[WebKit-https.git] / WebCore / ChangeLog-2005-08-23

2005-02-28  David Hyatt  <hyatt@apple.com>

        Fix for 4028999, safari crashes when resetting if mallocsribble is on.  Clip rects were being cleared using dead
        render objects.  Change the ordering.
        
        Reviewed by rjw

        * khtml/rendering/render_box.cpp:
        (RenderBox::detach):
        * khtml/rendering/render_layer.cpp:
        (RenderLayer::~RenderLayer):

2005-02-28  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/4024786> REGRESSION (Mail): "Smart paste" plain-text word into Blot leaves insertion point misplaced

        Reviewed by kocienda.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::completeHTMLReplacement): Call updateLayout so caretMinOffset and caretMaxOffset return correct values.

2005-02-28  John Sullivan  <sullivan@apple.com>

        Reviewed by Ken.
        
        - WebCore part of fix for <rdar://problem/4023490> REGRESSION (125-185): Tabbing through links 
        on frameset page gets stuck at end (tivofaq.com)

        * kwq/WebCoreBridge.h:
        add nextValidKeyViewOutsideWebFrameViews (code is in WebKit)

2005-02-28  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3972665> 8A360: HTML message partially truncated on left hand side, text-indent from Script Editor

        * khtml/css/css_computedstyle.cpp: inheritableProperties array now defined in css_valueimpl.cpp.
        * khtml/css/css_valueimpl.cpp: Define inheritableProperties array here.
        (DOM::CSSMutableStyleDeclarationImpl::copyBlockProperties): Use new name for blockProperties, and use the new
        constant for the number of items in the array.
        (DOM::CSSMutableStyleDeclarationImpl::removeBlockProperties): Ditto.
        (DOM::CSSMutableStyleDeclarationImpl::removeInheritableProperties): New function.
        * khtml/css/css_valueimpl.h: Declare inheritableProperties array and numInheritableProperties extern so they
        can be defined in css_valueimpl.cpp and used in css_computedstyle.cpp.
        * khtml/editing/htmlediting.cpp:
        (khtml::ReplacementFragment::removeStyleNodes): This code was misguided, and removed too much style from HTML
        elements. Now, it removes from HTML elements only the styles that we replace later with a call to applyStyle().
        Also, add ID_B to list of inline "style" nodes we are willing to remove. Leaving it off was an oversight.

        * layout-tests/editing/pasteboard/paste-text-011-expected.txt: ID_B fix made this result change, without any
        visible change in the test.
        
        New test:
        
        * layout-tests/editing/style/smoosh-styles-003.html
        * layout-tests/editing/style/smoosh-styles-003-expected.txt

2005-02-28  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/4026985> CrashTracer: ...14 crashes at com.apple.WebCore: -[KWQPageState invalidate] + 32

        Added more nil checking and ASSERTS.
        Without a reproducible case this is hard to definitively resolve.

        Reviewed by John Sullivan.

        * kwq/KWQPageState.mm:
        (-[KWQPageState invalidate]):

2005-02-28  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/4027702> 3.5% performance regression btwn Safari-188 and Safari-400

        I inadvertently checked in some debugging code that disabled
        style sharing.  Backed out that change.

        * khtml/css/cssstyleselector.cpp:
        (khtml::CSSStyleSelector::styleForElement):

2005-02-28  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/4026639> www.bmw.ca configurator does not work with Safari

        Reviewed by john.

        * khtml/ecma/kjs_html.cpp:
        (KJS::HTMLElement::tryGet): when frameset.<name of frame child> is called, return the window object of the frame child

2005-02-28  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Chris

        Fix for this bug:
        
        <rdar://problem/4026906> Paste of HTML table content can break table structure

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplacementFragment::pruneEmptyNodes): Call new isProbablyTableStructureNode() function to prevent
        removal of empty table structure nodes.
        (khtml::ReplacementFragment::removeUnrenderedNodesUsingTestRendering): Ditto.
        (khtml::isProbablyTableStructureNode): New helper function.
        * khtml/editing/htmlediting.h: Declare new helper.

2005-02-28  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/4023566> Stickies: Crash in ReplacementFragment::insertFragmentForTestRendering on paste

        Reviewed by kocienda.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): ref and deref the element while it is "floating"
        (khtml::ReplacementFragment::insertFragmentForTestRendering): ditto
        (khtml::floatRefdElement): new, keeps an element alive while its ref count is 0
        (khtml::createDefaultParagraphElement): removed commented out code
        (khtml::createBlockPlaceholderElement): ref the element and return it as "floating"
        (khtml::createFontElement): ditto
        (khtml::createStyleSpanElement): ditto
        * khtml/editing/htmlediting.h:

2005-02-27  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Vicki.

        <rdar://problem/3993557> REGRESSION (125-180-ish): getElementsByTagName no longer works with namespace designations
        
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeBaseImpl::getElementsByTagNameNS): When no namespace is specified, find elements
        of the specified name in any namespace to match Mozilla and earlier Safari behavior.

2005-02-25  Darin Adler  <darin@apple.com>

        Reviewed by Adele.

        - fixed <rdar://problem/4025618> Crash while searching at hollywoodvideo.com

        * khtml/html/html_formimpl.h: Added valueWithDefault.
        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLInputElementImpl::appendFormData): Call valueWithDefault instead of going at the render object
        to try to get the default value; there may be no render object if this is display:none.
        (DOM::HTMLInputElementImpl::valueWithDefault): Added. Knows about the defaults for "submit" and "reset"
        buttons; otherwise just returns the value as-is.

        * khtml/rendering/render_form.h: Removed the defaultLabel member functions.
        * khtml/rendering/render_form.cpp:
        (RenderSubmitButton::rawText): Call valueWithDefault instead of using defaultLabel function here on
        the render side. The DOM needs to know how to deal with the default anyway for form submission.
        (RenderSubmitButton::defaultLabel): Removed.
        (RenderResetButton::defaultLabel): Removed.
        (RenderPushButton::defaultLabel): Removed.

2005-02-25  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4025088> window onblur and onfocus don't fire when text field has focus

        * kwq/KWQKHTMLPart.h: Added setWindowHasFocus function and m_windowHasFocus data member.
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::setDisplaysWithFocusAttributes): Took out the code that sends the focus and blur events.
        (KWQKHTMLPart::setWindowHasFocus): Put that code here instead.

        * kwq/WebCoreBridge.h: Added setWindowHasFocus: method to the bridge.
        * kwq/WebCoreBridge.mm: (-[WebCoreBridge setWindowHasFocus:]): Added. Calls method on the part.

2005-02-25  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - re-fixed <rdar://problem/3665430> horizontal scroll bar of text area does not show, even when text is wide in "no wrap" mode

        * kwq/KWQTextArea.mm: (-[KWQTextArea _updateTextViewWidth]): Don't change the text view width
        to match the text area's width in the "wrap" case.

2005-02-25  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/4021518> 8A394 Mail crashes during paste: khtml::RootInlineBox::closestLeafChildForXPos

        * khtml/editing/visible_units.cpp:
        (khtml::previousLinePosition): Adding an updateLayout call at the start of the function fixes the crash, since
        we caught line boxes in a not-completely-updated state.
        (khtml::nextLinePosition): Ditto.

2005-02-25  David Hyatt  <hyatt@apple.com>

        Fix for 4010774, make sure to avoid an O(N^2) algorithm in nextRenderer() that is triggered when large
        DOM subtrees are inserted into documents via one insert/append call.
        
        Reviewed by kocienda

        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::nextRenderer):

2005-02-25  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3382926> Bidi neutrals at RTL/LTR boundaries not handled correctly.

        Second pass at fixing 3382926 w/o causing layout regressions.  Same concept:
        if directionality of text's element is RTL and first character has neutral directionality
        then set the initial directionality to RTL.

        Reviewed by Hyatt.

        * khtml/rendering/bidi.cpp:
        (khtml::RenderBlock::bidiReorderLine):
        (khtml::RenderBlock::determineStartPosition):

=== Safari-400 ===

2005-02-25  Ken Kocienda  <kocienda@apple.com>

        Reviewed by me

        Roll out Chris' change to fix this bug:
        
        <rdar://problem/4023566> Stickies: Crash in ReplacementFragment::insertFragmentForTestRendering on paste
        
        That code change is responsible for all these new crashers:
        
        <rdar://problem/4025177> crash copying safari.apple.com into Blot document
        <rdar://problem/4025184> crash in DOM::NodeImpl::parentNode copying "New!" from google.com to Blot
        <rdar://problem/4025214> crash in DOM::NodeImpl::getRect loading paste-match-style-001.html

        Since we wish to close the tree for a build right now, I am rolling out rather than investigating.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded)
        (khtml::ReplacementFragment::insertFragmentForTestRendering)
        (khtml::createDefaultParagraphElement)
        (khtml::createBlockPlaceholderElement)
        (khtml::createFontElement)
        (khtml::createStyleSpanElement)
        * khtml/editing/htmlediting.h

2005-02-25  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/4020108> Pasting text into message makes Mail crash reproducibly

        Note that, even with this fix, development build will crash until this bug is fixed:
        <rdar://problem/4024996> Applying block styles can cause assertion failure in inline style removal
        
        This will not crash deployment builds, so I am going to land.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::applyBlockStyle): Applying block styles can make the loop to reach beyondEnd
        fail since the structure of the document can change. Cache the next node first before operating on it,
        as we do elsewhere.

2005-02-25  Vicki Murley <vicki@apple.com>

        - recommit this change, since rolling it out did NOT fix the performance regression!

    2005-02-23  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4011405> REGRESSION (180-181): Unconfirmed text disappears when text focus moves

        The key was to change things around so that we don't push text from the DOM to the widget
        unless the DOM has actually been changed. This prevents the code path that wipes out inline input
        during the blur process.

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLInputElementImpl::HTMLInputElementImpl): Start m_valueMatchesRenderer as false.
        (DOM::HTMLInputElementImpl::parseHTMLAttribute): Set m_valueMatchesRenderer to false when a
        new value is set here.
        (DOM::HTMLInputElementImpl::setValue): Set m_valueMatchesRenderer to false when a new value
        is set here.
        (DOM::HTMLInputElementImpl::setValueFromRenderer): Added. Sets m_value, sets m_valueMatchesRenderer
        to true, and also sends out the input event. It's better to have this here than in the renderer code.
        (DOM::HTMLTextAreaElementImpl::HTMLTextAreaElementImpl): Start m_valueIsValid as false (replaces
        m_dirtyvalue) and m_valueMatchesRenderer as false.
        (DOM::HTMLTextAreaElementImpl::updateValue): Added. Factored this out from the value function. Uses
        the new booleans and keeps them up to date, specifically setting m_valueMatchesRenderer based on
        where the value came from.
        (DOM::HTMLTextAreaElementImpl::value): Updated to call updateValue to do most of the work.
        (DOM::HTMLTextAreaElementImpl::setValue): Set both m_valueIsValid and m_valueMatchesRenderer.
        (DOM::HTMLTextAreaElementImpl::setDefaultValue): Take parameter by reference.

        * khtml/html/html_formimpl.h: Added setValueFromRenderer, valueMatchesRenderer, setValueMatchesRenderer,
        and m_valueMatchesRenderer to input element. For textarea element, made some parameters pass DOMString
        by reference, and added invalidateValue, updateValue, valueMatchesRenderer, and setValueMatchesRenderer.

        * khtml/rendering/render_form.cpp:
        (RenderLineEdit::updateFromElement): Don't re-get the value from the DOM if valueMatchesRenderer
        is true.
        (RenderLineEdit::slotTextChanged): Call setValueFromRenderer instead of manipulating the DOM
        directly.
        (RenderTextArea::detach): Call updateValue instead of calling value for its side effect.
        (RenderTextArea::handleFocusOut): Ditto.
        (RenderTextArea::updateFromElement): Call updateValue and then not re-get the value from the
        DOM if valueMatchesRenderer is true.
        (RenderTextArea::slotTextChanged): Call invalidateValue instead of directly setting m_dirtyvalue to true.

2005-02-25  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/4023566> Stickies: Crash in ReplacementFragment::insertFragmentForTestRendering on paste

        Reviewed by darin.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded):
        (khtml::ReplacementFragment::insertFragmentForTestRendering):
        (khtml::floatRefdElement):
        (khtml::createDefaultParagraphElement):
        (khtml::createBlockPlaceholderElement):
        (khtml::createFontElement):
        (khtml::createStyleSpanElement):
        * khtml/editing/htmlediting.h:

2005-02-25  Darin Adler  <darin@apple.com>

        * kwq/character-sets.txt: Checked in updated file. This new file has no effect, because none of the
        changes affect character sets that we support, but it's good to have the latest file in here, and
        completely safe because I checked that the generated files have not changed.

2005-02-25  Vicki Murley <vicki@apple.com>

        Reviewed by Darin.
 
        - back out this change, since it causes a 3.5% performance regression

    2005-02-23  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4011405> REGRESSION (180-181): Unconfirmed text disappears when text focus moves

        The key was to change things around so that we don't push text from the DOM to the widget
        unless the DOM has actually been changed. This prevents the code path that wipes out inline input
        during the blur process.

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLInputElementImpl::HTMLInputElementImpl): Start m_valueMatchesRenderer as false.
        (DOM::HTMLInputElementImpl::parseHTMLAttribute): Set m_valueMatchesRenderer to false when a
        new value is set here.
        (DOM::HTMLInputElementImpl::setValue): Set m_valueMatchesRenderer to false when a new value
        is set here.
        (DOM::HTMLInputElementImpl::setValueFromRenderer): Added. Sets m_value, sets m_valueMatchesRenderer
        to true, and also sends out the input event. It's better to have this here than in the renderer code.
        (DOM::HTMLTextAreaElementImpl::HTMLTextAreaElementImpl): Start m_valueIsValid as false (replaces
        m_dirtyvalue) and m_valueMatchesRenderer as false.
        (DOM::HTMLTextAreaElementImpl::updateValue): Added. Factored this out from the value function. Uses
        the new booleans and keeps them up to date, specifically setting m_valueMatchesRenderer based on
        where the value came from.
        (DOM::HTMLTextAreaElementImpl::value): Updated to call updateValue to do most of the work.
        (DOM::HTMLTextAreaElementImpl::setValue): Set both m_valueIsValid and m_valueMatchesRenderer.
        (DOM::HTMLTextAreaElementImpl::setDefaultValue): Take parameter by reference.

        * khtml/html/html_formimpl.h: Added setValueFromRenderer, valueMatchesRenderer, setValueMatchesRenderer,
        and m_valueMatchesRenderer to input element. For textarea element, made some parameters pass DOMString
        by reference, and added invalidateValue, updateValue, valueMatchesRenderer, and setValueMatchesRenderer.

        * khtml/rendering/render_form.cpp:
        (RenderLineEdit::updateFromElement): Don't re-get the value from the DOM if valueMatchesRenderer
        is true.
        (RenderLineEdit::slotTextChanged): Call setValueFromRenderer instead of manipulating the DOM
        directly.
        (RenderTextArea::detach): Call updateValue instead of calling value for its side effect.
        (RenderTextArea::handleFocusOut): Ditto.
        (RenderTextArea::updateFromElement): Call updateValue and then not re-get the value from the
        DOM if valueMatchesRenderer is true.
        (RenderTextArea::slotTextChanged): Call invalidateValue instead of directly setting m_dirtyvalue to true.

2005-02-25  Darin Adler  <darin@apple.com>

        Reviewed by Chris.

        - fixed <rdar://problem/4024786> REGRESSION (Mail): "Smart paste" plain-text word into Blot leaves insertion point misplaced

        * khtml/editing/htmlediting.cpp: (khtml::ReplaceSelectionCommand::doApply):
        Use the insertNodeAfterAndUpdateNodesInserted and insertNodeBeforeAndUpdateNodesInserted
        functions to add the leading and trailing spaces for smart paste.

2005-02-25  David Hyatt  <hyatt@apple.com>

        Back out fix for 3382926, since it breaks LTR text inside RTL contexts.
        
        Reviewed by john

        * khtml/rendering/bidi.cpp:
        (khtml::BidiIterator::direction):

2005-02-25  David Hyatt  <hyatt@apple.com>

        Fix for 3975039, scrolling is slow in huge RSS views.  Optimize the calculation of clip rects for overflow:hidden
        layers.  Also optimize layer movement when scrolling overflow sections.
        
        Reviewed by darin

        * khtml/rendering/render_layer.cpp:
        (throw):
        (ClipRects::operator delete):
        (ClipRects::detach):
        (RenderLayer::RenderLayer):
        (RenderLayer::~RenderLayer):
        (RenderLayer::updateLayerPosition):
        (RenderLayer::removeOnlyThisLayer):
        (RenderLayer::insertOnlyThisLayer):
        (RenderLayer::scrollToOffset):
        (RenderLayer::hitTest):
        (RenderLayer::calculateClipRects):
        (RenderLayer::calculateRects):
        (RenderLayer::containsPoint):
        (RenderLayer::clearClipRects):
        (RenderLayer::clearClipRect):
        * khtml/rendering/render_layer.h:
        (khtml::ClipRects::m_refCnt):
        (khtml::ClipRects::overflowClipRect):
        (khtml::ClipRects::fixedClipRect):
        (khtml::ClipRects::posClipRect):
        (khtml::ClipRects::ref):
        (khtml::ClipRects::deref):
        (khtml::RenderLayer::clipRects):
        * khtml/rendering/render_object.cpp:
        (RenderObject::setStyle):

2005-02-24  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Darin and Dave a while ago.

        <rdar://problem/3996685> REGRESSION: Crash in KWQVectorImpl::at loading http://maps.google.com/mapfiles/homepanel.xsl
        
        * kwq/WebCoreBridge.mm:
        (formElementFromDOMElement): Check for isHTMLElement() as well as
        id() == ID_FORM. This seems like an impossible situation, but
        papering over it seems more expedient for the time being.

2005-02-24  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3382926> Bidi neutrals at RTL/LTR boundaries not handled correctly.

        Use mirror characters correctly when rendering with RTL directionality.

        Reviewed by Hyatt.

        * khtml/rendering/bidi.cpp:
        (khtml::BidiIterator::direction):

2005-02-24  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3985889> REGRESSION (125-180): setting <img> src to GIF that already animated does not animate; just shows final frame
        
        Reviewed by Hyatt.

        * WebCore.pbproj/project.pbxproj:
        * khtml/html/html_imageimpl.cpp:
        (HTMLImageLoader::updateFromElement):
        * khtml/rendering/render_image.cpp:
        (RenderImage::resetAnimation):
        * khtml/rendering/render_image.h:
        * khtml/rendering/render_list.cpp:
        (RenderListMarker::setStyle):
        (RenderListMarker::paint):
        * khtml/rendering/render_list.h:
        * kwq/KWQPixmap.h:
        * kwq/KWQPixmap.mm:
        (QPixmap::QPixmap):
        (QPixmap::resetAnimation):
        * kwq/WebCoreImageRenderer.h:

2005-02-24  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/4020110> Safari crashes in setAllData while taking a www.zoomerang.com survey

        Reviewed by rjw.

        * kwq/KWQResourceLoader.mm:
        (-[KWQResourceLoader finishJobAndHandle:]): clear the job after we've deleted to avoid reentrancy

2005-02-24  Darin Adler  <darin@apple.com>

        Reviewed by Vicki.

        - fixed <rdar://problem/4023360> REGRESSION (186-187): image file upload is broken at pep.apple.com

        * khtml/html/html_formimpl.cpp: (DOM::HTMLInputElementImpl::appendFormData): Return true after
        setting up the form data for an uploaded file. The old code would fall through to the ISINDEX
        case and send double form data (the filename instead of the file contents the second time).

2005-02-24  David Harrison  <harrison@apple.com>

        Reviewed by Ken.
        
        <rdar://problem/3990849> AX: textMarkerRange for an AXUIElement within an AXWebArea

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject textMarkerRange]):
        (-[KWQAccObject accessibilityParameterizedAttributeNames]):
        (-[KWQAccObject doAXTextMarkerRangeForUIElement:]):
        (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
        Added AXTextMarkerRangeForUIElement, or you can send textMarkerRange to the UIElement itself.

2005-02-24  David Harrison  <harrison@apple.com>

        Reviewed by Vicki.

        <rdar://problem/4004279> 3 AXSelectedTextChanged notifications are firing each time I type a character

        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::updateSelection):
        Send notification only if the selection is not null.  This safely ignores transitory selections set during editing.

2005-02-24  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3987619> in some cases, text doesn't resize with Format->Style->Bigger/Smaller

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::applyRelativeFontStyleChange): Only call nodeFullySelected for non-text nodes.
        Text nodes are already split so they're either in the range and full selected or out of the range.
        And nodeFullySelected doesn't work for text nodes.
        (khtml::ApplyStyleCommand::nodeFullySelected): Add an assertion, since this function only works for elements,
        not text nodes.
        (khtml::ApplyStyleCommand::nodeFullyUnselected): Ditto.

        - make big improvement in <rdar://problem/3953636> Mail hung for ~10sec changing font of 84328 characters: khtml::ApplyStyleCommand::nodeFullySelected

        * khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::compareBoundaryPoints): Improve algorithm based on suggestion
        from Nate Begeman.

        - fixed <rdar://problem/4020305> REGRESSION (185-186): loading image in new window using document.write fails

        * khtml/xml/dom_docimpl.cpp: (DocumentImpl::open): If there is no parent document, don't blow away the base URL.

        - fixed <rdar://problem/4021701> REGRESSION (188-188+): form not submitted after pressing <return> at http://hrweb.apple.com
        
        * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::dispatchMouseEvent): Send activate event in the case where the event
        sent is a KHTML_CLICK_EVENT, not CLICK_EVENT.

2005-02-23  Kevin Decker  <kdecker@apple.com>

        Reviewed by mjs  
        
        Fixed <rdar://problem/4020747> REGRESSION: stray </applet> tags crash Safari

        * khtml/html/htmlparser.cpp:
        (KHTMLParser::processCloseTag):  Made a typesafe check that prevents crashes
        whenever there is a closing applet tag without an actual <applet> in the first place. 
        Now verifies the current token is ID_APPLET before casting it to an HTMLAppletElementImpl.
         

2005-02-23  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3977962> font loses bold style after pasting next to existing text and pressing return

        * khtml/editing/htmlediting.cpp:
        (khtml::InsertParagraphSeparatorCommand::calculateStyleBeforeInsertion): Problem here was
        that we were doing work in cases where we should not, and content whose style would have
        been correct if we had done nothing was getting clobbered. It turns out that extra work
        to apply style to the new paragraph added in this command only needs to be done if we're 
        at the boundaries of a paragraph. Otherwise, content that is moved as part of the work 
        of the command will lend their styles to the new paragraph without any extra work needed.
        So, make this position check and return unless at a paragraph boundary.

        New test:

        * layout-tests/editing/style/block-styles-007-expected.txt
        * layout-tests/editing/style/block-styles-007.html

2005-02-23  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/4017641> REGRESSION (Mail): you can only bold/unbold a selection starting from end of line once

        Problem is with the way we figure out whether to add or remove a style based on
        the current selection. In this case, the code is looking at the end of the
        previous line, which is not bold, and deduces incorrectly that the operation is
        a "make bold". Then the style code runs to make bold, but there is nothing on
        the end of the previous line to embolden, so we get into a cycle where the same
        thing happens each time cmd-b is hit.

        * khtml/khtml_part.cpp:
        (KHTMLPart::selectionComputedStyle):  Call editingStartPosition() to get the right position for the font determination.
        * khtml/xml/dom2_rangeimpl.cpp:
        (DOM::RangeImpl::editingStartPosition): New helper function that "does the right thing" based on whether the
        selection is a caret or a range, moving upstream for the former, and downstream for the latter.
        * khtml/xml/dom2_rangeimpl.h:
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::fontForSelection): Call editingStartPosition() to get the right position for the font determination.

        New tests:

        * layout-tests/editing/style/style-boundary-001-expected.txt
        * layout-tests/editing/style/style-boundary-001.html
        * layout-tests/editing/style/style-boundary-002-expected.txt
        * layout-tests/editing/style/style-boundary-002.html
        * layout-tests/editing/style/style-boundary-003-expected.txt
        * layout-tests/editing/style/style-boundary-003.html
        * layout-tests/editing/style/style-boundary-004-expected.txt
        * layout-tests/editing/style/style-boundary-004.html

2005-02-23  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3985579> 8A367: Dashboard: Stock widget not visible when click remove to remove single char ticker symbol

        Explicitly remove scroll bar views when removing them from 
        the overflow.

        Don't paint synchronously when the scroll position changes,
        this caused funky clip problems.

        Reviewed by Hyatt.

        * khtml/rendering/render_layer.cpp:
        (RenderLayer::scrollToOffset):
        (RenderLayer::setHasHorizontalScrollbar):
        (RenderLayer::setHasVerticalScrollbar):
        (RenderLayer::updateScrollInfoAfterLayout):

2005-02-23  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3959996> REGRESSION (Mail): cursor moves to beginning of document when click is past end

        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::positionForCoordinates): Skip blocks that are invisible or have no height when 
        looking for a child to pass off to. And save away the last visible block with a height to pass off
        to if there is no child at the right y-coordinate.

2005-02-23  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/4010059> BoundsForTextMarkerRange does not update with scrolled web area

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject doAXBoundsForTextMarkerRange:]):
        Adjust for scrolling.

2005-02-22  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Ken.

        <rdar://problem/3949790> hitting return after pasted styled line results in extra content getting the style
        
        * khtml/editing/htmlediting.cpp:
        (khtml::InsertParagraphSeparatorCommand::doApply): In the case
        where the start block is the root, insert the newly created DIV at
        the end of the root block instead of after the last sibling in the
        start node, since the start node could be inside other
        style-affecting nodes and we don't want to reparent its cousins
        into that.

2005-02-23  David Harrison  <harrison@apple.com>

        Reviewed by Vicki.

        <rdar://problem/4014691> switch to correctly spelled NSAccessibilityForegroundColorTextAttribute constant
        
        Also removed two older, now unneeded, wrappers.

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject visiblePositionForStartOfTextMarkerRange:]):
        (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
        Removed these older, now unneeded, wrappers.
        
        (NSAccessibilityForegroundColorTextAttributeWrapper):
        New wrapper for NSAccessibilityForegroundColorTextAttribute.
        
        (AXAttributeStringSetStyle):
        Use NSAccessibilityForegroundColorTextAttributeWrapper.
        
        (-[KWQAccObject doAXTextMarkerRangeForUnorderedTextMarkers:]):
        Remove uses of visiblePositionForStartOfTextMarkerRange and visiblePositionForEndOfTextMarkerRange.

2005-02-23  David Harrison  <harrison@apple.com>

        Reviewed by Vicki.

        <rdar://problem/3524784> AX hit test doesn't return info when done in empty space of content area

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject accessibilityHitTest:]):
        Return unignored object.

2005-02-23  Darin Adler  <darin@apple.com>

        Reviewed by Ken.

        - fixed <rdar://problem/4006509> REGRESSION (171-172): Setting CSS -khtml-user-modify property triggers crash

        * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::applyProperty):
        Removed the code that changes the style of the element. This was never needed, and caused the
        style to be modified while we were iterating it.

2005-02-23  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3980209> Mail crashed when I pressed Cmd-Shift-[ (nil-deref in ApplyStyleCommand::addBlockStyleIfNeeded)

        * khtml/editing/htmlediting.cpp:
        (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): Reordered the new block
        insertion so that it come before the move. The logic stays exactly the same, however, with the old
        ordering, the new block could want to become a child of itself come insertion time. I considered
        making a more complicated code change to fix this problem, but the simple reordering works just
        as well, and seems less risky.
        
        These all changed in an insignificant way. It seems that with the new code, some empty text nodes
        got reordered in the document. This has no effect on anything visible to the user.
        
        * layout-tests/editing/style/create-block-for-style-003-expected.txt
        * layout-tests/editing/style/create-block-for-style-004-expected.txt
        * layout-tests/editing/style/create-block-for-style-009-expected.txt
        * layout-tests/editing/style/create-block-for-style-011-expected.txt
        * layout-tests/editing/style/create-block-for-style-013-expected.txt

2005-02-23  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4013986> REGRESSION (173-174): onclick event not sent when mouse click on checkbox is double-click

        We need to send an onclick event *and* an ondblclick event when we process a double click.

        * khtml/khtmlview.cpp: (KHTMLView::dispatchMouseEvent): In the case where we're sending a CLICK_EVENT,
        follow it by a KHTML_CLICK_EVENT, a KHTML_DBLCLICK_EVENT when handling a double click, and a DOMACTIVATE_EVENT.
        We no longer do the DOMACTIVATE_EVENT in dispatchGenericEvent.
        * khtml/rendering/render_form.cpp: (RenderFormElement::slotClicked): send only CLICK_EVENT here, and
        lets dispatchMouseEvent deal with the other subsequent events.
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::dispatchGenericEvent): Remove the code that sends a DOMACTIVATE_EVENT, since there's no longer
        a good way to figure out if this is the last event that should be sent before it is.
        (NodeImpl::dispatchMouseEvent): Set the meta key modifier here (as it already is set elsewhere), and
        follow up a CLICK_EVENT with KHTML_CLICK_EVENT, a KHTML_DBLCLICK_EVENT, and DOMACTIVATE_EVENT as above.

2005-02-23  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4011405> REGRESSION (180-181): Unconfirmed text disappears when text focus moves

        The key was to change things around so that we don't push text from the DOM to the widget
        unless the DOM has actually been changed. This prevents the code path that wipes out inline input
        during the blur process.

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLInputElementImpl::HTMLInputElementImpl): Start m_valueMatchesRenderer as false.
        (DOM::HTMLInputElementImpl::parseHTMLAttribute): Set m_valueMatchesRenderer to false when a
        new value is set here.
        (DOM::HTMLInputElementImpl::setValue): Set m_valueMatchesRenderer to false when a new value
        is set here.
        (DOM::HTMLInputElementImpl::setValueFromRenderer): Added. Sets m_value, sets m_valueMatchesRenderer
        to true, and also sends out the input event. It's better to have this here than in the renderer code.
        (DOM::HTMLTextAreaElementImpl::HTMLTextAreaElementImpl): Start m_valueIsValid as false (replaces
        m_dirtyvalue) and m_valueMatchesRenderer as false.
        (DOM::HTMLTextAreaElementImpl::updateValue): Added. Factored this out from the value function. Uses
        the new booleans and keeps them up to date, specifically setting m_valueMatchesRenderer based on
        where the value came from.
        (DOM::HTMLTextAreaElementImpl::value): Updated to call updateValue to do most of the work.
        (DOM::HTMLTextAreaElementImpl::setValue): Set both m_valueIsValid and m_valueMatchesRenderer.
        (DOM::HTMLTextAreaElementImpl::setDefaultValue): Take parameter by reference.

        * khtml/html/html_formimpl.h: Added setValueFromRenderer, valueMatchesRenderer, setValueMatchesRenderer,
        and m_valueMatchesRenderer to input element. For textarea element, made some parameters pass DOMString
        by reference, and added invalidateValue, updateValue, valueMatchesRenderer, and setValueMatchesRenderer.

        * khtml/rendering/render_form.cpp:
        (RenderLineEdit::updateFromElement): Don't re-get the value from the DOM if valueMatchesRenderer
        is true.
        (RenderLineEdit::slotTextChanged): Call setValueFromRenderer instead of manipulating the DOM
        directly.
        (RenderTextArea::detach): Call updateValue instead of calling value for its side effect.
        (RenderTextArea::handleFocusOut): Ditto.
        (RenderTextArea::updateFromElement): Call updateValue and then not re-get the value from the
        DOM if valueMatchesRenderer is true.
        (RenderTextArea::slotTextChanged): Call invalidateValue instead of directly setting m_dirtyvalue to true.

2005-02-22  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3937203> when a div adds a scrollbar (overflow:auto) we do not get regions

        Just set dashboard dirty bit when overflow scrolling changes.

        Don't do comparison of regions in before scroll regions are
        added, instead do it in WebKit after automatic scroll regions
        are added.

        Reviewed by Maciej.

        * khtml/css/cssparser.cpp:
        (CSSParser::parseDashboardRegions):  Cleaned up comments
        * khtml/css/cssstyleselector.cpp:
        (khtml::CSSStyleSelector::styleForElement): Cleaned up comments

        * khtml/khtmlview.cpp:
        (KHTMLView::updateDashboardRegions):
        * khtml/rendering/render_layer.cpp:
        (RenderLayer::updateScrollInfoAfterLayout):
        * kwq/WebDashboardRegion.m:
        (-[WebDashboardRegion isEqual:]):

2005-02-22  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3976872> Pasted plain text doesn't get the proper style if pasted into newlines

        Reviewed by mjs.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::doApply): don't clear the typing style when matching style
        (khtml::ReplaceSelectionCommand::completeHTMLReplacement): apply the typing style when matching style
        * khtml/editing/jsediting.cpp:
        * khtml/khtml_part.cpp:
        (KHTMLPart::pasteAndMatchStyle): new
        * khtml/khtml_part.h:
        * kwq/KWQKHTMLPart.h:
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::issuePasteAndMatchStyleCommand): new
        * kwq/WebCoreBridge.h:
        * layout-tests/editing/editing.js:

2005-02-22  Darin Adler  <darin@apple.com>

        Reviewed by Adele.

        - fixed <rdar://problem/4006596> REGRESSION (183-184): crash in DOM::DocumentImpl::setFocusNode(DOM::NodeImpl*)

        * khtml/xml/dom_docimpl.cpp:
        (widgetForNode): Added helper.
        (DocumentImpl::setFocusNode): Re-get the widget for the node after calling updateLayout.
        The updateLayout can destroy the old widget, so we can't keep a stale widget pointer around.

2005-02-22  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/4003463> Mail.app HTML uses inline styling markup not understood by Entourage and Eudora

        * khtml/editing/htmlediting.cpp:
        (khtml::isEmptyFontTag): Helper for removing <font> tags.
        (khtml::StyleChange::styleModeForParseMode): Helper to map a document parse mode to a use/don't use
        legacy-html-styles value.
        (khtml::StyleChange::checkForLegacyHTMLStyleChange): Add support for colors, font faces, and font sizes.
        (khtml::ApplyStyleCommand::isHTMLStyleNode):
        (khtml::ApplyStyleCommand::removeHTMLFontStyle):
        (khtml::ApplyStyleCommand::applyTextDecorationStyle): Now call styleModeForParseMode to determine
        whether to use legacy html styles or not.
        (khtml::ApplyStyleCommand::removeInlineStyle): Now properly removes <font> styles.
        (khtml::ApplyStyleCommand::addBlockStyleIfNeeded): Ditto.
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): Ditto.
        (khtml::createFontElement): Helper for applying <font> elements.
        * khtml/editing/htmlediting.h: All the following support the new bits of data we need to store.
        (khtml::StyleChange::applyFontColor)
        (khtml::StyleChange::applyFontFace)
        (khtml::StyleChange::applyFontSize)
        (khtml::StyleChange::fontColor)
        (khtml::StyleChange::fontFace)
        (khtml::StyleChange::fontSize)

        Test results updated now that we will write out <font> tags for quirks mode documents.        
        
        * layout-tests/editing/execCommand/modifyForeColorByCharacter-expected.txt
        * layout-tests/editing/style/block-style-004-expected.txt
        * layout-tests/editing/editing/pasteboard/paste-text-011-expected.txt
        * layout-tests/editing/style/block-style-005-expected.txt
        * layout-tests/editing/style/block-style-006-expected.txt
        * layout-tests/editing/style/smoosh-styles-001-expected.txt
        * layout-tests/editing/style/smoosh-styles-002-expected.txt

2005-02-22  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Kevin.

        <rdar://problem/4017066> crash in KJS::ValueImp::dispatchType() every time I load www.nytimes.com/pages/automobiles

        When creating option elements, use lowercase "option" instead of
        uppercase "OPTION" to create option elements, because only
        lowercase works for XHTML.
        
        * khtml/ecma/kjs_html.cpp:
        (KJS::HTMLSelectCollection::tryPut):
        (OptionConstructorImp::construct):

2005-02-22  Chris Blumenberg  <cblu@apple.com>

        WebCore fix for: <rdar://problem/3918056> Mail not line breaking my <pre> formatted emails on replies
        
        Mail must 4018993 to fully address the problem.

        Reviewed by kocienda.

        * khtml/editing/markup.cpp:
        (khtml::startMarkup): don't compute style for text in PRE tags
        (khtml::createMarkup): include PRE if it is an ancestor of the nodes in the range

2005-02-21  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/4008338> REGRESSION (125-178): opacity style not working, breaks fading images on okcupid.com

        There was a long standing bug in cssText().  Double value were always cast int!  So,
        opacity values values were incorrectly converted to text, i.e. 0.75 became 0.
        
        The problem was newly triggered because we parse the css for opacity *TWICE*.  This is a result of 
        incorrectly "invalidating" the style attribute (from fix for 3790449).  The second parse was from
        the cssText() of the style.

        Reviewed by Hyatt.

        * khtml/css/css_valueimpl.cpp:
        (DOM::CSSPrimitiveValueImpl::cssText):

2005-02-21  David Hyatt  <hyatt@apple.com>

        Fix for 4017204, apply the same fix to the base class nodeAtPoint that was already applied to RenderBlock.
        Skip elements with layers and inline flows.  Demoted <form>s can end up causing trouble otherwise.
        
        Reviewed by Richard Williamson

        * khtml/rendering/render_box.cpp:
        (RenderBox::nodeAtPoint):

2005-02-21  David Hyatt  <hyatt@apple.com>

        Fix for 4017033, CSS is being parsed twice.  Make sure to always validate the style attribute when it is
        initially parsed.  Add code to clean up decls when the style attribute is completely removed.  Add a new
        synchronizing boolean that avoids reparsing the style declaration when the attribute is simply
        being synced up to the declaration.

        Reviewed by John

        * khtml/css/css_valueimpl.cpp:
        (DOM::CSSMutableStyleDeclarationImpl::setChanged):
        * khtml/html/html_elementimpl.cpp:
        (HTMLElementImpl::invalidateStyleAttribute):
        (HTMLElementImpl::updateStyleAttributeIfNeeded):
        (HTMLElementImpl::HTMLElementImpl):
        (HTMLElementImpl::~HTMLElementImpl):
        (HTMLElementImpl::destroyInlineStyleDecl):
        (HTMLElementImpl::mapToEntry):
        (HTMLElementImpl::parseHTMLAttribute):
        * khtml/html/html_elementimpl.h:
        * khtml/xml/dom_elementimpl.cpp:
        (ElementImpl::ElementImpl):
        (ElementImpl::setAttribute):
        * khtml/xml/dom_elementimpl.h:
        (DOM::ElementImpl::updateStyleAttributeIfNeeded):

=== Safari-188 ===

2005-02-21  Darin Adler  <darin@apple.com>

        - fixed Panther deployment build

        * kwq/KWQTextEdit.mm: (QTextEdit::setScrollBarModes): Put more stuff inside #if.

2005-02-21  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3943090> REGRESSION (Mail): Spelling underline incompletely erased following certain steps

        * khtml/rendering/render_text.cpp:
        (InlineTextBox::paintMarker):
        Make sure underline is placed within the text bounds.
        
        * kwq/KWQPainter.h:
        * kwq/KWQPainter.mm:
        (QPainter::misspellingLineThickness):
        * kwq/WebCoreTextRenderer.h:
        Add misspellingLineThickness for use by InlineTextBox::paintMarker.

2005-02-21  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4012978> -[DOMRange markupString] crashes when range contains only a text node with a single space

        * khtml/editing/markup.cpp: (khtml::createMarkup): Added updateLayout calls, and added a missing
        nil check.

2005-02-21  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/4005435> Safari hung while pasting text into a <textarea> (Panther-only)

        * kwq/KWQTextEdit.mm: (QTextEdit::setScrollBarModes): Don't setAutohidesScrollers:YES on Panther.

2005-02-21  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/4015499> REGRESSION (186-187): pasted quoted text starting with a blank line increases quote level of pasted text when pasted

        * khtml/editing/markup.cpp:
        (khtml::markup): Changed over to ASSERT instead of assert.
        (khtml::createMarkup): The issue was that the code to add parents all the way back to the common ancestor
        block did not check for blocks whose markup had already been added, and could result in adding markup for
        nodes twice (hence the additional and erroneous quote level). Now there is a new check that will
        only add markup for those nodes before the start of the selection range. This fixes the bug.
        (khtml::createFragmentFromMarkup): Changed over to ASSERT instead of assert.
        (khtml::createFragmentFromText): Ditto.

2005-02-21  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/4015494> REGRESSION (186-187) <cr> removed if quoted word is copy/pasted on line immediately above quoted text

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::doApply): Relatively new check designed to remove a <br> element when
        that element was on a line by itself did not do an adequate check for this condition, causing the
        failure described in the bug. Fixed.

2005-02-19  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Maciej

        * khtml/editing/htmlediting.cpp:
        (khtml::matchNearestBlockquoteColorString): New function which abstracts away the CSS property we use
        for this pseudo-color.
        (khtml::ReplacementFragment::removeBlockquoteColorsIfNeeded): Set the new nearestBlockquoteColorString
        psuedo-color here, rather than removing colors as was done before.
        (khtml::ReplaceSelectionCommand::applyStyleToInsertedNodes): Check for the pseudo-color in the desired
        style, and determine the real color based on the nearest blockquote (or none) to the node.
        (khtml::nearestMailBlockquote): Name change from closestMailBlockquote to match new property better.

        All the rest of this change is the mechanical coding you need to do to add a new CSS property, in this
        case, -khtml-match-nearest-mail-blockquote-color.

        * khtml/css/css_computedstyle.cpp:
        (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue)
        * khtml/css/cssparser.cpp:
        (CSSParser::parseValue):
        * khtml/css/cssproperties.c:
        (hash_prop):
        (findProp):
        * khtml/css/cssproperties.h:
        * khtml/css/cssproperties.in:
        * khtml/css/cssstyleselector.cpp:
        (khtml::CSSStyleSelector::applyProperty)
        * khtml/css/cssvalues.c:
        (findValue)
        * khtml/css/cssvalues.h:
        * khtml/css/cssvalues.in:
        * khtml/rendering/render_style.h:
        (khtml::RenderStyle::matchNearestMailBlockquoteColor)
        (khtml::RenderStyle::setMatchNearestMailBlockquoteColor)
        (khtml::RenderStyle::initialMatchNearestMailBlockquoteColor)

2005-02-19  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Chris

        Fix for these bugs:
        
        <rdar://problem/4014228> REGRESSION (186-187) extra, uneditable lines inserted above and below a line of pasted quoted text
        <rdar://problem/4014393> REGRESSION (186-187) pasted quoted text gets extra <cr>s when pasted at top of document

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplacementFragment::ReplacementFragment): Part of a general refactoring of how
        the fragment is inserted into the document, rendered, and then tested for certain
        important pieces of information that are required for pasting.
        (khtml::ReplacementFragment::insertFragmentForTestRendering): New helper. Handles inserting
        the fragment nodes into the document.
        (khtml::ReplacementFragment::restoreTestRenderingNodesToFragment): Removes nodes from the
        document, and restores them to the fragment.
        (khtml::ReplacementFragment::computeStylesUsingTestRendering): Factored out code that
        did this before into its own function.
        (khtml::ReplacementFragment::removeUnrenderedNodesUsingTestRendering): Ditto.
        (khtml::ReplacementFragment::countRenderedBlocks): This is a real improvement, as it
        eliminates a major use of the isProbablyBlock() function. Now, the blocks that are
        counted are real, rendered blocks.
        (khtml::ReplacementFragment::removeStyleNodes): Made this function retain margin-zeroing
        CSS properties on paragraphs. This does two things: 1) It helps us to maintain good behavior
        in the short term while there are still versions of Mail out there that use <p> elements
        instead of <div> elements for new paragraphs; and 2) It will help to maintain the compatibility
        with other mail clients that use <p> elements for their paragraphs but render them themselves
        with no margins as the result of quirks.
        (khtml::ReplaceSelectionCommand::doApply): Do some work to fix up and improve the handling
        of blank lines, be they <p> elements or <br> elements, that can be removed after pasting. This, 
        coupled with the refactoring, fixes 4014393.
        * khtml/editing/htmlediting.h: Updated for new functions.
        * layout-tests/editing/pasteboard/paste-text-010-expected.txt: Updated results, actually improved with this change.
        * layout-tests/editing/pasteboard/paste-text-011-expected.txt: Ditto.

2005-02-19  Kevin Decker  <kdecker@apple.com>

        Reviewed by Chris.

        Fixed <rdar://problem/4010765> Flash player can be used to arbitrarily open popup windows without user permission
        
        Our window.open() policy is to refuse the <script>window.open(...)</script> case and allow the inline the <a href="javascript:window.open('foo')> case.  Clever advertisers at some point realized that by executing their Javascript through the Flash plugin, Safari would always treat their code as the inline case, and thus, they were able to work around our popup blocker.

        * kwq/WebCoreBridge.h:
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge stringByEvaluatingJavaScriptFromString:]): Changed this to invoke the new stringByEvaluatingJavaScriptFromString:forceUserGesture method below.
        (-[WebCoreBridge stringByEvaluatingJavaScriptFromString:forceUserGesture:]): Let the WebKit make the determination if this was a user originated gesture or not; we must no longer assume this is always the case.

2005-02-18  Chris Blumenberg  <cblu@apple.com>
        
        Fixed: <rdar://problem/3951196> REGRESSION (Mail): too many levels of reply quotes after certain steps

        Reviewed by kocienda.

        * khtml/editing/htmlediting.cpp:
        (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
        - If we find a new start node, update topBlockquote so we don't use too many block quotes for the contents following the new line.
        - Build up the list of ancestors after we've determined the actual topBlockquote.
        - Don't insert an extra new line if there is a new start node.

2005-02-19  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3978461> smart paste is broken

        Reviewed by kevin.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::doApply): properly check for leading and trailing whitespace. These checks were incorrectly reversed. Also check if we're pasting at the beginning or end of a line. We should not insert spaces in either case.

2005-02-18  Adele Amchan  <adele@apple.com>

        Reviewed by Darin.

        Fix for <rdar://problem/3975568> REGRESSION(125-180)Australian Open pages have drawing problem

        This patch fixes two problems caused by our added support for custom tags.  First, the layout problem at the sites mentioned in the bug
        was caused by custom tags within tables.  In checkChild, we needed to treat these tags as spans so they get placed correctly in the DOM tree.
        Also, we were indexing the tagPriority and endTag arrays with id values from the custom tags that were greater than the size of the array.  So now 
        we have functions to check for the custom tags, and again, treat them as spans.  To avoid confusion, we changed the names of the arrays to
        endTagArray and tagPriorityArray.

        * khtml/html/dtd.h: changed all uses of the endTag array to endTagArray for our new wrapper functions
        (DOM::tagPriority): added function to check array bounds and to treat custom tags as spans
        (DOM::endTagRequirement): added function to check array bounds and to treat custom tags as spans
        * khtml/html/dtd.cpp: changed name of endTag and tagPriority arrays to endTagArray and tagPriorityArray
        (DOM::checkChild): treat custom tags as spans during this check
        * khtml/html/html_elementimpl.cpp: changed all uses of the endTag array to the endTagRequirement function
        (HTMLElementImpl::createContextualFragment):
        (HTMLElementImpl::setInnerText):
        (HTMLElementImpl::setOuterText):
        (HTMLElementImpl::toString):
        * khtml/html/htmlparser.cpp: changed all uses of the endTag array to the endTagRequirement function and all uses of the tagPriority array to the tagPriority function.
        (KHTMLParser::parseToken):
        (KHTMLParser::insertNode):
        * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): changed all uses of the endTag array to the endTagRequirement function
       * khtml/editing/markup.cpp: changed all uses of the endTag array to the endTagRequirement function
        (khtml::endMarkup): 
        (khtml::markup):

2005-02-18  Jens Alfke  <jens@apple.com>

        Reviewed by Hyatt.

        Fixed build: Whoops, setNeedsLayout's parameter is NOT optional.

        * khtml/html/html_objectimpl.cpp:
        (HTMLAppletElementImpl::setAllParamsAvailable):

2005-02-18  Jens Alfke  <jens@apple.com>

        Reviewed by cblu, hyatt.

        Fixes <rdar://problem/3603191> "REGRESSION: Applets not receiving all of the Applet Parameters in Java 1.4.1/1.4.2"
        Defer instantiation of Java applet plugin until after all of the <applet> tag's nested <param> tags have been parsed, otherwise the list of parameters passed to the applet is incomplete. The regression was introduced (says Dave) when the parser's close-tag notifications were removed in the name of performance.

        * khtml/html/html_objectimpl.cpp:
        (HTMLAppletElementImpl::HTMLAppletElementImpl):
        (HTMLAppletElementImpl::getAppletInstance):
        (HTMLAppletElementImpl::setAllParamsAvailable):
        (HTMLAppletElementImpl::allParamsAvailable):
        * khtml/html/html_objectimpl.h:
        * khtml/html/htmlparser.cpp:
        (KHTMLParser::processCloseTag):
        * khtml/rendering/render_applet.cpp:
        (RenderApplet::createWidgetIfNecessary):

2005-02-18  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/4006161> Tiger8A380: Widgets leak dashboard regions
        
        We were leaking the DashboardRegionImpls.

        Reviewed by Kevin.

        * khtml/css/css_valueimpl.cpp:
        (DOM::CSSPrimitiveValueImpl::cleanup):

2005-02-18  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3945271> REGRESSION (Mail): pasted plain text should pick up typing style instead of being unstyled

        Reviewed by kocienda.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): don't set class on element returned by createStyleSpanElement since that's already done
        (khtml::ReplacementFragment::ReplacementFragment): take matchStyle param, don't call computeStylesAndRemoveUnrendered() if !matchStyle
        (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): take matchStyle param
        (khtml::ReplaceSelectionCommand::doApply): if m_matchStyle, use selection.start() as the insertion position and don't call applyStyleToInsertedNodes()
        * khtml/editing/htmlediting.h:
        * kwq/WebCoreBridge.h:
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge replaceSelectionWithFragment:selectReplacement:smartReplace:matchStyle:]): take matchStyle param and pass it
        (-[WebCoreBridge replaceSelectionWithNode:selectReplacement:smartReplace:]): pass NO for matchStyle
        (-[WebCoreBridge replaceSelectionWithMarkupString:baseURLString:selectReplacement:smartReplace:]): pass NO for matchStyle
        (-[WebCoreBridge replaceSelectionWithText:selectReplacement:smartReplace:]): pass YES for matchStyle

2005-02-18  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Chris

        Fix for these bugs:
        
        <rdar://problem/4013025> Copy/Paste of quoted word results in removal of any following <cr>
        <rdar://problem/4013100> Copy/Paste quoted text and then decrease quote level does not change text color

        For the most part, these bugs were caused by errors and lack of foresight on my part when
        I added the better paste code. Chalk these fixes up to the result of bake time.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplacementFragment::ReplacementFragment): Need to move count of number of blocks in
        fragment after the call to remove unrendered nodes. Meant to do this before, but forgot to.
        (khtml::ReplacementFragment::removeStyleNodes): Need to remove inline styles from elements!
        Terrible omission now fixed.
        (khtml::ReplacementFragment::removeBlockquoteColorsIfNeeded): Remove blockquote colors for now.
        Code has a more extensive comment in it now to explain the difficulty, and the need for more
        study and changes.
        (khtml::ReplaceSelectionCommand::doApply): Need to call applyStyleToInsertedNodes() in the 
        m_fragment.hasInterchangeNewline() case. This was just missed before.
        * layout-tests/editing/pasteboard/paste-text-011-expected.txt: Updated results, subtly different, but OK.
        * layout-tests/editing/pasteboard/paste-text-017-expected.txt: Updated for <p> to <div> change in test content.
        * layout-tests/editing/pasteboard/paste-text-017.html: Needed to change <p> to <div> to 
        make this test go with the new design of using <div> tags for default paragraphs.

2005-02-18  David Hyatt  <hyatt@apple.com>

        Fix for 3974263 (and possibly others).  Don't let fixed tables use maxint as their maxwidth when some
        cells have percentage values.
        
        Reviewed by kocienda

        * khtml/misc/arena.cpp:
        (ArenaAllocate):
        * khtml/rendering/table_layout.cpp:
        (FixedTableLayout::calcWidthArray):
        (FixedTableLayout::calcMinMaxWidth):
        (AutoTableLayout::layout):

2005-02-17  Darin Adler  <darin@apple.com>

        Reviewed by Kevin.

        - fixed <rdar://problem/3998627> WebKit crashes when deleting text in a modified div

        * khtml/khtml_part.cpp: (KHTMLPart::selectionComputedStyle): Add a ref and deref to keep
        the element alive until it's added to the DOM tree.
        * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::styleForSelectionStart): Ditto.

2005-02-17  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3999467> when Osaka-Mono is specified as fixed width font, Osaka used instead

        Fixed w/o introducing a performance regression.  Add early
        check for Osaka-Mono to avoid expensive call into WebKit.

        Reviewed by Vicki.

        * kwq/KWQFont.mm:
        (QFont::isFixedPitch):

2005-02-17  Darin Adler  <darin@apple.com>

        Reviewed by Kevin.

        - fixed <rdar://problem/4011210> REGRESSION (180-181): Maxlength property for INPUT object not working

        * kwq/KWQTextField.mm:
        (-[KWQTextField textView:shouldChangeTextInRange:replacementString:]): If controller returns
        YES, go on to call super, since super does the text formatter handling, which we need.
        (-[KWQSecureTextField textView:shouldChangeTextInRange:replacementString:]): Ditto.
        (-[KWQSearchField textView:shouldChangeTextInRange:replacementString:]): Ditto.

=== Safari-187 ===

2005-02-17  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/4012058> Copy from quoted text and paste results in blue text

        The relatively-new paste code tries hard to retain style of the content from the
        source location. However, in the case of quoted material in mail messages, we do 
        not want to carry the quoting color along. This fixes the problem by factoring
        out such colors.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplacementFragment::computeStylesAndRemoveUnrendered): Now calls removeBlockquoteColorsIfNeeded()
        to fix bug.
        (khtml::ReplacementFragment::removeBlockquoteColorsIfNeeded): New function. Factors out colors that are
        the result of text being quoted. 
        (khtml::isNodeRendered): Function moved in file. No other change.
        (khtml::isProbablyBlock): Function moved in file. No other change.
        (khtml::closestMailBlockquote): New function. Helps fix bug.
        (khtml::isMailBlockquote): Function moved to be free-floating instead of being a 
        member of CompositeEditCommand. No other change.
        * khtml/editing/htmlediting.h: Moved around some functions. Added removeBlockquoteColorsIfNeeded().
        * khtml/editing/markup.cpp: Remove redundant static implementation of isMailBlockquote.

2005-02-17  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/4008163> dynamic support for -apple-dashboard-region is flakey

        Style operator== wasn't including regions.

        Reviewed by Darin.

        * khtml/rendering/render_style.cpp:
        (StyleCSS3NonInheritedData::operator==):

2005-02-17  Adele Amchan  <adele@apple.com>

        Reviewed by Chris.

        fix for <rdar://problem/4010028> 8A383: Safari v185 crash loading united.com multi city fare finder page.

        * khtml/dom/dom_string.cpp: (DOM::DOMString::operator += ): prevent nil dereference when DOMString being added is nil

2005-02-17  Jens Alfke  <jens@apple.com>

        Reviewed by Hyatt.

        Fix for rdar://3963151 "Mail only pasted ~950 of 1407 text lines into my message!"
        Force tokenizer to run synchronously while parsing document fragments, so it doesn't stop halfway through and cause truncated content.

        * khtml/html/html_elementimpl.cpp:
        (HTMLElementImpl::createContextualFragment):
        * khtml/html/htmltokenizer.cpp:
        (khtml::HTMLTokenizer::reset):
        (khtml::HTMLTokenizer::begin):
        (khtml::HTMLTokenizer::setForceSynchronous):
        (khtml::HTMLTokenizer::continueProcessing):
        * khtml/html/htmltokenizer.h:

2005-02-17  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3982183> Mail's HTML paragraphs appear with unintended margins in Entourage and Eudora

        * khtml/editing/htmlediting.cpp:
        (khtml::createDefaultParagraphElement): Change element we make from <p> to <div>.
        This fixes the problem, as these other mailers show <div> elements without margins.
        
        Some test files needed to change to preserve the ability to "eyeball" the results.
        Many others changed just because their <p> elements changed to <div> elements.
        
        * layout-tests/editing/deleting/delete-3928305-fix-expected.txt
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002-expected.txt
        * layout-tests/editing/inserting/insert-div-013-expected.txt
        * layout-tests/editing/inserting/insert-div-013.html
        * layout-tests/editing/inserting/insert-div-014-expected.txt
        * layout-tests/editing/inserting/insert-div-014.html
        * layout-tests/editing/inserting/insert-div-018-expected.txt
        * layout-tests/editing/inserting/insert-div-018.html
        * layout-tests/editing/inserting/insert-div-019-expected.txt
        * layout-tests/editing/inserting/insert-div-019.html
        * layout-tests/editing/inserting/insert-div-020-expected.txt
        * layout-tests/editing/inserting/insert-div-020.html
        * layout-tests/editing/inserting/insert-div-021-expected.txt
        * layout-tests/editing/inserting/insert-div-021.html
        * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt
        * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt
        * layout-tests/editing/style/remove-underline-after-paragraph-expected.txt
        * layout-tests/editing/style/remove-underline-after-paragraph-in-bold-expected.txt

2005-02-17  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3996737> REGRESSION (Mail): Copy/paste in Mail inserts returns

        The root cause of the problem is that a couple of pieces of code in AppKit and Mail 
        insert newlines into markup as it is generated, and the paste code in WebCore was 
        not smart about nodes that do not render (as is the case with these added newlines). 
        So, the solution is to remove these unrendered nodes in the paste code in a pre-pass. 
        Fortunately, my recent addition of such a pass to the paste code to handle styles 
        gave me a convenient place to put this new logic.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplacementFragment::ReplacementFragment): Use new name for function below.
        (khtml::ReplacementFragment::computeStylesAndRemoveUnrendered): Name change from computeStylesForNodes().
        Now does the additional work of removing unrendered nodes, as mentioned above.
        (khtml::isNodeRendered): New helper function.
        * khtml/editing/htmlediting.h: Updated as necessary.

2005-02-17  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3998892> REGRESSION (Mail): bolding a selection from end of line changes unselected text on starting line.
        
        The styling code did not move to the next node when the starting position was 
        at the last offset of a node. Instead, it styled it. Clearly wrong. Solution 
        is to borrow a check from the delete algorithm. Though I call caretMaxOffset(),
        which I consider deprecated, it is still the simplest and most strightforward
        way to ask the right question for this kind of problem.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::applyInlineStyle): Fixed, as described above.
        * layout-tests/editing/style/style-3998892-fix-expected.txt: Added.
        * layout-tests/editing/style/style-3998892-fix.html: Added.

2005-02-16  Chris Blumenberg  <cblu@apple.com>
        
        Fixed: <rdar://problem/3954842> Forward/reply to an HTML email can result in nothing (in cases with <link> tags for CSS)

        Reviewed by john.

        * khtml/xml/dom_nodeimpl.cpp:
        (NodeBaseImpl::addChild): don't call insertedIntoDocument on the added child if "this" itself is not in the document.

2005-02-16  John Sullivan  <sullivan@apple.com>

        Written by Darin, reviewed by Maciej and me
        
        - change required to make previous checkin work with English language RSS pages

        * kwq/KWQKURL.h:
        new private method isHierarchical
        * kwq/KWQKURL.mm:
        (KURL::KURL):
        add hierarchical base URL check when determining whether the URL is absolute
        (KURL::isHierarchical):
        new method, returns true if this is a valid URL with a slash just past the scheme's trailing colon

2005-02-16  John Sullivan  <sullivan@apple.com>

        Written by Darin, reviewed by me.
        
        - WebCore part of fix for <rdar://problem/4007384> 
        FILTER: Bookmark of RSS with Japanese search word & multiple RSS pages loses filter

        * kwq/DOM.mm:
        (-[DOMElement _getURLAttribute:]):
        new SPI method, uses parseURL and completeURL to get valid URL from attribute value

        * kwq/DOMPrivate.h:
        declare new SPI method

2005-02-15  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Darin.

        <rdar://problem/3942428> reproducible crash loading cbs.sportsline.com
        
        This change reverts the fix for <rdar://problem/3805311>, and
        re-fixes it in a different (better) way. Instead of preventing
        programmatic open from setting the parsing flag, instead make sure
        that programmatic close resets it.

        * khtml/khtml_part.cpp:
        (KHTMLPart::openURL):
        (KHTMLPart::didExplicitOpen):
        (KHTMLPart::closeURL):
        (KHTMLPart::begin):
        (KHTMLPart::end):
        (KHTMLPart::endIfNotLoading):
        (KHTMLPart::slotFinishedParsing):
        (KHTMLPart::checkEmitLoadEvent):
        * khtml/khtml_part.h:
        * khtml/khtmlpart_p.h:
        (KHTMLPartPrivate::KHTMLPartPrivate):
        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::open):
        (DocumentImpl::implicitOpen):
        (DocumentImpl::close):
        (DocumentImpl::implicitClose):
        * khtml/xml/dom_docimpl.h:

2005-02-15  David Harrison  <harrison@apple.com>

        Reviewed by Maciej.

        <rdar://problem/3933665> smart delete seems to delete too much after expanding selection with arrow keys

        Fixed by updating the selection granularity.  Only byWord granularity enables smart delete, but in this
        case the granularity changed from byWord to byCharacter.

        * khtml/khtml_part.cpp:
        (KHTMLPart::setSelectionGranularity):
        * khtml/khtml_part.h:
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge rangeByExpandingSelectionWithGranularity:]):
        (-[WebCoreBridge rangeByAlteringCurrentSelection:direction:granularity:]):
        (-[WebCoreBridge alterCurrentSelection:direction:granularity:]):

2005-02-15  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        (continued) <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line

        This was supposed have been committed yesterday with the other part of the fix, but was not.
        
        * khtml/editing/visible_units.cpp:
        (khtml::endOfWord):

2005-02-15  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John
        
        Fix for this bug:
        
        <rdar://problem/3951178> REGRESSION (Mail): blank line lost after pasting as quotation
        
        Problem was that the blank line after the selection was getting deleted incorrectly since the
        paste code thought this was an unneeded placeholder rather than a placeholder outside of the
        selection.

        * khtml/editing/htmlediting.cpp:
        (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Now calls findBlockPlaceholder.
        (khtml::CompositeEditCommand::findBlockPlaceholder): Moved finding code formerly in 
        removeBlockPlaceholderIfNeeded to this new helper.
        (khtml::ReplaceSelectionCommand::doApply): Do not delete placeholder up front. Call 
        findBlockPlaceholder, and delete it later if needed in the already-existing cleanup step.
        * khtml/editing/htmlediting.h: Add new function.

        New layout test.

        * layout-tests/editing/pasteboard/paste-text-017-expected.txt: Added.
        * layout-tests/editing/pasteboard/paste-text-017.html: Added.

        Result changed for the better.

        * layout-tests/editing/pasteboard/paste-text-011-expected.txt

2005-02-14  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line

        Reviewed by Darin.

        <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line

        * khtml/editing/visible_text.cpp:
        (khtml::SimplifiedBackwardsTextIterator::advance):
        Add BR in for <rdar://problem/3917929> fix only if leaving a visible text node.
        
        * khtml/editing/visible_units.cpp:
        (khtml::endOfWord):
        Do not move left over a paragraph boundary.

2005-02-14  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3686434> Safari uses too much RAM on file upload, leading to malloc errors and crashes (HP printers)

        * khtml/html/html_formimpl.cpp:
        (DOM::FormDataList::begin): Updated to use a list of FormDataListItem instead of QCString.
        (DOM::FormDataList::end): Ditto.
        (DOM::HTMLFormElementImpl::formData): Same here, and also use appendFile rather than appendData
        when we encounter a path name rather than data.
        (DOM::HTMLInputElementImpl::appendFormData): Use appendFile here rather than reading the file
        in; the reading now happens inside WebKit.
        (DOM::FormDataList::appendString): Updated for FormDataListItem.
        (DOM::FormDataList::appendFile): Added.

        * ForwardingHeaders/kfileitem.h: Emptied out the file; no longer includes KWQKFileItem.h.
        * ForwardingHeaders/netaccess.h: Emptied out the file; no longer includes KWQKIONetAccess.h.
        * kwq/KWQKFileItem.h: Removed.
        * kwq/KWQKFileItem.mm: Removed.
        * kwq/KWQKIONetAccess.h: Removed.
        * kwq/KWQKIONetAccess.mm: Removed.
        * WebCore.pbproj/project.pbxproj: Removed the 4 files above.

        - small unrelated fix (not reviewed by John)

        * khtml/ecma/kjs_window.cpp: (Window::isSafeScript): Put some logging code outside if statements so it works all the time.

2005-02-14  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line

        Fixed nextLinePosition to calculate affinity rather than take it as a parameter.  Propagated the parameter change out to related methods.

        * khtml/editing/htmlediting.cpp:
        (khtml::DeleteSelectionCommand::initializePositionData):
        (khtml::InsertLineBreakCommand::doApply):
        (khtml::InsertParagraphSeparatorCommand::doApply):
        (khtml::InsertTextCommand::input):
        (khtml::ReplaceSelectionCommand::doApply):
        (khtml::ReplaceSelectionCommand::applyStyleToInsertedNodes):
        * khtml/editing/selection.cpp:
        (khtml::Selection::modifyExtendingRightForward):
        (khtml::Selection::modifyMovingRightForward):
        (khtml::Selection::modifyExtendingLeftBackward):
        (khtml::Selection::modifyMovingLeftBackward):
        (khtml::Selection::modify):
        (khtml::Selection::validate):
        * khtml/editing/visible_position.cpp:
        (khtml::visiblePositionsOnDifferentLines):
        * khtml/editing/visible_units.cpp:
        (khtml::rootBoxForLine):
        (khtml::startOfLine):
        (khtml::endOfLine):
        (khtml::inSameLine):
        (khtml::isStartOfLine):
        (khtml::isEndOfLine):
        (khtml::previousLinePosition):
        (khtml::nextLinePosition):
        (khtml::previousSentencePosition):
        (khtml::nextSentencePosition):
        (khtml::previousParagraphPosition):
        (khtml::nextParagraphPosition):
        * khtml/editing/visible_units.h:
        * khtml/khtml_events.cpp:
        (khtml::MouseEvent::offset):
        * khtml/khtml_part.cpp:
        (KHTMLPart::isPointInsideSelection):
        (KHTMLPart::selectClosestWordFromMouseEvent):
        (KHTMLPart::handleMousePressEventTripleClick):
        (KHTMLPart::handleMousePressEventSingleClick):
        (KHTMLPart::handleMouseMoveEventSelection):
        (KHTMLPart::khtmlMouseReleaseEvent):
        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::positionForCoordinates):
        * khtml/rendering/render_block.h:
        * khtml/rendering/render_br.cpp:
        (RenderBR::positionForCoordinates):
        * khtml/rendering/render_br.h:
        * khtml/rendering/render_container.cpp:
        (RenderContainer::positionForCoordinates):
        * khtml/rendering/render_container.h:
        * khtml/rendering/render_inline.cpp:
        (RenderInline::positionForCoordinates):
        * khtml/rendering/render_inline.h:
        * khtml/rendering/render_object.cpp:
        (RenderObject::caretRect):
        (RenderObject::positionForCoordinates):
        * khtml/rendering/render_object.h:
        * khtml/rendering/render_replaced.cpp:
        (RenderReplaced::positionForCoordinates):
        * khtml/rendering/render_replaced.h:
        * khtml/rendering/render_text.cpp:
        (RenderText::positionForCoordinates):
        * khtml/rendering/render_text.h:
        * khtml/xml/dom_position.cpp:
        (DOM::Position::previousCharacterPosition):
        (DOM::Position::nextCharacterPosition):
        (DOM::Position::leadingWhitespacePosition):
        (DOM::Position::trailingWhitespacePosition):
        * khtml/xml/dom_position.h:
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject value]):
        (-[KWQAccObject accessibilityAttributeValue:]):
        (-[KWQAccObject doAXLineForTextMarker:]):
        (-[KWQAccObject doAXTextMarkerRangeForLine:]):
        (-[KWQAccObject doAXTextMarkerForPosition:]):
        (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
        (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
        (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
        * kwq/KWQKHTMLPart.mm:
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge _visiblePositionForPoint:]):

05-02-07  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Ken and John.

        Re-fixed a specific case of the following:

        <rdar://problem/3790449> REGRESSION (Mail): underline behavior is flaky because of how CSS handles text-decoration

        Oddly, Cmd-B, Cmd-U, type some text, return, Cmd-U, Cmd-B, type
        some text, worked fine. But hitting the second Cmd-B before the
        second Cmd-U still failed to remove underlining. The reason for
        this is that our code to compute the style of the current position
        did not work when you had a typing style and were positioned right
        at a <br> element. For various reasons, this did not show up for
        bold and italic, since those are handled through the font manager.

        The following change fixes this - for elements that can't have
        children, we add the dummy span after the element of interest,
        rather than as a child of it.
        
        * khtml/khtml_part.cpp:
        (KHTMLPart::selectionComputedStyle): 

2005-02-11  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3978980> Double Clicking on a line in Mail selected the entire body

        * khtml/editing/visible_units.cpp:
        (khtml::startOfWord):
        (khtml::endOfWord):
        Pay attention to being at the end of a paragraph.
        
        (khtml::previousLinePosition):
        (khtml::nextLinePosition):
        (khtml::endOfParagraph):
        Use DOWNSTREAM per recent affinity changes.

2005-02-11  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3985118> DOM objects not being marshaled on JS->native calls

        Re-factored how 'native' wrappers for JS objects are created.  The interpreter now
        creates these wrappers.  The WebCore subclass of the interpreter now overrides
        createLanguageInstanceForValue() and creates a DOM ObjC wrapper for DOM objects.

        * WebCore.pbproj/project.pbxproj:
        * khtml/ecma/kjs_binding.cpp:
        (ScriptInterpreter::createLanguageInstanceForValue):
        * khtml/ecma/kjs_binding.h:
        * kwq/DOMUtility.mm: Added.
        (KJS::ScriptInterpreter::createObjcInstanceForValue):
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::getAppletInstanceForView):
        (getInstanceForView):
        (KWQKHTMLPart::getEmbedInstanceForView):
        (KWQKHTMLPart::getObjectInstanceForView):

2005-02-11  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3937352> Quote level not maintained when copied and pasted within a Mail message

        Reviewed by harrison.

        * khtml/editing/markup.cpp:
        (khtml::isMailBlockquote): new 
        (khtml::createMarkup): retain the Mail quote level by including all ancestor mail block quotes
        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::positionForCoordinates): default to downstream for the affinity since the affinity is only upstream when the cursor is clicked to the right of a wrapped line
        * khtml/rendering/render_text.cpp:
        (RenderText::positionForCoordinates): ditto

2005-02-11  Adele Amchan  <adele@apple.com>

        Reviewed by Darin.

        fix for <rdar://problem/4004004> no need to add body element for xml documents

        In the fix for <rdar://problem/3758785> we decided to add a body element when closing a document 
        to ensure that the onload handler would fire.  This is unnecessary for xml documents, 
        so now we also check to see if we're dealing with an html document before adding the body element.

        * khtml/xml/dom_docimpl.cpp: (DocumentImpl::close): 

2005-02-11  Adele Amchan  <adele@apple.com>

        Reviewed by Darin.

        * layout-tests/apple-only/base/www.excite.com/index-expected.txt: updates tests because of custom tag change.
        * layout-tests/apple-only/base/www.msnbc.com/index-expected.txt:
        * layout-tests/apple-only/base/www.time.com/index-expected.txt:
        * layout-tests/apple-only/base/www.travelocity.com/index-expected.txt:
        * layout-tests/apple-only/base/www.zdnet.com/index-expected.txt:
        * layout-tests/fast/overflow/003-expected.txt:

        * layout-tests/fast/js/window-object-cross-frame-calls-expected.txt: updated test because of added Body (see rdar://problem/3758785)

=== Safari-186 ===

2005-02-11  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3915449> paths are relative to the old src URL after document.open, which is supposed to clear the document, including the URL

        * khtml/xml/dom_docimpl.h: Added a new openInternal function for use by KHTMLPart.
        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::open): Changed to do everything we did before, but also clear the URL and set the
        base URL based on the enclosing document. This is the basic JavaScript/DOM operation of opening a
        document, which is supposed to clear the document, including the URL. In the long run we might want
        to do even more document "resetting and clearing" in here, but this URL clearing is what's needed
        now to fix the most important problem.
        (DocumentImpl::openInternal): Moved the old open code in here, except for the "parsing" check, which
        is unnecessary and inappropriate in the one place we call this.

        * khtml/khtml_part.cpp: (KHTMLPart::begin): Call openInternal instead of open.

2005-02-10  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3992092> 8A374: Mail crash while pasting - RemoveNodeCommand

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::doApply): Code tried to remove a node that was no longer in 
        the document. We try to do a good job of detecting all these cases, and generally do. This
        one was missed. Fixed.

2005-02-10  Darin Adler  <darin@apple.com>

        Reviewed by Harrison.

        - added support needed to fix <rdar://problem/3991225> Format->Style->Underline menu item does not get checked when selected text is underlined

        * kwq/WebCoreBridge.h: Added selectionHasStyle: method.
        * kwq/WebCoreBridge.mm: (-[WebCoreBridge selectionHasStyle:]): Added. Calls selectionHasStyle on the part.

2005-02-10  Darin Adler  <darin@apple.com>

        Reviewed by Harrison.

        - fixed <rdar://problem/3990484> cursor changes to I-beam when moving over text in widgets, even when selection off via CSS

        * khtml/khtmlview.cpp: (KHTMLView::viewportMouseMoveEvent): Changed code to check canSelect when deciding
        whether to show an I-beam cursor.

        * khtml/rendering/render_object.h: Added canSelect.
        * khtml/rendering/render_object.cpp:
        (selectStartNode): Added. Helper function with the guts of shouldSelect.
        (RenderObject::canSelect): Added. Like shouldSelect, but does not call the "start selecting" event handler.
        (RenderObject::shouldSelect): Refactored to use selectStartNode.

        * khtml/css/cssvalues.c: Regnerated with newer gperf.

2005-02-10  David Hyatt  <hyatt@apple.com>

        Fix for 3867759, .mac regression where scrollers don't show up.  Make sure you can set the .width and
        .height properties on the Image object.
        
        Reviewed by John Sullivan

        * khtml/ecma/kjs_html.cpp:
        (ImageConstructorImp::construct):
        (Image::getValueProperty):
        (Image::putValue):
        (Image::Image):
        * khtml/ecma/kjs_html.h:

2005-02-10  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Chris

        Fix for this bug:
        
        <rdar://problem/3965158> Drag-n-drop within a rich text message sometimes changes the color of the dragged text
        
        This change fixes the bug....and much more. Now, for the first time, the paste code can do "smart merging"
        or "smooshing" of styles during its operation. Since this new code is actively, rather than passively
        working with styles, it fixes the bug, and lays the groundwork for similar work we need to do to
        preserve quote levels in Mail.

        * khtml/css/css_valueimpl.cpp:
        (DOM::CSSMutableStyleDeclarationImpl::clear): New method.
        (DOM::CSSMutableStyleDeclarationImpl::removeBlockProperties): Ditto.
        (DOM::CSSMutableStyleDeclarationImpl::removePropertiesInSet): Ditto.
        (DOM::operator==): Add operator for CSSProperty.
        * khtml/css/css_valueimpl.h: Declare new functions.
        * khtml/editing/htmlediting.cpp:
        (khtml::isEmptyStyleSpan): Improved the test in this function, rolling together the old implementation
        with some code that did this work inline elsewhere. Sum of the parts is better than either test was by itself.
        (khtml::isStyleSpan): Check for ID_SPAN.
        (khtml::ApplyStyleCommand::removeCSSStyle): Call isEmptyStyleSpan. This was the place with an inline implementation before.
        (khtml::ReplacementFragment::ReplacementFragment): Now takes a DocumentImpl argument. No longer does a "default style"
        check, but rather calls functions which do a similar check to that, and much more.
        (khtml::ReplacementFragment::~ReplacementFragment): Deref document, and computed styles.
        (khtml::ReplacementFragment::styleForNode): New helper. Looks up and returns computed style for a node.
        (khtml::ReplacementFragment::removeNodePreservingChildren): New helper.
        (khtml::ReplacementFragment::computeStylesForNodes): New function which computes the "desired" style for
        every node in the fragment. This information is used later after paste is done as a reference for testing
        what styles need to be added, and which can be removed as redundant, from all the nodes inserted by the
        replacement code.
        (khtml::ReplacementFragment::removeStyleNodes): Clears out all style nodes from the fragment. They are
        no longer needed after the call to computeStylesForNodes(),
        (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): Add a document to the call to initialize the
        command's ReplacementFragment.
        (khtml::ReplaceSelectionCommand::doApply): Call applyStyleToInsertedNodes() after inserting nodes to make
        styles come out right.
        (khtml::ReplaceSelectionCommand::applyStyleToInsertedNodes): This is the "style smooshing" function. It 
        computes the styles that need to be added to each node inserted, comparing the style it gets from just
        being inserted into its correct destination with the computed "desired style" done in the 
        ReplacementFragment constructor. It then adds in all the necessary styles, and will also remove redundant styles.
        * khtml/editing/htmlediting.h: Update declarations and member variables as needed.
        * khtml/editing/markup.cpp:
        (khtml::startMarkup): Add additional style annotations to the markup we generate, so that paste code can preserve it.
        (khtml::markup): Ditto.
        (khtml::createMarkup): Ditto.
        
        These test results are subtly better with this change. They no longer have an unneeded empty span.
        Visually the same as before.
        
        * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt
        * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt
        * layout-tests/editing/style/remove-underline-expected.txt
        * layout-tests/editing/style/remove-underline-from-stylesheet-expected.txt

        New tests:

        * layout-tests/editing/style/smoosh-styles-001-expected.txt
        * layout-tests/editing/style/smoosh-styles-002-expected.txt
        * layout-tests/editing/style/smoosh-styles-001.html
        * layout-tests/editing/style/smoosh-styles-002.html

2005-02-10  Darin Adler  <darin@apple.com>

        Reviewed by Adele.

        - fixed <rdar://problem/3974988> WebHTMLView drops scroll wheel events when deltas are 0

        * kwq/WebCoreScrollView.m: (-[WebCoreScrollView autoforwardsScrollWheelEvents]): Added, for Tiger only.
        For Tiger, removed the override of scrollWheel:, which doesn't work quite right because of how the
        delta attributes return all 0 for newfangled scrolling events from the new trackpads.

2005-02-10  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3991532> REGRESSION (Mail): Triple-click on trailing blank line should select previous line but doesn't

        * khtml/css/cssvalues.c:
        (hash_val):
        (findValue):
        * khtml/editing/selection.cpp:
        (khtml::Selection::validate):
        Do moral equiavalent of LeftWordIfOnBoundary for PARAGRAPH.

2005-02-10  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3991848> Double-click on first character selects wrong item

        * khtml/editing/selection.cpp:
        (khtml::Selection::validate):
        Honor the fact that clicking on a character positions the cursor on the left side of the character.

2005-02-10  David Hyatt  <hyatt@apple.com>

        Fix for 3980778, repro crash in RootInlineBox::paint.  Make sure that when merging blocks we dirty them for layout
        up front. That way as we do the merge, we dont attempt to examine the lines in dirtyLinesForChangedChild.
        
        Reviewed by mjs

        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::removeChild):

2005-02-10  David Hyatt  <hyatt@apple.com>

        Fix for 3987010, fix a bug that caused the self-collapsing-block-cleared-float bit to never get set correctly.
        
        Reviewed by mjs

        * khtml/rendering/render_block.h:
        (khtml::RenderBlock::MarginInfo::setSelfCollapsingBlockClearedFloat):

2005-02-10  David Hyatt  <hyatt@apple.com>

        Fix  for 3992440, text not wrapping correctly on half moon bay web site.  Refine an incorrect optimization I
        added to be less restrictive.
        
        Reviewed by mjs

        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::layoutBlockChildren):

2005-02-08  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Darin.

        <rdar://problem/3977973> pages on ebay leak referenced JavaScript objects -- over time browsing becomes super-slow

        I fixed this by removing all event listeners for a document, it's
        children, and any disconnected nodes that used to be in the
        document at document detach time. Mozilla temporarily disables
        event listeners on such nodes, but re-enables them if you
        re-parant a node into a new document. However, in WebCore, you
        can't re-parent a node into another document, so there is no
        observable change in behavior.

        We have to do this to break the possible reference cycles between
        event listeners and the dom nodes they are attached to (e.g. via
        scope chain, as in this case).

        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::detach):
        (DocumentImpl::removeAllEventListenersFromAllNodesx):
        (DocumentImpl::registerDisconnectedNodeWithEventListeners):
        (DocumentImpl::unregisterDisconnectedNodeWithEventListeners):
        (DocumentImpl::removeAllDisconnectedNodeEventListeners):
        * khtml/xml/dom_docimpl.h:
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::~NodeImpl):
        (NodeImpl::addEventListener):
        (NodeImpl::removeEventListener):
        (NodeImpl::removeAllEventListeners):
        (NodeImpl::removeHTMLEventListener):
        (NodeImpl::insertedIntoDocument):
        (NodeImpl::removedFromDocument):
        * khtml/xml/dom_nodeimpl.h:

2005-02-09  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3999213> Sometimes 2 Windows Media Player plugin instances are loaded

        Reviewed by hyatt.

        * khtml/html/html_objectimpl.cpp:
        (HTMLObjectElementImpl::attach): Set needWidgetUpdate to false before calling updateWidget because updateWidget may cause this method or another method (which also calls updateWidget) to be called.
        (HTMLObjectElementImpl::recalcStyle): ditto

2005-02-09  David Harrison  <harrison@apple.com>

        Reviewed by Darin.
        
        <rdar://problem/3937447> Mail-689: Arrow key navigation in new message body stops working when a line ends with a bold word
        
        Added affinity to VisiblePosition.  Changed Selection code to use affinity more.
        
        (Partial) <rdar://problem/3982096> editing/pasteboard/paste-text-007 is failing
        
        Changed ReplaceSelectionCommand to also pay attention to the top children of the incoming fragment, not just the very last node inserted, when deciding whether to insert a paragraph for the Apple interchange newline.

        * khtml/ecma/kjs_window.cpp:
        (SelectionFunc::tryCall):
        * khtml/editing/htmlediting.cpp:
        (khtml::EditCommandPtr::setStartingSelection):
        (khtml::EditCommandPtr::setEndingSelection):
        (khtml::EditCommand::setStartingSelection):
        (khtml::EditCommand::setEndingSelection):
        (khtml::CompositeEditCommand::deleteInsignificantTextDownstream):
        (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary):
        (khtml::ApplyStyleCommand::splitTextAtStartIfNeeded):
        (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded):
        (khtml::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
        (khtml::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
        (khtml::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
        (khtml::ApplyStyleCommand::mergeEndWithNextIfIdentical):
        (khtml::ApplyStyleCommand::joinChildTextNodes):
        (khtml::DeleteSelectionCommand::initializePositionData):
        (khtml::DeleteSelectionCommand::handleGeneralDelete):
        (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete):
        (khtml::DeleteSelectionCommand::doApply):
        (khtml::InsertLineBreakCommand::doApply):
        (khtml::InsertParagraphSeparatorCommand::doApply):
        (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
        (khtml::InsertTextCommand::input):
        (khtml::MoveSelectionCommand::doApply):
        (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand):
        (khtml::ReplaceSelectionCommand::~ReplaceSelectionCommand):
        (khtml::ReplaceSelectionCommand::doApply):
        (khtml::ReplaceSelectionCommand::completeHTMLReplacement):
        (khtml::ReplaceSelectionCommand::updateNodesInserted):
        (khtml::TypingCommand::deleteKeyPressed):
        (khtml::TypingCommand::forwardDeleteKeyPressed):
        (khtml::TypingCommand::markMisspellingsAfterTyping):
        * khtml/editing/htmlediting.h:
        * khtml/editing/selection.cpp:
        (khtml::Selection::Selection):
        (khtml::Selection::init):
        (khtml::Selection::moveTo):
        (khtml::Selection::modifyExtendingRightForward):
        (khtml::Selection::modifyMovingRightForward):
        (khtml::Selection::modifyExtendingLeftBackward):
        (khtml::Selection::modifyMovingLeftBackward):
        (khtml::Selection::modify):
        (khtml::Selection::xPosForVerticalArrowNavigation):
        (khtml::Selection::clear):
        (khtml::Selection::setBase):
        (khtml::Selection::setExtent):
        (khtml::Selection::setBaseAndExtent):
        (khtml::Selection::layout):
        (khtml::Selection::validate):
        * khtml/editing/selection.h:
        (khtml::Selection::startAffinity):
        (khtml::Selection::endAffinity):
        (khtml::Selection::baseAffinity):
        (khtml::Selection::extentAffinity):
        (khtml::operator==):
        * khtml/editing/text_affinity.h:
        (khtml::):
        * khtml/editing/visible_position.cpp:
        (khtml::VisiblePosition::VisiblePosition):
        (khtml::VisiblePosition::init):
        (khtml::VisiblePosition::initUpstream):
        (khtml::VisiblePosition::initDownstream):
        (khtml::VisiblePosition::next):
        (khtml::VisiblePosition::previous):
        (khtml::startVisiblePosition):
        (khtml::endVisiblePosition):
        * khtml/editing/visible_position.h:
        (khtml::VisiblePosition::):
        (khtml::VisiblePosition::VisiblePosition):
        (khtml::VisiblePosition::affinity):
        (khtml::VisiblePosition::setAffinity):
        (khtml::operator==):
        * khtml/editing/visible_units.cpp:
        (khtml::previousBoundary):
        (khtml::nextBoundary):
        (khtml::startOfLine):
        (khtml::endOfLine):
        (khtml::previousLinePosition):
        (khtml::nextLinePosition):
        (khtml::startOfParagraph):
        (khtml::endOfParagraph):
        (khtml::previousParagraphPosition):
        (khtml::nextParagraphPosition):
        (khtml::startOfBlock):
        (khtml::endOfBlock):
        (khtml::startOfDocument):
        (khtml::endOfDocument):
        * khtml/editing/visible_units.h:
        * khtml/khtml_part.cpp:
        (KHTMLPart::findTextNext):
        (KHTMLPart::selectClosestWordFromMouseEvent):
        (KHTMLPart::handleMousePressEventTripleClick):
        (KHTMLPart::handleMousePressEventSingleClick):
        (KHTMLPart::handleMouseMoveEventSelection):
        (KHTMLPart::khtmlMouseReleaseEvent):
        (KHTMLPart::selectAll):
        (KHTMLPart::computeAndSetTypingStyle):
        (KHTMLPart::selectionComputedStyle):
        * khtml/rendering/render_br.cpp:
        (RenderBR::positionForCoordinates):
        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::updateSelection):
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeBaseImpl::setFocus):
        * khtml/xml/dom_position.cpp:
        (DOM::Position::previousCharacterPosition):
        (DOM::Position::nextCharacterPosition):
        * khtml/xml/dom_position.h:
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject value]):
        (-[KWQAccObject visiblePositionForStartOfTextMarkerRange:]):
        (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
        (-[KWQAccObject accessibilityAttributeValue:]):
        (-[KWQAccObject doAXLineForTextMarker:]):
        (-[KWQAccObject doAXTextMarkerRangeForLine:]):
        (-[KWQAccObject doAXTextMarkerForPosition:]):
        (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
        (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
        (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
        (-[KWQAccObject doSetAXSelectedTextMarkerRange:]):
        * kwq/KWQAccObjectCache.mm:
        (KWQAccObjectCache::textMarkerForVisiblePosition):
        (KWQAccObjectCache::visiblePositionForTextMarker):
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::findString):
        (KWQKHTMLPart::advanceToNextMisspelling):
        (KWQKHTMLPart::styleForSelectionStart):
        (KWQKHTMLPart::baseWritingDirectionForSelectionStart):
        (KWQKHTMLPart::setSelectionFromNone):
        (KWQKHTMLPart::respondToChangedSelection):
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge setSelectedDOMRange:affinity:]):
        (-[WebCoreBridge selectionAffinity]):
        (-[WebCoreBridge setMarkDOMRange:]):
        (-[WebCoreBridge _visiblePositionForPoint:]):
        (-[WebCoreBridge moveDragCaretToPoint:]):
        (-[WebCoreBridge editableDOMRangeForPoint:]):
        (-[WebCoreBridge ensureSelectionVisible]):
        (-[WebCoreBridge rangeOfCharactersAroundCaret]):

2005-02-09  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3985211> Seed: Mail: Drag-and-drop destination indicator / insertion point disappears

        Reviewed by kocienda.

        * khtml/rendering/render_block.cpp:
        (khtml::RenderBlock::paintCaret): new
        (khtml::RenderBlock::paintObject): call paintCaret for the cursor caret and the drag caret
        * khtml/rendering/render_block.h:

=== Safari-185 ===

2005-02-08  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3996344> Entire document content is deleted when only the first paragraph was supposed to be

        Problem stems from my attempt to fix this bug: <rdar://problem/3986155> Insertion point goes 
        to beginning of doc after deleting.
        
        I added an "optimization" that proved to do more harm than good. So, basically, the fix is to remove code.

        * khtml/editing/htmlediting.cpp:
        (khtml::DeleteSelectionCommand::handleGeneralDelete)

2005-02-07  Darin Adler  <darin@apple.com>

        Reviewed by Kevin.

        - fixed <rdar://problem/3890352> REGRESSION (125-170, Panther-only): pasted textarea contents disappear (exception in NSText)

        * kwq/KWQTextArea.mm: (-[KWQTextArea tile]): Don't resize the text view immediately here.
        Use performWithDelay to defer it to avoid the re-entrancy that bothers the Panther version
        of NSTextView. On Tiger, leave the code alone.

2005-02-07  Adele Amchan  <adele@apple.com>

        Reviewed by Darin.

        Fix for <rdar://problem/3993628> REGRESSION (180-181): Logitelnet bank's website doesn't work

        This fixes an oversight in the fix for <rdar://problem/3964286>.  
        We're now making sure there is a document before calling checkCompleted in slotLoaderRequestDone.
        That function is called even when other parts finish loads of their subresources, so we could end
        up calling checkCompleted before we had a document. We also looked over all the other calls to
        checkCompleted to convince ourselves no other call sites had a similar issue.

        * khtml/khtml_part.cpp: (KHTMLPart::slotLoaderRequestDone): Check if doc is nil before calling
        checkCompleted.

2005-02-07  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Darin

        Fix for this bug:
        
        <rdar://problem/3953302> Replacing quoted text ends up with blue-colored text that is not quoted

        * khtml/editing/htmlediting.cpp:
        (khtml::DeleteSelectionCommand::doApply): Removed special case delete code for select all. It was
        causing as many bugs as it fixed. What's more, the design concept is a bit shaky. By removing
        this special case, the bug goes away.
        * khtml/editing/htmlediting.h: Remove declaration.
        
        Test results using selectAll() updated to reflect changed behavior.
        
        * layout-tests/editing/deleting/delete-3775172-fix-expected.txt:
        * layout-tests/editing/deleting/delete-select-all-001-expected.txt:
        * layout-tests/editing/deleting/delete-select-all-003-expected.txt:
        * layout-tests/editing/inserting/insert-3654864-fix-expected.txt:
        * layout-tests/editing/inserting/insert-3775316-fix-expected.txt:
        * layout-tests/editing/inserting/insert-3851164-fix-expected.txt:
        * layout-tests/editing/inserting/insert-3907422-fix-expected.txt:
        * layout-tests/editing/pasteboard/paste-text-015-expected.txt:

2005-02-06  Darin Adler  <darin@apple.com>

        Reviewed by Ken.

        - fixed <rdar://problem/3994164> setting a new style attribute with setAttribute("style") doesn't blow away the old style

        * khtml/css/css_valueimpl.h: Renamed parseProperty to parseDeclaration, since it's used to parse
        an entire style declaration, not a single property.
        * khtml/css/css_valueimpl.cpp: (DOM::CSSMutableStyleDeclarationImpl::parseDeclaration): Renamed,
        and added code to clear m_values.
        * khtml/html/html_elementimpl.cpp: (HTMLElementImpl::parseHTMLAttribute): Call function by its
        new name, and remove some unnecessary comments.

2005-02-06  Darin Adler  <darin@apple.com>

        Reviewed by Maciej.

        - fixed <rdar://problem/3993722> need oncut and onpaste handlers implemented for <textarea> (needed for widgets)

        * kwq/KWQTextArea.mm:
        (-[KWQTextAreaTextView dispatchHTMLEvent:]): Added.
        (-[KWQTextAreaTextView cut:]): Dispatch "before cut" and "cut" events.
        (-[KWQTextAreaTextView copy:]): Dispatch "before copy" and "copy" events.
        (-[KWQTextAreaTextView paste:]): Dispatch "before paste" and "paste" events.
        (-[KWQTextAreaTextView pasteAsPlainText:]): Ditto.
        (-[KWQTextAreaTextView pasteAsRichText:]): Ditto.

2005-02-06  Darin Adler  <darin@apple.com>

        Reviewed by Maciej.

        - fixed <rdar://problem/3943038> <input type=search> that is focused in onload handler doesn't have a visible editor

        * khtml/xml/dom_docimpl.cpp: (DocumentImpl::setFocusNode): Always update layout before giving focus to a widget.
        This prevents the bad case where we give a widget focus before it has been positioned or sized, causing us to
        exercise edge cases AppKit doesn't handle well.

        * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Remove workaround code I added
        to fix bug 3943049. Updating the layout fixes the same problem in a cleaner, safer way, and solves another problem,
        with the way the clip view is set up.

2005-02-06  Darin Adler  <darin@apple.com>

        Reviewed by Maciej.

        - fixed <rdar://problem/3425232> textarea won't trigger onchange action when clicking to new field

        * khtml/rendering/render_form.h: Added new m_dirty data member.
        * khtml/rendering/render_form.cpp:
        (RenderTextArea::RenderTextArea): Initialize m_dirty to false.
        (RenderTextArea::detach): Call value() on the DOM element to cause it to fetch the value from this object.
        (RenderTextArea::handleFocusOut): Ditto. Also send the onchange event based only on the m_dirty flag.
        (RenderTextArea::updateFromElement): Clear the m_dirty flag if the edited value is blown away by a value
        that comes from the DOM element. Remove code to clear the element's m_dirtyvalue flag; that's an unnecessary
        optimization, and ideally we'll stop having code in the render object that knows about that flag later.
        (RenderTextArea::slotTextChanged): Set the m_dirty flag.

2005-02-06  Darin Adler  <darin@apple.com>

        Reviewed by Maciej.

        - re-fixed <rdar://problem/3760910> please add support for custom tag names in HTML

        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::tagId): Use ID_LAST_TAG + 1 for the ID of the first assigned tag ID. Before we'd use ID_LAST_TAG,
        which resulted in the same number being used for ID_COMMENT and the first custom tag.
        (DocumentImpl::tagName): Use ID_LAST_TAG + 1, as above.
        (DocumentImpl::namespaceURI): Change a < ID_LAST_TAG to <= ID_LAST_TAG in the same spirit as above.

2005-02-06  Darin Adler  <darin@apple.com>

        Reviewed by Maciej.

        - fixed <rdar://problem/3986639> Crash occurs after choosing Undo Typing from the Edit menu

        * kwq/KWQTextArea.mm: (-[KWQTextArea viewWillMoveToWindow:]): Remove actions from undo manager when the view is
        removed from the window. Unfortunately, to do this the code has to know that the target of NSTextView actions
        will be the text storage object, but given the NSText architecture, that's pretty clear.

2005-02-05  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3991974> REGRESSION: www.jabra.com world location screen does not work

        Reviewed by hyatt.

        * kwq/KWQKJavaAppletWidget.mm:
        (KJavaAppletWidget::KJavaAppletWidget): fallback to the document's base URL if there is no baseURL attribute
        * kwq/WebCoreBridge.h:
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge baseURL]): new

2005-02-04  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        Rolled out Ken's accidental checkin when committing <rdar://problem/3986155> fix yesterday.
        
        * khtml/editing/selection.cpp:
        (khtml::Selection::validate):

2005-02-04  Adele Amchan  <adele@apple.com>

        Reviewed by Darin.

        Fix for <rdar://problem/3758785> Safari no longer works at BankOfAmerica online banking for military users
  
        When closing the document, if a body doesn't exist we now create one.  This bug was preventing the onload handler from
        firing in cases where there wasn't a body.  In the BankOfAmerica case, the page was just a bit of script where the onload
        handler redirected to the correct page.

        * khtml/xml/dom_docimpl.cpp: (DocumentImpl::close):

2005-02-03  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Chris

        Fix for this bug:
        
        <rdar://problem/3986155> Insertion point goes to beginning of doc after deleting

        * khtml/editing/htmlediting.cpp:
        (khtml::DeleteSelectionCommand::setStartNode): New convenience to handle reference counting when setting.
        (khtml::DeleteSelectionCommand::handleGeneralDelete): This contains the crux of the bug fix. Improve tests
        that detect when a selected node needs to be retained, rather than deleted, to preserve the intent of the user.
        This has the side effect of causing the insertion point placement code to succeed rather than fail. Before
        this fix, the failure of the insertion point placement code caused the insertion point to jump to the start
        of the document, which is the symptom that can be perceived by users when editing.
        * khtml/editing/htmlediting.h: Add setStartNode declaration.
        * khtml/editing/visible_units.cpp:
        (khtml::startOfBlock): This function had a stubbed-in non-tested implementation. Implement and 
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-001-expected.txt: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-001.html: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002-expected.txt: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002.html: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-003-expected.txt: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-003.html: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-004-expected.txt: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-004.html: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-005-expected.txt: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-005.html: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-006-expected.txt: Added.
        * layout-tests/editing/deleting/delete-at-paragraph-boundaries-006.html: Added.

=== Safari-183 ===

2005-02-03  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3987419> Stocks and Weather leak what appears to XMLHTTPRequest results

        XMLHTTPRequests were causing massive leaks.  _webcore_initWithHeaderString: did funky things with
        self replacement.  Re-wrote to use more traditional factory constructor avoiding self replacement.
        
        Reviewed by David Harrison.

        * kwq/KWQLoader.mm:
        (+[NSDictionary _webcore_dictionaryWithHeaderString:]):
        (KWQServeRequest):
        (KWQServeSynchronousRequest):

2005-02-03  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3938763> Cannot view Windows Media Player videos (soundtrack is played twice with delay)

        Reviewed by hyatt.

        * khtml/html/html_objectimpl.cpp:
        (HTMLObjectElementImpl::attach): set needWidgetUpdate to false right after calling updateWidget because dispatchHTMLEvent can end up calling this method again
        (HTMLObjectElementImpl::recalcStyle): ditto

2005-02-02  John Sullivan  <sullivan@apple.com>

        Reviewed by Chris.

        - WebCore part of fix for <rdar://problem/3980651> REGRESSION (125-180): Huge number of pages printed from certain page, 
        iFrame involved
        
        This also fixes the problems with printing from GMail, yay!

        * khtml/rendering/render_flow.cpp:
        (RenderFlow::paintLines):
        If the current line is taller than the entire page height (e.g. tall iFrame), don't try
        to avoid splitting it across pages.
        
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge computePageRectsWithPrintWidthScaleFactor:printHeight:]):
        Handle error cases in a more obvious manner; this will cause future problems like this
        to print a blank page and complain to the console on debug builds, rather than print a 
        zillionty mostly-blank pages.

2005-02-02  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3960304> can't load a particular applet (at www.escape.de) unless it's the first applet to be loaded

        Reviewed by hyatt.

        * khtml/rendering/render_frames.cpp:
        (RenderPartObject::updateWidget): when getting the MIME type from the PARAM tag, make sure the MIME type is the text from the left of the semi-colon if there is one. We do this elsewhere as well.

2005-02-02  Chris Blumenberg  <cblu@apple.com>

        Fixed: <rdar://problem/3983628> control-click on WebView is not selecting the word under the cursor (Mail, non-editable WebView)

        Reviewed by rjw.

        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::sendContextMenuEvent): check if SPI to always enable selecting closest word is enabled
        * kwq/WebCoreBridge.h:

2005-02-02  Ken Kocienda  <kocienda@apple.com>

        Reviewed by me

        * khtml/editing/markup.cpp:
        (khtml::createMarkup): Removed debug spam I committed earlier in error.

2005-02-02  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Darin

        Fix for this bug:
        
        <rdar://problem/3823109> WebKit should support -toggleBaseWritingDirection: (bidi editing support)

        Since base writing direction is a paragraph-level property in AppKit, and we use the CSS direction
        property in WebCore, which can be applied to inline elements as well as blocks, a new notion has
        been added to the WebCore apply style logic. It is now possible to force all properties in a 
        style declaration to be applied as though they were block properties.

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::ApplyStyleCommand): Add an enum and a flag to this command that controls whether to force all
        properties in the style to be applied as block styles. Default is "no", retaining previous behavior.
        (khtml::ApplyStyleCommand::doApply): Switch on new flag to apply styles as before, or force all preoperties to be applied
        as block styles.
        * khtml/editing/htmlediting.h:
        (khtml::ApplyStyleCommand::): Add an enum and a flag, as above.
        * khtml/khtml_part.cpp:
        (KHTMLPart::applyParagraphStyle): New method to force application of all style properties as block styles.
        * khtml/khtml_part.h: Update header declarations.
        * kwq/KWQKHTMLPart.h: Update header declarations.
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::baseWritingDirectionForSelectionStart): Accessor to help WebKit do a toggle operation on 
        writing direction.
        * kwq/WebCoreBridge.h: Update header declarations.
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge applyParagraphStyle:withUndoAction:]): New method to force application of all style properties as block styles.
        (-[WebCoreBridge baseWritingDirectionForSelectionStart]): Accessor to help WebKit do a toggle operation on 
        writing direction.

        Fix for this bug:
        
        <rdar://problem/3985035> REGRESSION (Mail): Text copied from wrapped line contains extra character when pasted

        * khtml/editing/markup.cpp:
        (khtml::renderedText): Add an enum and a flag to this command that controls whether to force all
        This is the result of an error in the code that computes the rendered text that is selected 
        when copying. Since spaces collapse at the end of lines, and these spaces need to be copied 
        when the selection spans line endings, code runs to compute this text. However, this code 
        was also running incorrectly in cases where lines wrapped. I have now added the missing 
        test to check that the selection does indeed extend to the end of the line.


2005-02-02  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Darin

        Fix for these bugs:
        
        <rdar://problem/3984894> REGRESSION (Mail): Command-right-arrow in reply does the wrong thing (two ways)
        <rdar://problem/3985130> REGRESSION (Mail): command-right-arrow in pasted RTF selects only up to tab

        * khtml/editing/visible_units.cpp:
        (khtml::endOfLine): Two separate problems in this new function recently added to take the place of
        selectionForLine. In the first, endOfLine did not stop at BR elements, and returned the position
        beyond them. In the second, the VisiblePosition constructor gives the wrong answer in certain cases.
        Darin is going to work on a solution for that problem. In the meantime, I can fix the symptom of this
        bug by going down to the last leaf child of the root line box; a one line change we will roll out when
        Darin's fix is in.

2005-02-01  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3985535> QT Plug-in JavaScript support now fails in <embed> tag only case

        Reviewed by Maciej.

        * khtml/ecma/kjs_html.cpp:
        (KJS::HTMLDocument::tryGet):

2005-02-01  John Sullivan  <sullivan@apple.com>

        Written by Darin, reviewed and tested by me
        
        - fixed <rdar://problem/3969684> Panther-only: extra blank lines between 
        pasted content from Blot

        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::attributedString):
        don't emit paragraph breaks if the margins are tiny; matches what
        we do for plain-text conversion

=== Safari-182 ===

2005-02-01  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3985160> Deficiencies in pasting architecture blocking progress on other bugs

        * khtml/editing/html_interchange.h: Move style span text used to mark element added to 
        add style to this header.
        * khtml/editing/htmlediting.cpp:
        (khtml::styleSpanClassString): Change to use constant moved to html_interchange.h.
        (khtml::isStyleSpan): New helper function. Checks if this is a span we added to apply style.
        (khtml::CompositeEditCommand::insertNodeBefore): Added an assert to check that the node
        we are inserting before is not the body.
        (khtml::CompositeEditCommand::insertNodeAfter): Ditto, but check is for after.
        (khtml::ReplacementFragment::ReplacementFragment): Added code to process the "default style"
        that is added by the copy code.
        (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): first and last nodes inserted are
        now member variables instead of function locals. Initialize them here.
        (khtml::ReplaceSelectionCommand::~ReplaceSelectionCommand): Deref first and last nodes inserted
        if necessary.
        (khtml::ReplaceSelectionCommand::doApply): Change design to fix the bug. Major change is to
        separate out the code that inserts nodes into the tree so additional styling checks can
        be done in a centralized way. Also got rid of the notion of "merging into the end block." That
        concept was just wrong.
        (khtml::ReplaceSelectionCommand::completeHTMLReplacement): Tweak interface now that first and 
        last nodes inserted are member variables.
        (khtml::ReplaceSelectionCommand::insertNodeAfterAndUpdateNodesInserted): New helper used
        by replace code to do the stated DOM operation and update state internal to the command.
        This will also be a catch point to handle the kinds of additional style checks needed to
        make paste work right.
        (khtml::ReplaceSelectionCommand::insertNodeAtAndUpdateNodesInserted): Ditto.
        (khtml::ReplaceSelectionCommand::insertNodeBeforeAndUpdateNodesInserted): Ditto.
        (khtml::ReplaceSelectionCommand::updateNodesInserted): Ditto.
        * khtml/editing/htmlediting.h: Update declarations as needed.
        * khtml/editing/markup.cpp:
        (khtml::createMarkup): Adds a "default style" span to the content written to the pasteboard.
        This will help us to fix some of the bugs blocked by the bug above.
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::lastDescendent): New helper.
        * khtml/xml/dom_nodeimpl.h: Ditto.
        * khtml/xml/dom_position.cpp:
        (DOM::Position::upstream): Fixed a bug which would allow the upstream position returned to be
        in unrendered content.
        (DOM::Position::downstream): Ditto.

        * layout-tests/editing/deleting/delete-3775172-fix-expected.txt: Ending positions tweaked due to 
        changes in upstream() and downstream() functions.
        * layout-tests/editing/inserting/insert-3851164-fix-expected.txt: Ditto
        * layout-tests/editing/inserting/insert-3907422-fix-expected.txt: Ditto
        * layout-tests/editing/selection/extend-by-character-006-expected.txt: Ditto

        * layout-tests/editing/pasteboard/paste-text-003-expected.txt: Changed what we expect, given
        new behavior of paste code.

2005-01-31  John Sullivan  <sullivan@apple.com>

        Reviewed by Dave Hyatt.
        
        - fixed <rdar://problem/3983097> Tabbing on RSS pages gets stuck in search field 
        with "full keyboard navigation" off

        * kwq/KWQSlider.h:
        * kwq/KWQSlider.mm:
        (QSlider::focusPolicy):
        Implemented this method for KWQSlider. It is needed for our widget subclasses that
        can become focused in some situations. I missed this when fixing 3949203 because
        it only comes into play when tab-to-controls is on, and because the other relevant
        methods are objective-C methods.

2005-01-31  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3964164> AXWebArea needs to unregister on going invisible

        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::detach):
        Detach the AX UIElement for the AXWebArea.

2005-01-31  David Harrison  <harrison@apple.com>

        Reviewed by Darin

        <rdar://problem/3850876> AX: button exposed as only an AXImage at http://appleseed.apple.com/ site (<input type=image>)

        * khtml/rendering/render_form.h:
        (khtml::RenderImageButton::isImageButton):
        New.
        * khtml/rendering/render_image.h:
        (khtml::RenderImage::isImageButton):
        New.
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject isImageButton]):
        (-[KWQAccObject actionElement]):
        (-[KWQAccObject role]):
        (-[KWQAccObject accessibilityActionNames]):
        (-[KWQAccObject accessibilityPerformAction:]):
        Represent RenderImageButton elements as buttons instead of images.

2005-01-31  David Harrison  <harrison@apple.com>

        Prevent nil dereference in debug code.

        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::displayTree):
        Stop when parent is 0, too.

2005-01-31  John Sullivan  <sullivan@apple.com>

        Reviewed by Dave Hyatt.
        
        - fixed <rdar://problem/3949203> cannot tab to, within, or out of the RSS sidebar

        * kwq/KWQSlider.mm:
        (-[KWQSlider becomeFirstResponder]):
        (-[KWQSlider resignFirstResponder]):
        (-[KWQSlider nextKeyView]):
        (-[KWQSlider previousKeyView]):
        (-[KWQSlider canBecomeKeyView]):
        (-[KWQSlider nextValidKeyView]):
        (-[KWQSlider previousValidKeyView]):
        Copied these methods from KWQButton. The lack of special handling for first-responder-ness
        here was causing the focus to get "stuck" on the RSS page's slider widget. Added FIXME
        about how it would be nice to share more of this code rather than replicating it in each
        KWQ widget subclass.

2005-01-31  Darin Adler  <darin@apple.com>

        Reviewed by Harrison.

        - fixed <rdar://problem/3980066> Double-click on single character moves insertion point to previous line

        * khtml/khtml_part.cpp:
        (KHTMLPart::selectClosestWordFromMouseEvent): Set affinity too.
        (KHTMLPart::handleMousePressEventTripleClick): Ditto.
        (KHTMLPart::handleMouseMoveEventSelection): Ditto.
        (KHTMLPart::khtmlMouseReleaseEvent): Ditto.

        * khtml/editing/selection.cpp:
        (khtml::Selection::modifyExtendingRightForward): Use endOfLine and endOfDocument.
        (khtml::Selection::modifyMovingRightForward): Ditto.
        (khtml::Selection::modifyExtendingLeftBackward): Use startOfLine and startOfDocument.
        (khtml::Selection::modifyMovingLeftBackward): Ditto.
        (khtml::Selection::validate): Rewrote the section that handles double-click. Two main fixes: 1) use isStartOfLine to
        check for another case where we want to select the word to the right, and 2) use isEndOfParagraph, which seems
        to work correctly in cases where isLastVisiblePositionInParagraph is giving the wrong answer. Also changed the line
        code to use startOfLine/endOfLine and the document code to use startOfDocument/endOfDocument.

2005-01-31  Darin Adler  <darin@apple.com>

        Reviewed by Harrison.

        - fixed <rdar://problem/3935275> unexpected quit scrolling over link; last.fm (works in IE and Firefox)

        * kwq/WebCoreBridge.mm: (-[WebCoreBridge elementAtPoint:]): By using the correct document, fix the crash reported here.
        The page still crashes when you close the browser window, but it's not trivial to fix so I'll file a new bug after that.

2005-01-31  Darin Adler  <darin@apple.com>

        Reviewed by Ken and Harrison.

        - fixed <rdar://problem/3947901> REGRESSION (Mail): Pasting paragraph of rich text leaves insertion point before pasted text
        - fixed <rdar://problem/3949790> hitting return after underlined line results in too much or too little underlined
        - fixed <rdar://problem/3981759> nil-deref and crash when pasting just a paragraph break
        - fixed a couple problems I discovered while working with bug 3949790

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::applyInlineStyle): Pass StayInBlock to upstream. Without this, we end up going too far
        upstream in the test case in bug 3949790.
        (khtml::ApplyStyleCommand::removeInlineStyle): Pass StayInBlock to upstream and downstream. Same reason as above.
        (khtml::ReplaceSelectionCommand::doApply): Update endPos if inserting a new node and endPos is using that node's
        parent and an offset past the node being inserted. That change fixes a problem with the position of the insertion point
        after pasting into the top level of a document (from test cases in 3947901 and 3949790). When setting insertionPos, use
        code that works when lastNodeInserted is a block rather than a text node. That change fixes a problem where a newline is
        not added when pasting an entire paragraph into the end of a document (from test case in 3949790). Added nil check before
        checking if lastNodeInserted is a <br> element, which fixes the crash when pasting just a paragraph break.

        * khtml/editing/visible_units.h: Filled out the set of calls to add some boolean checks for lines (needed for the
        bug fix), and calls for blocks (not yet implemented), and documents. The document checks may need refinement to
        properly handle documents with a mix of editable and non-editable content, but for now they just refactor code
        and make things a little clearer. Also removed the "include line break" parameter from endOfSentence.
        * khtml/editing/visible_units.cpp:
        (khtml::rootBoxForLine): Added.
        (khtml::startOfLine): Added. Algorithm taken from selectionForLine in selection.cpp.
        (khtml::endOfLine): Ditto.
        (khtml::inSameLine): Added.
        (khtml::isStartOfLine): Added.
        (khtml::isEndOfLine): Added.
        (khtml::endOfSentence): Removed "include line break" parameter.
        (khtml::inSameParagraph): Added a null check.
        (khtml::isStartOfParagraph): Ditto.
        (khtml::isEndOfParagraph): Ditto.
        (khtml::startOfBlock): Added.
        (khtml::endOfBlock): Added.
        (khtml::inSameBlock): Added.
        (khtml::isStartOfBlock): Added.
        (khtml::isEndOfBlock): Added.
        (khtml::startOfDocument): Added.
        (khtml::endOfDocument): Added.
        (khtml::inSameDocument): Added.
        (khtml::isStartOfDocument): Added.
        (khtml::isEndOfDocument): Added.

2005-01-30  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3977000> form data set posted for <input type="image" ...> omits name/value pair

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLInputElementImpl::appendFormData): Append a name/value pair for type IMAGE in addition to the X/Y
        coordinates.

2005-01-30  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3974246> REGRESSION (125-180): popup menus are missing item text on Harmony Remote web site

        * khtml/html/html_formimpl.cpp: (DOM::HTMLOptionElementImpl::text): Change this function to traverse the entire tree
        and gather all the text rather than just looking at immediate children.

        * khtml/xml/dom_nodeimpl.h: Added const.
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::traverseNextNode): Added const.
        (NodeImpl::traverseNextSibling): Ditto.
        (NodeImpl::traversePreviousNodePostOrder): Ditto.
        (NodeImpl::detach): Add a missing nil check.

2005-01-30  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3864536> crash copying text from other apps and pasting in a sticky in dashboard

        * khtml/editing/markup.cpp: (khtml::createFragmentFromText): Added missing ref/deref of newly created element.

2005-01-28  Darin Adler  <darin@apple.com>

        * khtml/css/cssproperties.c: Regenerated with newer gperf.

2005-01-28  David Harrison  <harrison@apple.com>

        Reviewed by John Sullivan.

        <rdar://problem/3968144> AX need to be able to focus an AXLink by setting AXFocused to true
        
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject accessibilityPerformAction:]):
        (-[KWQAccObject accessibilityIsAttributeSettable:]):
        Add AXFocused for AXLinks.

2005-01-28  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3584942> AX: Safari Accessibility parent-child mismatch

        Use AppKit SPI _accessibilityParentForSubview to return KWQAccObject parent of AppKit AX object.
        
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject _accessibilityParentForSubview:]):
        New.

2005-01-27  Adele Amchan  <adele@apple.com>

        fixed by Darin, reviewed by me.

        <rdar://problem/3976314> REGRESSION (180-TOT): submitting password fields fail on Panther

        * kwq/KWQTextField.mm: (-[KWQTextFieldController string]): on panther, the secure text field's editor 
        does not contain the real string, so now we just call stringValue on the field in that case.

2005-01-26  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Darin, Hyatt and Ken.

        <rdar://problem/3790449> REGRESSION (Mail): underline behavior is flaky because of how CSS handles text-decoration
        
        * khtml/css/css_computedstyle.cpp:
        (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): Add support
        for -khtml-text-decorations-in-effect property. This works like text-docration
        but follows proper inline semantics, that is, if a parent is causing text
        decoration then -khtml-text-decorations-in-effect on this element will reflect that.
        * khtml/css/cssproperties.in: Add -khtml-text-decorations-in-effect property.
        * khtml/css/cssproperties.h: regenerated
        * khtml/css/cssvalues.c: regenerated
        * khtml/css/cssparser.cpp: regenerated
        * khtml/css/cssproperties.c: regenerated

        * khtml/css/css_valueimpl.h: Made copyPropertiesInSet public.

        * khtml/editing/htmlediting.cpp:
        (khtml::StyleChange::init): Handle -khtml-text-decorations-in-effect
        specially, translate to text-decoration
        (khtml::ApplyStyleCommand::applyInlineStyle): Fix a comment.
        (khtml::ApplyStyleCommand::removeInlineStyle): Add new way of pushing down
        text-decoration styles around the edges of the selected range.
        (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAroundNode): Helper for this.
        (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAtBoundaries): ditto
        (khtml::ApplyStyleCommand::nodeFullyUnselected): ditto
        (khtml::hasTextDecorationProperty): ditto
        (khtml::highestAncestorWithTextDecoration): ditto
        (khtml::ApplyStyleCommand::extractTextDecorationStyle): ditto
        (khtml::ApplyStyleCommand::extractAndNegateTextDecorationStyle): ditto
        (khtml::ApplyStyleCommand::applyTextDecorationStyle): ditto
        * khtml/editing/htmlediting.h: Prototype new methods.
        * khtml/editing/jsediting.cpp: Make underline execCommand use the new CSS property.

        - as a necessary part of the fix, fixed the longstanding problem
        that mutating an element's inlineStyleDecl would not property
        update its style attribute. now it does.
        
        * khtml/css/css_valueimpl.cpp:
        (DOM::CSSMutableStyleDeclarationImpl::setChanged): If this is an inline style
        declaration, tell the element it's style attribute needs updating.
        * khtml/html/html_elementimpl.cpp:
        (HTMLElementImpl::invalidateStyleAttribute): New method, just calls down
        to the dom element.
        (HTMLElementImpl::updateStyleAttribute): Regenerate style attribute from
        inline style declaration.
        * khtml/html/html_elementimpl.h: Prototype new stuff.
        * khtml/xml/dom_elementimpl.cpp:
        (ElementImpl::ElementImpl): Initialize new attribute.
        (ElementImpl::updateStyleAttributeIfNeeded): New method to check if the style
        is dirty and this is an html element, and if so call the html element to
        regenerate the style attribute.
        (ElementImpl::attributes): update style if needed
        (ElementImpl::getAttribute): ditto
        (ElementImpl::hasAttributes): ditto
        (ElementImpl::dump): ditto
        (XMLElementImpl::cloneNode): ditto
        * khtml/xml/dom_elementimpl.h:
        (DOM::ElementImpl::isStyleAttributeValid): new method to manage style attr
        dirty bit
        (DOM::ElementImpl::setStyleAttributeValid): ditto

        - added new layout tests for underlining:
        
        * layout-tests/editing/editing.js:
        * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-across-paragraph-in-bold.html: Added.
        * layout-tests/editing/style/remove-underline-across-paragraph.html: Added.
        * layout-tests/editing/style/remove-underline-after-paragraph-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-after-paragraph-in-bold-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-after-paragraph-in-bold.html: Added.
        * layout-tests/editing/style/remove-underline-after-paragraph.html: Added.
        * layout-tests/editing/style/remove-underline-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-from-stylesheet-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-from-stylesheet.html: Added.
        * layout-tests/editing/style/remove-underline-in-bold-expected.txt: Added.
        * layout-tests/editing/style/remove-underline-in-bold.html: Added.
        * layout-tests/editing/style/remove-underline.html: Added.
        * layout-tests/editing/style/underline-expected.txt: Added.
        * layout-tests/editing/style/underline.html: Added.

2005-01-27  David Hyatt  <hyatt@apple.com>

        Fix for 3875199, search field needs to be able to  show a magnifying  glass without a dropdown menu.
        
        Reviewed by kocienda

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLInputElementImpl::HTMLInputElementImpl):
        (DOM::HTMLInputElementImpl::parseHTMLAttribute):
        * kwq/KWQLineEdit.mm:
        (QLineEdit::setMaxResults):

2005-01-27  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Maciej

        Fix for this bug:
        
        <rdar://problem/3973254> Deletions of ranges does not coalesce correctly with subsequent typing

        * khtml/editing/htmlediting.cpp:
        (khtml::InsertTextCommand::deleteCharacter): Remove this dead code.
        (khtml::ReplaceSelectionCommand::editingAction): Moved this code to a better place in the file. 
        It was in with unrelated functions.
        (khtml::TypingCommand::TypingCommand): Reorganize initialization list so it is easier to read.
        (khtml::TypingCommand::issueCommandForDeleteKey): Remove dead code. Roll in remaining code into deleteKeyPressed.
        (khtml::TypingCommand::deleteKeyPressed): Add in support for carrying along smart delete flag.
        (khtml::TypingCommand::forwardDeleteKeyPressed): New function. Makes forward delete work like
        "regular" delete in terms of coalescing typing.
        (khtml::TypingCommand::doApply): Add case for ForwardDeleteKey.
        (khtml::TypingCommand::preservesTypingStyle): Ditto.
        * khtml/editing/htmlediting.h:
        (khtml::TypingCommand::): Add ForwardDeleteKey constant. Remove a couple declarations for now-dead code.
        (khtml::TypingCommand::smartDelete): New accessor.
        (khtml::TypingCommand::setSmartDelete): Ditto.
        * kwq/WebCoreBridge.h:
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge deleteKeyPressedWithSmartDelete:]): Add smart delete flag.
        (-[WebCoreBridge forwardDeleteKeyPressedWithSmartDelete:]): New method.
        * khtml/editing/jsediting.cpp: Add supprt for ForwardDelete command
        * layout-tests/editing/editing.js: Ditto.
        * layout-tests/editing/deleting/delete-and-undo-expected.txt: Added.
        * layout-tests/editing/deleting/delete-and-undo.html: Added.
        * layout-tests/editing/deleting/forward-delete-expected.txt: Added.
        * layout-tests/editing/deleting/forward-delete.html: Added.

=== Safari-181 ===

2005-01-27  David Harrison  <harrison@apple.com>

        Reviewed by Ken.

        <rdar://problem/3962214> AX: AXLeftLineTextMarkerRangeForTextMarker returns incorrect range

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
        Allowed for selection to end of line including the linebreak.
        
2005-01-27  David Harrison  <harrison@apple.com>

        Reviewed by Darin, Ken.

        <rdar://problem/3964470> AX: Include attachments in AXAttributedStringForTextMarkerRange

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject isAttachment]):
        (-[KWQAccObject attachmentView]):
        (-[KWQAccObject role]):
        (-[KWQAccObject subrole]):
        (-[KWQAccObject roleDescription]):
        (-[KWQAccObject value]):
        (-[KWQAccObject title]):
        (-[KWQAccObject accessibilityDescription]):
        (-[KWQAccObject accessibilityIsIgnored]):
        (-[KWQAccObject accessibilityAttributeNames]):
        (-[KWQAccObject accessibilityAttributeValue:]):
        Make attachments accessible.
        
        (CreateCGColorIfDifferent):
        (AXAttributeStringSetColor):
        (AXAttributeStringSetNumber):
        (AXAttributeStringSetFont):
        (AXAttributeStringSetStyle):
        (AXAttributeStringSetElement):
        (AXLinkElementForNode):
        (AXAttributedStringAppendText):
        (AXAttributedStringAppendReplaced):
        (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
        Add attachments and more text attributes to AXAttributedStringForTextMarkerRange.

2005-01-27  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3807935> DOM CSS computed style line-height is wrong in two ways

        * khtml/css/css_computedstyle.cpp: (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue):
        Use specified size rather than computed size when dealing with a percentage for line height.

2005-01-27  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3971372> SWB: template function DOM_cast() won't compile with gcc 4.0
        - fixed other gcc 4.0 compiling problems Patrick Beard pointed out on the phone

        * kwq/DOMInternal.h: (DOM_cast): Change the "failToCompile()" trick to use a dependent name.
        Nowadays, if a name is not dependent, it is checked at template definition time. But we want
        a failure only at template instantiation time.

        * ForwardingHeaders/editing/text_granularity.h: Added. Needed but a bug in the 3.3 compiler made
        it compile anyway without this.

        - other changes

        * khtml/html/html_miscimpl.h: For clarity, mark these functions virtual too. They are automatically
        virtual because the base class ones are virtual, but it's ugly to leave it this way.

2005-01-26  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3971609> REGRESSION (Mail): up/down arrow navigation broken after rewrapping text by resizing window
        <rdar://problem/3975661> REGRESSION (Mail): left/right arrow navigation can place insertion point after last character on line

        * khtml/editing/selection.cpp:
        (khtml::Selection::modifyAffinity): PARAGRAPH and LINE movements should not alter affinity. This was
        just a mistake before when it reset to upstream in some cases.
        * khtml/rendering/render_line.cpp:
        (khtml::RootInlineBox::closestLeafChildForXPos): Move this function down from InlineBox, and 
        rework the implementation. This fixes 3971609.
        * khtml/rendering/render_line.h:
        * khtml/rendering/render_text.cpp:
        (RenderText::caretRect): Rework the algorithm to fix 3975661. This regressed when we began placing
        the spaces on the end of lines in text boxes when lines wrap.

2005-01-26  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3972522> (179-180) 40% slowdown on iBench JavaScript test

        I added a member variable to ObjectImp.  This changed it's size and consequently
        hampered the optimizations built into the garbage collector.  Objects no longer
        fit within the allocators cell size, and thus allocation fell back to a slower
        allocator.

        As a result of this fix I also dramatically cleaned up how runtime objects are
        accessed.  The path mostly *removes* code.
        
        Reviewed by Chris.

        * khtml/ecma/kjs_dom.cpp:
        (DOMDocumentProtoFunc::tryCall):
        (DOMElementProtoFunc::tryCall):
        (KJS::getRuntimeObject):
        * khtml/ecma/kjs_dom.h:
        * khtml/ecma/kjs_html.cpp:
        (KJS::HTMLDocument::tryGet):
        (KJS::HTMLElement::tryGet):
        (KJS::HTMLElement::implementsCall):
        (KJS::HTMLElement::call):
        (KJS::HTMLElement::tryPut):
        (KJS::HTMLCollection::tryGet):
        (KJS::HTMLCollection::getNamedItems):
        * khtml/ecma/kjs_html.h:
        * khtml/ecma/kjs_window.cpp:
        (Window::get):

2005-01-26  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3757712> REGRESSION (Mail): WebCore does not allow Devanagari ligature input

        For now we are using ICU UBreakIterator to determine grapheme boundaries for
        cursor and deletion.  This does not match what Cocoa does exactly, but does match
        what Carbon does.  The areas are difference are obscure, but, according to
        Deborah Goldsmith, using the UBreakIterator is the reasonable approach.

        Reviewed by Ken.

        * khtml/rendering/render_text.cpp:
        (RenderText::previousOffset):
        (RenderText::nextOffset):

2005-01-25  David Harrison  <harrison@apple.com>

        Reviewed by Maciej and Richard.

        <rdar://problem/3963731> AX VO: Changing window sizes causes a crash using with Voice Over - KWQPtrDictImpl::clear

        * khtml/html/html_miscimpl.cpp:
        (HTMLCollectionImpl::HTMLCollectionImpl):
        Ensure that isHTMLDocument before calling HTMLDocument-only method.

2005-01-25  Richard Williamson   <rjw@apple.com>

        Part one of fix for <rdar://problem/3757712> REGRESSION (Mail): WebCore does not allow Devanagari ligature input

        This patch changes the semantics of next() and previous() on
        VisiblePosition to move a grapheme (character cluster) at a
        time.  This means that cursor navigation with correctly move
        over an entire cluster.

        However, the expected behavior for deleting a grapheme is to
        delete individual code points, thus decomposing the grapheme
        into it constituent parts.  That will be addressed in the next
        part of the fix.

        Reviewed by Ken.

        * khtml/editing/visible_position.cpp:
        (khtml::VisiblePosition::previousPosition):
        (khtml::VisiblePosition::nextPosition):
        * khtml/editing/visible_position.h:
        * khtml/rendering/render_object.cpp:
        (RenderObject::previousOffset):
        (RenderObject::nextOffset):
        * khtml/rendering/render_object.h:
        * khtml/rendering/render_text.cpp:
        (RenderText::previousOffset):
        (RenderText::nextOffset):
        (RenderText::findNextInlineTextBox):
        * khtml/rendering/render_text.h:
        * khtml/xml/dom_nodeimpl.cpp:
        (NodeImpl::previousOffset):
        (NodeImpl::nextOffset):
        * khtml/xml/dom_nodeimpl.h:

2005-01-25  David Harrison  <harrison@apple.com>

        Reviewed by Maciej.

        <rdar://problem/3973067> AXWebArea for http://apple.netscape.com/ has an empty AXLinkUIElements

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject accessibilityAttributeValue:]):
        Omit ignored AXLinks from AXLinkUIElements result

2005-01-25  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3972851> REGRESSION (179-180+): Repro crash in ApplyStyleCommand::cleanUpEmptyStyleSpans

        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::cleanUpEmptyStyleSpans): Added two null checks.

2005-01-25  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3954710> Mail crashed while editing signatures - NodeImpl::isBlockFlow

        The fix is more general than for this one bug, and may work to fix many crashers. The problem
        is that the ReplaceSelectionCommand never checked whether its starting selection is empty. If
        it is, then we need to bail before doing the work of the command, which we need to deref the
        start and end points of the selection in order to do its work. I think you can see the crash
        potential.

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::doApply): Assert selection is not empty.
        * kwq/WebCoreBridge.mm:
        (partHasSelection): New helper function to test that bridge has a part with a selection.
        
        Use new helper function to test part and selection; return from these function if this test fails.
        
        (-[WebCoreBridge rangeByExpandingSelectionWithGranularity:])
        (-[WebCoreBridge rangeByAlteringCurrentSelection:direction:granularity:])
        (-[WebCoreBridge alterCurrentSelection:direction:granularity:])
        (-[WebCoreBridge rangeByAlteringCurrentSelection:verticalDistance:])
        (-[WebCoreBridge alterCurrentSelection:verticalDistance:])
        (-[WebCoreBridge documentFragmentWithText:])
        (-[WebCoreBridge replaceSelectionWithFragment:selectReplacement:smartReplace:])
        (-[WebCoreBridge insertLineBreak])
        (-[WebCoreBridge insertParagraphSeparator])
        (-[WebCoreBridge insertParagraphSeparatorInQuotedContent])
        (-[WebCoreBridge insertText:selectInsertedText:])
        (-[WebCoreBridge deleteSelectionWithSmartDelete:])
        (-[WebCoreBridge ensureSelectionVisible])

2005-01-24  Kevin Decker  <kdecker@apple.com>

        Reviewed by Darin.

        Fixed <rdar://problem/3932374> REGRESSION: 'ReferenceError - Can't find variable' JavaScript error at webxpress.fidelity.com

        * khtml/html/html_miscimpl.h: Missing virtual identifiers for namedItem() and nextNamedItem() exposed a flaw that prevented finding variables in form [HTMLFormCollectionImpl] scopes.
        

2005-01-24  Vicki Murley <vicki@apple.com>

        Reviewed by kocienda.

        - fix <rdar://problem/3810661> computed style not handled for Apple extensions
 
        * khtml/css/css_computedstyle.cpp: (DOM::):
        (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): add -apple-line-clamp and -apple-text-size-adjust

2005-01-24  Darin Adler  <darin@apple.com>

        Reviewed by Kevin.

        - fixed <rdar://problem/3969884> REGRESSION (179-180): Typing password not echoed as "bullets" at paypal.com when field is first focused

        * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]):
        Sizing up the form editor, which fixes things for normal fields and search fields,
        screws things up for password fields. Eventually, I'll need to figure out why this is
        and file a bug so that the AppKit team fixes it. In the mean time, just check for the
        case of the secure text field, and don't size up in that case.

2005-01-24  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt

        Fix for this bug:
        
        <rdar://problem/3963560> 8A354: Inserting Japanese text to the line head causes line break

        * khtml/editing/htmlediting.cpp:
        (khtml::ReplaceSelectionCommand::doApply): Add one more case where we merge content into the
        existing line. This covers the situation see using the steps to reproduce this bug.

2005-01-21  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Darin.

        <rdar://problem/3967572> Editing should split elements before removing style
        
        * khtml/editing/htmlediting.cpp:
        (khtml::CompositeEditCommand::splitElement):
        (khtml::CompositeEditCommand::mergeIdenticalElements):
        (khtml::CompositeEditCommand::wrapContentsInDummySpan):
        (khtml::CompositeEditCommand::splitTextNodeContainingElement):
        (khtml::ApplyStyleCommand::applyInlineStyle):
        (khtml::ApplyStyleCommand::splitTextAtStartIfNeeded):
        (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded):
        (khtml::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
        (khtml::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
        (khtml::areIdenticalElements):
        (khtml::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
        (khtml::ApplyStyleCommand::mergeEndWithNextIfIdentical):
        (khtml::ApplyStyleCommand::cleanUpEmptyStyleSpans):
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded):
        (khtml::InsertParagraphSeparatorCommand::doApply):
        (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
        (khtml::InsertTextCommand::prepareForTextInsertion):
        (khtml::SplitTextNodeCommand::doUnapply):
        (khtml::SplitElementCommand::SplitElementCommand):
        (khtml::SplitElementCommand::~SplitElementCommand):
        (khtml::SplitElementCommand::doApply):
        (khtml::SplitElementCommand::doUnapply):
        (khtml::MergeIdenticalElementsCommand::MergeIdenticalElementsCommand):
        (khtml::MergeIdenticalElementsCommand::~MergeIdenticalElementsCommand):
        (khtml::MergeIdenticalElementsCommand::doApply):
        (khtml::MergeIdenticalElementsCommand::doUnapply):
        (khtml::WrapContentsInDummySpanCommand::WrapContentsInDummySpanCommand):
        (khtml::WrapContentsInDummySpanCommand::~WrapContentsInDummySpanCommand):
        (khtml::WrapContentsInDummySpanCommand::doApply):
        (khtml::WrapContentsInDummySpanCommand::doUnapply):
        (khtml::SplitTextNodeContainingElementCommand::SplitTextNodeContainingElementCommand):
        (khtml::SplitTextNodeContainingElementCommand::~SplitTextNodeContainingElementCommand):
        (khtml::SplitTextNodeContainingElementCommand::doApply):
        * khtml/editing/htmlediting.h:

        New layout test that shows the side benefit of this.
        
        * layout-tests/editing/style/unbold-in-bold-expected.txt: Added.
        * layout-tests/editing/style/unbold-in-bold.html: Added.

2005-01-24  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3933435> -[DOMRange cloneContents] sometimes alters the original DOMRange

        * khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::processContents): Fixed code that was collapsing
        the range after processing the contents. That's right for extract and delete, but not clone.

        - fixed <rdar://problem/3714184> inline input works incorrectly when I type Korean slowly in textfield in form on www.google.co.kr
        - fixed <rdar://problem/3193848> Can't enter SSN to sign up for Chase visa online account (changing focus inside key press handler)
        - fixed <rdar://problem/3874683> REGRESSION (161-162): Crash after typing command-Z to undo after pasting text with newline in it into a text field <input type=text>

        * kwq/KWQLineEdit.mm:
        (QLineEdit::text): Changed to call a new string method on the controller instead of calling
        stringValue, since stringValue has a bad side effect of ending inline input.
        (QLineEdit::selectAll): Changed to only call selectText: when the field is already selected.
        When you are giving the field focus, it automatically gets all selected, and this change
        is required in the case where we defer the focus change.

        * kwq/KWQTextField.h: Added a new string method to the controller.
        * kwq/KWQTextField.mm:
        (-[KWQTextFieldController setMaximumLength:]): Call the new string method instead of using stringValue.
        (-[KWQTextFieldController controlTextDidChange:]): Removed the code to truncate at the first CR or LF.
        That's now handled in the shouldChangeTextInRange method instead.
        (-[KWQTextFieldController textView:shouldHandleEvent:]): Add new logic to defer responder changes until
        after the event is handled. This makes sure the key gets into the field that's already handling it at
        rather than ending up in the newly-focused field if the key press handler focuses a different element.
        (-[KWQTextFieldController textView:didHandleEvent:]): Turn off deferral here, now that the key press
        has been handled.
        (-[KWQTextFieldController string]): Added. Calls stringValue only if there's no editor. If there is
        an editor, calls string on that instead.
        (-[KWQTextFieldController textView:shouldChangeTextInRange:replacementString:]): Added. Called by
        the three field subclasses. Truncates incoming strings at the first CR or LF character. This works
        properly with Undo, and the old technique did not.
        (-[KWQTextFieldController preprocessString:]): Added. Helper used for all the setStringValue methods
        so they all truncate at the first CR or LF as well as truncating to the appropriate length.
        (-[KWQTextField setStringValue:]): Changed to call the controller.
        (-[KWQTextField textView:shouldChangeTextInRange:replacementString:]): Added. Calls the controller.
        (-[KWQSecureTextField setStringValue:]): More of the same.
        (-[KWQSecureTextField textView:shouldChangeTextInRange:replacementString:]): Ditto.
        (-[KWQSearchField setStringValue:]): Ditto.
        (-[KWQSearchField textView:shouldChangeTextInRange:replacementString:]): Ditto.

        * kwq/KWQWidget.h: Added the new setDeferFirstResponderChanges function.
        * kwq/KWQWidget.mm:
        (QWidget::~QWidget): Clear out the deferredFirstResponder global just in case we are destroyed
        while we "have the ball".
        (QWidget::hasFocus): If we are in the deferred mode, use the global to decide who to say is focused.
        (QWidget::setFocus): If we are in the deferred mode, set the global instead of making something be
        the first responder.
        (QWidget::setDeferFirstResponderChanges): Set the boolean. If the boolean is being cleared, then
        get the deferred first responder and call setFocus on it again, which will do the work we didn't
        want to do earlier.

2005-01-24  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3964350> Deleting all content and typing in a message gets one letter, then beeps

        This is a problem with the way we turn key events, when a user starts typing, into editing commands. 
        Right now, that initial delete key in an empty window is turned into an editing command, and much 
        editing code runs that need not run. What's more, this is confusing the selection machinery. The
        solution is to detect the "delete in an empty window" case and don't turn this into an editing
        command at all.

        * khtml/editing/htmlediting.cpp:
        (khtml::TypingCommand::deleteKeyPressed)

2005-01-23  Adele Amchan  <adele@apple.com>

        Fixed <rdar://problem/3964286> REGRESSION (178-179): menus and "go to old site" link at Ameritrade's new site don't work

        Removed check for document from checkCompleted.  In this case, WebCore never knew that a WebImageView was complete because 
        it fell into this document check case in checkCompleted.  
        Added check for document to tokenizerProcessedData (see <rdar://problem/3807144>)

        Reviewed by Darin.

        * khtml/khtml_part.cpp:
        (KHTMLPart::stop):
        (KHTMLPart::checkCompleted):
        * khtml/khtml_part.h:
        * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::tokenizerProcessedData):

2005-01-21  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3759399> Javascript / Liveconnect problems ((event handler):Undefined value)

        Java applets specified with <object> or <embed> weren't scriptable.  Now they are.

        Reviewed by Chris.

        * khtml/html/html_objectimpl.cpp:
        (HTMLEmbedElementImpl::getEmbedInstance):
        (HTMLObjectElementImpl::getObjectInstance):

2005-01-21  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3959464> REGRESSION (Mail): Insertion point goes back to beginning of document after deleting

        * khtml/editing/htmlediting.cpp:
        (khtml::DeleteSelectionCommand::handleGeneralDelete): Add special case to handle retaining a fully-selected block.
        This fixes the bug.
        * layout-tests/editing/deleting/delete-3959464-fix-expected.txt: Added.
        * layout-tests/editing/deleting/delete-3959464-fix.html: Added.

2005-01-21  Richard Williamson   <rjw@apple.com>

        Fixed <rdar://problem/3966998> REGRESSION(179-TOT) clicking on gmail message brings me to blank screen

        getElementById() sometimes returns Undefined() instead of Null().

        Reviewed by Hyatt.

        * khtml/ecma/kjs_dom.cpp:
        (DOMDocumentProtoFunc::tryCall):

2005-01-21  David Hyatt  <hyatt@apple.com>

        Fix for 3773809, make sure that overflow regions never end up at an invalid scroll offset because of a layout change.
        
        Reviewed by kocienda

        * khtml/rendering/render_layer.cpp:
        (RenderLayer::updateScrollInfoAfterLayout):

2005-01-21  David Hyatt  <hyatt@apple.com>

        Fix for 3966349, hang loading page.  Make sure that loops using popOneBlock properly check for a null
        blockStack to avoid an infinite loop.
        
        Reviewed by kocienda

        * khtml/html/htmlparser.cpp:
        (KHTMLParser::parseToken):
        (KHTMLParser::insertNode):
        (KHTMLParser::popInlineBlocks):

2005-01-21  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3957204> Mail crashed when sending a message in ~InsertParagraphSeparatorInQuotedContentCommand

        * khtml/editing/htmlediting.cpp:
        (khtml::InsertParagraphSeparatorInQuotedContentCommand::InsertParagraphSeparatorInQuotedContentCommand):
        InsertParagraphSeparatorInQuotedContentCommand has this member variable: DOM::ElementImpl *m_breakNode;
        I failed to initialize this variable to null in the constructor, and there is an uncommonly-traveled 
        code path which does not set this variable to something good. In the destructor, we check for null, 
        and deref if non-null. Obvious "BOOM" potential. Fixed.

2005-01-21  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3966311> REGRESSION (Mail): Hitting return makes space character disappear

        * khtml/editing/htmlediting.cpp:
        (khtml::InsertParagraphSeparatorCommand::doApply): When writing the code to insert a block
        in response to the return key, I did not write code to cover the case described in the bug.
        Now I have.
        * layout-tests/editing/inserting/insert-div-025-expected.txt: Added.
        * layout-tests/editing/inserting/insert-div-025.html: Added.

2005-01-20  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Ken.

        <rdar://problem/3965196> security fix for javascript: exploit missed one case (already fixed in updates)
        
        * khtml/ecma/kjs_window.cpp:
        (WindowFunc::tryCall): correct mistake in earlier fix for the following bug, caught by Adele:

2005-01-20  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Ken.

        <rdar://problem/3965466> editing needs to insert text before applying typing style
        
        * khtml/editing/htmlediting.cpp:
        (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): Remove a FIXME comment.
        (khtml::InsertLineBreakCommand::doApply): Apply style to the BR
        node after inserting it.
        (khtml::InsertTextCommand::prepareForTextInsertion): Don't try to
        apply style to any new nodes created before inserting.
        (khtml::InsertTextCommand::input): Apply style to the inserted range
        after doing the text insert.
        * khtml/editing/htmlediting.h:

        This change leads to improved results on one of the layout tests:
        
        * layout-tests/editing/inserting/insert-div-024-expected.txt:

=== Safari-180 ===

2005-01-20  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3964646> REGRESSION (179-180): Typing space at end of line makes following paragraph disappear

        * khtml/editing/htmlediting.cpp:
        (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Recent change to block placeholder removal code 
        caused this regression. The code became too aggressive in removing block placeholders, and would remove them
        from blocks other than the block containing the selection.
        * layout-tests/editing/inserting/insert-div-023-expected.txt: This file had a spurious BR element in it that
        I did not notice earlier.
        * layout-tests/editing/inserting/insert-div-024-expected.txt: Ditto.

2005-01-20  David Hyatt  <hyatt@apple.com>

        Fix for oddness on albertsons.com.  Make sure not to crash when setting/removing style properties on a node
        with no document.  The bug # is 3813900.
        
        Reviewed by john

2005-01-20  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        PARTIAL fix for following bug.  Create attributed string with fonts and links.  Still need to add attachments.
        <rdar://problem/3942606> AX: Support kAXAttributedStringForTextMarkerRangeParameterizedAttribute
        
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject textUnderElement]):
        Touched up previous checkin to match code review comments.
        
        (-[KWQAccObject value]):
        Use plainText for this one instead of obsolete attributedString.
        
        (-[KWQAccObject accessibilityParameterizedAttributeNames]):
        Advertise AXAttributedStringForTextMarkerRange.
        
        (AXAttributeStringAddFont):
        (AXAttributeStringAddElement):
        (-[KWQAccObject linkUIElementForNode:]):
        (-[KWQAccObject _addAttachmentForElement:URL:needsParagraph:usePlaceholder:]):
        (-[KWQAccObject accessibilityAttributedStringForRange:]):
        New routines to support AXAttributedStringForTextMarkerRange.
        
        (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
        Use new routines instead of obsolete attributedString.

2005-01-20  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3960196> AX Crash in DOM::Range::setStartBefore

        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::getAccObjectCache):
        Adjust when detecting cache in non-top level WebArea.
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject textUnderElement]):
        (-[KWQAccObject accessibilityAttributeValue:]):
        Make sure the elements document is the current one for the part.

2005-01-20  Darin Adler  <darin@apple.com>

        Reviewed by John.

        - fixed <rdar://problem/3922980> Mail not crashing, just quitting itself suddenly in -[WebHTMLView(MailExtras) findString:options:]

        * khtml/dom/dom2_range.cpp: (DOM::operator==): Handle null and detached ranges without raising exceptions.
        The uncaught exception would make the entire program terminate.

        - fixed <rdar://problem/3885729> attempting to modify a computed style does nothing, but should raise a DOM exception

        * kwq/DOM-CSS.mm: (-[DOMCSSStyleDeclaration setProperty:::]): Uncommented this code which I had to disable
        back in November because Mail was stumbling over it.

        - fixed <rdar://problem/3943049> focus() called during onload handler results in square text field on <input type=search>        

        * kwq/KWQLineEdit.mm: (QLineEdit::baselinePosition): Change computation so it will work even for fields
        that position their text in a way that depends on vertical size. In the case of a search field, the text
        is centered, so the old logic was broken.
        * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Set the frame size to
        something large enough to accomodate the field editor. If we start the frame at size 0,0 we run into
        AppKit trouble when it insets the frame to figure out the frame for the field editor. In the case of this
        bug this happens because we become first responder before being sized and positioned by the HTML layout code.

        - improved debugging output when using "po" from gdb with Objective-C DOM

        * kwq/DOM.mm:
        (-[DOMNode description]): Added. Includes node name ("<tr>") and node value (e.g., string for text node).
        (-[DOMRange description]): Tweaked format.

2005-01-19  Richard Williamson   <rjw@apple.com>

        After further discussion with Real we have decided to NOT include the
        additional CLSID for the real plugin.  See 3958601.

        Reviewed by Darin.

        * khtml/rendering/render_frames.cpp:
        (RenderPartObject::updateWidget):

2005-01-19  David Harrison  <harrison@apple.com>

        Reviewed by Maciej.

        <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line

        Previous checkin for this bug was missing some of the patch.  Editing snafu with multiple changes in tree.
        
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
        (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):

2005-01-19  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        * khtml/editing/htmlediting.cpp:
        (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Small, cosmetic change John and I decided
        to do on my last checkin, but I forgot to do before landing.

2005-01-19  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3959727> REGRESSION (Mail): Style not preserved on blank lines

        * khtml/editing/htmlediting.cpp:
        (khtml::CompositeEditCommand::applyStyle):
        (khtml::CompositeEditCommand::insertBlockPlaceholder): New function that unconditionally adds a block placeholder.
        (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Now returns bool based on whether
        placeholder was added or not.
        (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Now searches all the descendents of a block
        looking for a placeholder. The old code, which just looked at the last child of a node, started missing
        once block placeholders became styled (which started happening with this patch).
        (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Now handles applying typing style
        to a block placeholder at call time, rather than setting the typing style as a latent style that
        might be applied later. This is an important part of the bug fix.
        (khtml::DeleteSelectionCommand::doApply): Now uses bool return value from insertBlockPlaceholderIfNeeded()
        and passes it along to calculateStyleBeforeInsertion, so the case where a block placeholder needs to
        be styled can be detected.
        (khtml::InsertParagraphSeparatorCommand::InsertParagraphSeparatorCommand): Changed the way this class
        managed style. Before it would calculate and set typing style for the block added. This is not
        sufficient. Added blocks need to styled immediately. Some name changes to instance variables in 
        this class due to the change to accommodate this change.
        (khtml::InsertParagraphSeparatorCommand::~InsertParagraphSeparatorCommand): Name changes, as above.
        (khtml::InsertParagraphSeparatorCommand::calculateStyleBeforeInsertion): Ditto.
        (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Ditto.
        (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
        (khtml::ReplaceSelectionCommand::doApply): Improve check for testing when a placeholder
        block can be removed in its entirety after the insertion.
        * khtml/editing/htmlediting.h: Update header accordingly.
        * khtml/khtml_part.cpp:
        (KHTMLPart::selectionComputedStyle): Move position for computed style check downstream before
        doing check when the position is in an empty block (this makes sure any style on any block
        placeholder is accounted for).

        New layout tests to check bug fix.

        * layout-tests/editing/style/block-style-004-expected.txt: Added.
        * layout-tests/editing/style/block-style-004.html: Added.
        * layout-tests/editing/style/block-style-005-expected.txt: Added.
        * layout-tests/editing/style/block-style-005.html: Added.
        * layout-tests/editing/style/block-style-006-expected.txt: Added.
        * layout-tests/editing/style/block-style-006.html: Added.

        Results updated to reflect new block placeholder code.

        * layout-tests/editing/inserting/insert-div-004-expected.txt 
        * layout-tests/editing/inserting/insert-div-005-expected.txt
        * layout-tests/editing/inserting/insert-div-006-expected.txt
        * layout-tests/editing/inserting/insert-div-008-expected.txt
        * layout-tests/editing/inserting/insert-div-011-expected.txt
        * layout-tests/editing/inserting/insert-div-012-expected.txt
        * layout-tests/editing/inserting/insert-div-013-expected.txt
        * layout-tests/editing/inserting/insert-div-014-expected.txt
        * layout-tests/editing/inserting/insert-div-015-expected.txt
        * layout-tests/editing/inserting/insert-div-016-expected.txt
        * layout-tests/editing/inserting/insert-div-017-expected.txt
        * layout-tests/editing/inserting/insert-div-018-expected.txt
        * layout-tests/editing/inserting/insert-div-019-expected.txt
        * layout-tests/editing/inserting/insert-div-021-expected.txt
        * layout-tests/editing/inserting/insert-div-022-expected.txt
        * layout-tests/editing/inserting/insert-div-023-expected.txt
        * layout-tests/editing/inserting/insert-div-024-expected.txt

2005-01-19  David Hyatt  <hyatt@apple.com>

        Dont null-check the renderer before submitting, since a script can set it to display:none and still expect the
        submission to occur.  Fixes bug #3477282.
        
        Reviewed by kdecker

        (DOM::HTMLInputElementImpl::defaultEventHandler):

2005-01-18  Richard Williamson   <rjw@apple.com>
        
        Fixed several issues all arising from analysis of plugin detection code at ifilm.com:

        Fixed <rdar://problem/3958592> can't script plug-ins if plug-in is invoked with <object> element instead of <embed>
        Fixed <rdar://problem/3958597> <object> elements with IDs do not show up as named properties of the document
        Fixed <rdar://problem/3960973> DOM objects for plugin elements are not accessible
        Fixed <rdar://problem/3958601> need an additional class ID in WebCore for the Real plug-in

        We now support accessing scriptable plugin objects that are specified with <applet>, <embed>, or <object>
        tags.  Also, if any of these elements are named they can be accessed from the document or window objects.
        Finally, DOM methods are properties will be forwarded appropriately for the plugin's root scriptable object.

        Reviewed by Chris.

        * khtml/dom/html_document.cpp:
        (HTMLDocument::objects):
        * khtml/dom/html_document.h:
        * khtml/ecma/kjs_dom.cpp:
        (DOMDocumentProtoFunc::tryCall):
        (DOMElementProtoFunc::tryCall):
        (KJS::getRuntimeObject):
        * khtml/ecma/kjs_dom.h:
        * khtml/ecma/kjs_html.cpp:
        (KJS::HTMLDocument::tryGet):
        (KJS::HTMLElement::tryGet):
        (KJS::HTMLCollection::tryGet):
        (KJS::HTMLCollection::getNamedItems):
        * khtml/ecma/kjs_window.cpp:
        (Window::get):
        * khtml/html/html_miscimpl.cpp:
        (HTMLCollectionImpl::traverseNextItem):
        * khtml/html/html_miscimpl.h:
        (DOM::HTMLCollectionImpl::):
        * khtml/html/html_objectimpl.cpp:
        (HTMLAppletElementImpl::getAppletInstance):
        (HTMLObjectElementImpl::HTMLObjectElementImpl):
        (HTMLObjectElementImpl::getObjectInstance):
        * khtml/html/html_objectimpl.h:
        * khtml/rendering/render_frames.cpp:
        (RenderPartObject::updateWidget):
        * kwq/KWQKHTMLPart.h:
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::getObjectInstanceForView):

2005-01-18  David Hyatt  <hyatt@apple.com>

        Fix for 3948123, rolling over link erases nearby text.  The repaint rect check for lines was wrong whenever
        two lines overlapped.
        
        Reviewed by kocienda

        * khtml/rendering/render_flow.cpp:
        (RenderFlow::paintLines):

2005-01-18  Ken Kocienda  <kocienda@apple.com>

        Reviewed by Hyatt
        
        Fix for this bug:
        
        <rdar://problem/3960116> Focus rings paint incorrectly for contenteditable blocks in web pages

        * khtml/rendering/render_flow.cpp:
        (RenderFlow::addFocusRingRects): Fix painting of focus rings so that ring only paints around
        outermost contenteditable elements.

2005-01-18  David Harrison  <harrison@apple.com>

        Reviewed by Darin.

        <rdar://problem/3959668> accessibilityFocusedUIElement sometimes returns an ignored element; it must not

        * kwq/KWQAccObject.mm:
        (-[KWQAccObject accessibilityFocusedUIElement]):
        Return parentObjectUnignored if focused object is ignored.

2005-01-18  Ken Kocienda  <kocienda@apple.com>

        Reviewed by John

        Fix for this bug:
        
        <rdar://problem/3952877> REGRESSION (Mail): Command-left/right-arrows don't work with file attachment

        * khtml/editing/selection.cpp:
        (khtml::nodeForInlineBox): New helper function used in reimplementation of function below.
        (khtml::selectionForLine): Reimplemented using line box smarts. I originally wrote this code when
        I had a less than full understanding of line layout. I can do better now, and my new version no
        longer fails to notice attachments when doing the kind of navigation mentioned in the bug.

2005-01-17  David Harrison  <harrison@apple.com>

        Reviewed by John Sullivan.

        <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
        <rdar://problem/3949848> AX: paragraph marker routines do not work when given a paragraph end marker

        Also fixed sentence support in the same way.
        
        * kwq/KWQAccObject.mm:
        (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
        (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
        Move one position in desired direction before calling visible_units.cpp code.

2005-01-17  David Harrison  <harrison@apple.com>

        Updated expected test results.
        
        <rdar://problem/3945880> line-ending space seems not to be present

        * layout-tests/editing/deleting/delete-ws-fixup-001-expected.txt:
        * layout-tests/editing/deleting/delete-ws-fixup-002-expected.txt:
        * layout-tests/editing/inserting/insert-div-020-expected.txt:
        * layout-tests/editing/inserting/insert-div-021-expected.txt:
        * layout-tests/editing/inserting/insert-div-022-expected.txt:
        * layout-tests/editing/inserting/insert-div-023-expected.txt:
        * layout-tests/editing/inserting/insert-div-024-expected.txt:

2005-01-17  David Harrison  <harrison@apple.com>

        Reviewed by Dave Hyatt (bidi.cpp) and Darin Adler (selection.cpp).

        <rdar://problem/3945880> line-ending space seems not to be present

        * khtml/editing/selection.cpp: