WebCore/ChangeLog-2005-08-23

   1 2005-02-09  Chris Blumenberg  <cblu@apple.com>
   2
   3         Fixed: <rdar://problem/3999213> Sometimes 2 Windows Media Player plugin instances are loaded
   4
   5         Reviewed by hyatt.
   6
   7         * khtml/html/html_objectimpl.cpp:
   8         (HTMLObjectElementImpl::attach): Set needWidgetUpdate to false before calling updateWidget because updateWidget may cause this method or another method (which also calls updateWidget) to be called.
   9         (HTMLObjectElementImpl::recalcStyle): ditto
  10
  11 2005-02-09  David Harrison  <harrison@apple.com>
  12
  13         Reviewed by Darin.
  14
  15         <rdar://problem/3937447> Mail-689: Arrow key navigation in new message body stops working when a line ends with a bold word
  16
  17         Added affinity to VisiblePosition.  Changed Selection code to use affinity more.
  18
  19         (Partial) <rdar://problem/3982096> editing/pasteboard/paste-text-007 is failing
  20
  21         Changed ReplaceSelectionCommand to also pay attention to the top children of the incoming fragment, not just the very last node inserted, when deciding whether to insert a paragraph for the Apple interchange newline.
  22
  23         * khtml/ecma/kjs_window.cpp:
  24         (SelectionFunc::tryCall):
  25         * khtml/editing/htmlediting.cpp:
  26         (khtml::EditCommandPtr::setStartingSelection):
  27         (khtml::EditCommandPtr::setEndingSelection):
  28         (khtml::EditCommand::setStartingSelection):
  29         (khtml::EditCommand::setEndingSelection):
  30         (khtml::CompositeEditCommand::deleteInsignificantTextDownstream):
  31         (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary):
  32         (khtml::ApplyStyleCommand::splitTextAtStartIfNeeded):
  33         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded):
  34         (khtml::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
  35         (khtml::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
  36         (khtml::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
  37         (khtml::ApplyStyleCommand::mergeEndWithNextIfIdentical):
  38         (khtml::ApplyStyleCommand::joinChildTextNodes):
  39         (khtml::DeleteSelectionCommand::initializePositionData):
  40         (khtml::DeleteSelectionCommand::handleGeneralDelete):
  41         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete):
  42         (khtml::DeleteSelectionCommand::doApply):
  43         (khtml::InsertLineBreakCommand::doApply):
  44         (khtml::InsertParagraphSeparatorCommand::doApply):
  45         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
  46         (khtml::InsertTextCommand::input):
  47         (khtml::MoveSelectionCommand::doApply):
  48         (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand):
  49         (khtml::ReplaceSelectionCommand::~ReplaceSelectionCommand):
  50         (khtml::ReplaceSelectionCommand::doApply):
  51         (khtml::ReplaceSelectionCommand::completeHTMLReplacement):
  52         (khtml::ReplaceSelectionCommand::updateNodesInserted):
  53         (khtml::TypingCommand::deleteKeyPressed):
  54         (khtml::TypingCommand::forwardDeleteKeyPressed):
  55         (khtml::TypingCommand::markMisspellingsAfterTyping):
  56         * khtml/editing/htmlediting.h:
  57         * khtml/editing/selection.cpp:
  58         (khtml::Selection::Selection):
  59         (khtml::Selection::init):
  60         (khtml::Selection::moveTo):
  61         (khtml::Selection::modifyExtendingRightForward):
  62         (khtml::Selection::modifyMovingRightForward):
  63         (khtml::Selection::modifyExtendingLeftBackward):
  64         (khtml::Selection::modifyMovingLeftBackward):
  65         (khtml::Selection::modify):
  66         (khtml::Selection::xPosForVerticalArrowNavigation):
  67         (khtml::Selection::clear):
  68         (khtml::Selection::setBase):
  69         (khtml::Selection::setExtent):
  70         (khtml::Selection::setBaseAndExtent):
  71         (khtml::Selection::layout):
  72         (khtml::Selection::validate):
  73         * khtml/editing/selection.h:
  74         (khtml::Selection::startAffinity):
  75         (khtml::Selection::endAffinity):
  76         (khtml::Selection::baseAffinity):
  77         (khtml::Selection::extentAffinity):
  78         (khtml::operator==):
  79         * khtml/editing/text_affinity.h:
  80         (khtml::):
  81         * khtml/editing/visible_position.cpp:
  82         (khtml::VisiblePosition::VisiblePosition):
  83         (khtml::VisiblePosition::init):
  84         (khtml::VisiblePosition::initUpstream):
  85         (khtml::VisiblePosition::initDownstream):
  86         (khtml::VisiblePosition::next):
  87         (khtml::VisiblePosition::previous):
  88         (khtml::startVisiblePosition):
  89         (khtml::endVisiblePosition):
  90         * khtml/editing/visible_position.h:
  91         (khtml::VisiblePosition::):
  92         (khtml::VisiblePosition::VisiblePosition):
  93         (khtml::VisiblePosition::affinity):
  94         (khtml::VisiblePosition::setAffinity):
  95         (khtml::operator==):
  96         * khtml/editing/visible_units.cpp:
  97         (khtml::previousBoundary):
  98         (khtml::nextBoundary):
  99         (khtml::startOfLine):
 100         (khtml::endOfLine):
 101         (khtml::previousLinePosition):
 102         (khtml::nextLinePosition):
 103         (khtml::startOfParagraph):
 104         (khtml::endOfParagraph):
 105         (khtml::previousParagraphPosition):
 106         (khtml::nextParagraphPosition):
 107         (khtml::startOfBlock):
 108         (khtml::endOfBlock):
 109         (khtml::startOfDocument):
 110         (khtml::endOfDocument):
 111         * khtml/editing/visible_units.h:
 112         * khtml/khtml_part.cpp:
 113         (KHTMLPart::findTextNext):
 114         (KHTMLPart::selectClosestWordFromMouseEvent):
 115         (KHTMLPart::handleMousePressEventTripleClick):
 116         (KHTMLPart::handleMousePressEventSingleClick):
 117         (KHTMLPart::handleMouseMoveEventSelection):
 118         (KHTMLPart::khtmlMouseReleaseEvent):
 119         (KHTMLPart::selectAll):
 120         (KHTMLPart::computeAndSetTypingStyle):
 121         (KHTMLPart::selectionComputedStyle):
 122         * khtml/rendering/render_br.cpp:
 123         (RenderBR::positionForCoordinates):
 124         * khtml/xml/dom_docimpl.cpp:
 125         (DocumentImpl::updateSelection):
 126         * khtml/xml/dom_nodeimpl.cpp:
 127         (NodeBaseImpl::setFocus):
 128         * khtml/xml/dom_position.cpp:
 129         (DOM::Position::previousCharacterPosition):
 130         (DOM::Position::nextCharacterPosition):
 131         * khtml/xml/dom_position.h:
 132         * kwq/KWQAccObject.mm:
 133         (-[KWQAccObject value]):
 134         (-[KWQAccObject visiblePositionForStartOfTextMarkerRange:]):
 135         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
 136         (-[KWQAccObject accessibilityAttributeValue:]):
 137         (-[KWQAccObject doAXLineForTextMarker:]):
 138         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
 139         (-[KWQAccObject doAXTextMarkerForPosition:]):
 140         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
 141         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
 142         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
 143         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
 144         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
 145         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
 146         (-[KWQAccObject doSetAXSelectedTextMarkerRange:]):
 147         * kwq/KWQAccObjectCache.mm:
 148         (KWQAccObjectCache::textMarkerForVisiblePosition):
 149         (KWQAccObjectCache::visiblePositionForTextMarker):
 150         * kwq/KWQKHTMLPart.mm:
 151         (KWQKHTMLPart::findString):
 152         (KWQKHTMLPart::advanceToNextMisspelling):
 153         (KWQKHTMLPart::styleForSelectionStart):
 154         (KWQKHTMLPart::baseWritingDirectionForSelectionStart):
 155         (KWQKHTMLPart::setSelectionFromNone):
 156         (KWQKHTMLPart::respondToChangedSelection):
 157         * kwq/WebCoreBridge.mm:
 158         (-[WebCoreBridge setSelectedDOMRange:affinity:]):
 159         (-[WebCoreBridge selectionAffinity]):
 160         (-[WebCoreBridge setMarkDOMRange:]):
 161         (-[WebCoreBridge _visiblePositionForPoint:]):
 162         (-[WebCoreBridge moveDragCaretToPoint:]):
 163         (-[WebCoreBridge editableDOMRangeForPoint:]):
 164         (-[WebCoreBridge ensureSelectionVisible]):
 165         (-[WebCoreBridge rangeOfCharactersAroundCaret]):
 166
 167 2005-02-09  Chris Blumenberg  <cblu@apple.com>
 168
 169         Fixed: <rdar://problem/3985211> Seed: Mail: Drag-and-drop destination indicator / insertion point disappears
 170
 171         Reviewed by kocienda.
 172
 173         * khtml/rendering/render_block.cpp:
 174         (khtml::RenderBlock::paintCaret): new
 175         (khtml::RenderBlock::paintObject): call paintCaret for the cursor caret and the drag caret
 176         * khtml/rendering/render_block.h:
 177
 178 === Safari-185 ===
 179
 180 2005-02-08  Ken Kocienda  <kocienda@apple.com>
 181
 182         Reviewed by John
 183
 184         Fix for this bug:
 185
 186         <rdar://problem/3996344> Entire document content is deleted when only the first paragraph was supposed to be
 187
 188         Problem stems from my attempt to fix this bug: <rdar://problem/3986155> Insertion point goes
 189         to beginning of doc after deleting.
 190
 191         I added an "optimization" that proved to do more harm than good. So, basically, the fix is to remove code.
 192
 193         * khtml/editing/htmlediting.cpp:
 194         (khtml::DeleteSelectionCommand::handleGeneralDelete)
 195
 196 2005-02-07  Darin Adler  <darin@apple.com>
 197
 198         Reviewed by Kevin.
 199
 200         - fixed <rdar://problem/3890352> REGRESSION (125-170, Panther-only): pasted textarea contents disappear (exception in NSText)
 201
 202         * kwq/KWQTextArea.mm: (-[KWQTextArea tile]): Don't resize the text view immediately here.
 203         Use performWithDelay to defer it to avoid the re-entrancy that bothers the Panther version
 204         of NSTextView. On Tiger, leave the code alone.
 205
 206 2005-02-07  Adele Amchan  <adele@apple.com>
 207
 208         Reviewed by Darin.
 209
 210         Fix for <rdar://problem/3993628> REGRESSION (180-181): Logitelnet bank's website doesn't work
 211
 212         This fixes an oversight in the fix for <rdar://problem/3964286>.
 213         We're now making sure there is a document before calling checkCompleted in slotLoaderRequestDone.
 214         That function is called even when other parts finish loads of their subresources, so we could end
 215         up calling checkCompleted before we had a document. We also looked over all the other calls to
 216         checkCompleted to convince ourselves no other call sites had a similar issue.
 217
 218         * khtml/khtml_part.cpp: (KHTMLPart::slotLoaderRequestDone): Check if doc is nil before calling
 219         checkCompleted.
 220
 221 2005-02-07  Ken Kocienda  <kocienda@apple.com>
 222
 223         Reviewed by Darin
 224
 225         Fix for this bug:
 226
 227         <rdar://problem/3953302> Replacing quoted text ends up with blue-colored text that is not quoted
 228
 229         * khtml/editing/htmlediting.cpp:
 230         (khtml::DeleteSelectionCommand::doApply): Removed special case delete code for select all. It was
 231         causing as many bugs as it fixed. What's more, the design concept is a bit shaky. By removing
 232         this special case, the bug goes away.
 233         * khtml/editing/htmlediting.h: Remove declaration.
 234
 235         Test results using selectAll() updated to reflect changed behavior.
 236
 237         * layout-tests/editing/deleting/delete-3775172-fix-expected.txt:
 238         * layout-tests/editing/deleting/delete-select-all-001-expected.txt:
 239         * layout-tests/editing/deleting/delete-select-all-003-expected.txt:
 240         * layout-tests/editing/inserting/insert-3654864-fix-expected.txt:
 241         * layout-tests/editing/inserting/insert-3775316-fix-expected.txt:
 242         * layout-tests/editing/inserting/insert-3851164-fix-expected.txt:
 243         * layout-tests/editing/inserting/insert-3907422-fix-expected.txt:
 244         * layout-tests/editing/pasteboard/paste-text-015-expected.txt:
 245
 246 2005-02-06  Darin Adler  <darin@apple.com>
 247
 248         Reviewed by Ken.
 249
 250         - fixed <rdar://problem/3994164> setting a new style attribute with setAttribute("style") doesn't blow away the old style
 251
 252         * khtml/css/css_valueimpl.h: Renamed parseProperty to parseDeclaration, since it's used to parse
 253         an entire style declaration, not a single property.
 254         * khtml/css/css_valueimpl.cpp: (DOM::CSSMutableStyleDeclarationImpl::parseDeclaration): Renamed,
 255         and added code to clear m_values.
 256         * khtml/html/html_elementimpl.cpp: (HTMLElementImpl::parseHTMLAttribute): Call function by its
 257         new name, and remove some unnecessary comments.
 258
 259 2005-02-06  Darin Adler  <darin@apple.com>
 260
 261         Reviewed by Maciej.
 262
 263         - fixed <rdar://problem/3993722> need oncut and onpaste handlers implemented for <textarea> (needed for widgets)
 264
 265         * kwq/KWQTextArea.mm:
 266         (-[KWQTextAreaTextView dispatchHTMLEvent:]): Added.
 267         (-[KWQTextAreaTextView cut:]): Dispatch "before cut" and "cut" events.
 268         (-[KWQTextAreaTextView copy:]): Dispatch "before copy" and "copy" events.
 269         (-[KWQTextAreaTextView paste:]): Dispatch "before paste" and "paste" events.
 270         (-[KWQTextAreaTextView pasteAsPlainText:]): Ditto.
 271         (-[KWQTextAreaTextView pasteAsRichText:]): Ditto.
 272
 273 2005-02-06  Darin Adler  <darin@apple.com>
 274
 275         Reviewed by Maciej.
 276
 277         - fixed <rdar://problem/3943038> <input type=search> that is focused in onload handler doesn't have a visible editor
 278
 279         * khtml/xml/dom_docimpl.cpp: (DocumentImpl::setFocusNode): Always update layout before giving focus to a widget.
 280         This prevents the bad case where we give a widget focus before it has been positioned or sized, causing us to
 281         exercise edge cases AppKit doesn't handle well.
 282
 283         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Remove workaround code I added
 284         to fix bug 3943049. Updating the layout fixes the same problem in a cleaner, safer way, and solves another problem,
 285         with the way the clip view is set up.
 286
 287 2005-02-06  Darin Adler  <darin@apple.com>
 288
 289         Reviewed by Maciej.
 290
 291         - fixed <rdar://problem/3425232> textarea won't trigger onchange action when clicking to new field
 292
 293         * khtml/rendering/render_form.h: Added new m_dirty data member.
 294         * khtml/rendering/render_form.cpp:
 295         (RenderTextArea::RenderTextArea): Initialize m_dirty to false.
 296         (RenderTextArea::detach): Call value() on the DOM element to cause it to fetch the value from this object.
 297         (RenderTextArea::handleFocusOut): Ditto. Also send the onchange event based only on the m_dirty flag.
 298         (RenderTextArea::updateFromElement): Clear the m_dirty flag if the edited value is blown away by a value
 299         that comes from the DOM element. Remove code to clear the element's m_dirtyvalue flag; that's an unnecessary
 300         optimization, and ideally we'll stop having code in the render object that knows about that flag later.
 301         (RenderTextArea::slotTextChanged): Set the m_dirty flag.
 302
 303 2005-02-06  Darin Adler  <darin@apple.com>
 304
 305         Reviewed by Maciej.
 306
 307         - re-fixed <rdar://problem/3760910> please add support for custom tag names in HTML
 308
 309         * khtml/xml/dom_docimpl.cpp:
 310         (DocumentImpl::tagId): Use ID_LAST_TAG + 1 for the ID of the first assigned tag ID. Before we'd use ID_LAST_TAG,
 311         which resulted in the same number being used for ID_COMMENT and the first custom tag.
 312         (DocumentImpl::tagName): Use ID_LAST_TAG + 1, as above.
 313         (DocumentImpl::namespaceURI): Change a < ID_LAST_TAG to <= ID_LAST_TAG in the same spirit as above.
 314
 315 2005-02-06  Darin Adler  <darin@apple.com>
 316
 317         Reviewed by Maciej.
 318
 319         - fixed <rdar://problem/3986639> Crash occurs after choosing Undo Typing from the Edit menu
 320
 321         * kwq/KWQTextArea.mm: (-[KWQTextArea viewWillMoveToWindow:]): Remove actions from undo manager when the view is
 322         removed from the window. Unfortunately, to do this the code has to know that the target of NSTextView actions
 323         will be the text storage object, but given the NSText architecture, that's pretty clear.
 324
 325 2005-02-05  Chris Blumenberg  <cblu@apple.com>
 326
 327         Fixed: <rdar://problem/3991974> REGRESSION: www.jabra.com world location screen does not work
 328
 329         Reviewed by hyatt.
 330
 331         * kwq/KWQKJavaAppletWidget.mm:
 332         (KJavaAppletWidget::KJavaAppletWidget): fallback to the document's base URL if there is no baseURL attribute
 333         * kwq/WebCoreBridge.h:
 334         * kwq/WebCoreBridge.mm:
 335         (-[WebCoreBridge baseURL]): new
 336
 337 2005-02-04  David Harrison  <harrison@apple.com>
 338
 339         Reviewed by Darin.
 340
 341         Rolled out Ken's accidental checkin when committing <rdar://problem/3986155> fix yesterday.
 342
 343         * khtml/editing/selection.cpp:
 344         (khtml::Selection::validate):
 345
 346 2005-02-04  Adele Amchan  <adele@apple.com>
 347
 348         Reviewed by Darin.
 349
 350         Fix for <rdar://problem/3758785> Safari no longer works at BankOfAmerica online banking for military users
 351
 352         When closing the document, if a body doesn't exist we now create one.  This bug was preventing the onload handler from
 353         firing in cases where there wasn't a body.  In the BankOfAmerica case, the page was just a bit of script where the onload
 354         handler redirected to the correct page.
 355
 356         * khtml/xml/dom_docimpl.cpp: (DocumentImpl::close):
 357
 358 2005-02-03  Ken Kocienda  <kocienda@apple.com>
 359
 360         Reviewed by Chris
 361
 362         Fix for this bug:
 363
 364         <rdar://problem/3986155> Insertion point goes to beginning of doc after deleting
 365
 366         * khtml/editing/htmlediting.cpp:
 367         (khtml::DeleteSelectionCommand::setStartNode): New convenience to handle reference counting when setting.
 368         (khtml::DeleteSelectionCommand::handleGeneralDelete): This contains the crux of the bug fix. Improve tests
 369         that detect when a selected node needs to be retained, rather than deleted, to preserve the intent of the user.
 370         This has the side effect of causing the insertion point placement code to succeed rather than fail. Before
 371         this fix, the failure of the insertion point placement code caused the insertion point to jump to the start
 372         of the document, which is the symptom that can be perceived by users when editing.
 373         * khtml/editing/htmlediting.h: Add setStartNode declaration.
 374         * khtml/editing/visible_units.cpp:
 375         (khtml::startOfBlock): This function had a stubbed-in non-tested implementation. Implement and
 376         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-001-expected.txt: Added.
 377         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-001.html: Added.
 378         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002-expected.txt: Added.
 379         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002.html: Added.
 380         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-003-expected.txt: Added.
 381         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-003.html: Added.
 382         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-004-expected.txt: Added.
 383         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-004.html: Added.
 384         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-005-expected.txt: Added.
 385         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-005.html: Added.
 386         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-006-expected.txt: Added.
 387         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-006.html: Added.
 388
 389 === Safari-183 ===
 390
 391 2005-02-03  Richard Williamson   <rjw@apple.com>
 392
 393         Fixed <rdar://problem/3987419> Stocks and Weather leak what appears to XMLHTTPRequest results
 394
 395         XMLHTTPRequests were causing massive leaks.  _webcore_initWithHeaderString: did funky things with
 396         self replacement.  Re-wrote to use more traditional factory constructor avoiding self replacement.
 397
 398         Reviewed by David Harrison.
 399
 400         * kwq/KWQLoader.mm:
 401         (+[NSDictionary _webcore_dictionaryWithHeaderString:]):
 402         (KWQServeRequest):
 403         (KWQServeSynchronousRequest):
 404
 405 2005-02-03  Chris Blumenberg  <cblu@apple.com>
 406
 407         Fixed: <rdar://problem/3938763> Cannot view Windows Media Player videos (soundtrack is played twice with delay)
 408
 409         Reviewed by hyatt.
 410
 411         * khtml/html/html_objectimpl.cpp:
 412         (HTMLObjectElementImpl::attach): set needWidgetUpdate to false right after calling updateWidget because dispatchHTMLEvent can end up calling this method again
 413         (HTMLObjectElementImpl::recalcStyle): ditto
 414
 415 2005-02-02  John Sullivan  <sullivan@apple.com>
 416
 417         Reviewed by Chris.
 418
 419         - WebCore part of fix for <rdar://problem/3980651> REGRESSION (125-180): Huge number of pages printed from certain page,
 420         iFrame involved
 421
 422         This also fixes the problems with printing from GMail, yay!
 423
 424         * khtml/rendering/render_flow.cpp:
 425         (RenderFlow::paintLines):
 426         If the current line is taller than the entire page height (e.g. tall iFrame), don't try
 427         to avoid splitting it across pages.
 428
 429         * kwq/WebCoreBridge.mm:
 430         (-[WebCoreBridge computePageRectsWithPrintWidthScaleFactor:printHeight:]):
 431         Handle error cases in a more obvious manner; this will cause future problems like this
 432         to print a blank page and complain to the console on debug builds, rather than print a
 433         zillionty mostly-blank pages.
 434
 435 2005-02-02  Chris Blumenberg  <cblu@apple.com>
 436
 437         Fixed: <rdar://problem/3960304> can't load a particular applet (at www.escape.de) unless it's the first applet to be loaded
 438
 439         Reviewed by hyatt.
 440
 441         * khtml/rendering/render_frames.cpp:
 442         (RenderPartObject::updateWidget): when getting the MIME type from the PARAM tag, make sure the MIME type is the text from the left of the semi-colon if there is one. We do this elsewhere as well.
 443
 444 2005-02-02  Chris Blumenberg  <cblu@apple.com>
 445
 446         Fixed: <rdar://problem/3983628> control-click on WebView is not selecting the word under the cursor (Mail, non-editable WebView)
 447
 448         Reviewed by rjw.
 449
 450         * kwq/KWQKHTMLPart.mm:
 451         (KWQKHTMLPart::sendContextMenuEvent): check if SPI to always enable selecting closest word is enabled
 452         * kwq/WebCoreBridge.h:
 453
 454 2005-02-02  Ken Kocienda  <kocienda@apple.com>
 455
 456         Reviewed by me
 457
 458         * khtml/editing/markup.cpp:
 459         (khtml::createMarkup): Removed debug spam I committed earlier in error.
 460
 461 2005-02-02  Ken Kocienda  <kocienda@apple.com>
 462
 463         Reviewed by Darin
 464
 465         Fix for this bug:
 466
 467         <rdar://problem/3823109> WebKit should support -toggleBaseWritingDirection: (bidi editing support)
 468
 469         Since base writing direction is a paragraph-level property in AppKit, and we use the CSS direction
 470         property in WebCore, which can be applied to inline elements as well as blocks, a new notion has
 471         been added to the WebCore apply style logic. It is now possible to force all properties in a
 472         style declaration to be applied as though they were block properties.
 473
 474         * khtml/editing/htmlediting.cpp:
 475         (khtml::ApplyStyleCommand::ApplyStyleCommand): Add an enum and a flag to this command that controls whether to force all
 476         properties in the style to be applied as block styles. Default is "no", retaining previous behavior.
 477         (khtml::ApplyStyleCommand::doApply): Switch on new flag to apply styles as before, or force all preoperties to be applied
 478         as block styles.
 479         * khtml/editing/htmlediting.h:
 480         (khtml::ApplyStyleCommand::): Add an enum and a flag, as above.
 481         * khtml/khtml_part.cpp:
 482         (KHTMLPart::applyParagraphStyle): New method to force application of all style properties as block styles.
 483         * khtml/khtml_part.h: Update header declarations.
 484         * kwq/KWQKHTMLPart.h: Update header declarations.
 485         * kwq/KWQKHTMLPart.mm:
 486         (KWQKHTMLPart::baseWritingDirectionForSelectionStart): Accessor to help WebKit do a toggle operation on
 487         writing direction.
 488         * kwq/WebCoreBridge.h: Update header declarations.
 489         * kwq/WebCoreBridge.mm:
 490         (-[WebCoreBridge applyParagraphStyle:withUndoAction:]): New method to force application of all style properties as block styles.
 491         (-[WebCoreBridge baseWritingDirectionForSelectionStart]): Accessor to help WebKit do a toggle operation on
 492         writing direction.
 493
 494         Fix for this bug:
 495
 496         <rdar://problem/3985035> REGRESSION (Mail): Text copied from wrapped line contains extra character when pasted
 497
 498         * khtml/editing/markup.cpp:
 499         (khtml::renderedText): Add an enum and a flag to this command that controls whether to force all
 500         This is the result of an error in the code that computes the rendered text that is selected
 501         when copying. Since spaces collapse at the end of lines, and these spaces need to be copied
 502         when the selection spans line endings, code runs to compute this text. However, this code
 503         was also running incorrectly in cases where lines wrapped. I have now added the missing
 504         test to check that the selection does indeed extend to the end of the line.
 505
 506
 507 2005-02-02  Ken Kocienda  <kocienda@apple.com>
 508
 509         Reviewed by Darin
 510
 511         Fix for these bugs:
 512
 513         <rdar://problem/3984894> REGRESSION (Mail): Command-right-arrow in reply does the wrong thing (two ways)
 514         <rdar://problem/3985130> REGRESSION (Mail): command-right-arrow in pasted RTF selects only up to tab
 515
 516         * khtml/editing/visible_units.cpp:
 517         (khtml::endOfLine): Two separate problems in this new function recently added to take the place of
 518         selectionForLine. In the first, endOfLine did not stop at BR elements, and returned the position
 519         beyond them. In the second, the VisiblePosition constructor gives the wrong answer in certain cases.
 520         Darin is going to work on a solution for that problem. In the meantime, I can fix the symptom of this
 521         bug by going down to the last leaf child of the root line box; a one line change we will roll out when
 522         Darin's fix is in.
 523
 524 2005-02-01  Richard Williamson   <rjw@apple.com>
 525
 526         Fixed <rdar://problem/3985535> QT Plug-in JavaScript support now fails in <embed> tag only case
 527
 528         Reviewed by Maciej.
 529
 530         * khtml/ecma/kjs_html.cpp:
 531         (KJS::HTMLDocument::tryGet):
 532
 533 2005-02-01  John Sullivan  <sullivan@apple.com>
 534
 535         Written by Darin, reviewed and tested by me
 536
 537         - fixed <rdar://problem/3969684> Panther-only: extra blank lines between
 538         pasted content from Blot
 539
 540         * kwq/KWQKHTMLPart.mm:
 541         (KWQKHTMLPart::attributedString):
 542         don't emit paragraph breaks if the margins are tiny; matches what
 543         we do for plain-text conversion
 544
 545 === Safari-182 ===
 546
 547 2005-02-01  Ken Kocienda  <kocienda@apple.com>
 548
 549         Reviewed by John
 550
 551         Fix for this bug:
 552
 553         <rdar://problem/3985160> Deficiencies in pasting architecture blocking progress on other bugs
 554
 555         * khtml/editing/html_interchange.h: Move style span text used to mark element added to
 556         add style to this header.
 557         * khtml/editing/htmlediting.cpp:
 558         (khtml::styleSpanClassString): Change to use constant moved to html_interchange.h.
 559         (khtml::isStyleSpan): New helper function. Checks if this is a span we added to apply style.
 560         (khtml::CompositeEditCommand::insertNodeBefore): Added an assert to check that the node
 561         we are inserting before is not the body.
 562         (khtml::CompositeEditCommand::insertNodeAfter): Ditto, but check is for after.
 563         (khtml::ReplacementFragment::ReplacementFragment): Added code to process the "default style"
 564         that is added by the copy code.
 565         (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): first and last nodes inserted are
 566         now member variables instead of function locals. Initialize them here.
 567         (khtml::ReplaceSelectionCommand::~ReplaceSelectionCommand): Deref first and last nodes inserted
 568         if necessary.
 569         (khtml::ReplaceSelectionCommand::doApply): Change design to fix the bug. Major change is to
 570         separate out the code that inserts nodes into the tree so additional styling checks can
 571         be done in a centralized way. Also got rid of the notion of "merging into the end block." That
 572         concept was just wrong.
 573         (khtml::ReplaceSelectionCommand::completeHTMLReplacement): Tweak interface now that first and
 574         last nodes inserted are member variables.
 575         (khtml::ReplaceSelectionCommand::insertNodeAfterAndUpdateNodesInserted): New helper used
 576         by replace code to do the stated DOM operation and update state internal to the command.
 577         This will also be a catch point to handle the kinds of additional style checks needed to
 578         make paste work right.
 579         (khtml::ReplaceSelectionCommand::insertNodeAtAndUpdateNodesInserted): Ditto.
 580         (khtml::ReplaceSelectionCommand::insertNodeBeforeAndUpdateNodesInserted): Ditto.
 581         (khtml::ReplaceSelectionCommand::updateNodesInserted): Ditto.
 582         * khtml/editing/htmlediting.h: Update declarations as needed.
 583         * khtml/editing/markup.cpp:
 584         (khtml::createMarkup): Adds a "default style" span to the content written to the pasteboard.
 585         This will help us to fix some of the bugs blocked by the bug above.
 586         * khtml/xml/dom_nodeimpl.cpp:
 587         (NodeImpl::lastDescendent): New helper.
 588         * khtml/xml/dom_nodeimpl.h: Ditto.
 589         * khtml/xml/dom_position.cpp:
 590         (DOM::Position::upstream): Fixed a bug which would allow the upstream position returned to be
 591         in unrendered content.
 592         (DOM::Position::downstream): Ditto.
 593
 594         * layout-tests/editing/deleting/delete-3775172-fix-expected.txt: Ending positions tweaked due to
 595         changes in upstream() and downstream() functions.
 596         * layout-tests/editing/inserting/insert-3851164-fix-expected.txt: Ditto
 597         * layout-tests/editing/inserting/insert-3907422-fix-expected.txt: Ditto
 598         * layout-tests/editing/selection/extend-by-character-006-expected.txt: Ditto
 599
 600         * layout-tests/editing/pasteboard/paste-text-003-expected.txt: Changed what we expect, given
 601         new behavior of paste code.
 602
 603 2005-01-31  John Sullivan  <sullivan@apple.com>
 604
 605         Reviewed by Dave Hyatt.
 606
 607         - fixed <rdar://problem/3983097> Tabbing on RSS pages gets stuck in search field
 608         with "full keyboard navigation" off
 609
 610         * kwq/KWQSlider.h:
 611         * kwq/KWQSlider.mm:
 612         (QSlider::focusPolicy):
 613         Implemented this method for KWQSlider. It is needed for our widget subclasses that
 614         can become focused in some situations. I missed this when fixing 3949203 because
 615         it only comes into play when tab-to-controls is on, and because the other relevant
 616         methods are objective-C methods.
 617
 618 2005-01-31  David Harrison  <harrison@apple.com>
 619
 620         Reviewed by Darin.
 621
 622         <rdar://problem/3964164> AXWebArea needs to unregister on going invisible
 623
 624         * khtml/xml/dom_docimpl.cpp:
 625         (DocumentImpl::detach):
 626         Detach the AX UIElement for the AXWebArea.
 627
 628 2005-01-31  David Harrison  <harrison@apple.com>
 629
 630         Reviewed by Darin
 631
 632         <rdar://problem/3850876> AX: button exposed as only an AXImage at http://appleseed.apple.com/ site (<input type=image>)
 633
 634         * khtml/rendering/render_form.h:
 635         (khtml::RenderImageButton::isImageButton):
 636         New.
 637         * khtml/rendering/render_image.h:
 638         (khtml::RenderImage::isImageButton):
 639         New.
 640         * kwq/KWQAccObject.mm:
 641         (-[KWQAccObject isImageButton]):
 642         (-[KWQAccObject actionElement]):
 643         (-[KWQAccObject role]):
 644         (-[KWQAccObject accessibilityActionNames]):
 645         (-[KWQAccObject accessibilityPerformAction:]):
 646         Represent RenderImageButton elements as buttons instead of images.
 647
 648 2005-01-31  David Harrison  <harrison@apple.com>
 649
 650         Prevent nil dereference in debug code.
 651
 652         * khtml/xml/dom_nodeimpl.cpp:
 653         (NodeImpl::displayTree):
 654         Stop when parent is 0, too.
 655
 656 2005-01-31  John Sullivan  <sullivan@apple.com>
 657
 658         Reviewed by Dave Hyatt.
 659
 660         - fixed <rdar://problem/3949203> cannot tab to, within, or out of the RSS sidebar
 661
 662         * kwq/KWQSlider.mm:
 663         (-[KWQSlider becomeFirstResponder]):
 664         (-[KWQSlider resignFirstResponder]):
 665         (-[KWQSlider nextKeyView]):
 666         (-[KWQSlider previousKeyView]):
 667         (-[KWQSlider canBecomeKeyView]):
 668         (-[KWQSlider nextValidKeyView]):
 669         (-[KWQSlider previousValidKeyView]):
 670         Copied these methods from KWQButton. The lack of special handling for first-responder-ness
 671         here was causing the focus to get "stuck" on the RSS page's slider widget. Added FIXME
 672         about how it would be nice to share more of this code rather than replicating it in each
 673         KWQ widget subclass.
 674
 675 2005-01-31  Darin Adler  <darin@apple.com>
 676
 677         Reviewed by Harrison.
 678
 679         - fixed <rdar://problem/3980066> Double-click on single character moves insertion point to previous line
 680
 681         * khtml/khtml_part.cpp:
 682         (KHTMLPart::selectClosestWordFromMouseEvent): Set affinity too.
 683         (KHTMLPart::handleMousePressEventTripleClick): Ditto.
 684         (KHTMLPart::handleMouseMoveEventSelection): Ditto.
 685         (KHTMLPart::khtmlMouseReleaseEvent): Ditto.
 686
 687         * khtml/editing/selection.cpp:
 688         (khtml::Selection::modifyExtendingRightForward): Use endOfLine and endOfDocument.
 689         (khtml::Selection::modifyMovingRightForward): Ditto.
 690         (khtml::Selection::modifyExtendingLeftBackward): Use startOfLine and startOfDocument.
 691         (khtml::Selection::modifyMovingLeftBackward): Ditto.
 692         (khtml::Selection::validate): Rewrote the section that handles double-click. Two main fixes: 1) use isStartOfLine to
 693         check for another case where we want to select the word to the right, and 2) use isEndOfParagraph, which seems
 694         to work correctly in cases where isLastVisiblePositionInParagraph is giving the wrong answer. Also changed the line
 695         code to use startOfLine/endOfLine and the document code to use startOfDocument/endOfDocument.
 696
 697 2005-01-31  Darin Adler  <darin@apple.com>
 698
 699         Reviewed by Harrison.
 700
 701         - fixed <rdar://problem/3935275> unexpected quit scrolling over link; last.fm (works in IE and Firefox)
 702
 703         * kwq/WebCoreBridge.mm: (-[WebCoreBridge elementAtPoint:]): By using the correct document, fix the crash reported here.
 704         The page still crashes when you close the browser window, but it's not trivial to fix so I'll file a new bug after that.
 705
 706 2005-01-31  Darin Adler  <darin@apple.com>
 707
 708         Reviewed by Ken and Harrison.
 709
 710         - fixed <rdar://problem/3947901> REGRESSION (Mail): Pasting paragraph of rich text leaves insertion point before pasted text
 711         - fixed <rdar://problem/3949790> hitting return after underlined line results in too much or too little underlined
 712         - fixed <rdar://problem/3981759> nil-deref and crash when pasting just a paragraph break
 713         - fixed a couple problems I discovered while working with bug 3949790
 714
 715         * khtml/editing/htmlediting.cpp:
 716         (khtml::ApplyStyleCommand::applyInlineStyle): Pass StayInBlock to upstream. Without this, we end up going too far
 717         upstream in the test case in bug 3949790.
 718         (khtml::ApplyStyleCommand::removeInlineStyle): Pass StayInBlock to upstream and downstream. Same reason as above.
 719         (khtml::ReplaceSelectionCommand::doApply): Update endPos if inserting a new node and endPos is using that node's
 720         parent and an offset past the node being inserted. That change fixes a problem with the position of the insertion point
 721         after pasting into the top level of a document (from test cases in 3947901 and 3949790). When setting insertionPos, use
 722         code that works when lastNodeInserted is a block rather than a text node. That change fixes a problem where a newline is
 723         not added when pasting an entire paragraph into the end of a document (from test case in 3949790). Added nil check before
 724         checking if lastNodeInserted is a <br> element, which fixes the crash when pasting just a paragraph break.
 725
 726         * khtml/editing/visible_units.h: Filled out the set of calls to add some boolean checks for lines (needed for the
 727         bug fix), and calls for blocks (not yet implemented), and documents. The document checks may need refinement to
 728         properly handle documents with a mix of editable and non-editable content, but for now they just refactor code
 729         and make things a little clearer. Also removed the "include line break" parameter from endOfSentence.
 730         * khtml/editing/visible_units.cpp:
 731         (khtml::rootBoxForLine): Added.
 732         (khtml::startOfLine): Added. Algorithm taken from selectionForLine in selection.cpp.
 733         (khtml::endOfLine): Ditto.
 734         (khtml::inSameLine): Added.
 735         (khtml::isStartOfLine): Added.
 736         (khtml::isEndOfLine): Added.
 737         (khtml::endOfSentence): Removed "include line break" parameter.
 738         (khtml::inSameParagraph): Added a null check.
 739         (khtml::isStartOfParagraph): Ditto.
 740         (khtml::isEndOfParagraph): Ditto.
 741         (khtml::startOfBlock): Added.
 742         (khtml::endOfBlock): Added.
 743         (khtml::inSameBlock): Added.
 744         (khtml::isStartOfBlock): Added.
 745         (khtml::isEndOfBlock): Added.
 746         (khtml::startOfDocument): Added.
 747         (khtml::endOfDocument): Added.
 748         (khtml::inSameDocument): Added.
 749         (khtml::isStartOfDocument): Added.
 750         (khtml::isEndOfDocument): Added.
 751
 752 2005-01-30  Darin Adler  <darin@apple.com>
 753
 754         Reviewed by John.
 755
 756         - fixed <rdar://problem/3977000> form data set posted for <input type="image" ...> omits name/value pair
 757
 758         * khtml/html/html_formimpl.cpp:
 759         (DOM::HTMLInputElementImpl::appendFormData): Append a name/value pair for type IMAGE in addition to the X/Y
 760         coordinates.
 761
 762 2005-01-30  Darin Adler  <darin@apple.com>
 763
 764         Reviewed by John.
 765
 766         - fixed <rdar://problem/3974246> REGRESSION (125-180): popup menus are missing item text on Harmony Remote web site
 767
 768         * khtml/html/html_formimpl.cpp: (DOM::HTMLOptionElementImpl::text): Change this function to traverse the entire tree
 769         and gather all the text rather than just looking at immediate children.
 770
 771         * khtml/xml/dom_nodeimpl.h: Added const.
 772         * khtml/xml/dom_nodeimpl.cpp:
 773         (NodeImpl::traverseNextNode): Added const.
 774         (NodeImpl::traverseNextSibling): Ditto.
 775         (NodeImpl::traversePreviousNodePostOrder): Ditto.
 776         (NodeImpl::detach): Add a missing nil check.
 777
 778 2005-01-30  Darin Adler  <darin@apple.com>
 779
 780         Reviewed by John.
 781
 782         - fixed <rdar://problem/3864536> crash copying text from other apps and pasting in a sticky in dashboard
 783
 784         * khtml/editing/markup.cpp: (khtml::createFragmentFromText): Added missing ref/deref of newly created element.
 785
 786 2005-01-28  Darin Adler  <darin@apple.com>
 787
 788         * khtml/css/cssproperties.c: Regenerated with newer gperf.
 789
 790 2005-01-28  David Harrison  <harrison@apple.com>
 791
 792         Reviewed by John Sullivan.
 793
 794         <rdar://problem/3968144> AX need to be able to focus an AXLink by setting AXFocused to true
 795
 796         * kwq/KWQAccObject.mm:
 797         (-[KWQAccObject accessibilityPerformAction:]):
 798         (-[KWQAccObject accessibilityIsAttributeSettable:]):
 799         Add AXFocused for AXLinks.
 800
 801 2005-01-28  David Harrison  <harrison@apple.com>
 802
 803         Reviewed by Darin.
 804
 805         <rdar://problem/3584942> AX: Safari Accessibility parent-child mismatch
 806
 807         Use AppKit SPI _accessibilityParentForSubview to return KWQAccObject parent of AppKit AX object.
 808
 809         * kwq/KWQAccObject.mm:
 810         (-[KWQAccObject _accessibilityParentForSubview:]):
 811         New.
 812
 813 2005-01-27  Adele Amchan  <adele@apple.com>
 814
 815         fixed by Darin, reviewed by me.
 816
 817         <rdar://problem/3976314> REGRESSION (180-TOT): submitting password fields fail on Panther
 818
 819         * kwq/KWQTextField.mm: (-[KWQTextFieldController string]): on panther, the secure text field's editor
 820         does not contain the real string, so now we just call stringValue on the field in that case.
 821
 822 2005-01-26  Maciej Stachowiak  <mjs@apple.com>
 823
 824         Reviewed by Darin, Hyatt and Ken.
 825
 826         <rdar://problem/3790449> REGRESSION (Mail): underline behavior is flaky because of how CSS handles text-decoration
 827
 828         * khtml/css/css_computedstyle.cpp:
 829         (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): Add support
 830         for -khtml-text-decorations-in-effect property. This works like text-docration
 831         but follows proper inline semantics, that is, if a parent is causing text
 832         decoration then -khtml-text-decorations-in-effect on this element will reflect that.
 833         * khtml/css/cssproperties.in: Add -khtml-text-decorations-in-effect property.
 834         * khtml/css/cssproperties.h: regenerated
 835         * khtml/css/cssvalues.c: regenerated
 836         * khtml/css/cssparser.cpp: regenerated
 837         * khtml/css/cssproperties.c: regenerated
 838
 839         * khtml/css/css_valueimpl.h: Made copyPropertiesInSet public.
 840
 841         * khtml/editing/htmlediting.cpp:
 842         (khtml::StyleChange::init): Handle -khtml-text-decorations-in-effect
 843         specially, translate to text-decoration
 844         (khtml::ApplyStyleCommand::applyInlineStyle): Fix a comment.
 845         (khtml::ApplyStyleCommand::removeInlineStyle): Add new way of pushing down
 846         text-decoration styles around the edges of the selected range.
 847         (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAroundNode): Helper for this.
 848         (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAtBoundaries): ditto
 849         (khtml::ApplyStyleCommand::nodeFullyUnselected): ditto
 850         (khtml::hasTextDecorationProperty): ditto
 851         (khtml::highestAncestorWithTextDecoration): ditto
 852         (khtml::ApplyStyleCommand::extractTextDecorationStyle): ditto
 853         (khtml::ApplyStyleCommand::extractAndNegateTextDecorationStyle): ditto
 854         (khtml::ApplyStyleCommand::applyTextDecorationStyle): ditto
 855         * khtml/editing/htmlediting.h: Prototype new methods.
 856         * khtml/editing/jsediting.cpp: Make underline execCommand use the new CSS property.
 857
 858         - as a necessary part of the fix, fixed the longstanding problem
 859         that mutating an element's inlineStyleDecl would not property
 860         update its style attribute. now it does.
 861
 862         * khtml/css/css_valueimpl.cpp:
 863         (DOM::CSSMutableStyleDeclarationImpl::setChanged): If this is an inline style
 864         declaration, tell the element it's style attribute needs updating.
 865         * khtml/html/html_elementimpl.cpp:
 866         (HTMLElementImpl::invalidateStyleAttribute): New method, just calls down
 867         to the dom element.
 868         (HTMLElementImpl::updateStyleAttribute): Regenerate style attribute from
 869         inline style declaration.
 870         * khtml/html/html_elementimpl.h: Prototype new stuff.
 871         * khtml/xml/dom_elementimpl.cpp:
 872         (ElementImpl::ElementImpl): Initialize new attribute.
 873         (ElementImpl::updateStyleAttributeIfNeeded): New method to check if the style
 874         is dirty and this is an html element, and if so call the html element to
 875         regenerate the style attribute.
 876         (ElementImpl::attributes): update style if needed
 877         (ElementImpl::getAttribute): ditto
 878         (ElementImpl::hasAttributes): ditto
 879         (ElementImpl::dump): ditto
 880         (XMLElementImpl::cloneNode): ditto
 881         * khtml/xml/dom_elementimpl.h:
 882         (DOM::ElementImpl::isStyleAttributeValid): new method to manage style attr
 883         dirty bit
 884         (DOM::ElementImpl::setStyleAttributeValid): ditto
 885
 886         - added new layout tests for underlining:
 887
 888         * layout-tests/editing/editing.js:
 889         * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt: Added.
 890         * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt: Added.
 891         * layout-tests/editing/style/remove-underline-across-paragraph-in-bold.html: Added.
 892         * layout-tests/editing/style/remove-underline-across-paragraph.html: Added.
 893         * layout-tests/editing/style/remove-underline-after-paragraph-expected.txt: Added.
 894         * layout-tests/editing/style/remove-underline-after-paragraph-in-bold-expected.txt: Added.
 895         * layout-tests/editing/style/remove-underline-after-paragraph-in-bold.html: Added.
 896         * layout-tests/editing/style/remove-underline-after-paragraph.html: Added.
 897         * layout-tests/editing/style/remove-underline-expected.txt: Added.
 898         * layout-tests/editing/style/remove-underline-from-stylesheet-expected.txt: Added.
 899         * layout-tests/editing/style/remove-underline-from-stylesheet.html: Added.
 900         * layout-tests/editing/style/remove-underline-in-bold-expected.txt: Added.
 901         * layout-tests/editing/style/remove-underline-in-bold.html: Added.
 902         * layout-tests/editing/style/remove-underline.html: Added.
 903         * layout-tests/editing/style/underline-expected.txt: Added.
 904         * layout-tests/editing/style/underline.html: Added.
 905
 906 2005-01-27  David Hyatt  <hyatt@apple.com>
 907
 908         Fix for 3875199, search field needs to be able to  show a magnifying  glass without a dropdown menu.
 909
 910         Reviewed by kocienda
 911
 912         * khtml/html/html_formimpl.cpp:
 913         (DOM::HTMLInputElementImpl::HTMLInputElementImpl):
 914         (DOM::HTMLInputElementImpl::parseHTMLAttribute):
 915         * kwq/KWQLineEdit.mm:
 916         (QLineEdit::setMaxResults):
 917
 918 2005-01-27  Ken Kocienda  <kocienda@apple.com>
 919
 920         Reviewed by Maciej
 921
 922         Fix for this bug:
 923
 924         <rdar://problem/3973254> Deletions of ranges does not coalesce correctly with subsequent typing
 925
 926         * khtml/editing/htmlediting.cpp:
 927         (khtml::InsertTextCommand::deleteCharacter): Remove this dead code.
 928         (khtml::ReplaceSelectionCommand::editingAction): Moved this code to a better place in the file.
 929         It was in with unrelated functions.
 930         (khtml::TypingCommand::TypingCommand): Reorganize initialization list so it is easier to read.
 931         (khtml::TypingCommand::issueCommandForDeleteKey): Remove dead code. Roll in remaining code into deleteKeyPressed.
 932         (khtml::TypingCommand::deleteKeyPressed): Add in support for carrying along smart delete flag.
 933         (khtml::TypingCommand::forwardDeleteKeyPressed): New function. Makes forward delete work like
 934         "regular" delete in terms of coalescing typing.
 935         (khtml::TypingCommand::doApply): Add case for ForwardDeleteKey.
 936         (khtml::TypingCommand::preservesTypingStyle): Ditto.
 937         * khtml/editing/htmlediting.h:
 938         (khtml::TypingCommand::): Add ForwardDeleteKey constant. Remove a couple declarations for now-dead code.
 939         (khtml::TypingCommand::smartDelete): New accessor.
 940         (khtml::TypingCommand::setSmartDelete): Ditto.
 941         * kwq/WebCoreBridge.h:
 942         * kwq/WebCoreBridge.mm:
 943         (-[WebCoreBridge deleteKeyPressedWithSmartDelete:]): Add smart delete flag.
 944         (-[WebCoreBridge forwardDeleteKeyPressedWithSmartDelete:]): New method.
 945         * khtml/editing/jsediting.cpp: Add supprt for ForwardDelete command
 946         * layout-tests/editing/editing.js: Ditto.
 947         * layout-tests/editing/deleting/delete-and-undo-expected.txt: Added.
 948         * layout-tests/editing/deleting/delete-and-undo.html: Added.
 949         * layout-tests/editing/deleting/forward-delete-expected.txt: Added.
 950         * layout-tests/editing/deleting/forward-delete.html: Added.
 951
 952 === Safari-181 ===
 953
 954 2005-01-27  David Harrison  <harrison@apple.com>
 955
 956         Reviewed by Ken.
 957
 958         <rdar://problem/3962214> AX: AXLeftLineTextMarkerRangeForTextMarker returns incorrect range
 959
 960         * kwq/KWQAccObject.mm:
 961         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
 962         Allowed for selection to end of line including the linebreak.
 963
 964 2005-01-27  David Harrison  <harrison@apple.com>
 965
 966         Reviewed by Darin, Ken.
 967
 968         <rdar://problem/3964470> AX: Include attachments in AXAttributedStringForTextMarkerRange
 969
 970         * kwq/KWQAccObject.mm:
 971         (-[KWQAccObject isAttachment]):
 972         (-[KWQAccObject attachmentView]):
 973         (-[KWQAccObject role]):
 974         (-[KWQAccObject subrole]):
 975         (-[KWQAccObject roleDescription]):
 976         (-[KWQAccObject value]):
 977         (-[KWQAccObject title]):
 978         (-[KWQAccObject accessibilityDescription]):
 979         (-[KWQAccObject accessibilityIsIgnored]):
 980         (-[KWQAccObject accessibilityAttributeNames]):
 981         (-[KWQAccObject accessibilityAttributeValue:]):
 982         Make attachments accessible.
 983
 984         (CreateCGColorIfDifferent):
 985         (AXAttributeStringSetColor):
 986         (AXAttributeStringSetNumber):
 987         (AXAttributeStringSetFont):
 988         (AXAttributeStringSetStyle):
 989         (AXAttributeStringSetElement):
 990         (AXLinkElementForNode):
 991         (AXAttributedStringAppendText):
 992         (AXAttributedStringAppendReplaced):
 993         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
 994         Add attachments and more text attributes to AXAttributedStringForTextMarkerRange.
 995
 996 2005-01-27  Darin Adler  <darin@apple.com>
 997
 998         Reviewed by John.
 999
1000         - fixed <rdar://problem/3807935> DOM CSS computed style line-height is wrong in two ways
1001
1002         * khtml/css/css_computedstyle.cpp: (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue):
1003         Use specified size rather than computed size when dealing with a percentage for line height.
1004
1005 2005-01-27  Darin Adler  <darin@apple.com>
1006
1007         Reviewed by John.
1008
1009         - fixed <rdar://problem/3971372> SWB: template function DOM_cast() won't compile with gcc 4.0
1010         - fixed other gcc 4.0 compiling problems Patrick Beard pointed out on the phone
1011
1012         * kwq/DOMInternal.h: (DOM_cast): Change the "failToCompile()" trick to use a dependent name.
1013         Nowadays, if a name is not dependent, it is checked at template definition time. But we want
1014         a failure only at template instantiation time.
1015
1016         * ForwardingHeaders/editing/text_granularity.h: Added. Needed but a bug in the 3.3 compiler made
1017         it compile anyway without this.
1018
1019         - other changes
1020
1021         * khtml/html/html_miscimpl.h: For clarity, mark these functions virtual too. They are automatically
1022         virtual because the base class ones are virtual, but it's ugly to leave it this way.
1023
1024 2005-01-26  Ken Kocienda  <kocienda@apple.com>
1025
1026         Reviewed by Hyatt
1027
1028         Fix for this bug:
1029
1030         <rdar://problem/3971609> REGRESSION (Mail): up/down arrow navigation broken after rewrapping text by resizing window
1031         <rdar://problem/3975661> REGRESSION (Mail): left/right arrow navigation can place insertion point after last character on line
1032
1033         * khtml/editing/selection.cpp:
1034         (khtml::Selection::modifyAffinity): PARAGRAPH and LINE movements should not alter affinity. This was
1035         just a mistake before when it reset to upstream in some cases.
1036         * khtml/rendering/render_line.cpp:
1037         (khtml::RootInlineBox::closestLeafChildForXPos): Move this function down from InlineBox, and
1038         rework the implementation. This fixes 3971609.
1039         * khtml/rendering/render_line.h:
1040         * khtml/rendering/render_text.cpp:
1041         (RenderText::caretRect): Rework the algorithm to fix 3975661. This regressed when we began placing
1042         the spaces on the end of lines in text boxes when lines wrap.
1043
1044 2005-01-26  Richard Williamson   <rjw@apple.com>
1045
1046         Fixed <rdar://problem/3972522> (179-180) 40% slowdown on iBench JavaScript test
1047
1048         I added a member variable to ObjectImp.  This changed it's size and consequently
1049         hampered the optimizations built into the garbage collector.  Objects no longer
1050         fit within the allocators cell size, and thus allocation fell back to a slower
1051         allocator.
1052
1053         As a result of this fix I also dramatically cleaned up how runtime objects are
1054         accessed.  The path mostly *removes* code.
1055
1056         Reviewed by Chris.
1057
1058         * khtml/ecma/kjs_dom.cpp:
1059         (DOMDocumentProtoFunc::tryCall):
1060         (DOMElementProtoFunc::tryCall):
1061         (KJS::getRuntimeObject):
1062         * khtml/ecma/kjs_dom.h:
1063         * khtml/ecma/kjs_html.cpp:
1064         (KJS::HTMLDocument::tryGet):
1065         (KJS::HTMLElement::tryGet):
1066         (KJS::HTMLElement::implementsCall):
1067         (KJS::HTMLElement::call):
1068         (KJS::HTMLElement::tryPut):
1069         (KJS::HTMLCollection::tryGet):
1070         (KJS::HTMLCollection::getNamedItems):
1071         * khtml/ecma/kjs_html.h:
1072         * khtml/ecma/kjs_window.cpp:
1073         (Window::get):
1074
1075 2005-01-26  Richard Williamson   <rjw@apple.com>
1076
1077         Fixed <rdar://problem/3757712> REGRESSION (Mail): WebCore does not allow Devanagari ligature input
1078
1079         For now we are using ICU UBreakIterator to determine grapheme boundaries for
1080         cursor and deletion.  This does not match what Cocoa does exactly, but does match
1081         what Carbon does.  The areas are difference are obscure, but, according to
1082         Deborah Goldsmith, using the UBreakIterator is the reasonable approach.
1083
1084         Reviewed by Ken.
1085
1086         * khtml/rendering/render_text.cpp:
1087         (RenderText::previousOffset):
1088         (RenderText::nextOffset):
1089
1090 2005-01-25  David Harrison  <harrison@apple.com>
1091
1092         Reviewed by Maciej and Richard.
1093
1094         <rdar://problem/3963731> AX VO: Changing window sizes causes a crash using with Voice Over - KWQPtrDictImpl::clear
1095
1096         * khtml/html/html_miscimpl.cpp:
1097         (HTMLCollectionImpl::HTMLCollectionImpl):
1098         Ensure that isHTMLDocument before calling HTMLDocument-only method.
1099
1100 2005-01-25  Richard Williamson   <rjw@apple.com>
1101
1102         Part one of fix for <rdar://problem/3757712> REGRESSION (Mail): WebCore does not allow Devanagari ligature input
1103
1104         This patch changes the semantics of next() and previous() on
1105         VisiblePosition to move a grapheme (character cluster) at a
1106         time.  This means that cursor navigation with correctly move
1107         over an entire cluster.
1108
1109         However, the expected behavior for deleting a grapheme is to
1110         delete individual code points, thus decomposing the grapheme
1111         into it constituent parts.  That will be addressed in the next
1112         part of the fix.
1113
1114         Reviewed by Ken.
1115
1116         * khtml/editing/visible_position.cpp:
1117         (khtml::VisiblePosition::previousPosition):
1118         (khtml::VisiblePosition::nextPosition):
1119         * khtml/editing/visible_position.h:
1120         * khtml/rendering/render_object.cpp:
1121         (RenderObject::previousOffset):
1122         (RenderObject::nextOffset):
1123         * khtml/rendering/render_object.h:
1124         * khtml/rendering/render_text.cpp:
1125         (RenderText::previousOffset):
1126         (RenderText::nextOffset):
1127         (RenderText::findNextInlineTextBox):
1128         * khtml/rendering/render_text.h:
1129         * khtml/xml/dom_nodeimpl.cpp:
1130         (NodeImpl::previousOffset):
1131         (NodeImpl::nextOffset):
1132         * khtml/xml/dom_nodeimpl.h:
1133
1134 2005-01-25  David Harrison  <harrison@apple.com>
1135
1136         Reviewed by Maciej.
1137
1138         <rdar://problem/3973067> AXWebArea for http://apple.netscape.com/ has an empty AXLinkUIElements
1139
1140         * kwq/KWQAccObject.mm:
1141         (-[KWQAccObject accessibilityAttributeValue:]):
1142         Omit ignored AXLinks from AXLinkUIElements result
1143
1144 2005-01-25  Ken Kocienda  <kocienda@apple.com>
1145
1146         Reviewed by John
1147
1148         Fix for this bug:
1149
1150         <rdar://problem/3972851> REGRESSION (179-180+): Repro crash in ApplyStyleCommand::cleanUpEmptyStyleSpans
1151
1152         * khtml/editing/htmlediting.cpp:
1153         (khtml::ApplyStyleCommand::cleanUpEmptyStyleSpans): Added two null checks.
1154
1155 2005-01-25  Ken Kocienda  <kocienda@apple.com>
1156
1157         Reviewed by John
1158
1159         Fix for this bug:
1160
1161         <rdar://problem/3954710> Mail crashed while editing signatures - NodeImpl::isBlockFlow
1162
1163         The fix is more general than for this one bug, and may work to fix many crashers. The problem
1164         is that the ReplaceSelectionCommand never checked whether its starting selection is empty. If
1165         it is, then we need to bail before doing the work of the command, which we need to deref the
1166         start and end points of the selection in order to do its work. I think you can see the crash
1167         potential.
1168
1169         * khtml/editing/htmlediting.cpp:
1170         (khtml::ReplaceSelectionCommand::doApply): Assert selection is not empty.
1171         * kwq/WebCoreBridge.mm:
1172         (partHasSelection): New helper function to test that bridge has a part with a selection.
1173
1174         Use new helper function to test part and selection; return from these function if this test fails.
1175
1176         (-[WebCoreBridge rangeByExpandingSelectionWithGranularity:])
1177         (-[WebCoreBridge rangeByAlteringCurrentSelection:direction:granularity:])
1178         (-[WebCoreBridge alterCurrentSelection:direction:granularity:])
1179         (-[WebCoreBridge rangeByAlteringCurrentSelection:verticalDistance:])
1180         (-[WebCoreBridge alterCurrentSelection:verticalDistance:])
1181         (-[WebCoreBridge documentFragmentWithText:])
1182         (-[WebCoreBridge replaceSelectionWithFragment:selectReplacement:smartReplace:])
1183         (-[WebCoreBridge insertLineBreak])
1184         (-[WebCoreBridge insertParagraphSeparator])
1185         (-[WebCoreBridge insertParagraphSeparatorInQuotedContent])
1186         (-[WebCoreBridge insertText:selectInsertedText:])
1187         (-[WebCoreBridge deleteSelectionWithSmartDelete:])
1188         (-[WebCoreBridge ensureSelectionVisible])
1189
1190 2005-01-24  Kevin Decker  <kdecker@apple.com>
1191
1192         Reviewed by Darin.
1193
1194         Fixed <rdar://problem/3932374> REGRESSION: 'ReferenceError - Can't find variable' JavaScript error at webxpress.fidelity.com
1195
1196         * khtml/html/html_miscimpl.h: Missing virtual identifiers for namedItem() and nextNamedItem() exposed a flaw that prevented finding variables in form [HTMLFormCollectionImpl] scopes.
1197
1198
1199 2005-01-24  Vicki Murley <vicki@apple.com>
1200
1201         Reviewed by kocienda.
1202
1203         - fix <rdar://problem/3810661> computed style not handled for Apple extensions
1204
1205         * khtml/css/css_computedstyle.cpp: (DOM::):
1206         (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): add -apple-line-clamp and -apple-text-size-adjust
1207
1208 2005-01-24  Darin Adler  <darin@apple.com>
1209
1210         Reviewed by Kevin.
1211
1212         - fixed <rdar://problem/3969884> REGRESSION (179-180): Typing password not echoed as "bullets" at paypal.com when field is first focused
1213
1214         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]):
1215         Sizing up the form editor, which fixes things for normal fields and search fields,
1216         screws things up for password fields. Eventually, I'll need to figure out why this is
1217         and file a bug so that the AppKit team fixes it. In the mean time, just check for the
1218         case of the secure text field, and don't size up in that case.
1219
1220 2005-01-24  Ken Kocienda  <kocienda@apple.com>
1221
1222         Reviewed by Hyatt
1223
1224         Fix for this bug:
1225
1226         <rdar://problem/3963560> 8A354: Inserting Japanese text to the line head causes line break
1227
1228         * khtml/editing/htmlediting.cpp:
1229         (khtml::ReplaceSelectionCommand::doApply): Add one more case where we merge content into the
1230         existing line. This covers the situation see using the steps to reproduce this bug.
1231
1232 2005-01-21  Maciej Stachowiak  <mjs@apple.com>
1233
1234         Reviewed by Darin.
1235
1236         <rdar://problem/3967572> Editing should split elements before removing style
1237
1238         * khtml/editing/htmlediting.cpp:
1239         (khtml::CompositeEditCommand::splitElement):
1240         (khtml::CompositeEditCommand::mergeIdenticalElements):
1241         (khtml::CompositeEditCommand::wrapContentsInDummySpan):
1242         (khtml::CompositeEditCommand::splitTextNodeContainingElement):
1243         (khtml::ApplyStyleCommand::applyInlineStyle):
1244         (khtml::ApplyStyleCommand::splitTextAtStartIfNeeded):
1245         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded):
1246         (khtml::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
1247         (khtml::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
1248         (khtml::areIdenticalElements):
1249         (khtml::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
1250         (khtml::ApplyStyleCommand::mergeEndWithNextIfIdentical):
1251         (khtml::ApplyStyleCommand::cleanUpEmptyStyleSpans):
1252         (khtml::ApplyStyleCommand::addInlineStyleIfNeeded):
1253         (khtml::InsertParagraphSeparatorCommand::doApply):
1254         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
1255         (khtml::InsertTextCommand::prepareForTextInsertion):
1256         (khtml::SplitTextNodeCommand::doUnapply):
1257         (khtml::SplitElementCommand::SplitElementCommand):
1258         (khtml::SplitElementCommand::~SplitElementCommand):
1259         (khtml::SplitElementCommand::doApply):
1260         (khtml::SplitElementCommand::doUnapply):
1261         (khtml::MergeIdenticalElementsCommand::MergeIdenticalElementsCommand):
1262         (khtml::MergeIdenticalElementsCommand::~MergeIdenticalElementsCommand):
1263         (khtml::MergeIdenticalElementsCommand::doApply):
1264         (khtml::MergeIdenticalElementsCommand::doUnapply):
1265         (khtml::WrapContentsInDummySpanCommand::WrapContentsInDummySpanCommand):
1266         (khtml::WrapContentsInDummySpanCommand::~WrapContentsInDummySpanCommand):
1267         (khtml::WrapContentsInDummySpanCommand::doApply):
1268         (khtml::WrapContentsInDummySpanCommand::doUnapply):
1269         (khtml::SplitTextNodeContainingElementCommand::SplitTextNodeContainingElementCommand):
1270         (khtml::SplitTextNodeContainingElementCommand::~SplitTextNodeContainingElementCommand):
1271         (khtml::SplitTextNodeContainingElementCommand::doApply):
1272         * khtml/editing/htmlediting.h:
1273
1274         New layout test that shows the side benefit of this.
1275
1276         * layout-tests/editing/style/unbold-in-bold-expected.txt: Added.
1277         * layout-tests/editing/style/unbold-in-bold.html: Added.
1278
1279 2005-01-24  Darin Adler  <darin@apple.com>
1280
1281         Reviewed by John.
1282
1283         - fixed <rdar://problem/3933435> -[DOMRange cloneContents] sometimes alters the original DOMRange
1284
1285         * khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::processContents): Fixed code that was collapsing
1286         the range after processing the contents. That's right for extract and delete, but not clone.
1287
1288         - fixed <rdar://problem/3714184> inline input works incorrectly when I type Korean slowly in textfield in form on www.google.co.kr
1289         - fixed <rdar://problem/3193848> Can't enter SSN to sign up for Chase visa online account (changing focus inside key press handler)
1290         - fixed <rdar://problem/3874683> REGRESSION (161-162): Crash after typing command-Z to undo after pasting text with newline in it into a text field <input type=text>
1291
1292         * kwq/KWQLineEdit.mm:
1293         (QLineEdit::text): Changed to call a new string method on the controller instead of calling
1294         stringValue, since stringValue has a bad side effect of ending inline input.
1295         (QLineEdit::selectAll): Changed to only call selectText: when the field is already selected.
1296         When you are giving the field focus, it automatically gets all selected, and this change
1297         is required in the case where we defer the focus change.
1298
1299         * kwq/KWQTextField.h: Added a new string method to the controller.
1300         * kwq/KWQTextField.mm:
1301         (-[KWQTextFieldController setMaximumLength:]): Call the new string method instead of using stringValue.
1302         (-[KWQTextFieldController controlTextDidChange:]): Removed the code to truncate at the first CR or LF.
1303         That's now handled in the shouldChangeTextInRange method instead.
1304         (-[KWQTextFieldController textView:shouldHandleEvent:]): Add new logic to defer responder changes until
1305         after the event is handled. This makes sure the key gets into the field that's already handling it at
1306         rather than ending up in the newly-focused field if the key press handler focuses a different element.
1307         (-[KWQTextFieldController textView:didHandleEvent:]): Turn off deferral here, now that the key press
1308         has been handled.
1309         (-[KWQTextFieldController string]): Added. Calls stringValue only if there's no editor. If there is
1310         an editor, calls string on that instead.
1311         (-[KWQTextFieldController textView:shouldChangeTextInRange:replacementString:]): Added. Called by
1312         the three field subclasses. Truncates incoming strings at the first CR or LF character. This works
1313         properly with Undo, and the old technique did not.
1314         (-[KWQTextFieldController preprocessString:]): Added. Helper used for all the setStringValue methods
1315         so they all truncate at the first CR or LF as well as truncating to the appropriate length.
1316         (-[KWQTextField setStringValue:]): Changed to call the controller.
1317         (-[KWQTextField textView:shouldChangeTextInRange:replacementString:]): Added. Calls the controller.
1318         (-[KWQSecureTextField setStringValue:]): More of the same.
1319         (-[KWQSecureTextField textView:shouldChangeTextInRange:replacementString:]): Ditto.
1320         (-[KWQSearchField setStringValue:]): Ditto.
1321         (-[KWQSearchField textView:shouldChangeTextInRange:replacementString:]): Ditto.
1322
1323         * kwq/KWQWidget.h: Added the new setDeferFirstResponderChanges function.
1324         * kwq/KWQWidget.mm:
1325         (QWidget::~QWidget): Clear out the deferredFirstResponder global just in case we are destroyed
1326         while we "have the ball".
1327         (QWidget::hasFocus): If we are in the deferred mode, use the global to decide who to say is focused.
1328         (QWidget::setFocus): If we are in the deferred mode, set the global instead of making something be
1329         the first responder.
1330         (QWidget::setDeferFirstResponderChanges): Set the boolean. If the boolean is being cleared, then
1331         get the deferred first responder and call setFocus on it again, which will do the work we didn't
1332         want to do earlier.
1333
1334 2005-01-24  Ken Kocienda  <kocienda@apple.com>
1335
1336         Reviewed by John
1337
1338         Fix for this bug:
1339
1340         <rdar://problem/3964350> Deleting all content and typing in a message gets one letter, then beeps
1341
1342         This is a problem with the way we turn key events, when a user starts typing, into editing commands.
1343         Right now, that initial delete key in an empty window is turned into an editing command, and much
1344         editing code runs that need not run. What's more, this is confusing the selection machinery. The
1345         solution is to detect the "delete in an empty window" case and don't turn this into an editing
1346         command at all.
1347
1348         * khtml/editing/htmlediting.cpp:
1349         (khtml::TypingCommand::deleteKeyPressed)
1350
1351 2005-01-23  Adele Amchan  <adele@apple.com>
1352
1353         Fixed <rdar://problem/3964286> REGRESSION (178-179): menus and "go to old site" link at Ameritrade's new site don't work
1354
1355         Removed check for document from checkCompleted.  In this case, WebCore never knew that a WebImageView was complete because
1356         it fell into this document check case in checkCompleted.
1357         Added check for document to tokenizerProcessedData (see <rdar://problem/3807144>)
1358
1359         Reviewed by Darin.
1360
1361         * khtml/khtml_part.cpp:
1362         (KHTMLPart::stop):
1363         (KHTMLPart::checkCompleted):
1364         * khtml/khtml_part.h:
1365         * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::tokenizerProcessedData):
1366
1367 2005-01-21  Richard Williamson   <rjw@apple.com>
1368
1369         Fixed <rdar://problem/3759399> Javascript / Liveconnect problems ((event handler):Undefined value)
1370
1371         Java applets specified with <object> or <embed> weren't scriptable.  Now they are.
1372
1373         Reviewed by Chris.
1374
1375         * khtml/html/html_objectimpl.cpp:
1376         (HTMLEmbedElementImpl::getEmbedInstance):
1377         (HTMLObjectElementImpl::getObjectInstance):
1378
1379 2005-01-21  Ken Kocienda  <kocienda@apple.com>
1380
1381         Reviewed by John
1382
1383         Fix for this bug:
1384
1385         <rdar://problem/3959464> REGRESSION (Mail): Insertion point goes back to beginning of document after deleting
1386
1387         * khtml/editing/htmlediting.cpp:
1388         (khtml::DeleteSelectionCommand::handleGeneralDelete): Add special case to handle retaining a fully-selected block.
1389         This fixes the bug.
1390         * layout-tests/editing/deleting/delete-3959464-fix-expected.txt: Added.
1391         * layout-tests/editing/deleting/delete-3959464-fix.html: Added.
1392
1393 2005-01-21  Richard Williamson   <rjw@apple.com>
1394
1395         Fixed <rdar://problem/3966998> REGRESSION(179-TOT) clicking on gmail message brings me to blank screen
1396
1397         getElementById() sometimes returns Undefined() instead of Null().
1398
1399         Reviewed by Hyatt.
1400
1401         * khtml/ecma/kjs_dom.cpp:
1402         (DOMDocumentProtoFunc::tryCall):
1403
1404 2005-01-21  David Hyatt  <hyatt@apple.com>
1405
1406         Fix for 3773809, make sure that overflow regions never end up at an invalid scroll offset because of a layout change.
1407
1408         Reviewed by kocienda
1409
1410         * khtml/rendering/render_layer.cpp:
1411         (RenderLayer::updateScrollInfoAfterLayout):
1412
1413 2005-01-21  David Hyatt  <hyatt@apple.com>
1414
1415         Fix for 3966349, hang loading page.  Make sure that loops using popOneBlock properly check for a null
1416         blockStack to avoid an infinite loop.
1417
1418         Reviewed by kocienda
1419
1420         * khtml/html/htmlparser.cpp:
1421         (KHTMLParser::parseToken):
1422         (KHTMLParser::insertNode):
1423         (KHTMLParser::popInlineBlocks):
1424
1425 2005-01-21  Ken Kocienda  <kocienda@apple.com>
1426
1427         Reviewed by John
1428
1429         Fix for this bug:
1430
1431         <rdar://problem/3957204> Mail crashed when sending a message in ~InsertParagraphSeparatorInQuotedContentCommand
1432
1433         * khtml/editing/htmlediting.cpp:
1434         (khtml::InsertParagraphSeparatorInQuotedContentCommand::InsertParagraphSeparatorInQuotedContentCommand):
1435         InsertParagraphSeparatorInQuotedContentCommand has this member variable: DOM::ElementImpl *m_breakNode;
1436         I failed to initialize this variable to null in the constructor, and there is an uncommonly-traveled
1437         code path which does not set this variable to something good. In the destructor, we check for null,
1438         and deref if non-null. Obvious "BOOM" potential. Fixed.
1439
1440 2005-01-21  Ken Kocienda  <kocienda@apple.com>
1441
1442         Reviewed by John
1443
1444         Fix for this bug:
1445
1446         <rdar://problem/3966311> REGRESSION (Mail): Hitting return makes space character disappear
1447
1448         * khtml/editing/htmlediting.cpp:
1449         (khtml::InsertParagraphSeparatorCommand::doApply): When writing the code to insert a block
1450         in response to the return key, I did not write code to cover the case described in the bug.
1451         Now I have.
1452         * layout-tests/editing/inserting/insert-div-025-expected.txt: Added.
1453         * layout-tests/editing/inserting/insert-div-025.html: Added.
1454
1455 2005-01-20  Maciej Stachowiak  <mjs@apple.com>
1456
1457         Reviewed by Ken.
1458
1459         <rdar://problem/3965196> security fix for javascript: exploit missed one case (already fixed in updates)
1460
1461         * khtml/ecma/kjs_window.cpp:
1462         (WindowFunc::tryCall): correct mistake in earlier fix for the following bug, caught by Adele:
1463
1464 2005-01-20  Maciej Stachowiak  <mjs@apple.com>
1465
1466         Reviewed by Ken.
1467
1468         <rdar://problem/3965466> editing needs to insert text before applying typing style
1469
1470         * khtml/editing/htmlediting.cpp:
1471         (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): Remove a FIXME comment.
1472         (khtml::InsertLineBreakCommand::doApply): Apply style to the BR
1473         node after inserting it.
1474         (khtml::InsertTextCommand::prepareForTextInsertion): Don't try to
1475         apply style to any new nodes created before inserting.
1476         (khtml::InsertTextCommand::input): Apply style to the inserted range
1477         after doing the text insert.
1478         * khtml/editing/htmlediting.h:
1479
1480         This change leads to improved results on one of the layout tests:
1481
1482         * layout-tests/editing/inserting/insert-div-024-expected.txt:
1483
1484 === Safari-180 ===
1485
1486 2005-01-20  Ken Kocienda  <kocienda@apple.com>
1487
1488         Reviewed by John
1489
1490         Fix for this bug:
1491
1492         <rdar://problem/3964646> REGRESSION (179-180): Typing space at end of line makes following paragraph disappear
1493
1494         * khtml/editing/htmlediting.cpp:
1495         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Recent change to block placeholder removal code
1496         caused this regression. The code became too aggressive in removing block placeholders, and would remove them
1497         from blocks other than the block containing the selection.
1498         * layout-tests/editing/inserting/insert-div-023-expected.txt: This file had a spurious BR element in it that
1499         I did not notice earlier.
1500         * layout-tests/editing/inserting/insert-div-024-expected.txt: Ditto.
1501
1502 2005-01-20  David Hyatt  <hyatt@apple.com>
1503
1504         Fix for oddness on albertsons.com.  Make sure not to crash when setting/removing style properties on a node
1505         with no document.  The bug # is 3813900.
1506
1507         Reviewed by john
1508
1509 2005-01-20  David Harrison  <harrison@apple.com>
1510
1511         Reviewed by Darin.
1512
1513         PARTIAL fix for following bug.  Create attributed string with fonts and links.  Still need to add attachments.
1514         <rdar://problem/3942606> AX: Support kAXAttributedStringForTextMarkerRangeParameterizedAttribute
1515
1516         * kwq/KWQAccObject.mm:
1517         (-[KWQAccObject textUnderElement]):
1518         Touched up previous checkin to match code review comments.
1519
1520         (-[KWQAccObject value]):
1521         Use plainText for this one instead of obsolete attributedString.
1522
1523         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
1524         Advertise AXAttributedStringForTextMarkerRange.
1525
1526         (AXAttributeStringAddFont):
1527         (AXAttributeStringAddElement):
1528         (-[KWQAccObject linkUIElementForNode:]):
1529         (-[KWQAccObject _addAttachmentForElement:URL:needsParagraph:usePlaceholder:]):
1530         (-[KWQAccObject accessibilityAttributedStringForRange:]):
1531         New routines to support AXAttributedStringForTextMarkerRange.
1532
1533         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
1534         Use new routines instead of obsolete attributedString.
1535
1536 2005-01-20  David Harrison  <harrison@apple.com>
1537
1538         Reviewed by Darin.
1539
1540         <rdar://problem/3960196> AX Crash in DOM::Range::setStartBefore
1541
1542         * khtml/xml/dom_docimpl.cpp:
1543         (DocumentImpl::getAccObjectCache):
1544         Adjust when detecting cache in non-top level WebArea.
1545         * kwq/KWQAccObject.mm:
1546         (-[KWQAccObject textUnderElement]):
1547         (-[KWQAccObject accessibilityAttributeValue:]):
1548         Make sure the elements document is the current one for the part.
1549
1550 2005-01-20  Darin Adler  <darin@apple.com>
1551
1552         Reviewed by John.
1553
1554         - fixed <rdar://problem/3922980> Mail not crashing, just quitting itself suddenly in -[WebHTMLView(MailExtras) findString:options:]
1555
1556         * khtml/dom/dom2_range.cpp: (DOM::operator==): Handle null and detached ranges without raising exceptions.
1557         The uncaught exception would make the entire program terminate.
1558
1559         - fixed <rdar://problem/3885729> attempting to modify a computed style does nothing, but should raise a DOM exception
1560
1561         * kwq/DOM-CSS.mm: (-[DOMCSSStyleDeclaration setProperty:::]): Uncommented this code which I had to disable
1562         back in November because Mail was stumbling over it.
1563
1564         - fixed <rdar://problem/3943049> focus() called during onload handler results in square text field on <input type=search>
1565
1566         * kwq/KWQLineEdit.mm: (QLineEdit::baselinePosition): Change computation so it will work even for fields
1567         that position their text in a way that depends on vertical size. In the case of a search field, the text
1568         is centered, so the old logic was broken.
1569         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Set the frame size to
1570         something large enough to accomodate the field editor. If we start the frame at size 0,0 we run into
1571         AppKit trouble when it insets the frame to figure out the frame for the field editor. In the case of this
1572         bug this happens because we become first responder before being sized and positioned by the HTML layout code.
1573
1574         - improved debugging output when using "po" from gdb with Objective-C DOM
1575
1576         * kwq/DOM.mm:
1577         (-[DOMNode description]): Added. Includes node name ("<tr>") and node value (e.g., string for text node).
1578         (-[DOMRange description]): Tweaked format.
1579
1580 2005-01-19  Richard Williamson   <rjw@apple.com>
1581
1582         After further discussion with Real we have decided to NOT include the
1583         additional CLSID for the real plugin.  See 3958601.
1584
1585         Reviewed by Darin.
1586
1587         * khtml/rendering/render_frames.cpp:
1588         (RenderPartObject::updateWidget):
1589
1590 2005-01-19  David Harrison  <harrison@apple.com>
1591
1592         Reviewed by Maciej.
1593
1594         <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
1595
1596         Previous checkin for this bug was missing some of the patch.  Editing snafu with multiple changes in tree.
1597
1598         * kwq/KWQAccObject.mm:
1599         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
1600         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
1601         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
1602         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
1603         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
1604         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
1605         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
1606
1607 2005-01-19  Ken Kocienda  <kocienda@apple.com>
1608
1609         Reviewed by John
1610
1611         * khtml/editing/htmlediting.cpp:
1612         (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Small, cosmetic change John and I decided
1613         to do on my last checkin, but I forgot to do before landing.
1614
1615 2005-01-19  Ken Kocienda  <kocienda@apple.com>
1616
1617         Reviewed by John
1618
1619         Fix for this bug:
1620
1621         <rdar://problem/3959727> REGRESSION (Mail): Style not preserved on blank lines
1622
1623         * khtml/editing/htmlediting.cpp:
1624         (khtml::CompositeEditCommand::applyStyle):
1625         (khtml::CompositeEditCommand::insertBlockPlaceholder): New function that unconditionally adds a block placeholder.
1626         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Now returns bool based on whether
1627         placeholder was added or not.
1628         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Now searches all the descendents of a block
1629         looking for a placeholder. The old code, which just looked at the last child of a node, started missing
1630         once block placeholders became styled (which started happening with this patch).
1631         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Now handles applying typing style
1632         to a block placeholder at call time, rather than setting the typing style as a latent style that
1633         might be applied later. This is an important part of the bug fix.
1634         (khtml::DeleteSelectionCommand::doApply): Now uses bool return value from insertBlockPlaceholderIfNeeded()
1635         and passes it along to calculateStyleBeforeInsertion, so the case where a block placeholder needs to
1636         be styled can be detected.
1637         (khtml::InsertParagraphSeparatorCommand::InsertParagraphSeparatorCommand): Changed the way this class
1638         managed style. Before it would calculate and set typing style for the block added. This is not
1639         sufficient. Added blocks need to styled immediately. Some name changes to instance variables in
1640         this class due to the change to accommodate this change.
1641         (khtml::InsertParagraphSeparatorCommand::~InsertParagraphSeparatorCommand): Name changes, as above.
1642         (khtml::InsertParagraphSeparatorCommand::calculateStyleBeforeInsertion): Ditto.
1643         (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Ditto.
1644         (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
1645         (khtml::ReplaceSelectionCommand::doApply): Improve check for testing when a placeholder
1646         block can be removed in its entirety after the insertion.
1647         * khtml/editing/htmlediting.h: Update header accordingly.
1648         * khtml/khtml_part.cpp:
1649         (KHTMLPart::selectionComputedStyle): Move position for computed style check downstream before
1650         doing check when the position is in an empty block (this makes sure any style on any block
1651         placeholder is accounted for).
1652
1653         New layout tests to check bug fix.
1654
1655         * layout-tests/editing/style/block-style-004-expected.txt: Added.
1656         * layout-tests/editing/style/block-style-004.html: Added.
1657         * layout-tests/editing/style/block-style-005-expected.txt: Added.
1658         * layout-tests/editing/style/block-style-005.html: Added.
1659         * layout-tests/editing/style/block-style-006-expected.txt: Added.
1660         * layout-tests/editing/style/block-style-006.html: Added.
1661
1662         Results updated to reflect new block placeholder code.
1663
1664         * layout-tests/editing/inserting/insert-div-004-expected.txt
1665         * layout-tests/editing/inserting/insert-div-005-expected.txt
1666         * layout-tests/editing/inserting/insert-div-006-expected.txt
1667         * layout-tests/editing/inserting/insert-div-008-expected.txt
1668         * layout-tests/editing/inserting/insert-div-011-expected.txt
1669         * layout-tests/editing/inserting/insert-div-012-expected.txt
1670         * layout-tests/editing/inserting/insert-div-013-expected.txt
1671         * layout-tests/editing/inserting/insert-div-014-expected.txt
1672         * layout-tests/editing/inserting/insert-div-015-expected.txt
1673         * layout-tests/editing/inserting/insert-div-016-expected.txt
1674         * layout-tests/editing/inserting/insert-div-017-expected.txt
1675         * layout-tests/editing/inserting/insert-div-018-expected.txt
1676         * layout-tests/editing/inserting/insert-div-019-expected.txt
1677         * layout-tests/editing/inserting/insert-div-021-expected.txt
1678         * layout-tests/editing/inserting/insert-div-022-expected.txt
1679         * layout-tests/editing/inserting/insert-div-023-expected.txt
1680         * layout-tests/editing/inserting/insert-div-024-expected.txt
1681
1682 2005-01-19  David Hyatt  <hyatt@apple.com>
1683
1684         Dont null-check the renderer before submitting, since a script can set it to display:none and still expect the
1685         submission to occur.  Fixes bug #3477282.
1686
1687         Reviewed by kdecker
1688
1689         (DOM::HTMLInputElementImpl::defaultEventHandler):
1690
1691 2005-01-18  Richard Williamson   <rjw@apple.com>
1692
1693         Fixed several issues all arising from analysis of plugin detection code at ifilm.com:
1694
1695         Fixed <rdar://problem/3958592> can't script plug-ins if plug-in is invoked with <object> element instead of <embed>
1696         Fixed <rdar://problem/3958597> <object> elements with IDs do not show up as named properties of the document
1697         Fixed <rdar://problem/3960973> DOM objects for plugin elements are not accessible
1698         Fixed <rdar://problem/3958601> need an additional class ID in WebCore for the Real plug-in
1699
1700         We now support accessing scriptable plugin objects that are specified with <applet>, <embed>, or <object>
1701         tags.  Also, if any of these elements are named they can be accessed from the document or window objects.
1702         Finally, DOM methods are properties will be forwarded appropriately for the plugin's root scriptable object.
1703
1704         Reviewed by Chris.
1705
1706         * khtml/dom/html_document.cpp:
1707         (HTMLDocument::objects):
1708         * khtml/dom/html_document.h:
1709         * khtml/ecma/kjs_dom.cpp:
1710         (DOMDocumentProtoFunc::tryCall):
1711         (DOMElementProtoFunc::tryCall):
1712         (KJS::getRuntimeObject):
1713         * khtml/ecma/kjs_dom.h:
1714         * khtml/ecma/kjs_html.cpp:
1715         (KJS::HTMLDocument::tryGet):
1716         (KJS::HTMLElement::tryGet):
1717         (KJS::HTMLCollection::tryGet):
1718         (KJS::HTMLCollection::getNamedItems):
1719         * khtml/ecma/kjs_window.cpp:
1720         (Window::get):
1721         * khtml/html/html_miscimpl.cpp:
1722         (HTMLCollectionImpl::traverseNextItem):
1723         * khtml/html/html_miscimpl.h:
1724         (DOM::HTMLCollectionImpl::):
1725         * khtml/html/html_objectimpl.cpp:
1726         (HTMLAppletElementImpl::getAppletInstance):
1727         (HTMLObjectElementImpl::HTMLObjectElementImpl):
1728         (HTMLObjectElementImpl::getObjectInstance):
1729         * khtml/html/html_objectimpl.h:
1730         * khtml/rendering/render_frames.cpp:
1731         (RenderPartObject::updateWidget):
1732         * kwq/KWQKHTMLPart.h:
1733         * kwq/KWQKHTMLPart.mm:
1734         (KWQKHTMLPart::getObjectInstanceForView):
1735
1736 2005-01-18  David Hyatt  <hyatt@apple.com>
1737
1738         Fix for 3948123, rolling over link erases nearby text.  The repaint rect check for lines was wrong whenever
1739         two lines overlapped.
1740
1741         Reviewed by kocienda
1742
1743         * khtml/rendering/render_flow.cpp:
1744         (RenderFlow::paintLines):
1745
1746 2005-01-18  Ken Kocienda  <kocienda@apple.com>
1747
1748         Reviewed by Hyatt
1749
1750         Fix for this bug:
1751
1752         <rdar://problem/3960116> Focus rings paint incorrectly for contenteditable blocks in web pages
1753
1754         * khtml/rendering/render_flow.cpp:
1755         (RenderFlow::addFocusRingRects): Fix painting of focus rings so that ring only paints around
1756         outermost contenteditable elements.
1757
1758 2005-01-18  David Harrison  <harrison@apple.com>
1759
1760         Reviewed by Darin.
1761
1762         <rdar://problem/3959668> accessibilityFocusedUIElement sometimes returns an ignored element; it must not
1763
1764         * kwq/KWQAccObject.mm:
1765         (-[KWQAccObject accessibilityFocusedUIElement]):
1766         Return parentObjectUnignored if focused object is ignored.
1767
1768 2005-01-18  Ken Kocienda  <kocienda@apple.com>
1769
1770         Reviewed by John
1771
1772         Fix for this bug:
1773
1774         <rdar://problem/3952877> REGRESSION (Mail): Command-left/right-arrows don't work with file attachment
1775
1776         * khtml/editing/selection.cpp:
1777         (khtml::nodeForInlineBox): New helper function used in reimplementation of function below.
1778         (khtml::selectionForLine): Reimplemented using line box smarts. I originally wrote this code when
1779         I had a less than full understanding of line layout. I can do better now, and my new version no
1780         longer fails to notice attachments when doing the kind of navigation mentioned in the bug.
1781
1782 2005-01-17  David Harrison  <harrison@apple.com>
1783
1784         Reviewed by John Sullivan.
1785
1786         <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
1787         <rdar://problem/3949848> AX: paragraph marker routines do not work when given a paragraph end marker
1788
1789         Also fixed sentence support in the same way.
1790
1791         * kwq/KWQAccObject.mm:
1792         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
1793         (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
1794         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
1795         (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
1796         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
1797         Move one position in desired direction before calling visible_units.cpp code.
1798
1799 2005-01-17  David Harrison  <harrison@apple.com>
1800
1801         Updated expected test results.
1802
1803         <rdar://problem/3945880> line-ending space seems not to be present
1804
1805         * layout-tests/editing/deleting/delete-ws-fixup-001-expected.txt:
1806         * layout-tests/editing/deleting/delete-ws-fixup-002-expected.txt:
1807         * layout-tests/editing/inserting/insert-div-020-expected.txt:
1808         * layout-tests/editing/inserting/insert-div-021-expected.txt:
1809         * layout-tests/editing/inserting/insert-div-022-expected.txt:
1810         * layout-tests/editing/inserting/insert-div-023-expected.txt:
1811         * layout-tests/editing/inserting/insert-div-024-expected.txt:
1812
1813 2005-01-17  David Harrison  <harrison@apple.com>
1814
1815         Reviewed by Dave Hyatt (bidi.cpp) and Darin Adler (selection.cpp).
1816
1817         <rdar://problem/3945880> line-ending space seems not to be present
1818
1819         * khtml/editing/selection.cpp:
1820         (khtml::Selection::validate):
1821         Extend AFTER_WHITE_SPACE code to support white spac in the middle of paragraphs, not just the end.
1822         * khtml/rendering/bidi.cpp:
1823         (khtml::RenderBlock::findNextLineBreak):
1824         Pick left/rightness of word selection based on being at the end of paragraph (i.e. after a hard line break).
1825
1826 2005-01-17  Darin Adler  <darin@apple.com>
1827
1828         Reviewed by John Louch.
1829
1830         - fixed <rdar://problem/3958503> need screenX and screenY to use WebKit windowFrame delegate
1831
1832         * khtml/ecma/kjs_window.cpp: (Window::get): Change screenX and screenY to use frameGeometry instead
1833         of using mapToGlobal and screen in a complicated way.
1834         * kwq/KWQKHTMLView.mm: Removed unused mapToGlobal function.
1835         * kwq/KWQWindowWidget.h: Ditto.
1836         * kwq/KWQWindowWidget.mm: Ditto.
1837
1838 2005-01-17  David Hyatt  <hyatt@apple.com>
1839
1840         Fix a screwup in rightmost/lowets position computation.  3955207.  Make sure floats with layers are still checked.
1841
1842         Reviewed by kocienda
1843
1844         * khtml/rendering/render_block.cpp
1845
1846 2005-01-17  Ken Kocienda  <kocienda@apple.com>
1847
1848         Reviewed by John
1849
1850         <rdar://problem/3953366> Problems with typing attributes in HTML compose
1851
1852         * khtml/editing/htmlediting.cpp:
1853         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Remove misguided "optimization"
1854         that tried to sense when typing style could be cleared without actually doing a style diff between
1855         before-delete and after-delete positions. Removing this extra check and running the
1856         general-purpose code fixes the bug.
1857
1858 2005-01-17  Richard Williamson   <rjw@apple.com>
1859
1860         Fixed <rdar://problem/3753030> Need to ensure same origin for plugin binding invocations (origin security rules)
1861
1862         Keep track of originating execution context and target execution
1863         context for native JS object wrappers, and perform appropriate
1864         security checks.
1865
1866         Reviewed by David Harrison.
1867
1868         * khtml/ecma/kjs_binding.cpp:
1869         (ScriptInterpreter::isGlobalObject):
1870         (ScriptInterpreter::isSafeScript):
1871         (ScriptInterpreter::interpreterForGlobalObject):
1872         * khtml/ecma/kjs_binding.h:
1873         * khtml/ecma/kjs_window.cpp:
1874         (Window::interpreter):
1875         (Window::isSafeScript):
1876         * khtml/ecma/kjs_window.h:
1877         * kwq/DOMInternal.mm:
1878         (-[WebScriptObject _initializeScriptDOMNodeImp]):
1879         * kwq/KWQKHTMLPart.mm:
1880         (KWQKHTMLPart::windowScriptObject):
1881         (KWQKHTMLPart::windowScriptNPObject):
1882         * kwq/WebCoreBridge.mm:
1883         (rootForView):
1884         (-[WebCoreBridge init]):
1885
1886 2005-01-17  Ken Kocienda  <kocienda@apple.com>
1887
1888         Reviewed by John
1889
1890         Fix for this bug:
1891
1892         <rdar://problem/3786659> REGRESSION (Mail): editable WebViews don't work with
1893         "size up" and "size down" NSFontManager changes
1894
1895         * khtml/css/cssparser.cpp:
1896         (CSSParser::parseValue): Add support for parsing new font size delta property.
1897         * khtml/css/cssproperties.c: Generated file.
1898         * khtml/css/cssproperties.h: Ditto.
1899         * khtml/css/cssproperties.in: Add support for parsing new font size delta property.
1900         * khtml/editing/htmlediting.cpp:
1901         (khtml::isEmptyStyleSpan): New helper function. Code existed before, but now factored out for convenient use.
1902         (khtml::CompositeEditCommand::removeNodeAttribute): Does check on value to see it exists before creating
1903         and running command to do the removal.
1904         (khtml::ApplyStyleCommand::doApply): Now calls new applyRelativeFontStyleChange() function as part of its work.
1905         (khtml::ApplyStyleCommand::applyRelativeFontStyleChange): New function that does the "heavy lifting" to handle
1906         relative font size changes.
1907         (khtml::ApplyStyleCommand::applyInlineStyle): Range check the start and end positions to make sure the start is
1908         before or equal to the end. Swap them if not true. I ran into this problem in some similar code in
1909         applyRelativeFontStyleChange(). Moving that goodness here too.
1910         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded): Uses a local variable to save a value used often.
1911         Code is the same, but reads better now. Function now returns bool as well, just like splitTextAtStartIfNeeded.
1912         I use the bool return value now (I obviously did not need it before).
1913         (khtml::ApplyStyleCommand::computedFontSize): New helper function.
1914         (khtml::ApplyStyleCommand::joinChildTextNodes): Ditto.
1915         (khtml::createStyleSpanElement): Ditto.
1916         * khtml/editing/htmlediting.h: Update header accordingly.
1917         * khtml/editing/jsediting.cpp: Add new command string to enable relative font size changes.
1918         * kwq/DOM-CSS.mm:
1919         (-[DOMCSSStyleDeclaration _fontSizeDelta]): New convenience.
1920         (-[DOMCSSStyleDeclaration _setFontSizeDelta:]): Ditto.
1921         * kwq/DOMPrivate.h: Declare new conveniences.
1922         * layout-tests/editing/editing.js: Add new JS to enable relative font size changes, as well as explicit font size setting.
1923         * layout-tests/editing/style/relative-font-size-change-001-expected.txt: Added.
1924         * layout-tests/editing/style/relative-font-size-change-001.html: Added.
1925         * layout-tests/editing/style/relative-font-size-change-002-expected.txt: Added.
1926         * layout-tests/editing/style/relative-font-size-change-002.html: Added.
1927         * layout-tests/editing/style/relative-font-size-change-003-expected.txt: Added.
1928         * layout-tests/editing/style/relative-font-size-change-003.html: Added.
1929         * layout-tests/editing/style/relative-font-size-change-004-expected.txt: Added.
1930         * layout-tests/editing/style/relative-font-size-change-004.html: Added.
1931
1932 2005-01-14  Darin Adler  <darin@apple.com>
1933
1934         * khtml/css/cssproperties.c: Regenerated with gperf 3.0.1.
1935         * khtml/css/cssvalues.c: Regenerated with gperf 3.0.1.
1936         * khtml/misc/htmlattrs.c: Regenerated with gperf 3.0.1.
1937         * khtml/misc/htmltags.c: Regenerated with gperf 3.0.1.
1938
1939 2005-01-14  Chris Blumenberg  <cblu@apple.com>
1940
1941         Fixed: <rdar://problem/3886415> arrow keys don't work when the user hits Back to return to RSS page
1942
1943         Reviewed by hyatt.
1944
1945         (KWQKHTMLPart::openURLFromPageCache): restore mousePressNode
1946         (KWQKHTMLPart::mousePressNode): new
1947         * kwq/KWQPageState.h:
1948         * kwq/KWQPageState.mm:
1949         (-[KWQPageState initWithDocument:URL:windowProperties:locationProperties:interpreterBuiltins:]): save the mousePressNode
1950         (-[KWQPageState clear]): clear the mousePressNode
1951         (-[KWQPageState dealloc]): deref the mousePressNode
1952         (-[KWQPageState mousePressNode]): new
1953
1954 === Safari-179 ===
1955
1956 2005-01-13  Vicki Murley <vicki@apple.com>
1957
1958         Reviewed by Adele.
1959
1960         - fix <rdar://problem/3946836> Safari about box lists 2004 instead of 2005
1961
1962         * WebCore.pbproj/project.pbxproj: bump "2004" to "2005"
1963
1964 2005-01-13  David Harrison  <harrison@apple.com>
1965
1966         Reviewed by Ken Kocienda.
1967
1968         Better fix for 3905066.
1969
1970         * khtml/editing/htmlediting.cpp:
1971         (khtml::InsertParagraphSeparatorCommand::doApply):
1972
1973 2005-01-10  Maciej Stachowiak  <mjs@apple.com>
1974
1975         Reviewed by Darin.
1976
1977         <rdar://problem/3758033> REGRESSION (Mail): Support attributes in marked text (International input)
1978
1979         * khtml/rendering/render_text.cpp:
1980         (InlineTextBox::paint): Support painting custom underline markers for
1981         marked text in place of generic yellow.
1982         (InlineTextBox::paintMarkedTextUnderline): New method that handles this.
1983         * khtml/rendering/render_text.h:
1984         * kwq/KWQKHTMLPart.h: Declare new methods and structs.
1985         * kwq/KWQKHTMLPart.mm:
1986         (KWQKHTMLPart::clear): Clear marked test underlines.
1987         (KWQKHTMLPart::setMarkedTextRange): Takes attributes and ranges now.
1988         (convertAttributesToUnderlines): Converts NSAttributedString attributes
1989         to simplified and C++-friendly form.
1990         (KWQKHTMLPart::markedTextUsesUnderlines): New method.
1991         (KWQKHTMLPart::markedTextUnderlines): New method.
1992         * kwq/KWQPainter.mm:
1993         (QPainter::drawLineForText): Handle pen width.
1994         * kwq/WebCoreBridge.h:
1995         * kwq/WebCoreBridge.mm:
1996         (-[WebCoreBridge setMarkedTextDOMRange:customAttributes:ranges:]): Take attributes
1997         and ranges.
1998         * kwq/WebCoreTextRenderer.h:
1999
2000 2005-01-12  David Harrison  <harrison@apple.com>
2001
2002         Reviewed by Dave Hyatt.
2003
2004         <rdar://problem/3888973> AX: Parent AXWebArea of nested AXWebAreas is messed up
2005
2006         * kwq/KWQAccObject.mm:
2007         (-[KWQAccObject addChildrenToArray:]):
2008         Use the widget's outer view.
2009
2010 2005-01-12  David Harrison  <harrison@apple.com>
2011
2012         Reviewed by Darin Adler.
2013
2014         <rdar://problem/3949908> Crash when asking for the kAXLengthForTextMarkerRangeParameterizedAttribute
2015
2016         * khtml/xml/dom_nodeimpl.cpp:
2017         (NodeImpl::displayNode):
2018         Add quotes around text node content.
2019         (NodeBaseImpl::childNode):
2020         Add nil check to return nil rather than crash when child node not found.
2021         * kwq/KWQAccObject.mm:
2022         (-[KWQAccObject doAXStringForTextMarkerRange:]):
2023         Pass range compliant positions to TextIterator.
2024
2025 2005-01-12  David Hyatt  <hyatt@apple.com>
2026
2027         Fix for 3951203, CSS border style incorrectly clipped on inline elements.  Make sure the repaint rect set during
2028         line layout is smarter about including the overflow for both old states and new states.
2029
2030         Reviewed by john
2031
2032         * khtml/rendering/bidi.cpp:
2033         (khtml::RenderBlock::layoutInlineChildren):
2034
2035 2005-01-12  David Hyatt  <hyatt@apple.com>
2036
2037         Fix for bug 3937608, versiontracker.com flashes and displays the right column below content.  Make sure to move
2038         tables/overflows that dont fit within a block only in strict mode.
2039
2040         Fix for bug 3931049, characters dont show up when typing.  Make sure to dirty the right lines when this specific
2041         case in editing is hit.
2042
2043         Reviewed by darin (first one), kocienda (second one)
2044
2045         * khtml/rendering/render_block.cpp:
2046         (khtml::RenderBlock::getClearDelta):
2047         * khtml/rendering/render_flow.cpp:
2048         (RenderFlow::dirtyLinesFromChangedChild):
2049         * khtml/rendering/render_line.cpp:
2050         (khtml::InlineFlowBox::verticallyAlignBoxes):
2051
2052 2005-01-12  Ken Kocienda  <kocienda@apple.com>
2053
2054         Reviewed by John
2055
2056         Fix for this bug:
2057
2058         <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
2059
2060         * khtml/editing/htmlediting.cpp:
2061         (khtml::DeleteSelectionCommand::initializePositionData): Change test that will prevent block
2062         merging. End-of-line test was just wrong. Call new start-of and end-of paragraph functions
2063         added to visible_position files.
2064         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Add tests for BR elements, which makes
2065         tests to determine when to stop moving nodes more complete and correct. Also improved comments.
2066         * khtml/editing/visible_position.cpp:
2067         (khtml::isFirstVisiblePositionInParagraph): New function.
2068         (khtml::isLastVisiblePositionInParagraph): New function.
2069         * khtml/editing/visible_position.h: Update header accordingly.
2070         * layout-tests/editing/deleting/delete-block-merge-contents-002-expected.txt: New results, slightly different from former
2071         results but still correct.
2072         * layout-tests/editing/deleting/delete-block-merge-contents-007-expected.txt: Ditto.
2073         * layout-tests/editing/deleting/delete-block-merge-contents-018-expected.txt: Added.
2074         * layout-tests/editing/deleting/delete-block-merge-contents-018.html: Added.
2075         * layout-tests/editing/deleting/delete-block-merge-contents-019-expected.txt: Added.
2076         * layout-tests/editing/deleting/delete-block-merge-contents-019.html: Added.
2077         * layout-tests/editing/deleting/delete-block-merge-contents-020-expected.txt: Added.
2078         * layout-tests/editing/deleting/delete-block-merge-contents-020.html: Added.
2079         * layout-tests/editing/deleting/delete-block-merge-contents-021-expected.txt: Added.
2080         * layout-tests/editing/deleting/delete-block-merge-contents-021.html: Added.
2081         * layout-tests/editing/deleting/delete-line-014-expected.txt: These new results are actually better, and fix a bug.
2082
2083 2005-01-11  Richard Williamson   <rjw@apple.com>
2084
2085         Fixed 3922875.  Fall back to DOM object is EMBED element
2086         has no associated runtime object.
2087
2088         Reviewed by Chris.
2089
2090         * khtml/ecma/kjs_dom.cpp:
2091         (KJS::getRuntimeObject):
2092         * khtml/ecma/kjs_html.cpp:
2093         (KJS::HTMLDocument::tryGet):
2094         (KJS::HTMLElement::tryGet):
2095         (KJS::HTMLCollection::tryGet):
2096         (KJS::HTMLCollection::getNamedItems):
2097
2098 2005-01-11  David Hyatt  <hyatt@apple.com>
2099
2100         Fix for 3882299, missing content on gibson.com.  Change our handling of " and ' in certain states of the parser to match
2101         other browsers.
2102
2103         Reviewed by Maciej
2104
2105         * khtml/html/htmltokenizer.cpp:
2106         (khtml::HTMLTokenizer::parseTag):
2107
2108 2005-01-11  Chris Blumenberg  <cblu@apple.com>
2109
2110         Fixed: <rdar://problem/3930733> Mail prints second page of email blank
2111
2112         Reviewed by dave.
2113
2114         * khtml/rendering/render_canvas.cpp:
2115         (RenderCanvas::paint): cache the print rect since the dirty rect can get changed during printing
2116         * khtml/rendering/render_flow.cpp:
2117         (RenderFlow::paintLines): removed null check since the print rect should never be null
2118         * khtml/rendering/render_list.cpp:
2119         (RenderListMarker::paint): ditto
2120         * kwq/KWQKHTMLPart.mm:
2121         (KWQKHTMLPart::adjustPageHeight): don't set the print rect here since this method is not called for the last page
2122
2123 2005-01-10  Ken Kocienda  <kocienda@apple.com>
2124
2125         Reviewed by Hyatt
2126
2127         Fix for this bug:
2128
2129         <rdar://problem/3943648> extra line is inserted after pressing return within quoted text of reply
2130
2131         * khtml/editing/htmlediting.cpp:
2132         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): This bug was the result of a
2133         simple coding mistake. A local variable was erroneously redefined in a deeper scope, and so the
2134         result calculated in that deeper scope was not available when tested.
2135
2136 2005-01-10  Ken Kocienda  <kocienda@apple.com>
2137
2138         Reviewed by John
2139
2140         Fix for this bug:
2141
2142         <rdar://problem/3946852> Option-e goes to next line
2143
2144         * khtml/editing/htmlediting.cpp:
2145         (khtml::ReplaceSelectionCommand::doApply): Tweak case used to determine when merging content
2146         into the start line is done. We plan to change pretty substantially soon to better handle
2147         the problem described in <rdar://problem/3937352> Quote level not maintained when copied
2148         and pasted within a Mail message. In the meantime, this change does no harm, and fixes the bug.
2149
2150 2005-01-10  Ken Kocienda  <kocienda@apple.com>
2151
2152         Reviewed by Darin
2153
2154         Fix for this bug:
2155
2156         <rdar://problem/3907005> Applying block styles to a line of text can unexpectedly affect other lines
2157
2158         * khtml/editing/htmlediting.cpp:
2159         (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): New function which moves "paragraphs"
2160         to their own blocks if needed so that a block style can be applied.
2161         (khtml::CompositeEditCommand::isMailBlockquote): Moved this function to base class so it can be used more generally.
2162         (khtml::ApplyStyleCommand::applyBlockStyle): Pass a node instead of a block to addBlockStyleIfNeeded().
2163         (khtml::ApplyStyleCommand::addBlockStyleIfNeeded): Now accepts a node instead of a block for styling.
2164         This function also now calls moveParagraphContentsToNewBlockIfNecessary() to make sure that only the current
2165         paragraph is styled.
2166         (khtml::SetNodeAttributeCommand::doUnapply): Should not assert on undo if old value of attribute was null.
2167         Should remove attributue instead.
2168         * khtml/editing/htmlediting.h: Touch function declarations accordingly.
2169         * layout-tests/editing/style/create-block-for-style-001-expected.txt: Added.
2170         * layout-tests/editing/style/create-block-for-style-001.html: Added.
2171         * layout-tests/editing/style/create-block-for-style-002-expected.txt: Added.
2172         * layout-tests/editing/style/create-block-for-style-002.html: Added.
2173         * layout-tests/editing/style/create-block-for-style-003-expected.txt: Added.
2174         * layout-tests/editing/style/create-block-for-style-003.html: Added.
2175         * layout-tests/editing/style/create-block-for-style-004-expected.txt: Added.
2176         * layout-tests/editing/style/create-block-for-style-004.html: Added.
2177         * layout-tests/editing/style/create-block-for-style-005-expected.txt: Added.
2178         * layout-tests/editing/style/create-block-for-style-005.html: Added.
2179         * layout-tests/editing/style/create-block-for-style-006-expected.txt: Added.
2180         * layout-tests/editing/style/create-block-for-style-006.html: Added.
2181         * layout-tests/editing/style/create-block-for-style-007-expected.txt: Added.
2182         * layout-tests/editing/style/create-block-for-style-007.html: Added.
2183         * layout-tests/editing/style/create-block-for-style-008-expected.txt: Added.
2184         * layout-tests/editing/style/create-block-for-style-008.html: Added.
2185         * layout-tests/editing/style/create-block-for-style-009-expected.txt: Added.
2186         * layout-tests/editing/style/create-block-for-style-009.html: Added.
2187         * layout-tests/editing/style/create-block-for-style-010-expected.txt: Added.
2188         * layout-tests/editing/style/create-block-for-style-010.html: Added.
2189         * layout-tests/editing/style/create-block-for-style-011-expected.txt: Added.
2190         * layout-tests/editing/style/create-block-for-style-011.html: Added.
2191         * layout-tests/editing/style/create-block-for-style-012-expected.txt: Added.
2192         * layout-tests/editing/style/create-block-for-style-012.html: Added.
2193         * layout-tests/editing/style/create-block-for-style-013-expected.txt: Added.
2194         * layout-tests/editing/style/create-block-for-style-013.html: Added.
2195
2196         Unrelated updates to these expected results.
2197         * layout-tests/editing/inserting/insert-div-007-expected.txt
2198         * layout-tests/editing/pasteboard/paste-text-013-expected.txt
2199
2200 2005-01-10  Chris Blumenberg  <cblu@apple.com>
2201
2202         Fixed: <rdar://problem/3936844> Mail: Messages with rich text do not print
2203
2204         Reviewed by john.
2205
2206         * khtml/rendering/render_flow.cpp:
2207         (RenderFlow::paintLines): don't do pagination work if printRect is not set
2208         * khtml/rendering/render_list.cpp:
2209         (RenderListMarker::paint): ditto
2210
2211 2005-01-10  David Harrison  <harrison@apple.com>
2212
2213         Reviewed by Darin.
2214
2215         * kwq/KWQTextUtilities.mm:
2216         (currentTextBreakLocaleID):
2217         Return empty string (AKA root locale) if locale pref can not be canonicalized.
2218
2219 2005-01-10  John Sullivan  <sullivan@apple.com>
2220
2221         Fixed broken Panther build.
2222
2223         * kwq/KWQTextUtilities.mm:
2224         (currentTextBreakLocaleID):
2225         This recently-added function was using code copy/pasted from CarbonCore UnicodeUtilities.
2226         That code used a Tiger-only function, CFLocaleCreateCanonicalLanguageIdentifierFromString.
2227         To fix the build, I added a BUILDING_ON_PANTHER #ifdef that avoids the call to the Tiger
2228         function. However, the Tiger-only code was wrong; the string generated using
2229         CFLocaleCreateCanonicalLanguageIdentifierFromString was not used at all, so I fixed that
2230         as well.
2231
2232 2005-01-09  David Harrison  <harrison@apple.com>
2233
2234         Reviewed by Ken Kocienda.
2235
2236         <rdar://problem/3786362> REGRESSION (Mail): pasted text loses one newline
2237
2238         * khtml/editing/htmlediting.cpp:
2239         (khtml::ReplaceSelectionCommand::doApply):
2240         Tune check for need for insertParagraphSeparator when hasInterchangeNewline.
2241
2242 2005-01-09  Darin Adler  <darin@apple.com>
2243
2244         Reviewed by Harrison.
2245
2246         - fixed <rdar://problem/3939176> select() method does not work on <input type=search>
2247
2248         * khtml/html/html_formimpl.cpp: (DOM::HTMLInputElementImpl::select): Changed if statement
2249         to a switch statement. Added SEARCH to the set of types that treat the renderer as a
2250         RenderLineEdit.
2251
2252 2005-01-09  David Harrison  <harrison@apple.com>
2253
2254         Reviewed by Ken Kocienda.
2255
2256         <rdar://problem/3905066> REGRESSION (Mail): Hitting return key with caret in front of space causes space to become lost (resize fixes)
2257
2258         Problem was that InsertParagraphSeparatorCommand::doApply() was not collapsing
2259         whitespace to a single non-breaking space when splitting a text node.
2260
2261         * khtml/editing/htmlediting.cpp:
2262         (khtml::InsertParagraphSeparatorCommand::doApply):
2263
2264 2005-01-08  Kevin Decker  <kdecker@apple.com>
2265
2266         Reviewed by Ken.
2267
2268         Fixed: <rdar://problem/3924219> Calling setOuterHTML: on a DOMHTMLHtmlElement can crash a program
2269
2270         * khtml/html/html_elementimpl.cpp:
2271         (HTMLElementImpl::setOuterHTML): Added a nil check for fragments who don't have parents.
2272
2273 2005-01-07  Maciej Stachowiak  <mjs@apple.com>
2274
2275         Reviewed by Darin.
2276
2277         <rdar://problem/3807144> REGRESSION (125-TOT): my bank's website doesn't work (Societe Generale, socgen.com)
2278
2279         * Khtml/khtml_part.cpp:
2280         (KHTMLPart::checkCompleted): If the document is NULL, assume this frame has
2281         not started loading yet so it could not possibly be finishing here...
2282         (KHTMLPart::stop): ...unless the part is explicitly stopped and there is still
2283         no document, in this case we must have hit an error or been loading a non-HTML
2284         frame.
2285         * khtml/khtml_part.h:
2286
2287 2005-01-08  David Harrison  <harrison@apple.com>
2288
2289         Reviewed by Maciej.
2290
2291         <rdar://problem/3943415> REGRESSION (Mail): double-clicking past word selects previous word instead of only blank space
2292
2293         * khtml/editing/selection.cpp:
2294         (khtml::Selection::validate):
2295         Tune word selections left/right choice to use right if on empty last line.
2296
2297 2005-01-07  David Harrison  <harrison@apple.com>
2298
2299         Reviewed by Darin.
2300
2301         <rdar://problem/3942619> AX: Support sentence ax attributes
2302
2303         Needed to use the unicode utilities properly.  Twas lame before.
2304
2305         * khtml/editing/visible_units.cpp:
2306         (khtml::previousBoundary):
2307         (khtml::nextBoundary):
2308         (khtml::startOfWord):
2309         (khtml::endOfWord):
2310         (khtml::previousWordPosition):
2311         (khtml::nextWordPosition):
2312         (khtml::startOfSentence):
2313         (khtml::endOfSentence):
2314         (khtml::previousSentencePosition):
2315         (khtml::nextSentencePosition):
2316         * kwq/KWQAccObject.mm:
2317         (-[KWQAccObject accessibilityAttributeNames]):
2318         (-[KWQAccObject accessibilityAttributeValue:]):
2319         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
2320         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
2321         * kwq/KWQTextUtilities.mm:
2322         (currentTextBreakLocaleID):
2323         (KWQFindSentenceBoundary):
2324         (KWQFindNextSentenceFromIndex):
2325
2326 2005-01-07  Ken Kocienda  <kocienda@apple.com>
2327
2328         Reviewed by Kevin
2329
2330         Fix for these bugs:
2331
2332         <rdar://problem/3939523> in some cases, text does not retain style info after pressing return twice
2333         <rdar://problem/3944492> after pressing return twice, text is bold when it shouldn't be
2334
2335         * khtml/editing/htmlediting.cpp:
2336         (khtml::InsertParagraphSeparatorCommand::setFullTypingStyleBeforeInsertion):
2337         Merge the typing style with the computed style for the current position. Fixes both bugs.
2338         * khtml/editing/htmlediting.h:
2339         * layout-tests/editing/inserting/insert-div-023-expected.txt: Added.
2340         * layout-tests/editing/inserting/insert-div-023.html: Added.
2341         * layout-tests/editing/inserting/insert-div-024-expected.txt: Added.
2342         * layout-tests/editing/inserting/insert-div-024.html: Added.
2343
2344 2005-01-07  David Hyatt  <hyatt@apple.com>
2345
2346         Fix for 3941364, make sure tables reset overflowWidth/Height when they lay out again.  Fixes the odd scrolling
2347         behavior on worldofwarcraft.com.
2348
2349         Reviewed by kevin
2350
2351         * khtml/rendering/render_table.cpp:
2352         (RenderTable::layout):
2353
2354 2005-01-06  David Hyatt  <hyatt@apple.com>
2355
2356         Fix for 3932418 and 3920998, assertion failures caused by stray inline content inside tables.  Bulletproof the hit testing to
2357         ignore inline flows in this case.
2358
2359         Reviewed by kevin
2360
2361         * khtml/rendering/render_block.cpp:
2362         (khtml::RenderBlock::nodeAtPoint):
2363
2364 2005-01-07  Ken Kocienda  <kocienda@apple.com>
2365
2366         Reviewed by Hyatt
2367
2368         Fix for this bug:
2369
2370         <rdar://problem/3848412> for forwarded message, tabbing to message view scrolls to bottom of view
2371
2372         * kwq/KWQKHTMLPart.mm:
2373         (KWQKHTMLPart::nextKeyViewInFrame): Don't scroll the focus node to visible if it is not
2374         in the document, or if it is not a descendent of the document element. In the case of the bug,
2375         since the selection has not yet been set up, the focus node passed here is the HTML element, and
2376         that does not yield a rect that is useful to us here. So now, in the case the bug mentions, we do nothing.
2377
2378 2005-01-06  Kevin Decker  <kdecker@apple.com>
2379
2380         Reviewed by mjs.
2381
2382         Fixed: <rdar://problem/3932215> REGRESSION (125-177): iFrame example at developer.apple.com fails in Safari
2383
2384         * khtml/khtml_part.cpp:
2385         (KHTMLPart::processObjectRequest): m_bComplete was never true for frames generated by Javascript due to our synchronous loading and as a result, scheduled redirects wouldn't fire in KHTMLPart::scheduleLocationChange().  By virtue of being an empty document, a document is complete.  In this special case it's safe at this point to call checkCompleted() which sets m_bComplete true.
2386
2387 === Safari-178 ===
2388
2389 2005-01-06  David Harrison  <harrison@apple.com>
2390
2391         Reviewed by Chris.
2392
2393         (addendum to previous checkin for this bug)
2394         <rdar://problem/3776056> AX: Editable HTML should not be viewed as AXStaticText
2395
2396         Fix line navigation.  Add AXUIElementForTextMarker.
2397
2398         * kwq/KWQAccObject.mm:
2399         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
2400         (-[KWQAccObject doAXUIElementForTextMarker:]):
2401         (-[KWQAccObject doAXLineForTextMarker:]):
2402         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
2403         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
2404         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
2405         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
2406         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
2407
2408 2005-01-06  Darin Adler  <darin@apple.com>
2409
2410         Reviewed by Ken.
2411
2412         - re-fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
2413
2414         (turns out the PLT regression was a false alarm)
2415
2416         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
2417         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
2418         much larger number.
2419         * khtml/html/htmlparser.cpp:
2420         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
2421         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
2422         to eliminate code that used ID_CLOSE_TAG for an array size.
2423         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
2424         that manages isindex to use deref instead of delete.
2425         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
2426         mistake of using ID_CLOSE_TAG for the array size too.
2427         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
2428         there and it would prevent custom tags from working. Added range check before using the forbidden
2429         tag array with the token ID since custom tags will use index values past the end of the array.
2430         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
2431         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
2432         createElement call is still here. Last time I left out a few form element types from this switch;
2433         fixed now.
2434         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
2435         document if getTagID fails; this creates a unique per-document ID.
2436
2437         * khtml/misc/htmltags.c: Regenerated.
2438         * khtml/misc/htmltags.h: Regenerated.
2439
2440         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
2441         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
2442         Also rewrote getTagName to work with the new scheme.
2443
2444 2005-01-06  David Harrison  <harrison@apple.com>
2445
2446         Fixed Panther build.  Also, do not advertize sentence support since it is incomplete.
2447
2448         * kwq/KWQAccObject.mm:
2449         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
2450         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
2451         (-[KWQAccObject accessibilityIsAttributeSettable:]):
2452
2453 2005-01-06  David Harrison  <harrison@apple.com>
2454
2455         Reviewed by Dave Hyatt.
2456
2457         <rdar://problem/3776056> AX: Editable HTML should not be viewed as AXStaticText
2458         <rdar://problem/3588548> AX: tabbing does not work correctly with the screen reader and a focused link; need AXFocusedUIElement to work
2459
2460         Many more AX attributes supported.  Numerous fixes to previous AX work.
2461
2462         * khtml/editing/visible_units.cpp:
2463         (khtml::startSentenceBoundary):
2464         (khtml::startOfSentence):
2465         (khtml::endSentenceBoundary):
2466         (khtml::endOfSentence):
2467         (khtml::previousSentencePositionBoundary):
2468         (khtml::previousSentencePosition):
2469         (khtml::nextSentencePositionBoundary):
2470         (khtml::nextSentencePosition):
2471         * khtml/editing/visible_units.h:
2472         * khtml/khtmlview.cpp:
2473         (KHTMLView::layout):
2474         * khtml/misc/helper.cpp:
2475         (khtml::findSentenceBoundary):
2476         (khtml::nextSentenceFromIndex):
2477         * khtml/misc/helper.h:
2478         * khtml/misc/htmltags.c:
2479         (hash_tag):
2480         (findTag):
2481         * khtml/rendering/render_container.cpp:
2482         (RenderContainer::removeChildNode):
2483         (RenderContainer::appendChildNode):
2484         (RenderContainer::insertChildNode):
2485         * khtml/rendering/render_object.cpp:
2486         (RenderObject::remove):
2487         * khtml/xml/dom_docimpl.cpp:
2488         (DocumentImpl::getAccObjectCache):
2489         (DocumentImpl::updateSelection):
2490         (DocumentImpl::close):
2491         (DocumentImpl::setFocusNode):
2492         (DocumentImpl::parentDocument):
2493         (DocumentImpl::topDocument):
2494         * khtml/xml/dom_docimpl.h:
2495         * kwq/KWQAccObject.mm:
2496         (-[KWQAccObject accessibilityShouldUseUniqueId]):
2497         (-[KWQAccObject detach]):
2498         (-[KWQAccObject anchorElement]):
2499         (-[KWQAccObject firstChild]):
2500         (-[KWQAccObject lastChild]):
2501         (-[KWQAccObject previousSibling]):
2502         (-[KWQAccObject nextSibling]):
2503         (-[KWQAccObject parentObject]):
2504         (-[KWQAccObject value]):
2505         (-[KWQAccObject accessibilityAttributeNames]):
2506         (-[KWQAccObject accessibilityPerformAction:]):
2507         (-[KWQAccObject textMarkerForVisiblePosition:]):
2508         (-[KWQAccObject visiblePositionForTextMarker:]):
2509         (-[KWQAccObject textMarkerRangeFromVisiblePositions:andEndPos:]):
2510         (-[KWQAccObject topDocument]):
2511         (-[KWQAccObject topRenderer]):
2512         (-[KWQAccObject topView]):
2513         (-[KWQAccObject accessibilityAttributeValue:]):
2514         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
2515         (-[KWQAccObject doAXLineForTextMarker:]):
2516         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
2517         (-[KWQAccObject doAXStringForTextMarkerRange:]):
2518         (-[KWQAccObject doAXTextMarkerForPosition:]):
2519         (-[KWQAccObject doAXBoundsForTextMarkerRange:]):
2520         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
2521         (-[KWQAccObject doAXTextMarkerRangeForUnorderedTextMarkers:]):
2522         (-[KWQAccObject doAXNextTextMarkerForTextMarker:]):
2523         (-[KWQAccObject doAXPreviousTextMarkerForTextMarker:]):
2524         (-[KWQAccObject doAXLeftWordTextMarkerRangeForTextMarker:]):
2525         (-[KWQAccObject doAXRightWordTextMarkerRangeForTextMarker:]):
2526         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
2527         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
2528         (-[KWQAccObject doAXSentenceTextMarkerRangeForTextMarker:]):
2529         (-[KWQAccObject doAXParagraphTextMarkerRangeForTextMarker:]):
2530         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
2531         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
2532         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
2533         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
2534         (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
2535         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
2536         (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
2537         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
2538         (-[KWQAccObject doAXLengthForTextMarkerRange:]):
2539         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
2540         (-[KWQAccObject accessibilityHitTest:]):
2541         (-[KWQAccObject accessibilityFocusedUIElement]):
2542         (-[KWQAccObject accessibilityIsAttributeSettable:]):
2543         (-[KWQAccObject doSetAXSelectedTextMarkerRange:]):
2544         (-[KWQAccObject setAccObjectID:]):
2545         (-[KWQAccObject removeAccObjectID]):
2546         * kwq/KWQAccObjectCache.h:
2547         * kwq/KWQAccObjectCache.mm:
2548         (KWQAccObjectCache::setAccObject):
2549         (KWQAccObjectCache::removeAccObject):
2550         (KWQAccObjectCache::visiblePositionForTextMarker):
2551         (KWQAccObjectCache::postNotificationToTopWebArea):
2552         (KWQAccObjectCache::postNotification):
2553         (KWQAccObjectCache::handleFocusedUIElementChanged):
2554         * kwq/KWQKHTMLPart.mm:
2555         (KWQKHTMLPart::respondToChangedContents):
2556         * kwq/KWQTextUtilities.h:
2557         * kwq/KWQTextUtilities.mm:
2558         (KWQFindNextWordFromIndex):
2559         (KWQFindSentenceBoundary):
2560         (KWQFindNextSentenceFromIndex):
2561         * kwq/WebCoreBridge.mm:
2562         (-[WebCoreBridge accessibilityTree]):
2563
2564 2005-01-05  Darin Adler  <darin@apple.com>
2565
2566         Reviewed by Ken.
2567
2568         - re-landing a subset of my custom tag change that does not fix the bug, but also does
2569           not introduce a performance regression
2570
2571         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
2572         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
2573
2574         * khtml/editing/htmlediting.cpp:
2575         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
2576         per-document tags and is just better all around for things like the document.
2577         (khtml::debugNode): Ditto.
2578         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
2579         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
2580         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
2581         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
2582
2583         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
2584         non-HTML elements to be nested as desired.
2585
2586         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
2587         * khtml/misc/htmlhashes.cpp:
2588         (khtml::getTagID): Changed return type to unsigned short.
2589         (khtml::getAttrID): Ditto.
2590
2591         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
2592         * khtml/xml/dom_docimpl.cpp:
2593         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
2594         Also updated for a few tags that the parser handled but this did not.
2595
2596         * kwq/KWQRenderTreeDebug.cpp:
2597         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
2598         the tag ID directly, which only works for standard nodes.
2599         (operator<<): Update to call getTagName.
2600         (nodePositionRelativeToRoot): Ditto.
2601         (writeSelection): Ditto.
2602
2603 2005-01-05  Ken Kocienda  <kocienda@apple.com>
2604
2605         Reviewed by Hyatt
2606
2607         Fix for this bug:
2608
2609         <rdar://problem/3941203> REGRESSION (Mail): Paste inserts content in wrong place
2610
2611         * khtml/editing/htmlediting.cpp:
2612         (khtml::ReplaceSelectionCommand::doApply): Some cleanup and refinement of the concepts used to make
2613         this operation work correctly, particularly in the logic to figure out whether to merge content, and
2614         also performing merges.
2615         * khtml/editing/visible_position.cpp:
2616         (khtml::isFirstVisiblePositionInBlock): Simplification of test used to make this determination.
2617         * khtml/editing/visible_units.cpp:
2618         (khtml::isStartOfParagraph): New helper, used in khtml::ReplaceSelectionCommand::doApply().
2619         (khtml::isEndOfParagraph): Ditto.
2620         * khtml/editing/visible_units.h: Declare new functions.
2621
2622 2005-01-04  Ken Kocienda  <kocienda@apple.com>
2623
2624         Reviewed by John
2625
2626         Fix for this bug:
2627
2628         <rdar://problem/3926522> Pressing return in a quoted block inserts too many newlines
2629
2630         * khtml/editing/htmlediting.cpp:
2631         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Call to insertBlockPlaceholderIfNeeded()
2632         for the block we are adding was being done at the wrong time, which led to the placeholder remaining
2633         in the document when it was not needed. This resulted in the extra space reported in the bug.
2634
2635 2005-01-04  Kevin Decker  <kdecker@apple.com>
2636
2637         Reviewed by Hyatt.
2638
2639         Fixed: <rdar://problem/3936879> nil-deref, crash in InlineFlowBox::nodeAtPoint closing DHTML menus at hrweb.apple.com
2640
2641         * khtml/rendering/render_block.cpp:
2642         (khtml::RenderBlock::fillInlineSelectionGaps): Added a nil check. If there is no selection, don't try to get the selection's containing block. If we do, we'll crash.
2643
2644 2005-01-04  David Hyatt  <hyatt@apple.com>
2645
2646         Fix for 3937203, force an update of the dashboard regions when scrollbars are added/removed.
2647
2648         Reviewed by kevin
2649
2650         * khtml/rendering/render_layer.cpp:
2651         (RenderLayer::updateScrollInfoAfterLayout):
2652
2653 2005-01-04  Ken Kocienda  <kocienda@apple.com>
2654
2655         Reviewed by John
2656
2657         Fix for these two bugs:
2658
2659         <rdar://problem/3938935> REGRESSION (Mail): Pasting into an empty document mangles content
2660         <rdar://problem/3939148> REGRESSION (Mail): Pasting mistakenly reverses lines
2661
2662         * khtml/editing/htmlediting.cpp:
2663         (khtml::ReplaceSelectionCommand::doApply): For 3938935, add one more case to handle an empty document; merge
2664         neither start nor end. For 3939148, improve the code which adjusts the insertion point during
2665         the process of pasting. It formerly handled only one of the possible cases.
2666         * layout-tests/editing/pasteboard/paste-text-015-expected.txt: Added.
2667         * layout-tests/editing/pasteboard/paste-text-015.html: Added.
2668
2669 2005-01-04  David Hyatt  <hyatt@apple.com>
2670
2671         Fix for 3936571, placeholder attribute should work for normal inputs for Dashboard.
2672
2673         Reviewed by john
2674
2675         * khtml/rendering/render_form.cpp:
2676         (RenderLineEdit::updateFromElement):
2677         * kwq/KWQLineEdit.mm:
2678         (QLineEdit::setPlaceholderString):
2679
2680 2005-01-04  David Hyatt  <hyatt@apple.com>
2681
2682         Fix for 3830936, hang on changeforamerica.com.  Make sure to ignore the style not yet available option when
2683         returning pseudo-styles.
2684
2685         Reviewed by john
2686
2687         * khtml/css/cssstyleselector.cpp:
2688         (khtml::CSSStyleSelector::matchRulesForList):
2689         (khtml::CSSStyleSelector::pseudoStyleForElement):
2690         * khtml/rendering/render_style.cpp:
2691         (RenderStyle::addPseudoStyle):
2692
2693 2005-01-04  Darin Adler  <darin@apple.com>
2694
2695         - rolled out my custom tag name change again -- it caused a 1 ms PLT regression
2696
2697         * khtml/css/cssstyleselector.cpp:
2698         * khtml/editing/htmlediting.cpp:
2699         * khtml/editing/selection.cpp:
2700         * khtml/editing/visible_position.cpp:
2701         * khtml/html/dtd.cpp:
2702         * khtml/html/htmlparser.cpp:
2703         * khtml/html/htmlparser.h:
2704         * khtml/html/htmltokenizer.cpp:
2705         * khtml/misc/htmlhashes.cpp:
2706         * khtml/misc/htmlhashes.h:
2707         * khtml/misc/htmltags.c:
2708         * khtml/misc/htmltags.h:
2709         * khtml/misc/maketags:
2710         * khtml/xml/dom_docimpl.cpp:
2711         * khtml/xml/dom_docimpl.h:
2712         * khtml/xml/dom_nodeimpl.cpp:
2713         * khtml/xml/dom_position.cpp:
2714         * kwq/KWQRenderTreeDebug.cpp:
2715
2716 2005-01-04  Ken Kocienda  <kocienda@apple.com>
2717
2718         Reviewed by John
2719
2720         Fix for this bug:
2721
2722         <rdar://problem/3927554> Style info applied to remainder of document after a newline is entered
2723
2724         * khtml/editing/htmlediting.cpp:
2725         (khtml::InsertParagraphSeparatorCommand::doApply): Clean up and simplification in code that inserts
2726         a paragraph separator. The bug was all about applying styles to the new paragraph that did not need
2727         to be applied. Now the code will detect when at the end of a style run and will not move and apply
2728         that ending style to the new paragraph, though it will place that style into the typing style. This
2729         seems to match NSText behavior.
2730         * layout-tests/editing/inserting/insert-div-021-expected.txt: This test result exhibited the bug fixed
2731         here. The only reason it was not noticed is that the erroneously copied inline was a span, and so did
2732         not have any visible effect on the document.
2733         * layout-tests/editing/inserting/insert-div-022-expected.txt: Added.
2734         * layout-tests/editing/inserting/insert-div-022.html: Added.
2735
2736 2005-01-04  David Hyatt  <hyatt@apple.com>
2737
2738         Fix for 3904562, make sure to clear the outline dictionary when drawing so that random focus rings dont get drawn
2739         all over the place.
2740
2741         Reviewed by john
2742
2743         * khtml/rendering/render_flow.cpp:
2744         (RenderFlow::paintLines):
2745
2746 2005-01-03  David Hyatt  <hyatt@apple.com>
2747
2748         Fix for float painting regressions 3932524, 3931664, and 3933068.  Make the noPaint flag setting more
2749         robust and make it work regardless of which objects get a layout or not.
2750
2751         Reviewed by mjs
2752
2753         * khtml/rendering/render_block.cpp:
2754         (khtml::RenderBlock::insertFloatingObject):
2755         (khtml::RenderBlock::addOverhangingFloats):
2756         (khtml::RenderBlock::addIntrudingFloats):
2757         * khtml/rendering/render_block.h:
2758         (khtml::RenderBlock::FloatingObject::FloatingObject):
2759
2760 2005-01-03  Maciej Stachowiak  <mjs@apple.com>
2761
2762         Reviewed by Kevin.
2763
2764         <rdar://problem/3935390> Tiger 8A341: nil deref crash in DOM::NodeImpl::traverseNextNode
2765
2766         * khtml/html/html_miscimpl.cpp:
2767         (HTMLCollectionImpl::traverseNextItem): Assert that the starting point is not nil,
2768         it should not ever be (now).
2769         (HTMLCollectionImpl::item): When traversing items stop when we hit
2770         nil, meaning the end to avoid triggering above assert (formerly crash).
2771
2772 2005-01-03  Maciej Stachowiak  <mjs@apple.com>
2773
2774         Reviewed by John and Kevin.
2775
2776         <rdar://problem/3870317> REGRESSION(125.9-125.11) broken behavior at test.profoundlearning.com - used to
2777
2778         * khtml/ecma/kjs_window.cpp:
2779         (Window::get): Look up frame names before buitin window properties
2780         to match other browsers. This regressed because we added a builtin
2781         "toolbar" property but this site had a frame with that name.
2782
2783 2004-12-21  Maciej Stachowiak  <mjs@apple.com>
2784
2785         Reviewed by Darin.
2786
2787         <rdar://problem/3888931> frame naming allows malicious site to bring up a window when you click on a link in another
2788
2789         Added opener bridge method to help WebKit implement security check
2790         for named frame visibility.
2791
2792         * khtml/khtml_part.h:
2793         * kwq/WebCoreBridge.h:
2794         * kwq/WebCoreBridge.mm:
2795         (-[WebCoreBridge opener]):
2796
2797 2005-01-03  Ken Kocienda  <kocienda@apple.com>
2798
2799         Reviewed by John
2800
2801         Fix for this bug:
2802
2803         <rdar://problem/3933926> Tiger8A341: Mail crashes while forwarding embedded HTML message in -[WebCoreBridge ensureSelectionVisible]
2804
2805         * kwq/WebCoreBridge.mm:
2806         (-[WebCoreBridge ensureSelectionVisible]): Put in some null checks to prevent crash experienced in bug.
2807
2808 2005-01-03  David Hyatt  <hyatt@apple.com>
2809
2810         Fix for 3936881, make sure positioned objects prooperly update y-position.
2811
2812         Reviewed by john
2813
2814         * khtml/rendering/render_block.cpp:
2815         (khtml::RenderBlock::layoutPositionedObjects):
2816
2817 2005-01-03  Ken Kocienda  <kocienda@apple.com>
2818
2819         Reviewed by Harrison
2820
2821         Fix for this bug:
2822
2823         <rdar://problem/3928250> REGRESSION (Mail): Typing style lost after hitting return key
2824
2825         * khtml/editing/htmlediting.cpp:
2826         (khtml::InsertLineBreakCommand::preservesTypingStyle): Now implemented, returning yes for this command.
2827         (khtml::InsertParagraphSeparatorCommand::InsertParagraphSeparatorCommand): Initialize new typing style member variable.
2828         (khtml::InsertParagraphSeparatorCommand::~InsertParagraphSeparatorCommand): Deref new typing style member variable.
2829         (khtml::InsertParagraphSeparatorCommand::preservesTypingStyle): Now implemented, returning yes for this command.
2830         (khtml::InsertParagraphSeparatorCommand::setFullTypingStyleBeforeInsertion): New function to set the typing style
2831         (khtml::InsertParagraphSeparatorCommand::calculateAndSetTypingStyleAfterInsertion): Function called after the <p>
2832         insertion is done. This function diffs the style created in setFullTypingStyleBeforeInsertion() with the style
2833         of the new <p> and only sets those styles needed to preserve the style in effect before the insertion.
2834         (khtml::InsertParagraphSeparatorCommand::doApply): Call new functions.
2835         (khtml::TypingCommand::preservesTypingStyle): Now yes for inserting line breaks and paragraphs.
2836          * khtml/editing/htmlediting.h: Declare new functions.
2837
2838 2004-12-25  Kevin Decker  <kdecker@apple.com>
2839
2840         Reviewed by Hyatt.
2841
2842         Fixed: <rdar://problem/3505072> hang in KHTMLParser::parseToken (consulting.soroos.net)
2843
2844         * khtml/html/dtd.cpp: <label> needed a much higher priority such that it will close block elements.  Malformed HTML (ugh) is the only reason why we need to do this.
2845
2846 2004-12-23  Darin Adler  <darin@apple.com>
2847
2848         Reviewed by Ken.
2849
2850         - re-fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
2851
2852         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
2853         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
2854
2855         * khtml/editing/htmlediting.cpp:
2856         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
2857         per-document tags and is just better all around for things like the document.
2858         (khtml::debugNode): Ditto.
2859         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
2860         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
2861         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
2862         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
2863
2864         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
2865         non-HTML elements to be nested as desired.
2866
2867         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
2868         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
2869         much larger number.
2870         * khtml/html/htmlparser.cpp:
2871         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
2872         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
2873         to eliminate code that used ID_CLOSE_TAG for an array size.
2874         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
2875         that manages isindex to use deref instead of delete.
2876         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
2877         mistake of using ID_CLOSE_TAG for the array size too.
2878         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
2879         there and it would prevent custom tags from working. Added range check before using the forbidden
2880         tag array with the token ID since custom tags will use index values past the end of the array.
2881         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
2882         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
2883         createElement call is still here. Last time I left out a few form element types from this switch;
2884         fixed now.
2885         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
2886         document if getTagID fails; this creates a unique per-document ID.
2887
2888         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
2889         * khtml/misc/htmlhashes.cpp:
2890         (khtml::getTagID): Changed return type to unsigned short.
2891         (khtml::getAttrID): Ditto.
2892
2893         * khtml/misc/htmltags.c: Regenerated.
2894         * khtml/misc/htmltags.h: Regenerated.
2895
2896         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
2897         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
2898         Also rewrote getTagName to work with the new scheme.
2899
2900         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
2901         * khtml/xml/dom_docimpl.cpp:
2902         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
2903         Also updated for a few tags that the parser handled but this did not.
2904
2905         * kwq/KWQRenderTreeDebug.cpp:
2906         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
2907         the tag ID directly, which only works for standard nodes.
2908         (operator<<): Update to call getTagName.
2909         (nodePositionRelativeToRoot): Ditto.
2910         (writeSelection): Ditto.
2911
2912 === Safari-177 ===
2913
2914 2004-12-22  Darin Adler  <darin@apple.com>
2915
2916         - rolled out my custom tag name change -- it broke amazon.com
2917
2918         * khtml/css/cssstyleselector.cpp:
2919         * khtml/editing/htmlediting.cpp:
2920         * khtml/editing/selection.cpp:
2921         * khtml/editing/visible_position.cpp:
2922         * khtml/html/dtd.cpp:
2923         * khtml/html/htmlparser.cpp:
2924         * khtml/html/htmlparser.h:
2925         * khtml/html/htmltokenizer.cpp:
2926         * khtml/misc/htmlhashes.cpp:
2927         * khtml/misc/htmlhashes.h:
2928         * khtml/misc/htmltags.c:
2929         * khtml/misc/htmltags.h:
2930         * khtml/misc/maketags:
2931         * khtml/xml/dom_docimpl.cpp:
2932         * khtml/xml/dom_docimpl.h:
2933         * khtml/xml/dom_nodeimpl.cpp:
2934         * khtml/xml/dom_position.cpp:
2935         * kwq/KWQRenderTreeDebug.cpp:
2936
2937 2004-12-22  David Harrison  <harrison@apple.com>
2938
2939         Reviewed by Darin Adler.
2940
2941         * khtml/editing/selection.cpp:
2942         (khtml::Selection::validate):
2943         The selecting/deselecting bad behavior is because the Selection code that expands by words
2944         had an inaccurate test for being at the end of the document (where double-clicking needs
2945         to select the last word).  Fixed that check.
2946
2947 2004-12-22  Adele Amchan  <adele@apple.com>
2948
2949         Reviewed by Chris.
2950
2951         Fix for <rdar://problem/3911650> tabs at safeway.com stop working after a while
2952
2953         * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::scrollToAnchor): removed call to cancelRedirection
2954           so that we match Firefox and WinIE behavior.
2955
2956 2004-12-22  Darin Adler  <darin@apple.com>
2957
2958         Reviewed by Ken.
2959
2960         - fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
2961
2962         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
2963         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
2964
2965         * khtml/editing/htmlediting.cpp:
2966         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
2967         per-document tags and is just better all around for things like the document.
2968         (khtml::debugNode): Ditto.
2969         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
2970         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
2971         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
2972         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
2973
2974         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
2975         non-HTML elements to be nested as desired.
2976
2977         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
2978         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
2979         much larger number.
2980         * khtml/html/htmlparser.cpp:
2981         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
2982         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
2983         to eliminate code that used ID_CLOSE_TAG for an array size.
2984         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
2985         that manages isindex to use deref instead of delete.
2986         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
2987         mistake of using ID_CLOSE_TAG for the array size too.
2988         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
2989         there and it would prevent custom tags from working. Added range check before using the forbidden
2990         tag array with the token ID since custom tags will use index values past the end of the array.
2991         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
2992         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
2993         createElement call is still here.
2994         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
2995         document if getTagID fails; this creates a unique per-document ID.
2996
2997         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
2998         * khtml/misc/htmlhashes.cpp:
2999         (khtml::getTagID): Changed return type to unsigned short.
3000         (khtml::getAttrID): Ditto.
3001
3002         * khtml/misc/htmltags.c: Regenerated.
3003         * khtml/misc/htmltags.h: Regenerated.
3004
3005         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
3006         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
3007         Also rewrote getTagName to work with the new scheme.
3008
3009         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
3010         * khtml/xml/dom_docimpl.cpp:
3011         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
3012         Also updated for a few tags that the parser handled but this did not.
3013
3014         * kwq/KWQRenderTreeDebug.cpp:
3015         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
3016         the tag ID directly, which only works for standard nodes.
3017         (operator<<): Update to call getTagName.
3018         (nodePositionRelativeToRoot): Ditto.
3019         (writeSelection): Ditto.
3020
3021 2004-12-21  David Harrison  <harrison@apple.com>
3022
3023         Reviewed by Ken Kocienda.
3024
3025         <rdar://problem/3924934> REGRESSION: double click at end of line selects start of next line
3026
3027         Problem was the TextIterator was not handling exitNode() from a P block properly.
3028
3029         * khtml/editing/visible_text.cpp:
3030         (khtml::TextIterator::TextIterator):
3031         Add new param that specifies whether the iterator is for content or for searching.
3032         Search iterators do not prevent newlines at the beginning.
3033         (khtml::TextIterator::advance):
3034         Added some comments.
3035         (khtml::TextIterator::handleTextNode):
3036         Added some comments.
3037         (khtml::TextIterator::exitNode):
3038         Emit newline for P (and other) blocks with position following the block, instead of the m_lastTextNode.
3039         (khtml::TextIterator::emitCharacter):
3040         Added some comments.
3041         (khtml::TextIterator::range):
3042         Added some comments.
3043         (khtml::CharacterIterator::CharacterIterator):
3044         Specify search type TextIterator.
3045         (khtml::CharacterIterator::advance):
3046         * khtml/editing/visible_text.h:
3047         (khtml::):
3048         Add new TextIterator::TextIterator param that specifies whether the iterator is for content or for searching.
3049         * khtml/editing/visible_units.cpp:
3050         (khtml::nextWordBoundary):
3051         Specify search type TextIterator.
3052
3053 2004-12-21  David Harrison  <harrison@apple.com>
3054
3055         Reviewed by Ken Kocienda.
3056
3057         <rdar://problem/3924695> REGRESSION (Mail): double-clicking past end of line shows no selection, should select to EOL
3058
3059         Problem was that RenderText::setSelectionState did not handle the SelectionStart case where start and end are the end of the line.
3060         Fixed by pretending the start == end-1 in that situation, as long as end > 0.
3061
3062         * khtml/rendering/render_text.cpp:
3063         (RenderText::setSelectionState)
3064
3065 2004-12-21  Maciej Stachowiak  <mjs@apple.com>
3066
3067         Reviewed by John.
3068
3069         <rdar://problem/3929187> WebKit needs to restrict access to certain window operations by domain
3070
3071         * khtml/ecma/kjs_window.cpp:
3072         (Window::get): Change most window functions to be restricted by
3073         XSS domain check.
3074
3075 2004-12-21  Ken Kocienda  <kocienda@apple.com>
3076
3077         Reviewed by John
3078
3079         Fix for this bug:
3080
3081         <rdar://problem/3928305> selecting an entire line and typing over causes new inserted text at top of document
3082
3083         * khtml/editing/htmlediting.cpp:
3084         (khtml::DeleteSelectionCommand::insertPlaceholderForAncestorBlockContent): New function to detect case
3085         formerly undetected and unhandled. This is the crux of the bug fix.
3086         (khtml::DeleteSelectionCommand::doApply): Call insertPlaceholderForAncestorBlockContent() during
3087         execution of command.
3088         * khtml/editing/htmlediting.h: Declare new function.
3089         * layout-tests/editing/deleting/delete-3928305-fix-expected.txt: Added.
3090         * layout-tests/editing/deleting/delete-3928305-fix.html: Added.
3091
3092 2004-12-21  Ken Kocienda  <kocienda@apple.com>
3093
3094         Reviewed by me
3095
3096         * khtml/editing/htmlediting.cpp:
3097         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Note to self: Must compile code before
3098         checking in (aka must return false from function returning bool).
3099
3100 2004-12-21  Ken Kocienda  <kocienda@apple.com>
3101
3102         Reviewed by John
3103
3104         Fix for this bug:
3105
3106         <rdar://problem/3927752> Crash in khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded(DOM::NodeImpl*)
3107
3108         * khtml/editing/htmlediting.cpp:
3109         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Added some null checks.
3110         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Ditto.
3111
3112 2004-12-21  Ken Kocienda  <kocienda@apple.com>
3113
3114         Reviewed by Darin
3115
3116         * khtml/editing/htmlediting.cpp:
3117         (khtml::ReplacementFragment::mergeStartNode): Refine concept of how this node is found based on
3118         further experiements.
3119         (khtml::ReplaceSelectionCommand::doApply): Add a special case for determining merges that need to
3120         be done if the insertion point is in an empty block.
3121         * layout-tests/editing/pasteboard/paste-text-012-expected.txt: Added.
3122         * layout-tests/editing/pasteboard/paste-text-012.html: Added.
3123         * layout-tests/editing/pasteboard/paste-text-013-expected.txt: Added.
3124         * layout-tests/editing/pasteboard/paste-text-013.html: Added.
3125         * layout-tests/editing/pasteboard/paste-text-014-expected.txt: Added.
3126         * layout-tests/editing/pasteboard/paste-text-014.html: Added.
3127
3128 2004-12-21  Darin Adler  <darin@apple.com>
3129
3130         Reviewed by Ken.
3131
3132         - fixed <rdar://problem/3899133> text search in a Safari window takes a very long time on Tiger updates page (and some other pages)
3133
3134         * khtml/editing/visible_text.h: Add an "offset base node" parameter to emitCharacter, and also
3135         add a field to track it. Must make a few things mutable so we can update them in the range accessor.
3136         * khtml/editing/visible_text.cpp:
3137         (khtml::TextIterator::advance): Pass in base node and offsets rather than computing actual offsets
3138         using the node's index. We only compute the node index if actually asked for the range.
3139         (khtml::TextIterator::handleTextNode): Pass 0 for base node and set base node to 0 when setting
3140         up the offsets.
3141         (khtml::TextIterator::handleTextBox): Ditto.
3142         (khtml::TextIterator::handleReplacedElement): Pass base node and set base node instead of calling
3143         nodeIndex.
3144         (khtml::TextIterator::handleNonTextNode): Pass 0 for offset.
3145         (khtml::TextIterator::exitNode): More of the same.
3146         (khtml::TextIterator::emitCharacter): Ditto.
3147         (khtml::TextIterator::range): If an offset base node is stored, then get its node index, and then
3148         add that in to the offsets. Doing the work here guarantees it's done only once when doing a text search.
3149
3150         - another small fix
3151
3152         * khtml/khtml_part.cpp: Removed SPEED_DEBUG define. Not sure why it was on.
3153
3154 === Safari-176 ===
3155
3156 2004-12-20  Ken Kocienda  <kocienda@apple.com>
3157
3158         Reviewed by Darin
3159
3160         * khtml/editing/htmlediting.cpp:
3161         (khtml::DeleteSelectionCommand::moveNodesAfterNode): My one-liner for this bug introduced layout test
3162         regressions: <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
3163         Rolling out until I can develop a real fix.
3164
3165 2004-12-20  David Harrison  <harrison@apple.com>
3166
3167         Reviewed by Dave Hyatt.
3168
3169         Initial checkin of AXTextMarkerRef support.
3170
3171         * khtml/xml/dom_docimpl.cpp:
3172         (DocumentImpl::updateSelection):
3173         (DocumentImpl::setFocusNode):
3174         * kwq/KWQAccObject.h:
3175         * kwq/KWQAccObject.mm:
3176         (-[KWQAccObject detach]):
3177         (-[KWQAccObject anchorElement]):
3178         (-[KWQAccObject addChildrenToArray:]):
3179         (-[KWQAccObject accessibilityAttributeNames]):
3180         (-[KWQAccObject accessibilityActionDescription:]):
3181         (-[KWQAccObject accessibilityPerformAction:]):
3182         (-[KWQAccObject textMarkerRangeFromMarkers:andEndMarker:]):
3183         (-[KWQAccObject textMarkerForVisiblePosition:]):
3184         (-[KWQAccObject visiblePositionForTextMarker:]):
3185         (-[KWQAccObject AXTextMarkerRangeCopyStartMarkerWrapper:]):
3186         (-[KWQAccObject AXTextMarkerRangeCopyEndMarkerWrapper:]):
3187         (-[KWQAccObject visiblePositionForStartOfTextMarkerRange:]):
3188         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
3189         (-[KWQAccObject accessibilityAttributeValue:]):
3190         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
3191         (-[KWQAccObject textMarkerRangeFromVisiblePositions:andEndPos:]):
3192         (-[KWQAccObject getSelectedTextMarkerRange]):
3193         (-[KWQAccObject doAXLineForTextMarker:]):
3194         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
3195         (-[KWQAccObject doAXStringForTextMarkerRange:]):
3196         (-[KWQAccObject doAXNextTextMarkerForTextMarker:]):
3197         (-[KWQAccObject doAXPreviousTextMarkerForTextMarker:]):
3198         (-[KWQAccObject doAXLeftWordTextMarkerRangeForTextMarker:]):
3199         (-[KWQAccObject doAXRightWordTextMarkerRangeForTextMarker:]):
3200         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
3201         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
3202         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
3203         (-[KWQAccObject accessibilityFocusedUIElement]):
3204         (-[KWQAccObject clearChildren]):
3205         (-[KWQAccObject accObjectID]):
3206         (-[KWQAccObject setAccObjectID:]):
3207         (-[KWQAccObject removeAccObjectID]):
3208         * kwq/KWQAccObjectCache.h:
3209         * kwq/KWQAccObjectCache.mm:
3210         (KWQAccObjectCache::KWQAccObjectCache):
3211         (KWQAccObjectCache::~KWQAccObjectCache):
3212         (KWQAccObjectCache::getAccObjectID):
3213         (KWQAccObjectCache::removeAccObjectID):
3214         (KWQAccObjectCache::textMarkerForVisiblePosition):
3215         (KWQAccObjectCache::visiblePositionForTextMarker):
3216
3217 2004-12-19  Darin Adler  <darin@apple.com>
3218
3219         * kwq/KWQFoundationExtras.h: (KWQCFAutorelease): Added nil check.
3220
3221 2004-12-19  Darin Adler  <darin@apple.com>
3222
3223         Reviewed by Kevin.
3224
3225         - a garbage collection fix
3226
3227         * kwq/KWQFoundationExtras.h: (KWQCFAutorelease): Added a more-extensive comment and fixed the BUILDING_ON_PANTHER
3228         #ifndef to use #if (in practice, either is OK).
3229         * kwq/KWQKURL.mm: (KURL::getNSURL): Use KWQCFAutorelease instead of autorelease.
3230
3231 2004-12-17  David Hyatt  <hyatt@apple.com>
3232
3233         Fix for 3923255, specified percentage heights of divs with overflow auto inside tables not honored.
3234
3235         Reviewed by kocienda
3236
3237         * khtml/rendering/render_box.cpp:
3238         (RenderBox::calcPercentageHeight):
3239         * khtml/rendering/render_table.cpp:
3240         (RenderTableSection::layoutRows):
3241
3242 2004-12-17  David Harrison  <harrison@apple.com>
3243
3244         Reviewed by Ken Kocienda.
3245
3246         <rdar://problem/3924930> REGRESSION: triple click does not select to end of line
3247
3248         * khtml/editing/visible_units.cpp:
3249         (khtml::endOfParagraph):
3250         When includeLineBreak is true, allow traversal to next node after enclosingBlockFlowElement.
3251
3252 2004-12-17  Ken Kocienda  <kocienda@apple.com>
3253
3254         Reviewed by Vicki
3255
3256         * khtml/editing/htmlediting.cpp:
3257         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Missed deleting some now-obsolete expcetion
3258         codes when I added the createBreakElement() function.
3259
3260 2004-12-17  Richard Williamson   <rjw@apple.com>
3261
3262         Set the floor of max cacheable object size to 40K.  This restores
3263         the long standing floor.  Lower floor deleteriously impacts the PLT.
3264         Reviewed by Hyatt.
3265
3266         * khtml/khtml_part.cpp:
3267         (KHTMLPart::checkCompleted):
3268         * khtml/misc/loader.cpp:
3269         (Cache::setSize):
3270
3271 2004-12-17  Ken Kocienda  <kocienda@apple.com>
3272
3273         Reviewed by Hyatt
3274
3275         Fix for this bug:
3276
3277         <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
3278
3279         * khtml/editing/htmlediting.cpp:
3280         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Stop merging nodes when a <br> is hit.
3281         Formerly checked only for block boundary.
3282
3283 2004-12-17  David Harrison  <harrison@apple.com>
3284
3285         Reviewed by Darin.
3286
3287                 Add KWQCFAutorelease for autoreleasing CF objects.
3288         * kwq/KWQFoundationExtras.h:
3289         (KWQCFAutorelease):
3290         New.
3291
3292 2004-12-17  David Harrison  <harrison@apple.com>
3293
3294         Reviewed by Darin.
3295
3296                 Fix GC compatibility in getNSString.
3297
3298         * kwq/KWQString.mm:
3299         (QString::getNSString):
3300                 Use NSString allocator instead of CFString, so that autorelease works under GC.
3301
3302 2004-12-17  Ken Kocienda  <kocienda@apple.com>
3303
3304         Reviewed by Hyatt
3305
3306         Fix for this bug:
3307
3308         <rdar://problem/3890973> REGRESSION (Mail): Deleting reorders remaining text
3309
3310         * khtml/editing/htmlediting.cpp:
3311         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Added check to ensure that moving content will not
3312         move it after the <body> element.
3313
3314 2004-12-17  Ken Kocienda  <kocienda@apple.com>
3315
3316         Reviewed by me
3317
3318         Added new layout tests covering cases from recent bug fixes.
3319
3320         * layout-tests/editing/inserting/insert-div-018-expected.txt: Added.
3321         * layout-tests/editing/inserting/insert-div-018.html: Added.
3322         * layout-tests/editing/inserting/insert-div-019-expected.txt: Added.
3323         * layout-tests/editing/inserting/insert-div-019.html: Added.
3324         * layout-tests/editing/inserting/insert-div-020-expected.txt: Added.
3325         * layout-tests/editing/inserting/insert-div-020.html: Added.
3326         * layout-tests/editing/inserting/insert-div-021-expected.txt: Added.
3327         * layout-tests/editing/inserting/insert-div-021.html: Added.
3328
3329 2004-12-17  Ken Kocienda  <kocienda@apple.com>
3330
3331         Reviewed by me
3332
3333         Changes in layout resulting from giving <p> elements no margin rather than 0.1em margin.
3334
3335         * layout-tests/editing/inserting/insert-div-013-expected.txt
3336         * layout-tests/editing/inserting/insert-div-014-expected.txt
3337
3338 2004-12-17  Ken Kocienda  <kocienda@apple.com>
3339
3340         Reviewed by John
3341
3342         Fix for this bug:
3343
3344         <rdar://problem/3925317> Increase quote in email and with cursor below the quote the new line appears above the quote
3345
3346         * khtml/editing/htmlediting.cpp:
3347         (khtml::InsertParagraphSeparatorCommand::doApply):
3348         * khtml/editing/visible_position.cpp: Refine rules for whether to use the starting node or the starting
3349         block as the reference node for the insertion of the new block.
3350         (khtml::isLastVisiblePositionInBlock): Tweak rules again. Descendants of following blocks should answer true.
3351         I worked this all out on the whiteboard this time. This should be the last tweak.
3352
3353 2004-12-17  Kevin Decker  <kdecker@apple.com>
3354
3355         Reviewed by Ken.
3356
3357         Fixed <rdar://problem/3824438> Need a clean way for Dashboard to detect when an XML parsing error occurs
3358
3359         * khtml/xml/xml_tokenizer.cpp:
3360         (khtml::XMLTokenizer::insertErrorMessageBlock): In the xml error report, instead of a generic <div>, use <parsererror> to match Mozilla.
3361
3362 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3363
3364         Reviewed by Chris
3365
3366         Fix for this bug:
3367
3368         <rdar://problem/3924888> REGRESSION (Mail): Hitting return key at end of line does not insert visible newline
3369
3370         * khtml/editing/htmlediting.cpp:
3371         (khtml::InsertParagraphSeparatorCommand::doApply): Tweaked code I added just before to fix 3924486, so that
3372         it checks the downstream node for the starting position.
3373
3374 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3375
3376         Reviewed by Chris
3377
3378         Added new createBreakElement() function that makes creating a <br> element a one-liner.
3379         Converted all the code that used to call the DocumentImpl, and dealt with the exceptionCode,
3380         over to this new helper.
3381
3382         * khtml/editing/htmlediting.cpp:
3383         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Now uses new helper function.
3384         (khtml::InsertLineBreakCommand::doApply): Ditto.
3385         (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
3386         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Ditto.
3387         (khtml::ReplaceSelectionCommand::doApply): Ditto.
3388         (khtml::createBreakElement): New helper.
3389         * khtml/editing/htmlediting.h: Ditto.
3390
3391 2004-12-16  David Hyatt  <hyatt@apple.com>
3392
3393         Fix a bug where vertical-align values that depended on their parent's value went crazy if the parent was
3394         vertical-align top or bottom.  The bug is 3771007, bankofamerica's ebills page.
3395
3396         Reviewed by kocienda
3397
3398         * khtml/rendering/render_object.cpp:
3399         (RenderObject::getVerticalPosition):
3400
3401 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3402
3403         Reviewed by John
3404
3405         Fix for these bugs:
3406
3407         <rdar://problem/3924486> REGRESSION (Mail): Hitting return key does not insert visible newline
3408         <rdar://problem/3924579> REGRESSION (Mail): After deleting, hitting return key does not insert visible newline
3409
3410         * khtml/editing/htmlediting.cpp:
3411         (khtml::InsertParagraphSeparatorCommand::doApply): For the first bug, detect when inserting a <p> will
3412         make a visible <br> collapse. For the second bug, move the code that inserts the new <p> after
3413         a check which may move the node used as the reference node for the insertion. Also change this code to
3414         insert the <p> after the last sibling of the starting position, to move it past the nodes we're going
3415         to want to shift into the new <p>.
3416
3417 2004-12-16  Darin Adler  <darin@apple.com>
3418
3419         Reviewed by Ken.
3420
3421         - use <p> elements rather than <br> when pasting plain text; also use margin of 0 rather than 0.1 em
3422           after talking this over with Dave and Ken
3423
3424         * khtml/editing/htmlediting.h: Added createDefaultParagraphElement and createBlockPlaceholderElement.
3425         * khtml/editing/htmlediting.cpp:
3426         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Use new createBlockPlaceholderElement
3427         helper so this can share code with the plain-text conversion code.
3428         (khtml::InsertParagraphSeparatorCommand::createParagraphElement): Use new createDefaultParagraphElement
3429         helper so this can share code with the plain-text conversion code.
3430         (khtml::createDefaultParagraphElement): Added.
3431         (khtml::createBlockPlaceholderElement): Added.
3432
3433         * khtml/editing/markup.cpp: (khtml::createFragmentFromText): Use paragraphs rather than
3434         <br> elements for pasting plain text.
3435
3436         - regenerated these files with the newer gperf
3437
3438         * khtml/css/cssproperties.c: Regenerated.
3439         * khtml/css/cssvalues.c: Regenerated.
3440         * khtml/misc/htmlattrs.c: Regenerated.
3441         * khtml/misc/htmltags.c: Regenerated.
3442
3443 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3444
3445         Reviewed by Hyatt
3446
3447         Fix for this bug:
3448
3449         <rdar://problem/3924291> REGRESSION (Mail): Crash deleting content following a <p> element
3450
3451         * khtml/xml/dom_nodeimpl.cpp:
3452         (NodeImpl::enclosingInlineElement): Fixed problem in this function where recursive search for
3453         parent that is not an inline would skip past previous siblings of nodes which were blocks. Now
3454         it stops looking in this situation. This problem eventually caused the delete code to try to
3455         merge a node under a descendent.
3456
3457 2004-12-16  Adele Amchan  <adele@apple.com>
3458
3459         Change by Richard, reviewed by me.
3460
3461         Fix for: <rdar://problem/3923983> background image decoding prevents page with 0Kb image from finishing to load
3462
3463         * khtml/misc/loader.cpp: (CachedImage::data): Added call to notifyFinished so the 0Kb image case finishes loading.
3464
3465 2004-12-16  Adele Amchan  <adele@apple.com>
3466
3467         Reviewed by Chris.
3468
3469         Fix for: <rdar://problem/3534824> VIP: some navigation links on safeway.com don't work due to a JavaScript quirk
3470
3471         Added a call to ObjectImp::get before Window::get just returns Undefined.
3472         This was preventing us from getting the toString function from a Window object.
3473
3474         * khtml/ecma/kjs_window.cpp: (Window::get):
3475
3476 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3477
3478         Reviewed by John
3479
3480         Fix for this bug:
3481
3482         <rdar://problem/3918351> REGRESSION (Mail, 173-175+): Return before first char of line leaves insertion point in wrong place
3483
3484         * khtml/editing/htmlediting.cpp:
3485         (khtml::InsertParagraphSeparatorCommand::doApply): Basically, did a rewrite of this function
3486         to do a better job than it was doing before. Added several test cases to prove I am on a
3487         better track.
3488         * khtml/editing/visible_position.cpp:
3489         (khtml::isFirstVisiblePositionInBlock): Tweaked the rules a bit to fix an issue very similar to the
3490         leaving-the-bar-node case problem I just fixed in a recent checkin (relevant markup: <p>foo</p>bar).
3491         This function was returning true for the first position in "bar". Wrong. Also tightened up other
3492         rule: Should not report true when relationship between blocks cannot be determined.
3493         (khtml::isLastVisiblePositionInBlock): Tightened up rule as above: Should not report true
3494         when relationship between blocks cannot be determined.
3495         * layout-tests/editing/inserting/insert-div-010-expected.txt: Added.
3496         * layout-tests/editing/inserting/insert-div-010.html: Added.
3497         * layout-tests/editing/inserting/insert-div-011-expected.txt: Added.
3498         * layout-tests/editing/inserting/insert-div-011.html: Added.
3499         * layout-tests/editing/inserting/insert-div-012-expected.txt: Added.
3500         * layout-tests/editing/inserting/insert-div-012.html: Added.
3501         * layout-tests/editing/inserting/insert-div-013-expected.txt: Added.
3502         * layout-tests/editing/inserting/insert-div-013.html: Added.
3503         * layout-tests/editing/inserting/insert-div-014-expected.txt: Added.
3504         * layout-tests/editing/inserting/insert-div-014.html: Added.
3505         * layout-tests/editing/inserting/insert-div-015-expected.txt: Added.
3506         * layout-tests/editing/inserting/insert-div-015.html: Added.
3507         * layout-tests/editing/inserting/insert-div-016-expected.txt: Added.
3508         * layout-tests/editing/inserting/insert-div-016.html: Added.
3509         * layout-tests/editing/inserting/insert-div-017-expected.txt: Added.
3510         * layout-tests/editing/inserting/insert-div-017.html: Added.
3511
3512 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3513
3514         Reviewed by me
3515
3516         Added a layout test based on my last checkin.
3517
3518         * layout-tests/editing/selection/move-by-word-001-expected.txt: Added.
3519         * layout-tests/editing/selection/move-by-word-001.html: Added.
3520
3521 2004-12-16  Ken Kocienda  <kocienda@apple.com>
3522
3523         Reviewed by John
3524
3525         Fix for this bug:
3526
3527         <rdar://problem/3917929> REGRESSION (Mail): Command-left-arrow leaves insertion point too high after specific steps
3528
3529         * khtml/editing/visible_text.cpp:
3530         (khtml::SimplifiedBackwardsTextIterator::advance): Add new check for leaving a text node and iterating
3531         backwards into a different block that is an descendent of the block containing the text node (as in leaving
3532         the "bar" node in this example: <p>foo</p>bar).
3533         (khtml::SimplifiedBackwardsTextIterator::handleNonTextNode): Call new emitNewlineForBROrText helper.
3534         (khtml::SimplifiedBackwardsTextIterator::emitNewlineForBROrText): Factor out code from handleNonTextNode, since
3535         it is called from there, and now also from advance().
3536
3537 2004-12-15  Darin Adler  <darin@apple.com>
3538
3539         Reviewed by Ken.
3540
3541         - fixed problem where plain-text would put a blank line between each <p> even when they have no margins
3542
3543         * khtml/editing/visible_text.cpp: (khtml::TextIterator::exitNode): Add more checks and only set the
3544         "add one more newline" flag if the margin is sufficient. A more complete fix would ignore the node
3545         type altogether and use the render tree instead.
3546
3547 2004-12-14  John Sullivan  <sullivan@apple.com>
3548
3549         Reviewed by Ken.
3550
3551         - rest of WebCore fix for <rdar://problem/3790011> undoable operations all say "Undo"
3552         in the menu, no specific action names
3553
3554         * khtml/editing/edit_actions.h: new header, contains EditAction enum
3555         (renamed from HTMLEditAction, formerly in htmlediting.h)
3556
3557         * ForwardingHeaders/editing/edit_actions.h: new forwarding header
3558
3559         * khtml/editing/htmlediting.h:
3560         removed HTMLEditAction enum in favor of including edit_actions.h; added khtml::
3561         namespace to EditAction usages; added m_editingAction ivar to ApplyStyleCommand
3562
3563         * khtml/editing/htmlediting.cpp:
3564         (khtml::EditCommandPtr::editingAction):
3565         update for enum name change
3566         (khtml::EditCommand::editingAction):
3567         ditto
3568         (khtml::ApplyStyleCommand::ApplyStyleCommand):
3569         added editingAction parameter to this constructor
3570         (khtml::ApplyStyleCommand::editingAction):
3571         return new ivar
3572         (khtml::DeleteSelectionCommand::editingAction):
3573         update for enum name change
3574         (khtml::MoveSelectionCommand::editingAction):
3575         ditto
3576         (khtml::TypingCommand::editingAction):
3577         ditto
3578         (khtml::ReplaceSelectionCommand::editingAction):
3579         ditto
3580
3581         * khtml/khtml_part.h:
3582         added EditAction parameter to applyStyle and computeAndSetTypingStyle
3583         * khtml/khtml_part.cpp:
3584         (KHTMLPart::computeAndSetTypingStyle):
3585         added EditAction parameter
3586         (KHTMLPart::applyStyle):
3587         ditto
3588
3589         * kwq/KWQKHTMLPart.mm:
3590         (KWQKHTMLPart::registerCommandForUndoOrRedo):
3591         do the cast from EditAction to WebUndoAction a different way to match other code
3592
3593         * kwq/WebCoreBridge.h:
3594         * kwq/WebCoreBridge.mm:
3595         (-[WebCoreBridge setTypingStyle:withUndoAction:]):
3596         added WebUndoAction parameter, passed into ApplyStyleCommand constructor
3597         (-[WebCoreBridge applyStyle:withUndoAction:]):
3598         ditto
3599
3600         * WebCore.pbproj/project.pbxproj:
3601         updated for new files
3602
3603 2004-12-14  David Hyatt  <hyatt@apple.com>
3604
3605         Fix for 3562458, rowspan and colspan converted to ints so that large values will work for them.  Remove
3606         the ridiculous 1024 limit on the span values.
3607
3608         * khtml/html/html_tableimpl.cpp:
3609         (HTMLTableCellElementImpl::parseHTMLAttribute):
3610         * khtml/rendering/render_table.cpp:
3611         (RenderTableCell::collapsedBottomBorder):
3612         * khtml/rendering/render_table.h:
3613         (khtml::RenderTableCell::colSpan):
3614         (khtml::RenderTableCell::setColSpan):
3615         (khtml::RenderTableCell::rowSpan):
3616         (khtml::RenderTableCell::setRowSpan):
3617         (khtml::RenderTableCol::span):
3618         (khtml::RenderTableCol::setSpan):
3619
3620 2004-12-14  David Hyatt  <hyatt@apple.com>
3621
3622         Make sure <col> and <colgroup> can have spans updated dynamically as well.
3623
3624         Reviewed by rjw
3625
3626         * khtml/html/html_tableimpl.cpp:
3627         (HTMLTableColElementImpl::parseHTMLAttribute):
3628         * khtml/rendering/render_table.cpp:
3629         (RenderTableCell::updateFromElement):
3630         (RenderTableCol::updateFromElement):
3631
3632 2004-12-14  David Hyatt  <hyatt@apple.com>
3633
3634         Fix for 3833123, setting a cell's colspan does not update rendering like it should.
3635
3636         Reviewed by rjw
3637
3638         * khtml/html/html_tableimpl.cpp:
3639         (HTMLTableCellElementImpl::parseHTMLAttribute):
3640         * khtml/rendering/render_table.cpp:
3641         (RenderTableCell::RenderTableCell):
3642         (RenderTableCell::updateFromElement):
3643
3644 2004-12-14  Chris Blumenberg  <cblu@apple.com>
3645
3646         Fixed: <rdar://problem/3864536> crash copying text from other apps and pasting in a sticky in dashboard
3647
3648         Reviewed by hyatt.
3649
3650         * khtml/editing/markup.cpp:
3651         (khtml::createFragmentFromText): ref and deref the fragment since calling appendChild can completely deref it
3652
3653 2004-12-14  David Hyatt  <hyatt@apple.com>
3654
3655         Fix for 3785211. Make sure to do a layout with the old position before doing a layout after changing to the new position.
3656         This is actually a regression from my positioned object DHTML optimization.
3657
3658         Reviewed by kocienda
3659
3660         * khtml/rendering/render_object.cpp:
3661         (RenderObject::setStyle):
3662
3663 2004-12-14  David Hyatt  <hyatt@apple.com>
3664
3665         Make sure the class attribute works when newlines are present in the attribute.
3666
3667         Reviewed by kocienda
3668
3669         * khtml/html/html_elementimpl.cpp:
3670         (HTMLNamedAttrMapImpl::parseClassAttribute):
3671
3672 2004-12-14  David Hyatt  <hyatt@apple.com>
3673
3674         Fix for 3724938, float element is duplicated and paints twice.  The logic for when to paint floats was
3675         actually fairly screwed up.  This patch simplifies the logic and makes addOverhangingFloats easier to
3676         read by splitting it into two separate functions.
3677
3678         Reviewed by kocienda
3679
3680         * khtml/rendering/render_block.cpp:
3681         (khtml::RenderBlock::layoutBlockChildren):
3682         (khtml::RenderBlock::clearFloats):
3683         (khtml::RenderBlock::addOverhangingFloats):
3684         (khtml::RenderBlock::addIntrudingFloats):
3685         * khtml/rendering/render_block.h:
3686
3687 2004-12-14  John Sullivan  <sullivan@apple.com>
3688
3689         Reviewed by Ken.
3690
3691         - added Undo action names for Cut, Paste, and Drag
3692
3693         * khtml/editing/htmlediting.h:
3694         * khtml/editing/htmlediting.cpp:
3695         (khtml::DeleteSelectionCommand::editingAction):
3696         overridden to return HTMLEditActionCut
3697         (khtml::MoveSelectionCommand::editingAction):
3698         overridden to return HTMLEditActionDrag
3699         (khtml::ReplaceSelectionCommand::editingAction):
3700         overridden to return HTMLEditActionPaste
3701
3702 2004-12-14  John Sullivan  <sullivan@apple.com>
3703
3704         Reviewed by Ken.
3705
3706         - architecture for WebCore part of fix for <rdar://problem/3790011> undoable operations all say "Undo" in the menu,
3707         no specific action names
3708
3709         The remaining step is to make each EditCommand subclass override editingAction() to return an
3710         appropriate value. (Unfortunately the mapping between subclass and user-distinguishable action
3711         is not completely straightforward, so this next step isn't trivial.)
3712
3713         * khtml/editing/htmlediting.h:
3714         new enum for HTMLEditAction
3715
3716         * khtml/editing/htmlediting.cpp:
3717         (khtml::EditCommandPtr::editingAction):
3718         new method, calls through to EditCommand
3719         (khtml::EditCommand::editingAction):
3720         new method for subclasses to override, returns HTMLEditActionUnspecified at this level
3721         (khtml::TypingCommand::editingAction):
3722         proof of concept override, returns HTMLEditActionTyping
3723
3724         * kwq/KWQKHTMLPart.h:
3725         declare new private bottleneck method registerCommandForUndoOrRedo
3726         * kwq/KWQKHTMLPart.mm:
3727         (KWQKHTMLPart::registerCommandForUndoOrRedo):
3728         new bottleneck method to reduce code duplication; now calls over the bridge
3729         to get the localized string to use for the Undo action name
3730         (KWQKHTMLPart::registerCommandForUndo):
3731         now calls new bottleneck method
3732         (KWQKHTMLPart::registerCommandForRedo):
3733         now calls new bottleneck method
3734
3735         * kwq/WebCoreBridge.h:
3736         new enum for WebUndoAction, maps directly to HTMLEditAction.
3737         Declaration of nameForUndoAction:
3738
3739 2004-12-14  Darin Adler  <darin@apple.com>
3740
3741         Reviewed by John.
3742
3743         - added a bunch of missing nil checks; our old version of inherits used to work for nil (by accident)
3744
3745         * khtml/rendering/render_frames.cpp:
3746         (RenderPart::~RenderPart): Check widget for nil.
3747         (RenderPart::setWidget): Ditto.
3748         (RenderFrame::slotViewCleared): Ditto.
3749         (RenderPartObject::slotViewCleared): Ditto.
3750
3751 2004-12-13  Darin Adler  <darin@apple.com>
3752
3753         Reviewed by Ken.
3754
3755         - moved markup-related functions into new sources files in the editing directory
3756         - removed all of the uses of dynamic_cast, preparing to turn off RTTI to make our code smaller and slightly faster
3757
3758         * ForwardingHeaders/editing/markup.h: Added.
3759         * khtml/editing/markup.h: Added.
3760         * khtml/editing/markup.cpp: Added.
3761
3762         * WebCore.pbproj/project.pbxproj: Added markup.h/cpp.
3763
3764         * khtml/dom/dom_node.cpp: (Node::toHTML): Call createMarkup since there's no toHTML in NodeImpl any more.
3765         * khtml/html/html_elementimpl.cpp:
3766         (HTMLElementImpl::innerHTML): Changed to call createMarkup.
3767         (HTMLElementImpl::outerHTML): Ditto.
3768
3769         * khtml/ecma/kjs_window.cpp:
3770         (Window::retrieveWindow): Comment out assert that uses dynamic_cast.
3771         (Window::retrieveActive): Ditto.
3772
3773         * khtml/editing/htmlediting.h: Added forward class declaration needed now that I removed one elsewhere.
3774         * khtml/xml/dom_docimpl.h: Ditto.
3775
3776         * khtml/khtml_part.cpp:
3777         (KHTMLPart::slotDebugDOMTree): Use createMarkup instead of toHTML.
3778         (KHTMLPart::processObjectRequest): Use inherits instead of dynamic_cast.
3779
3780         * khtml/rendering/render_image.cpp: (RenderImage::paint): Add an explicit QChar conversion so this code
3781         still works even with the additional replace overloads added to QString.
3782         * kwq/KWQTextCodec.mm: (QTextCodec::fromUnicode): Ditto.
3783
3784         * khtml/rendering/render_object.h: Removed the version of arenaDelete that does not take an object
3785         base pointer, because it used dynamic_cast in its implementation. Made the other version public.
3786         * khtml/rendering/render_object.cpp: Ditto.
3787         * khtml/rendering/render_replaced.cpp: (RenderWidget::deref): Pass object base pointer to arenaDelete.
3788
3789         * khtml/xml/dom2_rangeimpl.h: Removed extra parameters from toHTML, and unneeded includes and declarations.
3790         * khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::toHTML): Changed to call createMarkup, and moved all
3791         the support code into markup.cpp.
3792
3793         * khtml/xml/dom_nodeimpl.h: Moved toHTML and related functions into markup.cpp.
3794         * khtml/xml/dom_nodeimpl.cpp: Ditto.
3795
3796         * khtml/xml/dom_position.cpp:
3797         (DOM::startPosition): Implemented the version of this that takes a RangeImpl. Also added null checks
3798         so these return null positions rather than raising exceptions.
3799         (DOM::endPosition): Ditto.
3800
3801         * khtml/khtmlview.h: Added an APPLE_CHANGES function so inherits can detect this class without dynamic_cast.
3802         * kwq/KWQFrame.h: Ditto.
3803         * kwq/KWQFrame.mm: (QFrame::isQFrame): Ditto.
3804         * kwq/KWQKPartsPart.h: Ditto.
3805         * kwq/KWQKPartsPart.mm: (KParts::ReadOnlyPart::isKPartsReadOnlyPart): Ditto.
3806         * kwq/KWQScrollView.h: Ditto.
3807         * kwq/KWQScrollView.mm: (QScrollView::isQScrollView): Ditto.
3808         * kwq/KWQKHTMLPart.h: Ditto.
3809         * kwq/KWQKHTMLPart.mm:
3810         (KHTMLView::isKHTMLView): Ditto.
3811         (KWQKHTMLPart::setTitle): Added an explicit QChar conversion so this code still works even with the additional
3812         replace overloads added to QString.
3813         (KWQKHTMLPart::setStatusBarText): Ditto.
3814         (KWQKHTMLPart::runJavaScriptAlert): Ditto.
3815         (KWQKHTMLPart::runJavaScriptConfirm): Ditto.
3816         (KWQKHTMLPart::runJavaScriptPrompt): Ditto.
3817         (KWQKHTMLPart::attributedString): Ditto.
3818         (KWQKHTMLPart::isCharacterSmartReplaceExempt): Ditto.
3819         (KWQKHTMLPart::isKHTMLPart): That dynamic_cast thing (see above).
3820
3821         * kwq/KWQObject.h: Added virtual methods for the few cases where we need dynamic_cast-like behavior.
3822         * kwq/KWQObject.mm:
3823         (QObject::inherits): Changed to not use dynamic cast.
3824         (QObject::isKHTMLPart): Added. Returns false.
3825         (QObject::isKHTMLView): Ditto.
3826         (QObject::isKPartsReadOnlyPart): Ditto.
3827         (QObject::isQFrame): Ditto.
3828         (QObject::isQScrollView): Ditto.
3829
3830         * kwq/KWQRenderTreeDebug.cpp:
3831         (write): Changed to use inherits rather than dynamic_cast.
3832         (writeSelection): Ditto.
3833
3834         * kwq/KWQSlot.mm: (KWQSlot::call): Call through to the version with just a job pointer parameter rather
3835         than going straight on to the "no parameters at all" version.
3836
3837         * kwq/KWQString.h:
3838         * kwq/KWQString.mm: (QString::replace): Added overloads.
3839         * kwq/WebCoreBridge.mm:
3840         (-[WebCoreBridge markupStringFromNode:nodes:]): Changed to call functions in markup.h.
3841         (-[WebCoreBridge markupStringFromRange:nodes:]): Ditto.
3842         (-[WebCoreBridge selectedString]): Added an explicit QChar conversion so this code still works even with
3843         the additional replace overloads added to QString.
3844         (-[WebCoreBridge stringForRange:]): Ditto.
3845         (-[WebCoreBridge copyDOMNode:copier:]): Changed to call functions in markup.h.
3846         (-[WebCoreBridge elementAtPoint:]): QChar conversion.
3847         (-[WebCoreBridge documentFragmentWithMarkupString:baseURLString:]): Changed to call functions in markup.h.
3848         (-[WebCoreBridge documentFragmentWithText:]): Changed to call functions in markup.h.
3849
3850 2004-12-13  Ken Kocienda  <kocienda@apple.com>
3851
3852         Reviewed by John
3853
3854         Fix for this bug:
3855
3856         <rdar://problem/3917956> REGRESSION (Mail): pasting can leave insertion point inside pasted text
3857
3858         * khtml/editing/htmlediting.cpp:
3859         (khtml::ReplaceSelectionCommand::doApply): Fix coding mistake. Calculations of bool flag based on
3860         leading and trailing whitespace positions was reversed! I must have introduced this error recently
3861         when changing around this code.
3862
3863 2004-12-13  David Hyatt  <hyatt@apple.com>
3864