +0000)]
Land the new "NSView-less" buttons.  With this landing both
<input type=button> and <button> will now look like OS X-style
widgets.  Customization of background and border is not yet
enabled, but this is still a big improvement over the old
AppKit widgets.

        Reviewed by mjs

        * ForwardingHeaders/qpushbutton.h: Removed.
        * WebCore.xcodeproj/project.pbxproj:
        * khtml/css/cssstyleselector.cpp:
        (khtml::CSSStyleSelector::adjustRenderStyle):
        * khtml/css/cssstyleselector.h:
        (khtml::CSSStyleSelector::paintMetrics):
        * khtml/css/cssvalues.in:
        * khtml/css/html4.css:
        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLButtonElementImpl::createRenderer):
        (DOM::HTMLInputElementImpl::click):
        (DOM::HTMLInputElementImpl::createRenderer):
        * khtml/html/html_formimpl.h:
        (DOM::HTMLInputElementImpl::isTextButton):
        * khtml/rendering/render_button.cpp: Added.
        (khtml::m_inner):
        (khtml::RenderButton::~RenderButton):
        (khtml::RenderButton::addChild):
        (khtml::RenderButton::removeChild):
        (khtml::RenderButton::setStyle):
        (khtml::RenderButton::updateFromElement):
        * khtml/rendering/render_button.h: Added.
        (khtml::RenderButton::removeLeftoverAnonymousBoxes):
        (khtml::RenderButton::renderName):
        * khtml/rendering/render_form.cpp:
        * khtml/rendering/render_form.h:
        (khtml::RenderFileButton::calcReplacedHeight):
        (khtml::RenderFileButton::isEditable):
        * khtml/rendering/render_style.h:
        (khtml::):
        * khtml/rendering/render_theme.cpp:
        (khtml::RenderTheme::adjustStyle):
        (khtml::RenderTheme::paint):
        (khtml::RenderTheme::adjustCheckboxStyle):
        (khtml::RenderTheme::adjustRadioStyle):
        (khtml::RenderTheme::adjustButtonStyle):
        * khtml/rendering/render_theme.h:
        (khtml::RenderTheme::setButtonSize):
        * khtml/rendering/render_theme_mac.h:
        * khtml/rendering/render_theme_mac.mm:
        (khtml::RenderThemeMac::setSizeFromFont):
        (khtml::RenderThemeMac::setFontFromControlSize):
        (khtml::RenderThemeMac::addIntrinsicMargins):
        (khtml::RenderThemeMac::setButtonPaddingFromControlSize):
        (khtml::RenderThemeMac::adjustButtonStyle):
        (khtml::RenderThemeMac::buttonSizes):
        (khtml::RenderThemeMac::buttonMargins):
        (khtml::RenderThemeMac::setButtonSize):
        (khtml::RenderThemeMac::setButtonCellState):
        (khtml::RenderThemeMac::paintButton):
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject actionElement]):
        (-[KWQAccObject role]):
        (-[KWQAccObject title]):
        * kwq/KWQButton.h: Removed.
        * kwq/KWQButton.mm: Removed.
        * kwq/KWQComboBox.mm:
        * kwq/KWQLineEdit.h:
        * kwq/KWQLineEdit.mm:
        * kwq/KWQPushButton.h: Removed.
        * kwq/KWQPushButton.mm: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10888 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Patch by TimO, Reviewed by hyatt, tested and landed by me.
ggaren [Fri, 21 Oct 2005 18:43:11 +0000 (18:43 +0000)]
    Patch by TimO, Reviewed by hyatt, tested and landed by me.

        Found what appears to be a misguided optimization that actually causes a measurable performance problem.
        A fixed-size buffer was allocated on the stack to pass into CFURLGetBytes(), presumably to avoid malloc()
        for URLs less than 2048 bytes.  There was also a fallback which malloc()'ed a buffer in case the fixed-size
        buffer was too small to hold the URL's bytes.  This malloc()'ed buffer was then wrapped in an NSData using
        +dataWithBytesNoCopy:length:, avoiding a memory copy (yay!)

        The problem with this approach is two-fold:

        1. Regardless of how the buffer was allocated and filled, it is immediately wrapped in an NSData using
        +dataWithBytes:length:, which copies the input bytes.  This is pretty much unavoidable; we need to get
        the data into a malloc()'ed buffer to return it to the caller, unless the caller provides its own storage
        (which would be super inconvenient).

        2. The size of the fixed buffer was large enough that it fit most (if not all) URLs involved in our Page
        Load Test.  This means the unintentionally-inefficient case was by far the most *common* case!

        My fix is to malloc() the buffer from the start, and then use +[NSData dataWithBytes:length:freeWhenDone:]
        to wrap the buffer in an NSData.  This avoids a memory copy for the normal case where a URL is less than
        2048 bytes, and keeps the efficient behavior for the uncommon long URL case.

        * Misc.subproj/WebNSURLExtras.m:
        (-[NSURL _web_originalData]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10882 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed and landed by Darin.
darin [Fri, 21 Oct 2005 16:35:25 +0000 (16:35 +0000)]
    Reviewed and landed by Darin.

        - fixed a couple regressions caused by my last check-in
          http://bugzilla.opendarwin.org/show_bug.cgi?id=5437
          http://bugzilla.opendarwin.org/show_bug.cgi?id=5443

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (-[WebTextRenderer CG_drawHighlightForRun:style:geometry:]): Compute background width
        correctly, by subtracting position after run from position before run.
        (addDirectionalOverride): Make the range include only the characters between the
        directional override characters, not the directional override characters themselves.
        (initializeWidthIterator): Correctly compute "widthToStart" based on the offset to
        the beginning of the run, not to the end of the run!

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoWebCore:
darin [Thu, 20 Oct 2005 06:02:38 +0000 (06:02 +0000)]
WebCore:

        Reviewed by Maciej.

        - optimizations for a total of about 1% speed-up on PLT

        * khtml/html/htmltokenizer.cpp:
        (khtml::fixUpChar): Changed to be more inlinable.
        (khtml::HTMLTokenizer::processListing): Simplified the skipLF handling.
        (khtml::HTMLTokenizer::parseSpecial): Changed to call the new fixUpChar.
        (khtml::HTMLTokenizer::parseText): Simplified the skipLF handling and
        changed to call the new fixUpChar.
        (khtml::HTMLTokenizer::parseEntity): Changed to call the new fixUpChar.
        (khtml::HTMLTokenizer::parseTag): Changed to call the new fixUpChar.
        (khtml::HTMLTokenizer::write): Changed to call the new fixUpChar.

        * khtml/rendering/font.h: Removed the floatCharacterWidths function,
        since it's never needed. Made some more functions inline.
        * khtml/rendering/font.cpp: Removed lots of stuff that was only for the
        !APPLE_CHANGES case.

        * kwq/KWQFontMetrics.h: Removed the single-character width overloads, charWidth,
        and floatCharacterWidths.
        * kwq/KWQFontMetrics.mm:
        (QFontMetrics::width): Use lroundf instead of the ROUND_TO_INT macro.
        (QFontMetrics::floatWidth): Removed the bogus unneeded call to the ROUND_TO_INT macro.

        * khtml/rendering/render_text.cpp:
        (RenderText::cacheWidths): Use floatWidth instead of floatCharacterWidths.

        * kwq/KWQComboBox.mm: (QComboBox::sizeHint): Update since the floatWidthForRun
        method no longer takes a widths parameter.
        * kwq/KWQLineEdit.mm: (QLineEdit::sizeForCharacterWidth): Ditto.
        * kwq/KWQListBox.mm:
        (QListBox::sizeForNumberOfLines): Ditto.
        (-[KWQTableView drawRow:clipRect:]): Ditto.
        * kwq/KWQPainter.mm: (QPainter::drawText): Ditto. Also use lroundf instead of the
        ROUND_TO_INT macro.

        * kwq/WebCoreTextRenderer.h: Removed the ROUND_TO_INT macro. Changed to use bool
        instead of bit fields. Removed the widths parameter from the floatWidthForRun method.

WebKit:

        Reviewed by Maciej.

        - optimizations for a total of about 1% speed-up on PLT

        * WebCoreSupport.subproj/WebTextRenderer.h: Updated to use bool instead of BOOL,
        since BOOL is a signed char (which is not so efficient, at least on PPC).
        * WebCoreSupport.subproj/WebTextRenderer.m:
        (isSpace): Changed BOOL to bool and UniChar to UChar32. This actually fixes a
        potential bug when the passed-in character is a non-BMP character (> FFFF).
        (isRoundingHackCharacter): Ditto.
        (widthForGlyph): Merged getUncachedWidth, widthFromMap, and widthForGlyph into
        one function. Marked it inline. Changed to include syntheticBoldOffset in the
        cached widths to save an add in the cached case. Instead of the special constant
        UNINITIALIZED_GLYPH_WIDTH, just check for a width >= 0. This allows us to use
        a negative number or NAN for the uninitialized width value -- I chose NAN.
        (overrideLayoutOperation): Use bool instead of Boolean in one place.
        (-[WebTextRenderer initWithFont:]): Use lroundf instead of ROUND_TO_INT.
        (-[WebTextRenderer floatWidthForRun:style:]): Put the code to choose the ATSU
        vs. CG code path back in here, because there are no callers inside the class
        that need to call both.
        (-[WebTextRenderer drawLineForCharacters:yOffset:width:color:thickness:]):
        Use bool instead of BOOL.
        (+[WebTextRenderer setAlwaysUseATSU:]): Ditto.
        (fontContainsString): Ditto.
        (-[WebTextRenderer computeWidthForSpace]): Ditto. Also use roundf instead of
        using ROUND_TO_INT.
        (-[WebTextRenderer setUpFont]): Use bool instead of BOOL.
        (drawGlyphs): Ditto.
        (-[WebTextRenderer CG_drawHighlightForRun:style:geometry:]): Restructure the
        code so it can use the new advanceWidthIterator function instead of the old
        widthForNextCharacter function.
        (-[WebTextRenderer CG_drawRun:style:geometry:]): Use malloc instead of calloc
        since we don't need initialization. Call CG_floatWidthForRun instead of
        floatWidthForRun -- no need to re-check whether to use the CG or ATSU code path.
        Removed code to handle a renderer of 0 since we no longer generate that in
        the renderers array in advanceWidthIterator.
        (CG_floatWidthForRun): Changed to call the new advanceWidthIterator instead of
        the old widthForNextCharacter.
        (-[WebTextRenderer extendCharacterToGlyphMapToInclude:]): Use malloc instead of
        calloc and explicitly initialize the one field that needs it. Fixed a potential
        storage leak by adding a call to WKClearGlyphVector. Initialize the renderers
        to self instead of to 0.
        (-[WebTextRenderer extendGlyphToWidthMapToInclude:]): Initialize the widths to
        NAN instead of UNINITIALIZED_GLYPH_WIDTH.
        (addDirectionalOverride): Fixed bug where the first and last character in the buffer
        could be uninitialized and where characters before and after the direction override
        could be incorrect.
        (-[WebTextRenderer ATSU_drawRun:style:geometry:]): Use bool instead of BOOL.
        (-[WebTextRenderer ATSU_pointToOffset:style:position:reversed:includePartialGlyphs:]):
        Ditto.
        (advanceWidthIteratorOneCharacter): Added new helper function for CG_pointToOffset.
        (-[WebTextRenderer CG_pointToOffset:style:position:reversed:includePartialGlyphs:]):
        Reimplemented to use advanceWidthIteratorOneCharacter instead of widthForNextCharacter.
        Also call CG_floatWidthForRun instead of floatWidthForRun since we don't need to
        reconsider whether to use CG or ATSU.
        (glyphForCharacter): Removed the map parameter and changed the renderer parameter to
        be an in-out one. Removed uneeded special case for when map is 0 and always get the
        renderer from the map. Also call extendCharacterToGlyphMapToInclude in here instead of
        making that the caller's responsibility.
        (initializeWidthIterator): Renamed to make the name shorter (removed "Character").
        Streamlned common cases like "no padding" and removed some unneeded casts. Changed to
        use advanceWidthIterator to compute width fo the first part of the run.
        (normalizeVoicingMarks): Factored this out into a separate function, since it's not part
        of the common case.
        (advanceWidthIterator): Changed widthForNextCharacter to this, eliminating per-character
        function overhead for iterating past a few characters. Merged the handling of surrogate
        pairs and of voicing marks so that we typically only have to do one "if" to rule out both.
        Merged the mirroring for RTL and uppercasing for small caps into a single boolean so that
        we only need one "if" to rule out both. Call the new glyphForCharacter. Check for the
        character '\t' first since that's cheaper than looking at tabWidth. Check tabWidth for 0
        first so that we don't have to convert it to floating point when not using it. Changed
        the special case for spaces to first check width, so that we don't bother with the rest
        of the code for glyphs not the same width as spaces. Fixed substitution code to call
        CG_floatWidthForRun -- no need to reconsider whether to use CG or ATSU. Also changed to
        properly get width from the result of that function. Merged the handling of letter spacing,
        padding, and word spacing into a single boolean so that we typically only have to do one
        "if" to rule out all three. Check for letterSpacing of 0 first so that we don't have to
        convert it to floating point when not using it. Same for padding and wordSpacing.
        Move the work from ceilCurrentWidth in line into this function. Assume that either we have
        all three pointers (widths, renderers, glyphs), or none of the three, to cut down on branches.
        (fillStyleWithAttributes): Use bool instead of BOOL.
        (shouldUseATSU): Ditto.

        * Misc.subproj/WebKitNSStringExtras.m: (-[NSString _web_widthWithFont:]): Update since
        the floatWidthForRun method no longer takes a widths parameter.
        * Misc.subproj/WebStringTruncator.m: (stringWidth): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10880 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: none
eseidel [Thu, 20 Oct 2005 05:03:15 +0000 (05:03 +0000)]
Bug #: none
Submitted by: timo
Reviewed by: eseidel & darin
Changed some of the run measurement methods to C functions
        to avoid overhead associated with objc_msgSend().

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (-[WebTextRenderer floatWidthForRun:style:widths:]):
Updated to call new run measurement functions instead
        of calling ObjC methods.
        (-[WebTextRenderer CG_drawRun:style:geometry:]): ditto
        (floatWidthForRun): ditto
        (CG_floatWidthForRun): ditto
        (ATSU_floatWidthForRun): ditto
        (widthForNextCharacter): ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Dave.
mjs [Wed, 19 Oct 2005 09:15:14 +0000 (09:15 +0000)]
    Reviewed by Dave.

        - instead of walking the render tree to update widget positions, instead
keep a set of widgets per canvas, maintained by the constructor and destroy method
for RenderWidget. About a .5% speedup.

        * khtml/khtmlview.cpp:
        (KHTMLView::layout):
        * khtml/rendering/render_canvas.cpp:
        (RenderCanvas::updateWidgetPositions):
        (RenderCanvas::addWidget):
        (RenderCanvas::removeWidget):
        * khtml/rendering/render_canvas.h:
        * khtml/rendering/render_frames.cpp:
        (RenderPart::updateWidgetPosition):
        (RenderPart::needWidgetPositionUpdating):
        * khtml/rendering/render_frames.h:
        * khtml/rendering/render_layer.cpp:
        (khtml::RenderLayer::scrollToOffset):
        * khtml/rendering/render_object.cpp:
        (RenderObject::updateWidgetPosition):
        (RenderObject::needWidgetPositionUpdating):
        * khtml/rendering/render_object.h:
        * khtml/rendering/render_replaced.cpp:
        (RenderWidget::RenderWidget):
        (RenderWidget::destroy):
        (RenderWidget::updateWidgetPosition):
        (RenderWidget::needWidgetPositionUpdating):
        * khtml/rendering/render_replaced.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10871 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed and landed by Maciej.
mjs [Wed, 19 Oct 2005 00:03:47 +0000 (00:03 +0000)]
    Reviewed and landed by Maciej.

        - some simple changes that amount to a < 1% speedup.

        * khtml/css/cssstyleselector.cpp:
        (khtml::CSSStyleSelector::applyProperty):
        * khtml/rendering/bidi.cpp:
        (khtml::BidiIterator::direction):
        * kwq/KWQKURL.mm:
        (hasSlashDotOrDotDot):
        * kwq/KWQString.mm:
        (QString::~QString):
        * kwq/WebCoreTextRendererFactory.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - back out the last change, it caused a regression with painting of offscreen...
mjs [Tue, 18 Oct 2005 20:45:38 +0000 (20:45 +0000)]
    - back out the last change, it caused a regression with painting of offscreen plugins

        * khtml/khtmlview.cpp:
        (KHTMLView::layout):
        * khtml/rendering/render_form.cpp:
        (RenderFormElement::layout):
        * khtml/rendering/render_frames.cpp:
        (RenderPartObject::layout):
        (RenderPart::updateWidgetPositions):
        * khtml/rendering/render_frames.h:
        * khtml/rendering/render_layer.cpp:
        (khtml::RenderLayer::scrollToOffset):
        * khtml/rendering/render_object.cpp:
        (RenderObject::updateWidgetPositions):
        * khtml/rendering/render_object.h:
        * khtml/rendering/render_replaced.cpp:
        (RenderWidget::layout):
        (RenderWidget::updateWidgetPositions):
        * khtml/rendering/render_replaced.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10869 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Dave Hyatt.
sullivan [Tue, 18 Oct 2005 17:08:17 +0000 (17:08 +0000)]
    Reviewed by Dave Hyatt.
        Committed by John Sullivan.

        - .5% or so speedup by removing updateWidgetPositions

Instead, widget sizes are calculated in layout and widget positions are set at
paint time, with the same kind of special handling for incremental repaint that
RenderImage does.

        * khtml/khtmlview.cpp:
        (KHTMLView::layout):
        * khtml/rendering/render_form.cpp:
        (RenderFormElement::layout):
        * khtml/rendering/render_frames.cpp:
        (RenderPartObject::layout):
        * khtml/rendering/render_frames.h:
        * khtml/rendering/render_layer.cpp:
        (khtml::RenderLayer::scrollToOffset):
        * khtml/rendering/render_object.cpp:
        * khtml/rendering/render_object.h:
        * khtml/rendering/render_replaced.cpp:
        (RenderWidget::layout):
        * khtml/rendering/render_replaced.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoJavaScriptCore:
mjs [Tue, 18 Oct 2005 03:15:31 +0000 (03:15 +0000)]
JavaScriptCore:

        Reviewed by Geoff. Code changes by Darin.

- some micro-optimizations to FastMalloc to reduce math and branches.

        * kxmlcore/FastMalloc.cpp:
        (KXMLCore::TCMalloc_Central_FreeList::Populate):
        (KXMLCore::fastMallocRegisterThread):
        (KXMLCore::TCMalloc_ThreadCache::GetCache):
        (KXMLCore::TCMalloc_ThreadCache::GetCacheIfPresent):

WebCore:

        Reviewed by Geoff.

        Speed up the tokenizer by keeping more state on the stack instead of in the object,
to avoid load-store traffic. About a .5% speedup.

        * khtml/html/htmltokenizer.cpp:
        (khtml::HTMLTokenizer::HTMLTokenizer):
        (khtml::HTMLTokenizer::reset):
        (khtml::HTMLTokenizer::begin):
        (khtml::HTMLTokenizer::setForceSynchronous):
        (khtml::HTMLTokenizer::processListing):
        (khtml::HTMLTokenizer::parseSpecial):
        (khtml::HTMLTokenizer::scriptHandler):
        (khtml::HTMLTokenizer::scriptExecution):
        (khtml::HTMLTokenizer::parseComment):
        (khtml::HTMLTokenizer::parseServer):
        (khtml::HTMLTokenizer::parseProcessingInstruction):
        (khtml::HTMLTokenizer::parseText):
        (khtml::HTMLTokenizer::parseEntity):
        (khtml::HTMLTokenizer::parseTag):
        (khtml::HTMLTokenizer::continueProcessing):
        (khtml::HTMLTokenizer::write):
        (khtml::HTMLTokenizer::allDataProcessed):
        (khtml::HTMLTokenizer::end):
        (khtml::HTMLTokenizer::finish):
        (khtml::HTMLTokenizer::notifyFinished):
        (khtml::HTMLTokenizer::isWaitingForScripts):
        * khtml/html/htmltokenizer.h:
        (khtml::HTMLTokenizer::):
        (khtml::HTMLTokenizer::State::State):
        (khtml::HTMLTokenizer::State::tagState):
        (khtml::HTMLTokenizer::State::setTagState):
        (khtml::HTMLTokenizer::State::entityState):
        (khtml::HTMLTokenizer::State::setEntityState):
        (khtml::HTMLTokenizer::State::inScript):
        (khtml::HTMLTokenizer::State::setInScript):
        (khtml::HTMLTokenizer::State::inStyle):
        (khtml::HTMLTokenizer::State::setInStyle):
        (khtml::HTMLTokenizer::State::inSelect):
        (khtml::HTMLTokenizer::State::setInSelect):
        (khtml::HTMLTokenizer::State::inXmp):
        (khtml::HTMLTokenizer::State::setInXmp):
        (khtml::HTMLTokenizer::State::inTitle):
        (khtml::HTMLTokenizer::State::setInTitle):
        (khtml::HTMLTokenizer::State::inPlainText):
        (khtml::HTMLTokenizer::State::setInPlainText):
        (khtml::HTMLTokenizer::State::inProcessingInstruction):
        (khtml::HTMLTokenizer::State::setInProcessingInstruction):
        (khtml::HTMLTokenizer::State::inComment):
        (khtml::HTMLTokenizer::State::setInComment):
        (khtml::HTMLTokenizer::State::inTextArea):
        (khtml::HTMLTokenizer::State::setInTextArea):
        (khtml::HTMLTokenizer::State::escaped):
        (khtml::HTMLTokenizer::State::setEscaped):
        (khtml::HTMLTokenizer::State::inServer):
        (khtml::HTMLTokenizer::State::setInServer):
        (khtml::HTMLTokenizer::State::skipLF):
        (khtml::HTMLTokenizer::State::setSkipLF):
        (khtml::HTMLTokenizer::State::startTag):
        (khtml::HTMLTokenizer::State::setStartTag):
        (khtml::HTMLTokenizer::State::discardLF):
        (khtml::HTMLTokenizer::State::setDiscardLF):
        (khtml::HTMLTokenizer::State::allowYield):
        (khtml::HTMLTokenizer::State::setAllowYield):
        (khtml::HTMLTokenizer::State::loadingExtScript):
        (khtml::HTMLTokenizer::State::setLoadingExtScript):
        (khtml::HTMLTokenizer::State::forceSynchronous):
        (khtml::HTMLTokenizer::State::setForceSynchronous):
        (khtml::HTMLTokenizer::State::inAnySpecial):
        (khtml::HTMLTokenizer::State::hasTagState):
        (khtml::HTMLTokenizer::State::hasEntityState):
        (khtml::HTMLTokenizer::State::):
        (khtml::HTMLTokenizer::State::setBit):
        (khtml::HTMLTokenizer::State::testBit):
        * khtml/rendering/bidi.cpp:
        (khtml::RenderBlock::checkLinesForTextOverflow):
        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::updateFirstLetter):
        * khtml/rendering/render_flow.cpp:
        (RenderFlow::caretRect):
        * khtml/rendering/render_line.cpp:
        (khtml::EllipsisBox::paint):
        * khtml/rendering/render_object.cpp:
        (RenderObject::firstLineStyle):
        * khtml/rendering/render_object.h:
        (khtml::RenderObject::style):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10867 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLayoutTests:
mjs [Tue, 18 Oct 2005 00:16:48 +0000 (00:16 +0000)]
LayoutTests:

        - temporarily back out hyatt's recent changes since the tree was closed

        * fast/clip/008-expected.txt:
        * fast/css/acid2-expected.txt:
        * fast/css/acid2-pixel-expected.txt:
        * fast/dom/gc-3.html:
        * fast/table/overflowHidden-expected.txt:

WebCore:

        - temporarily back out hyatt's recent changes since the tree was closed

        * khtml/rendering/render_container.cpp:
        (RenderContainer::destroy):
        (RenderContainer::destroyChildren):
        * khtml/rendering/render_container.h:
        * khtml/rendering/render_flow.cpp:
        (RenderFlow::destroy):
        * khtml/rendering/render_layer.cpp:
        (khtml::RenderLayer::isTransparent):
        (khtml::RenderLayer::paintLayer):
        (khtml::sortByZOrder):
        * khtml/rendering/render_layer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10866 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agocheck in fixed overflow:hidden table case.
hyatt [Mon, 17 Oct 2005 07:37:01 +0000 (07:37 +0000)]
check in fixed overflow:hidden table case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10863 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoCheck in fixed acid2.
hyatt [Mon, 17 Oct 2005 07:36:37 +0000 (07:36 +0000)]
Check in fixed acid2.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10862 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoFix clip test.
hyatt [Mon, 17 Oct 2005 07:35:59 +0000 (07:35 +0000)]
Fix clip test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Fix for bugzilla bug 5283. Make overflow layers lose to other kinds of
hyatt [Mon, 17 Oct 2005 07:29:22 +0000 (07:29 +0000)]
Fix for bugzilla bug 5283.  Make overflow layers lose to other kinds of
layers if z-index is equivalent.  Technically overflow isn't even supposed
to establish a stacking context, so the use of RenderLayer for overflow,
although elegant and simple, isn't correct.  This patch is essentially a hack
to make the common problem go away, but the deeper mistake remains.

        Reviewed by eseidel

        * khtml/rendering/render_layer.cpp:
        (khtml::RenderLayer::isTransparent):
        (khtml::RenderLayer::paintLayer):
        (khtml::isOverflowOnly):
        (khtml::compare):
        (khtml::sortByZOrder):
        * khtml/rendering/render_layer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Clean up the deletion of anonymous boxes in the render tree.
hyatt [Mon, 17 Oct 2005 06:20:03 +0000 (06:20 +0000)]
Clean up the deletion of anonymous boxes in the render tree.
Renamed methods to make it more clear and also added comments.

        Reviewed by bdakin

        * khtml/rendering/render_container.cpp:
        (RenderContainer::destroy):
        (RenderContainer::destroyLeftoverAnonymousChildren):
        * khtml/rendering/render_container.h:
        * khtml/rendering/render_flow.cpp:
        (RenderFlow::destroy):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Backed out the following changes, since they are in a range that
mjs [Sun, 16 Oct 2005 03:36:07 +0000 (03:36 +0000)]
Backed out the following changes, since they are in a range that
provably caused a performance regression:

<rdar://problem/4302874> Denver Regression: crash repeatedly reloading www.supermanhomepage.com
<rdar://problem/4302879> Re-land SVG object element fix
<rdar://problem/4302880> Re-land isSameNode change
<rdar://problem/4065748> Repro crash at http://www.vanaqua.org using menu system (KHTMLParser::setCurrent(DOM::NodeImpl*))

        * khtml/ecma/kjs_dom.cpp:
        (KJS::DOMNodeProtoFunc::callAsFunction):
        * khtml/ecma/kjs_dom.h:
        (KJS::DOMNode::):
        * khtml/ecma/kjs_events.cpp:
        (KJS::JSAbstractEventListener::handleEvent):
        * khtml/html/html_objectimpl.cpp:
        (DOM::HTMLObjectElementImpl::isImageType):
        * khtml/html/htmlparser.cpp:
        (HTMLStackElem::HTMLStackElem):
        (HTMLParser::popNestedHeaderTag):
        (HTMLParser::handleResidualStyleCloseTagAcrossBlocks):
        (HTMLParser::popOneBlock):
        * khtml/xml/dom_nodeimpl.h:
        * kwq/DOM.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10858 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reverted fix for this bug, because it was part of a time range that caused a performance
mjs [Sun, 16 Oct 2005 00:46:25 +0000 (00:46 +0000)]
Reverted fix for this bug, because it was part of a time range that caused a performance
regression:

<rdar://problem/4260481> Remove Reference type from JavaScriptCore

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago * kxmlcore/HashTable.cpp: Fixed build failure (said hashtable.h instead of...
darin [Sat, 15 Oct 2005 23:27:28 +0000 (23:27 +0000)]
    * kxmlcore/HashTable.cpp: Fixed build failure (said hashtable.h instead of HashTable.h).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10856 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Style changes recommended by Darin.
ggaren [Fri, 14 Oct 2005 21:03:56 +0000 (21:03 +0000)]
    Style changes recommended by Darin.

        Changed to camelCase, changed ValueImp* to ValueImp *.

        * kjs/simple_number.h:
        (KJS::SimpleNumber::make):
        (KJS::SimpleNumber::value):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Style change suggested by Darin.
ggaren [Fri, 14 Oct 2005 20:57:54 +0000 (20:57 +0000)]
    Style change suggested by Darin.

        * khtml/html/htmltokenizer.cpp:
        (khtml::HTMLTokenizer::scriptHandler): Added "Bugzilla" prefix to bug #.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Changes by Mitz Pettel, reviewed by Maciej.
vicki [Fri, 14 Oct 2005 20:31:46 +0000 (20:31 +0000)]
Changes by Mitz Pettel, reviewed by Maciej.

Fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5029 (Assertion failure in -[NSPasteboard(WebExtras)
_web_writeImage:URL:title:archive:types:] when trying to drag an image from a site with no favicon)

        * Misc.subproj/WebNSPasteboardExtras.m:
        (-[NSPasteboard _web_writeImage:URL:title:archive:types:]): Prefer the main resource if it is an image

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5336
eseidel [Fri, 14 Oct 2005 20:20:54 +0000 (20:20 +0000)]
Bug #: 5336
Submitted by: Julien Palmas <julien.palmas@mac.com>
Reviewed by: eseidel
        Fixed crash when animating unsupported elements.
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5336

        * ksvg2/svg/SVGAnimationElementImpl.cpp:
        (SVGAnimationElementImpl::targetElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: none
eseidel [Fri, 14 Oct 2005 06:10:01 +0000 (06:10 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: none, removed newline.
        Committed file w/ extra newline by mistake.

        * fast/js/resources/xmlhttprequest-get-data.xml:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: none
eseidel [Thu, 13 Oct 2005 23:54:38 +0000 (23:54 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: mjs
        Adding the mozilla xslt tests.

        * fast/xsl/mozilla-tests.xml: Added.
        * fast/xsl/mozilla-tests.xsl: Added.
        * fast/xsl/mozilla-tests-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: none
eseidel [Thu, 13 Oct 2005 23:51:01 +0000 (23:51 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: mjs
        Adding a small xmlhttprequest test.

        * fast/js/resources/xmlhttprequest-get-data.xml: Added.
        * fast/js/xmlhttprequest-get-expected.txt: Added.
        * fast/js/xmlhttprequest-get.xhtml: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: none
eseidel [Thu, 13 Oct 2005 23:39:46 +0000 (23:39 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: mjs
        Adding a small xmlhttprequest test.

        * fast/js/resources/xmlhttprequest-get-data.xml: Added.
        * fast/js/xmlhttprequest-get-expected.txt: Added.
        * fast/js/xmlhttprequest-get.xhtml: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by hyatt
justing [Thu, 13 Oct 2005 22:51:00 +0000 (22:51 +0000)]
    Reviewed by hyatt

        <rdar://problem/3643259> contentEditable=true elements do not acquire the caret or focus halo when overflow:scroll

        Updated/added layout tests:

        * fast/clip/001-expected.txt:
        * fast/clip/009-expected.txt:
        * fast/clip/010-expected.txt:
        * fast/clip/011-expected.txt:
        * fast/clip/012-expected.txt:
        * fast/clip/013-expected.txt:
        * fast/clip/014-expected.txt:
        * fast/clip/016-expected.txt:
        * fast/clip/outline-overflowClip-expected.checksum: Added.
        * fast/clip/outline-overflowClip-expected.png: Added.
        * fast/clip/outline-overflowClip-expected.txt: Added.
        * fast/clip/outline-overflowClip.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by hyatt
justing [Thu, 13 Oct 2005 22:45:56 +0000 (22:45 +0000)]
    Reviewed by hyatt

        <rdar://problem/3643259> contentEditable=true elements do not acquire the caret or focus halo when overflow:scroll

        Test cases added:
        * fast/clip/outline-overflowClip
        Test cases changed (since I added the outline rect to the render tree dump)
        * fast/clip/001
        * fast/clip/009
        * fast/clip/010
        * fast/clip/011
        * fast/clip/012
        * fast/clip/013
        * fast/clip/014
        * fast/clip/016

        * khtml/rendering/render_layer.cpp:
        Clip the outline by a new rect, which is equal to the foreground rect w/o clipping
        by the overflow rect.

        (khtml::RenderLayer::paintLayer):
        (khtml::RenderLayer::hitTestLayer):
        (khtml::RenderLayer::calculateRects):
        * khtml/rendering/render_layer.h:
        * kwq/KWQRenderTreeDebug.cpp:
        (write):
        (writeLayers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Darin.
thatcher [Thu, 13 Oct 2005 22:07:51 +0000 (22:07 +0000)]
    Reviewed by Darin.

        no need to calculate linewidth multiple times in findnextlinebreak()
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5319

        No test case needed, this is a simple optimization.

        * khtml/rendering/bidi.cpp:
        (khtml::RenderBlock::findNextLineBreak):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10845 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Darin.
vicki [Thu, 13 Oct 2005 19:37:31 +0000 (19:37 +0000)]
    Reviewed by Darin.

- fix <rdar://problem/4043643> iframe swallows events for overlapping elements (3449)

        * WebView.subproj/WebHTMLView.m:
        (-[WebHTMLView hitTest:]):
        (-[WebHTMLView _updateMouseoverWithEvent:]): eliminate _hitViewForEvent hackery and self
dependency from this function

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10844 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - Fixed <rdar://problem/4259434> Safari crashes in HTMLTokenizer::~HTMLTokeni...
ggaren [Thu, 13 Oct 2005 19:11:23 +0000 (19:11 +0000)]
    - Fixed <rdar://problem/4259434> Safari crashes in HTMLTokenizer::~HTMLTokenizer()
          at http://www.timewarner.com/corp/careers/jobtools_us/index.html

        I changed the test for whether to put a script in the "to be executed" queue to
        match the test for whether to ref a script, so that scripts can't end up in the
        queue without being refed.

        I also renamed cachedScript to pendingScripts.

        Reviewed by Darin.

        No test case because the crash isn't deterministically reproducible.
        However, I did add assertions that should catch the underlying bug
        in the future.

        * khtml/html/htmltokenizer.cpp:
        (khtml::HTMLTokenizer::reset):
        (khtml::HTMLTokenizer::scriptHandler):
        (khtml::HTMLTokenizer::write):
        (khtml::HTMLTokenizer::notifyFinished):
        * khtml/html/htmltokenizer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10843 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5326
eseidel [Thu, 13 Oct 2005 08:29:21 +0000 (08:29 +0000)]
Bug #: 5326
Submitted by: Rob Buis  <rwlbuis@xs4all.nl>
Reviewed by: eseidel
        Test cases updated:
        * svg-tests/W3C-SVG-1.1/filters-color-01-b-expected.checksum:
        * svg-tests/W3C-SVG-1.1/filters-color-01-b-expected.png:
        * svg-tests/W3C-SVG-1.1/filters-color-01-b-expected.txt:
        * svg-tests/W3C-SVG-1.1/struct-symbol-01-b-expected.checksum:
        * svg-tests/W3C-SVG-1.1/struct-symbol-01-b-expected.png:
        * svg-tests/W3C-SVG-1.1/struct-symbol-01-b-expected.txt:

        Fixed <svg> to base %lengths off of the nearest viewbox.
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5326

        * ksvg2/svg/SVGPreserveAspectRatioImpl.cpp:
        (SVGPreserveAspectRatioImpl::SVGPreserveAspectRatioImpl):
        * ksvg2/svg/SVGSVGElementImpl.cpp:
        (SVGSVGElementImpl::x):
        (SVGSVGElementImpl::y):
        (SVGSVGElementImpl::width):
        (SVGSVGElementImpl::height):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10842 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Darin.
thatcher [Thu, 13 Oct 2005 01:08:26 +0000 (01:08 +0000)]
    Reviewed by Darin.

        Define WebNSInt and WebNSUInt to wrap around NSInt on Leopard and still build on Tiger
        Once building on Tiger isn't needed we will drop WebNSInt and use NSInt

        * WebView.subproj/WebDefaultResourceLoadDelegate.m:
        (-[WebDefaultResourceLoadDelegate webView:resource:didReceiveContentLength:fromDataSource:]):
        * WebView.subproj/WebFrame.m:
        (-[WebFrame _sendRemainingDelegateMessagesWithIdentifier:response:length:error:]):
        * WebView.subproj/WebLoader.m:
        (-[NSURLProtocol didReceiveData:lengthReceived:]):
        * WebView.subproj/WebResourceLoadDelegate.h:
        * WebView.subproj/WebUIDelegate.h:
        * WebView.subproj/WebView.h:
        * WebView.subproj/WebView.m:
        (-[WebView _mouseDidMoveOverElement:modifierFlags:]):
        (-[WebView spellCheckerDocumentTag]):
        * WebView.subproj/WebViewInternal.h:
        * WebView.subproj/WebViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10839 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - update expected results for <rdar://problem/4288266> createElement does not validate
vicki [Thu, 13 Oct 2005 00:53:10 +0000 (00:53 +0000)]
- update expected results for <rdar://problem/4288266> createElement does not validate
names when called in an HTML document (4566)

        * dom/html/level1/core/hc_documentinvalidcharacterexceptioncreateelement-expected.txt:
        * dom/html/level1/core/hc_documentinvalidcharacterexceptioncreateelement1-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10838 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Hyatt.
vicki [Thu, 13 Oct 2005 00:49:29 +0000 (00:49 +0000)]
    Reviewed by Hyatt.

- fix <rdar://problem/4288266> createElement does not validate names when called in an HTML document (4566)

        * khtml/html/html_documentimpl.cpp:
        (DOM::HTMLDocumentImpl::createElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago*** empty log message ***
bdakin [Thu, 13 Oct 2005 00:35:47 +0000 (00:35 +0000)]
*** empty log message ***

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10836 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago * WebView.subproj/WebPolicyDelegate.h: Fix a comment.
darin [Wed, 12 Oct 2005 21:02:50 +0000 (21:02 +0000)]
    * WebView.subproj/WebPolicyDelegate.h: Fix a comment.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10835 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - removed some old results for tests that are no longer present
darin [Wed, 12 Oct 2005 20:58:20 +0000 (20:58 +0000)]
    - removed some old results for tests that are no longer present
          http://bugzilla.opendarwin.org/show_bug.cgi?id=4850

        * dom/html/level2/html/frame-expected.checksum: Removed.
        * dom/html/level2/html/frame-expected.png: Removed.
        * dom/html/level2/html/iframe-expected.checksum: Removed.
        * dom/html/level2/html/iframe-expected.png: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10834 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Eric.
thatcher [Wed, 12 Oct 2005 06:55:39 +0000 (06:55 +0000)]
    Reviewed by Eric.

        Crash: Drag element with CSS ::Before absolute positioning applied to it
        http://bugzilla.opendarwin.org/show_bug.cgi?id=4088

        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::khtmlMouseMoveEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10832 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Darin.
thatcher [Wed, 12 Oct 2005 01:00:28 +0000 (01:00 +0000)]
    Reviewed by Darin.

        Test for 10.4 because of <rdar://problem/4243463>

        * WebView.subproj/WebHTMLView.m:
        (-[WebHTMLView conversationIdentifier]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10831 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Darin.
thatcher [Wed, 12 Oct 2005 00:44:53 +0000 (00:44 +0000)]
    Reviewed by Darin.

        workaround for <rdar://problem/4294625> to fix our build

        * kwq/KWQPointArray.h:
        * kwq/KWQSize.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10830 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Rolling out fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=5195
adele [Tue, 11 Oct 2005 22:37:52 +0000 (22:37 +0000)]
    Rolling out fix for bugzilla.opendarwin.org/show_bug.cgi?id=5195
        since it caused:

        REGRESSION text areas draw focus ring around each glyph, no caret in text fields
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5335

        * WebView.subproj/WebHTMLView.m:
        (-[WebHTMLView _propagateDirtyRectsToOpaqueAncestors]):
        (-[WebHTMLView _recursiveDisplayRectIfNeededIgnoringOpacity:isVisibleRect:rectIsVisibleRectForView:topView:]):
        (-[WebHTMLView _recursiveDisplayAllDirtyWithLockFocus:visRect:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10827 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Added regexp_object.lut.h build phase from JavaScriptCore
ggaren [Tue, 11 Oct 2005 22:26:37 +0000 (22:26 +0000)]
    Added regexp_object.lut.h build phase from JavaScriptCore
        to JavaScriptCore+SVG.

        Reviewed by mitz.

        * JavaScriptCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10825 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Fixed build bustage from last checkin (stray characters
ggaren [Tue, 11 Oct 2005 22:01:42 +0000 (22:01 +0000)]
    Fixed build bustage from last checkin (stray characters
        in the project file).

        Reviewed by mitz.

        * JavaScriptCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10824 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago New JavaScriptCore test results to reflect the last change.
ggaren [Tue, 11 Oct 2005 21:49:37 +0000 (21:49 +0000)]
    New JavaScriptCore test results to reflect the last change.

        * tests/mozilla/expected.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10823 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoSmall fixup to last Changelog entry.
ggaren [Tue, 11 Oct 2005 21:00:13 +0000 (21:00 +0000)]
Small fixup to last Changelog entry.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10821 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Added a layout test for the new special properties of the global RegExp object.
ggaren [Tue, 11 Oct 2005 20:57:39 +0000 (20:57 +0000)]
    Added a layout test for the new special properties of the global RegExp object.
        (See JavaSriptCore changelog.)

        * fast/js/regexp-caching-expected.txt: Added.
        * fast/js/regexp-caching.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - Implemented caching of match state inside the global RegExp object
ggaren [Tue, 11 Oct 2005 20:43:49 +0000 (20:43 +0000)]
    - Implemented caching of match state inside the global RegExp object
          (lastParen, leftContext, rightContext, lastMatch, input).

          exec(), test(), match(), search(), and replace() now dipatch regular
          expression matching through the RegExp object's performMatch function,
          to facilitate caching. This replaces registerRegexp and
          setSubPatterns.

        - Implemented the special '$' aliases (e.g. RegExp.input aliases to
          RegExp.$_).

        - Moved support for backreferences into the new static hash table
          used for other special RegExp properties. Truncated backreferences
          at $9 to match IE, FF, and the "What's New in Netscape 1.2?" doc.
          (String.replace still supports double-digit backreferences.)

        - Tweaked RegExp.prototype.exec to handle ginormous values in lastIndex.

        Fixes 11 -- count em, 11 -- JavaScriptCore tests.

        Reviewed by NOBODY (OOPS!).

        * JavaScriptCore.xcodeproj/project.pbxproj: Added regexp_object.lut.h
        * kjs/create_hash_table: Tweaked to allow for more exotic characters.
                                 We now rely on the compiler to catch illegal
                                 identifiers.
        * kjs/regexp.cpp:
        (KJS::RegExp::RegExp):
        * kjs/regexp_object.cpp:
        (RegExpProtoFuncImp::callAsFunction):
        (RegExpObjectImp::RegExpObjectImp):
        (RegExpObjectImp::performMatch):
        (RegExpObjectImp::arrayOfMatches):
        (RegExpObjectImp::backrefGetter):
        (RegExpObjectImp::getLastMatch):
        (RegExpObjectImp::getLastParen):
        (RegExpObjectImp::getLeftContext):
        (RegExpObjectImp::getRightContext):
        (RegExpObjectImp::getOwnPropertySlot):
        (RegExpObjectImp::getValueProperty):
        (RegExpObjectImp::put):
        (RegExpObjectImp::putValueProperty):
        * kjs/regexp_object.h:
        (KJS::RegExpObjectImp::):
        * kjs/string_object.cpp:
        (substituteBackreferences):
        (replace):
        (StringProtoFuncImp::callAsFunction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10818 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by eric
justing [Tue, 11 Oct 2005 20:33:00 +0000 (20:33 +0000)]
    Reviewed by eric

        <rdar://problem/4283218> M-JPEG / Server Push and JavaScript

        Ignore multipart loads in numRequests, call checkCompleted() after each section is complete.

        * khtml/misc/loader.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLayoutTests:
adele [Tue, 11 Oct 2005 19:16:45 +0000 (19:16 +0000)]
LayoutTests:

        Rolling out Darin's previous change.

        * fast/clip/014-expected.checksum: Updated.
        * fast/clip/014-expected.png: Updated.

WebCore:

        Rolling out fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=5195
        since it caused:

        REGRESSION text areas draw focus ring around each glyph, no caret in text fields
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5335

        And updated expected results.
        * fast/clip/014-expected.checksum: Updated.
        * fast/clip/014-expected.png: Updated.

        * kwq/KWQTextArea.h:
        * kwq/KWQTextArea.mm:
        (-[KWQTextAreaTextView displayRectIgnoringOpacity:]):
        (-[KWQTextAreaTextView textView:shouldDrawInsertionPointInRect:color:turnedOn:]):
        * kwq/KWQTextField.h:
        * kwq/KWQTextField.mm:
        (-[KWQTextFieldController setInDrawingMachinery:]):
        (-[KWQTextFieldController textView:shouldDrawInsertionPointInRect:color:turnedOn:]):
        (-[KWQTextField displayRectIgnoringOpacity:]):
        (-[KWQSecureTextField displayRectIgnoringOpacity:]):
        (-[KWQSearchField displayRectIgnoringOpacity:]):
        * kwq/KWQWidget.h:
        * kwq/KWQWidget.mm:
        (QWidget::QWidget):
        (QWidget::~QWidget):
        (QWidget::paint):
        (QWidget::setDeferFirstResponderChanges):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10816 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: none
eseidel [Tue, 11 Oct 2005 09:33:29 +0000 (09:33 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: hyatt
        Checks for leaks in ImageDiff too.

        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10815 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5278
eseidel [Tue, 11 Oct 2005 09:19:23 +0000 (09:19 +0000)]
Bug #: 5278
Submitted by: Rob Buis  <rwlbuis@xs4all.nl>
Reviewed by: eseidel
        Fix SVGSymbolElementImpl to not warn when inheriting viewbox.
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5278

        * ksvg2/svg/SVGUseElementImpl.cpp:
        (SVGUseElementImpl::close):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10814 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5301
eseidel [Tue, 11 Oct 2005 09:11:49 +0000 (09:11 +0000)]
Bug #: 5301
Submitted by: Julien Palmas <julien.palmas@mac.com>
Reviewed by: eseidel
        Fixes WebCore+SVG build under gcc 3.3
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5301

        * WebCore+SVG/DrawCanvasItem.mm:
        * WebCore+SVG/DrawView.mm:
        (-[DrawDocument drawRect:initWithDrawView:]):
        * kwq/KWQSlot.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10813 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - Fixed <rdar://problem/4245682> Regression: CrashTracer: 923 crashes
ggaren [Tue, 11 Oct 2005 06:55:29 +0000 (06:55 +0000)]
    - Fixed <rdar://problem/4245682> Regression: CrashTracer: 923 crashes
          in Safari at com.apple.WebCore: KJS::Window::clear + 132

        Reviewed by mjs.

        No test case because this crash isn't reliably reproducible.

        * khtml/ecma/kjs_window.cpp:
        (KJS::Window::Window): Initialize m_returnValueSlot to 0, so that we
        know not to write to it until it's explicitly set to a valid address.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10812 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - updated expected results due to fix for http://bugzilla.opendarwin.org...
darin [Tue, 11 Oct 2005 03:45:15 +0000 (03:45 +0000)]
    - updated expected results due to fix for bugzilla.opendarwin.org/show_bug.cgi?id=5195

        The old results were incorrectly clipped. This always worked fine in the
        browser, but didn't draw correctly in DumpRenderTree until the bug was fixed.

        * fast/clip/014-expected.checksum: Updated.
        * fast/clip/014-expected.png: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10811 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Maciej.
darin [Tue, 11 Oct 2005 02:49:35 +0000 (02:49 +0000)]
    Reviewed by Maciej.

        - finished fixing http://bugzilla.opendarwin.org/show_bug.cgi?id=5195
          Would like API to flush rendering of pending DOM changes

        The first half of the fix alone broke DumpRenderTree. Better now.

        * kwq/KWQWidget.h: Made paint non-virtual (it had no reason to be virtual in KWQ) and
        added an isPainting function.
        * kwq/KWQWidget.mm:
        (QWidget::QWidget): Added painting boolean (actually a count).
        (QWidget::~QWidget): Check that painting is 0 when the widget is destroyed.
        (QWidget::paint): Use SPI to draw in a way that works no matter what context we are in.
        (QWidget::isPainting): Return true if inside the paint function.

        * kwq/KWQTextArea.h: Remove inDrawingMachinery field since we can use isPainting now.
        * kwq/KWQTextArea.mm:
        (-[KWQTextAreaTextView textView:shouldDrawInsertionPointInRect:color:turnedOn:]):
        Use isPainting() instead of inDrawingMachinery. inDrawingMachinery required that we
        use displayRectIgnoringOpacity: to paint, which we no longer do.
        * kwq/KWQTextField.h: More of the same.
        * kwq/KWQTextField.mm:
        (-[KWQTextFieldController textView:shouldDrawInsertionPointInRect:color:turnedOn:]):
        Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10810 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5320
eseidel [Mon, 10 Oct 2005 23:48:55 +0000 (23:48 +0000)]
Bug #: 5320
Submitted by: darin
Reviewed by: eseidel
        Mem leak fixes in xslt code.
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5320

        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::applyXSLTransform):
        (DocumentImpl::setTransformSourceDocument):
        * khtml/xml/xml_tokenizer.cpp:
        (khtml::parseXMLDocumentFragment):
        * khtml/xsl/xslt_processorimpl.cpp:
        (DOM::XSLTProcessorImpl::XSLTProcessorImpl):
        (DOM::XSLTProcessorImpl::~XSLTProcessorImpl):
        (DOM::stylesheetLoadFunc):
        (DOM::XSLTProcessorImpl::transformDocument):
        (DOM::XSLTProcessorImpl::documentFromXMLDocPtr):
        * khtml/xsl/xslt_processorimpl.h:
        (DOM::XSLTProcessorImpl::stylesheet):
        (DOM::XSLTProcessorImpl::sourceDocument):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10809 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #:
hyatt [Mon, 10 Oct 2005 07:16:23 +0000 (07:16 +0000)]
Bug #:
Submitted by:
Reviewed by:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10808 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago * Scripts/check-dom-results: Dump ".xhtml" for tests in the xhtml directory.
darin [Mon, 10 Oct 2005 06:35:12 +0000 (06:35 +0000)]
    * Scripts/check-dom-results: Dump ".xhtml" for tests in the xhtml directory.
        * Scripts/cvs-apply: Handle added/deleted files in the current directory.
        * Scripts/cvs-unapply: Ditto.
        * Scripts/run-webkit-tests: Add a few more false-positive leaks.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10806 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Land fix for bugzilla bug 4974, make sure to only move below floats in line layout...
hyatt [Mon, 10 Oct 2005 03:50:22 +0000 (03:50 +0000)]
Land fix for bugzilla bug 4974, make sure to only move below floats in line layout if white-space is normal.
One case was using !isPre and including nowrap incorrectly as a result.  Patch from koivisto.

        Reviewed by hyatt

        * khtml/rendering/bidi.cpp:
        (khtml::RenderBlock::findNextLineBreak):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10805 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Maciej.
darin [Mon, 10 Oct 2005 01:28:59 +0000 (01:28 +0000)]
    Reviewed by Maciej.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5195
          Would like API to flush rendering of pending DOM changes

        This was actually a Tiger regression. When AppKit added a new code path for rendering
        NSView, our special hack for doing layout when we draw didn't work any more. So we
        were able to fix this without adding any API.

        * WebView.subproj/WebHTMLView.m:
        (-[WebHTMLView _recursiveDisplayRectIgnoringOpacity:inContext:topView:]): Added.
        Does the same thing that other _recursiveDisplay methods do.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10804 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoFixes the following build error when building Deployment:
thatcher [Sun, 9 Oct 2005 23:15:32 +0000 (23:15 +0000)]
Fixes the following build error when building Deployment:
DumpRenderTree/TextInputController.m: In function '-[TextInputController textInput]':
DumpRenderTree/TextInputController.m:87: warning: class 'NSView <WebDocumentView>' does not implement the 'NSTextInput' protocol

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10803 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoUpdated change log to mention one more bug fixed.
darin [Sun, 9 Oct 2005 22:57:59 +0000 (22:57 +0000)]
Updated change log to mention one more bug fixed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10802 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoJavaScriptCore:
darin [Sun, 9 Oct 2005 22:56:30 +0000 (22:56 +0000)]
JavaScriptCore:

        Reviewed by Maciej; some changes done after review.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5280
          Date.setMonth fails with negative values
        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5154
          JSC should switch to _r variants of unix time/date functions
        - fixed a few possible overflow cases

        Retested all tests to be sure nothing broke; added layout test for bug 5280.

        * kjs/config.h: Removed TIME_WITH_SYS_TIME define. Also set HAVE_SYS_TIMEB_H
        for the __APPLE__ case (the latter is accurate but irrelevant).

        * kjs/date_object.h: Reformatted. Removed unnecessary include of "function_object.h".
        Moved declarations of helper classes and functions into the cpp file.

        * kjs/date_object.cpp: Removed code at top to define macros to use CoreFoundation instead of
        POSIX date functions.
        (KJS::styleFromArgString): Tweaked to return early instead of using a variable.
        (KJS::formatLocaleDate): Tweaked to check for undefined rather than checking argument count.
        (KJS::formatDate): Made parameter const.
        (KJS::formatDateUTCVariant): Ditto.
        (KJS::formatTime): Ditto.
        (KJS::DateProtoFuncImp::callAsFunction): Use gmtime_r and localtime_r instead of gmtime and
        localtime.
        (KJS::DateObjectImp::callAsFunction): Use localtime_r instead of localtime.
        (KJS::ymdhmsToSeconds): Renamed from ymdhms_to_seconds. Changed computation to avoid possible
        overflow if year is an extremely large or small number.
        (KJS::makeTime): Removed code to move large month numbers from tm_mon to tm_year; this was
        to accomodate CFGregorianDate, which is no longer used (and didn't handle negative values).
        (KJS::parseDate): Renamed from KRFCDate_parseDate; changed to return a value in milliseconds
        rather than in seconds. Reformatted the code. Changed to use UTF8String() instead of ascii(),
        since ascii() is not thread safe. Changed some variables back from int to long to avoid
        trouble if the result of strtol does not fit in an int (64-bit issue only).

LayoutTests:

        - added a test for http://bugzilla.opendarwin.org/show_bug.cgi?id=5280
          Date.setMonth fails with negative values

        * fast/js/date-negative-setmonth-expected.txt: Added.
        * fast/js/date-negative-setmonth.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10801 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoJavaScriptCore:
darin [Sun, 9 Oct 2005 04:45:29 +0000 (04:45 +0000)]
JavaScriptCore:

        Reviewed by Geoff.
        Tweaked and landed by Darin.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5266
          Support parenthesized comments in Date.parse()

        * kjs/date_object.cpp:
        (KJS::skipSpacesAndComments): Take a pointer, and advance it past spaces,
        and also past anything enclosed in parentheses.
        (KJS::KRFCDate_parseDate): Use skipSpacesAndComments wherever we formerly had
        code to skip spaces.

LayoutTests:

        - added a test for http://bugzilla.opendarwin.org/show_bug.cgi?id=5266
          Support parenthesized comments in Date.parse()

        * fast/js/date-parse-comments-test-expected.txt: Added.
        * fast/js/date-parse-comments-test.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10800 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago - added a test case for XSLT charset support
darin [Sun, 9 Oct 2005 04:36:58 +0000 (04:36 +0000)]
    - added a test case for XSLT charset support
          http://bugzilla.opendarwin.org/show_bug.cgi?id=5244

        * fast/xsl/xslt-entity-enc.xml: Added.
        * fast/xsl/resources/xslt-entity-enc.xsl: Added.
        * fast/xsl/xslt-entity-enc-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10799 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed, tweaked, and landed by Darin.
darin [Sun, 9 Oct 2005 04:25:00 +0000 (04:25 +0000)]
    Reviewed, tweaked, and landed by Darin.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5189
          pcre_exec.c fails to compile using MSVC
        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5190
          KJS config.h adjustment for Win32

        * kjs/config.h: Make sure HAVE_MMAP and HAVE_SBRK are off for Win32.
        Turn HAVE_ERRNO_H on for Mac OS X. Sort defines so they are easy to compare
        with each other. Remove #undef of DEBUG_COLLECTOR.
        * pcre/pcre_exec.c: (match): Work around strange MSVC complaint by splitting
        the definition of a local variable into a separate declaration and
        initialization.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10798 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoWebCore:
darin [Sun, 9 Oct 2005 04:17:42 +0000 (04:17 +0000)]
WebCore:

        Reviewed by Maciej.

        - support for WebKit fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=5187
          UTF-8 in long text files breaks at some point

        No layout test added: only affects plain text view and requires loading the
        file in multiple chunks.

        * WebCore.exp: Added WebCoreTextDecoder.
        * kwq/WebCoreTextDecoder.h: Added.
        * kwq/WebCoreTextDecoder.mm: Added.
        * WebCore.xcodeproj/project.pbxproj: Added files, made header private (SPI).

WebKit:

        Reviewed by Maciej.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5187
          UTF-8 in long text files breaks at some point

        * WebView.subproj/WebTextRepresentation.m:
        (-[WebTextRepresentation finishedLoadingWithDataSource:]):
        Call flushReceivedData on the WebTextView so it can decode any final bytes.

        * WebView.subproj/WebTextView.h: Added WebCoreTextDecoder field and flushReceivedData method.
        * WebView.subproj/WebTextView.m:
        (-[WebTextView dealloc]): Release WebCoreTextDecoder.
        (-[WebTextView appendReceivedData:fromDataSource:]): Create a WebCoreTextDecoder to decode
        the text; use the textEncodingName from the data source. Use it to decode instead of the
        data source's stringWithData.
        (-[WebTextView flushReceivedData]): Call flush on the decoder and append any last bytes to
        the text view.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10797 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Hyatt.
darin [Sun, 9 Oct 2005 03:50:13 +0000 (03:50 +0000)]
    Reviewed by Hyatt.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5148
          Rendering regressions in the Acid2 test cannot be detected

        * fast/css/acid2-pixel-expected.checksum: Added.
        * fast/css/acid2-pixel-expected.png: Added.
        * fast/css/acid2-pixel-expected.txt: Added.
        * fast/css/acid2-pixel.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10796 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Geoff Garen.
darin [Sun, 9 Oct 2005 03:37:08 +0000 (03:37 +0000)]
    Reviewed by Geoff Garen.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5061
          date-big-setmonth test is not timezone-independent

        * fast/js/date-big-setmonth.html: Changed to compare expected result with a Date value rather
        than with a constant. Also write out final result as a date rather than as a constant.
        * fast/js/date-big-setmonth-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10795 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLayoutTests:
darin [Sun, 9 Oct 2005 03:31:11 +0000 (03:31 +0000)]
LayoutTests:

        - Added a test for http://bugzilla.opendarwin.org/show_bug.cgi?id=4821
          Text in submitted forms should be entity-encoded if the current encoding doesn't support it

        * fast/forms/form-data-encoding-expected.txt: Added.
        * fast/forms/form-data-encoding.html: Added.

WebCore:

        Reviewed by Maciej.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=4821
          Text in submitted forms should be entity-encoded if the current encoding doesn't support it

        Added one layout test.

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLFormElementImpl::formData): Pass true to the fromUnicode function to request entity encoding.
        (DOM::FormDataList::appendString): Ditto.

        * kwq/KWQTextCodec.h: Add boolean parameter to request entity-style encoding for characters that can't
        be expressed in the desired encoding.
        * kwq/KWQTextCodec.mm: (QTextCodec::fromUnicode): Added code to entity-encode characters that can't be
        expressed in the desired encoding.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLayoutTests:
darin [Sun, 9 Oct 2005 02:57:21 +0000 (02:57 +0000)]
LayoutTests:

        Reviewed by Darin.

        - Added a test for http://bugzilla.opendarwin.org/show_bug.cgi?id=4707
          Need a way to automatically test for regressions in NSTextInput implementation

        * editing/input/text-input-controller-expected.txt: Added.
        * editing/input/text-input-controller.html: Added.

WebKitTools:

        Reviewed, rearranged and landed by Darin.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=4707
          Need a way to automatically test for regressions in NSTextInput implementation

        * DumpRenderTree/DumpRenderTree.m:
        (-[WaitUntilDoneDelegate webView:windowScriptObjectAvailable:]):
        Create a text input controller and put it in a property of the window object.

        * DumpRenderTree/TextInputController.h: Added.
        * DumpRenderTree/TextInputController.m: Added.
        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added files.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10793 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Darin.
darin [Sun, 9 Oct 2005 02:36:14 +0000 (02:36 +0000)]
    Reviewed by Darin.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=3852
          typeahead doesn't work in multiple row select boxes.

        * kwq/KWQListBox.mm:
        (KWQTableViewTypeSelectCallback): Added.
        (-[KWQTableView finalize]): Deallocate the UCTypeSelect object.
        (-[KWQTableView dealloc]): Ditto.
        (-[KWQTableView keyUp:]): If character typed is a graphic character, pass it along to
        the UCTypeSelect object and use it to type select, otherwise flush the object.

        * manual-tests/select-element-type-select.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10792 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLayoutTests:
darin [Sun, 9 Oct 2005 02:14:07 +0000 (02:14 +0000)]
LayoutTests:

        - Updated test results for fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=3245
          Support the 'bdo' element

        * fast/dom/HTMLElement/bdo.html: Updated test, no longer says to expect failure.
        * fast/dom/HTMLElement/bdo-expected.checksum: Correct results.
        * fast/dom/HTMLElement/bdo-expected.png: Ditto.
        * fast/dom/HTMLElement/bdo-expected.txt: Ditto.

        * fast/text/international/bidi-LDB-2-HTML-expected.checksum: Updated to expect correct
        results for the section done with <bdo>.
        * fast/text/international/bidi-LDB-2-HTML-expected.png: Ditto.
        * fast/text/international/bidi-LDB-2-HTML-expected.txt: Ditto.

WebCore:

        Reviewed by Darin.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=3245
          Support the 'bdo' element

        Fixes the results of some existing layout tests.

        * khtml/xml/dom_elementimpl.h: Add eBDO to the list of entry types so there's a way to
        keep <bdo> dir attribute mappings separate from those of other elements.

        * khtml/html/html_elementimpl.cpp:
        (HTMLElementImpl::mapToEntry): Add special case to use eBDO for the dir attribute on
        elements with the <bdo> tag.
        (HTMLElementImpl::parseMappedAttribute): Use bidi-override for the unicode-bidi CSS property
        when setting it based on the dir attribute on a <bdo> tag. Along with the change above this
        obviates the need for an HTMLElementImpl subclass for <bdo>, but we could refactor it that
        way later if we decide there's an advantage to doing it that way.

        * khtml/css/html4.css: Remove the rules about <bdo>. They are insufficient, and now
        they are also unnecessary with the changes above.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10791 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoLayoutTests:
darin [Sun, 9 Oct 2005 01:46:17 +0000 (01:46 +0000)]
LayoutTests:

        Test cases for http://bugzilla.opendarwin.org/show_bug.cgi?id=4898
        (Failures in dbaron's bidi ordering tests).

        * fast/dom/HTMLElement/bdo-expected.txt: Updated for differently-breaking text runs.
        * fast/text/international/bidi-override-expected.txt: Updated because override characters no
        longer show up in the render tree.

        * fast/text/international/bidi-LDB-2-CSS-expected.checksum: Added.
        * fast/text/international/bidi-LDB-2-CSS-expected.png: Added.
        * fast/text/international/bidi-LDB-2-CSS-expected.txt: Added.
        * fast/text/international/bidi-LDB-2-CSS.html: Added.

        * fast/text/international/bidi-LDB-2-HTML-expected.checksum: Added.
        * fast/text/international/bidi-LDB-2-HTML-expected.png: Added.
        * fast/text/international/bidi-LDB-2-HTML-expected.txt: Added.
        * fast/text/international/bidi-LDB-2-HTML.html: Added.

        * fast/text/international/bidi-LDB-2-formatting-characters-expected.checksum: Added.
        * fast/text/international/bidi-LDB-2-formatting-characters-expected.png: Added.
        * fast/text/international/bidi-LDB-2-formatting-characters-expected.txt: Added.
        * fast/text/international/bidi-LDB-2-formatting-characters.html: Added.

        * fast/text/international/bidi-explicit-embedding-expected.checksum: Updated to now-correct results.
        * fast/text/international/bidi-explicit-embedding-expected.png: Ditto.
        * fast/text/international/bidi-explicit-embedding-expected.txt: Ditto.

WebCore:

        Reviewed by Darin.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=4898
          Failures in dbaron's bidi ordering tests

        Added 3 dbaron bidi ordering tests to LayoutTests.

        * khtml/rendering/bidi.cpp:
        (khtml::deleteBidiRuns):
        (khtml::Bidinext):
        (khtml::appendRunsForObject):
        (khtml::embed):
        (khtml::RenderBlock::bidiReorderLine):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5240
eseidel [Sat, 8 Oct 2005 10:45:06 +0000 (10:45 +0000)]
Bug #: 5240
Submitted by: Rob Buis  <rwlbuis@xs4all.nl>
Reviewed by: eseidel
        Almost forgot the layout test for the last commit. ;)
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5240

        * svg-tests/custom/use-forward-refs-expected.checksum: Added.
        * svg-tests/custom/use-forward-refs-expected.png: Added.
        * svg-tests/custom/use-forward-refs-expected.txt: Added.
        * svg-tests/custom/use-forward-refs.svg: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10789 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years agoBug #: 5240
eseidel [Sat, 8 Oct 2005 09:18:15 +0000 (09:18 +0000)]
Bug #: 5240
Submitted by: Rob Buis  <rwlbuis@xs4all.nl>
Reviewed by: eseidel
        Fix to handle forward referencing <use> elements.
        http://bugzilla.opendarwin.org/show_bug.cgi?id=5240

        * ksvg2/svg/SVGDocumentImpl.cpp:
        (SVGDocumentImpl::executeScripts): delayed close.
        (SVGDocumentImpl::addForwardReference): new function.
        * ksvg2/svg/SVGDocumentImpl.h:
        * ksvg2/svg/SVGUseElementImpl.cpp:
        (SVGUseElementImpl::close): delay close, if missing reference.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10788 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Justin.
harrison [Sat, 8 Oct 2005 00:51:25 +0000 (00:51 +0000)]
    Reviewed by Justin.

        Adjusted expected results after fixing position code in
        <rdar://problem/4064017> "Safari crashes at -[WebCoreBridge firstRectForDOMRange:] + 92"

        * editing/deleting/delete-at-paragraph-boundaries-002-expected.txt:
        * editing/deleting/delete-at-paragraph-boundaries-004-expected.txt:
        * editing/deleting/delete-at-paragraph-boundaries-005-expected.txt:
        * editing/selection/3690703-expected.txt:
        * editing/selection/3690719-expected.txt:
        * editing/selection/select-all-004-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Justin.
harrison [Sat, 8 Oct 2005 00:50:30 +0000 (00:50 +0000)]
    Reviewed by Justin.

        "<rdar://problem/4064017> Safari crashes at -[WebCoreBridge firstRectForDOMRange:] + 92"

        * khtml/editing/delete_selection_command.cpp:
        (khtml::DeleteSelectionCommand::insertPlaceholderForAncestorBlockContent):
        Do not insert placeholder if selection ends at a BR.

        (khtml::DeleteSelectionCommand::handleGeneralDelete):
        No need to preserve starting BR because insertPlaceholderForAncestorBlockContent already did.

        * khtml/xml/dom_position.cpp:
        (DOM::Position::upstream):
        (DOM::Position::downstream):
        Fixed to return original position instead of invisible position when no suitable position found upstream.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

17 years ago Reviewed by Hyatt.
vicki [Fri, 7 Oct 2005 18:38:46 +0000 (18:38 +0000)]
    Reviewed by Hyatt.

- pass the current form to the isindex constructor, so that isindex doesn't malfunction in the
case of a misnested form

        * khtml/html/htmlparser.cpp:
        (HTMLParser::handleIsindex):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@10785 268f45cc-cd09-0410-ab3c-d52691b4dbfc