WebCore/ChangeLog-2005-08-23

   1 2005-02-21  Darin Adler  <darin@apple.com>
   2
   3         - fixed Panther deployment build
   4
   5         * kwq/KWQTextEdit.mm: (QTextEdit::setScrollBarModes): Put more stuff inside #if.
   6
   7 2005-02-21  David Harrison  <harrison@apple.com>
   8
   9         Reviewed by Darin.
  10
  11         <rdar://problem/3943090> REGRESSION (Mail): Spelling underline incompletely erased following certain steps
  12
  13         * khtml/rendering/render_text.cpp:
  14         (InlineTextBox::paintMarker):
  15         Make sure underline is placed within the text bounds.
  16
  17         * kwq/KWQPainter.h:
  18         * kwq/KWQPainter.mm:
  19         (QPainter::misspellingLineThickness):
  20         * kwq/WebCoreTextRenderer.h:
  21         Add misspellingLineThickness for use by InlineTextBox::paintMarker.
  22
  23 2005-02-21  Darin Adler  <darin@apple.com>
  24
  25         Reviewed by John.
  26
  27         - fixed <rdar://problem/4012978> -[DOMRange markupString] crashes when range contains only a text node with a single space
  28
  29         * khtml/editing/markup.cpp: (khtml::createMarkup): Added updateLayout calls, and added a missing
  30         nil check.
  31
  32 2005-02-21  Darin Adler  <darin@apple.com>
  33
  34         Reviewed by John.
  35
  36         - fixed <rdar://problem/4005435> Safari hung while pasting text into a <textarea> (Panther-only)
  37
  38         * kwq/KWQTextEdit.mm: (QTextEdit::setScrollBarModes): Don't setAutohidesScrollers:YES on Panther.
  39
  40 2005-02-21  Ken Kocienda  <kocienda@apple.com>
  41
  42         Reviewed by John
  43
  44         Fix for this bug:
  45
  46         <rdar://problem/4015499> REGRESSION (186-187): pasted quoted text starting with a blank line increases quote level of pasted text when pasted
  47
  48         * khtml/editing/markup.cpp:
  49         (khtml::markup): Changed over to ASSERT instead of assert.
  50         (khtml::createMarkup): The issue was that the code to add parents all the way back to the common ancestor
  51         block did not check for blocks whose markup had already been added, and could result in adding markup for
  52         nodes twice (hence the additional and erroneous quote level). Now there is a new check that will
  53         only add markup for those nodes before the start of the selection range. This fixes the bug.
  54         (khtml::createFragmentFromMarkup): Changed over to ASSERT instead of assert.
  55         (khtml::createFragmentFromText): Ditto.
  56
  57 2005-02-21  Ken Kocienda  <kocienda@apple.com>
  58
  59         Reviewed by John
  60
  61         Fix for this bug:
  62
  63         <rdar://problem/4015494> REGRESSION (186-187) <cr> removed if quoted word is copy/pasted on line immediately above quoted text
  64
  65         * khtml/editing/htmlediting.cpp:
  66         (khtml::ReplaceSelectionCommand::doApply): Relatively new check designed to remove a <br> element when
  67         that element was on a line by itself did not do an adequate check for this condition, causing the
  68         failure described in the bug. Fixed.
  69
  70 2005-02-19  Ken Kocienda  <kocienda@apple.com>
  71
  72         Reviewed by Maciej
  73
  74         * khtml/editing/htmlediting.cpp:
  75         (khtml::matchNearestBlockquoteColorString): New function which abstracts away the CSS property we use
  76         for this pseudo-color.
  77         (khtml::ReplacementFragment::removeBlockquoteColorsIfNeeded): Set the new nearestBlockquoteColorString
  78         psuedo-color here, rather than removing colors as was done before.
  79         (khtml::ReplaceSelectionCommand::applyStyleToInsertedNodes): Check for the pseudo-color in the desired
  80         style, and determine the real color based on the nearest blockquote (or none) to the node.
  81         (khtml::nearestMailBlockquote): Name change from closestMailBlockquote to match new property better.
  82
  83         All the rest of this change is the mechanical coding you need to do to add a new CSS property, in this
  84         case, -khtml-match-nearest-mail-blockquote-color.
  85
  86         * khtml/css/css_computedstyle.cpp:
  87         (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue)
  88         * khtml/css/cssparser.cpp:
  89         (CSSParser::parseValue):
  90         * khtml/css/cssproperties.c:
  91         (hash_prop):
  92         (findProp):
  93         * khtml/css/cssproperties.h:
  94         * khtml/css/cssproperties.in:
  95         * khtml/css/cssstyleselector.cpp:
  96         (khtml::CSSStyleSelector::applyProperty)
  97         * khtml/css/cssvalues.c:
  98         (findValue)
  99         * khtml/css/cssvalues.h:
 100         * khtml/css/cssvalues.in:
 101         * khtml/rendering/render_style.h:
 102         (khtml::RenderStyle::matchNearestMailBlockquoteColor)
 103         (khtml::RenderStyle::setMatchNearestMailBlockquoteColor)
 104         (khtml::RenderStyle::initialMatchNearestMailBlockquoteColor)
 105
 106 2005-02-19  Ken Kocienda  <kocienda@apple.com>
 107
 108         Reviewed by Chris
 109
 110         Fix for these bugs:
 111
 112         <rdar://problem/4014228> REGRESSION (186-187) extra, uneditable lines inserted above and below a line of pasted quoted text
 113         <rdar://problem/4014393> REGRESSION (186-187) pasted quoted text gets extra <cr>s when pasted at top of document
 114
 115         * khtml/editing/htmlediting.cpp:
 116         (khtml::ReplacementFragment::ReplacementFragment): Part of a general refactoring of how
 117         the fragment is inserted into the document, rendered, and then tested for certain
 118         important pieces of information that are required for pasting.
 119         (khtml::ReplacementFragment::insertFragmentForTestRendering): New helper. Handles inserting
 120         the fragment nodes into the document.
 121         (khtml::ReplacementFragment::restoreTestRenderingNodesToFragment): Removes nodes from the
 122         document, and restores them to the fragment.
 123         (khtml::ReplacementFragment::computeStylesUsingTestRendering): Factored out code that
 124         did this before into its own function.
 125         (khtml::ReplacementFragment::removeUnrenderedNodesUsingTestRendering): Ditto.
 126         (khtml::ReplacementFragment::countRenderedBlocks): This is a real improvement, as it
 127         eliminates a major use of the isProbablyBlock() function. Now, the blocks that are
 128         counted are real, rendered blocks.
 129         (khtml::ReplacementFragment::removeStyleNodes): Made this function retain margin-zeroing
 130         CSS properties on paragraphs. This does two things: 1) It helps us to maintain good behavior
 131         in the short term while there are still versions of Mail out there that use <p> elements
 132         instead of <div> elements for new paragraphs; and 2) It will help to maintain the compatibility
 133         with other mail clients that use <p> elements for their paragraphs but render them themselves
 134         with no margins as the result of quirks.
 135         (khtml::ReplaceSelectionCommand::doApply): Do some work to fix up and improve the handling
 136         of blank lines, be they <p> elements or <br> elements, that can be removed after pasting. This,
 137         coupled with the refactoring, fixes 4014393.
 138         * khtml/editing/htmlediting.h: Updated for new functions.
 139         * layout-tests/editing/pasteboard/paste-text-010-expected.txt: Updated results, actually improved with this change.
 140         * layout-tests/editing/pasteboard/paste-text-011-expected.txt: Ditto.
 141
 142 2005-02-19  Kevin Decker  <kdecker@apple.com>
 143
 144         Reviewed by Chris.
 145
 146         Fixed <rdar://problem/4010765> Flash player can be used to arbitrarily open popup windows without user permission
 147
 148         Our window.open() policy is to refuse the <script>window.open(...)</script> case and allow the inline the <a href="javascript:window.open('foo')> case.  Clever advertisers at some point realized that by executing their Javascript through the Flash plugin, Safari would always treat their code as the inline case, and thus, they were able to work around our popup blocker.
 149
 150         * kwq/WebCoreBridge.h:
 151         * kwq/WebCoreBridge.mm:
 152         (-[WebCoreBridge stringByEvaluatingJavaScriptFromString:]): Changed this to invoke the new stringByEvaluatingJavaScriptFromString:forceUserGesture method below.
 153         (-[WebCoreBridge stringByEvaluatingJavaScriptFromString:forceUserGesture:]): Let the WebKit make the determination if this was a user originated gesture or not; we must no longer assume this is always the case.
 154
 155 2005-02-18  Chris Blumenberg  <cblu@apple.com>
 156
 157         Fixed: <rdar://problem/3951196> REGRESSION (Mail): too many levels of reply quotes after certain steps
 158
 159         Reviewed by kocienda.
 160
 161         * khtml/editing/htmlediting.cpp:
 162         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
 163         - If we find a new start node, update topBlockquote so we don't use too many block quotes for the contents following the new line.
 164         - Build up the list of ancestors after we've determined the actual topBlockquote.
 165         - Don't insert an extra new line if there is a new start node.
 166
 167 2005-02-19  Chris Blumenberg  <cblu@apple.com>
 168
 169         Fixed: <rdar://problem/3978461> smart paste is broken
 170
 171         Reviewed by kevin.
 172
 173         * khtml/editing/htmlediting.cpp:
 174         (khtml::ReplaceSelectionCommand::doApply): properly check for leading and trailing whitespace. These checks were incorrectly reversed. Also check if we're pasting at the beginning or end of a line. We should not insert spaces in either case.
 175
 176 2005-02-18  Adele Amchan  <adele@apple.com>
 177
 178         Reviewed by Darin.
 179
 180         Fix for <rdar://problem/3975568> REGRESSION(125-180)Australian Open pages have drawing problem
 181
 182         This patch fixes two problems caused by our added support for custom tags.  First, the layout problem at the sites mentioned in the bug
 183         was caused by custom tags within tables.  In checkChild, we needed to treat these tags as spans so they get placed correctly in the DOM tree.
 184         Also, we were indexing the tagPriority and endTag arrays with id values from the custom tags that were greater than the size of the array.  So now
 185         we have functions to check for the custom tags, and again, treat them as spans.  To avoid confusion, we changed the names of the arrays to
 186         endTagArray and tagPriorityArray.
 187
 188         * khtml/html/dtd.h: changed all uses of the endTag array to endTagArray for our new wrapper functions
 189         (DOM::tagPriority): added function to check array bounds and to treat custom tags as spans
 190         (DOM::endTagRequirement): added function to check array bounds and to treat custom tags as spans
 191         * khtml/html/dtd.cpp: changed name of endTag and tagPriority arrays to endTagArray and tagPriorityArray
 192         (DOM::checkChild): treat custom tags as spans during this check
 193         * khtml/html/html_elementimpl.cpp: changed all uses of the endTag array to the endTagRequirement function
 194         (HTMLElementImpl::createContextualFragment):
 195         (HTMLElementImpl::setInnerText):
 196         (HTMLElementImpl::setOuterText):
 197         (HTMLElementImpl::toString):
 198         * khtml/html/htmlparser.cpp: changed all uses of the endTag array to the endTagRequirement function and all uses of the tagPriority array to the tagPriority function.
 199         (KHTMLParser::parseToken):
 200         (KHTMLParser::insertNode):
 201         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): changed all uses of the endTag array to the endTagRequirement function
 202        * khtml/editing/markup.cpp: changed all uses of the endTag array to the endTagRequirement function
 203         (khtml::endMarkup):
 204         (khtml::markup):
 205
 206 2005-02-18  Jens Alfke  <jens@apple.com>
 207
 208         Reviewed by Hyatt.
 209
 210         Fixed build: Whoops, setNeedsLayout's parameter is NOT optional.
 211
 212         * khtml/html/html_objectimpl.cpp:
 213         (HTMLAppletElementImpl::setAllParamsAvailable):
 214
 215 2005-02-18  Jens Alfke  <jens@apple.com>
 216
 217         Reviewed by cblu, hyatt.
 218
 219         Fixes <rdar://problem/3603191> "REGRESSION: Applets not receiving all of the Applet Parameters in Java 1.4.1/1.4.2"
 220         Defer instantiation of Java applet plugin until after all of the <applet> tag's nested <param> tags have been parsed, otherwise the list of parameters passed to the applet is incomplete. The regression was introduced (says Dave) when the parser's close-tag notifications were removed in the name of performance.
 221
 222         * khtml/html/html_objectimpl.cpp:
 223         (HTMLAppletElementImpl::HTMLAppletElementImpl):
 224         (HTMLAppletElementImpl::getAppletInstance):
 225         (HTMLAppletElementImpl::setAllParamsAvailable):
 226         (HTMLAppletElementImpl::allParamsAvailable):
 227         * khtml/html/html_objectimpl.h:
 228         * khtml/html/htmlparser.cpp:
 229         (KHTMLParser::processCloseTag):
 230         * khtml/rendering/render_applet.cpp:
 231         (RenderApplet::createWidgetIfNecessary):
 232
 233 2005-02-18  Richard Williamson   <rjw@apple.com>
 234
 235         Fixed <rdar://problem/4006161> Tiger8A380: Widgets leak dashboard regions
 236
 237         We were leaking the DashboardRegionImpls.
 238
 239         Reviewed by Kevin.
 240
 241         * khtml/css/css_valueimpl.cpp:
 242         (DOM::CSSPrimitiveValueImpl::cleanup):
 243
 244 2005-02-18  Chris Blumenberg  <cblu@apple.com>
 245
 246         Fixed: <rdar://problem/3945271> REGRESSION (Mail): pasted plain text should pick up typing style instead of being unstyled
 247
 248         Reviewed by kocienda.
 249
 250         * khtml/editing/htmlediting.cpp:
 251         (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): don't set class on element returned by createStyleSpanElement since that's already done
 252         (khtml::ReplacementFragment::ReplacementFragment): take matchStyle param, don't call computeStylesAndRemoveUnrendered() if !matchStyle
 253         (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): take matchStyle param
 254         (khtml::ReplaceSelectionCommand::doApply): if m_matchStyle, use selection.start() as the insertion position and don't call applyStyleToInsertedNodes()
 255         * khtml/editing/htmlediting.h:
 256         * kwq/WebCoreBridge.h:
 257         * kwq/WebCoreBridge.mm:
 258         (-[WebCoreBridge replaceSelectionWithFragment:selectReplacement:smartReplace:matchStyle:]): take matchStyle param and pass it
 259         (-[WebCoreBridge replaceSelectionWithNode:selectReplacement:smartReplace:]): pass NO for matchStyle
 260         (-[WebCoreBridge replaceSelectionWithMarkupString:baseURLString:selectReplacement:smartReplace:]): pass NO for matchStyle
 261         (-[WebCoreBridge replaceSelectionWithText:selectReplacement:smartReplace:]): pass YES for matchStyle
 262
 263 2005-02-18  Ken Kocienda  <kocienda@apple.com>
 264
 265         Reviewed by Chris
 266
 267         Fix for these bugs:
 268
 269         <rdar://problem/4013025> Copy/Paste of quoted word results in removal of any following <cr>
 270         <rdar://problem/4013100> Copy/Paste quoted text and then decrease quote level does not change text color
 271
 272         For the most part, these bugs were caused by errors and lack of foresight on my part when
 273         I added the better paste code. Chalk these fixes up to the result of bake time.
 274
 275         * khtml/editing/htmlediting.cpp:
 276         (khtml::ReplacementFragment::ReplacementFragment): Need to move count of number of blocks in
 277         fragment after the call to remove unrendered nodes. Meant to do this before, but forgot to.
 278         (khtml::ReplacementFragment::removeStyleNodes): Need to remove inline styles from elements!
 279         Terrible omission now fixed.
 280         (khtml::ReplacementFragment::removeBlockquoteColorsIfNeeded): Remove blockquote colors for now.
 281         Code has a more extensive comment in it now to explain the difficulty, and the need for more
 282         study and changes.
 283         (khtml::ReplaceSelectionCommand::doApply): Need to call applyStyleToInsertedNodes() in the
 284         m_fragment.hasInterchangeNewline() case. This was just missed before.
 285         * layout-tests/editing/pasteboard/paste-text-011-expected.txt: Updated results, subtly different, but OK.
 286         * layout-tests/editing/pasteboard/paste-text-017-expected.txt: Updated for <p> to <div> change in test content.
 287         * layout-tests/editing/pasteboard/paste-text-017.html: Needed to change <p> to <div> to
 288         make this test go with the new design of using <div> tags for default paragraphs.
 289
 290 2005-02-18  David Hyatt  <hyatt@apple.com>
 291
 292         Fix for 3974263 (and possibly others).  Don't let fixed tables use maxint as their maxwidth when some
 293         cells have percentage values.
 294
 295         Reviewed by kocienda
 296
 297         * khtml/misc/arena.cpp:
 298         (ArenaAllocate):
 299         * khtml/rendering/table_layout.cpp:
 300         (FixedTableLayout::calcWidthArray):
 301         (FixedTableLayout::calcMinMaxWidth):
 302         (AutoTableLayout::layout):
 303
 304 2005-02-17  Darin Adler  <darin@apple.com>
 305
 306         Reviewed by Kevin.
 307
 308         - fixed <rdar://problem/3998627> WebKit crashes when deleting text in a modified div
 309
 310         * khtml/khtml_part.cpp: (KHTMLPart::selectionComputedStyle): Add a ref and deref to keep
 311         the element alive until it's added to the DOM tree.
 312         * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::styleForSelectionStart): Ditto.
 313
 314 2005-02-17  Richard Williamson   <rjw@apple.com>
 315
 316         Fixed <rdar://problem/3999467> when Osaka-Mono is specified as fixed width font, Osaka used instead
 317
 318         Fixed w/o introducing a performance regression.  Add early
 319         check for Osaka-Mono to avoid expensive call into WebKit.
 320
 321         Reviewed by Vicki.
 322
 323         * kwq/KWQFont.mm:
 324         (QFont::isFixedPitch):
 325
 326 2005-02-17  Darin Adler  <darin@apple.com>
 327
 328         Reviewed by Kevin.
 329
 330         - fixed <rdar://problem/4011210> REGRESSION (180-181): Maxlength property for INPUT object not working
 331
 332         * kwq/KWQTextField.mm:
 333         (-[KWQTextField textView:shouldChangeTextInRange:replacementString:]): If controller returns
 334         YES, go on to call super, since super does the text formatter handling, which we need.
 335         (-[KWQSecureTextField textView:shouldChangeTextInRange:replacementString:]): Ditto.
 336         (-[KWQSearchField textView:shouldChangeTextInRange:replacementString:]): Ditto.
 337
 338 === Safari-187 ===
 339
 340 2005-02-17  Ken Kocienda  <kocienda@apple.com>
 341
 342         Reviewed by Hyatt
 343
 344         Fix for this bug:
 345
 346         <rdar://problem/4012058> Copy from quoted text and paste results in blue text
 347
 348         The relatively-new paste code tries hard to retain style of the content from the
 349         source location. However, in the case of quoted material in mail messages, we do
 350         not want to carry the quoting color along. This fixes the problem by factoring
 351         out such colors.
 352
 353         * khtml/editing/htmlediting.cpp:
 354         (khtml::ReplacementFragment::computeStylesAndRemoveUnrendered): Now calls removeBlockquoteColorsIfNeeded()
 355         to fix bug.
 356         (khtml::ReplacementFragment::removeBlockquoteColorsIfNeeded): New function. Factors out colors that are
 357         the result of text being quoted.
 358         (khtml::isNodeRendered): Function moved in file. No other change.
 359         (khtml::isProbablyBlock): Function moved in file. No other change.
 360         (khtml::closestMailBlockquote): New function. Helps fix bug.
 361         (khtml::isMailBlockquote): Function moved to be free-floating instead of being a
 362         member of CompositeEditCommand. No other change.
 363         * khtml/editing/htmlediting.h: Moved around some functions. Added removeBlockquoteColorsIfNeeded().
 364         * khtml/editing/markup.cpp: Remove redundant static implementation of isMailBlockquote.
 365
 366 2005-02-17  Richard Williamson   <rjw@apple.com>
 367
 368         Fixed <rdar://problem/4008163> dynamic support for -apple-dashboard-region is flakey
 369
 370         Style operator== wasn't including regions.
 371
 372         Reviewed by Darin.
 373
 374         * khtml/rendering/render_style.cpp:
 375         (StyleCSS3NonInheritedData::operator==):
 376
 377 2005-02-17  Adele Amchan  <adele@apple.com>
 378
 379         Reviewed by Chris.
 380
 381         fix for <rdar://problem/4010028> 8A383: Safari v185 crash loading united.com multi city fare finder page.
 382
 383         * khtml/dom/dom_string.cpp: (DOM::DOMString::operator += ): prevent nil dereference when DOMString being added is nil
 384
 385 2005-02-17  Jens Alfke  <jens@apple.com>
 386
 387         Reviewed by Hyatt.
 388
 389         Fix for rdar://3963151 "Mail only pasted ~950 of 1407 text lines into my message!"
 390         Force tokenizer to run synchronously while parsing document fragments, so it doesn't stop halfway through and cause truncated content.
 391
 392         * khtml/html/html_elementimpl.cpp:
 393         (HTMLElementImpl::createContextualFragment):
 394         * khtml/html/htmltokenizer.cpp:
 395         (khtml::HTMLTokenizer::reset):
 396         (khtml::HTMLTokenizer::begin):
 397         (khtml::HTMLTokenizer::setForceSynchronous):
 398         (khtml::HTMLTokenizer::continueProcessing):
 399         * khtml/html/htmltokenizer.h:
 400
 401 2005-02-17  Ken Kocienda  <kocienda@apple.com>
 402
 403         Reviewed by Hyatt
 404
 405         Fix for this bug:
 406
 407         <rdar://problem/3982183> Mail's HTML paragraphs appear with unintended margins in Entourage and Eudora
 408
 409         * khtml/editing/htmlediting.cpp:
 410         (khtml::createDefaultParagraphElement): Change element we make from <p> to <div>.
 411         This fixes the problem, as these other mailers show <div> elements without margins.
 412
 413         Some test files needed to change to preserve the ability to "eyeball" the results.
 414         Many others changed just because their <p> elements changed to <div> elements.
 415
 416         * layout-tests/editing/deleting/delete-3928305-fix-expected.txt
 417         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002-expected.txt
 418         * layout-tests/editing/inserting/insert-div-013-expected.txt
 419         * layout-tests/editing/inserting/insert-div-013.html
 420         * layout-tests/editing/inserting/insert-div-014-expected.txt
 421         * layout-tests/editing/inserting/insert-div-014.html
 422         * layout-tests/editing/inserting/insert-div-018-expected.txt
 423         * layout-tests/editing/inserting/insert-div-018.html
 424         * layout-tests/editing/inserting/insert-div-019-expected.txt
 425         * layout-tests/editing/inserting/insert-div-019.html
 426         * layout-tests/editing/inserting/insert-div-020-expected.txt
 427         * layout-tests/editing/inserting/insert-div-020.html
 428         * layout-tests/editing/inserting/insert-div-021-expected.txt
 429         * layout-tests/editing/inserting/insert-div-021.html
 430         * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt
 431         * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt
 432         * layout-tests/editing/style/remove-underline-after-paragraph-expected.txt
 433         * layout-tests/editing/style/remove-underline-after-paragraph-in-bold-expected.txt
 434
 435 2005-02-17  Ken Kocienda  <kocienda@apple.com>
 436
 437         Reviewed by Hyatt
 438
 439         Fix for this bug:
 440
 441         <rdar://problem/3996737> REGRESSION (Mail): Copy/paste in Mail inserts returns
 442
 443         The root cause of the problem is that a couple of pieces of code in AppKit and Mail
 444         insert newlines into markup as it is generated, and the paste code in WebCore was
 445         not smart about nodes that do not render (as is the case with these added newlines).
 446         So, the solution is to remove these unrendered nodes in the paste code in a pre-pass.
 447         Fortunately, my recent addition of such a pass to the paste code to handle styles
 448         gave me a convenient place to put this new logic.
 449
 450         * khtml/editing/htmlediting.cpp:
 451         (khtml::ReplacementFragment::ReplacementFragment): Use new name for function below.
 452         (khtml::ReplacementFragment::computeStylesAndRemoveUnrendered): Name change from computeStylesForNodes().
 453         Now does the additional work of removing unrendered nodes, as mentioned above.
 454         (khtml::isNodeRendered): New helper function.
 455         * khtml/editing/htmlediting.h: Updated as necessary.
 456
 457 2005-02-17  Ken Kocienda  <kocienda@apple.com>
 458
 459         Reviewed by Hyatt
 460
 461         Fix for this bug:
 462
 463         <rdar://problem/3998892> REGRESSION (Mail): bolding a selection from end of line changes unselected text on starting line.
 464
 465         The styling code did not move to the next node when the starting position was
 466         at the last offset of a node. Instead, it styled it. Clearly wrong. Solution
 467         is to borrow a check from the delete algorithm. Though I call caretMaxOffset(),
 468         which I consider deprecated, it is still the simplest and most strightforward
 469         way to ask the right question for this kind of problem.
 470
 471         * khtml/editing/htmlediting.cpp:
 472         (khtml::ApplyStyleCommand::applyInlineStyle): Fixed, as described above.
 473         * layout-tests/editing/style/style-3998892-fix-expected.txt: Added.
 474         * layout-tests/editing/style/style-3998892-fix.html: Added.
 475
 476 2005-02-16  Chris Blumenberg  <cblu@apple.com>
 477
 478         Fixed: <rdar://problem/3954842> Forward/reply to an HTML email can result in nothing (in cases with <link> tags for CSS)
 479
 480         Reviewed by john.
 481
 482         * khtml/xml/dom_nodeimpl.cpp:
 483         (NodeBaseImpl::addChild): don't call insertedIntoDocument on the added child if "this" itself is not in the document.
 484
 485 2005-02-16  John Sullivan  <sullivan@apple.com>
 486
 487         Written by Darin, reviewed by Maciej and me
 488
 489         - change required to make previous checkin work with English language RSS pages
 490
 491         * kwq/KWQKURL.h:
 492         new private method isHierarchical
 493         * kwq/KWQKURL.mm:
 494         (KURL::KURL):
 495         add hierarchical base URL check when determining whether the URL is absolute
 496         (KURL::isHierarchical):
 497         new method, returns true if this is a valid URL with a slash just past the scheme's trailing colon
 498
 499 2005-02-16  John Sullivan  <sullivan@apple.com>
 500
 501         Written by Darin, reviewed by me.
 502
 503         - WebCore part of fix for <rdar://problem/4007384>
 504         FILTER: Bookmark of RSS with Japanese search word & multiple RSS pages loses filter
 505
 506         * kwq/DOM.mm:
 507         (-[DOMElement _getURLAttribute:]):
 508         new SPI method, uses parseURL and completeURL to get valid URL from attribute value
 509
 510         * kwq/DOMPrivate.h:
 511         declare new SPI method
 512
 513 2005-02-15  Maciej Stachowiak  <mjs@apple.com>
 514
 515         Reviewed by Darin.
 516
 517         <rdar://problem/3942428> reproducible crash loading cbs.sportsline.com
 518
 519         This change reverts the fix for <rdar://problem/3805311>, and
 520         re-fixes it in a different (better) way. Instead of preventing
 521         programmatic open from setting the parsing flag, instead make sure
 522         that programmatic close resets it.
 523
 524         * khtml/khtml_part.cpp:
 525         (KHTMLPart::openURL):
 526         (KHTMLPart::didExplicitOpen):
 527         (KHTMLPart::closeURL):
 528         (KHTMLPart::begin):
 529         (KHTMLPart::end):
 530         (KHTMLPart::endIfNotLoading):
 531         (KHTMLPart::slotFinishedParsing):
 532         (KHTMLPart::checkEmitLoadEvent):
 533         * khtml/khtml_part.h:
 534         * khtml/khtmlpart_p.h:
 535         (KHTMLPartPrivate::KHTMLPartPrivate):
 536         * khtml/xml/dom_docimpl.cpp:
 537         (DocumentImpl::open):
 538         (DocumentImpl::implicitOpen):
 539         (DocumentImpl::close):
 540         (DocumentImpl::implicitClose):
 541         * khtml/xml/dom_docimpl.h:
 542
 543 2005-02-15  David Harrison  <harrison@apple.com>
 544
 545         Reviewed by Maciej.
 546
 547         <rdar://problem/3933665> smart delete seems to delete too much after expanding selection with arrow keys
 548
 549         Fixed by updating the selection granularity.  Only byWord granularity enables smart delete, but in this
 550         case the granularity changed from byWord to byCharacter.
 551
 552         * khtml/khtml_part.cpp:
 553         (KHTMLPart::setSelectionGranularity):
 554         * khtml/khtml_part.h:
 555         * kwq/WebCoreBridge.mm:
 556         (-[WebCoreBridge rangeByExpandingSelectionWithGranularity:]):
 557         (-[WebCoreBridge rangeByAlteringCurrentSelection:direction:granularity:]):
 558         (-[WebCoreBridge alterCurrentSelection:direction:granularity:]):
 559
 560 2005-02-15  David Harrison  <harrison@apple.com>
 561
 562         Reviewed by Darin.
 563
 564         (continued) <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line
 565
 566         This was supposed have been committed yesterday with the other part of the fix, but was not.
 567
 568         * khtml/editing/visible_units.cpp:
 569         (khtml::endOfWord):
 570
 571 2005-02-15  Ken Kocienda  <kocienda@apple.com>
 572
 573         Reviewed by John
 574
 575         Fix for this bug:
 576
 577         <rdar://problem/3951178> REGRESSION (Mail): blank line lost after pasting as quotation
 578
 579         Problem was that the blank line after the selection was getting deleted incorrectly since the
 580         paste code thought this was an unneeded placeholder rather than a placeholder outside of the
 581         selection.
 582
 583         * khtml/editing/htmlediting.cpp:
 584         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Now calls findBlockPlaceholder.
 585         (khtml::CompositeEditCommand::findBlockPlaceholder): Moved finding code formerly in
 586         removeBlockPlaceholderIfNeeded to this new helper.
 587         (khtml::ReplaceSelectionCommand::doApply): Do not delete placeholder up front. Call
 588         findBlockPlaceholder, and delete it later if needed in the already-existing cleanup step.
 589         * khtml/editing/htmlediting.h: Add new function.
 590
 591         New layout test.
 592
 593         * layout-tests/editing/pasteboard/paste-text-017-expected.txt: Added.
 594         * layout-tests/editing/pasteboard/paste-text-017.html: Added.
 595
 596         Result changed for the better.
 597
 598         * layout-tests/editing/pasteboard/paste-text-011-expected.txt
 599
 600 2005-02-14  David Harrison  <harrison@apple.com>
 601
 602         Reviewed by Darin.
 603
 604         <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line
 605
 606         Reviewed by Darin.
 607
 608         <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line
 609
 610         * khtml/editing/visible_text.cpp:
 611         (khtml::SimplifiedBackwardsTextIterator::advance):
 612         Add BR in for <rdar://problem/3917929> fix only if leaving a visible text node.
 613
 614         * khtml/editing/visible_units.cpp:
 615         (khtml::endOfWord):
 616         Do not move left over a paragraph boundary.
 617
 618 2005-02-14  Darin Adler  <darin@apple.com>
 619
 620         Reviewed by John.
 621
 622         - fixed <rdar://problem/3686434> Safari uses too much RAM on file upload, leading to malloc errors and crashes (HP printers)
 623
 624         * khtml/html/html_formimpl.cpp:
 625         (DOM::FormDataList::begin): Updated to use a list of FormDataListItem instead of QCString.
 626         (DOM::FormDataList::end): Ditto.
 627         (DOM::HTMLFormElementImpl::formData): Same here, and also use appendFile rather than appendData
 628         when we encounter a path name rather than data.
 629         (DOM::HTMLInputElementImpl::appendFormData): Use appendFile here rather than reading the file
 630         in; the reading now happens inside WebKit.
 631         (DOM::FormDataList::appendString): Updated for FormDataListItem.
 632         (DOM::FormDataList::appendFile): Added.
 633
 634         * ForwardingHeaders/kfileitem.h: Emptied out the file; no longer includes KWQKFileItem.h.
 635         * ForwardingHeaders/netaccess.h: Emptied out the file; no longer includes KWQKIONetAccess.h.
 636         * kwq/KWQKFileItem.h: Removed.
 637         * kwq/KWQKFileItem.mm: Removed.
 638         * kwq/KWQKIONetAccess.h: Removed.
 639         * kwq/KWQKIONetAccess.mm: Removed.
 640         * WebCore.pbproj/project.pbxproj: Removed the 4 files above.
 641
 642         - small unrelated fix (not reviewed by John)
 643
 644         * khtml/ecma/kjs_window.cpp: (Window::isSafeScript): Put some logging code outside if statements so it works all the time.
 645
 646 2005-02-14  David Harrison  <harrison@apple.com>
 647
 648         Reviewed by Darin.
 649
 650         <rdar://problem/4004305> REGRESSION (Mail): Command-right-arrow on wrapped text goes to end of previous line
 651
 652         Fixed nextLinePosition to calculate affinity rather than take it as a parameter.  Propagated the parameter change out to related methods.
 653
 654         * khtml/editing/htmlediting.cpp:
 655         (khtml::DeleteSelectionCommand::initializePositionData):
 656         (khtml::InsertLineBreakCommand::doApply):
 657         (khtml::InsertParagraphSeparatorCommand::doApply):
 658         (khtml::InsertTextCommand::input):
 659         (khtml::ReplaceSelectionCommand::doApply):
 660         (khtml::ReplaceSelectionCommand::applyStyleToInsertedNodes):
 661         * khtml/editing/selection.cpp:
 662         (khtml::Selection::modifyExtendingRightForward):
 663         (khtml::Selection::modifyMovingRightForward):
 664         (khtml::Selection::modifyExtendingLeftBackward):
 665         (khtml::Selection::modifyMovingLeftBackward):
 666         (khtml::Selection::modify):
 667         (khtml::Selection::validate):
 668         * khtml/editing/visible_position.cpp:
 669         (khtml::visiblePositionsOnDifferentLines):
 670         * khtml/editing/visible_units.cpp:
 671         (khtml::rootBoxForLine):
 672         (khtml::startOfLine):
 673         (khtml::endOfLine):
 674         (khtml::inSameLine):
 675         (khtml::isStartOfLine):
 676         (khtml::isEndOfLine):
 677         (khtml::previousLinePosition):
 678         (khtml::nextLinePosition):
 679         (khtml::previousSentencePosition):
 680         (khtml::nextSentencePosition):
 681         (khtml::previousParagraphPosition):
 682         (khtml::nextParagraphPosition):
 683         * khtml/editing/visible_units.h:
 684         * khtml/khtml_events.cpp:
 685         (khtml::MouseEvent::offset):
 686         * khtml/khtml_part.cpp:
 687         (KHTMLPart::isPointInsideSelection):
 688         (KHTMLPart::selectClosestWordFromMouseEvent):
 689         (KHTMLPart::handleMousePressEventTripleClick):
 690         (KHTMLPart::handleMousePressEventSingleClick):
 691         (KHTMLPart::handleMouseMoveEventSelection):
 692         (KHTMLPart::khtmlMouseReleaseEvent):
 693         * khtml/rendering/render_block.cpp:
 694         (khtml::RenderBlock::positionForCoordinates):
 695         * khtml/rendering/render_block.h:
 696         * khtml/rendering/render_br.cpp:
 697         (RenderBR::positionForCoordinates):
 698         * khtml/rendering/render_br.h:
 699         * khtml/rendering/render_container.cpp:
 700         (RenderContainer::positionForCoordinates):
 701         * khtml/rendering/render_container.h:
 702         * khtml/rendering/render_inline.cpp:
 703         (RenderInline::positionForCoordinates):
 704         * khtml/rendering/render_inline.h:
 705         * khtml/rendering/render_object.cpp:
 706         (RenderObject::caretRect):
 707         (RenderObject::positionForCoordinates):
 708         * khtml/rendering/render_object.h:
 709         * khtml/rendering/render_replaced.cpp:
 710         (RenderReplaced::positionForCoordinates):
 711         * khtml/rendering/render_replaced.h:
 712         * khtml/rendering/render_text.cpp:
 713         (RenderText::positionForCoordinates):
 714         * khtml/rendering/render_text.h:
 715         * khtml/xml/dom_position.cpp:
 716         (DOM::Position::previousCharacterPosition):
 717         (DOM::Position::nextCharacterPosition):
 718         (DOM::Position::leadingWhitespacePosition):
 719         (DOM::Position::trailingWhitespacePosition):
 720         * khtml/xml/dom_position.h:
 721         * kwq/KWQAccObject.mm:
 722         (-[KWQAccObject value]):
 723         (-[KWQAccObject accessibilityAttributeValue:]):
 724         (-[KWQAccObject doAXLineForTextMarker:]):
 725         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
 726         (-[KWQAccObject doAXTextMarkerForPosition:]):
 727         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
 728         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
 729         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
 730         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
 731         * kwq/KWQKHTMLPart.mm:
 732         * kwq/WebCoreBridge.mm:
 733         (-[WebCoreBridge _visiblePositionForPoint:]):
 734
 735 05-02-07  Maciej Stachowiak  <mjs@apple.com>
 736
 737         Reviewed by Ken and John.
 738
 739         Re-fixed a specific case of the following:
 740
 741         <rdar://problem/3790449> REGRESSION (Mail): underline behavior is flaky because of how CSS handles text-decoration
 742
 743         Oddly, Cmd-B, Cmd-U, type some text, return, Cmd-U, Cmd-B, type
 744         some text, worked fine. But hitting the second Cmd-B before the
 745         second Cmd-U still failed to remove underlining. The reason for
 746         this is that our code to compute the style of the current position
 747         did not work when you had a typing style and were positioned right
 748         at a <br> element. For various reasons, this did not show up for
 749         bold and italic, since those are handled through the font manager.
 750
 751         The following change fixes this - for elements that can't have
 752         children, we add the dummy span after the element of interest,
 753         rather than as a child of it.
 754
 755         * khtml/khtml_part.cpp:
 756         (KHTMLPart::selectionComputedStyle):
 757
 758 2005-02-11  David Harrison  <harrison@apple.com>
 759
 760         Reviewed by Darin.
 761
 762         <rdar://problem/3978980> Double Clicking on a line in Mail selected the entire body
 763
 764         * khtml/editing/visible_units.cpp:
 765         (khtml::startOfWord):
 766         (khtml::endOfWord):
 767         Pay attention to being at the end of a paragraph.
 768
 769         (khtml::previousLinePosition):
 770         (khtml::nextLinePosition):
 771         (khtml::endOfParagraph):
 772         Use DOWNSTREAM per recent affinity changes.
 773
 774 2005-02-11  Richard Williamson   <rjw@apple.com>
 775
 776         Fixed <rdar://problem/3985118> DOM objects not being marshaled on JS->native calls
 777
 778         Re-factored how 'native' wrappers for JS objects are created.  The interpreter now
 779         creates these wrappers.  The WebCore subclass of the interpreter now overrides
 780         createLanguageInstanceForValue() and creates a DOM ObjC wrapper for DOM objects.
 781
 782         * WebCore.pbproj/project.pbxproj:
 783         * khtml/ecma/kjs_binding.cpp:
 784         (ScriptInterpreter::createLanguageInstanceForValue):
 785         * khtml/ecma/kjs_binding.h:
 786         * kwq/DOMUtility.mm: Added.
 787         (KJS::ScriptInterpreter::createObjcInstanceForValue):
 788         * kwq/KWQKHTMLPart.mm:
 789         (KWQKHTMLPart::getAppletInstanceForView):
 790         (getInstanceForView):
 791         (KWQKHTMLPart::getEmbedInstanceForView):
 792         (KWQKHTMLPart::getObjectInstanceForView):
 793
 794 2005-02-11  Chris Blumenberg  <cblu@apple.com>
 795
 796         Fixed: <rdar://problem/3937352> Quote level not maintained when copied and pasted within a Mail message
 797
 798         Reviewed by harrison.
 799
 800         * khtml/editing/markup.cpp:
 801         (khtml::isMailBlockquote): new
 802         (khtml::createMarkup): retain the Mail quote level by including all ancestor mail block quotes
 803         * khtml/rendering/render_block.cpp:
 804         (khtml::RenderBlock::positionForCoordinates): default to downstream for the affinity since the affinity is only upstream when the cursor is clicked to the right of a wrapped line
 805         * khtml/rendering/render_text.cpp:
 806         (RenderText::positionForCoordinates): ditto
 807
 808 2005-02-11  Adele Amchan  <adele@apple.com>
 809
 810         Reviewed by Darin.
 811
 812         fix for <rdar://problem/4004004> no need to add body element for xml documents
 813
 814         In the fix for <rdar://problem/3758785> we decided to add a body element when closing a document
 815         to ensure that the onload handler would fire.  This is unnecessary for xml documents,
 816         so now we also check to see if we're dealing with an html document before adding the body element.
 817
 818         * khtml/xml/dom_docimpl.cpp: (DocumentImpl::close):
 819
 820 2005-02-11  Adele Amchan  <adele@apple.com>
 821
 822         Reviewed by Darin.
 823
 824         * layout-tests/apple-only/base/www.excite.com/index-expected.txt: updates tests because of custom tag change.
 825         * layout-tests/apple-only/base/www.msnbc.com/index-expected.txt:
 826         * layout-tests/apple-only/base/www.time.com/index-expected.txt:
 827         * layout-tests/apple-only/base/www.travelocity.com/index-expected.txt:
 828         * layout-tests/apple-only/base/www.zdnet.com/index-expected.txt:
 829         * layout-tests/fast/overflow/003-expected.txt:
 830
 831         * layout-tests/fast/js/window-object-cross-frame-calls-expected.txt: updated test because of added Body (see rdar://problem/3758785)
 832
 833 === Safari-186 ===
 834
 835 2005-02-11  Darin Adler  <darin@apple.com>
 836
 837         Reviewed by John.
 838
 839         - fixed <rdar://problem/3915449> paths are relative to the old src URL after document.open, which is supposed to clear the document, including the URL
 840
 841         * khtml/xml/dom_docimpl.h: Added a new openInternal function for use by KHTMLPart.
 842         * khtml/xml/dom_docimpl.cpp:
 843         (DocumentImpl::open): Changed to do everything we did before, but also clear the URL and set the
 844         base URL based on the enclosing document. This is the basic JavaScript/DOM operation of opening a
 845         document, which is supposed to clear the document, including the URL. In the long run we might want
 846         to do even more document "resetting and clearing" in here, but this URL clearing is what's needed
 847         now to fix the most important problem.
 848         (DocumentImpl::openInternal): Moved the old open code in here, except for the "parsing" check, which
 849         is unnecessary and inappropriate in the one place we call this.
 850
 851         * khtml/khtml_part.cpp: (KHTMLPart::begin): Call openInternal instead of open.
 852
 853 2005-02-10  Ken Kocienda  <kocienda@apple.com>
 854
 855         Reviewed by Hyatt
 856
 857         Fix for this bug:
 858
 859         <rdar://problem/3992092> 8A374: Mail crash while pasting - RemoveNodeCommand
 860
 861         * khtml/editing/htmlediting.cpp:
 862         (khtml::ReplaceSelectionCommand::doApply): Code tried to remove a node that was no longer in
 863         the document. We try to do a good job of detecting all these cases, and generally do. This
 864         one was missed. Fixed.
 865
 866 2005-02-10  Darin Adler  <darin@apple.com>
 867
 868         Reviewed by Harrison.
 869
 870         - added support needed to fix <rdar://problem/3991225> Format->Style->Underline menu item does not get checked when selected text is underlined
 871
 872         * kwq/WebCoreBridge.h: Added selectionHasStyle: method.
 873         * kwq/WebCoreBridge.mm: (-[WebCoreBridge selectionHasStyle:]): Added. Calls selectionHasStyle on the part.
 874
 875 2005-02-10  Darin Adler  <darin@apple.com>
 876
 877         Reviewed by Harrison.
 878
 879         - fixed <rdar://problem/3990484> cursor changes to I-beam when moving over text in widgets, even when selection off via CSS
 880
 881         * khtml/khtmlview.cpp: (KHTMLView::viewportMouseMoveEvent): Changed code to check canSelect when deciding
 882         whether to show an I-beam cursor.
 883
 884         * khtml/rendering/render_object.h: Added canSelect.
 885         * khtml/rendering/render_object.cpp:
 886         (selectStartNode): Added. Helper function with the guts of shouldSelect.
 887         (RenderObject::canSelect): Added. Like shouldSelect, but does not call the "start selecting" event handler.
 888         (RenderObject::shouldSelect): Refactored to use selectStartNode.
 889
 890         * khtml/css/cssvalues.c: Regnerated with newer gperf.
 891
 892 2005-02-10  David Hyatt  <hyatt@apple.com>
 893
 894         Fix for 3867759, .mac regression where scrollers don't show up.  Make sure you can set the .width and
 895         .height properties on the Image object.
 896
 897         Reviewed by John Sullivan
 898
 899         * khtml/ecma/kjs_html.cpp:
 900         (ImageConstructorImp::construct):
 901         (Image::getValueProperty):
 902         (Image::putValue):
 903         (Image::Image):
 904         * khtml/ecma/kjs_html.h:
 905
 906 2005-02-10  Ken Kocienda  <kocienda@apple.com>
 907
 908         Reviewed by Chris
 909
 910         Fix for this bug:
 911
 912         <rdar://problem/3965158> Drag-n-drop within a rich text message sometimes changes the color of the dragged text
 913
 914         This change fixes the bug....and much more. Now, for the first time, the paste code can do "smart merging"
 915         or "smooshing" of styles during its operation. Since this new code is actively, rather than passively
 916         working with styles, it fixes the bug, and lays the groundwork for similar work we need to do to
 917         preserve quote levels in Mail.
 918
 919         * khtml/css/css_valueimpl.cpp:
 920         (DOM::CSSMutableStyleDeclarationImpl::clear): New method.
 921         (DOM::CSSMutableStyleDeclarationImpl::removeBlockProperties): Ditto.
 922         (DOM::CSSMutableStyleDeclarationImpl::removePropertiesInSet): Ditto.
 923         (DOM::operator==): Add operator for CSSProperty.
 924         * khtml/css/css_valueimpl.h: Declare new functions.
 925         * khtml/editing/htmlediting.cpp:
 926         (khtml::isEmptyStyleSpan): Improved the test in this function, rolling together the old implementation
 927         with some code that did this work inline elsewhere. Sum of the parts is better than either test was by itself.
 928         (khtml::isStyleSpan): Check for ID_SPAN.
 929         (khtml::ApplyStyleCommand::removeCSSStyle): Call isEmptyStyleSpan. This was the place with an inline implementation before.
 930         (khtml::ReplacementFragment::ReplacementFragment): Now takes a DocumentImpl argument. No longer does a "default style"
 931         check, but rather calls functions which do a similar check to that, and much more.
 932         (khtml::ReplacementFragment::~ReplacementFragment): Deref document, and computed styles.
 933         (khtml::ReplacementFragment::styleForNode): New helper. Looks up and returns computed style for a node.
 934         (khtml::ReplacementFragment::removeNodePreservingChildren): New helper.
 935         (khtml::ReplacementFragment::computeStylesForNodes): New function which computes the "desired" style for
 936         every node in the fragment. This information is used later after paste is done as a reference for testing
 937         what styles need to be added, and which can be removed as redundant, from all the nodes inserted by the
 938         replacement code.
 939         (khtml::ReplacementFragment::removeStyleNodes): Clears out all style nodes from the fragment. They are
 940         no longer needed after the call to computeStylesForNodes(),
 941         (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): Add a document to the call to initialize the
 942         command's ReplacementFragment.
 943         (khtml::ReplaceSelectionCommand::doApply): Call applyStyleToInsertedNodes() after inserting nodes to make
 944         styles come out right.
 945         (khtml::ReplaceSelectionCommand::applyStyleToInsertedNodes): This is the "style smooshing" function. It
 946         computes the styles that need to be added to each node inserted, comparing the style it gets from just
 947         being inserted into its correct destination with the computed "desired style" done in the
 948         ReplacementFragment constructor. It then adds in all the necessary styles, and will also remove redundant styles.
 949         * khtml/editing/htmlediting.h: Update declarations and member variables as needed.
 950         * khtml/editing/markup.cpp:
 951         (khtml::startMarkup): Add additional style annotations to the markup we generate, so that paste code can preserve it.
 952         (khtml::markup): Ditto.
 953         (khtml::createMarkup): Ditto.
 954
 955         These test results are subtly better with this change. They no longer have an unneeded empty span.
 956         Visually the same as before.
 957
 958         * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt
 959         * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt
 960         * layout-tests/editing/style/remove-underline-expected.txt
 961         * layout-tests/editing/style/remove-underline-from-stylesheet-expected.txt
 962
 963         New tests:
 964
 965         * layout-tests/editing/style/smoosh-styles-001-expected.txt
 966         * layout-tests/editing/style/smoosh-styles-002-expected.txt
 967         * layout-tests/editing/style/smoosh-styles-001.html
 968         * layout-tests/editing/style/smoosh-styles-002.html
 969
 970 2005-02-10  Darin Adler  <darin@apple.com>
 971
 972         Reviewed by Adele.
 973
 974         - fixed <rdar://problem/3974988> WebHTMLView drops scroll wheel events when deltas are 0
 975
 976         * kwq/WebCoreScrollView.m: (-[WebCoreScrollView autoforwardsScrollWheelEvents]): Added, for Tiger only.
 977         For Tiger, removed the override of scrollWheel:, which doesn't work quite right because of how the
 978         delta attributes return all 0 for newfangled scrolling events from the new trackpads.
 979
 980 2005-02-10  David Harrison  <harrison@apple.com>
 981
 982         Reviewed by Darin.
 983
 984         <rdar://problem/3991532> REGRESSION (Mail): Triple-click on trailing blank line should select previous line but doesn't
 985
 986         * khtml/css/cssvalues.c:
 987         (hash_val):
 988         (findValue):
 989         * khtml/editing/selection.cpp:
 990         (khtml::Selection::validate):
 991         Do moral equiavalent of LeftWordIfOnBoundary for PARAGRAPH.
 992
 993 2005-02-10  David Harrison  <harrison@apple.com>
 994
 995         Reviewed by Darin.
 996
 997         <rdar://problem/3991848> Double-click on first character selects wrong item
 998
 999         * khtml/editing/selection.cpp:
1000         (khtml::Selection::validate):
1001         Honor the fact that clicking on a character positions the cursor on the left side of the character.
1002
1003 2005-02-10  David Hyatt  <hyatt@apple.com>
1004
1005         Fix for 3980778, repro crash in RootInlineBox::paint.  Make sure that when merging blocks we dirty them for layout
1006         up front. That way as we do the merge, we dont attempt to examine the lines in dirtyLinesForChangedChild.
1007
1008         Reviewed by mjs
1009
1010         * khtml/rendering/render_block.cpp:
1011         (khtml::RenderBlock::removeChild):
1012
1013 2005-02-10  David Hyatt  <hyatt@apple.com>
1014
1015         Fix for 3987010, fix a bug that caused the self-collapsing-block-cleared-float bit to never get set correctly.
1016
1017         Reviewed by mjs
1018
1019         * khtml/rendering/render_block.h:
1020         (khtml::RenderBlock::MarginInfo::setSelfCollapsingBlockClearedFloat):
1021
1022 2005-02-10  David Hyatt  <hyatt@apple.com>
1023
1024         Fix  for 3992440, text not wrapping correctly on half moon bay web site.  Refine an incorrect optimization I
1025         added to be less restrictive.
1026
1027         Reviewed by mjs
1028
1029         * khtml/rendering/render_block.cpp:
1030         (khtml::RenderBlock::layoutBlockChildren):
1031
1032 2005-02-08  Maciej Stachowiak  <mjs@apple.com>
1033
1034         Reviewed by Darin.
1035
1036         <rdar://problem/3977973> pages on ebay leak referenced JavaScript objects -- over time browsing becomes super-slow
1037
1038         I fixed this by removing all event listeners for a document, it's
1039         children, and any disconnected nodes that used to be in the
1040         document at document detach time. Mozilla temporarily disables
1041         event listeners on such nodes, but re-enables them if you
1042         re-parant a node into a new document. However, in WebCore, you
1043         can't re-parent a node into another document, so there is no
1044         observable change in behavior.
1045
1046         We have to do this to break the possible reference cycles between
1047         event listeners and the dom nodes they are attached to (e.g. via
1048         scope chain, as in this case).
1049
1050         * khtml/xml/dom_docimpl.cpp:
1051         (DocumentImpl::detach):
1052         (DocumentImpl::removeAllEventListenersFromAllNodesx):
1053         (DocumentImpl::registerDisconnectedNodeWithEventListeners):
1054         (DocumentImpl::unregisterDisconnectedNodeWithEventListeners):
1055         (DocumentImpl::removeAllDisconnectedNodeEventListeners):
1056         * khtml/xml/dom_docimpl.h:
1057         * khtml/xml/dom_nodeimpl.cpp:
1058         (NodeImpl::~NodeImpl):
1059         (NodeImpl::addEventListener):
1060         (NodeImpl::removeEventListener):
1061         (NodeImpl::removeAllEventListeners):
1062         (NodeImpl::removeHTMLEventListener):
1063         (NodeImpl::insertedIntoDocument):
1064         (NodeImpl::removedFromDocument):
1065         * khtml/xml/dom_nodeimpl.h:
1066
1067 2005-02-09  Chris Blumenberg  <cblu@apple.com>
1068
1069         Fixed: <rdar://problem/3999213> Sometimes 2 Windows Media Player plugin instances are loaded
1070
1071         Reviewed by hyatt.
1072
1073         * khtml/html/html_objectimpl.cpp:
1074         (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.
1075         (HTMLObjectElementImpl::recalcStyle): ditto
1076
1077 2005-02-09  David Harrison  <harrison@apple.com>
1078
1079         Reviewed by Darin.
1080
1081         <rdar://problem/3937447> Mail-689: Arrow key navigation in new message body stops working when a line ends with a bold word
1082
1083         Added affinity to VisiblePosition.  Changed Selection code to use affinity more.
1084
1085         (Partial) <rdar://problem/3982096> editing/pasteboard/paste-text-007 is failing
1086
1087         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.
1088
1089         * khtml/ecma/kjs_window.cpp:
1090         (SelectionFunc::tryCall):
1091         * khtml/editing/htmlediting.cpp:
1092         (khtml::EditCommandPtr::setStartingSelection):
1093         (khtml::EditCommandPtr::setEndingSelection):
1094         (khtml::EditCommand::setStartingSelection):
1095         (khtml::EditCommand::setEndingSelection):
1096         (khtml::CompositeEditCommand::deleteInsignificantTextDownstream):
1097         (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary):
1098         (khtml::ApplyStyleCommand::splitTextAtStartIfNeeded):
1099         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded):
1100         (khtml::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
1101         (khtml::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
1102         (khtml::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
1103         (khtml::ApplyStyleCommand::mergeEndWithNextIfIdentical):
1104         (khtml::ApplyStyleCommand::joinChildTextNodes):
1105         (khtml::DeleteSelectionCommand::initializePositionData):
1106         (khtml::DeleteSelectionCommand::handleGeneralDelete):
1107         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete):
1108         (khtml::DeleteSelectionCommand::doApply):
1109         (khtml::InsertLineBreakCommand::doApply):
1110         (khtml::InsertParagraphSeparatorCommand::doApply):
1111         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
1112         (khtml::InsertTextCommand::input):
1113         (khtml::MoveSelectionCommand::doApply):
1114         (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand):
1115         (khtml::ReplaceSelectionCommand::~ReplaceSelectionCommand):
1116         (khtml::ReplaceSelectionCommand::doApply):
1117         (khtml::ReplaceSelectionCommand::completeHTMLReplacement):
1118         (khtml::ReplaceSelectionCommand::updateNodesInserted):
1119         (khtml::TypingCommand::deleteKeyPressed):
1120         (khtml::TypingCommand::forwardDeleteKeyPressed):
1121         (khtml::TypingCommand::markMisspellingsAfterTyping):
1122         * khtml/editing/htmlediting.h:
1123         * khtml/editing/selection.cpp:
1124         (khtml::Selection::Selection):
1125         (khtml::Selection::init):
1126         (khtml::Selection::moveTo):
1127         (khtml::Selection::modifyExtendingRightForward):
1128         (khtml::Selection::modifyMovingRightForward):
1129         (khtml::Selection::modifyExtendingLeftBackward):
1130         (khtml::Selection::modifyMovingLeftBackward):
1131         (khtml::Selection::modify):
1132         (khtml::Selection::xPosForVerticalArrowNavigation):
1133         (khtml::Selection::clear):
1134         (khtml::Selection::setBase):
1135         (khtml::Selection::setExtent):
1136         (khtml::Selection::setBaseAndExtent):
1137         (khtml::Selection::layout):
1138         (khtml::Selection::validate):
1139         * khtml/editing/selection.h:
1140         (khtml::Selection::startAffinity):
1141         (khtml::Selection::endAffinity):
1142         (khtml::Selection::baseAffinity):
1143         (khtml::Selection::extentAffinity):
1144         (khtml::operator==):
1145         * khtml/editing/text_affinity.h:
1146         (khtml::):
1147         * khtml/editing/visible_position.cpp:
1148         (khtml::VisiblePosition::VisiblePosition):
1149         (khtml::VisiblePosition::init):
1150         (khtml::VisiblePosition::initUpstream):
1151         (khtml::VisiblePosition::initDownstream):
1152         (khtml::VisiblePosition::next):
1153         (khtml::VisiblePosition::previous):
1154         (khtml::startVisiblePosition):
1155         (khtml::endVisiblePosition):
1156         * khtml/editing/visible_position.h:
1157         (khtml::VisiblePosition::):
1158         (khtml::VisiblePosition::VisiblePosition):
1159         (khtml::VisiblePosition::affinity):
1160         (khtml::VisiblePosition::setAffinity):
1161         (khtml::operator==):
1162         * khtml/editing/visible_units.cpp:
1163         (khtml::previousBoundary):
1164         (khtml::nextBoundary):
1165         (khtml::startOfLine):
1166         (khtml::endOfLine):
1167         (khtml::previousLinePosition):
1168         (khtml::nextLinePosition):
1169         (khtml::startOfParagraph):
1170         (khtml::endOfParagraph):
1171         (khtml::previousParagraphPosition):
1172         (khtml::nextParagraphPosition):
1173         (khtml::startOfBlock):
1174         (khtml::endOfBlock):
1175         (khtml::startOfDocument):
1176         (khtml::endOfDocument):
1177         * khtml/editing/visible_units.h:
1178         * khtml/khtml_part.cpp:
1179         (KHTMLPart::findTextNext):
1180         (KHTMLPart::selectClosestWordFromMouseEvent):
1181         (KHTMLPart::handleMousePressEventTripleClick):
1182         (KHTMLPart::handleMousePressEventSingleClick):
1183         (KHTMLPart::handleMouseMoveEventSelection):
1184         (KHTMLPart::khtmlMouseReleaseEvent):
1185         (KHTMLPart::selectAll):
1186         (KHTMLPart::computeAndSetTypingStyle):
1187         (KHTMLPart::selectionComputedStyle):
1188         * khtml/rendering/render_br.cpp:
1189         (RenderBR::positionForCoordinates):
1190         * khtml/xml/dom_docimpl.cpp:
1191         (DocumentImpl::updateSelection):
1192         * khtml/xml/dom_nodeimpl.cpp:
1193         (NodeBaseImpl::setFocus):
1194         * khtml/xml/dom_position.cpp:
1195         (DOM::Position::previousCharacterPosition):
1196         (DOM::Position::nextCharacterPosition):
1197         * khtml/xml/dom_position.h:
1198         * kwq/KWQAccObject.mm:
1199         (-[KWQAccObject value]):
1200         (-[KWQAccObject visiblePositionForStartOfTextMarkerRange:]):
1201         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
1202         (-[KWQAccObject accessibilityAttributeValue:]):
1203         (-[KWQAccObject doAXLineForTextMarker:]):
1204         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
1205         (-[KWQAccObject doAXTextMarkerForPosition:]):
1206         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
1207         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
1208         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
1209         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
1210         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
1211         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
1212         (-[KWQAccObject doSetAXSelectedTextMarkerRange:]):
1213         * kwq/KWQAccObjectCache.mm:
1214         (KWQAccObjectCache::textMarkerForVisiblePosition):
1215         (KWQAccObjectCache::visiblePositionForTextMarker):
1216         * kwq/KWQKHTMLPart.mm:
1217         (KWQKHTMLPart::findString):
1218         (KWQKHTMLPart::advanceToNextMisspelling):
1219         (KWQKHTMLPart::styleForSelectionStart):
1220         (KWQKHTMLPart::baseWritingDirectionForSelectionStart):
1221         (KWQKHTMLPart::setSelectionFromNone):
1222         (KWQKHTMLPart::respondToChangedSelection):
1223         * kwq/WebCoreBridge.mm:
1224         (-[WebCoreBridge setSelectedDOMRange:affinity:]):
1225         (-[WebCoreBridge selectionAffinity]):
1226         (-[WebCoreBridge setMarkDOMRange:]):
1227         (-[WebCoreBridge _visiblePositionForPoint:]):
1228         (-[WebCoreBridge moveDragCaretToPoint:]):
1229         (-[WebCoreBridge editableDOMRangeForPoint:]):
1230         (-[WebCoreBridge ensureSelectionVisible]):
1231         (-[WebCoreBridge rangeOfCharactersAroundCaret]):
1232
1233 2005-02-09  Chris Blumenberg  <cblu@apple.com>
1234
1235         Fixed: <rdar://problem/3985211> Seed: Mail: Drag-and-drop destination indicator / insertion point disappears
1236
1237         Reviewed by kocienda.
1238
1239         * khtml/rendering/render_block.cpp:
1240         (khtml::RenderBlock::paintCaret): new
1241         (khtml::RenderBlock::paintObject): call paintCaret for the cursor caret and the drag caret
1242         * khtml/rendering/render_block.h:
1243
1244 === Safari-185 ===
1245
1246 2005-02-08  Ken Kocienda  <kocienda@apple.com>
1247
1248         Reviewed by John
1249
1250         Fix for this bug:
1251
1252         <rdar://problem/3996344> Entire document content is deleted when only the first paragraph was supposed to be
1253
1254         Problem stems from my attempt to fix this bug: <rdar://problem/3986155> Insertion point goes
1255         to beginning of doc after deleting.
1256
1257         I added an "optimization" that proved to do more harm than good. So, basically, the fix is to remove code.
1258
1259         * khtml/editing/htmlediting.cpp:
1260         (khtml::DeleteSelectionCommand::handleGeneralDelete)
1261
1262 2005-02-07  Darin Adler  <darin@apple.com>
1263
1264         Reviewed by Kevin.
1265
1266         - fixed <rdar://problem/3890352> REGRESSION (125-170, Panther-only): pasted textarea contents disappear (exception in NSText)
1267
1268         * kwq/KWQTextArea.mm: (-[KWQTextArea tile]): Don't resize the text view immediately here.
1269         Use performWithDelay to defer it to avoid the re-entrancy that bothers the Panther version
1270         of NSTextView. On Tiger, leave the code alone.
1271
1272 2005-02-07  Adele Amchan  <adele@apple.com>
1273
1274         Reviewed by Darin.
1275
1276         Fix for <rdar://problem/3993628> REGRESSION (180-181): Logitelnet bank's website doesn't work
1277
1278         This fixes an oversight in the fix for <rdar://problem/3964286>.
1279         We're now making sure there is a document before calling checkCompleted in slotLoaderRequestDone.
1280         That function is called even when other parts finish loads of their subresources, so we could end
1281         up calling checkCompleted before we had a document. We also looked over all the other calls to
1282         checkCompleted to convince ourselves no other call sites had a similar issue.
1283
1284         * khtml/khtml_part.cpp: (KHTMLPart::slotLoaderRequestDone): Check if doc is nil before calling
1285         checkCompleted.
1286
1287 2005-02-07  Ken Kocienda  <kocienda@apple.com>
1288
1289         Reviewed by Darin
1290
1291         Fix for this bug:
1292
1293         <rdar://problem/3953302> Replacing quoted text ends up with blue-colored text that is not quoted
1294
1295         * khtml/editing/htmlediting.cpp:
1296         (khtml::DeleteSelectionCommand::doApply): Removed special case delete code for select all. It was
1297         causing as many bugs as it fixed. What's more, the design concept is a bit shaky. By removing
1298         this special case, the bug goes away.
1299         * khtml/editing/htmlediting.h: Remove declaration.
1300
1301         Test results using selectAll() updated to reflect changed behavior.
1302
1303         * layout-tests/editing/deleting/delete-3775172-fix-expected.txt:
1304         * layout-tests/editing/deleting/delete-select-all-001-expected.txt:
1305         * layout-tests/editing/deleting/delete-select-all-003-expected.txt:
1306         * layout-tests/editing/inserting/insert-3654864-fix-expected.txt:
1307         * layout-tests/editing/inserting/insert-3775316-fix-expected.txt:
1308         * layout-tests/editing/inserting/insert-3851164-fix-expected.txt:
1309         * layout-tests/editing/inserting/insert-3907422-fix-expected.txt:
1310         * layout-tests/editing/pasteboard/paste-text-015-expected.txt:
1311
1312 2005-02-06  Darin Adler  <darin@apple.com>
1313
1314         Reviewed by Ken.
1315
1316         - fixed <rdar://problem/3994164> setting a new style attribute with setAttribute("style") doesn't blow away the old style
1317
1318         * khtml/css/css_valueimpl.h: Renamed parseProperty to parseDeclaration, since it's used to parse
1319         an entire style declaration, not a single property.
1320         * khtml/css/css_valueimpl.cpp: (DOM::CSSMutableStyleDeclarationImpl::parseDeclaration): Renamed,
1321         and added code to clear m_values.
1322         * khtml/html/html_elementimpl.cpp: (HTMLElementImpl::parseHTMLAttribute): Call function by its
1323         new name, and remove some unnecessary comments.
1324
1325 2005-02-06  Darin Adler  <darin@apple.com>
1326
1327         Reviewed by Maciej.
1328
1329         - fixed <rdar://problem/3993722> need oncut and onpaste handlers implemented for <textarea> (needed for widgets)
1330
1331         * kwq/KWQTextArea.mm:
1332         (-[KWQTextAreaTextView dispatchHTMLEvent:]): Added.
1333         (-[KWQTextAreaTextView cut:]): Dispatch "before cut" and "cut" events.
1334         (-[KWQTextAreaTextView copy:]): Dispatch "before copy" and "copy" events.
1335         (-[KWQTextAreaTextView paste:]): Dispatch "before paste" and "paste" events.
1336         (-[KWQTextAreaTextView pasteAsPlainText:]): Ditto.
1337         (-[KWQTextAreaTextView pasteAsRichText:]): Ditto.
1338
1339 2005-02-06  Darin Adler  <darin@apple.com>
1340
1341         Reviewed by Maciej.
1342
1343         - fixed <rdar://problem/3943038> <input type=search> that is focused in onload handler doesn't have a visible editor
1344
1345         * khtml/xml/dom_docimpl.cpp: (DocumentImpl::setFocusNode): Always update layout before giving focus to a widget.
1346         This prevents the bad case where we give a widget focus before it has been positioned or sized, causing us to
1347         exercise edge cases AppKit doesn't handle well.
1348
1349         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Remove workaround code I added
1350         to fix bug 3943049. Updating the layout fixes the same problem in a cleaner, safer way, and solves another problem,
1351         with the way the clip view is set up.
1352
1353 2005-02-06  Darin Adler  <darin@apple.com>
1354
1355         Reviewed by Maciej.
1356
1357         - fixed <rdar://problem/3425232> textarea won't trigger onchange action when clicking to new field
1358
1359         * khtml/rendering/render_form.h: Added new m_dirty data member.
1360         * khtml/rendering/render_form.cpp:
1361         (RenderTextArea::RenderTextArea): Initialize m_dirty to false.
1362         (RenderTextArea::detach): Call value() on the DOM element to cause it to fetch the value from this object.
1363         (RenderTextArea::handleFocusOut): Ditto. Also send the onchange event based only on the m_dirty flag.
1364         (RenderTextArea::updateFromElement): Clear the m_dirty flag if the edited value is blown away by a value
1365         that comes from the DOM element. Remove code to clear the element's m_dirtyvalue flag; that's an unnecessary
1366         optimization, and ideally we'll stop having code in the render object that knows about that flag later.
1367         (RenderTextArea::slotTextChanged): Set the m_dirty flag.
1368
1369 2005-02-06  Darin Adler  <darin@apple.com>
1370
1371         Reviewed by Maciej.
1372
1373         - re-fixed <rdar://problem/3760910> please add support for custom tag names in HTML
1374
1375         * khtml/xml/dom_docimpl.cpp:
1376         (DocumentImpl::tagId): Use ID_LAST_TAG + 1 for the ID of the first assigned tag ID. Before we'd use ID_LAST_TAG,
1377         which resulted in the same number being used for ID_COMMENT and the first custom tag.
1378         (DocumentImpl::tagName): Use ID_LAST_TAG + 1, as above.
1379         (DocumentImpl::namespaceURI): Change a < ID_LAST_TAG to <= ID_LAST_TAG in the same spirit as above.
1380
1381 2005-02-06  Darin Adler  <darin@apple.com>
1382
1383         Reviewed by Maciej.
1384
1385         - fixed <rdar://problem/3986639> Crash occurs after choosing Undo Typing from the Edit menu
1386
1387         * kwq/KWQTextArea.mm: (-[KWQTextArea viewWillMoveToWindow:]): Remove actions from undo manager when the view is
1388         removed from the window. Unfortunately, to do this the code has to know that the target of NSTextView actions
1389         will be the text storage object, but given the NSText architecture, that's pretty clear.
1390
1391 2005-02-05  Chris Blumenberg  <cblu@apple.com>
1392
1393         Fixed: <rdar://problem/3991974> REGRESSION: www.jabra.com world location screen does not work
1394
1395         Reviewed by hyatt.
1396
1397         * kwq/KWQKJavaAppletWidget.mm:
1398         (KJavaAppletWidget::KJavaAppletWidget): fallback to the document's base URL if there is no baseURL attribute
1399         * kwq/WebCoreBridge.h:
1400         * kwq/WebCoreBridge.mm:
1401         (-[WebCoreBridge baseURL]): new
1402
1403 2005-02-04  David Harrison  <harrison@apple.com>
1404
1405         Reviewed by Darin.
1406
1407         Rolled out Ken's accidental checkin when committing <rdar://problem/3986155> fix yesterday.
1408
1409         * khtml/editing/selection.cpp:
1410         (khtml::Selection::validate):
1411
1412 2005-02-04  Adele Amchan  <adele@apple.com>
1413
1414         Reviewed by Darin.
1415
1416         Fix for <rdar://problem/3758785> Safari no longer works at BankOfAmerica online banking for military users
1417
1418         When closing the document, if a body doesn't exist we now create one.  This bug was preventing the onload handler from
1419         firing in cases where there wasn't a body.  In the BankOfAmerica case, the page was just a bit of script where the onload
1420         handler redirected to the correct page.
1421
1422         * khtml/xml/dom_docimpl.cpp: (DocumentImpl::close):
1423
1424 2005-02-03  Ken Kocienda  <kocienda@apple.com>
1425
1426         Reviewed by Chris
1427
1428         Fix for this bug:
1429
1430         <rdar://problem/3986155> Insertion point goes to beginning of doc after deleting
1431
1432         * khtml/editing/htmlediting.cpp:
1433         (khtml::DeleteSelectionCommand::setStartNode): New convenience to handle reference counting when setting.
1434         (khtml::DeleteSelectionCommand::handleGeneralDelete): This contains the crux of the bug fix. Improve tests
1435         that detect when a selected node needs to be retained, rather than deleted, to preserve the intent of the user.
1436         This has the side effect of causing the insertion point placement code to succeed rather than fail. Before
1437         this fix, the failure of the insertion point placement code caused the insertion point to jump to the start
1438         of the document, which is the symptom that can be perceived by users when editing.
1439         * khtml/editing/htmlediting.h: Add setStartNode declaration.
1440         * khtml/editing/visible_units.cpp:
1441         (khtml::startOfBlock): This function had a stubbed-in non-tested implementation. Implement and
1442         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-001-expected.txt: Added.
1443         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-001.html: Added.
1444         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002-expected.txt: Added.
1445         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-002.html: Added.
1446         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-003-expected.txt: Added.
1447         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-003.html: Added.
1448         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-004-expected.txt: Added.
1449         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-004.html: Added.
1450         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-005-expected.txt: Added.
1451         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-005.html: Added.
1452         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-006-expected.txt: Added.
1453         * layout-tests/editing/deleting/delete-at-paragraph-boundaries-006.html: Added.
1454
1455 === Safari-183 ===
1456
1457 2005-02-03  Richard Williamson   <rjw@apple.com>
1458
1459         Fixed <rdar://problem/3987419> Stocks and Weather leak what appears to XMLHTTPRequest results
1460
1461         XMLHTTPRequests were causing massive leaks.  _webcore_initWithHeaderString: did funky things with
1462         self replacement.  Re-wrote to use more traditional factory constructor avoiding self replacement.
1463
1464         Reviewed by David Harrison.
1465
1466         * kwq/KWQLoader.mm:
1467         (+[NSDictionary _webcore_dictionaryWithHeaderString:]):
1468         (KWQServeRequest):
1469         (KWQServeSynchronousRequest):
1470
1471 2005-02-03  Chris Blumenberg  <cblu@apple.com>
1472
1473         Fixed: <rdar://problem/3938763> Cannot view Windows Media Player videos (soundtrack is played twice with delay)
1474
1475         Reviewed by hyatt.
1476
1477         * khtml/html/html_objectimpl.cpp:
1478         (HTMLObjectElementImpl::attach): set needWidgetUpdate to false right after calling updateWidget because dispatchHTMLEvent can end up calling this method again
1479         (HTMLObjectElementImpl::recalcStyle): ditto
1480
1481 2005-02-02  John Sullivan  <sullivan@apple.com>
1482
1483         Reviewed by Chris.
1484
1485         - WebCore part of fix for <rdar://problem/3980651> REGRESSION (125-180): Huge number of pages printed from certain page,
1486         iFrame involved
1487
1488         This also fixes the problems with printing from GMail, yay!
1489
1490         * khtml/rendering/render_flow.cpp:
1491         (RenderFlow::paintLines):
1492         If the current line is taller than the entire page height (e.g. tall iFrame), don't try
1493         to avoid splitting it across pages.
1494
1495         * kwq/WebCoreBridge.mm:
1496         (-[WebCoreBridge computePageRectsWithPrintWidthScaleFactor:printHeight:]):
1497         Handle error cases in a more obvious manner; this will cause future problems like this
1498         to print a blank page and complain to the console on debug builds, rather than print a
1499         zillionty mostly-blank pages.
1500
1501 2005-02-02  Chris Blumenberg  <cblu@apple.com>
1502
1503         Fixed: <rdar://problem/3960304> can't load a particular applet (at www.escape.de) unless it's the first applet to be loaded
1504
1505         Reviewed by hyatt.
1506
1507         * khtml/rendering/render_frames.cpp:
1508         (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.
1509
1510 2005-02-02  Chris Blumenberg  <cblu@apple.com>
1511
1512         Fixed: <rdar://problem/3983628> control-click on WebView is not selecting the word under the cursor (Mail, non-editable WebView)
1513
1514         Reviewed by rjw.
1515
1516         * kwq/KWQKHTMLPart.mm:
1517         (KWQKHTMLPart::sendContextMenuEvent): check if SPI to always enable selecting closest word is enabled
1518         * kwq/WebCoreBridge.h:
1519
1520 2005-02-02  Ken Kocienda  <kocienda@apple.com>
1521
1522         Reviewed by me
1523
1524         * khtml/editing/markup.cpp:
1525         (khtml::createMarkup): Removed debug spam I committed earlier in error.
1526
1527 2005-02-02  Ken Kocienda  <kocienda@apple.com>
1528
1529         Reviewed by Darin
1530
1531         Fix for this bug:
1532
1533         <rdar://problem/3823109> WebKit should support -toggleBaseWritingDirection: (bidi editing support)
1534
1535         Since base writing direction is a paragraph-level property in AppKit, and we use the CSS direction
1536         property in WebCore, which can be applied to inline elements as well as blocks, a new notion has
1537         been added to the WebCore apply style logic. It is now possible to force all properties in a
1538         style declaration to be applied as though they were block properties.
1539
1540         * khtml/editing/htmlediting.cpp:
1541         (khtml::ApplyStyleCommand::ApplyStyleCommand): Add an enum and a flag to this command that controls whether to force all
1542         properties in the style to be applied as block styles. Default is "no", retaining previous behavior.
1543         (khtml::ApplyStyleCommand::doApply): Switch on new flag to apply styles as before, or force all preoperties to be applied
1544         as block styles.
1545         * khtml/editing/htmlediting.h:
1546         (khtml::ApplyStyleCommand::): Add an enum and a flag, as above.
1547         * khtml/khtml_part.cpp:
1548         (KHTMLPart::applyParagraphStyle): New method to force application of all style properties as block styles.
1549         * khtml/khtml_part.h: Update header declarations.
1550         * kwq/KWQKHTMLPart.h: Update header declarations.
1551         * kwq/KWQKHTMLPart.mm:
1552         (KWQKHTMLPart::baseWritingDirectionForSelectionStart): Accessor to help WebKit do a toggle operation on
1553         writing direction.
1554         * kwq/WebCoreBridge.h: Update header declarations.
1555         * kwq/WebCoreBridge.mm:
1556         (-[WebCoreBridge applyParagraphStyle:withUndoAction:]): New method to force application of all style properties as block styles.
1557         (-[WebCoreBridge baseWritingDirectionForSelectionStart]): Accessor to help WebKit do a toggle operation on
1558         writing direction.
1559
1560         Fix for this bug:
1561
1562         <rdar://problem/3985035> REGRESSION (Mail): Text copied from wrapped line contains extra character when pasted
1563
1564         * khtml/editing/markup.cpp:
1565         (khtml::renderedText): Add an enum and a flag to this command that controls whether to force all
1566         This is the result of an error in the code that computes the rendered text that is selected
1567         when copying. Since spaces collapse at the end of lines, and these spaces need to be copied
1568         when the selection spans line endings, code runs to compute this text. However, this code
1569         was also running incorrectly in cases where lines wrapped. I have now added the missing
1570         test to check that the selection does indeed extend to the end of the line.
1571
1572
1573 2005-02-02  Ken Kocienda  <kocienda@apple.com>
1574
1575         Reviewed by Darin
1576
1577         Fix for these bugs:
1578
1579         <rdar://problem/3984894> REGRESSION (Mail): Command-right-arrow in reply does the wrong thing (two ways)
1580         <rdar://problem/3985130> REGRESSION (Mail): command-right-arrow in pasted RTF selects only up to tab
1581
1582         * khtml/editing/visible_units.cpp:
1583         (khtml::endOfLine): Two separate problems in this new function recently added to take the place of
1584         selectionForLine. In the first, endOfLine did not stop at BR elements, and returned the position
1585         beyond them. In the second, the VisiblePosition constructor gives the wrong answer in certain cases.
1586         Darin is going to work on a solution for that problem. In the meantime, I can fix the symptom of this
1587         bug by going down to the last leaf child of the root line box; a one line change we will roll out when
1588         Darin's fix is in.
1589
1590 2005-02-01  Richard Williamson   <rjw@apple.com>
1591
1592         Fixed <rdar://problem/3985535> QT Plug-in JavaScript support now fails in <embed> tag only case
1593
1594         Reviewed by Maciej.
1595
1596         * khtml/ecma/kjs_html.cpp:
1597         (KJS::HTMLDocument::tryGet):
1598
1599 2005-02-01  John Sullivan  <sullivan@apple.com>
1600
1601         Written by Darin, reviewed and tested by me
1602
1603         - fixed <rdar://problem/3969684> Panther-only: extra blank lines between
1604         pasted content from Blot
1605
1606         * kwq/KWQKHTMLPart.mm:
1607         (KWQKHTMLPart::attributedString):
1608         don't emit paragraph breaks if the margins are tiny; matches what
1609         we do for plain-text conversion
1610
1611 === Safari-182 ===
1612
1613 2005-02-01  Ken Kocienda  <kocienda@apple.com>
1614
1615         Reviewed by John
1616
1617         Fix for this bug:
1618
1619         <rdar://problem/3985160> Deficiencies in pasting architecture blocking progress on other bugs
1620
1621         * khtml/editing/html_interchange.h: Move style span text used to mark element added to
1622         add style to this header.
1623         * khtml/editing/htmlediting.cpp:
1624         (khtml::styleSpanClassString): Change to use constant moved to html_interchange.h.
1625         (khtml::isStyleSpan): New helper function. Checks if this is a span we added to apply style.
1626         (khtml::CompositeEditCommand::insertNodeBefore): Added an assert to check that the node
1627         we are inserting before is not the body.
1628         (khtml::CompositeEditCommand::insertNodeAfter): Ditto, but check is for after.
1629         (khtml::ReplacementFragment::ReplacementFragment): Added code to process the "default style"
1630         that is added by the copy code.
1631         (khtml::ReplaceSelectionCommand::ReplaceSelectionCommand): first and last nodes inserted are
1632         now member variables instead of function locals. Initialize them here.
1633         (khtml::ReplaceSelectionCommand::~ReplaceSelectionCommand): Deref first and last nodes inserted
1634         if necessary.
1635         (khtml::ReplaceSelectionCommand::doApply): Change design to fix the bug. Major change is to
1636         separate out the code that inserts nodes into the tree so additional styling checks can
1637         be done in a centralized way. Also got rid of the notion of "merging into the end block." That
1638         concept was just wrong.
1639         (khtml::ReplaceSelectionCommand::completeHTMLReplacement): Tweak interface now that first and
1640         last nodes inserted are member variables.
1641         (khtml::ReplaceSelectionCommand::insertNodeAfterAndUpdateNodesInserted): New helper used
1642         by replace code to do the stated DOM operation and update state internal to the command.
1643         This will also be a catch point to handle the kinds of additional style checks needed to
1644         make paste work right.
1645         (khtml::ReplaceSelectionCommand::insertNodeAtAndUpdateNodesInserted): Ditto.
1646         (khtml::ReplaceSelectionCommand::insertNodeBeforeAndUpdateNodesInserted): Ditto.
1647         (khtml::ReplaceSelectionCommand::updateNodesInserted): Ditto.
1648         * khtml/editing/htmlediting.h: Update declarations as needed.
1649         * khtml/editing/markup.cpp:
1650         (khtml::createMarkup): Adds a "default style" span to the content written to the pasteboard.
1651         This will help us to fix some of the bugs blocked by the bug above.
1652         * khtml/xml/dom_nodeimpl.cpp:
1653         (NodeImpl::lastDescendent): New helper.
1654         * khtml/xml/dom_nodeimpl.h: Ditto.
1655         * khtml/xml/dom_position.cpp:
1656         (DOM::Position::upstream): Fixed a bug which would allow the upstream position returned to be
1657         in unrendered content.
1658         (DOM::Position::downstream): Ditto.
1659
1660         * layout-tests/editing/deleting/delete-3775172-fix-expected.txt: Ending positions tweaked due to
1661         changes in upstream() and downstream() functions.
1662         * layout-tests/editing/inserting/insert-3851164-fix-expected.txt: Ditto
1663         * layout-tests/editing/inserting/insert-3907422-fix-expected.txt: Ditto
1664         * layout-tests/editing/selection/extend-by-character-006-expected.txt: Ditto
1665
1666         * layout-tests/editing/pasteboard/paste-text-003-expected.txt: Changed what we expect, given
1667         new behavior of paste code.
1668
1669 2005-01-31  John Sullivan  <sullivan@apple.com>
1670
1671         Reviewed by Dave Hyatt.
1672
1673         - fixed <rdar://problem/3983097> Tabbing on RSS pages gets stuck in search field
1674         with "full keyboard navigation" off
1675
1676         * kwq/KWQSlider.h:
1677         * kwq/KWQSlider.mm:
1678         (QSlider::focusPolicy):
1679         Implemented this method for KWQSlider. It is needed for our widget subclasses that
1680         can become focused in some situations. I missed this when fixing 3949203 because
1681         it only comes into play when tab-to-controls is on, and because the other relevant
1682         methods are objective-C methods.
1683
1684 2005-01-31  David Harrison  <harrison@apple.com>
1685
1686         Reviewed by Darin.
1687
1688         <rdar://problem/3964164> AXWebArea needs to unregister on going invisible
1689
1690         * khtml/xml/dom_docimpl.cpp:
1691         (DocumentImpl::detach):
1692         Detach the AX UIElement for the AXWebArea.
1693
1694 2005-01-31  David Harrison  <harrison@apple.com>
1695
1696         Reviewed by Darin
1697
1698         <rdar://problem/3850876> AX: button exposed as only an AXImage at http://appleseed.apple.com/ site (<input type=image>)
1699
1700         * khtml/rendering/render_form.h:
1701         (khtml::RenderImageButton::isImageButton):
1702         New.
1703         * khtml/rendering/render_image.h:
1704         (khtml::RenderImage::isImageButton):
1705         New.
1706         * kwq/KWQAccObject.mm:
1707         (-[KWQAccObject isImageButton]):
1708         (-[KWQAccObject actionElement]):
1709         (-[KWQAccObject role]):
1710         (-[KWQAccObject accessibilityActionNames]):
1711         (-[KWQAccObject accessibilityPerformAction:]):
1712         Represent RenderImageButton elements as buttons instead of images.
1713
1714 2005-01-31  David Harrison  <harrison@apple.com>
1715
1716         Prevent nil dereference in debug code.
1717
1718         * khtml/xml/dom_nodeimpl.cpp:
1719         (NodeImpl::displayTree):
1720         Stop when parent is 0, too.
1721
1722 2005-01-31  John Sullivan  <sullivan@apple.com>
1723
1724         Reviewed by Dave Hyatt.
1725
1726         - fixed <rdar://problem/3949203> cannot tab to, within, or out of the RSS sidebar
1727
1728         * kwq/KWQSlider.mm:
1729         (-[KWQSlider becomeFirstResponder]):
1730         (-[KWQSlider resignFirstResponder]):
1731         (-[KWQSlider nextKeyView]):
1732         (-[KWQSlider previousKeyView]):
1733         (-[KWQSlider canBecomeKeyView]):
1734         (-[KWQSlider nextValidKeyView]):
1735         (-[KWQSlider previousValidKeyView]):
1736         Copied these methods from KWQButton. The lack of special handling for first-responder-ness
1737         here was causing the focus to get "stuck" on the RSS page's slider widget. Added FIXME
1738         about how it would be nice to share more of this code rather than replicating it in each
1739         KWQ widget subclass.
1740
1741 2005-01-31  Darin Adler  <darin@apple.com>
1742
1743         Reviewed by Harrison.
1744
1745         - fixed <rdar://problem/3980066> Double-click on single character moves insertion point to previous line
1746
1747         * khtml/khtml_part.cpp:
1748         (KHTMLPart::selectClosestWordFromMouseEvent): Set affinity too.
1749         (KHTMLPart::handleMousePressEventTripleClick): Ditto.
1750         (KHTMLPart::handleMouseMoveEventSelection): Ditto.
1751         (KHTMLPart::khtmlMouseReleaseEvent): Ditto.
1752
1753         * khtml/editing/selection.cpp:
1754         (khtml::Selection::modifyExtendingRightForward): Use endOfLine and endOfDocument.
1755         (khtml::Selection::modifyMovingRightForward): Ditto.
1756         (khtml::Selection::modifyExtendingLeftBackward): Use startOfLine and startOfDocument.
1757         (khtml::Selection::modifyMovingLeftBackward): Ditto.
1758         (khtml::Selection::validate): Rewrote the section that handles double-click. Two main fixes: 1) use isStartOfLine to
1759         check for another case where we want to select the word to the right, and 2) use isEndOfParagraph, which seems
1760         to work correctly in cases where isLastVisiblePositionInParagraph is giving the wrong answer. Also changed the line
1761         code to use startOfLine/endOfLine and the document code to use startOfDocument/endOfDocument.
1762
1763 2005-01-31  Darin Adler  <darin@apple.com>
1764
1765         Reviewed by Harrison.
1766
1767         - fixed <rdar://problem/3935275> unexpected quit scrolling over link; last.fm (works in IE and Firefox)
1768
1769         * kwq/WebCoreBridge.mm: (-[WebCoreBridge elementAtPoint:]): By using the correct document, fix the crash reported here.
1770         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.
1771
1772 2005-01-31  Darin Adler  <darin@apple.com>
1773
1774         Reviewed by Ken and Harrison.
1775
1776         - fixed <rdar://problem/3947901> REGRESSION (Mail): Pasting paragraph of rich text leaves insertion point before pasted text
1777         - fixed <rdar://problem/3949790> hitting return after underlined line results in too much or too little underlined
1778         - fixed <rdar://problem/3981759> nil-deref and crash when pasting just a paragraph break
1779         - fixed a couple problems I discovered while working with bug 3949790
1780
1781         * khtml/editing/htmlediting.cpp:
1782         (khtml::ApplyStyleCommand::applyInlineStyle): Pass StayInBlock to upstream. Without this, we end up going too far
1783         upstream in the test case in bug 3949790.
1784         (khtml::ApplyStyleCommand::removeInlineStyle): Pass StayInBlock to upstream and downstream. Same reason as above.
1785         (khtml::ReplaceSelectionCommand::doApply): Update endPos if inserting a new node and endPos is using that node's
1786         parent and an offset past the node being inserted. That change fixes a problem with the position of the insertion point
1787         after pasting into the top level of a document (from test cases in 3947901 and 3949790). When setting insertionPos, use
1788         code that works when lastNodeInserted is a block rather than a text node. That change fixes a problem where a newline is
1789         not added when pasting an entire paragraph into the end of a document (from test case in 3949790). Added nil check before
1790         checking if lastNodeInserted is a <br> element, which fixes the crash when pasting just a paragraph break.
1791
1792         * khtml/editing/visible_units.h: Filled out the set of calls to add some boolean checks for lines (needed for the
1793         bug fix), and calls for blocks (not yet implemented), and documents. The document checks may need refinement to
1794         properly handle documents with a mix of editable and non-editable content, but for now they just refactor code
1795         and make things a little clearer. Also removed the "include line break" parameter from endOfSentence.
1796         * khtml/editing/visible_units.cpp:
1797         (khtml::rootBoxForLine): Added.
1798         (khtml::startOfLine): Added. Algorithm taken from selectionForLine in selection.cpp.
1799         (khtml::endOfLine): Ditto.
1800         (khtml::inSameLine): Added.
1801         (khtml::isStartOfLine): Added.
1802         (khtml::isEndOfLine): Added.
1803         (khtml::endOfSentence): Removed "include line break" parameter.
1804         (khtml::inSameParagraph): Added a null check.
1805         (khtml::isStartOfParagraph): Ditto.
1806         (khtml::isEndOfParagraph): Ditto.
1807         (khtml::startOfBlock): Added.
1808         (khtml::endOfBlock): Added.
1809         (khtml::inSameBlock): Added.
1810         (khtml::isStartOfBlock): Added.
1811         (khtml::isEndOfBlock): Added.
1812         (khtml::startOfDocument): Added.
1813         (khtml::endOfDocument): Added.
1814         (khtml::inSameDocument): Added.
1815         (khtml::isStartOfDocument): Added.
1816         (khtml::isEndOfDocument): Added.
1817
1818 2005-01-30  Darin Adler  <darin@apple.com>
1819
1820         Reviewed by John.
1821
1822         - fixed <rdar://problem/3977000> form data set posted for <input type="image" ...> omits name/value pair
1823
1824         * khtml/html/html_formimpl.cpp:
1825         (DOM::HTMLInputElementImpl::appendFormData): Append a name/value pair for type IMAGE in addition to the X/Y
1826         coordinates.
1827
1828 2005-01-30  Darin Adler  <darin@apple.com>
1829
1830         Reviewed by John.
1831
1832         - fixed <rdar://problem/3974246> REGRESSION (125-180): popup menus are missing item text on Harmony Remote web site
1833
1834         * khtml/html/html_formimpl.cpp: (DOM::HTMLOptionElementImpl::text): Change this function to traverse the entire tree
1835         and gather all the text rather than just looking at immediate children.
1836
1837         * khtml/xml/dom_nodeimpl.h: Added const.
1838         * khtml/xml/dom_nodeimpl.cpp:
1839         (NodeImpl::traverseNextNode): Added const.
1840         (NodeImpl::traverseNextSibling): Ditto.
1841         (NodeImpl::traversePreviousNodePostOrder): Ditto.
1842         (NodeImpl::detach): Add a missing nil check.
1843
1844 2005-01-30  Darin Adler  <darin@apple.com>
1845
1846         Reviewed by John.
1847
1848         - fixed <rdar://problem/3864536> crash copying text from other apps and pasting in a sticky in dashboard
1849
1850         * khtml/editing/markup.cpp: (khtml::createFragmentFromText): Added missing ref/deref of newly created element.
1851
1852 2005-01-28  Darin Adler  <darin@apple.com>
1853
1854         * khtml/css/cssproperties.c: Regenerated with newer gperf.
1855
1856 2005-01-28  David Harrison  <harrison@apple.com>
1857
1858         Reviewed by John Sullivan.
1859
1860         <rdar://problem/3968144> AX need to be able to focus an AXLink by setting AXFocused to true
1861
1862         * kwq/KWQAccObject.mm:
1863         (-[KWQAccObject accessibilityPerformAction:]):
1864         (-[KWQAccObject accessibilityIsAttributeSettable:]):
1865         Add AXFocused for AXLinks.
1866
1867 2005-01-28  David Harrison  <harrison@apple.com>
1868
1869         Reviewed by Darin.
1870
1871         <rdar://problem/3584942> AX: Safari Accessibility parent-child mismatch
1872
1873         Use AppKit SPI _accessibilityParentForSubview to return KWQAccObject parent of AppKit AX object.
1874
1875         * kwq/KWQAccObject.mm:
1876         (-[KWQAccObject _accessibilityParentForSubview:]):
1877         New.
1878
1879 2005-01-27  Adele Amchan  <adele@apple.com>
1880
1881         fixed by Darin, reviewed by me.
1882
1883         <rdar://problem/3976314> REGRESSION (180-TOT): submitting password fields fail on Panther
1884
1885         * kwq/KWQTextField.mm: (-[KWQTextFieldController string]): on panther, the secure text field's editor
1886         does not contain the real string, so now we just call stringValue on the field in that case.
1887
1888 2005-01-26  Maciej Stachowiak  <mjs@apple.com>
1889
1890         Reviewed by Darin, Hyatt and Ken.
1891
1892         <rdar://problem/3790449> REGRESSION (Mail): underline behavior is flaky because of how CSS handles text-decoration
1893
1894         * khtml/css/css_computedstyle.cpp:
1895         (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): Add support
1896         for -khtml-text-decorations-in-effect property. This works like text-docration
1897         but follows proper inline semantics, that is, if a parent is causing text
1898         decoration then -khtml-text-decorations-in-effect on this element will reflect that.
1899         * khtml/css/cssproperties.in: Add -khtml-text-decorations-in-effect property.
1900         * khtml/css/cssproperties.h: regenerated
1901         * khtml/css/cssvalues.c: regenerated
1902         * khtml/css/cssparser.cpp: regenerated
1903         * khtml/css/cssproperties.c: regenerated
1904
1905         * khtml/css/css_valueimpl.h: Made copyPropertiesInSet public.
1906
1907         * khtml/editing/htmlediting.cpp:
1908         (khtml::StyleChange::init): Handle -khtml-text-decorations-in-effect
1909         specially, translate to text-decoration
1910         (khtml::ApplyStyleCommand::applyInlineStyle): Fix a comment.
1911         (khtml::ApplyStyleCommand::removeInlineStyle): Add new way of pushing down
1912         text-decoration styles around the edges of the selected range.
1913         (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAroundNode): Helper for this.
1914         (khtml::ApplyStyleCommand::pushDownTextDecorationStyleAtBoundaries): ditto
1915         (khtml::ApplyStyleCommand::nodeFullyUnselected): ditto
1916         (khtml::hasTextDecorationProperty): ditto
1917         (khtml::highestAncestorWithTextDecoration): ditto
1918         (khtml::ApplyStyleCommand::extractTextDecorationStyle): ditto
1919         (khtml::ApplyStyleCommand::extractAndNegateTextDecorationStyle): ditto
1920         (khtml::ApplyStyleCommand::applyTextDecorationStyle): ditto
1921         * khtml/editing/htmlediting.h: Prototype new methods.
1922         * khtml/editing/jsediting.cpp: Make underline execCommand use the new CSS property.
1923
1924         - as a necessary part of the fix, fixed the longstanding problem
1925         that mutating an element's inlineStyleDecl would not property
1926         update its style attribute. now it does.
1927
1928         * khtml/css/css_valueimpl.cpp:
1929         (DOM::CSSMutableStyleDeclarationImpl::setChanged): If this is an inline style
1930         declaration, tell the element it's style attribute needs updating.
1931         * khtml/html/html_elementimpl.cpp:
1932         (HTMLElementImpl::invalidateStyleAttribute): New method, just calls down
1933         to the dom element.
1934         (HTMLElementImpl::updateStyleAttribute): Regenerate style attribute from
1935         inline style declaration.
1936         * khtml/html/html_elementimpl.h: Prototype new stuff.
1937         * khtml/xml/dom_elementimpl.cpp:
1938         (ElementImpl::ElementImpl): Initialize new attribute.
1939         (ElementImpl::updateStyleAttributeIfNeeded): New method to check if the style
1940         is dirty and this is an html element, and if so call the html element to
1941         regenerate the style attribute.
1942         (ElementImpl::attributes): update style if needed
1943         (ElementImpl::getAttribute): ditto
1944         (ElementImpl::hasAttributes): ditto
1945         (ElementImpl::dump): ditto
1946         (XMLElementImpl::cloneNode): ditto
1947         * khtml/xml/dom_elementimpl.h:
1948         (DOM::ElementImpl::isStyleAttributeValid): new method to manage style attr
1949         dirty bit
1950         (DOM::ElementImpl::setStyleAttributeValid): ditto
1951
1952         - added new layout tests for underlining:
1953
1954         * layout-tests/editing/editing.js:
1955         * layout-tests/editing/style/remove-underline-across-paragraph-expected.txt: Added.
1956         * layout-tests/editing/style/remove-underline-across-paragraph-in-bold-expected.txt: Added.
1957         * layout-tests/editing/style/remove-underline-across-paragraph-in-bold.html: Added.
1958         * layout-tests/editing/style/remove-underline-across-paragraph.html: Added.
1959         * layout-tests/editing/style/remove-underline-after-paragraph-expected.txt: Added.
1960         * layout-tests/editing/style/remove-underline-after-paragraph-in-bold-expected.txt: Added.
1961         * layout-tests/editing/style/remove-underline-after-paragraph-in-bold.html: Added.
1962         * layout-tests/editing/style/remove-underline-after-paragraph.html: Added.
1963         * layout-tests/editing/style/remove-underline-expected.txt: Added.
1964         * layout-tests/editing/style/remove-underline-from-stylesheet-expected.txt: Added.
1965         * layout-tests/editing/style/remove-underline-from-stylesheet.html: Added.
1966         * layout-tests/editing/style/remove-underline-in-bold-expected.txt: Added.
1967         * layout-tests/editing/style/remove-underline-in-bold.html: Added.
1968         * layout-tests/editing/style/remove-underline.html: Added.
1969         * layout-tests/editing/style/underline-expected.txt: Added.
1970         * layout-tests/editing/style/underline.html: Added.
1971
1972 2005-01-27  David Hyatt  <hyatt@apple.com>
1973
1974         Fix for 3875199, search field needs to be able to  show a magnifying  glass without a dropdown menu.
1975
1976         Reviewed by kocienda
1977
1978         * khtml/html/html_formimpl.cpp:
1979         (DOM::HTMLInputElementImpl::HTMLInputElementImpl):
1980         (DOM::HTMLInputElementImpl::parseHTMLAttribute):
1981         * kwq/KWQLineEdit.mm:
1982         (QLineEdit::setMaxResults):
1983
1984 2005-01-27  Ken Kocienda  <kocienda@apple.com>
1985
1986         Reviewed by Maciej
1987
1988         Fix for this bug:
1989
1990         <rdar://problem/3973254> Deletions of ranges does not coalesce correctly with subsequent typing
1991
1992         * khtml/editing/htmlediting.cpp:
1993         (khtml::InsertTextCommand::deleteCharacter): Remove this dead code.
1994         (khtml::ReplaceSelectionCommand::editingAction): Moved this code to a better place in the file.
1995         It was in with unrelated functions.
1996         (khtml::TypingCommand::TypingCommand): Reorganize initialization list so it is easier to read.
1997         (khtml::TypingCommand::issueCommandForDeleteKey): Remove dead code. Roll in remaining code into deleteKeyPressed.
1998         (khtml::TypingCommand::deleteKeyPressed): Add in support for carrying along smart delete flag.
1999         (khtml::TypingCommand::forwardDeleteKeyPressed): New function. Makes forward delete work like
2000         "regular" delete in terms of coalescing typing.
2001         (khtml::TypingCommand::doApply): Add case for ForwardDeleteKey.
2002         (khtml::TypingCommand::preservesTypingStyle): Ditto.
2003         * khtml/editing/htmlediting.h:
2004         (khtml::TypingCommand::): Add ForwardDeleteKey constant. Remove a couple declarations for now-dead code.
2005         (khtml::TypingCommand::smartDelete): New accessor.
2006         (khtml::TypingCommand::setSmartDelete): Ditto.
2007         * kwq/WebCoreBridge.h:
2008         * kwq/WebCoreBridge.mm:
2009         (-[WebCoreBridge deleteKeyPressedWithSmartDelete:]): Add smart delete flag.
2010         (-[WebCoreBridge forwardDeleteKeyPressedWithSmartDelete:]): New method.
2011         * khtml/editing/jsediting.cpp: Add supprt for ForwardDelete command
2012         * layout-tests/editing/editing.js: Ditto.
2013         * layout-tests/editing/deleting/delete-and-undo-expected.txt: Added.
2014         * layout-tests/editing/deleting/delete-and-undo.html: Added.
2015         * layout-tests/editing/deleting/forward-delete-expected.txt: Added.
2016         * layout-tests/editing/deleting/forward-delete.html: Added.
2017
2018 === Safari-181 ===
2019
2020 2005-01-27  David Harrison  <harrison@apple.com>
2021
2022         Reviewed by Ken.
2023
2024         <rdar://problem/3962214> AX: AXLeftLineTextMarkerRangeForTextMarker returns incorrect range
2025
2026         * kwq/KWQAccObject.mm:
2027         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
2028         Allowed for selection to end of line including the linebreak.
2029
2030 2005-01-27  David Harrison  <harrison@apple.com>
2031
2032         Reviewed by Darin, Ken.
2033
2034         <rdar://problem/3964470> AX: Include attachments in AXAttributedStringForTextMarkerRange
2035
2036         * kwq/KWQAccObject.mm:
2037         (-[KWQAccObject isAttachment]):
2038         (-[KWQAccObject attachmentView]):
2039         (-[KWQAccObject role]):
2040         (-[KWQAccObject subrole]):
2041         (-[KWQAccObject roleDescription]):
2042         (-[KWQAccObject value]):
2043         (-[KWQAccObject title]):
2044         (-[KWQAccObject accessibilityDescription]):
2045         (-[KWQAccObject accessibilityIsIgnored]):
2046         (-[KWQAccObject accessibilityAttributeNames]):
2047         (-[KWQAccObject accessibilityAttributeValue:]):
2048         Make attachments accessible.
2049
2050         (CreateCGColorIfDifferent):
2051         (AXAttributeStringSetColor):
2052         (AXAttributeStringSetNumber):
2053         (AXAttributeStringSetFont):
2054         (AXAttributeStringSetStyle):
2055         (AXAttributeStringSetElement):
2056         (AXLinkElementForNode):
2057         (AXAttributedStringAppendText):
2058         (AXAttributedStringAppendReplaced):
2059         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
2060         Add attachments and more text attributes to AXAttributedStringForTextMarkerRange.
2061
2062 2005-01-27  Darin Adler  <darin@apple.com>
2063
2064         Reviewed by John.
2065
2066         - fixed <rdar://problem/3807935> DOM CSS computed style line-height is wrong in two ways
2067
2068         * khtml/css/css_computedstyle.cpp: (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue):
2069         Use specified size rather than computed size when dealing with a percentage for line height.
2070
2071 2005-01-27  Darin Adler  <darin@apple.com>
2072
2073         Reviewed by John.
2074
2075         - fixed <rdar://problem/3971372> SWB: template function DOM_cast() won't compile with gcc 4.0
2076         - fixed other gcc 4.0 compiling problems Patrick Beard pointed out on the phone
2077
2078         * kwq/DOMInternal.h: (DOM_cast): Change the "failToCompile()" trick to use a dependent name.
2079         Nowadays, if a name is not dependent, it is checked at template definition time. But we want
2080         a failure only at template instantiation time.
2081
2082         * ForwardingHeaders/editing/text_granularity.h: Added. Needed but a bug in the 3.3 compiler made
2083         it compile anyway without this.
2084
2085         - other changes
2086
2087         * khtml/html/html_miscimpl.h: For clarity, mark these functions virtual too. They are automatically
2088         virtual because the base class ones are virtual, but it's ugly to leave it this way.
2089
2090 2005-01-26  Ken Kocienda  <kocienda@apple.com>
2091
2092         Reviewed by Hyatt
2093
2094         Fix for this bug:
2095
2096         <rdar://problem/3971609> REGRESSION (Mail): up/down arrow navigation broken after rewrapping text by resizing window
2097         <rdar://problem/3975661> REGRESSION (Mail): left/right arrow navigation can place insertion point after last character on line
2098
2099         * khtml/editing/selection.cpp:
2100         (khtml::Selection::modifyAffinity): PARAGRAPH and LINE movements should not alter affinity. This was
2101         just a mistake before when it reset to upstream in some cases.
2102         * khtml/rendering/render_line.cpp:
2103         (khtml::RootInlineBox::closestLeafChildForXPos): Move this function down from InlineBox, and
2104         rework the implementation. This fixes 3971609.
2105         * khtml/rendering/render_line.h:
2106         * khtml/rendering/render_text.cpp:
2107         (RenderText::caretRect): Rework the algorithm to fix 3975661. This regressed when we began placing
2108         the spaces on the end of lines in text boxes when lines wrap.
2109
2110 2005-01-26  Richard Williamson   <rjw@apple.com>
2111
2112         Fixed <rdar://problem/3972522> (179-180) 40% slowdown on iBench JavaScript test
2113
2114         I added a member variable to ObjectImp.  This changed it's size and consequently
2115         hampered the optimizations built into the garbage collector.  Objects no longer
2116         fit within the allocators cell size, and thus allocation fell back to a slower
2117         allocator.
2118
2119         As a result of this fix I also dramatically cleaned up how runtime objects are
2120         accessed.  The path mostly *removes* code.
2121
2122         Reviewed by Chris.
2123
2124         * khtml/ecma/kjs_dom.cpp:
2125         (DOMDocumentProtoFunc::tryCall):
2126         (DOMElementProtoFunc::tryCall):
2127         (KJS::getRuntimeObject):
2128         * khtml/ecma/kjs_dom.h:
2129         * khtml/ecma/kjs_html.cpp:
2130         (KJS::HTMLDocument::tryGet):
2131         (KJS::HTMLElement::tryGet):
2132         (KJS::HTMLElement::implementsCall):
2133         (KJS::HTMLElement::call):
2134         (KJS::HTMLElement::tryPut):
2135         (KJS::HTMLCollection::tryGet):
2136         (KJS::HTMLCollection::getNamedItems):
2137         * khtml/ecma/kjs_html.h:
2138         * khtml/ecma/kjs_window.cpp:
2139         (Window::get):
2140
2141 2005-01-26  Richard Williamson   <rjw@apple.com>
2142
2143         Fixed <rdar://problem/3757712> REGRESSION (Mail): WebCore does not allow Devanagari ligature input
2144
2145         For now we are using ICU UBreakIterator to determine grapheme boundaries for
2146         cursor and deletion.  This does not match what Cocoa does exactly, but does match
2147         what Carbon does.  The areas are difference are obscure, but, according to
2148         Deborah Goldsmith, using the UBreakIterator is the reasonable approach.
2149
2150         Reviewed by Ken.
2151
2152         * khtml/rendering/render_text.cpp:
2153         (RenderText::previousOffset):
2154         (RenderText::nextOffset):
2155
2156 2005-01-25  David Harrison  <harrison@apple.com>
2157
2158         Reviewed by Maciej and Richard.
2159
2160         <rdar://problem/3963731> AX VO: Changing window sizes causes a crash using with Voice Over - KWQPtrDictImpl::clear
2161
2162         * khtml/html/html_miscimpl.cpp:
2163         (HTMLCollectionImpl::HTMLCollectionImpl):
2164         Ensure that isHTMLDocument before calling HTMLDocument-only method.
2165
2166 2005-01-25  Richard Williamson   <rjw@apple.com>
2167
2168         Part one of fix for <rdar://problem/3757712> REGRESSION (Mail): WebCore does not allow Devanagari ligature input
2169
2170         This patch changes the semantics of next() and previous() on
2171         VisiblePosition to move a grapheme (character cluster) at a
2172         time.  This means that cursor navigation with correctly move
2173         over an entire cluster.
2174
2175         However, the expected behavior for deleting a grapheme is to
2176         delete individual code points, thus decomposing the grapheme
2177         into it constituent parts.  That will be addressed in the next
2178         part of the fix.
2179
2180         Reviewed by Ken.
2181
2182         * khtml/editing/visible_position.cpp:
2183         (khtml::VisiblePosition::previousPosition):
2184         (khtml::VisiblePosition::nextPosition):
2185         * khtml/editing/visible_position.h:
2186         * khtml/rendering/render_object.cpp:
2187         (RenderObject::previousOffset):
2188         (RenderObject::nextOffset):
2189         * khtml/rendering/render_object.h:
2190         * khtml/rendering/render_text.cpp:
2191         (RenderText::previousOffset):
2192         (RenderText::nextOffset):
2193         (RenderText::findNextInlineTextBox):
2194         * khtml/rendering/render_text.h:
2195         * khtml/xml/dom_nodeimpl.cpp:
2196         (NodeImpl::previousOffset):
2197         (NodeImpl::nextOffset):
2198         * khtml/xml/dom_nodeimpl.h:
2199
2200 2005-01-25  David Harrison  <harrison@apple.com>
2201
2202         Reviewed by Maciej.
2203
2204         <rdar://problem/3973067> AXWebArea for http://apple.netscape.com/ has an empty AXLinkUIElements
2205
2206         * kwq/KWQAccObject.mm:
2207         (-[KWQAccObject accessibilityAttributeValue:]):
2208         Omit ignored AXLinks from AXLinkUIElements result
2209
2210 2005-01-25  Ken Kocienda  <kocienda@apple.com>
2211
2212         Reviewed by John
2213
2214         Fix for this bug:
2215
2216         <rdar://problem/3972851> REGRESSION (179-180+): Repro crash in ApplyStyleCommand::cleanUpEmptyStyleSpans
2217
2218         * khtml/editing/htmlediting.cpp:
2219         (khtml::ApplyStyleCommand::cleanUpEmptyStyleSpans): Added two null checks.
2220
2221 2005-01-25  Ken Kocienda  <kocienda@apple.com>
2222
2223         Reviewed by John
2224
2225         Fix for this bug:
2226
2227         <rdar://problem/3954710> Mail crashed while editing signatures - NodeImpl::isBlockFlow
2228
2229         The fix is more general than for this one bug, and may work to fix many crashers. The problem
2230         is that the ReplaceSelectionCommand never checked whether its starting selection is empty. If
2231         it is, then we need to bail before doing the work of the command, which we need to deref the
2232         start and end points of the selection in order to do its work. I think you can see the crash
2233         potential.
2234
2235         * khtml/editing/htmlediting.cpp:
2236         (khtml::ReplaceSelectionCommand::doApply): Assert selection is not empty.
2237         * kwq/WebCoreBridge.mm:
2238         (partHasSelection): New helper function to test that bridge has a part with a selection.
2239
2240         Use new helper function to test part and selection; return from these function if this test fails.
2241
2242         (-[WebCoreBridge rangeByExpandingSelectionWithGranularity:])
2243         (-[WebCoreBridge rangeByAlteringCurrentSelection:direction:granularity:])
2244         (-[WebCoreBridge alterCurrentSelection:direction:granularity:])
2245         (-[WebCoreBridge rangeByAlteringCurrentSelection:verticalDistance:])
2246         (-[WebCoreBridge alterCurrentSelection:verticalDistance:])
2247         (-[WebCoreBridge documentFragmentWithText:])
2248         (-[WebCoreBridge replaceSelectionWithFragment:selectReplacement:smartReplace:])
2249         (-[WebCoreBridge insertLineBreak])
2250         (-[WebCoreBridge insertParagraphSeparator])
2251         (-[WebCoreBridge insertParagraphSeparatorInQuotedContent])
2252         (-[WebCoreBridge insertText:selectInsertedText:])
2253         (-[WebCoreBridge deleteSelectionWithSmartDelete:])
2254         (-[WebCoreBridge ensureSelectionVisible])
2255
2256 2005-01-24  Kevin Decker  <kdecker@apple.com>
2257
2258         Reviewed by Darin.
2259
2260         Fixed <rdar://problem/3932374> REGRESSION: 'ReferenceError - Can't find variable' JavaScript error at webxpress.fidelity.com
2261
2262         * khtml/html/html_miscimpl.h: Missing virtual identifiers for namedItem() and nextNamedItem() exposed a flaw that prevented finding variables in form [HTMLFormCollectionImpl] scopes.
2263
2264
2265 2005-01-24  Vicki Murley <vicki@apple.com>
2266
2267         Reviewed by kocienda.
2268
2269         - fix <rdar://problem/3810661> computed style not handled for Apple extensions
2270
2271         * khtml/css/css_computedstyle.cpp: (DOM::):
2272         (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): add -apple-line-clamp and -apple-text-size-adjust
2273
2274 2005-01-24  Darin Adler  <darin@apple.com>
2275
2276         Reviewed by Kevin.
2277
2278         - fixed <rdar://problem/3969884> REGRESSION (179-180): Typing password not echoed as "bullets" at paypal.com when field is first focused
2279
2280         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]):
2281         Sizing up the form editor, which fixes things for normal fields and search fields,
2282         screws things up for password fields. Eventually, I'll need to figure out why this is
2283         and file a bug so that the AppKit team fixes it. In the mean time, just check for the
2284         case of the secure text field, and don't size up in that case.
2285
2286 2005-01-24  Ken Kocienda  <kocienda@apple.com>
2287
2288         Reviewed by Hyatt
2289
2290         Fix for this bug:
2291
2292         <rdar://problem/3963560> 8A354: Inserting Japanese text to the line head causes line break
2293
2294         * khtml/editing/htmlediting.cpp:
2295         (khtml::ReplaceSelectionCommand::doApply): Add one more case where we merge content into the
2296         existing line. This covers the situation see using the steps to reproduce this bug.
2297
2298 2005-01-21  Maciej Stachowiak  <mjs@apple.com>
2299
2300         Reviewed by Darin.
2301
2302         <rdar://problem/3967572> Editing should split elements before removing style
2303
2304         * khtml/editing/htmlediting.cpp:
2305         (khtml::CompositeEditCommand::splitElement):
2306         (khtml::CompositeEditCommand::mergeIdenticalElements):
2307         (khtml::CompositeEditCommand::wrapContentsInDummySpan):
2308         (khtml::CompositeEditCommand::splitTextNodeContainingElement):
2309         (khtml::ApplyStyleCommand::applyInlineStyle):
2310         (khtml::ApplyStyleCommand::splitTextAtStartIfNeeded):
2311         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded):
2312         (khtml::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
2313         (khtml::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
2314         (khtml::areIdenticalElements):
2315         (khtml::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
2316         (khtml::ApplyStyleCommand::mergeEndWithNextIfIdentical):
2317         (khtml::ApplyStyleCommand::cleanUpEmptyStyleSpans):
2318         (khtml::ApplyStyleCommand::addInlineStyleIfNeeded):
2319         (khtml::InsertParagraphSeparatorCommand::doApply):
2320         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply):
2321         (khtml::InsertTextCommand::prepareForTextInsertion):
2322         (khtml::SplitTextNodeCommand::doUnapply):
2323         (khtml::SplitElementCommand::SplitElementCommand):
2324         (khtml::SplitElementCommand::~SplitElementCommand):
2325         (khtml::SplitElementCommand::doApply):
2326         (khtml::SplitElementCommand::doUnapply):
2327         (khtml::MergeIdenticalElementsCommand::MergeIdenticalElementsCommand):
2328         (khtml::MergeIdenticalElementsCommand::~MergeIdenticalElementsCommand):
2329         (khtml::MergeIdenticalElementsCommand::doApply):
2330         (khtml::MergeIdenticalElementsCommand::doUnapply):
2331         (khtml::WrapContentsInDummySpanCommand::WrapContentsInDummySpanCommand):
2332         (khtml::WrapContentsInDummySpanCommand::~WrapContentsInDummySpanCommand):
2333         (khtml::WrapContentsInDummySpanCommand::doApply):
2334         (khtml::WrapContentsInDummySpanCommand::doUnapply):
2335         (khtml::SplitTextNodeContainingElementCommand::SplitTextNodeContainingElementCommand):
2336         (khtml::SplitTextNodeContainingElementCommand::~SplitTextNodeContainingElementCommand):
2337         (khtml::SplitTextNodeContainingElementCommand::doApply):
2338         * khtml/editing/htmlediting.h:
2339
2340         New layout test that shows the side benefit of this.
2341
2342         * layout-tests/editing/style/unbold-in-bold-expected.txt: Added.
2343         * layout-tests/editing/style/unbold-in-bold.html: Added.
2344
2345 2005-01-24  Darin Adler  <darin@apple.com>
2346
2347         Reviewed by John.
2348
2349         - fixed <rdar://problem/3933435> -[DOMRange cloneContents] sometimes alters the original DOMRange
2350
2351         * khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::processContents): Fixed code that was collapsing
2352         the range after processing the contents. That's right for extract and delete, but not clone.
2353
2354         - fixed <rdar://problem/3714184> inline input works incorrectly when I type Korean slowly in textfield in form on www.google.co.kr
2355         - fixed <rdar://problem/3193848> Can't enter SSN to sign up for Chase visa online account (changing focus inside key press handler)
2356         - 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>
2357
2358         * kwq/KWQLineEdit.mm:
2359         (QLineEdit::text): Changed to call a new string method on the controller instead of calling
2360         stringValue, since stringValue has a bad side effect of ending inline input.
2361         (QLineEdit::selectAll): Changed to only call selectText: when the field is already selected.
2362         When you are giving the field focus, it automatically gets all selected, and this change
2363         is required in the case where we defer the focus change.
2364
2365         * kwq/KWQTextField.h: Added a new string method to the controller.
2366         * kwq/KWQTextField.mm:
2367         (-[KWQTextFieldController setMaximumLength:]): Call the new string method instead of using stringValue.
2368         (-[KWQTextFieldController controlTextDidChange:]): Removed the code to truncate at the first CR or LF.
2369         That's now handled in the shouldChangeTextInRange method instead.
2370         (-[KWQTextFieldController textView:shouldHandleEvent:]): Add new logic to defer responder changes until
2371         after the event is handled. This makes sure the key gets into the field that's already handling it at
2372         rather than ending up in the newly-focused field if the key press handler focuses a different element.
2373         (-[KWQTextFieldController textView:didHandleEvent:]): Turn off deferral here, now that the key press
2374         has been handled.
2375         (-[KWQTextFieldController string]): Added. Calls stringValue only if there's no editor. If there is
2376         an editor, calls string on that instead.
2377         (-[KWQTextFieldController textView:shouldChangeTextInRange:replacementString:]): Added. Called by
2378         the three field subclasses. Truncates incoming strings at the first CR or LF character. This works
2379         properly with Undo, and the old technique did not.
2380         (-[KWQTextFieldController preprocessString:]): Added. Helper used for all the setStringValue methods
2381         so they all truncate at the first CR or LF as well as truncating to the appropriate length.
2382         (-[KWQTextField setStringValue:]): Changed to call the controller.
2383         (-[KWQTextField textView:shouldChangeTextInRange:replacementString:]): Added. Calls the controller.
2384         (-[KWQSecureTextField setStringValue:]): More of the same.
2385         (-[KWQSecureTextField textView:shouldChangeTextInRange:replacementString:]): Ditto.
2386         (-[KWQSearchField setStringValue:]): Ditto.
2387         (-[KWQSearchField textView:shouldChangeTextInRange:replacementString:]): Ditto.
2388
2389         * kwq/KWQWidget.h: Added the new setDeferFirstResponderChanges function.
2390         * kwq/KWQWidget.mm:
2391         (QWidget::~QWidget): Clear out the deferredFirstResponder global just in case we are destroyed
2392         while we "have the ball".
2393         (QWidget::hasFocus): If we are in the deferred mode, use the global to decide who to say is focused.
2394         (QWidget::setFocus): If we are in the deferred mode, set the global instead of making something be
2395         the first responder.
2396         (QWidget::setDeferFirstResponderChanges): Set the boolean. If the boolean is being cleared, then
2397         get the deferred first responder and call setFocus on it again, which will do the work we didn't
2398         want to do earlier.
2399
2400 2005-01-24  Ken Kocienda  <kocienda@apple.com>
2401
2402         Reviewed by John
2403
2404         Fix for this bug:
2405
2406         <rdar://problem/3964350> Deleting all content and typing in a message gets one letter, then beeps
2407
2408         This is a problem with the way we turn key events, when a user starts typing, into editing commands.
2409         Right now, that initial delete key in an empty window is turned into an editing command, and much
2410         editing code runs that need not run. What's more, this is confusing the selection machinery. The
2411         solution is to detect the "delete in an empty window" case and don't turn this into an editing
2412         command at all.
2413
2414         * khtml/editing/htmlediting.cpp:
2415         (khtml::TypingCommand::deleteKeyPressed)
2416
2417 2005-01-23  Adele Amchan  <adele@apple.com>
2418
2419         Fixed <rdar://problem/3964286> REGRESSION (178-179): menus and "go to old site" link at Ameritrade's new site don't work
2420
2421         Removed check for document from checkCompleted.  In this case, WebCore never knew that a WebImageView was complete because
2422         it fell into this document check case in checkCompleted.
2423         Added check for document to tokenizerProcessedData (see <rdar://problem/3807144>)
2424
2425         Reviewed by Darin.
2426
2427         * khtml/khtml_part.cpp:
2428         (KHTMLPart::stop):
2429         (KHTMLPart::checkCompleted):
2430         * khtml/khtml_part.h:
2431         * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::tokenizerProcessedData):
2432
2433 2005-01-21  Richard Williamson   <rjw@apple.com>
2434
2435         Fixed <rdar://problem/3759399> Javascript / Liveconnect problems ((event handler):Undefined value)
2436
2437         Java applets specified with <object> or <embed> weren't scriptable.  Now they are.
2438
2439         Reviewed by Chris.
2440
2441         * khtml/html/html_objectimpl.cpp:
2442         (HTMLEmbedElementImpl::getEmbedInstance):
2443         (HTMLObjectElementImpl::getObjectInstance):
2444
2445 2005-01-21  Ken Kocienda  <kocienda@apple.com>
2446
2447         Reviewed by John
2448
2449         Fix for this bug:
2450
2451         <rdar://problem/3959464> REGRESSION (Mail): Insertion point goes back to beginning of document after deleting
2452
2453         * khtml/editing/htmlediting.cpp:
2454         (khtml::DeleteSelectionCommand::handleGeneralDelete): Add special case to handle retaining a fully-selected block.
2455         This fixes the bug.
2456         * layout-tests/editing/deleting/delete-3959464-fix-expected.txt: Added.
2457         * layout-tests/editing/deleting/delete-3959464-fix.html: Added.
2458
2459 2005-01-21  Richard Williamson   <rjw@apple.com>
2460
2461         Fixed <rdar://problem/3966998> REGRESSION(179-TOT) clicking on gmail message brings me to blank screen
2462
2463         getElementById() sometimes returns Undefined() instead of Null().
2464
2465         Reviewed by Hyatt.
2466
2467         * khtml/ecma/kjs_dom.cpp:
2468         (DOMDocumentProtoFunc::tryCall):
2469
2470 2005-01-21  David Hyatt  <hyatt@apple.com>
2471
2472         Fix for 3773809, make sure that overflow regions never end up at an invalid scroll offset because of a layout change.
2473
2474         Reviewed by kocienda
2475
2476         * khtml/rendering/render_layer.cpp:
2477         (RenderLayer::updateScrollInfoAfterLayout):
2478
2479 2005-01-21  David Hyatt  <hyatt@apple.com>
2480
2481         Fix for 3966349, hang loading page.  Make sure that loops using popOneBlock properly check for a null
2482         blockStack to avoid an infinite loop.
2483
2484         Reviewed by kocienda
2485
2486         * khtml/html/htmlparser.cpp:
2487         (KHTMLParser::parseToken):
2488         (KHTMLParser::insertNode):
2489         (KHTMLParser::popInlineBlocks):
2490
2491 2005-01-21  Ken Kocienda  <kocienda@apple.com>
2492
2493         Reviewed by John
2494
2495         Fix for this bug:
2496
2497         <rdar://problem/3957204> Mail crashed when sending a message in ~InsertParagraphSeparatorInQuotedContentCommand
2498
2499         * khtml/editing/htmlediting.cpp:
2500         (khtml::InsertParagraphSeparatorInQuotedContentCommand::InsertParagraphSeparatorInQuotedContentCommand):
2501         InsertParagraphSeparatorInQuotedContentCommand has this member variable: DOM::ElementImpl *m_breakNode;
2502         I failed to initialize this variable to null in the constructor, and there is an uncommonly-traveled
2503         code path which does not set this variable to something good. In the destructor, we check for null,
2504         and deref if non-null. Obvious "BOOM" potential. Fixed.
2505
2506 2005-01-21  Ken Kocienda  <kocienda@apple.com>
2507
2508         Reviewed by John
2509
2510         Fix for this bug:
2511
2512         <rdar://problem/3966311> REGRESSION (Mail): Hitting return makes space character disappear
2513
2514         * khtml/editing/htmlediting.cpp:
2515         (khtml::InsertParagraphSeparatorCommand::doApply): When writing the code to insert a block
2516         in response to the return key, I did not write code to cover the case described in the bug.
2517         Now I have.
2518         * layout-tests/editing/inserting/insert-div-025-expected.txt: Added.
2519         * layout-tests/editing/inserting/insert-div-025.html: Added.
2520
2521 2005-01-20  Maciej Stachowiak  <mjs@apple.com>
2522
2523         Reviewed by Ken.
2524
2525         <rdar://problem/3965196> security fix for javascript: exploit missed one case (already fixed in updates)
2526
2527         * khtml/ecma/kjs_window.cpp:
2528         (WindowFunc::tryCall): correct mistake in earlier fix for the following bug, caught by Adele:
2529
2530 2005-01-20  Maciej Stachowiak  <mjs@apple.com>
2531
2532         Reviewed by Ken.
2533
2534         <rdar://problem/3965466> editing needs to insert text before applying typing style
2535
2536         * khtml/editing/htmlediting.cpp:
2537         (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): Remove a FIXME comment.
2538         (khtml::InsertLineBreakCommand::doApply): Apply style to the BR
2539         node after inserting it.
2540         (khtml::InsertTextCommand::prepareForTextInsertion): Don't try to
2541         apply style to any new nodes created before inserting.
2542         (khtml::InsertTextCommand::input): Apply style to the inserted range
2543         after doing the text insert.
2544         * khtml/editing/htmlediting.h:
2545
2546         This change leads to improved results on one of the layout tests:
2547
2548         * layout-tests/editing/inserting/insert-div-024-expected.txt:
2549
2550 === Safari-180 ===
2551
2552 2005-01-20  Ken Kocienda  <kocienda@apple.com>
2553
2554         Reviewed by John
2555
2556         Fix for this bug:
2557
2558         <rdar://problem/3964646> REGRESSION (179-180): Typing space at end of line makes following paragraph disappear
2559
2560         * khtml/editing/htmlediting.cpp:
2561         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Recent change to block placeholder removal code
2562         caused this regression. The code became too aggressive in removing block placeholders, and would remove them
2563         from blocks other than the block containing the selection.
2564         * layout-tests/editing/inserting/insert-div-023-expected.txt: This file had a spurious BR element in it that
2565         I did not notice earlier.
2566         * layout-tests/editing/inserting/insert-div-024-expected.txt: Ditto.
2567
2568 2005-01-20  David Hyatt  <hyatt@apple.com>
2569
2570         Fix for oddness on albertsons.com.  Make sure not to crash when setting/removing style properties on a node
2571         with no document.  The bug # is 3813900.
2572
2573         Reviewed by john
2574
2575 2005-01-20  David Harrison  <harrison@apple.com>
2576
2577         Reviewed by Darin.
2578
2579         PARTIAL fix for following bug.  Create attributed string with fonts and links.  Still need to add attachments.
2580         <rdar://problem/3942606> AX: Support kAXAttributedStringForTextMarkerRangeParameterizedAttribute
2581
2582         * kwq/KWQAccObject.mm:
2583         (-[KWQAccObject textUnderElement]):
2584         Touched up previous checkin to match code review comments.
2585
2586         (-[KWQAccObject value]):
2587         Use plainText for this one instead of obsolete attributedString.
2588
2589         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
2590         Advertise AXAttributedStringForTextMarkerRange.
2591
2592         (AXAttributeStringAddFont):
2593         (AXAttributeStringAddElement):
2594         (-[KWQAccObject linkUIElementForNode:]):
2595         (-[KWQAccObject _addAttachmentForElement:URL:needsParagraph:usePlaceholder:]):
2596         (-[KWQAccObject accessibilityAttributedStringForRange:]):
2597         New routines to support AXAttributedStringForTextMarkerRange.
2598
2599         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
2600         Use new routines instead of obsolete attributedString.
2601
2602 2005-01-20  David Harrison  <harrison@apple.com>
2603
2604         Reviewed by Darin.
2605
2606         <rdar://problem/3960196> AX Crash in DOM::Range::setStartBefore
2607
2608         * khtml/xml/dom_docimpl.cpp:
2609         (DocumentImpl::getAccObjectCache):
2610         Adjust when detecting cache in non-top level WebArea.
2611         * kwq/KWQAccObject.mm:
2612         (-[KWQAccObject textUnderElement]):
2613         (-[KWQAccObject accessibilityAttributeValue:]):
2614         Make sure the elements document is the current one for the part.
2615
2616 2005-01-20  Darin Adler  <darin@apple.com>
2617
2618         Reviewed by John.
2619
2620         - fixed <rdar://problem/3922980> Mail not crashing, just quitting itself suddenly in -[WebHTMLView(MailExtras) findString:options:]
2621
2622         * khtml/dom/dom2_range.cpp: (DOM::operator==): Handle null and detached ranges without raising exceptions.
2623         The uncaught exception would make the entire program terminate.
2624
2625         - fixed <rdar://problem/3885729> attempting to modify a computed style does nothing, but should raise a DOM exception
2626
2627         * kwq/DOM-CSS.mm: (-[DOMCSSStyleDeclaration setProperty:::]): Uncommented this code which I had to disable
2628         back in November because Mail was stumbling over it.
2629
2630         - fixed <rdar://problem/3943049> focus() called during onload handler results in square text field on <input type=search>
2631
2632         * kwq/KWQLineEdit.mm: (QLineEdit::baselinePosition): Change computation so it will work even for fields
2633         that position their text in a way that depends on vertical size. In the case of a search field, the text
2634         is centered, so the old logic was broken.
2635         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Set the frame size to
2636         something large enough to accomodate the field editor. If we start the frame at size 0,0 we run into
2637         AppKit trouble when it insets the frame to figure out the frame for the field editor. In the case of this
2638         bug this happens because we become first responder before being sized and positioned by the HTML layout code.
2639
2640         - improved debugging output when using "po" from gdb with Objective-C DOM
2641
2642         * kwq/DOM.mm:
2643         (-[DOMNode description]): Added. Includes node name ("<tr>") and node value (e.g., string for text node).
2644         (-[DOMRange description]): Tweaked format.
2645
2646 2005-01-19  Richard Williamson   <rjw@apple.com>
2647
2648         After further discussion with Real we have decided to NOT include the
2649         additional CLSID for the real plugin.  See 3958601.
2650
2651         Reviewed by Darin.
2652
2653         * khtml/rendering/render_frames.cpp:
2654         (RenderPartObject::updateWidget):
2655
2656 2005-01-19  David Harrison  <harrison@apple.com>
2657
2658         Reviewed by Maciej.
2659
2660         <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
2661
2662         Previous checkin for this bug was missing some of the patch.  Editing snafu with multiple changes in tree.
2663
2664         * kwq/KWQAccObject.mm:
2665         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
2666         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
2667         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
2668         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
2669         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
2670         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
2671         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
2672
2673 2005-01-19  Ken Kocienda  <kocienda@apple.com>
2674
2675         Reviewed by John
2676
2677         * khtml/editing/htmlediting.cpp:
2678         (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Small, cosmetic change John and I decided
2679         to do on my last checkin, but I forgot to do before landing.
2680
2681 2005-01-19  Ken Kocienda  <kocienda@apple.com>
2682
2683         Reviewed by John
2684
2685         Fix for this bug:
2686
2687         <rdar://problem/3959727> REGRESSION (Mail): Style not preserved on blank lines
2688
2689         * khtml/editing/htmlediting.cpp:
2690         (khtml::CompositeEditCommand::applyStyle):
2691         (khtml::CompositeEditCommand::insertBlockPlaceholder): New function that unconditionally adds a block placeholder.
2692         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Now returns bool based on whether
2693         placeholder was added or not.
2694         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Now searches all the descendents of a block
2695         looking for a placeholder. The old code, which just looked at the last child of a node, started missing
2696         once block placeholders became styled (which started happening with this patch).
2697         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Now handles applying typing style
2698         to a block placeholder at call time, rather than setting the typing style as a latent style that
2699         might be applied later. This is an important part of the bug fix.
2700         (khtml::DeleteSelectionCommand::doApply): Now uses bool return value from insertBlockPlaceholderIfNeeded()
2701         and passes it along to calculateStyleBeforeInsertion, so the case where a block placeholder needs to
2702         be styled can be detected.
2703         (khtml::InsertParagraphSeparatorCommand::InsertParagraphSeparatorCommand): Changed the way this class
2704         managed style. Before it would calculate and set typing style for the block added. This is not
2705         sufficient. Added blocks need to styled immediately. Some name changes to instance variables in
2706         this class due to the change to accommodate this change.
2707         (khtml::InsertParagraphSeparatorCommand::~InsertParagraphSeparatorCommand): Name changes, as above.
2708         (khtml::InsertParagraphSeparatorCommand::calculateStyleBeforeInsertion): Ditto.
2709         (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Ditto.
2710         (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
2711         (khtml::ReplaceSelectionCommand::doApply): Improve check for testing when a placeholder
2712         block can be removed in its entirety after the insertion.
2713         * khtml/editing/htmlediting.h: Update header accordingly.
2714         * khtml/khtml_part.cpp:
2715         (KHTMLPart::selectionComputedStyle): Move position for computed style check downstream before
2716         doing check when the position is in an empty block (this makes sure any style on any block
2717         placeholder is accounted for).
2718
2719         New layout tests to check bug fix.
2720
2721         * layout-tests/editing/style/block-style-004-expected.txt: Added.
2722         * layout-tests/editing/style/block-style-004.html: Added.
2723         * layout-tests/editing/style/block-style-005-expected.txt: Added.
2724         * layout-tests/editing/style/block-style-005.html: Added.
2725         * layout-tests/editing/style/block-style-006-expected.txt: Added.
2726         * layout-tests/editing/style/block-style-006.html: Added.
2727
2728         Results updated to reflect new block placeholder code.
2729
2730         * layout-tests/editing/inserting/insert-div-004-expected.txt
2731         * layout-tests/editing/inserting/insert-div-005-expected.txt
2732         * layout-tests/editing/inserting/insert-div-006-expected.txt
2733         * layout-tests/editing/inserting/insert-div-008-expected.txt
2734         * layout-tests/editing/inserting/insert-div-011-expected.txt
2735         * layout-tests/editing/inserting/insert-div-012-expected.txt
2736         * layout-tests/editing/inserting/insert-div-013-expected.txt
2737         * layout-tests/editing/inserting/insert-div-014-expected.txt
2738         * layout-tests/editing/inserting/insert-div-015-expected.txt
2739         * layout-tests/editing/inserting/insert-div-016-expected.txt
2740         * layout-tests/editing/inserting/insert-div-017-expected.txt
2741         * layout-tests/editing/inserting/insert-div-018-expected.txt
2742         * layout-tests/editing/inserting/insert-div-019-expected.txt
2743         * layout-tests/editing/inserting/insert-div-021-expected.txt
2744         * layout-tests/editing/inserting/insert-div-022-expected.txt
2745         * layout-tests/editing/inserting/insert-div-023-expected.txt
2746         * layout-tests/editing/inserting/insert-div-024-expected.txt
2747
2748 2005-01-19  David Hyatt  <hyatt@apple.com>
2749
2750         Dont null-check the renderer before submitting, since a script can set it to display:none and still expect the
2751         submission to occur.  Fixes bug #3477282.
2752
2753         Reviewed by kdecker
2754
2755         (DOM::HTMLInputElementImpl::defaultEventHandler):
2756
2757 2005-01-18  Richard Williamson   <rjw@apple.com>
2758
2759         Fixed several issues all arising from analysis of plugin detection code at ifilm.com:
2760
2761         Fixed <rdar://problem/3958592> can't script plug-ins if plug-in is invoked with <object> element instead of <embed>
2762         Fixed <rdar://problem/3958597> <object> elements with IDs do not show up as named properties of the document
2763         Fixed <rdar://problem/3960973> DOM objects for plugin elements are not accessible
2764         Fixed <rdar://problem/3958601> need an additional class ID in WebCore for the Real plug-in
2765
2766         We now support accessing scriptable plugin objects that are specified with <applet>, <embed>, or <object>
2767         tags.  Also, if any of these elements are named they can be accessed from the document or window objects.
2768         Finally, DOM methods are properties will be forwarded appropriately for the plugin's root scriptable object.
2769
2770         Reviewed by Chris.
2771
2772         * khtml/dom/html_document.cpp:
2773         (HTMLDocument::objects):
2774         * khtml/dom/html_document.h:
2775         * khtml/ecma/kjs_dom.cpp:
2776         (DOMDocumentProtoFunc::tryCall):
2777         (DOMElementProtoFunc::tryCall):
2778         (KJS::getRuntimeObject):
2779         * khtml/ecma/kjs_dom.h:
2780         * khtml/ecma/kjs_html.cpp:
2781         (KJS::HTMLDocument::tryGet):
2782         (KJS::HTMLElement::tryGet):
2783         (KJS::HTMLCollection::tryGet):
2784         (KJS::HTMLCollection::getNamedItems):
2785         * khtml/ecma/kjs_window.cpp:
2786         (Window::get):
2787         * khtml/html/html_miscimpl.cpp:
2788         (HTMLCollectionImpl::traverseNextItem):
2789         * khtml/html/html_miscimpl.h:
2790         (DOM::HTMLCollectionImpl::):
2791         * khtml/html/html_objectimpl.cpp:
2792         (HTMLAppletElementImpl::getAppletInstance):
2793         (HTMLObjectElementImpl::HTMLObjectElementImpl):
2794         (HTMLObjectElementImpl::getObjectInstance):
2795         * khtml/html/html_objectimpl.h:
2796         * khtml/rendering/render_frames.cpp:
2797         (RenderPartObject::updateWidget):
2798         * kwq/KWQKHTMLPart.h:
2799         * kwq/KWQKHTMLPart.mm:
2800         (KWQKHTMLPart::getObjectInstanceForView):
2801
2802 2005-01-18  David Hyatt  <hyatt@apple.com>
2803
2804         Fix for 3948123, rolling over link erases nearby text.  The repaint rect check for lines was wrong whenever
2805         two lines overlapped.
2806
2807         Reviewed by kocienda
2808
2809         * khtml/rendering/render_flow.cpp:
2810         (RenderFlow::paintLines):
2811
2812 2005-01-18  Ken Kocienda  <kocienda@apple.com>
2813
2814         Reviewed by Hyatt
2815
2816         Fix for this bug:
2817
2818         <rdar://problem/3960116> Focus rings paint incorrectly for contenteditable blocks in web pages
2819
2820         * khtml/rendering/render_flow.cpp:
2821         (RenderFlow::addFocusRingRects): Fix painting of focus rings so that ring only paints around
2822         outermost contenteditable elements.
2823
2824 2005-01-18  David Harrison  <harrison@apple.com>
2825
2826         Reviewed by Darin.
2827
2828         <rdar://problem/3959668> accessibilityFocusedUIElement sometimes returns an ignored element; it must not
2829
2830         * kwq/KWQAccObject.mm:
2831         (-[KWQAccObject accessibilityFocusedUIElement]):
2832         Return parentObjectUnignored if focused object is ignored.
2833
2834 2005-01-18  Ken Kocienda  <kocienda@apple.com>
2835
2836         Reviewed by John
2837
2838         Fix for this bug:
2839
2840         <rdar://problem/3952877> REGRESSION (Mail): Command-left/right-arrows don't work with file attachment
2841
2842         * khtml/editing/selection.cpp:
2843         (khtml::nodeForInlineBox): New helper function used in reimplementation of function below.
2844         (khtml::selectionForLine): Reimplemented using line box smarts. I originally wrote this code when
2845         I had a less than full understanding of line layout. I can do better now, and my new version no
2846         longer fails to notice attachments when doing the kind of navigation mentioned in the bug.
2847
2848 2005-01-17  David Harrison  <harrison@apple.com>
2849
2850         Reviewed by John Sullivan.
2851
2852         <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
2853         <rdar://problem/3949848> AX: paragraph marker routines do not work when given a paragraph end marker
2854
2855         Also fixed sentence support in the same way.
2856
2857         * kwq/KWQAccObject.mm:
2858         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
2859         (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
2860         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
2861         (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
2862         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
2863         Move one position in desired direction before calling visible_units.cpp code.
2864
2865 2005-01-17  David Harrison  <harrison@apple.com>
2866
2867         Updated expected test results.
2868
2869         <rdar://problem/3945880> line-ending space seems not to be present
2870
2871         * layout-tests/editing/deleting/delete-ws-fixup-001-expected.txt:
2872         * layout-tests/editing/deleting/delete-ws-fixup-002-expected.txt:
2873         * layout-tests/editing/inserting/insert-div-020-expected.txt:
2874         * layout-tests/editing/inserting/insert-div-021-expected.txt:
2875         * layout-tests/editing/inserting/insert-div-022-expected.txt:
2876         * layout-tests/editing/inserting/insert-div-023-expected.txt:
2877         * layout-tests/editing/inserting/insert-div-024-expected.txt:
2878
2879 2005-01-17  David Harrison  <harrison@apple.com>
2880
2881         Reviewed by Dave Hyatt (bidi.cpp) and Darin Adler (selection.cpp).
2882
2883         <rdar://problem/3945880> line-ending space seems not to be present
2884
2885         * khtml/editing/selection.cpp:
2886         (khtml::Selection::validate):
2887         Extend AFTER_WHITE_SPACE code to support white spac in the middle of paragraphs, not just the end.
2888         * khtml/rendering/bidi.cpp:
2889         (khtml::RenderBlock::findNextLineBreak):
2890         Pick left/rightness of word selection based on being at the end of paragraph (i.e. after a hard line break).
2891
2892 2005-01-17  Darin Adler  <darin@apple.com>
2893
2894         Reviewed by John Louch.
2895
2896         - fixed <rdar://problem/3958503> need screenX and screenY to use WebKit windowFrame delegate
2897
2898         * khtml/ecma/kjs_window.cpp: (Window::get): Change screenX and screenY to use frameGeometry instead
2899         of using mapToGlobal and screen in a complicated way.
2900         * kwq/KWQKHTMLView.mm: Removed unused mapToGlobal function.
2901         * kwq/KWQWindowWidget.h: Ditto.
2902         * kwq/KWQWindowWidget.mm: Ditto.
2903
2904 2005-01-17  David Hyatt  <hyatt@apple.com>
2905
2906         Fix a screwup in rightmost/lowets position computation.  3955207.  Make sure floats with layers are still checked.
2907
2908         Reviewed by kocienda
2909
2910         * khtml/rendering/render_block.cpp
2911
2912 2005-01-17  Ken Kocienda  <kocienda@apple.com>
2913
2914         Reviewed by John
2915
2916         <rdar://problem/3953366> Problems with typing attributes in HTML compose
2917
2918         * khtml/editing/htmlediting.cpp:
2919         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Remove misguided "optimization"
2920         that tried to sense when typing style could be cleared without actually doing a style diff between
2921         before-delete and after-delete positions. Removing this extra check and running the
2922         general-purpose code fixes the bug.
2923
2924 2005-01-17  Richard Williamson   <rjw@apple.com>
2925
2926         Fixed <rdar://problem/3753030> Need to ensure same origin for plugin binding invocations (origin security rules)
2927
2928         Keep track of originating execution context and target execution
2929         context for native JS object wrappers, and perform appropriate
2930         security checks.
2931
2932         Reviewed by David Harrison.
2933
2934         * khtml/ecma/kjs_binding.cpp:
2935         (ScriptInterpreter::isGlobalObject):
2936         (ScriptInterpreter::isSafeScript):
2937         (ScriptInterpreter::interpreterForGlobalObject):
2938         * khtml/ecma/kjs_binding.h:
2939         * khtml/ecma/kjs_window.cpp:
2940         (Window::interpreter):
2941         (Window::isSafeScript):
2942         * khtml/ecma/kjs_window.h:
2943         * kwq/DOMInternal.mm:
2944         (-[WebScriptObject _initializeScriptDOMNodeImp]):
2945         * kwq/KWQKHTMLPart.mm:
2946         (KWQKHTMLPart::windowScriptObject):
2947         (KWQKHTMLPart::windowScriptNPObject):
2948         * kwq/WebCoreBridge.mm:
2949         (rootForView):
2950         (-[WebCoreBridge init]):
2951
2952 2005-01-17  Ken Kocienda  <kocienda@apple.com>
2953
2954         Reviewed by John
2955
2956         Fix for this bug:
2957
2958         <rdar://problem/3786659> REGRESSION (Mail): editable WebViews don't work with
2959         "size up" and "size down" NSFontManager changes
2960
2961         * khtml/css/cssparser.cpp:
2962         (CSSParser::parseValue): Add support for parsing new font size delta property.
2963         * khtml/css/cssproperties.c: Generated file.
2964         * khtml/css/cssproperties.h: Ditto.
2965         * khtml/css/cssproperties.in: Add support for parsing new font size delta property.
2966         * khtml/editing/htmlediting.cpp:
2967         (khtml::isEmptyStyleSpan): New helper function. Code existed before, but now factored out for convenient use.
2968         (khtml::CompositeEditCommand::removeNodeAttribute): Does check on value to see it exists before creating
2969         and running command to do the removal.
2970         (khtml::ApplyStyleCommand::doApply): Now calls new applyRelativeFontStyleChange() function as part of its work.
2971         (khtml::ApplyStyleCommand::applyRelativeFontStyleChange): New function that does the "heavy lifting" to handle
2972         relative font size changes.
2973         (khtml::ApplyStyleCommand::applyInlineStyle): Range check the start and end positions to make sure the start is
2974         before or equal to the end. Swap them if not true. I ran into this problem in some similar code in
2975         applyRelativeFontStyleChange(). Moving that goodness here too.
2976         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded): Uses a local variable to save a value used often.
2977         Code is the same, but reads better now. Function now returns bool as well, just like splitTextAtStartIfNeeded.
2978         I use the bool return value now (I obviously did not need it before).
2979         (khtml::ApplyStyleCommand::computedFontSize): New helper function.
2980         (khtml::ApplyStyleCommand::joinChildTextNodes): Ditto.
2981         (khtml::createStyleSpanElement): Ditto.
2982         * khtml/editing/htmlediting.h: Update header accordingly.
2983         * khtml/editing/jsediting.cpp: Add new command string to enable relative font size changes.
2984         * kwq/DOM-CSS.mm:
2985         (-[DOMCSSStyleDeclaration _fontSizeDelta]): New convenience.
2986         (-[DOMCSSStyleDeclaration _setFontSizeDelta:]): Ditto.
2987         * kwq/DOMPrivate.h: Declare new conveniences.
2988         * layout-tests/editing/editing.js: Add new JS to enable relative font size changes, as well as explicit font size setting.
2989         * layout-tests/editing/style/relative-font-size-change-001-expected.txt: Added.
2990         * layout-tests/editing/style/relative-font-size-change-001.html: Added.
2991         * layout-tests/editing/style/relative-font-size-change-002-expected.txt: Added.
2992         * layout-tests/editing/style/relative-font-size-change-002.html: Added.
2993         * layout-tests/editing/style/relative-font-size-change-003-expected.txt: Added.
2994         * layout-tests/editing/style/relative-font-size-change-003.html: Added.
2995         * layout-tests/editing/style/relative-font-size-change-004-expected.txt: Added.
2996         * layout-tests/editing/style/relative-font-size-change-004.html: Added.
2997
2998 2005-01-14  Darin Adler  <darin@apple.com>
2999
3000         * khtml/css/cssproperties.c: Regenerated with gperf 3.0.1.
3001         * khtml/css/cssvalues.c: Regenerated with gperf 3.0.1.
3002         * khtml/misc/htmlattrs.c: Regenerated with gperf 3.0.1.
3003         * khtml/misc/htmltags.c: Regenerated with gperf 3.0.1.
3004
3005 2005-01-14  Chris Blumenberg  <cblu@apple.com>
3006
3007         Fixed: <rdar://problem/3886415> arrow keys don't work when the user hits Back to return to RSS page
3008
3009         Reviewed by hyatt.
3010
3011         (KWQKHTMLPart::openURLFromPageCache): restore mousePressNode
3012         (KWQKHTMLPart::mousePressNode): new
3013         * kwq/KWQPageState.h:
3014         * kwq/KWQPageState.mm:
3015         (-[KWQPageState initWithDocument:URL:windowProperties:locationProperties:interpreterBuiltins:]): save the mousePressNode
3016         (-[KWQPageState clear]): clear the mousePressNode
3017         (-[KWQPageState dealloc]): deref the mousePressNode
3018         (-[KWQPageState mousePressNode]): new
3019
3020 === Safari-179 ===
3021
3022 2005-01-13  Vicki Murley <vicki@apple.com>
3023
3024         Reviewed by Adele.
3025
3026         - fix <rdar://problem/3946836> Safari about box lists 2004 instead of 2005
3027
3028         * WebCore.pbproj/project.pbxproj: bump "2004" to "2005"
3029
3030 2005-01-13  David Harrison  <harrison@apple.com>
3031
3032         Reviewed by Ken Kocienda.
3033
3034         Better fix for 3905066.
3035
3036         * khtml/editing/htmlediting.cpp:
3037         (khtml::InsertParagraphSeparatorCommand::doApply):
3038
3039 2005-01-10  Maciej Stachowiak  <mjs@apple.com>
3040
3041         Reviewed by Darin.
3042
3043         <rdar://problem/3758033> REGRESSION (Mail): Support attributes in marked text (International input)
3044
3045         * khtml/rendering/render_text.cpp:
3046         (InlineTextBox::paint): Support painting custom underline markers for
3047         marked text in place of generic yellow.
3048         (InlineTextBox::paintMarkedTextUnderline): New method that handles this.
3049         * khtml/rendering/render_text.h:
3050         * kwq/KWQKHTMLPart.h: Declare new methods and structs.
3051         * kwq/KWQKHTMLPart.mm:
3052         (KWQKHTMLPart::clear): Clear marked test underlines.
3053         (KWQKHTMLPart::setMarkedTextRange): Takes attributes and ranges now.
3054         (convertAttributesToUnderlines): Converts NSAttributedString attributes
3055         to simplified and C++-friendly form.
3056         (KWQKHTMLPart::markedTextUsesUnderlines): New method.
3057         (KWQKHTMLPart::markedTextUnderlines): New method.
3058         * kwq/KWQPainter.mm:
3059         (QPainter::drawLineForText): Handle pen width.
3060         * kwq/WebCoreBridge.h:
3061         * kwq/WebCoreBridge.mm:
3062         (-[WebCoreBridge setMarkedTextDOMRange:customAttributes:ranges:]): Take attributes
3063         and ranges.
3064         * kwq/WebCoreTextRenderer.h:
3065
3066 2005-01-12  David Harrison  <harrison@apple.com>
3067
3068         Reviewed by Dave Hyatt.
3069
3070         <rdar://problem/3888973> AX: Parent AXWebArea of nested AXWebAreas is messed up
3071
3072         * kwq/KWQAccObject.mm:
3073         (-[KWQAccObject addChildrenToArray:]):
3074         Use the widget's outer view.
3075
3076 2005-01-12  David Harrison  <harrison@apple.com>
3077
3078         Reviewed by Darin Adler.
3079
3080         <rdar://problem/3949908> Crash when asking for the kAXLengthForTextMarkerRangeParameterizedAttribute
3081
3082         * khtml/xml/dom_nodeimpl.cpp:
3083         (NodeImpl::displayNode):
3084         Add quotes around text node content.
3085         (NodeBaseImpl::childNode):
3086         Add nil check to return nil rather than crash when child node not found.
3087         * kwq/KWQAccObject.mm:
3088         (-[KWQAccObject doAXStringForTextMarkerRange:]):
3089         Pass range compliant positions to TextIterator.
3090
3091 2005-01-12  David Hyatt  <hyatt@apple.com>
3092
3093         Fix for 3951203, CSS border style incorrectly clipped on inline elements.  Make sure the repaint rect set during
3094         line layout is smarter about including the overflow for both old states and new states.
3095
3096         Reviewed by john
3097
3098         * khtml/rendering/bidi.cpp:
3099         (khtml::RenderBlock::layoutInlineChildren):
3100
3101 2005-01-12  David Hyatt  <hyatt@apple.com>
3102
3103         Fix for bug 3937608, versiontracker.com flashes and displays the right column below content.  Make sure to move
3104         tables/overflows that dont fit within a block only in strict mode.
3105
3106         Fix for bug 3931049, characters dont show up when typing.  Make sure to dirty the right lines when this specific
3107         case in editing is hit.
3108
3109         Reviewed by darin (first one), kocienda (second one)
3110
3111         * khtml/rendering/render_block.cpp:
3112         (khtml::RenderBlock::getClearDelta):
3113         * khtml/rendering/render_flow.cpp:
3114         (RenderFlow::dirtyLinesFromChangedChild):
3115         * khtml/rendering/render_line.cpp:
3116         (khtml::InlineFlowBox::verticallyAlignBoxes):
3117
3118 2005-01-12  Ken Kocienda  <kocienda@apple.com>
3119
3120         Reviewed by John
3121
3122         Fix for this bug:
3123
3124         <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
3125
3126         * khtml/editing/htmlediting.cpp:
3127         (khtml::DeleteSelectionCommand::initializePositionData): Change test that will prevent block
3128         merging. End-of-line test was just wrong. Call new start-of and end-of paragraph functions
3129         added to visible_position files.
3130         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Add tests for BR elements, which makes
3131         tests to determine when to stop moving nodes more complete and correct. Also improved comments.
3132         * khtml/editing/visible_position.cpp:
3133         (khtml::isFirstVisiblePositionInParagraph): New function.
3134         (khtml::isLastVisiblePositionInParagraph): New function.
3135         * khtml/editing/visible_position.h: Update header accordingly.
3136         * layout-tests/editing/deleting/delete-block-merge-contents-002-expected.txt: New results, slightly different from former
3137         results but still correct.
3138         * layout-tests/editing/deleting/delete-block-merge-contents-007-expected.txt: Ditto.
3139         * layout-tests/editing/deleting/delete-block-merge-contents-018-expected.txt: Added.
3140         * layout-tests/editing/deleting/delete-block-merge-contents-018.html: Added.
3141         * layout-tests/editing/deleting/delete-block-merge-contents-019-expected.txt: Added.
3142         * layout-tests/editing/deleting/delete-block-merge-contents-019.html: Added.
3143         * layout-tests/editing/deleting/delete-block-merge-contents-020-expected.txt: Added.
3144         * layout-tests/editing/deleting/delete-block-merge-contents-020.html: Added.
3145         * layout-tests/editing/deleting/delete-block-merge-contents-021-expected.txt: Added.
3146         * layout-tests/editing/deleting/delete-block-merge-contents-021.html: Added.
3147         * layout-tests/editing/deleting/delete-line-014-expected.txt: These new results are actually better, and fix a bug.
3148
3149 2005-01-11  Richard Williamson   <rjw@apple.com>
3150
3151         Fixed 3922875.  Fall back to DOM object is EMBED element
3152         has no associated runtime object.
3153
3154         Reviewed by Chris.
3155
3156         * khtml/ecma/kjs_dom.cpp:
3157         (KJS::getRuntimeObject):
3158         * khtml/ecma/kjs_html.cpp:
3159         (KJS::HTMLDocument::tryGet):
3160         (KJS::HTMLElement::tryGet):
3161         (KJS::HTMLCollection::tryGet):
3162         (KJS::HTMLCollection::getNamedItems):
3163
3164 2005-01-11  David Hyatt  <hyatt@apple.com>
3165
3166         Fix for 3882299, missing content on gibson.com.  Change our handling of " and ' in certain states of the parser to match
3167         other browsers.
3168
3169         Reviewed by Maciej
3170
3171         * khtml/html/htmltokenizer.cpp:
3172         (khtml::HTMLTokenizer::parseTag):
3173
3174 2005-01-11  Chris Blumenberg  <cblu@apple.com>
3175
3176         Fixed: <rdar://problem/3930733> Mail prints second page of email blank
3177
3178         Reviewed by dave.
3179
3180         * khtml/rendering/render_canvas.cpp:
3181         (RenderCanvas::paint): cache the print rect since the dirty rect can get changed during printing
3182         * khtml/rendering/render_flow.cpp:
3183         (RenderFlow::paintLines): removed null check since the print rect should never be null
3184         * khtml/rendering/render_list.cpp:
3185         (RenderListMarker::paint): ditto
3186         * kwq/KWQKHTMLPart.mm:
3187         (KWQKHTMLPart::adjustPageHeight): don't set the print rect here since this method is not called for the last page
3188
3189 2005-01-10  Ken Kocienda  <kocienda@apple.com>
3190
3191         Reviewed by Hyatt
3192
3193         Fix for this bug:
3194
3195         <rdar://problem/3943648> extra line is inserted after pressing return within quoted text of reply
3196
3197         * khtml/editing/htmlediting.cpp:
3198         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): This bug was the result of a
3199         simple coding mistake. A local variable was erroneously redefined in a deeper scope, and so the
3200         result calculated in that deeper scope was not available when tested.
3201
3202 2005-01-10  Ken Kocienda  <kocienda@apple.com>
3203
3204         Reviewed by John
3205
3206         Fix for this bug:
3207
3208         <rdar://problem/3946852> Option-e goes to next line
3209
3210         * khtml/editing/htmlediting.cpp:
3211         (khtml::ReplaceSelectionCommand::doApply): Tweak case used to determine when merging content
3212         into the start line is done. We plan to change pretty substantially soon to better handle
3213         the problem described in <rdar://problem/3937352> Quote level not maintained when copied
3214         and pasted within a Mail message. In the meantime, this change does no harm, and fixes the bug.
3215
3216 2005-01-10  Ken Kocienda  <kocienda@apple.com>
3217
3218         Reviewed by Darin
3219
3220         Fix for this bug:
3221
3222         <rdar://problem/3907005> Applying block styles to a line of text can unexpectedly affect other lines
3223
3224         * khtml/editing/htmlediting.cpp:
3225         (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): New function which moves "paragraphs"
3226         to their own blocks if needed so that a block style can be applied.
3227         (khtml::CompositeEditCommand::isMailBlockquote): Moved this function to base class so it can be used more generally.
3228         (khtml::ApplyStyleCommand::applyBlockStyle): Pass a node instead of a block to addBlockStyleIfNeeded().
3229         (khtml::ApplyStyleCommand::addBlockStyleIfNeeded): Now accepts a node instead of a block for styling.
3230         This function also now calls moveParagraphContentsToNewBlockIfNecessary() to make sure that only the current
3231         paragraph is styled.
3232         (khtml::SetNodeAttributeCommand::doUnapply): Should not assert on undo if old value of attribute was null.
3233         Should remove attributue instead.
3234         * khtml/editing/htmlediting.h: Touch function declarations accordingly.
3235         * layout-tests/editing/style/create-block-for-style-001-expected.txt: Added.
3236         * layout-tests/editing/style/create-block-for-style-001.html: Added.
3237         * layout-tests/editing/style/create-block-for-style-002-expected.txt: Added.
3238         * layout-tests/editing/style/create-block-for-style-002.html: Added.
3239         * layout-tests/editing/style/create-block-for-style-003-expected.txt: Added.
3240         * layout-tests/editing/style/create-block-for-style-003.html: Added.
3241         * layout-tests/editing/style/create-block-for-style-004-expected.txt: Added.
3242         * layout-tests/editing/style/create-block-for-style-004.html: Added.
3243         * layout-tests/editing/style/create-block-for-style-005-expected.txt: Added.
3244         * layout-tests/editing/style/create-block-for-style-005.html: Added.
3245         * layout-tests/editing/style/create-block-for-style-006-expected.txt: Added.
3246         * layout-tests/editing/style/create-block-for-style-006.html: Added.
3247         * layout-tests/editing/style/create-block-for-style-007-expected.txt: Added.
3248         * layout-tests/editing/style/create-block-for-style-007.html: Added.
3249         * layout-tests/editing/style/create-block-for-style-008-expected.txt: Added.
3250         * layout-tests/editing/style/create-block-for-style-008.html: Added.
3251         * layout-tests/editing/style/create-block-for-style-009-expected.txt: Added.
3252         * layout-tests/editing/style/create-block-for-style-009.html: Added.
3253         * layout-tests/editing/style/create-block-for-style-010-expected.txt: Added.
3254         * layout-tests/editing/style/create-block-for-style-010.html: Added.
3255         * layout-tests/editing/style/create-block-for-style-011-expected.txt: Added.
3256         * layout-tests/editing/style/create-block-for-style-011.html: Added.
3257         * layout-tests/editing/style/create-block-for-style-012-expected.txt: Added.
3258         * layout-tests/editing/style/create-block-for-style-012.html: Added.
3259         * layout-tests/editing/style/create-block-for-style-013-expected.txt: Added.
3260         * layout-tests/editing/style/create-block-for-style-013.html: Added.
3261
3262         Unrelated updates to these expected results.
3263         * layout-tests/editing/inserting/insert-div-007-expected.txt
3264         * layout-tests/editing/pasteboard/paste-text-013-expected.txt
3265
3266 2005-01-10  Chris Blumenberg  <cblu@apple.com>
3267
3268         Fixed: <rdar://problem/3936844> Mail: Messages with rich text do not print
3269
3270         Reviewed by john.
3271
3272         * khtml/rendering/render_flow.cpp:
3273         (RenderFlow::paintLines): don't do pagination work if printRect is not set
3274         * khtml/rendering/render_list.cpp:
3275         (RenderListMarker::paint): ditto
3276
3277 2005-01-10  David Harrison  <harrison@apple.com>
3278
3279         Reviewed by Darin.
3280
3281         * kwq/KWQTextUtilities.mm:
3282         (currentTextBreakLocaleID):
3283         Return empty string (AKA root locale) if locale pref can not be canonicalized.
3284
3285 2005-01-10  John Sullivan  <sullivan@apple.com>
3286
3287         Fixed broken Panther build.
3288
3289         * kwq/KWQTextUtilities.mm:
3290         (currentTextBreakLocaleID):
3291         This recently-added function was using code copy/pasted from CarbonCore UnicodeUtilities.
3292         That code used a Tiger-only function, CFLocaleCreateCanonicalLanguageIdentifierFromString.
3293         To fix the build, I added a BUILDING_ON_PANTHER #ifdef that avoids the call to the Tiger
3294         function. However, the Tiger-only code was wrong; the string generated using
3295         CFLocaleCreateCanonicalLanguageIdentifierFromString was not used at all, so I fixed that
3296         as well.
3297
3298 2005-01-09  David Harrison  <harrison@apple.com>
3299
3300         Reviewed by Ken Kocienda.
3301
3302         <rdar://problem/3786362> REGRESSION (Mail): pasted text loses one newline
3303
3304         * khtml/editing/htmlediting.cpp:
3305         (khtml::ReplaceSelectionCommand::doApply):
3306         Tune check for need for insertParagraphSeparator when hasInterchangeNewline.
3307
3308 2005-01-09  Darin Adler  <darin@apple.com>
3309
3310         Reviewed by Harrison.
3311
3312         - fixed <rdar://problem/3939176> select() method does not work on <input type=search>
3313
3314         * khtml/html/html_formimpl.cpp: (DOM::HTMLInputElementImpl::select): Changed if statement
3315         to a switch statement. Added SEARCH to the set of types that treat the renderer as a
3316         RenderLineEdit.
3317
3318 2005-01-09  David Harrison  <harrison@apple.com>
3319
3320         Reviewed by Ken Kocienda.
3321
3322         <rdar://problem/3905066> REGRESSION (Mail): Hitting return key with caret in front of space causes space to become lost (resize fixes)
3323
3324         Problem was that InsertParagraphSeparatorCommand::doApply() was not collapsing
3325         whitespace to a single non-breaking space when splitting a text node.
3326
3327         * khtml/editing/htmlediting.cpp:
3328         (khtml::InsertParagraphSeparatorCommand::doApply):
3329
3330 2005-01-08  Kevin Decker  <kdecker@apple.com>
3331
3332         Reviewed by Ken.
3333
3334         Fixed: <rdar://problem/3924219> Calling setOuterHTML: on a DOMHTMLHtmlElement can crash a program
3335
3336         * khtml/html/html_elementimpl.cpp:
3337         (HTMLElementImpl::setOuterHTML): Added a nil check for fragments who don't have parents.
3338
3339 2005-01-07  Maciej Stachowiak  <mjs@apple.com>
3340
3341         Reviewed by Darin.
3342
3343         <rdar://problem/3807144> REGRESSION (125-TOT): my bank's website doesn't work (Societe Generale, socgen.com)
3344
3345         * Khtml/khtml_part.cpp:
3346         (KHTMLPart::checkCompleted): If the document is NULL, assume this frame has
3347         not started loading yet so it could not possibly be finishing here...
3348         (KHTMLPart::stop): ...unless the part is explicitly stopped and there is still
3349         no document, in this case we must have hit an error or been loading a non-HTML
3350         frame.
3351         * khtml/khtml_part.h:
3352
3353 2005-01-08  David Harrison  <harrison@apple.com>
3354
3355         Reviewed by Maciej.
3356
3357         <rdar://problem/3943415> REGRESSION (Mail): double-clicking past word selects previous word instead of only blank space
3358
3359         * khtml/editing/selection.cpp:
3360         (khtml::Selection::validate):
3361         Tune word selections left/right choice to use right if on empty last line.
3362
3363 2005-01-07  David Harrison  <harrison@apple.com>
3364
3365         Reviewed by Darin.
3366
3367         <rdar://problem/3942619> AX: Support sentence ax attributes
3368
3369         Needed to use the unicode utilities properly.  Twas lame before.
3370
3371         * khtml/editing/visible_units.cpp:
3372         (khtml::previousBoundary):
3373         (khtml::nextBoundary):
3374         (khtml::startOfWord):
3375         (khtml::endOfWord):
3376         (khtml::previousWordPosition):
3377         (khtml::nextWordPosition):
3378         (khtml::startOfSentence):
3379         (khtml::endOfSentence):
3380         (khtml::previousSentencePosition):
3381         (khtml::nextSentencePosition):
3382         * kwq/KWQAccObject.mm:
3383         (-[KWQAccObject accessibilityAttributeNames]):
3384         (-[KWQAccObject accessibilityAttributeValue:]):
3385         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
3386         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
3387         * kwq/KWQTextUtilities.mm:
3388         (currentTextBreakLocaleID):
3389         (KWQFindSentenceBoundary):
3390         (KWQFindNextSentenceFromIndex):
3391
3392 2005-01-07  Ken Kocienda  <kocienda@apple.com>
3393
3394         Reviewed by Kevin
3395
3396         Fix for these bugs:
3397
3398         <rdar://problem/3939523> in some cases, text does not retain style info after pressing return twice
3399         <rdar://problem/3944492> after pressing return twice, text is bold when it shouldn't be
3400
3401         * khtml/editing/htmlediting.cpp:
3402         (khtml::InsertParagraphSeparatorCommand::setFullTypingStyleBeforeInsertion):
3403         Merge the typing style with the computed style for the current position. Fixes both bugs.
3404         * khtml/editing/htmlediting.h:
3405         * layout-tests/editing/inserting/insert-div-023-expected.txt: Added.
3406         * layout-tests/editing/inserting/insert-div-023.html: Added.
3407         * layout-tests/editing/inserting/insert-div-024-expected.txt: Added.
3408         * layout-tests/editing/inserting/insert-div-024.html: Added.
3409
3410 2005-01-07  David Hyatt  <hyatt@apple.com>
3411
3412         Fix for 3941364, make sure tables reset overflowWidth/Height when they lay out again.  Fixes the odd scrolling
3413         behavior on worldofwarcraft.com.
3414
3415         Reviewed by kevin
3416
3417         * khtml/rendering/render_table.cpp:
3418         (RenderTable::layout):
3419
3420 2005-01-06  David Hyatt  <hyatt@apple.com>
3421
3422         Fix for 3932418 and 3920998, assertion failures caused by stray inline content inside tables.  Bulletproof the hit testing to
3423         ignore inline flows in this case.
3424
3425         Reviewed by kevin
3426
3427         * khtml/rendering/render_block.cpp:
3428         (khtml::RenderBlock::nodeAtPoint):
3429
3430 2005-01-07  Ken Kocienda  <kocienda@apple.com>
3431
3432         Reviewed by Hyatt
3433
3434         Fix for this bug:
3435
3436         <rdar://problem/3848412> for forwarded message, tabbing to message view scrolls to bottom of view
3437
3438         * kwq/KWQKHTMLPart.mm:
3439         (KWQKHTMLPart::nextKeyViewInFrame): Don't scroll the focus node to visible if it is not
3440         in the document, or if it is not a descendent of the document element. In the case of the bug,
3441         since the selection has not yet been set up, the focus node passed here is the HTML element, and
3442         that does not yield a rect that is useful to us here. So now, in the case the bug mentions, we do nothing.
3443
3444 2005-01-06  Kevin Decker  <kdecker@apple.com>
3445
3446         Reviewed by mjs.
3447
3448         Fixed: <rdar://problem/3932215> REGRESSION (125-177): iFrame example at developer.apple.com fails in Safari
3449
3450         * khtml/khtml_part.cpp:
3451         (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.
3452
3453 === Safari-178 ===
3454
3455 2005-01-06  David Harrison  <harrison@apple.com>
3456
3457         Reviewed by Chris.
3458
3459         (addendum to previous checkin for this bug)
3460         <rdar://problem/3776056> AX: Editable HTML should not be viewed as AXStaticText
3461
3462         Fix line navigation.  Add AXUIElementForTextMarker.
3463
3464         * kwq/KWQAccObject.mm:
3465         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
3466         (-[KWQAccObject doAXUIElementForTextMarker:]):
3467         (-[KWQAccObject doAXLineForTextMarker:]):
3468         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
3469         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
3470         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
3471         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
3472         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
3473
3474 2005-01-06  Darin Adler  <darin@apple.com>
3475
3476         Reviewed by Ken.
3477
3478         - re-fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
3479
3480         (turns out the PLT regression was a false alarm)
3481
3482         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
3483         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
3484         much larger number.
3485         * khtml/html/htmlparser.cpp:
3486         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
3487         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
3488         to eliminate code that used ID_CLOSE_TAG for an array size.
3489         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
3490         that manages isindex to use deref instead of delete.
3491         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
3492         mistake of using ID_CLOSE_TAG for the array size too.
3493         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
3494         there and it would prevent custom tags from working. Added range check before using the forbidden
3495         tag array with the token ID since custom tags will use index values past the end of the array.
3496         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
3497         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
3498         createElement call is still here. Last time I left out a few form element types from this switch;
3499         fixed now.
3500         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
3501         document if getTagID fails; this creates a unique per-document ID.
3502
3503         * khtml/misc/htmltags.c: Regenerated.
3504         * khtml/misc/htmltags.h: Regenerated.
3505
3506         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
3507         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
3508         Also rewrote getTagName to work with the new scheme.
3509
3510 2005-01-06  David Harrison  <harrison@apple.com>
3511
3512         Fixed Panther build.  Also, do not advertize sentence support since it is incomplete.
3513
3514         * kwq/KWQAccObject.mm:
3515         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
3516         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
3517         (-[KWQAccObject accessibilityIsAttributeSettable:]):
3518
3519 2005-01-06  David Harrison  <harrison@apple.com>
3520
3521         Reviewed by Dave Hyatt.
3522
3523         <rdar://problem/3776056> AX: Editable HTML should not be viewed as AXStaticText
3524         <rdar://problem/3588548> AX: tabbing does not work correctly with the screen reader and a focused link; need AXFocusedUIElement to work
3525
3526         Many more AX attributes supported.  Numerous fixes to previous AX work.
3527
3528         * khtml/editing/visible_units.cpp:
3529         (khtml::startSentenceBoundary):
3530         (khtml::startOfSentence):
3531         (khtml::endSentenceBoundary):
3532         (khtml::endOfSentence):
3533         (khtml::previousSentencePositionBoundary):
3534         (khtml::previousSentencePosition):
3535         (khtml::nextSentencePositionBoundary):
3536         (khtml::nextSentencePosition):
3537         * khtml/editing/visible_units.h:
3538         * khtml/khtmlview.cpp:
3539         (KHTMLView::layout):
3540         * khtml/misc/helper.cpp:
3541         (khtml::findSentenceBoundary):
3542         (khtml::nextSentenceFromIndex):
3543         * khtml/misc/helper.h:
3544         * khtml/misc/htmltags.c:
3545         (hash_tag):
3546         (findTag):
3547         * khtml/rendering/render_container.cpp:
3548         (RenderContainer::removeChildNode):
3549         (RenderContainer::appendChildNode):
3550         (RenderContainer::insertChildNode):
3551         * khtml/rendering/render_object.cpp:
3552         (RenderObject::remove):
3553         * khtml/xml/dom_docimpl.cpp:
3554         (DocumentImpl::getAccObjectCache):
3555         (DocumentImpl::updateSelection):
3556         (DocumentImpl::close):
3557         (DocumentImpl::setFocusNode):
3558         (DocumentImpl::parentDocument):
3559         (DocumentImpl::topDocument):
3560         * khtml/xml/dom_docimpl.h:
3561         * kwq/KWQAccObject.mm:
3562         (-[KWQAccObject accessibilityShouldUseUniqueId]):
3563         (-[KWQAccObject detach]):
3564         (-[KWQAccObject anchorElement]):
3565         (-[KWQAccObject firstChild]):
3566         (-[KWQAccObject lastChild]):
3567         (-[KWQAccObject previousSibling]):
3568         (-[KWQAccObject nextSibling]):
3569         (-[KWQAccObject parentObject]):
3570         (-[KWQAccObject value]):
3571         (-[KWQAccObject accessibilityAttributeNames]):
3572         (-[KWQAccObject accessibilityPerformAction:]):
3573         (-[KWQAccObject textMarkerForVisiblePosition:]):
3574         (-[KWQAccObject visiblePositionForTextMarker:]):
3575         (-[KWQAccObject textMarkerRangeFromVisiblePositions:andEndPos:]):
3576         (-[KWQAccObject topDocument]):
3577         (-[KWQAccObject topRenderer]):
3578         (-[KWQAccObject topView]):
3579         (-[KWQAccObject accessibilityAttributeValue:]):
3580         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
3581         (-[KWQAccObject doAXLineForTextMarker:]):
3582         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
3583         (-[KWQAccObject doAXStringForTextMarkerRange:]):
3584         (-[KWQAccObject doAXTextMarkerForPosition:]):
3585         (-[KWQAccObject doAXBoundsForTextMarkerRange:]):
3586         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
3587         (-[KWQAccObject doAXTextMarkerRangeForUnorderedTextMarkers:]):
3588         (-[KWQAccObject doAXNextTextMarkerForTextMarker:]):
3589         (-[KWQAccObject doAXPreviousTextMarkerForTextMarker:]):
3590         (-[KWQAccObject doAXLeftWordTextMarkerRangeForTextMarker:]):
3591         (-[KWQAccObject doAXRightWordTextMarkerRangeForTextMarker:]):
3592         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
3593         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
3594         (-[KWQAccObject doAXSentenceTextMarkerRangeForTextMarker:]):
3595         (-[KWQAccObject doAXParagraphTextMarkerRangeForTextMarker:]):
3596         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
3597         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
3598         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
3599         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
3600         (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
3601         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
3602         (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
3603         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
3604         (-[KWQAccObject doAXLengthForTextMarkerRange:]):
3605         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
3606         (-[KWQAccObject accessibilityHitTest:]):
3607         (-[KWQAccObject accessibilityFocusedUIElement]):
3608         (-[KWQAccObject accessibilityIsAttributeSettable:]):
3609         (-[KWQAccObject doSetAXSelectedTextMarkerRange:]):
3610         (-[KWQAccObject setAccObjectID:]):
3611         (-[KWQAccObject removeAccObjectID]):
3612         * kwq/KWQAccObjectCache.h:
3613         * kwq/KWQAccObjectCache.mm:
3614         (KWQAccObjectCache::setAccObject):
3615         (KWQAccObjectCache::removeAccObject):
3616         (KWQAccObjectCache::visiblePositionForTextMarker):
3617         (KWQAccObjectCache::postNotificationToTopWebArea):
3618         (KWQAccObjectCache::postNotification):
3619         (KWQAccObjectCache::handleFocusedUIElementChanged):
3620         * kwq/KWQKHTMLPart.mm:
3621         (KWQKHTMLPart::respondToChangedContents):
3622         * kwq/KWQTextUtilities.h:
3623         * kwq/KWQTextUtilities.mm:
3624         (KWQFindNextWordFromIndex):
3625         (KWQFindSentenceBoundary):
3626         (KWQFindNextSentenceFromIndex):
3627         * kwq/WebCoreBridge.mm:
3628         (-[WebCoreBridge accessibilityTree]):
3629
3630 2005-01-05  Darin Adler  <darin@apple.com>
3631
3632         Reviewed by Ken.
3633
3634         - re-landing a subset of my custom tag change that does not fix the bug, but also does
3635           not introduce a performance regression
3636
3637         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
3638         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
3639
3640         * khtml/editing/htmlediting.cpp:
3641         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
3642         per-document tags and is just better all around for things like the document.
3643         (khtml::debugNode): Ditto.
3644         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
3645         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
3646         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
3647         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
3648
3649         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
3650         non-HTML elements to be nested as desired.
3651
3652         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
3653         * khtml/misc/htmlhashes.cpp:
3654         (khtml::getTagID): Changed return type to unsigned short.
3655         (khtml::getAttrID): Ditto.
3656
3657         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
3658         * khtml/xml/dom_docimpl.cpp:
3659         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
3660         Also updated for a few tags that the parser handled but this did not.
3661
3662         * kwq/KWQRenderTreeDebug.cpp:
3663         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
3664         the tag ID directly, which only works for standard nodes.
3665         (operator<<): Update to call getTagName.
3666         (nodePositionRelativeToRoot): Ditto.
3667         (writeSelection): Ditto.
3668
3669 2005-01-05  Ken Kocienda  <kocienda@apple.com>
3670
3671         Reviewed by Hyatt
3672
3673         Fix for this bug:
3674
3675         <rdar://problem/3941203> REGRESSION (Mail): Paste inserts content in wrong place
3676
3677         * khtml/editing/htmlediting.cpp:
3678         (khtml::ReplaceSelectionCommand::doApply): Some cleanup and refinement of the concepts used to make
3679         this operation work correctly, particularly in the logic to figure out whether to merge content, and
3680         also performing merges.
3681         * khtml/editing/visible_position.cpp:
3682         (khtml::isFirstVisiblePositionInBlock): Simplification of test used to make this determination.
3683         * khtml/editing/visible_units.cpp:
3684         (khtml::isStartOfParagraph): New helper, used in khtml::ReplaceSelectionCommand::doApply().
3685         (khtml::isEndOfParagraph): Ditto.
3686         * khtml/editing/visible_units.h: Declare new functions.
3687
3688 2005-01-04  Ken Kocienda  <kocienda@apple.com>
3689
3690         Reviewed by John
3691
3692         Fix for this bug:
3693
3694         <rdar://problem/3926522> Pressing return in a quoted block inserts too many newlines
3695
3696         * khtml/editing/htmlediting.cpp:
3697         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Call to insertBlockPlaceholderIfNeeded()
3698         for the block we are adding was being done at the wrong time, which led to the placeholder remaining
3699         in the document when it was not needed. This resulted in the extra space reported in the bug.
3700
3701 2005-01-04  Kevin Decker  <kdecker@apple.com>
3702
3703         Reviewed by Hyatt.
3704
3705         Fixed: <rdar://problem/3936879> nil-deref, crash in InlineFlowBox::nodeAtPoint closing DHTML menus at hrweb.apple.com
3706
3707         * khtml/rendering/render_block.cpp:
3708         (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.
3709
3710 2005-01-04  David Hyatt  <hyatt@apple.com>
3711
3712         Fix for 3937203, force an update of the dashboard regions when scrollbars are added/removed.
3713
3714         Reviewed by kevin
3715
3716         * khtml/rendering/render_layer.cpp:
3717         (RenderLayer::updateScrollInfoAfterLayout):
3718
3719 2005-01-04  Ken Kocienda  <kocienda@apple.com>
3720
3721         Reviewed by John
3722
3723         Fix for these two bugs:
3724
3725         <rdar://problem/3938935> REGRESSION (Mail): Pasting into an empty document mangles content
3726         <rdar://problem/3939148> REGRESSION (Mail): Pasting mistakenly reverses lines
3727
3728         * khtml/editing/htmlediting.cpp:
3729         (khtml::ReplaceSelectionCommand::doApply): For 3938935, add one more case to handle an empty document; merge
3730         neither start nor end. For 3939148, improve the code which adjusts the insertion point during
3731         the process of pasting. It formerly handled only one of the possible cases.
3732         * layout-tests/editing/pasteboard/paste-text-015-expected.txt: Added.
3733         * layout-tests/editing/pasteboard/paste-text-015.html: Added.
3734
3735 2005-01-04  David Hyatt  <hyatt@apple.com>
3736
3737         Fix for 3936571, placeholder attribute should work for normal inputs for Dashboard.
3738
3739         Reviewed by john
3740
3741         * khtml/rendering/render_form.cpp:
3742         (RenderLineEdit::updateFromElement):
3743         * kwq/KWQLineEdit.mm:
3744         (QLineEdit::setPlaceholderString):
3745
3746 2005-01-04  David Hyatt  <hyatt@apple.com>
3747
3748         Fix for 3830936, hang on changeforamerica.com.  Make sure to ignore the style not yet available option when
3749         returning pseudo-styles.
3750
3751         Reviewed by john
3752
3753         * khtml/css/cssstyleselector.cpp:
3754         (khtml::CSSStyleSelector::matchRulesForList):
3755         (khtml::CSSStyleSelector::pseudoStyleForElement):
3756         * khtml/rendering/render_style.cpp:
3757         (RenderStyle::addPseudoStyle):
3758
3759 2005-01-04  Darin Adler  <darin@apple.com>
3760
3761         - rolled out my custom tag name change again -- it caused a 1 ms PLT regression
3762
3763         * khtml/css/cssstyleselector.cpp:
3764         * khtml/editing/htmlediting.cpp:
3765         * khtml/editing/selection.cpp:
3766         * khtml/editing/visible_position.cpp:
3767         * khtml/html/dtd.cpp:
3768         * khtml/html/htmlparser.cpp:
3769         * khtml/html/htmlparser.h:
3770         * khtml/html/htmltokenizer.cpp:
3771         * khtml/misc/htmlhashes.cpp:
3772         * khtml/misc/htmlhashes.h:
3773         * khtml/misc/htmltags.c:
3774         * khtml/misc/htmltags.h:
3775         * khtml/misc/maketags:
3776         * khtml/xml/dom_docimpl.cpp:
3777         * khtml/xml/dom_docimpl.h:
3778         * khtml/xml/dom_nodeimpl.cpp:
3779         * khtml/xml/dom_position.cpp:
3780         * kwq/KWQRenderTreeDebug.cpp:
3781
3782 2005-01-04  Ken Kocienda  <kocienda@apple.com>
3783
3784         Reviewed by John
3785
3786         Fix for this bug:
3787
3788         <rdar://problem/3927554> Style info applied to remainder of document after a newline is entered
3789
3790         * khtml/editing/htmlediting.cpp:
3791         (khtml::InsertParagraphSeparatorCommand::doApply): Clean up and simplification in code that inserts
3792         a paragraph separator. The bug was all about applying styles to the new paragraph that did not need
3793         to be applied. Now the code will detect when at the end of a style run and will not move and apply
3794         that ending style to the new paragraph, though it will place that style into the typing style. This
3795         seems to match NSText behavior.
3796         * layout-tests/editing/inserting/insert-div-021-expected.txt: This test result exhibited the bug fixed
3797         here. The only reason it was not noticed is that the erroneously copied inline was a span, and so did
3798         not have any visible effect on the document.
3799         * layout-tests/editing/inserting/insert-div-022-expected.txt: Added.
3800         * layout-tests/editing/inserting/insert-div-022.html: Added.
3801
3802 2005-01-04  David Hyatt  <hyatt@apple.com>
3803
3804         Fix for 3904562, make sure to clear the outline dictionary when drawing so that random focus rings dont get drawn
3805         all over the place.
3806
3807         Reviewed by john
3808
3809         * khtml/rendering/render_flow.cpp:
3810         (RenderFlow::paintLines):
3811
3812 2005-01-03  David Hyatt  <hyatt@apple.com>
3813
3814         Fix for float painting regressions 3932524, 3931664, and 3933068.  Make the noPaint flag setting more
3815         robust and make it work regardless of which objects get a layout or not.
3816
3817         Reviewed by mjs
3818
3819         * khtml/rendering/render_block.cpp:
3820         (khtml::RenderBlock::insertFloatingObject):
3821         (khtml::RenderBlock::addOverhangingFloats):
3822         (khtml::RenderBlock::addIntrudingFloats):
3823         * khtml/rendering/render_block.h:
3824         (khtml::RenderBlock::FloatingObject::FloatingObject):
3825
3826 2005-01-03  Maciej Stachowiak  <mjs@apple.com>
3827
3828         Reviewed by Kevin.
3829
3830         <rdar://problem/3935390> Tiger 8A341: nil deref crash in DOM::NodeImpl::traverseNextNode
3831
3832         * khtml/html/html_miscimpl.cpp:
3833         (HTMLCollectionImpl::traverseNextItem): Assert that the starting point is not nil,
3834         it should not ever be (now).
3835         (HTMLCollectionImpl::item): When traversing items stop when we hit
3836         nil, meaning the end to avoid triggering above assert (formerly crash).
3837
3838 2005-01-03  Maciej Stachowiak  <mjs@apple.com>
3839
3840         Reviewed by John and Kevin.
3841
3842         <rdar://problem/3870317> REGRESSION(125.9-125.11) broken behavior at test.profoundlearning.com - used to
3843
3844         * khtml/ecma/kjs_window.cpp:
3845         (Window::get): Look up frame names before buitin window properties
3846         to match other browsers. This regressed because we added a builtin
3847         "toolbar" property but this site had a frame with that name.
3848
3849 2004-12-21  Maciej Stachowiak  <mjs@apple.com>
3850
3851         Reviewed by Darin.
3852
3853         <rdar://problem/3888931> frame naming allows malicious site to br