Timeline

Apr 25, 2005:

4:35 PM Changeset in webkit [9049] by sullivan

• 2 edits in trunk/WebCore

Reviewed by Dave Harrison.

• fixed <rdar://problem/4098731> [DOMHTMLTextAreaElement cols] returns wrong value (usually zero)

• kwq/DOMHTML.mm: (-[DOMHTMLTextAreaElement cols]): use ATTR_COLS instead of ATTR_ACCESSKEY (D'oh!) (-[DOMHTMLTextAreaElement setCols:]): fixed whitespace

11:37 AM Changeset in webkit [9048]

• 3 copies
  3 deletes in tags/JavaScriptCore-412~1

This commit was manufactured by cvs2svn to create tag
'JavaScriptCore-412~1'.

11:37 AM Changeset in webkit [9047] by vicki

• 2 edits in trunk/JavaScriptCore

JavaScriptCore-412.1

8:55 AM Changeset in webkit [9046] by darin

• 6 edits in trunk/WebCore

Reviewed by John.

• fixed <rdar://problem/4097849> REGRESSION (162-163): importNode creates non-HTML elements, thus style attributes (and some others) don't work

• khtml/xml/dom_docimpl.h: Add virtual functions HTMLElementNamespace and isHTMLNamespace.
• khtml/xml/dom_docimpl.cpp: (DocumentImpl::importNode): Rearranged this function and fixed the following problems: 1) made sure to ref node while attributes are being set on it so it doesn't get destroyed; 2) fixed code to get namespace from the element we are importing to use the IDs from the source document, not the destination document; 3) removed unneeded getDocument() call which just returns this; 4) fixed error handling for cases where an exception happens while processing the children. (DocumentImpl::HTMLElementNamespace): Added. Returns XHTML_NAMESPACE. (DocumentImpl::isHTMLNamespace): Added. Returns true for any namespace that matches XHTML_NAMESPACE (case insensitive). (DocumentImpl::createElementNS): Changed to call isHTMLNamespace, which will cause it to accept the null namespace in an HTML document. This is the change that fixes the bug. Also fixed the code path to do a little less wasteful work in the non-XHTML case. (DocumentImpl::createHTMLElement): Pass in HTMLElementNamespace() rather than 0 to tagId. (DocumentImpl::attrId): Use isHTMLNamespace instead of allowing the null namespace explicitly. (DocumentImpl::tagId): Ditto.

• khtml/html/html_documentimpl.h: Add overrides for HTMLElementNamespace and isHTMLNamespace.
• khtml/html/html_documentimpl.cpp: (HTMLDocumentImpl::HTMLElementNamespace): Added. Returns 0 so we use the null string for HTML elements inside HTML documents (as before). (HTMLDocumentImpl::isHTMLNamespace): Added. Allows 0, and then calls base class to check for the actual XHTML namespace. Thus, we allow both no namespace at all and the XHTML namespace inside HTML documents.

• khtml/html/html_elementimpl.cpp: (HTMLElementImpl::namespaceURI): Changed to call HTMLElementNamespace rather than checking isHTMLDocument. Same result as before, but better division of responsibilities.

8:43 AM Changeset in webkit [9045] by darin

• 2 edits in trunk/WebCore

Reviewed by John.

• fixed <rdar://problem/4091956> JavaScript drop handlers don't receive more than one dropped item

• kwq/KWQClipboard.mm: (cocoaTypeFromMIMEType): Remove some use of compare -- less efficient than a simpler "==" check. Fixed non-GC-safe code to use KWQCFAutorelease instead. (MIMETypeFromCocoaType): Use fromCFString instead of fromNSString to avoid a cast. (KWQClipboard::getData): Rearrange so that we'll use filenames if both filenames and a URL are present, since filenames can accomodate multiple items. Fix bug where we'd return multiple filenames when the type requested is "URL". Fixed loop that computed the count over and over again for the loop termination condition. Check that the data for NSFilenamesPboardType is an NSArray instead of assuming it is. (KWQClipboard::setData): Use isEqualToString: instead of == when comparing types.