16 years ago Reviewed by Darin Adler.
ap [Fri, 23 Dec 2005 08:28:44 +0000 (08:28 +0000)]
    Reviewed by Darin Adler.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6118
        Investigate not using the frameset charset as a default for frames

        * fast/encoding/frame-default-enc-expected.checksum: Added.
        * fast/encoding/frame-default-enc-expected.png: Added.
        * fast/encoding/frame-default-enc-expected.txt: Added.
        * fast/encoding/frame-default-enc.html: Added.
        * fast/encoding/resources/frame-default-enc-frame.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11746 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
mjs [Fri, 23 Dec 2005 08:19:32 +0000 (08:19 +0000)]
    Reviewed by Eric.

- move a few more methods from WebView to WebFrame.

        * WebView.subproj/WebFrame.m:
        (-[WebFrame _hasSelection]): Renamed from _frameIsSelected: and moved from
        (-[WebFrame _clearSelection]): Renamed from _deselectFrame: and moved from
        * WebView.subproj/WebFrameInternal.h:
        * WebView.subproj/WebView.m:
        (-[WebView searchFor:direction:caseSensitive:wrap:]): Update for renames.
        (-[WebView _findSelectedFrameStartingFromFrame:skippingFrame:]): Ditto.
        (-[WebView _debugCollectSelectedFramesIntoArray:startingFromFrame:]): Ditto.
        (-[WebView _selectedFrameDidChange]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11745 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago <http://bugzilla.opendarwin.org/show_bug.cgi?id=6211>
justing [Fri, 23 Dec 2005 08:05:14 +0000 (08:05 +0000)]
        fast/text/attributed-substring-from-range-lines needs LayoutTests/editing/editing.js

        * editing/input/attributed-substring-from-range-lines-expected.txt: Added.
        * editing/input/attributed-substring-from-range-lines.html: Added.
        * fast/text/attributed-substring-from-range-lines-expected.txt: Removed.
        * fast/text/attributed-substring-from-range-lines.html: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11744 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
mjs [Fri, 23 Dec 2005 08:02:17 +0000 (08:02 +0000)]
    Reviewed by Eric.

- updated for WebCore changes to make XSLT params partially work

        * fast/xsl/xslt-processer-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11743 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin Adler.
ap [Fri, 23 Dec 2005 07:55:26 +0000 (07:55 +0000)]
    Reviewed by Darin Adler.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6118
        Investigate not using the frameset charset as a default for frames

        * khtml/khtml_part.cpp (KHTMLPart::write):
        Do not use the parent frame encoding as a default for sub-frames (revert

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11742 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago - fix build problem from last commit.
mjs [Fri, 23 Dec 2005 07:35:00 +0000 (07:35 +0000)]
    - fix build problem from last commit.

        * kxmlcore/RefPtr.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11741 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
mjs [Fri, 23 Dec 2005 03:57:25 +0000 (03:57 +0000)]
    Reviewed by Eric.

        - made RefPtr<DOMStringImpl> usable as a hashtable key/value
- changed XSLTProcessorImpl to make use of this instead of a QDict.
- partially fixed broken param handling for xslt

See LayoutTests for updated test case.

        * khtml/xml/dom_stringimpl.cpp:
        * khtml/xml/dom_stringimpl.h:
        (KXMLCore::): Set things up so you can use RefPtr<DOMStringImpl> as
a hashtable key (should already be usable as a value).
        * khtml/xsl/xslt_processorimpl.cpp:
        (DOM::xsltParamArrayFromParameterMap): Updated to use a HashMap of RefPtrs instead
of QDict. Also, fixed bugs that would have kept this from ever working at all.
        (DOM::XSLTProcessorImpl::transformToString): ditto
        (DOM::XSLTProcessorImpl::setParameter): ditto
        (DOM::XSLTProcessorImpl::getParameter): ditto
        (DOM::XSLTProcessorImpl::removeParameter): ditto
        * khtml/xsl/xslt_processorimpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11740 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoJavaScriptCore:
mjs [Fri, 23 Dec 2005 01:52:43 +0000 (01:52 +0000)]

        Reviewed by Darin.

        - Make HashMap/HashSet support non-POD types

        The changes for support are relatively simple, but I also made extensive changes to
        avoid copying, so that there isn't refcount thrash when you put RefPtrs into a HashMap.

        * kxmlcore/HashTable.h:
        (KXMLCore::swap): specialize swap for pairs, to swap elements individually,
so that excess copies can be avoided.
        (KXMLCore::Mover::move): Template function to either copy or swap, used
when transferring elements from old table to new.
        (KXMLCore::IdentityHashTranslator::hash): The old "converting lookup" templates
that took two or three function parameters now take a class parameter, this is
the class used to do a normal lookup.
        (KXMLCore::IdentityHashTranslator::equal): Ditto.
        (KXMLCore::IdentityHashTranslator::translate): Ditto. Translate now takes a reference
to write into instead of returning a value to avoid redundant copies.
        (KXMLCore::HashTable::~HashTable): Use deallocateTable instead of freeing directly.
        (KXMLCore::HashTable::insert): Based on HashTranslator now instead of separate
functions. Added a FIXME about a remaining rare excess copy.
        (KXMLCore::HashTable::isEmptyBucket): Use KeyTraits directly instead of unwrapping
the key from Traits, to avoid creating and destroying pair, which copies.
        (KXMLCore::HashTable::isDeletedBucket): ditto
        (KXMLCore::HashTable::lookup): Use HashTranslator now instead of separate functions.
        (KXMLCore::HashTable::initializeBucket): Renamed from emptyBucket. Use placement new to
work right for non-POD types.
        (KXMLCore::HashTable::deleteBucket): Use assignDeleted to avoid excess copies.
        (KXMLCore::HashTable::reinsert): use Mover template to copy or swap as appropriate
        (KXMLCore::HashTable::allocateTable): Initialize every bucket if calloc won't do.
        (KXMLCore::HashTable::deallocateTable): Destruct every bucket if needed.
        (KXMLCore::HashTable::rehash): Avoid copy before reinserting, so that swap can do its magic.
        (KXMLCore::HashTable::clear): use deallocateTable instead of freeing directly.
        (KXMLCore::HashTable::HashTable): be more dumb when copying to ensure that non-POD types
work right
        * kxmlcore/HashFunctions.h:
        (KXMLCore::PointerHash): Specialize PointerHash for RefPtr
        * kxmlcore/HashMap.h:
        (KXMLCore::extractFirst): Return a reference not a full object to avoid
        (KXMLCore::HashMapTranslator::hash): Use a special translator for insertion
to defer making the pair as long as possible, thus avoiding needless copies.
        (KXMLCore::HashMapTranslator::equal): ditto
        (KXMLCore::HashMapTranslator::translate): ditto
        (KXMLCore::::inlineAdd): Shared by set and add to insert using HashMapTranslator
        (KXMLCore::::set): Use inlineAdd
        (KXMLCore::::add): Use inlineAdd
        * kxmlcore/HashMapPtrSpec.h:
        (KXMLCore::): Pass KeyTraits along
        * kxmlcore/HashSet.h:
        (KXMLCore::identityExtract): Return a reference not a full object to avoid copies.
        (KXMLCore::HashSetTranslatorAdapter::hash): Redo adapter stuff to work with
the new HashTranslator approach.
        (KXMLCore::HashSetTranslatorAdapter::equal): ditto
        (KXMLCore::HashSetTranslatorAdapter::translate): ditto
        (KXMLCore::::insert): ditto
        * kxmlcore/HashTraits.h:
        (KXMLCore::GenericHashTraits): This is intended be used as a base class for
customized traits: sensible defaults.
        (KXMLCore::): Use it a bunch
        (KXMLCore::assignDeleted): template function to allow pairs to be assigned the
deleted value w/o excess copies.
        (KXMLCore::PairHashTraits::emptyValue): Updated
        (KXMLCore::PairHashTraits::deletedValue): Updated
        (KXMLCore::PairHashTraits::assignDeletedValue): part of assignDeleted hack
        (KXMLCore::DeletedValueAssigner::assignDeletedValue): Use template magic
to either use use deletedValue or assignDeletedValue for the cases where we care.
        * kxmlcore/RefPtr.h:
        (KXMLCore::RefPtr::swap): Added swap method.
        (KXMLCore::swap): Added swap free function.
        * kjs/identifier.cpp:
        (KJS::CStringTranslator::hash): Use new HashTranslator class approach to
alternate type based insertion.
        (KJS::CStringTranslator::equal): ditto
        (KJS::CStringTranslator::translate): ditto
        (KJS::Identifier::add): ditto
        (KJS::UCharBufferTranslator::hash): ditto
        (KJS::UCharBufferTranslator::equal): ditto
        (KJS::UCharBufferTranslator::translate): ditto

- irrelevant change:

        * kjs/array_object.cpp:
        (ArrayProtoFunc::callAsFunction): Removed a stray space.


        Reviewed by Darin.

        - update for new HashTranslator stuff

        * khtml/xml/dom_atomicstring.cpp:
        * khtml/xml/dom_qname.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11739 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin.
adele [Fri, 23 Dec 2005 01:13:50 +0000 (01:13 +0000)]
    Reviewed by Darin.

        Cleaned up a previous checkin by defining global const defaultForm.

        * khtml/xml/dom_docimpl.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11738 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
adele [Thu, 22 Dec 2005 23:06:21 +0000 (23:06 +0000)]

        Reviewed by Tim Hatcher.

        Tests for <rdar://problem/4387433> Seed: Radio buttons behave incorrectly in Gmail settings

        * fast/forms/radio_checked_dynamic-expected.checksum: Added.
        * fast/forms/radio_checked_dynamic-expected.png: Added.
        * fast/forms/radio_checked_dynamic-expected.txt: Added.
        * fast/forms/radio_checked_dynamic.html: Added.


        Reviewed by Tim Hatcher.

        Fixed <rdar://problem/4387433> Seed: Radio buttons behave incorrectly in Gmail settings

        * khtml/html/html_formimpl.cpp:
        (DOM::HTMLFormElementImpl::registerFormElement): update radio button hash map when moving form elements around.
        (DOM::HTMLGenericFormElementImpl::insertedIntoTree): ditto.
        * khtml/html/html_formimpl.h:
        (DOM::HTMLGenericFormElementImpl::isRadioButton): Added.
        (DOM::HTMLInputElementImpl::isRadioButton): Added.
        * khtml/xml/dom_docimpl.cpp:
        Added comment explaining that we use 1 for the default form.  We can't use a null pointer as a key for the hash map.
        (DocumentImpl::checkedRadioButtonForGroup): Added conversion of null form pointer to 1.
        (DocumentImpl::removeRadioButtonGroup): ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11736 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago2005-12-22 Anders Carlsson <andersca@mac.com>
andersca [Thu, 22 Dec 2005 21:10:54 +0000 (21:10 +0000)]
2005-12-22  Anders Carlsson  <andersca@mac.com>

        Reviewed by Eric.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6196
        Would like to be able to define prototypes in headers

        * khtml/ecma/XSLTProcessor.cpp:
        * khtml/ecma/domparser.cpp:
        * khtml/ecma/kjs_css.cpp:
        * khtml/ecma/kjs_dom.cpp:
        * khtml/ecma/kjs_events.cpp:
        * khtml/ecma/kjs_html.cpp:
        * khtml/ecma/kjs_range.cpp:
        * khtml/ecma/kjs_traversal.cpp:
        * khtml/ecma/kjs_views.cpp:
        * khtml/ecma/xmlhttprequest.cpp:
        * khtml/ecma/xmlserializer.cpp:
        Update for changes to JSC.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11733 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoRemove conflict markers
andersca [Thu, 22 Dec 2005 21:08:59 +0000 (21:08 +0000)]
Remove conflict markers

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11732 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago2005-12-22 Anders Carlsson <andersca@mac.com>
andersca [Thu, 22 Dec 2005 21:07:37 +0000 (21:07 +0000)]
2005-12-22  Anders Carlsson  <andersca@mac.com>

        Reviewed by Eric and Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6196
        Would like to be able to define prototypes in headers

        * kjs/lookup.h:
        Also, namespace all macros by prefixing them with KJS_.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11731 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
darin [Thu, 22 Dec 2005 16:53:57 +0000 (16:53 +0000)]
    Reviewed by Eric.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6193
          remove some unused KWQ code

        * kcanvas/KCanvasCreator.cpp: (KCanvasCreator::self): Don't use static deleter for
        two reasons: (1) We don't want any globals with constructors in any of our frameworks
        because they slow down framework load time. (2) There's no need to destroy this object at
        process termination time -- it's extra work with no benefit. Because of both these
        reasons, our KStaticDeleter implementation was basically a no-op.
        * ksvg2/svg/SVGDOMImplementationImpl.cpp: (SVGDOMImplementationImpl::self): Ditto.
        * khtml/xml/dom_docimpl.cpp: (DocumentImpl::setDocumentChanged): Ditto.

        * khtml/html/html_canvasimpl.cpp: Removed include of kstringhandler.h.
        * khtml/html/html_imageimpl.cpp: Ditto.

        * khtml/khtmlpart_p.h: Removed something that was in an "APPLE_CANGES" ifdef.

        * ForwardingHeaders/kstaticdeleter.h: Removed.
        * ForwardingHeaders/kstringhandler.h: Removed.
        * kwq/KWQKStaticDeleter.h: Removed.
        * kwq/KWQKStringHandler.h: Removed.
        * kwq/KWQKStringHandler.mm: Removed.

        * WebCore.xcodeproj/project.pbxproj: Removed files.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11729 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
darin [Thu, 22 Dec 2005 16:50:09 +0000 (16:50 +0000)]

        Reviewed by Eric.

        - test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6192
          add support for non-standard &nsup; entity (implemented in other browsers)

        * fast/parser/nsup-entity-expected.txt: Added.
        * fast/parser/nsup-entity.html: Added.


        Reviewed by Eric.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6192
          add support for non-standard &nsup; entity (implemented in other browsers)

        * khtml/html/kentities.gperf: Add "nsup" to table.
        * khtml/html/htmltokenizer.cpp: Removed old workaround for inlining issue that no
        longer seems to be necessary; also touches the file which is important because
        Xcode doesn't seem to understand the dependency on kentities.gperf and the need
        to recompile.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11728 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Maciej.
darin [Thu, 22 Dec 2005 16:48:08 +0000 (16:48 +0000)]
    Reviewed by Maciej.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6191
          RefPtr/PassRefPtr have a leak issue, operator== issues

        * kxmlcore/PassRefPtr.h:
        (KXMLCore::PassRefPtr::PassRefPtr): Remove non-template constructor that takes RefPtr
        since the constructor template that takes RefPtr should be sufficient. Add a constructor
        template that takes PassRefPtr&.
        (KXMLCore::PassRefPtr::adopt): Use PassRefPtr_Ref to avoid setting pointer first to
        0 and then to the pointer.
        (KXMLCore::PassRefPtr::operator=): Added template versions that take PassRefPtr& and
        RefPtr parameters.
        (KXMLCore::PassRefPtr::operator PassRefPtr<U>): Changed to fix leak -- old version
        would release and then ref.
        (KXMLCore::operator==): Make templates have two parameters so you can mix types.
        Also remove unneeded const in raw pointer versions.
        (KXMLCore::operator!=): Ditto.

        * kxmlcore/RefPtr.h:
        (KXMLCore::RefPtr::RefPtr): Add constructor template that takes PassRefPtr.
        (KXMLCore::RefPtr::operator=): Add assignment operator templates that take
        RefPtr and PassRefPtr.
        (KXMLCore::operator==): Make templates have two parameters so you can mix types.
        Also remove unneeded const in raw pointer versions.
        (KXMLCore::operator!=): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
darin [Thu, 22 Dec 2005 16:46:39 +0000 (16:46 +0000)]
    Reviewed by Eric.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6167
          RenderStyle default constructor should initialize its members for speed

        * khtml/rendering/DataRef.h: (khtml::DataRef::operator=): Eliminate an extra
        branch by doing ref before deref instead of == check.

        * khtml/rendering/render_style.cpp:
        (khtml::initDefaultStyle): Added. Function to initialize the default style
        for use in constructor.
        (khtml::RenderStyle::RenderStyle): Changed constructor to initalize all the
        members with constructor syntax instead of using assignment on all of them.
        * khtml/rendering/render_style.h: Removed static data member _default --
        it's now a file scope global instead.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11726 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin Adler.
ap [Thu, 22 Dec 2005 07:39:45 +0000 (07:39 +0000)]
    Reviewed by Darin Adler.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5846
          cvs-create-patch --include produces incorrect paths

        * Scripts/cvs-create-patch: Handle newly-added directories using chdir.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11725 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6184
darin [Thu, 22 Dec 2005 06:55:50 +0000 (06:55 +0000)]
    - fix bugzilla.opendarwin.org/show_bug.cgi?id=6184
          fast/text/attributed-substring-from-range* failing

        * fast/text/attributed-substring-from-range-expected.txt: Updated.
        * fast/text/attributed-substring-from-range-lines-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11724 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Justin.
darin [Thu, 22 Dec 2005 04:11:39 +0000 (04:11 +0000)]
    Reviewed by Justin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6142
          intermittent failures in some paste tests

        * khtml/editing/apply_style_command.cpp:
        (khtml::ApplyStyleCommand::applyBlockStyle): Call new updateLayout member function.
        (khtml::ApplyStyleCommand::applyInlineStyle): Ditto.
        (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAtBoundaries): Ditto.
        * khtml/editing/composite_edit_command.cpp:
        (khtml::CompositeEditCommand::addBlockPlaceholderIfNeeded): Ditto.
        (khtml::CompositeEditCommand::findBlockPlaceholder): Ditto.
        (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): Ditto.
        * khtml/editing/delete_selection_command.cpp:
        (khtml::DeleteSelectionCommand::fixupWhitespace): Ditto.
        (khtml::DeleteSelectionCommand::moveNodesAfterNode): Ditto.
        * khtml/editing/edit_command.cpp:
        (khtml::EditCommand::apply): Ditto.
        (khtml::EditCommand::unapply): Ditto.
        (khtml::EditCommand::reapply): Ditto.
        (khtml::EditCommand::updateLayout): Added. Calls updateLayoutIgnorePendingStylesheets
        on the document.
        * khtml/editing/edit_command.h: Added updateLayout member function.
        * khtml/editing/insert_line_break_command.cpp:
        (khtml::InsertLineBreakCommand::doApply): Call new updateLayout member function.
        * khtml/editing/insert_paragraph_separator_command.cpp:
        (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
        * khtml/editing/jsediting.cpp:
        (DOM::JSEditor::execCommand): Call updateLayoutIgnorePendingStylesheets instead of updateLayout.
        (DOM::JSEditor::queryCommandEnabled): Ditto.
        (DOM::JSEditor::queryCommandIndeterm): Ditto.
        (DOM::JSEditor::queryCommandState): Ditto.
        (DOM::JSEditor::queryCommandValue): Ditto.
        * khtml/editing/markup.cpp: (khtml::createMarkup): Ditto.
        * khtml/editing/replace_selection_command.cpp:
        (khtml::ReplacementFragment::insertFragmentForTestRendering): Ditto.
        (khtml::ReplaceSelectionCommand::fixupNodeStyles): Call new updateLayout member function.
        (khtml::ReplacementFragment::computeStylesUsingTestRendering): Call
        updateLayoutIgnorePendingStylesheets instead of updateLayout.
        (khtml::ReplaceSelectionCommand::doApply): Call new updateLayout member function.
        (khtml::ReplaceSelectionCommand::removeLinePlaceholderIfNeeded): Ditto.
        (khtml::ReplaceSelectionCommand::completeHTMLReplacement): Ditto.
        * khtml/editing/visible_units.cpp:
        (khtml::previousLinePosition): Call updateLayoutIgnorePendingStylesheets instead of updateLayout.
        (khtml::nextLinePosition): Ditto.
        * khtml/html/html_elementimpl.cpp:
        (HTMLElementImpl::innerText): Ditto.
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge setSelectedDOMRange:affinity:closeTyping:]): Ditto.
        (-[WebCoreBridge smartDeleteRangeForProposedRange:]): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11723 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Geoff.
darin [Thu, 22 Dec 2005 04:06:46 +0000 (04:06 +0000)]
    Reviewed by Geoff.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6177
          move event code from JavaScript binding into DOM implementation

        * khtml/ecma/kjs_events.cpp:
        (KJS::DOMMouseEvent::getValueProperty): Change everything to just call through instead of doing
        the work here.
        (KJS::DOMWheelEvent::getValueProperty): Ditto.

        * khtml/xml/dom2_eventsimpl.h:
        * khtml/xml/dom2_eventsimpl.cpp:
        (DOM::MouseRelatedEventImpl::MouseRelatedEventImpl): Initialize new m_pageX, m_pageY, m_layerX,
        m_layerY, m_offsetX, and m_offsetY.
        (DOM::MouseRelatedEventImpl::computePositions): Renamed from computeLayerPos since it handles
        page position andoffset position too now.
        (DOM::MouseRelatedEventImpl::pageX): Added. Not in header since it's virtual.
        (DOM::MouseRelatedEventImpl::pageY): Ditto.
        (DOM::MouseRelatedEventImpl::x): Added, with FIXME since it should change eventually.
        (DOM::MouseRelatedEventImpl::y): Ditto.
        (DOM::MouseEventImpl::initMouseEvent): Call computePositions instead of computeLayerPos.
        (DOM::MouseEventImpl::toElement): Added.
        (DOM::MouseEventImpl::fromElement): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11722 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoJavaScriptCore:
thatcher [Thu, 22 Dec 2005 01:40:58 +0000 (01:40 +0000)]

        * JavaScriptCore.xcodeproj/project.pbxproj:
          Set tab width to 8, indent width to 4 and uses tabs to false per file.


        * WebCore.xcodeproj/project.pbxproj:
          Set tab width to 8, indent width to 4 and uses tabs to false per file.


        * WebKit.xcodeproj/project.pbxproj:
          Set tab width to 8, indent width to 4 and uses tabs to false per file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11721 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Tim Omernick and Darin Adler.
sullivan [Thu, 22 Dec 2005 01:25:13 +0000 (01:25 +0000)]
    Reviewed by Tim Omernick and Darin Adler.

        - fixed HiDPI problem with forms auto-fill menu width

        No test cases added; this code is used for "chrome" only.

        * kwq/DOMHTML.mm:
        (-[DOMHTMLInputElement _rectOnScreen]):
        convert entire rect to window coordinates, not just origin

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11720 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin.
ggaren [Thu, 22 Dec 2005 00:55:34 +0000 (00:55 +0000)]
    Reviewed by Darin.

        Removed evil hack for determining if a type is an integer, replaced
        with template metaprogramming.

        * JavaScriptCore.xcodeproj/project.pbxproj: Set tab size to 2 for
        * kjs/testkjs.cpp:
        (main): Inserted asserts to test IsInteger. FIXME: Move these to
        KXMLCore unit tests directory when we create one.
        * kxmlcore/HashTraits.h:
        (KXMLCore::): Added IsInteger class for querying types.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11719 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Justin).
harrison [Wed, 21 Dec 2005 22:51:12 +0000 (22:51 +0000)]
    Reviewed by Justin).

        <rdar://problem/4039777> Pasting particular snippet of HTML containing list items and a link creates too many list items
        - Fixed paste crash by making calling RenderBox::deleteLineBoxWrapper() from RenderObject::remove(),
        so that the connection is broken before the InlineBox's parent gets deleted.
        - Fixed overzealous style changes when setting the style on a specific range of elements.  These specific
        ranges are derived programmatically, e.g. as a part of pasting, so they must be respected exactly rather
        than modified by converting the endpoints to VisiblePositions.

        Test cases coming soon.

        * khtml/editing/apply_style_command.cpp:
        * khtml/editing/apply_style_command.h:
        * khtml/editing/composite_edit_command.cpp:
        * khtml/editing/composite_edit_command.h:
        * khtml/editing/insert_line_break_command.cpp:
        * khtml/editing/replace_selection_command.cpp:
        * khtml/rendering/render_box.cpp:
        * khtml/rendering/render_box.h:
        * khtml/rendering/render_list.cpp:
        * khtml/rendering/render_object.cpp:
        * khtml/rendering/render_object.h:
        * khtml/rendering/render_replaced.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11718 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
adele [Wed, 21 Dec 2005 16:50:59 +0000 (16:50 +0000)]

        Reviewed by Darin.

        Updating results for:
        REGRESSION: Page scroll position jumps when clicking on word in editable div

        These results all changed because we now select all for contenteditable elements when focus() is called on them.

        * editing/selection/focus_editable_html-expected.checksum:
        * editing/selection/focus_editable_html-expected.png:
        * editing/selection/focus_editable_html-expected.txt:
        * fast/dom/blur-contenteditable-expected.txt:
        * fast/dom/focus-contenteditable-expected.checksum:
        * fast/dom/focus-contenteditable-expected.png:
        * fast/dom/focus-contenteditable-expected.txt:
        * fast/overflow/overflow-focus-ring-expected.checksum:
        * fast/overflow/overflow-focus-ring-expected.png:
        * fast/overflow/overflow-focus-ring-expected.txt:


        Reviewed by Darin.

        Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=5911
        REGRESSION: Page scroll position jumps when clicking on word in editable div

        I moved the scrolling code out of setFocusNode, and consolidated some of the focus code to scroll when necessary.

        * khtml/html/html_formimpl.cpp: Removed HTMLButtonElementImpl::blur, HTMLButtonElementImpl::focus, HTMLInputElementImpl::blur, HTMLInputElementImpl::focus.
        New code in ElementImpl::focus will now handle these cases.  This allows tabbing through the elements to go through the same code path as calling focus() on an element.
        Before, focus() would scroll to reveal for any form elements that had a RenderWidget, but wouldn't scroll for anchor elements, or any of the new form elements.
        Now the behavior will be more consistent.
        (DOM::HTMLLabelElementImpl::focus): calls ElementImpl::focus.
        (DOM::HTMLLegendElementImpl::focus): ditto.
        * khtml/html/html_formimpl.h:
        * khtml/khtml_part.cpp:
        (KHTMLPart::selectAll): calls new function selectContentsOfNode
        (KHTMLPart::selectContentsOfNode): factored out code to selectAll for a particular node- which is useful for contenteditable elements.
        * khtml/khtml_part.h: added selectContentsOfNode
        * khtml/xml/dom_docimpl.cpp:
        (DocumentImpl::setFocusNode): removed scrolling code.
        * khtml/xml/dom_elementimpl.cpp:
        (ElementImpl::focus): Calls updateLayout in case focus() is called before there's a renderer.
        Makes a selection for editable elements (right now we select all, but this will change).
        Doesn't scroll if the renderer is a RenderWidget, since that is handled when the view becomes first responder.  This will go away when we convert the rest of our form elements.
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::revealSelection): Made this more like centerSelectionInVisibleArea  where we get the right rectangle if the selection is a caret.
        (KWQKHTMLPart::nextKeyViewInFrame): call ElementImpl::focus() for the node.  This will set the selection too, which used to be done here.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11711 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: none
eseidel [Wed, 21 Dec 2005 10:53:16 +0000 (10:53 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: none needed.
        Fixed ChangeLog typo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11710 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: none
eseidel [Wed, 21 Dec 2005 10:44:40 +0000 (10:44 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: mjs
        Beef up our XSLT processor tests to include a test which passes
        a parameter into the transform.  I also took this opportunity
        to clean up the output a bit, and in the process found another
        bug and file: http://bugzilla.opendarwin.org/show_bug.cgi?id=6180

        * fast/xsl/resources/xslt-param.xml: Added.
        * fast/xsl/resources/xslt-param.xsl: Added.
        * fast/xsl/xslt-processer-expected.txt:
        * fast/xsl/xslt-processer.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11709 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by justin
justing [Wed, 21 Dec 2005 02:06:18 +0000 (02:06 +0000)]
    Reviewed by justin

        New layout tests for:
        -[WebHTMLView firstRectForCharacterRange:] is using _selectedRange instead of the given range if no marked text

        * editing/input/firstrectforcharacterrange-plain-expected.txt: Added.
        * editing/input/firstrectforcharacterrange-plain.html: Added.
        * editing/input/firstrectforcharacterrange-styled-expected.txt: Added.
        * editing/input/firstrectforcharacterrange-styled.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11708 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by justin
justing [Wed, 21 Dec 2005 02:05:12 +0000 (02:05 +0000)]
    Reviewed by justin

        -[WebHTMLView firstRectForCharacterRange:] is using _selectedRange instead of the given range if no marked text

        Added layout tests:
        * editing/input/firstrectforcharacterrange-styled
        * editing/input/firstrectforcharacterrange-plain

        * khtml/editing/visible_text.cpp:
        Return null if the range isn't found, instead of a startless/endless
        range.  Set the end if the requested location+length is out of bounds.
        * kwq/WebCoreBridge.mm:
        (-[WebCoreBridge convertToDOMRange:]): Handle larged unsigned values
        before calling rangeWithLocationAndLength, which expects signed ints.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11707 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by justin
justing [Wed, 21 Dec 2005 02:04:38 +0000 (02:04 +0000)]
    Reviewed by justin

        -[WebHTMLView firstRectForCharacterRange:] is using _selectedRange instead of the given range if no marked text

        * WebView.subproj/WebHTMLView.m:
        (-[WebHTMLView firstRectForCharacterRange:]):
        Handle some large unsigned values the way NSTextView does.
        Actually use the range passed in instead of _selectedRange,
        use of _selectedRange was a workaround that is no longer necessary.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11706 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin Adler.
sullivan [Wed, 21 Dec 2005 01:16:47 +0000 (01:16 +0000)]
    Reviewed by Darin Adler.

        - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=6146
          (REGRESSION: Bold font used for Google search result pages is too thick)

        This is a problem with a particular font that was installed by Microsoft Office X. Though the
        font and/or lower levels of font-handling code in the system are buggy, this bad symptom will
        occur for users of Safari and other WebKit clients who happen to have one of these bad fonts.
        This adds a workaround to avoid the problem.

        * WebCoreSupport.subproj/WebTextRendererFactory.m:
        (-[WebTextRendererFactory fontWithFamilies:traits:size:]):
        When we're going to synthesize bold or italic, yet the font we looked up was apparently a
        match for the traits, try to look up a font that without the to-be-synthesized traits.
        This way, instead of applying synthetic bold over Arial Bold, we'll apply synthetic
        bold over Arial Regular, which is uglier than just using Arial Bold, but far less ugly
        than using Arial Bold with synthetic bold too.

 2005-12-16  Justin Garcia  <justin.garcia@apple.com>

         <rdar://problem/4103393> Frequent Safari crash on lexisnexus.com (khtml::Selection::xPosForVerticalArrowNavigation)
         <rdar://problem/4330451> CrashTracer: [REGRESSION] 2235 crashes in Safari at com.apple.WebCore: khtml::Selection::xPosForVerticalArrowNavigation const  436

         Reviewed by darin

         WebCore will crash when a selection that starts or ends in a node
         that has been removed from the document is modify()d.  This can occur:
         (1) in non-editable regions (4103393 and 4330451), (2) in editable
         regions (4383146) as the result of arbitrary DOM operations, and (3) in
         Mail (4099739) as the result of an editing operation that sets a
         bad ending selection.

         Crashes of type (1) can occur when the user uses the arrow keys
         to interact with a web app, or when the user tries to use
         command-shift-arrow to switch tabs (this is a depricated
         combo that will work if no one else responds to it). The easiest
         way to fix these crashes is to disallow editing'ish selection changes
         like moveDown:, selectWord:, pageDown:, etc, when the selection
         is in a non-editable region.

         Crashes of type (2) will require a more complicated fix (but occur
         much less often than type (1)).  Crashes of type (3) must be
         fixed by tracking down the editing operation that sets bad selections.

         Added a layout-test:
         * editing/selection/selection-actions.html

         * WebView.subproj/WebHTMLView.m:
         (-[WebHTMLView _canAlterCurrentSelection]):
         (-[WebHTMLView _alterCurrentSelection:direction:granularity:]):
         (-[WebHTMLView _alterCurrentSelection:verticalDistance:]):
         (-[WebHTMLView _expandSelectionToGranularity:]):
         * WebView.subproj/WebHTMLViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11701 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin.
adele [Wed, 21 Dec 2005 00:48:54 +0000 (00:48 +0000)]
    Reviewed by Darin.

        Fix for <rdar://problem/4387630> REGRESSION: <select> element's onClick event doesn't fire @ bugweb.apple.com

        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::passSubframeEventToSubframe): only pass mouse down for khtmlviews.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11698 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by darin
justing [Wed, 21 Dec 2005 00:23:52 +0000 (00:23 +0000)]
    Reviewed by darin

        Layout test for:
        <rdar://problem/4103393> Frequent Safari crash on lexisnexus.com (khtml::Selection::xPosForVerticalArrowNavigation)
         <rdar://problem/4330451> CrashTracer: [REGRESSION] 2235 crashes in Safari at com.apple.WebCore: khtml::Selection::xPosForVerticalArrowNavigation const  436

        * editing/selection/selection-actions-expected.checksum: Added.
        * editing/selection/selection-actions-expected.png: Added.
        * editing/selection/selection-actions-expected.txt: Added.
        * editing/selection/selection-actions.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago 2005-12-16 Justin Garcia <justin.garcia@apple.com>
justing [Wed, 21 Dec 2005 00:23:08 +0000 (00:23 +0000)]
 2005-12-16  Justin Garcia  <justin.garcia@apple.com>

         <rdar://problem/4103393> Frequent Safari crash on lexisnexus.com (khtml::Selection::xPosForVerticalArrowNavigation)
         <rdar://problem/4330451> CrashTracer: [REGRESSION] 2235 crashes in Safari at com.apple.WebCore: khtml::Selection::xPosForVerticalArrowNavigation const  436

         Reviewed by darin

         WebCore will crash when a selection that starts or ends in a node
         that has been removed from the document is modify()d.  This can occur:
         (1) in non-editable regions (4103393 and 4330451), (2) in editable
         regions (4383146) as the result of arbitrary DOM operations, and (3) in
         Mail (4099739) as the result of an editing operation that sets a
         bad ending selection.

         Crashes of type (1) can occur when the user uses the arrow keys
         to interact with a web app, or when the user tries to use
         command-shift-arrow to switch tabs (this is a depricated
         combo that will work if no one else responds to it). The easiest
         way to fix these crashes is to disallow editing'ish selection changes
         like moveDown:, selectWord:, pageDown:, etc, when the selection
         is in a non-editable region.

         Crashes of type (2) will require a more complicated fix (but occur
         much less often than type (1)).  Crashes of type (3) must be
         fixed by tracking down the editing operation that sets bad selections.

         Added a layout-test:
         * editing/selection/selection-actions.html

         * WebView.subproj/WebHTMLView.m:
         (-[WebHTMLView _canAlterCurrentSelection]):
         (-[WebHTMLView _alterCurrentSelection:direction:granularity:]):
         (-[WebHTMLView _alterCurrentSelection:verticalDistance:]):
         (-[WebHTMLView _expandSelectionToGranularity:]):
         * WebView.subproj/WebHTMLViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by: eric, thatcher
justing [Wed, 21 Dec 2005 00:00:21 +0000 (00:00 +0000)]
Reviewed by: eric, thatcher

        <rdar://problem/4172984> KWQExceptions needs to use @try/@catch instead of relying on NSException.h internals

        Needed to undef try/catch because of 4333439.  Moved the
        declarations of variables that are returned from within a @try
        block outside the @try block (because of "might be clobbered by
        a longjmp or vfork" warnings).  Moved some return statements
        inside the @try block to fix volatilization errors with gcc4.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by eric, thatcher
justing [Tue, 20 Dec 2005 23:05:44 +0000 (23:05 +0000)]
    Reviewed by eric, thatcher

        <rdar://problem/4172984> KWQExceptions needs to use @try/@catch instead of relying on NSException.h internals

        Needed to undef try/catch because of 4333439.  Moved the
        declarations of variables that are returned from within a @try
        block outside the @try block (because of "might be clobbered by
        a longjmp or vfork" warnings).  Moved some return statements
        inside the @try block to fix volatilization errors with gcc4.

        * WebCore.xcodeproj/project.pbxproj:
        * WebCorePrefix.h:
        * kcanvas/device/quartz/KCanvasFilterQuartz.mm:
        * kwq/KWQExceptions.h:
        * kwq/KWQExceptions.mm:
        * kwq/KWQFileButton.mm:
        * kwq/KWQKCursor.mm:
        (+[NSCursor _WebCore_cursorWithName:hotSpot:_WebCore_cursorWithName:hotSpot:]):
        * kwq/KWQKHTMLPart.mm:
        * kwq/KWQKHTMLPartBrowserExtension.mm:
        * kwq/KWQLineEdit.mm:
        * kwq/KWQListBox.mm:
        * kwq/KWQLoader.mm:
        * kwq/KWQWidget.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago <rdar://problem/4294417> Cannot un-italicize some text after triple clicking it
harrison [Tue, 20 Dec 2005 22:31:44 +0000 (22:31 +0000)]
    <rdar://problem/4294417> Cannot un-italicize some text after triple clicking it

        Reviewed by Justin.

        * khtml/editing/apply_style_command.cpp:
        Removed check for tab span because ApplyStyleCommand::removeCSSStyle() already makes the same check.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by mitz
justing [Tue, 20 Dec 2005 21:36:40 +0000 (21:36 +0000)]
    Reviewed by mitz

        Fixed more uninitialized variable warnings, and
        removed an extra semicolon.

        * WebCoreSupport.subproj/WebTextRenderer.m:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago <rdar://problem/4387270> editing/deleting/delete-3800834-fix failing
justing [Tue, 20 Dec 2005 21:18:28 +0000 (21:18 +0000)]
    <rdar://problem/4387270> editing/deleting/delete-3800834-fix failing
        <http://bugzilla.opendarwin.org/show_bug.cgi?id=6160> REGRESSION: Crash when running editing/deleting/delete-3800834-fix.html
        <http://bugzilla.opendarwin.org/show_bug.cgi?id=6161> REGRESSION: crash when pressing tab in editable WebHTMLView

        Reviewed by darin

        Some callers call setAttribute on a floating element.  So, using
        a RefPtr for the element inside addAttribute can destroy

        * khtml/xml/dom_elementimpl.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoFix bug numbers in last two entries.
mjs [Tue, 20 Dec 2005 20:48:55 +0000 (20:48 +0000)]
Fix bug numbers in last two entries.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoAdded bugzilla link.
mjs [Tue, 20 Dec 2005 20:42:41 +0000 (20:42 +0000)]
Added bugzilla link.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin.
mjs [Tue, 20 Dec 2005 20:41:55 +0000 (20:41 +0000)]
    Reviewed by Darin.

- push more of frame lookup and management from WebView to
WebFrame, this is in preparation for shifting this to WebCore

        * WebCoreSupport.subproj/WebBridge.m:
        (-[WebBridge closeWindowSoon]): Adjust for change to WebFrameNamespaces
        (-[WebBridge runModal]): ditto
        * WebView.subproj/WebControllerSets.h:
        * WebView.subproj/WebControllerSets.m:
        (+[WebFrameNamespaces addFrame:toNamespace:]): This now operates in terms
of WebFrames (expected to be the main frame) not WebViews.
        (+[WebFrameNamespaces framesInNamespace:]): Ditto.
        * WebView.subproj/WebFrame.m:
        (-[WebFrame _setFrameNamespace:]): Set self, not WebView.
        (-[WebFrame _shouldAllowAccessFrom:]): Moved this code above use
to avoid prototyping the method.
        (-[WebFrame _descendantFrameNamed:sourceFrame:]): Ditto.
        (-[WebFrame _frameInAnyWindowNamed:sourceFrame:]): Copied logic
over from WebView.
        (-[WebFrame findFrameNamed:]): Do it all here, don't call WebView.
        * WebView.subproj/WebFramePrivate.h:
        * WebView.subproj/WebView.m:
        * WebView.subproj/WebViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin.
mjs [Tue, 20 Dec 2005 20:40:43 +0000 (20:40 +0000)]
    Reviewed by Darin.

        - fixed leak in createAttributeMap

        * khtml/xml/dom_elementimpl.cpp:
        (StyledElementImpl::createAttributeMap): the new attribute map
is going in a RefPtr, don't also ref it manually.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoRemove accidentally committed change (which I am about to recommit, so no ChangeLog
mjs [Tue, 20 Dec 2005 20:27:15 +0000 (20:27 +0000)]
Remove accidentally committed change (which I am about to recommit, so no ChangeLog
comment for the revert).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin.
mjs [Tue, 20 Dec 2005 20:18:54 +0000 (20:18 +0000)]
    Reviewed by Darin.

- made ALWAYS_INLINE declare things inline as well as __attribute__((always_inline))

        * kxmlcore/AlwaysInline.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoJavaScriptCore:
mjs [Tue, 20 Dec 2005 20:12:50 +0000 (20:12 +0000)]

        Reviewed by Darin.

- fixed a leak in the assignment operator from PassRefPtr to RefPtr

        * kxmlcore/RefPtr.h:

- fix problem with PassRefPtr that darin spotted - it lacked a copy constructor
and therefore was using the default one, which can lead to excess derefs

I fixed this by adding a copy constructor from non-const
reference, and by adding a template pass() function that you have
to use when raw pointer or RefPtr are passed where PassRefPtr is

        * kjs/identifier.cpp:
        (KJS::Identifier::add): Changed to have PassRefPtr return type and
pass() the results.
        * kjs/identifier.h:
        * kjs/property_map.cpp:
        (KJS::PropertyMap::addSparseArrayPropertiesToReferenceList): Use pass()
where required.
        * kjs/ustring.cpp:
        (KJS::UString::UString): Use pass() as needed.
        (KJS::UString::append): ditto
        (KJS::UString::substr): ditto
        * kjs/ustring.h:
        (KJS::UString::UString): Use initializer instead of assignment
        * kxmlcore/PassRefPtr.h:
        (KXMLCore::PassRefPtr::PassRefPtr): Added copy constructor
        (KXMLCore::pass): new template function to make it convenient to pass
a PassRefPtr


        Reviewed by Darin.

        - change an assignment to a contructor declaration to build with PassRefPtr
leak fix changes

        * ksvg2/svg/SVGTransformableImpl.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11684 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by adele.
ggaren [Tue, 20 Dec 2005 18:47:15 +0000 (18:47 +0000)]
    Reviewed by adele.

        Fixed build failure due to missing 'b's in my last checkin.

        * WebCoreSupport.subproj/WebBridge.m:
        (-[WebBridge isStatusbarVisible]): changed 'B' to 'b'
        (-[WebBridge setStatusbarVisible:]): ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11683 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoJavaScriptCore:
ggaren [Tue, 20 Dec 2005 18:34:31 +0000 (18:34 +0000)]

        Reviewed by Maciej.

        Fixed <rdar://problem/4370397> Missing return statement in

        JSMethodNameToObjcMethodName had a check for a name being too long, but
        the check was missing a return statement.

        A lot of this code was confusing and some of it was wrong, so I fixed
        it up, added some asserts to catch this type of bug in the future,
        changed some comments, and renamed some variables.

        The two advantages of the new algorithm are (1) It makes writing past
        the end of the buffer virtually impossible because the test on the main
        loop is "while (not past end of buffer)" and (2) It's twice as fast
        because it doesn't call strlen. (There's no need to call strlen when
        we're walking the string ourselves.)

        methodsNamed also supports arbitrary-length method names now. Just in
        case the AppKit folks start getting REALLY verbose...

        * bindings/objc/objc_class.mm:
        * bindings/objc/objc_utility.h:
        * bindings/objc/objc_utility.mm:


        Layout test for <rdar://problem/4370397> Missing return statement in

        * fast/js/objc-big-method-name-expected.txt: Added.
        * fast/js/objc-big-method-name.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin, committed by Adele.
adele [Tue, 20 Dec 2005 18:29:13 +0000 (18:29 +0000)]
    Reviewed by Darin, committed by Adele.

        - fix deployment builds broken by the ATSUI multiple renderers patch

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (createATSULayoutParameters): Assign initial values, which will never be used,
        to substituteRenderer and firstSmallCap, to avoid uninitialized variable warnings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoWebCore:
ggaren [Tue, 20 Dec 2005 17:06:04 +0000 (17:06 +0000)]

        Reviewed by John.

        Fixed <rdar://problem/4310363> JavaScript window.open: Height is 1
        pixel short, and related bugs.

        There were a few bugs here.
        (1) Our code took size arguments and applied them to the window's
            content rect. That's incorrect. The Rhino book says the arguments
            should apply to the WebView. Other things that occupy the content
            rect include the tab bar, the status bar, and the 1 pixel border
            between brushed metal and document. All of these used to impinge
            on the web page's display area.

            The fix is to calculate sizing based on the WebView instead of
            the content rect. This means that the webViewContentRect and
            setContentRect delegate methods are obsolete and no longer called
            by any of our code. (setContentRect was never called in the
            first place.)

        (2) None of our sizing accounted for scaled resolutions.

            The fix is to ask the WebView to scale all coordintes for us.

        (3) Our code assumed that all window accoutrements were on by default.
            Safari works that way, but other WebKit clients might not.

            The fix is always to explicitly set an on/off state.

        (a) To facilitate scaling, I added a new bridge method, webView, to
        access the webView.

        (b) For internal consistency, I changed ___Bars to ___bars in bridge
        methods, and ___bars to ___Bars in WinArgs data members. (Interestingly,
        the different classes in our code are evenly divided on which format to

        Added manual test:
        * manual-tests/window-open-features.html: Added.
        * manual-tests/resources/200x200.png: Added.
        * manual-tests/resources/popup200x200.html: Added.

        * khtml/ecma/kjs_window.cpp:
        (KJS::showModalDialog): see (b)
        (KJS::WindowFunc::callAsFunction): see (b)
        * kwq/KWQKHTMLPart.mm:
        (KWQKHTMLPart::statusbarVisible): see (b)
        * kwq/KWQKHTMLPartBrowserExtension.mm:
        At the top of this method, I just did some formatting cleanup and
        moved the 'referrer' variable closer to where it's used.
        The changes in the middle of the method are (3), the bottom, (2).
        * kwq/KWQKPartsBrowserExtension.h:
        (KParts::WindowArgs::WindowArgs): see (b)
        * kwq/WebCoreBridge.h: see (a)


        Reviewed by John.

        Part of fix for <rdar://problem/4310363> JavaScript window.open: Height
        is 1 pixel short, and related bugs. See WebCore ChageLog.

        * WebCoreSupport.subproj/WebBridge.m:
        (-[WebBridge webView]): Added method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoSubmitted by: eseidel
eseidel [Tue, 20 Dec 2005 10:08:04 +0000 (10:08 +0000)]
Submitted by: eseidel
Reviewed by: mjs
        Development-only build fix.

        * WebView.subproj/WebFrame.m:
        (-[WebFramePrivate dealloc]): fixed typo

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Not reviewed.
mjs [Tue, 20 Dec 2005 09:59:11 +0000 (09:59 +0000)]
    Not reviewed.

- revert accidental commit of this file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: 6156
eseidel [Tue, 20 Dec 2005 09:30:24 +0000 (09:30 +0000)]
Bug #: 6156
Submitted by: eseidel
Reviewed by: mjs
        Leaks when running SVG tests
        No additional tests necessary, leaks already caught by other tests.

        * kcanvas/KCanvasFilters.cpp:
        (KCanvasFEDiffuseLighting::setLightSource): takes ownership
        (KCanvasFESpecularLighting::setLightSource): takes ownership
        * kcanvas/KCanvasFilters.h:
        (KCanvasFEDiffuseLighting::KCanvasFEDiffuseLighting): added
        (KCanvasFEDiffuseLighting::~KCanvasFEDiffuseLighting): added
        (KCanvasFEDiffuseLighting::lightSource): fixed spacing
        (KCanvasFESpecularLighting::KCanvasFESpecularLighting): added
        (KCanvasFESpecularLighting::~KCanvasFESpecularLighting): added
        (KCanvasFESpecularLighting::lightSource): fixed spacing
        * kcanvas/device/quartz/KRenderingDeviceQuartz.mm:
        (KRenderingDeviceQuartz::stringForPath): added missing CFRelease

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11677 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: 5931
eseidel [Tue, 20 Dec 2005 09:16:45 +0000 (09:16 +0000)]
Bug #: 5931
Submitted by: eseidel
Reviewed by: darin
        Remove additional bit-rotted DEBUG* ifdefs from WebCore.
        No tests possible, only removing dead code.

        * khtml/css/css_stylesheetimpl.cpp:
        * khtml/html/html_formimpl.cpp:
        * khtml/html/htmlparser.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11676 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: 6107
eseidel [Tue, 20 Dec 2005 08:49:51 +0000 (08:49 +0000)]
Bug #: 6107
Submitted by: eseidel
Reviewed by: darin
        Move Decoder onto Shared<T> and clients onto RefPtr.
        No test cases possible, no functional changes.

        * khtml/ecma/XSLTProcessor.cpp:
        * khtml/ecma/xmlhttprequest.cpp:
        * khtml/ecma/xmlhttprequest.h:
        * khtml/khtml_part.cpp:
        * khtml/khtmlpart_p.h:
        * khtml/misc/decoder.cpp:
        * khtml/misc/decoder.h:
        * khtml/misc/loader.h:
        * khtml/xml/dom_docimpl.cpp:
        * khtml/xml/dom_docimpl.h:
        * khtml/xsl/xslt_processorimpl.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11675 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
mjs [Tue, 20 Dec 2005 08:32:56 +0000 (08:32 +0000)]
    Reviewed by Eric.

- Move handling of frame namespaces down to WebFrame.
- Put some internal class declarations in the implementation file.

        * WebView.subproj/WebControllerSets.m:
        (+[WebFrameNamespaces addWebView:toFrameNamespace:]):
        (+[WebFrameNamespaces webViewsInFrameNamespace:]):
        * WebView.subproj/WebFrame.m:
        (-[WebFramePrivate dealloc]):
        (-[WebFrame _setFrameNamespace:]):
        (-[WebFrame _frameNamespace]):
        * WebView.subproj/WebFrameInternal.h:
        * WebView.subproj/WebFramePrivate.h:
        * WebView.subproj/WebView.m:
        (-[WebView _close]):
        (-[WebView _findFrameNamed:sourceFrame:]):
        (-[WebView setGroupName:]):
        (-[WebView groupName]):
        * WebView.subproj/WebViewInternal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11674 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
adele [Mon, 19 Dec 2005 22:16:29 +0000 (22:16 +0000)]

        Reviewed by Darin, committed by Adele.

        - test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6139
          ATSUI code path should implement small caps, synthetic bold and oblique and correct metrics for fallback fonts

        * fast/text/atsui-multiple-renderers-expected.checksum: Added.
        * fast/text/atsui-multiple-renderers-expected.png: Added.
        * fast/text/atsui-multiple-renderers-expected.txt: Added.
        * fast/text/atsui-multiple-renderers.html: Added.

        * css1/text_properties/text-transCapitalize-expected.checksum:
        * css1/text_properties/text-transCapitalize-expected.png:
        * fast/text/international/bidi-AN-after-L-expected.checksum:
        * fast/text/international/bidi-AN-after-L-expected.png:
        * fast/text/international/bidi-AN-after-L-expected.txt:
        * fast/text/international/bidi-L2-run-reordering-expected.checksum:
        * fast/text/international/bidi-L2-run-reordering-expected.png:
        * fast/text/international/bidi-L2-run-reordering-expected.txt:
        * fast/text/international/bidi-LDB-2-CSS-expected.checksum:
        * fast/text/international/bidi-LDB-2-CSS-expected.png:
        * fast/text/international/bidi-LDB-2-CSS-expected.txt:
        * fast/text/international/bidi-LDB-2-HTML-expected.checksum:
        * fast/text/international/bidi-LDB-2-HTML-expected.png:
        * fast/text/international/bidi-LDB-2-HTML-expected.txt:
        * fast/text/international/bidi-LDB-2-formatting-characters-expected.checksum:
        * fast/text/international/bidi-LDB-2-formatting-characters-expected.png:
        * fast/text/international/bidi-LDB-2-formatting-characters-expected.txt:
        * fast/text/international/bidi-european-terminators-expected.checksum:
        * fast/text/international/bidi-european-terminators-expected.png:
        * fast/text/international/bidi-european-terminators-expected.txt:
        * fast/text/international/bidi-fallback-font-weight-expected.checksum:
        * fast/text/international/bidi-fallback-font-weight-expected.png:
        * fast/text/international/bidi-fallback-font-weight-expected.txt:
        * fast/text/international/bidi-ignored-for-first-child-inline-expected.checksum:
        * fast/text/international/bidi-ignored-for-first-child-inline-expected.png:
        * fast/text/international/bidi-ignored-for-first-child-inline-expected.txt:
        * fast/text/international/bidi-innertext-expected.checksum:
        * fast/text/international/bidi-innertext-expected.png:
        * fast/text/international/bidi-innertext-expected.txt:
        * fast/text/international/bidi-override-expected.checksum:
        * fast/text/international/bidi-override-expected.png:
        * fast/text/international/bidi-override-expected.txt:


        Reviewed by Darin, committed by Adele.

        Test: fast/text/atsui-multiple-renderers.html

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6139
          ATSUI code path should implement small caps, synthetic bold and oblique and correct metrics for fallback fonts

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (overrideLayoutOperation): Fetch renderer info from the renderers array;
        add synthetic bold offset; render only synthetic bold in the synthetic bold pass.
        (drawGlyphs): Replaced 14 with new SYNTHETIC_OBLIQUE_ANGLE define.
        (initializeATSUStyle): Apply a skewing transform for synthetic oblique.
        (createATSUTextLayout): Merged into createATSUTextLayout.
        (createATSULayoutParameters): Merged in createATSUTextLayout; allocate
        and fill a renderers array and a character buffer for small caps and mirroring;
        (applyMirroringToRun): Merged into createATSULayoutParameters.
        (ATSU_drawHighlight): Deleted mirroring code.
        (ATSU_draw): Deleted mirroring code; added second pass for synthetic bold.
        (ATSU_pointToOffset): Deleted mirroring code.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11673 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: none
eseidel [Mon, 19 Dec 2005 21:03:53 +0000 (21:03 +0000)]
Bug #: none
Submitted by: eseidel
Reviewed by: darin
        * Scripts/run-webkit-tests: stop /etc/catalog warnings

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11672 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
adele [Mon, 19 Dec 2005 20:54:15 +0000 (20:54 +0000)]

        Reviewed by Darin, committed by Adele.

        - test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6137
          Disable kerning and some ligatures in the ATSUI code path

        * fast/text/atsui-kerning-and-ligatures-expected.checksum: Added.
        * fast/text/atsui-kerning-and-ligatures-expected.png: Added.
        * fast/text/atsui-kerning-and-ligatures-expected.txt: Added.
        * fast/text/atsui-kerning-and-ligatures.html: Added.


        Reviewed by Darin, committed by Adele.

        Test: fast/text/atsui-kerning-and-ligatures.html

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6137
          Disable kerning and some ligatures in the ATSUI code path

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (initializeATSUStyle): Disable kerning; disable ligatures unless the font
        does not contain 'a', in which case it is assumed to never be rendered
        by the CG code path.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11671 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Originally done by both George Staikos and Alexey Proskuryakov.
darin [Mon, 19 Dec 2005 20:48:23 +0000 (20:48 +0000)]
    Originally done by both George Staikos and Alexey Proskuryakov.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5706
          Sharedptr dependency can be removed

        Our coding guidelines say "use 0 instead of NULL" and both RefPtr and
        PassRefPtr were using NULL, which required including a header that
        defines NULL.

        * kxmlcore/PassRefPtr.h:
        (KXMLCore::PassRefPtr::PassRefPtr): Use 0 instead of NULL.
        (KXMLCore::PassRefPtr::operator!): Use ! instead of == NULL.
        * kxmlcore/RefPtr.h:
        (KXMLCore::RefPtr::RefPtr): Use 0 instead of NULL.
        (KXMLCore::RefPtr::operator!): Use ! instead of == NULL.
        Also did some reformatting.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11670 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Maciej.
darin [Mon, 19 Dec 2005 20:41:47 +0000 (20:41 +0000)]
    Reviewed by Maciej.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6143
          DOM::ElementImpl should use a RefPtr for the attribute map

        * khtml/xml/dom_elementimpl.cpp:
        (ElementImpl::ElementImpl): Remove code to initialize the pointer; not needed since
        RefPtr gets initialized to 0.
        (ElementImpl::~ElementImpl): Remove code to deref the pointer; RefPtr handles that.
        (ElementImpl::attributes): Add get() call to get raw pointer.
        (ElementImpl::setAttributeMap): Remove code to deref the old map and set the new map.
        But added code to clear the element pointer from the old map (missing in the old
        version). Also added a FIXME.
        (ElementImpl::createAttributeMap): Remove ref(); RefPtr handles that.
        (NamedAttrMapImpl::addAttribute): Use a RefPtr to guarantee the element does not go
        away in the middle of dispatching DOM events.
        (StyledElementImpl::attributeChanged): Clean up code by using the inline function
        mappedAttributes() instead of doing type casts.
        (StyledElementImpl::parseMappedAttribute): Ditto.
        (StyledElementImpl::getClassList): Ditto.

        * khtml/xml/dom_elementimpl.h: Make ElementImpl::namedAttrMap be a RefPtr instead
        of raw pointer. Added an overload of StyledElementImpl::mappedAttributes for both
        const and non-const.

        * khtml/xml/dom_nodeimpl.cpp: (DOM::NodeImpl::addChild): Use a RefPtr to ref/deref
        the child so that it doesn't leak.

        * khtml/html/htmlparser.h: Changed isindex to use a RefPtr.
        * khtml/html/htmlparser.cpp:
        (HTMLParser::~HTMLParser): Removed now-unneeded ref.
        (HTMLParser::isindexCreateErrorCheck): Remove now-unneeded deref/ref.
        (HTMLParser::handleIsindex): Put isindex element into a RefPtr. This prevents a
        crash that was otherwise happening during layout tests (caused indirectly by
        the changes above).
        (HTMLParser::startBody): Added call to get().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11669 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric Seidel.
darin [Mon, 19 Dec 2005 19:56:25 +0000 (19:56 +0000)]
    Reviewed by Eric Seidel.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=4990
          WebKit needs to use a local pasteboard during testing

        * DumpRenderTree/DumpRenderTree.m:
        (main): Call poseAs to substitute our NSPasteboard class for the default one.
        Create a local pasteboard (really a global one with a unique name) and release
        it when exiting from the function so we don't leave it in the pasteboard server.
        (dumpRenderTree): Added an autorelease pool around one small bit of code that
        ddn't have one. Fixed a leak in an unlikely error case.
        (+[DumpRenderTreePasteboard generalPasteboard]): Override the default version
        of this method to return our local pasteboard.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11668 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoJavaScriptCore:
darin [Mon, 19 Dec 2005 19:53:12 +0000 (19:53 +0000)]

        Reviewed by Geoff Garen and Eric Seidel.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=4923
          stop using <ostream> in WebCore, eliminating the <cmath> troubles it causes

        * kjs/simple_number.h: Removed many unnecessary includes, including
        the <cmath> one to work around GCC library header bugs. We may have to
        add some includes elsewhere for platforms other than OS X, since our
        prefix header takes care of some things.

        * kxmlcore/AlwaysInline.h: Added. Now clients that don't include
        simple_number.h can still get the ALWAYS_INLINE macro.
        * JavaScriptCore.xcodeproj/project.pbxproj: Added AlwaysInline.h.

        * bindings/NP_jsobject.h: Removed a lot of unnecessary includes
        and removed C-specific stuff from this C++-only header.
        * bindings/jni/jni_jsobject.h: Removed a lot of unnecessary includes
        and did some reformatting.
        * bindings/objc/objc_runtime.h:  Removed an unnecessary include.
        * bindings/runtime.h: Removed some unneeded includes. Reformatted.
        * bindings/runtime.cpp: Updated to compile with header changes,
        including a lot of reformatting.
        * bindings/runtime_object.h: Removed an unnecessary include.


        Reviewed by Geoff Garen and Eric Seidel.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=4923
          stop using <ostream> in WebCore, eliminating the <cmath> troubles it causes

        * ForwardingHeaders/kxmlcore/AlwaysInline.h: Added.
        * WebCorePrefix.h: Removed the use of <ostream>.

        * kwq/KWQDef.h: Removed now-unused Q_INT64, Q_INT16, Q_UINT16, qRound, and _KWQ_IOSTREAM_.

        * kwq/KWQKHTMLPart.h: Removed some unneeded headers and added forward-declarations of classes
        instead. Corrected some incorrect member function declarations.
        * kwq/KWQKHTMLPart.mm: Added a now-needed header.

        * khtml/ecma/kjs_window.cpp: Removed the <cmath> workaround.
        * khtml/rendering/bidi.cpp: Added include of AlwaysInline.h and removed a
        lot of unnecessary includes.
        * khtml/rendering/render_canvasimage.cpp: Removed the <cmath> workaround.
        * khtml/rendering/render_image.cpp: Removed the <cmath> workaround.
        * khtml/rendering/render_text.cpp: Added include of AlwaysInline.h and removed a
        lot of unnecessary includes.

        * ksvg2/css/SVGCSSStyleSelector.cpp: (KDOM::CSSStyleSelector::applySVGProperty):
        * ksvg2/svg/SVGAnimateColorElementImpl.cpp: (SVGAnimateColorElementImpl::calculateColor):
        * ksvg2/svg/SVGAnimateTransformElementImpl.cpp: (SVGAnimateTransformElementImpl::handleTimerEvent):
        * ksvg2/svg/SVGAnimationElementImpl.cpp: (SVGAnimationElementImpl::closeRenderer):
        * ksvg2/svg/SVGLinearGradientElementImpl.cpp: (SVGLinearGradientElementImpl::buildGradient):
        * ksvg2/svg/SVGPatternElementImpl.cpp: (SVGPatternElementImpl::notifyAttributeChange):
        * ksvg2/svg/SVGRadialGradientElementImpl.cpp: (SVGRadialGradientElementImpl::buildGradient):
        Replaced use of qRound with use of lroundf or lround as appropriate.

        * kwq/KWQCString.h:
        * kwq/KWQCString.mm:
        * kwq/KWQDateTime.h:
        * kwq/KWQDateTime.mm:
        * kwq/KWQMap.h:
        * kwq/KWQMemArray.h:
        * kwq/KWQPoint.mm:
        * kwq/KWQPointArray.h:
        * kwq/KWQPtrList.h:
        * kwq/KWQPtrStack.h:
        * kwq/KWQPtrVector.h:
        * kwq/KWQRect.h:
        * kwq/KWQRect.mm:
        * kwq/KWQSize.h:
        * kwq/KWQSize.mm:
        * kwq/KWQValueList.h:
        Removed _KWQ_IOSTREAM_ code that was used at one time for unit tests, but is now unneeded,
        and requires <ostream>.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11667 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Geoff Garen and John Sullivan.
darin [Mon, 19 Dec 2005 19:50:45 +0000 (19:50 +0000)]
    Reviewed by Geoff Garen and John Sullivan.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=4312
          XMLHttpRequest headers that have two CRLF sequences lead to Obj-C exception

        * kwq/KWQLoader.mm:
        (+[NSDictionary _webcore_dictionaryWithHeaderString:_webcore_dictionaryWithHeaderString:]):
        Check length of string before calling characterAtIndex:0 since it will fail for an empty string.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11666 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #:
bdakin [Mon, 19 Dec 2005 19:41:39 +0000 (19:41 +0000)]
Bug #:
Submitted by:
Reviewed by:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
adele [Mon, 19 Dec 2005 19:37:58 +0000 (19:37 +0000)]

        Reviewed by Darin, committed by Adele.

        - test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6132
          Incorrect selection highlighting for ATSUI text when selected range is "CG-safe"

        * fast/text/should-use-atsui-expected.checksum: Added.
        * fast/text/should-use-atsui-expected.png: Added.
        * fast/text/should-use-atsui-expected.txt: Added.
        * fast/text/should-use-atsui.html: Added.


        Reviewed by Darin, committed by Adele.

        Test: fast/text/should-use-atsui.html

        - fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=6132
          Incorrect selection highlighting for ATSUI text when selected range is "CG-safe"

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (shouldUseATSU): Always start scanning from 0 since drawing and
        highlighting also measure everything up to run->from.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11664 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Darin, committed by Adele.
adele [Mon, 19 Dec 2005 19:17:32 +0000 (19:17 +0000)]
    Reviewed by Darin, committed by Adele.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5744
          XMLHttpRequest does not apply page encoding after assigning via innerHtml

        * khtml/ecma/xmlhttprequest.cpp:
        (getMIMEType): A helper function to get MIME type from a Content-Type string.
        (getCharset): A helper function to get charset from a Content-Type string.
        (KJS::XMLHttpRequest::getValueProperty): Factored out responseIsXML().
        (KJS::XMLHttpRequest::getResponseHeader): Return QString instead of JSValue
          (to get rid of unnecessary JSLocks).
        (KJS::XMLHttpRequest::responseIsXML): A new method that analyses Content-Type.
        (KJS::XMLHttpRequest::slotData): Use a correct charset for responses, see bug for details.
        (KJS::XMLHttpRequestProtoFunc::callAsFunction): Update for getResponseHeader() changes
        * khtml/ecma/xmlhttprequest.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11663 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago== Rolled over to ChangeLog-2005-12-19 ==
thatcher [Mon, 19 Dec 2005 19:16:44 +0000 (19:16 +0000)]
== Rolled over to ChangeLog-2005-12-19 ==

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11662 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Geoff and Adele
mjs [Mon, 19 Dec 2005 10:38:07 +0000 (10:38 +0000)]
    Reviewed by Geoff and Adele

- replaced custom Identifier hashtable with HashSet

        * kjs/identifier.cpp:
        * kjs/identifier.h:
        * kjs/internal.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11661 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoWebCore:
mjs [Mon, 19 Dec 2005 10:32:11 +0000 (10:32 +0000)]

        Rubber stamped by Eric.

        - removed unused SPI headers

        * WebCore.xcodeproj/project.pbxproj:
        * kwq/DOM-compat.h: Removed.


        Rubber stamped by Eric.

- remove some unused SPI headers.

        * DOM.subproj/WebDOMDocument.h: Removed.
        * DOM.subproj/WebDOMElement.h: Removed.
        * DOM.subproj/WebDOMNode.h: Removed.
        * WebKit.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11660 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoJavaScriptCore:
darin [Mon, 19 Dec 2005 00:27:29 +0000 (00:27 +0000)]

        Reviewed, tweaked, and landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5227
          Array indexOf() extension for JavaScript 1.5 Core

        * kjs/array_object.h:
        * kjs/array_object.cpp: (ArrayProtoFunc::callAsFunction): Added implementation of indexOf.


        Reviewed, tweaked, and landed by Darin.

        - test for http://bugzilla.opendarwin.org/show_bug.cgi?id=5227
          Array indexOf() extension for JavaScript 1.5 Core

        * fast/js/array-indexof-expected.txt: Added.
        * fast/js/array-indexof.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11659 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Geoff.
darin [Sun, 18 Dec 2005 23:41:48 +0000 (23:41 +0000)]
    Reviewed by Geoff.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6144
          REGRESSION: DOMStringImpl leak in code to make DOMString from NSString

        * kwq/DOMInternal.mm: (DOMString::DOMString): Remove now-unneeded ref.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11658 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Geoff.
darin [Sun, 18 Dec 2005 23:40:13 +0000 (23:40 +0000)]
    Reviewed by Geoff.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5680
          containsOnlyWhitespace does not need to be a virtual function

        * khtml/rendering/render_frames.cpp: (RenderPartObject::updateWidget): Cast
        pointer to a TextImpl before calling containsOnlyWhitespace; the code already
        checked isTextNode.
        * khtml/rendering/render_line.cpp: (khtml::shouldDrawDecoration): Expanded an
        if statement into a few separate ones for slight additional clarity. Added a
        check of isTextNode and cast pointer to a TextImpl.

        * khtml/xml/dom_nodeimpl.h: Remove containsOnlyWhitespace function.
        * khtml/xml/dom_textimpl.h: Remove virtual keyword from containsOnlyWhitespace.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11657 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
darin [Sun, 18 Dec 2005 22:55:35 +0000 (22:55 +0000)]

        Reviewed by Hyatt, landed by Darin.

        Test for: http://bugzilla.opendarwin.org/show_bug.cgi?id=5980
        Bidi properties of an inline container whose first child is an
        inline container are ignored

        * fast/text/international/bidi-ignored-for-first-child-inline-expected.checksum: Added.
        * fast/text/international/bidi-ignored-for-first-child-inline-expected.png: Added.
        * fast/text/international/bidi-ignored-for-first-child-inline-expected.txt: Added.
        * fast/text/international/bidi-ignored-for-first-child-inline.html: Added.


        Reviewed by Hyatt, landed by Darin.

        Test: fast/text/international/bidi-ignored-for-first-child-inline.html

        Fix for: http://bugzilla.opendarwin.org/show_bug.cgi?id=5980
        Bidi properties of an inline container whose first child is an
        inline container are ignored

        * khtml/rendering/bidi.cpp:
        Adjust embedding at inline containers' boundaries instead of inside.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
darin [Sun, 18 Dec 2005 22:37:26 +0000 (22:37 +0000)]

        Reviewed and landed by Darin.

        - test of a particular character set name that is not in the WebCore table
          but is in the ICU table (see bugzilla 5932)

        * fast/encoding/charset-cp1251-expected.txt: Added.
        * fast/encoding/charset-cp1251.html: Added.


        Reviewed and landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5932
          Wrong encoding used for http://www.miel.ru

        * kwq/KWQCharsets.mm: (KWQCFStringEncodingFromIANACharsetName):
        Fall back on ICU encoding names if we don't find the name in the WebCore
        encoding names table.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11655 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
darin [Sun, 18 Dec 2005 22:23:29 +0000 (22:23 +0000)]

        Reviewed and landed by Darin.

        - XSLT document encoding tests

        * fast/xsl/xslt-doc-enc-expected.txt: Added.
        * fast/xsl/xslt-doc-enc.xml: Added.
        * fast/xsl/xslt-doc-enc.xsl: Added.
        * fast/xsl/xslt-doc-noenc-expected.txt: Added.
        * fast/xsl/xslt-doc-noenc.xml: Added.
        * fast/xsl/xslt-doc-noenc.xsl: Added.


        Reviewed and landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5927
          Result of XSLT transformation has no charset defined

        * khtml/xml/dom_docimpl.cpp: (DocumentImpl::applyXSLTransform):
        Get encoding from transform and pass in to XSLTProcessor.

        * khtml/xsl/xslt_processorimpl.h:
        * khtml/xsl/xslt_processorimpl.cpp:
        (DOM::XSLTProcessorImpl::createDocumentFromSource): Added encoding parameter.
        Create a Decoder and use it to decode the source document (default encoding UTF-8).
        (DOM::XSLTProcessorImpl::transformToString): Add an encoding result. Set it to the
        encoding from the libxml2 document object.
        (DOM::XSLTProcessorImpl::transformToDocument): Get encoding from transform and pass
        in to XSLTProcessor.
        (DOM::XSLTProcessorImpl::transformToFragment): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11654 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric, landed by Darin.
darin [Sun, 18 Dec 2005 22:13:20 +0000 (22:13 +0000)]
    Reviewed by Eric, landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5631
          KWQKHTMLPart::attributedString ignores many tags

        * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::attributedString):
        Fix incorrectly nested else.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoFix date
andersca [Sun, 18 Dec 2005 13:28:33 +0000 (13:28 +0000)]
Fix date

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago2005-12-18 Anders Carlsson <andersca@mac.com>
andersca [Sun, 18 Dec 2005 13:27:35 +0000 (13:27 +0000)]
2005-12-18  Anders Carlsson  <andersca@mac.com>

        Reviewed by Darin.

        Test case for http://bugzilla.opendarwin.org/show_bug.cgi?id=4000
        Object.prototype is missing isPrototypeOf

        * fast/js/test-isPrototypeOf-expected.txt: Added.
        * fast/js/test-isPrototypeOf.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago2005-12-18 Anders Carlsson <andersca@mac.com>
andersca [Sun, 18 Dec 2005 13:27:00 +0000 (13:27 +0000)]
2005-12-18  Anders Carlsson  <andersca@mac.com>

        Reviewed by Darin and Geoffrey.

        - fix for <http://bugzilla.opendarwin.org/show_bug.cgi?id=4000>
        Object.prototype is missing isPrototypeOf

        * kjs/object_object.cpp:
        Add isPrototypeOf to object prototype.

        Implement isPrototypeOf

        * kjs/object_object.h:
        Add id for isPrototypeOf.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11650 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoupdaated changelog to remove previous suggestion that darin was a nobody
ggaren [Sun, 18 Dec 2005 09:20:20 +0000 (09:20 +0000)]
updaated changelog to remove previous suggestion that darin was a nobody

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed and landed by Darin.
darin [Sun, 18 Dec 2005 05:01:35 +0000 (05:01 +0000)]
    Reviewed and landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5768
          pages with frames that are all "fixed" get no layout and are blank (like www.farnell.nl)

        * khtml/rendering/render_frames.cpp: (RenderFrameSet::layout):
        * manual-tests/bugzilla-5768.html: Added. Can't seem to get it to work (fail) as an automated test.
        * manual-tests/resources/frame1.html: Added.
        * manual-tests/resources/frame2.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11648 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Layout test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6119
ggaren [Sun, 18 Dec 2005 03:09:41 +0000 (03:09 +0000)]
    Layout test for bugzilla.opendarwin.org/show_bug.cgi?id=6119
        split() function ignores case insensitive modifier.

        * fast/js/string-split-ignore-case-expected.txt: Added.
        * fast/js/string-split-ignore-case.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11647 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by NOBODY (OOPS!).
ggaren [Sun, 18 Dec 2005 03:09:31 +0000 (03:09 +0000)]
    Reviewed by NOBODY (OOPS!).

        Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=6119
        split() function ignores case insensitive modifier.

        RegExpImp: The C++ object you get when JavaScript executes
                   "new RegExp()".
        RegExp:    A C++ wrapper object that performs regular expression
                   matching on behalf of a RegExpImp.

        Instead of unnecessarily constructing a RegExp which (wrongly) lacks
        any modifiers, String.split() now uses the RegExp built in to the
        RegExpImp passed to it, which has the right modifiers already.

        I also cleaned up other bits of the string code to standardized how
        we handle RegExpImp arguments.

        * ChangeLog:
        * kjs/string_object.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11646 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago2005-12-17 Anders Carlsson <andersca@mac.com>
andersca [Sat, 17 Dec 2005 20:12:35 +0000 (20:12 +0000)]
2005-12-17  Anders Carlsson  <andersca@mac.com>

        Reviewed by Darin.

        - Fix <http://bugzilla.opendarwin.org/show_bug.cgi?id=6120>
        Random crashes when running the test suite

        * khtml/css/css_ruleimpl.h:
        Make m_lstMedia a smart pointer.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11645 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago2005-12-17 Anders Carlsson <andersca@mac.com>
andersca [Sat, 17 Dec 2005 20:09:46 +0000 (20:09 +0000)]
2005-12-17  Anders Carlsson  <andersca@mac.com>

        Reviewed by Eric.

        * khtml/css/css_ruleimpl.h:
        Make m_lstMedia a smart pointer.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11644 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoLayoutTests:
darin [Sat, 17 Dec 2005 18:10:36 +0000 (18:10 +0000)]

        Reviewed and landed by Darin.

        - test for http://bugzilla.opendarwin.org/show_bug.cgi?id=5479
          Can't select text with RTL override rendered by ATSUI

        * fast/text/atsui-rtl-override-selection-expected.checksum: Added.
        * fast/text/atsui-rtl-override-selection-expected.png: Added.
        * fast/text/atsui-rtl-override-selection-expected.txt: Added.
        * fast/text/atsui-rtl-override-selection.html: Added.


        Reviewed and landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5479
          Can't select text with RTL override rendered by ATSUI

        * khtml/rendering/font.h: Remove reversed parameter from checkSelectionPoint, replacing
        it with a text direction parameter.
        * khtml/rendering/render_text.cpp: (InlineTextBox::offsetForPosition): Ditto.
        * kwq/KWQFontMetrics.h: Add a directional override parameter to checkSelectionPoint.
        * kwq/KWQFontMetrics.mm: (QFontMetrics::checkSelectionPoint): Pass the directional override
        through to the text renderer. Don't pass the reversed flag redundantly -- use the rtl flag
        in the style instead.
        * kwq/WebCoreTextRenderer.h: Remove reversed parameter from pointToOffset.


        Reviewed and landed by Darin.

        - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=5479
          Can't select text with RTL override rendered by ATSUI

        * WebCoreSupport.subproj/WebTextRenderer.m:
        (overrideLayoutOperation): Update since the layout only includes from -> to now.
        (-[WebTextRenderer pointToOffset:style:position:includePartialGlyphs:]): Remove
        reversed parameter.
        (CG_floatWidthForRun): Add code to handle RTL case.
        (addDirectionalOverride): Put the override around the entire run.
        (ATSU_drawHighlight): Rearrange and reuse ATSU_floatWidthForRun for more of the work.
        Also round.
        (ATSU_pointToOffset): Remove reversed parameter and run swapping.
        (CG_pointToOffset): Remove reversed parameter, using rtl flag in style instead.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11643 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoFix the calls to strftime on the non-apple code path to take const tm* and not just tm
hyatt [Fri, 16 Dec 2005 23:01:33 +0000 (23:01 +0000)]
Fix the calls to strftime on the non-apple code path to take const tm* and not just tm

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11640 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoFix the calls to strftime on the non-apple code path to take const tm* and not just tm
hyatt [Fri, 16 Dec 2005 22:56:41 +0000 (22:56 +0000)]
Fix the calls to strftime on the non-apple code path to take const tm* and not just tm

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11639 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoAdd assert.h include to ustring.cpp
hyatt [Fri, 16 Dec 2005 22:48:19 +0000 (22:48 +0000)]
Add assert.h include to ustring.cpp

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11638 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoifdef out the getpid call on windows.
hyatt [Fri, 16 Dec 2005 22:45:35 +0000 (22:45 +0000)]
ifdef out the getpid call on windows.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11637 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Remove unused RefPtr constructors that can create an ambiguity in ustring on some...
hyatt [Fri, 16 Dec 2005 22:27:26 +0000 (22:27 +0000)]
Remove unused RefPtr constructors that can create an ambiguity in ustring on some platforms.

        Reviewed by mjs

        * kxmlcore/RefPtr.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11636 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: 6106
eseidel [Fri, 16 Dec 2005 22:26:12 +0000 (22:26 +0000)]
Bug #: 6106
Submitted by: eseidel
Reviewed by: darin
        DOMString should hold its impl in a RefPtr
        I also cleaned up spacing in dom_string.cpp
        No tests needed (no functionality changes).

        * khtml/dom/dom_string.cpp:
        (DOM::DOMString::operator += ):
        (DOM::DOMString::operator []):
        * khtml/dom/dom_string.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11635 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoBug #: 6110
eseidel [Fri, 16 Dec 2005 21:43:48 +0000 (21:43 +0000)]
Bug #: 6110
Submitted by: eseidel
Reviewed by: darin
        No test cases needed (no functionality changes).

        * khtml/xsl/xsl_stylesheetimpl.cpp:
        * khtml/xsl/xsl_stylesheetimpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Maciej.
adele [Fri, 16 Dec 2005 21:30:35 +0000 (21:30 +0000)]
    Reviewed by Maciej.

        Fix for <rdar://problem/4381546> Safari crash at www.dallascowboys.com in KWQListIteratorImpl::current()

        * khtml/khtml_part.cpp:
        (KHTMLPart::slotFinishedParsing): ref the part before calling checkCompleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11631 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Tim O.
vicki [Fri, 16 Dec 2005 21:21:27 +0000 (21:21 +0000)]
    Reviewed by Tim O.

- fix <rdar://problem/4380428> window.onbeforeunload works on IE but not on Safari

        * khtml/ecma/kjs_window.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11630 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years agoDon't include bits/cpp_type_traits on win32
hyatt [Fri, 16 Dec 2005 20:33:21 +0000 (20:33 +0000)]
Don't include bits/cpp_type_traits on win32

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11628 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Reviewed by Eric.
ggaren [Fri, 16 Dec 2005 18:35:19 +0000 (18:35 +0000)]
    Reviewed by Eric.

        Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=6112
        REGRESSION: navigator.plugins empty.

        Layout test added:
        * fast/js/Plug-ins-expected.checksum: Added.
        * fast/js/Plug-ins-expected.png: Added.
        * fast/js/Plug-ins-expected.txt: Added.
        * fast/js/Plug-ins.html: Added.

        * khtml/ecma/kjs_navigator.cpp:
        (KJS::PluginBase::PluginBase): Changed initialization of KConfig from
        empty string to pluginsinfo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11624 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 years ago Layout test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6112
ggaren [Fri, 16 Dec 2005 18:34:56 +0000 (18:34 +0000)]
    Layout test for bugzilla.opendarwin.org/show_bug.cgi?id=6112
        REGRESSION: navigator.plugins empty.

        * fast/js/Plug-ins-expected.checksum: Added.
        * fast/js/Plug-ins-expected.png: Added.
        * fast/js/Plug-ins-expected.txt: Added.
        * fast/js/Plug-ins.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@11623 268f45cc-cd09-0410-ab3c-d52691b4dbfc