WebCore/ChangeLog-2005-08-23

   1 2005-01-20  Maciej Stachowiak  <mjs@apple.com>
   2
   3         Reviewed by Ken.
   4
   5         <rdar://problem/3965466> editing needs to insert text before applying typing style
   6
   7         * khtml/editing/htmlediting.cpp:
   8         (khtml::ApplyStyleCommand::addInlineStyleIfNeeded): Remove a FIXME comment.
   9         (khtml::InsertLineBreakCommand::doApply): Apply style to the BR
  10         node after inserting it.
  11         (khtml::InsertTextCommand::prepareForTextInsertion): Don't try to
  12         apply style to any new nodes created before inserting.
  13         (khtml::InsertTextCommand::input): Apply style to the inserted range
  14         after doing the text insert.
  15         * khtml/editing/htmlediting.h:
  16
  17         This change leads to improved results on one of the layout tests:
  18
  19         * layout-tests/editing/inserting/insert-div-024-expected.txt:
  20
  21 === Safari-180 ===
  22
  23 2005-01-20  Ken Kocienda  <kocienda@apple.com>
  24
  25         Reviewed by John
  26
  27         Fix for this bug:
  28
  29         <rdar://problem/3964646> REGRESSION (179-180): Typing space at end of line makes following paragraph disappear
  30
  31         * khtml/editing/htmlediting.cpp:
  32         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Recent change to block placeholder removal code
  33         caused this regression. The code became too aggressive in removing block placeholders, and would remove them
  34         from blocks other than the block containing the selection.
  35         * layout-tests/editing/inserting/insert-div-023-expected.txt: This file had a spurious BR element in it that
  36         I did not notice earlier.
  37         * layout-tests/editing/inserting/insert-div-024-expected.txt: Ditto.
  38
  39 2005-01-20  David Hyatt  <hyatt@apple.com>
  40
  41         Fix for oddness on albertsons.com.  Make sure not to crash when setting/removing style properties on a node
  42         with no document.  The bug # is 3813900.
  43
  44         Reviewed by john
  45
  46 2005-01-20  David Harrison  <harrison@apple.com>
  47
  48         Reviewed by Darin.
  49
  50         PARTIAL fix for following bug.  Create attributed string with fonts and links.  Still need to add attachments.
  51         <rdar://problem/3942606> AX: Support kAXAttributedStringForTextMarkerRangeParameterizedAttribute
  52
  53         * kwq/KWQAccObject.mm:
  54         (-[KWQAccObject textUnderElement]):
  55         Touched up previous checkin to match code review comments.
  56
  57         (-[KWQAccObject value]):
  58         Use plainText for this one instead of obsolete attributedString.
  59
  60         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
  61         Advertise AXAttributedStringForTextMarkerRange.
  62
  63         (AXAttributeStringAddFont):
  64         (AXAttributeStringAddElement):
  65         (-[KWQAccObject linkUIElementForNode:]):
  66         (-[KWQAccObject _addAttachmentForElement:URL:needsParagraph:usePlaceholder:]):
  67         (-[KWQAccObject accessibilityAttributedStringForRange:]):
  68         New routines to support AXAttributedStringForTextMarkerRange.
  69
  70         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
  71         Use new routines instead of obsolete attributedString.
  72
  73 2005-01-20  David Harrison  <harrison@apple.com>
  74
  75         Reviewed by Darin.
  76
  77         <rdar://problem/3960196> AX Crash in DOM::Range::setStartBefore
  78
  79         * khtml/xml/dom_docimpl.cpp:
  80         (DocumentImpl::getAccObjectCache):
  81         Adjust when detecting cache in non-top level WebArea.
  82         * kwq/KWQAccObject.mm:
  83         (-[KWQAccObject textUnderElement]):
  84         (-[KWQAccObject accessibilityAttributeValue:]):
  85         Make sure the elements document is the current one for the part.
  86
  87 2005-01-20  Darin Adler  <darin@apple.com>
  88
  89         Reviewed by John.
  90
  91         - fixed <rdar://problem/3922980> Mail not crashing, just quitting itself suddenly in -[WebHTMLView(MailExtras) findString:options:]
  92
  93         * khtml/dom/dom2_range.cpp: (DOM::operator==): Handle null and detached ranges without raising exceptions.
  94         The uncaught exception would make the entire program terminate.
  95
  96         - fixed <rdar://problem/3885729> attempting to modify a computed style does nothing, but should raise a DOM exception
  97
  98         * kwq/DOM-CSS.mm: (-[DOMCSSStyleDeclaration setProperty:::]): Uncommented this code which I had to disable
  99         back in November because Mail was stumbling over it.
 100
 101         - fixed <rdar://problem/3943049> focus() called during onload handler results in square text field on <input type=search>
 102
 103         * kwq/KWQLineEdit.mm: (QLineEdit::baselinePosition): Change computation so it will work even for fields
 104         that position their text in a way that depends on vertical size. In the case of a search field, the text
 105         is centered, so the old logic was broken.
 106         * kwq/KWQTextField.mm: (-[KWQTextFieldController initWithTextField:QLineEdit:]): Set the frame size to
 107         something large enough to accomodate the field editor. If we start the frame at size 0,0 we run into
 108         AppKit trouble when it insets the frame to figure out the frame for the field editor. In the case of this
 109         bug this happens because we become first responder before being sized and positioned by the HTML layout code.
 110
 111         - improved debugging output when using "po" from gdb with Objective-C DOM
 112
 113         * kwq/DOM.mm:
 114         (-[DOMNode description]): Added. Includes node name ("<tr>") and node value (e.g., string for text node).
 115         (-[DOMRange description]): Tweaked format.
 116
 117 2005-01-19  Richard Williamson   <rjw@apple.com>
 118
 119         After further discussion with Real we have decided to NOT include the
 120         additional CLSID for the real plugin.  See 3958601.
 121
 122         Reviewed by Darin.
 123
 124         * khtml/rendering/render_frames.cpp:
 125         (RenderPartObject::updateWidget):
 126
 127 2005-01-19  David Harrison  <harrison@apple.com>
 128
 129         Reviewed by Maciej.
 130
 131         <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
 132
 133         Previous checkin for this bug was missing some of the patch.  Editing snafu with multiple changes in tree.
 134
 135         * kwq/KWQAccObject.mm:
 136         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
 137         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
 138         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
 139         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
 140         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
 141         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
 142         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
 143
 144 2005-01-19  Ken Kocienda  <kocienda@apple.com>
 145
 146         Reviewed by John
 147
 148         * khtml/editing/htmlediting.cpp:
 149         (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Small, cosmetic change John and I decided
 150         to do on my last checkin, but I forgot to do before landing.
 151
 152 2005-01-19  Ken Kocienda  <kocienda@apple.com>
 153
 154         Reviewed by John
 155
 156         Fix for this bug:
 157
 158         <rdar://problem/3959727> REGRESSION (Mail): Style not preserved on blank lines
 159
 160         * khtml/editing/htmlediting.cpp:
 161         (khtml::CompositeEditCommand::applyStyle):
 162         (khtml::CompositeEditCommand::insertBlockPlaceholder): New function that unconditionally adds a block placeholder.
 163         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Now returns bool based on whether
 164         placeholder was added or not.
 165         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Now searches all the descendents of a block
 166         looking for a placeholder. The old code, which just looked at the last child of a node, started missing
 167         once block placeholders became styled (which started happening with this patch).
 168         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Now handles applying typing style
 169         to a block placeholder at call time, rather than setting the typing style as a latent style that
 170         might be applied later. This is an important part of the bug fix.
 171         (khtml::DeleteSelectionCommand::doApply): Now uses bool return value from insertBlockPlaceholderIfNeeded()
 172         and passes it along to calculateStyleBeforeInsertion, so the case where a block placeholder needs to
 173         be styled can be detected.
 174         (khtml::InsertParagraphSeparatorCommand::InsertParagraphSeparatorCommand): Changed the way this class
 175         managed style. Before it would calculate and set typing style for the block added. This is not
 176         sufficient. Added blocks need to styled immediately. Some name changes to instance variables in
 177         this class due to the change to accommodate this change.
 178         (khtml::InsertParagraphSeparatorCommand::~InsertParagraphSeparatorCommand): Name changes, as above.
 179         (khtml::InsertParagraphSeparatorCommand::calculateStyleBeforeInsertion): Ditto.
 180         (khtml::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Ditto.
 181         (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
 182         (khtml::ReplaceSelectionCommand::doApply): Improve check for testing when a placeholder
 183         block can be removed in its entirety after the insertion.
 184         * khtml/editing/htmlediting.h: Update header accordingly.
 185         * khtml/khtml_part.cpp:
 186         (KHTMLPart::selectionComputedStyle): Move position for computed style check downstream before
 187         doing check when the position is in an empty block (this makes sure any style on any block
 188         placeholder is accounted for).
 189
 190         New layout tests to check bug fix.
 191
 192         * layout-tests/editing/style/block-style-004-expected.txt: Added.
 193         * layout-tests/editing/style/block-style-004.html: Added.
 194         * layout-tests/editing/style/block-style-005-expected.txt: Added.
 195         * layout-tests/editing/style/block-style-005.html: Added.
 196         * layout-tests/editing/style/block-style-006-expected.txt: Added.
 197         * layout-tests/editing/style/block-style-006.html: Added.
 198
 199         Results updated to reflect new block placeholder code.
 200
 201         * layout-tests/editing/inserting/insert-div-004-expected.txt
 202         * layout-tests/editing/inserting/insert-div-005-expected.txt
 203         * layout-tests/editing/inserting/insert-div-006-expected.txt
 204         * layout-tests/editing/inserting/insert-div-008-expected.txt
 205         * layout-tests/editing/inserting/insert-div-011-expected.txt
 206         * layout-tests/editing/inserting/insert-div-012-expected.txt
 207         * layout-tests/editing/inserting/insert-div-013-expected.txt
 208         * layout-tests/editing/inserting/insert-div-014-expected.txt
 209         * layout-tests/editing/inserting/insert-div-015-expected.txt
 210         * layout-tests/editing/inserting/insert-div-016-expected.txt
 211         * layout-tests/editing/inserting/insert-div-017-expected.txt
 212         * layout-tests/editing/inserting/insert-div-018-expected.txt
 213         * layout-tests/editing/inserting/insert-div-019-expected.txt
 214         * layout-tests/editing/inserting/insert-div-021-expected.txt
 215         * layout-tests/editing/inserting/insert-div-022-expected.txt
 216         * layout-tests/editing/inserting/insert-div-023-expected.txt
 217         * layout-tests/editing/inserting/insert-div-024-expected.txt
 218
 219 2005-01-19  David Hyatt  <hyatt@apple.com>
 220
 221         Dont null-check the renderer before submitting, since a script can set it to display:none and still expect the
 222         submission to occur.  Fixes bug #3477282.
 223
 224         Reviewed by kdecker
 225
 226         (DOM::HTMLInputElementImpl::defaultEventHandler):
 227
 228 2005-01-18  Richard Williamson   <rjw@apple.com>
 229
 230         Fixed several issues all arising from analysis of plugin detection code at ifilm.com:
 231
 232         Fixed <rdar://problem/3958592> can't script plug-ins if plug-in is invoked with <object> element instead of <embed>
 233         Fixed <rdar://problem/3958597> <object> elements with IDs do not show up as named properties of the document
 234         Fixed <rdar://problem/3960973> DOM objects for plugin elements are not accessible
 235         Fixed <rdar://problem/3958601> need an additional class ID in WebCore for the Real plug-in
 236
 237         We now support accessing scriptable plugin objects that are specified with <applet>, <embed>, or <object>
 238         tags.  Also, if any of these elements are named they can be accessed from the document or window objects.
 239         Finally, DOM methods are properties will be forwarded appropriately for the plugin's root scriptable object.
 240
 241         Reviewed by Chris.
 242
 243         * khtml/dom/html_document.cpp:
 244         (HTMLDocument::objects):
 245         * khtml/dom/html_document.h:
 246         * khtml/ecma/kjs_dom.cpp:
 247         (DOMDocumentProtoFunc::tryCall):
 248         (DOMElementProtoFunc::tryCall):
 249         (KJS::getRuntimeObject):
 250         * khtml/ecma/kjs_dom.h:
 251         * khtml/ecma/kjs_html.cpp:
 252         (KJS::HTMLDocument::tryGet):
 253         (KJS::HTMLElement::tryGet):
 254         (KJS::HTMLCollection::tryGet):
 255         (KJS::HTMLCollection::getNamedItems):
 256         * khtml/ecma/kjs_window.cpp:
 257         (Window::get):
 258         * khtml/html/html_miscimpl.cpp:
 259         (HTMLCollectionImpl::traverseNextItem):
 260         * khtml/html/html_miscimpl.h:
 261         (DOM::HTMLCollectionImpl::):
 262         * khtml/html/html_objectimpl.cpp:
 263         (HTMLAppletElementImpl::getAppletInstance):
 264         (HTMLObjectElementImpl::HTMLObjectElementImpl):
 265         (HTMLObjectElementImpl::getObjectInstance):
 266         * khtml/html/html_objectimpl.h:
 267         * khtml/rendering/render_frames.cpp:
 268         (RenderPartObject::updateWidget):
 269         * kwq/KWQKHTMLPart.h:
 270         * kwq/KWQKHTMLPart.mm:
 271         (KWQKHTMLPart::getObjectInstanceForView):
 272
 273 2005-01-18  David Hyatt  <hyatt@apple.com>
 274
 275         Fix for 3948123, rolling over link erases nearby text.  The repaint rect check for lines was wrong whenever
 276         two lines overlapped.
 277
 278         Reviewed by kocienda
 279
 280         * khtml/rendering/render_flow.cpp:
 281         (RenderFlow::paintLines):
 282
 283 2005-01-18  Ken Kocienda  <kocienda@apple.com>
 284
 285         Reviewed by Hyatt
 286
 287         Fix for this bug:
 288
 289         <rdar://problem/3960116> Focus rings paint incorrectly for contenteditable blocks in web pages
 290
 291         * khtml/rendering/render_flow.cpp:
 292         (RenderFlow::addFocusRingRects): Fix painting of focus rings so that ring only paints around
 293         outermost contenteditable elements.
 294
 295 2005-01-18  David Harrison  <harrison@apple.com>
 296
 297         Reviewed by Darin.
 298
 299         <rdar://problem/3959668> accessibilityFocusedUIElement sometimes returns an ignored element; it must not
 300
 301         * kwq/KWQAccObject.mm:
 302         (-[KWQAccObject accessibilityFocusedUIElement]):
 303         Return parentObjectUnignored if focused object is ignored.
 304
 305 2005-01-18  Ken Kocienda  <kocienda@apple.com>
 306
 307         Reviewed by John
 308
 309         Fix for this bug:
 310
 311         <rdar://problem/3952877> REGRESSION (Mail): Command-left/right-arrows don't work with file attachment
 312
 313         * khtml/editing/selection.cpp:
 314         (khtml::nodeForInlineBox): New helper function used in reimplementation of function below.
 315         (khtml::selectionForLine): Reimplemented using line box smarts. I originally wrote this code when
 316         I had a less than full understanding of line layout. I can do better now, and my new version no
 317         longer fails to notice attachments when doing the kind of navigation mentioned in the bug.
 318
 319 2005-01-17  David Harrison  <harrison@apple.com>
 320
 321         Reviewed by John Sullivan.
 322
 323         <rdar://problem/3949429> AX: word marker routines returns incorrect data for empty line
 324         <rdar://problem/3949848> AX: paragraph marker routines do not work when given a paragraph end marker
 325
 326         Also fixed sentence support in the same way.
 327
 328         * kwq/KWQAccObject.mm:
 329         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
 330         (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
 331         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
 332         (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
 333         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
 334         Move one position in desired direction before calling visible_units.cpp code.
 335
 336 2005-01-17  David Harrison  <harrison@apple.com>
 337
 338         Updated expected test results.
 339
 340         <rdar://problem/3945880> line-ending space seems not to be present
 341
 342         * layout-tests/editing/deleting/delete-ws-fixup-001-expected.txt:
 343         * layout-tests/editing/deleting/delete-ws-fixup-002-expected.txt:
 344         * layout-tests/editing/inserting/insert-div-020-expected.txt:
 345         * layout-tests/editing/inserting/insert-div-021-expected.txt:
 346         * layout-tests/editing/inserting/insert-div-022-expected.txt:
 347         * layout-tests/editing/inserting/insert-div-023-expected.txt:
 348         * layout-tests/editing/inserting/insert-div-024-expected.txt:
 349
 350 2005-01-17  David Harrison  <harrison@apple.com>
 351
 352         Reviewed by Dave Hyatt (bidi.cpp) and Darin Adler (selection.cpp).
 353
 354         <rdar://problem/3945880> line-ending space seems not to be present
 355
 356         * khtml/editing/selection.cpp:
 357         (khtml::Selection::validate):
 358         Extend AFTER_WHITE_SPACE code to support white spac in the middle of paragraphs, not just the end.
 359         * khtml/rendering/bidi.cpp:
 360         (khtml::RenderBlock::findNextLineBreak):
 361         Pick left/rightness of word selection based on being at the end of paragraph (i.e. after a hard line break).
 362
 363 2005-01-17  Darin Adler  <darin@apple.com>
 364
 365         Reviewed by John Louch.
 366
 367         - fixed <rdar://problem/3958503> need screenX and screenY to use WebKit windowFrame delegate
 368
 369         * khtml/ecma/kjs_window.cpp: (Window::get): Change screenX and screenY to use frameGeometry instead
 370         of using mapToGlobal and screen in a complicated way.
 371         * kwq/KWQKHTMLView.mm: Removed unused mapToGlobal function.
 372         * kwq/KWQWindowWidget.h: Ditto.
 373         * kwq/KWQWindowWidget.mm: Ditto.
 374
 375 2005-01-17  David Hyatt  <hyatt@apple.com>
 376
 377         Fix a screwup in rightmost/lowets position computation.  3955207.  Make sure floats with layers are still checked.
 378
 379         Reviewed by kocienda
 380
 381         * khtml/rendering/render_block.cpp
 382
 383 2005-01-17  Ken Kocienda  <kocienda@apple.com>
 384
 385         Reviewed by John
 386
 387         <rdar://problem/3953366> Problems with typing attributes in HTML compose
 388
 389         * khtml/editing/htmlediting.cpp:
 390         (khtml::DeleteSelectionCommand::calculateTypingStyleAfterDelete): Remove misguided "optimization"
 391         that tried to sense when typing style could be cleared without actually doing a style diff between
 392         before-delete and after-delete positions. Removing this extra check and running the
 393         general-purpose code fixes the bug.
 394
 395 2005-01-17  Richard Williamson   <rjw@apple.com>
 396
 397         Fixed <rdar://problem/3753030> Need to ensure same origin for plugin binding invocations (origin security rules)
 398
 399         Keep track of originating execution context and target execution
 400         context for native JS object wrappers, and perform appropriate
 401         security checks.
 402
 403         Reviewed by David Harrison.
 404
 405         * khtml/ecma/kjs_binding.cpp:
 406         (ScriptInterpreter::isGlobalObject):
 407         (ScriptInterpreter::isSafeScript):
 408         (ScriptInterpreter::interpreterForGlobalObject):
 409         * khtml/ecma/kjs_binding.h:
 410         * khtml/ecma/kjs_window.cpp:
 411         (Window::interpreter):
 412         (Window::isSafeScript):
 413         * khtml/ecma/kjs_window.h:
 414         * kwq/DOMInternal.mm:
 415         (-[WebScriptObject _initializeScriptDOMNodeImp]):
 416         * kwq/KWQKHTMLPart.mm:
 417         (KWQKHTMLPart::windowScriptObject):
 418         (KWQKHTMLPart::windowScriptNPObject):
 419         * kwq/WebCoreBridge.mm:
 420         (rootForView):
 421         (-[WebCoreBridge init]):
 422
 423 2005-01-17  Ken Kocienda  <kocienda@apple.com>
 424
 425         Reviewed by John
 426
 427         Fix for this bug:
 428
 429         <rdar://problem/3786659> REGRESSION (Mail): editable WebViews don't work with
 430         "size up" and "size down" NSFontManager changes
 431
 432         * khtml/css/cssparser.cpp:
 433         (CSSParser::parseValue): Add support for parsing new font size delta property.
 434         * khtml/css/cssproperties.c: Generated file.
 435         * khtml/css/cssproperties.h: Ditto.
 436         * khtml/css/cssproperties.in: Add support for parsing new font size delta property.
 437         * khtml/editing/htmlediting.cpp:
 438         (khtml::isEmptyStyleSpan): New helper function. Code existed before, but now factored out for convenient use.
 439         (khtml::CompositeEditCommand::removeNodeAttribute): Does check on value to see it exists before creating
 440         and running command to do the removal.
 441         (khtml::ApplyStyleCommand::doApply): Now calls new applyRelativeFontStyleChange() function as part of its work.
 442         (khtml::ApplyStyleCommand::applyRelativeFontStyleChange): New function that does the "heavy lifting" to handle
 443         relative font size changes.
 444         (khtml::ApplyStyleCommand::applyInlineStyle): Range check the start and end positions to make sure the start is
 445         before or equal to the end. Swap them if not true. I ran into this problem in some similar code in
 446         applyRelativeFontStyleChange(). Moving that goodness here too.
 447         (khtml::ApplyStyleCommand::splitTextAtEndIfNeeded): Uses a local variable to save a value used often.
 448         Code is the same, but reads better now. Function now returns bool as well, just like splitTextAtStartIfNeeded.
 449         I use the bool return value now (I obviously did not need it before).
 450         (khtml::ApplyStyleCommand::computedFontSize): New helper function.
 451         (khtml::ApplyStyleCommand::joinChildTextNodes): Ditto.
 452         (khtml::createStyleSpanElement): Ditto.
 453         * khtml/editing/htmlediting.h: Update header accordingly.
 454         * khtml/editing/jsediting.cpp: Add new command string to enable relative font size changes.
 455         * kwq/DOM-CSS.mm:
 456         (-[DOMCSSStyleDeclaration _fontSizeDelta]): New convenience.
 457         (-[DOMCSSStyleDeclaration _setFontSizeDelta:]): Ditto.
 458         * kwq/DOMPrivate.h: Declare new conveniences.
 459         * layout-tests/editing/editing.js: Add new JS to enable relative font size changes, as well as explicit font size setting.
 460         * layout-tests/editing/style/relative-font-size-change-001-expected.txt: Added.
 461         * layout-tests/editing/style/relative-font-size-change-001.html: Added.
 462         * layout-tests/editing/style/relative-font-size-change-002-expected.txt: Added.
 463         * layout-tests/editing/style/relative-font-size-change-002.html: Added.
 464         * layout-tests/editing/style/relative-font-size-change-003-expected.txt: Added.
 465         * layout-tests/editing/style/relative-font-size-change-003.html: Added.
 466         * layout-tests/editing/style/relative-font-size-change-004-expected.txt: Added.
 467         * layout-tests/editing/style/relative-font-size-change-004.html: Added.
 468
 469 2005-01-14  Darin Adler  <darin@apple.com>
 470
 471         * khtml/css/cssproperties.c: Regenerated with gperf 3.0.1.
 472         * khtml/css/cssvalues.c: Regenerated with gperf 3.0.1.
 473         * khtml/misc/htmlattrs.c: Regenerated with gperf 3.0.1.
 474         * khtml/misc/htmltags.c: Regenerated with gperf 3.0.1.
 475
 476 2005-01-14  Chris Blumenberg  <cblu@apple.com>
 477
 478         Fixed: <rdar://problem/3886415> arrow keys don't work when the user hits Back to return to RSS page
 479
 480         Reviewed by hyatt.
 481
 482         (KWQKHTMLPart::openURLFromPageCache): restore mousePressNode
 483         (KWQKHTMLPart::mousePressNode): new
 484         * kwq/KWQPageState.h:
 485         * kwq/KWQPageState.mm:
 486         (-[KWQPageState initWithDocument:URL:windowProperties:locationProperties:interpreterBuiltins:]): save the mousePressNode
 487         (-[KWQPageState clear]): clear the mousePressNode
 488         (-[KWQPageState dealloc]): deref the mousePressNode
 489         (-[KWQPageState mousePressNode]): new
 490
 491 === Safari-179 ===
 492
 493 2005-01-13  Vicki Murley <vicki@apple.com>
 494
 495         Reviewed by Adele.
 496
 497         - fix <rdar://problem/3946836> Safari about box lists 2004 instead of 2005
 498
 499         * WebCore.pbproj/project.pbxproj: bump "2004" to "2005"
 500
 501 2005-01-13  David Harrison  <harrison@apple.com>
 502
 503         Reviewed by Ken Kocienda.
 504
 505         Better fix for 3905066.
 506
 507         * khtml/editing/htmlediting.cpp:
 508         (khtml::InsertParagraphSeparatorCommand::doApply):
 509
 510 2005-01-10  Maciej Stachowiak  <mjs@apple.com>
 511
 512         Reviewed by Darin.
 513
 514         <rdar://problem/3758033> REGRESSION (Mail): Support attributes in marked text (International input)
 515
 516         * khtml/rendering/render_text.cpp:
 517         (InlineTextBox::paint): Support painting custom underline markers for
 518         marked text in place of generic yellow.
 519         (InlineTextBox::paintMarkedTextUnderline): New method that handles this.
 520         * khtml/rendering/render_text.h:
 521         * kwq/KWQKHTMLPart.h: Declare new methods and structs.
 522         * kwq/KWQKHTMLPart.mm:
 523         (KWQKHTMLPart::clear): Clear marked test underlines.
 524         (KWQKHTMLPart::setMarkedTextRange): Takes attributes and ranges now.
 525         (convertAttributesToUnderlines): Converts NSAttributedString attributes
 526         to simplified and C++-friendly form.
 527         (KWQKHTMLPart::markedTextUsesUnderlines): New method.
 528         (KWQKHTMLPart::markedTextUnderlines): New method.
 529         * kwq/KWQPainter.mm:
 530         (QPainter::drawLineForText): Handle pen width.
 531         * kwq/WebCoreBridge.h:
 532         * kwq/WebCoreBridge.mm:
 533         (-[WebCoreBridge setMarkedTextDOMRange:customAttributes:ranges:]): Take attributes
 534         and ranges.
 535         * kwq/WebCoreTextRenderer.h:
 536
 537 2005-01-12  David Harrison  <harrison@apple.com>
 538
 539         Reviewed by Dave Hyatt.
 540
 541         <rdar://problem/3888973> AX: Parent AXWebArea of nested AXWebAreas is messed up
 542
 543         * kwq/KWQAccObject.mm:
 544         (-[KWQAccObject addChildrenToArray:]):
 545         Use the widget's outer view.
 546
 547 2005-01-12  David Harrison  <harrison@apple.com>
 548
 549         Reviewed by Darin Adler.
 550
 551         <rdar://problem/3949908> Crash when asking for the kAXLengthForTextMarkerRangeParameterizedAttribute
 552
 553         * khtml/xml/dom_nodeimpl.cpp:
 554         (NodeImpl::displayNode):
 555         Add quotes around text node content.
 556         (NodeBaseImpl::childNode):
 557         Add nil check to return nil rather than crash when child node not found.
 558         * kwq/KWQAccObject.mm:
 559         (-[KWQAccObject doAXStringForTextMarkerRange:]):
 560         Pass range compliant positions to TextIterator.
 561
 562 2005-01-12  David Hyatt  <hyatt@apple.com>
 563
 564         Fix for 3951203, CSS border style incorrectly clipped on inline elements.  Make sure the repaint rect set during
 565         line layout is smarter about including the overflow for both old states and new states.
 566
 567         Reviewed by john
 568
 569         * khtml/rendering/bidi.cpp:
 570         (khtml::RenderBlock::layoutInlineChildren):
 571
 572 2005-01-12  David Hyatt  <hyatt@apple.com>
 573
 574         Fix for bug 3937608, versiontracker.com flashes and displays the right column below content.  Make sure to move
 575         tables/overflows that dont fit within a block only in strict mode.
 576
 577         Fix for bug 3931049, characters dont show up when typing.  Make sure to dirty the right lines when this specific
 578         case in editing is hit.
 579
 580         Reviewed by darin (first one), kocienda (second one)
 581
 582         * khtml/rendering/render_block.cpp:
 583         (khtml::RenderBlock::getClearDelta):
 584         * khtml/rendering/render_flow.cpp:
 585         (RenderFlow::dirtyLinesFromChangedChild):
 586         * khtml/rendering/render_line.cpp:
 587         (khtml::InlineFlowBox::verticallyAlignBoxes):
 588
 589 2005-01-12  Ken Kocienda  <kocienda@apple.com>
 590
 591         Reviewed by John
 592
 593         Fix for this bug:
 594
 595         <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
 596
 597         * khtml/editing/htmlediting.cpp:
 598         (khtml::DeleteSelectionCommand::initializePositionData): Change test that will prevent block
 599         merging. End-of-line test was just wrong. Call new start-of and end-of paragraph functions
 600         added to visible_position files.
 601         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Add tests for BR elements, which makes
 602         tests to determine when to stop moving nodes more complete and correct. Also improved comments.
 603         * khtml/editing/visible_position.cpp:
 604         (khtml::isFirstVisiblePositionInParagraph): New function.
 605         (khtml::isLastVisiblePositionInParagraph): New function.
 606         * khtml/editing/visible_position.h: Update header accordingly.
 607         * layout-tests/editing/deleting/delete-block-merge-contents-002-expected.txt: New results, slightly different from former
 608         results but still correct.
 609         * layout-tests/editing/deleting/delete-block-merge-contents-007-expected.txt: Ditto.
 610         * layout-tests/editing/deleting/delete-block-merge-contents-018-expected.txt: Added.
 611         * layout-tests/editing/deleting/delete-block-merge-contents-018.html: Added.
 612         * layout-tests/editing/deleting/delete-block-merge-contents-019-expected.txt: Added.
 613         * layout-tests/editing/deleting/delete-block-merge-contents-019.html: Added.
 614         * layout-tests/editing/deleting/delete-block-merge-contents-020-expected.txt: Added.
 615         * layout-tests/editing/deleting/delete-block-merge-contents-020.html: Added.
 616         * layout-tests/editing/deleting/delete-block-merge-contents-021-expected.txt: Added.
 617         * layout-tests/editing/deleting/delete-block-merge-contents-021.html: Added.
 618         * layout-tests/editing/deleting/delete-line-014-expected.txt: These new results are actually better, and fix a bug.
 619
 620 2005-01-11  Richard Williamson   <rjw@apple.com>
 621
 622         Fixed 3922875.  Fall back to DOM object is EMBED element
 623         has no associated runtime object.
 624
 625         Reviewed by Chris.
 626
 627         * khtml/ecma/kjs_dom.cpp:
 628         (KJS::getRuntimeObject):
 629         * khtml/ecma/kjs_html.cpp:
 630         (KJS::HTMLDocument::tryGet):
 631         (KJS::HTMLElement::tryGet):
 632         (KJS::HTMLCollection::tryGet):
 633         (KJS::HTMLCollection::getNamedItems):
 634
 635 2005-01-11  David Hyatt  <hyatt@apple.com>
 636
 637         Fix for 3882299, missing content on gibson.com.  Change our handling of " and ' in certain states of the parser to match
 638         other browsers.
 639
 640         Reviewed by Maciej
 641
 642         * khtml/html/htmltokenizer.cpp:
 643         (khtml::HTMLTokenizer::parseTag):
 644
 645 2005-01-11  Chris Blumenberg  <cblu@apple.com>
 646
 647         Fixed: <rdar://problem/3930733> Mail prints second page of email blank
 648
 649         Reviewed by dave.
 650
 651         * khtml/rendering/render_canvas.cpp:
 652         (RenderCanvas::paint): cache the print rect since the dirty rect can get changed during printing
 653         * khtml/rendering/render_flow.cpp:
 654         (RenderFlow::paintLines): removed null check since the print rect should never be null
 655         * khtml/rendering/render_list.cpp:
 656         (RenderListMarker::paint): ditto
 657         * kwq/KWQKHTMLPart.mm:
 658         (KWQKHTMLPart::adjustPageHeight): don't set the print rect here since this method is not called for the last page
 659
 660 2005-01-10  Ken Kocienda  <kocienda@apple.com>
 661
 662         Reviewed by Hyatt
 663
 664         Fix for this bug:
 665
 666         <rdar://problem/3943648> extra line is inserted after pressing return within quoted text of reply
 667
 668         * khtml/editing/htmlediting.cpp:
 669         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): This bug was the result of a
 670         simple coding mistake. A local variable was erroneously redefined in a deeper scope, and so the
 671         result calculated in that deeper scope was not available when tested.
 672
 673 2005-01-10  Ken Kocienda  <kocienda@apple.com>
 674
 675         Reviewed by John
 676
 677         Fix for this bug:
 678
 679         <rdar://problem/3946852> Option-e goes to next line
 680
 681         * khtml/editing/htmlediting.cpp:
 682         (khtml::ReplaceSelectionCommand::doApply): Tweak case used to determine when merging content
 683         into the start line is done. We plan to change pretty substantially soon to better handle
 684         the problem described in <rdar://problem/3937352> Quote level not maintained when copied
 685         and pasted within a Mail message. In the meantime, this change does no harm, and fixes the bug.
 686
 687 2005-01-10  Ken Kocienda  <kocienda@apple.com>
 688
 689         Reviewed by Darin
 690
 691         Fix for this bug:
 692
 693         <rdar://problem/3907005> Applying block styles to a line of text can unexpectedly affect other lines
 694
 695         * khtml/editing/htmlediting.cpp:
 696         (khtml::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): New function which moves "paragraphs"
 697         to their own blocks if needed so that a block style can be applied.
 698         (khtml::CompositeEditCommand::isMailBlockquote): Moved this function to base class so it can be used more generally.
 699         (khtml::ApplyStyleCommand::applyBlockStyle): Pass a node instead of a block to addBlockStyleIfNeeded().
 700         (khtml::ApplyStyleCommand::addBlockStyleIfNeeded): Now accepts a node instead of a block for styling.
 701         This function also now calls moveParagraphContentsToNewBlockIfNecessary() to make sure that only the current
 702         paragraph is styled.
 703         (khtml::SetNodeAttributeCommand::doUnapply): Should not assert on undo if old value of attribute was null.
 704         Should remove attributue instead.
 705         * khtml/editing/htmlediting.h: Touch function declarations accordingly.
 706         * layout-tests/editing/style/create-block-for-style-001-expected.txt: Added.
 707         * layout-tests/editing/style/create-block-for-style-001.html: Added.
 708         * layout-tests/editing/style/create-block-for-style-002-expected.txt: Added.
 709         * layout-tests/editing/style/create-block-for-style-002.html: Added.
 710         * layout-tests/editing/style/create-block-for-style-003-expected.txt: Added.
 711         * layout-tests/editing/style/create-block-for-style-003.html: Added.
 712         * layout-tests/editing/style/create-block-for-style-004-expected.txt: Added.
 713         * layout-tests/editing/style/create-block-for-style-004.html: Added.
 714         * layout-tests/editing/style/create-block-for-style-005-expected.txt: Added.
 715         * layout-tests/editing/style/create-block-for-style-005.html: Added.
 716         * layout-tests/editing/style/create-block-for-style-006-expected.txt: Added.
 717         * layout-tests/editing/style/create-block-for-style-006.html: Added.
 718         * layout-tests/editing/style/create-block-for-style-007-expected.txt: Added.
 719         * layout-tests/editing/style/create-block-for-style-007.html: Added.
 720         * layout-tests/editing/style/create-block-for-style-008-expected.txt: Added.
 721         * layout-tests/editing/style/create-block-for-style-008.html: Added.
 722         * layout-tests/editing/style/create-block-for-style-009-expected.txt: Added.
 723         * layout-tests/editing/style/create-block-for-style-009.html: Added.
 724         * layout-tests/editing/style/create-block-for-style-010-expected.txt: Added.
 725         * layout-tests/editing/style/create-block-for-style-010.html: Added.
 726         * layout-tests/editing/style/create-block-for-style-011-expected.txt: Added.
 727         * layout-tests/editing/style/create-block-for-style-011.html: Added.
 728         * layout-tests/editing/style/create-block-for-style-012-expected.txt: Added.
 729         * layout-tests/editing/style/create-block-for-style-012.html: Added.
 730         * layout-tests/editing/style/create-block-for-style-013-expected.txt: Added.
 731         * layout-tests/editing/style/create-block-for-style-013.html: Added.
 732
 733         Unrelated updates to these expected results.
 734         * layout-tests/editing/inserting/insert-div-007-expected.txt
 735         * layout-tests/editing/pasteboard/paste-text-013-expected.txt
 736
 737 2005-01-10  Chris Blumenberg  <cblu@apple.com>
 738
 739         Fixed: <rdar://problem/3936844> Mail: Messages with rich text do not print
 740
 741         Reviewed by john.
 742
 743         * khtml/rendering/render_flow.cpp:
 744         (RenderFlow::paintLines): don't do pagination work if printRect is not set
 745         * khtml/rendering/render_list.cpp:
 746         (RenderListMarker::paint): ditto
 747
 748 2005-01-10  David Harrison  <harrison@apple.com>
 749
 750         Reviewed by Darin.
 751
 752         * kwq/KWQTextUtilities.mm:
 753         (currentTextBreakLocaleID):
 754         Return empty string (AKA root locale) if locale pref can not be canonicalized.
 755
 756 2005-01-10  John Sullivan  <sullivan@apple.com>
 757
 758         Fixed broken Panther build.
 759
 760         * kwq/KWQTextUtilities.mm:
 761         (currentTextBreakLocaleID):
 762         This recently-added function was using code copy/pasted from CarbonCore UnicodeUtilities.
 763         That code used a Tiger-only function, CFLocaleCreateCanonicalLanguageIdentifierFromString.
 764         To fix the build, I added a BUILDING_ON_PANTHER #ifdef that avoids the call to the Tiger
 765         function. However, the Tiger-only code was wrong; the string generated using
 766         CFLocaleCreateCanonicalLanguageIdentifierFromString was not used at all, so I fixed that
 767         as well.
 768
 769 2005-01-09  David Harrison  <harrison@apple.com>
 770
 771         Reviewed by Ken Kocienda.
 772
 773         <rdar://problem/3786362> REGRESSION (Mail): pasted text loses one newline
 774
 775         * khtml/editing/htmlediting.cpp:
 776         (khtml::ReplaceSelectionCommand::doApply):
 777         Tune check for need for insertParagraphSeparator when hasInterchangeNewline.
 778
 779 2005-01-09  Darin Adler  <darin@apple.com>
 780
 781         Reviewed by Harrison.
 782
 783         - fixed <rdar://problem/3939176> select() method does not work on <input type=search>
 784
 785         * khtml/html/html_formimpl.cpp: (DOM::HTMLInputElementImpl::select): Changed if statement
 786         to a switch statement. Added SEARCH to the set of types that treat the renderer as a
 787         RenderLineEdit.
 788
 789 2005-01-09  David Harrison  <harrison@apple.com>
 790
 791         Reviewed by Ken Kocienda.
 792
 793         <rdar://problem/3905066> REGRESSION (Mail): Hitting return key with caret in front of space causes space to become lost (resize fixes)
 794
 795         Problem was that InsertParagraphSeparatorCommand::doApply() was not collapsing
 796         whitespace to a single non-breaking space when splitting a text node.
 797
 798         * khtml/editing/htmlediting.cpp:
 799         (khtml::InsertParagraphSeparatorCommand::doApply):
 800
 801 2005-01-08  Kevin Decker  <kdecker@apple.com>
 802
 803         Reviewed by Ken.
 804
 805         Fixed: <rdar://problem/3924219> Calling setOuterHTML: on a DOMHTMLHtmlElement can crash a program
 806
 807         * khtml/html/html_elementimpl.cpp:
 808         (HTMLElementImpl::setOuterHTML): Added a nil check for fragments who don't have parents.
 809
 810 2005-01-07  Maciej Stachowiak  <mjs@apple.com>
 811
 812         Reviewed by Darin.
 813
 814         <rdar://problem/3807144> REGRESSION (125-TOT): my bank's website doesn't work (Societe Generale, socgen.com)
 815
 816         * Khtml/khtml_part.cpp:
 817         (KHTMLPart::checkCompleted): If the document is NULL, assume this frame has
 818         not started loading yet so it could not possibly be finishing here...
 819         (KHTMLPart::stop): ...unless the part is explicitly stopped and there is still
 820         no document, in this case we must have hit an error or been loading a non-HTML
 821         frame.
 822         * khtml/khtml_part.h:
 823
 824 2005-01-08  David Harrison  <harrison@apple.com>
 825
 826         Reviewed by Maciej.
 827
 828         <rdar://problem/3943415> REGRESSION (Mail): double-clicking past word selects previous word instead of only blank space
 829
 830         * khtml/editing/selection.cpp:
 831         (khtml::Selection::validate):
 832         Tune word selections left/right choice to use right if on empty last line.
 833
 834 2005-01-07  David Harrison  <harrison@apple.com>
 835
 836         Reviewed by Darin.
 837
 838         <rdar://problem/3942619> AX: Support sentence ax attributes
 839
 840         Needed to use the unicode utilities properly.  Twas lame before.
 841
 842         * khtml/editing/visible_units.cpp:
 843         (khtml::previousBoundary):
 844         (khtml::nextBoundary):
 845         (khtml::startOfWord):
 846         (khtml::endOfWord):
 847         (khtml::previousWordPosition):
 848         (khtml::nextWordPosition):
 849         (khtml::startOfSentence):
 850         (khtml::endOfSentence):
 851         (khtml::previousSentencePosition):
 852         (khtml::nextSentencePosition):
 853         * kwq/KWQAccObject.mm:
 854         (-[KWQAccObject accessibilityAttributeNames]):
 855         (-[KWQAccObject accessibilityAttributeValue:]):
 856         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
 857         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
 858         * kwq/KWQTextUtilities.mm:
 859         (currentTextBreakLocaleID):
 860         (KWQFindSentenceBoundary):
 861         (KWQFindNextSentenceFromIndex):
 862
 863 2005-01-07  Ken Kocienda  <kocienda@apple.com>
 864
 865         Reviewed by Kevin
 866
 867         Fix for these bugs:
 868
 869         <rdar://problem/3939523> in some cases, text does not retain style info after pressing return twice
 870         <rdar://problem/3944492> after pressing return twice, text is bold when it shouldn't be
 871
 872         * khtml/editing/htmlediting.cpp:
 873         (khtml::InsertParagraphSeparatorCommand::setFullTypingStyleBeforeInsertion):
 874         Merge the typing style with the computed style for the current position. Fixes both bugs.
 875         * khtml/editing/htmlediting.h:
 876         * layout-tests/editing/inserting/insert-div-023-expected.txt: Added.
 877         * layout-tests/editing/inserting/insert-div-023.html: Added.
 878         * layout-tests/editing/inserting/insert-div-024-expected.txt: Added.
 879         * layout-tests/editing/inserting/insert-div-024.html: Added.
 880
 881 2005-01-07  David Hyatt  <hyatt@apple.com>
 882
 883         Fix for 3941364, make sure tables reset overflowWidth/Height when they lay out again.  Fixes the odd scrolling
 884         behavior on worldofwarcraft.com.
 885
 886         Reviewed by kevin
 887
 888         * khtml/rendering/render_table.cpp:
 889         (RenderTable::layout):
 890
 891 2005-01-06  David Hyatt  <hyatt@apple.com>
 892
 893         Fix for 3932418 and 3920998, assertion failures caused by stray inline content inside tables.  Bulletproof the hit testing to
 894         ignore inline flows in this case.
 895
 896         Reviewed by kevin
 897
 898         * khtml/rendering/render_block.cpp:
 899         (khtml::RenderBlock::nodeAtPoint):
 900
 901 2005-01-07  Ken Kocienda  <kocienda@apple.com>
 902
 903         Reviewed by Hyatt
 904
 905         Fix for this bug:
 906
 907         <rdar://problem/3848412> for forwarded message, tabbing to message view scrolls to bottom of view
 908
 909         * kwq/KWQKHTMLPart.mm:
 910         (KWQKHTMLPart::nextKeyViewInFrame): Don't scroll the focus node to visible if it is not
 911         in the document, or if it is not a descendent of the document element. In the case of the bug,
 912         since the selection has not yet been set up, the focus node passed here is the HTML element, and
 913         that does not yield a rect that is useful to us here. So now, in the case the bug mentions, we do nothing.
 914
 915 2005-01-06  Kevin Decker  <kdecker@apple.com>
 916
 917         Reviewed by mjs.
 918
 919         Fixed: <rdar://problem/3932215> REGRESSION (125-177): iFrame example at developer.apple.com fails in Safari
 920
 921         * khtml/khtml_part.cpp:
 922         (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.
 923
 924 === Safari-178 ===
 925
 926 2005-01-06  David Harrison  <harrison@apple.com>
 927
 928         Reviewed by Chris.
 929
 930         (addendum to previous checkin for this bug)
 931         <rdar://problem/3776056> AX: Editable HTML should not be viewed as AXStaticText
 932
 933         Fix line navigation.  Add AXUIElementForTextMarker.
 934
 935         * kwq/KWQAccObject.mm:
 936         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
 937         (-[KWQAccObject doAXUIElementForTextMarker:]):
 938         (-[KWQAccObject doAXLineForTextMarker:]):
 939         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
 940         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
 941         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
 942         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
 943         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
 944
 945 2005-01-06  Darin Adler  <darin@apple.com>
 946
 947         Reviewed by Ken.
 948
 949         - re-fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
 950
 951         (turns out the PLT regression was a false alarm)
 952
 953         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
 954         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
 955         much larger number.
 956         * khtml/html/htmlparser.cpp:
 957         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
 958         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
 959         to eliminate code that used ID_CLOSE_TAG for an array size.
 960         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
 961         that manages isindex to use deref instead of delete.
 962         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
 963         mistake of using ID_CLOSE_TAG for the array size too.
 964         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
 965         there and it would prevent custom tags from working. Added range check before using the forbidden
 966         tag array with the token ID since custom tags will use index values past the end of the array.
 967         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
 968         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
 969         createElement call is still here. Last time I left out a few form element types from this switch;
 970         fixed now.
 971         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
 972         document if getTagID fails; this creates a unique per-document ID.
 973
 974         * khtml/misc/htmltags.c: Regenerated.
 975         * khtml/misc/htmltags.h: Regenerated.
 976
 977         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
 978         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
 979         Also rewrote getTagName to work with the new scheme.
 980
 981 2005-01-06  David Harrison  <harrison@apple.com>
 982
 983         Fixed Panther build.  Also, do not advertize sentence support since it is incomplete.
 984
 985         * kwq/KWQAccObject.mm:
 986         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
 987         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
 988         (-[KWQAccObject accessibilityIsAttributeSettable:]):
 989
 990 2005-01-06  David Harrison  <harrison@apple.com>
 991
 992         Reviewed by Dave Hyatt.
 993
 994         <rdar://problem/3776056> AX: Editable HTML should not be viewed as AXStaticText
 995         <rdar://problem/3588548> AX: tabbing does not work correctly with the screen reader and a focused link; need AXFocusedUIElement to work
 996
 997         Many more AX attributes supported.  Numerous fixes to previous AX work.
 998
 999         * khtml/editing/visible_units.cpp:
1000         (khtml::startSentenceBoundary):
1001         (khtml::startOfSentence):
1002         (khtml::endSentenceBoundary):
1003         (khtml::endOfSentence):
1004         (khtml::previousSentencePositionBoundary):
1005         (khtml::previousSentencePosition):
1006         (khtml::nextSentencePositionBoundary):
1007         (khtml::nextSentencePosition):
1008         * khtml/editing/visible_units.h:
1009         * khtml/khtmlview.cpp:
1010         (KHTMLView::layout):
1011         * khtml/misc/helper.cpp:
1012         (khtml::findSentenceBoundary):
1013         (khtml::nextSentenceFromIndex):
1014         * khtml/misc/helper.h:
1015         * khtml/misc/htmltags.c:
1016         (hash_tag):
1017         (findTag):
1018         * khtml/rendering/render_container.cpp:
1019         (RenderContainer::removeChildNode):
1020         (RenderContainer::appendChildNode):
1021         (RenderContainer::insertChildNode):
1022         * khtml/rendering/render_object.cpp:
1023         (RenderObject::remove):
1024         * khtml/xml/dom_docimpl.cpp:
1025         (DocumentImpl::getAccObjectCache):
1026         (DocumentImpl::updateSelection):
1027         (DocumentImpl::close):
1028         (DocumentImpl::setFocusNode):
1029         (DocumentImpl::parentDocument):
1030         (DocumentImpl::topDocument):
1031         * khtml/xml/dom_docimpl.h:
1032         * kwq/KWQAccObject.mm:
1033         (-[KWQAccObject accessibilityShouldUseUniqueId]):
1034         (-[KWQAccObject detach]):
1035         (-[KWQAccObject anchorElement]):
1036         (-[KWQAccObject firstChild]):
1037         (-[KWQAccObject lastChild]):
1038         (-[KWQAccObject previousSibling]):
1039         (-[KWQAccObject nextSibling]):
1040         (-[KWQAccObject parentObject]):
1041         (-[KWQAccObject value]):
1042         (-[KWQAccObject accessibilityAttributeNames]):
1043         (-[KWQAccObject accessibilityPerformAction:]):
1044         (-[KWQAccObject textMarkerForVisiblePosition:]):
1045         (-[KWQAccObject visiblePositionForTextMarker:]):
1046         (-[KWQAccObject textMarkerRangeFromVisiblePositions:andEndPos:]):
1047         (-[KWQAccObject topDocument]):
1048         (-[KWQAccObject topRenderer]):
1049         (-[KWQAccObject topView]):
1050         (-[KWQAccObject accessibilityAttributeValue:]):
1051         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
1052         (-[KWQAccObject doAXLineForTextMarker:]):
1053         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
1054         (-[KWQAccObject doAXStringForTextMarkerRange:]):
1055         (-[KWQAccObject doAXTextMarkerForPosition:]):
1056         (-[KWQAccObject doAXBoundsForTextMarkerRange:]):
1057         (-[KWQAccObject doAXAttributedStringForTextMarkerRange:]):
1058         (-[KWQAccObject doAXTextMarkerRangeForUnorderedTextMarkers:]):
1059         (-[KWQAccObject doAXNextTextMarkerForTextMarker:]):
1060         (-[KWQAccObject doAXPreviousTextMarkerForTextMarker:]):
1061         (-[KWQAccObject doAXLeftWordTextMarkerRangeForTextMarker:]):
1062         (-[KWQAccObject doAXRightWordTextMarkerRangeForTextMarker:]):
1063         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
1064         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
1065         (-[KWQAccObject doAXSentenceTextMarkerRangeForTextMarker:]):
1066         (-[KWQAccObject doAXParagraphTextMarkerRangeForTextMarker:]):
1067         (-[KWQAccObject doAXNextWordEndTextMarkerForTextMarker:]):
1068         (-[KWQAccObject doAXPreviousWordStartTextMarkerForTextMarker:]):
1069         (-[KWQAccObject doAXNextLineEndTextMarkerForTextMarker:]):
1070         (-[KWQAccObject doAXPreviousLineStartTextMarkerForTextMarker:]):
1071         (-[KWQAccObject doAXNextSentenceEndTextMarkerForTextMarker:]):
1072         (-[KWQAccObject doAXPreviousSentenceStartTextMarkerForTextMarker:]):
1073         (-[KWQAccObject doAXNextParagraphEndTextMarkerForTextMarker:]):
1074         (-[KWQAccObject doAXPreviousParagraphStartTextMarkerForTextMarker:]):
1075         (-[KWQAccObject doAXLengthForTextMarkerRange:]):
1076         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
1077         (-[KWQAccObject accessibilityHitTest:]):
1078         (-[KWQAccObject accessibilityFocusedUIElement]):
1079         (-[KWQAccObject accessibilityIsAttributeSettable:]):
1080         (-[KWQAccObject doSetAXSelectedTextMarkerRange:]):
1081         (-[KWQAccObject setAccObjectID:]):
1082         (-[KWQAccObject removeAccObjectID]):
1083         * kwq/KWQAccObjectCache.h:
1084         * kwq/KWQAccObjectCache.mm:
1085         (KWQAccObjectCache::setAccObject):
1086         (KWQAccObjectCache::removeAccObject):
1087         (KWQAccObjectCache::visiblePositionForTextMarker):
1088         (KWQAccObjectCache::postNotificationToTopWebArea):
1089         (KWQAccObjectCache::postNotification):
1090         (KWQAccObjectCache::handleFocusedUIElementChanged):
1091         * kwq/KWQKHTMLPart.mm:
1092         (KWQKHTMLPart::respondToChangedContents):
1093         * kwq/KWQTextUtilities.h:
1094         * kwq/KWQTextUtilities.mm:
1095         (KWQFindNextWordFromIndex):
1096         (KWQFindSentenceBoundary):
1097         (KWQFindNextSentenceFromIndex):
1098         * kwq/WebCoreBridge.mm:
1099         (-[WebCoreBridge accessibilityTree]):
1100
1101 2005-01-05  Darin Adler  <darin@apple.com>
1102
1103         Reviewed by Ken.
1104
1105         - re-landing a subset of my custom tag change that does not fix the bug, but also does
1106           not introduce a performance regression
1107
1108         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
1109         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
1110
1111         * khtml/editing/htmlediting.cpp:
1112         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
1113         per-document tags and is just better all around for things like the document.
1114         (khtml::debugNode): Ditto.
1115         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
1116         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
1117         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
1118         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
1119
1120         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
1121         non-HTML elements to be nested as desired.
1122
1123         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
1124         * khtml/misc/htmlhashes.cpp:
1125         (khtml::getTagID): Changed return type to unsigned short.
1126         (khtml::getAttrID): Ditto.
1127
1128         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
1129         * khtml/xml/dom_docimpl.cpp:
1130         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
1131         Also updated for a few tags that the parser handled but this did not.
1132
1133         * kwq/KWQRenderTreeDebug.cpp:
1134         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
1135         the tag ID directly, which only works for standard nodes.
1136         (operator<<): Update to call getTagName.
1137         (nodePositionRelativeToRoot): Ditto.
1138         (writeSelection): Ditto.
1139
1140 2005-01-05  Ken Kocienda  <kocienda@apple.com>
1141
1142         Reviewed by Hyatt
1143
1144         Fix for this bug:
1145
1146         <rdar://problem/3941203> REGRESSION (Mail): Paste inserts content in wrong place
1147
1148         * khtml/editing/htmlediting.cpp:
1149         (khtml::ReplaceSelectionCommand::doApply): Some cleanup and refinement of the concepts used to make
1150         this operation work correctly, particularly in the logic to figure out whether to merge content, and
1151         also performing merges.
1152         * khtml/editing/visible_position.cpp:
1153         (khtml::isFirstVisiblePositionInBlock): Simplification of test used to make this determination.
1154         * khtml/editing/visible_units.cpp:
1155         (khtml::isStartOfParagraph): New helper, used in khtml::ReplaceSelectionCommand::doApply().
1156         (khtml::isEndOfParagraph): Ditto.
1157         * khtml/editing/visible_units.h: Declare new functions.
1158
1159 2005-01-04  Ken Kocienda  <kocienda@apple.com>
1160
1161         Reviewed by John
1162
1163         Fix for this bug:
1164
1165         <rdar://problem/3926522> Pressing return in a quoted block inserts too many newlines
1166
1167         * khtml/editing/htmlediting.cpp:
1168         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Call to insertBlockPlaceholderIfNeeded()
1169         for the block we are adding was being done at the wrong time, which led to the placeholder remaining
1170         in the document when it was not needed. This resulted in the extra space reported in the bug.
1171
1172 2005-01-04  Kevin Decker  <kdecker@apple.com>
1173
1174         Reviewed by Hyatt.
1175
1176         Fixed: <rdar://problem/3936879> nil-deref, crash in InlineFlowBox::nodeAtPoint closing DHTML menus at hrweb.apple.com
1177
1178         * khtml/rendering/render_block.cpp:
1179         (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.
1180
1181 2005-01-04  David Hyatt  <hyatt@apple.com>
1182
1183         Fix for 3937203, force an update of the dashboard regions when scrollbars are added/removed.
1184
1185         Reviewed by kevin
1186
1187         * khtml/rendering/render_layer.cpp:
1188         (RenderLayer::updateScrollInfoAfterLayout):
1189
1190 2005-01-04  Ken Kocienda  <kocienda@apple.com>
1191
1192         Reviewed by John
1193
1194         Fix for these two bugs:
1195
1196         <rdar://problem/3938935> REGRESSION (Mail): Pasting into an empty document mangles content
1197         <rdar://problem/3939148> REGRESSION (Mail): Pasting mistakenly reverses lines
1198
1199         * khtml/editing/htmlediting.cpp:
1200         (khtml::ReplaceSelectionCommand::doApply): For 3938935, add one more case to handle an empty document; merge
1201         neither start nor end. For 3939148, improve the code which adjusts the insertion point during
1202         the process of pasting. It formerly handled only one of the possible cases.
1203         * layout-tests/editing/pasteboard/paste-text-015-expected.txt: Added.
1204         * layout-tests/editing/pasteboard/paste-text-015.html: Added.
1205
1206 2005-01-04  David Hyatt  <hyatt@apple.com>
1207
1208         Fix for 3936571, placeholder attribute should work for normal inputs for Dashboard.
1209
1210         Reviewed by john
1211
1212         * khtml/rendering/render_form.cpp:
1213         (RenderLineEdit::updateFromElement):
1214         * kwq/KWQLineEdit.mm:
1215         (QLineEdit::setPlaceholderString):
1216
1217 2005-01-04  David Hyatt  <hyatt@apple.com>
1218
1219         Fix for 3830936, hang on changeforamerica.com.  Make sure to ignore the style not yet available option when
1220         returning pseudo-styles.
1221
1222         Reviewed by john
1223
1224         * khtml/css/cssstyleselector.cpp:
1225         (khtml::CSSStyleSelector::matchRulesForList):
1226         (khtml::CSSStyleSelector::pseudoStyleForElement):
1227         * khtml/rendering/render_style.cpp:
1228         (RenderStyle::addPseudoStyle):
1229
1230 2005-01-04  Darin Adler  <darin@apple.com>
1231
1232         - rolled out my custom tag name change again -- it caused a 1 ms PLT regression
1233
1234         * khtml/css/cssstyleselector.cpp:
1235         * khtml/editing/htmlediting.cpp:
1236         * khtml/editing/selection.cpp:
1237         * khtml/editing/visible_position.cpp:
1238         * khtml/html/dtd.cpp:
1239         * khtml/html/htmlparser.cpp:
1240         * khtml/html/htmlparser.h:
1241         * khtml/html/htmltokenizer.cpp:
1242         * khtml/misc/htmlhashes.cpp:
1243         * khtml/misc/htmlhashes.h:
1244         * khtml/misc/htmltags.c:
1245         * khtml/misc/htmltags.h:
1246         * khtml/misc/maketags:
1247         * khtml/xml/dom_docimpl.cpp:
1248         * khtml/xml/dom_docimpl.h:
1249         * khtml/xml/dom_nodeimpl.cpp:
1250         * khtml/xml/dom_position.cpp:
1251         * kwq/KWQRenderTreeDebug.cpp:
1252
1253 2005-01-04  Ken Kocienda  <kocienda@apple.com>
1254
1255         Reviewed by John
1256
1257         Fix for this bug:
1258
1259         <rdar://problem/3927554> Style info applied to remainder of document after a newline is entered
1260
1261         * khtml/editing/htmlediting.cpp:
1262         (khtml::InsertParagraphSeparatorCommand::doApply): Clean up and simplification in code that inserts
1263         a paragraph separator. The bug was all about applying styles to the new paragraph that did not need
1264         to be applied. Now the code will detect when at the end of a style run and will not move and apply
1265         that ending style to the new paragraph, though it will place that style into the typing style. This
1266         seems to match NSText behavior.
1267         * layout-tests/editing/inserting/insert-div-021-expected.txt: This test result exhibited the bug fixed
1268         here. The only reason it was not noticed is that the erroneously copied inline was a span, and so did
1269         not have any visible effect on the document.
1270         * layout-tests/editing/inserting/insert-div-022-expected.txt: Added.
1271         * layout-tests/editing/inserting/insert-div-022.html: Added.
1272
1273 2005-01-04  David Hyatt  <hyatt@apple.com>
1274
1275         Fix for 3904562, make sure to clear the outline dictionary when drawing so that random focus rings dont get drawn
1276         all over the place.
1277
1278         Reviewed by john
1279
1280         * khtml/rendering/render_flow.cpp:
1281         (RenderFlow::paintLines):
1282
1283 2005-01-03  David Hyatt  <hyatt@apple.com>
1284
1285         Fix for float painting regressions 3932524, 3931664, and 3933068.  Make the noPaint flag setting more
1286         robust and make it work regardless of which objects get a layout or not.
1287
1288         Reviewed by mjs
1289
1290         * khtml/rendering/render_block.cpp:
1291         (khtml::RenderBlock::insertFloatingObject):
1292         (khtml::RenderBlock::addOverhangingFloats):
1293         (khtml::RenderBlock::addIntrudingFloats):
1294         * khtml/rendering/render_block.h:
1295         (khtml::RenderBlock::FloatingObject::FloatingObject):
1296
1297 2005-01-03  Maciej Stachowiak  <mjs@apple.com>
1298
1299         Reviewed by Kevin.
1300
1301         <rdar://problem/3935390> Tiger 8A341: nil deref crash in DOM::NodeImpl::traverseNextNode
1302
1303         * khtml/html/html_miscimpl.cpp:
1304         (HTMLCollectionImpl::traverseNextItem): Assert that the starting point is not nil,
1305         it should not ever be (now).
1306         (HTMLCollectionImpl::item): When traversing items stop when we hit
1307         nil, meaning the end to avoid triggering above assert (formerly crash).
1308
1309 2005-01-03  Maciej Stachowiak  <mjs@apple.com>
1310
1311         Reviewed by John and Kevin.
1312
1313         <rdar://problem/3870317> REGRESSION(125.9-125.11) broken behavior at test.profoundlearning.com - used to
1314
1315         * khtml/ecma/kjs_window.cpp:
1316         (Window::get): Look up frame names before buitin window properties
1317         to match other browsers. This regressed because we added a builtin
1318         "toolbar" property but this site had a frame with that name.
1319
1320 2004-12-21  Maciej Stachowiak  <mjs@apple.com>
1321
1322         Reviewed by Darin.
1323
1324         <rdar://problem/3888931> frame naming allows malicious site to bring up a window when you click on a link in another
1325
1326         Added opener bridge method to help WebKit implement security check
1327         for named frame visibility.
1328
1329         * khtml/khtml_part.h:
1330         * kwq/WebCoreBridge.h:
1331         * kwq/WebCoreBridge.mm:
1332         (-[WebCoreBridge opener]):
1333
1334 2005-01-03  Ken Kocienda  <kocienda@apple.com>
1335
1336         Reviewed by John
1337
1338         Fix for this bug:
1339
1340         <rdar://problem/3933926> Tiger8A341: Mail crashes while forwarding embedded HTML message in -[WebCoreBridge ensureSelectionVisible]
1341
1342         * kwq/WebCoreBridge.mm:
1343         (-[WebCoreBridge ensureSelectionVisible]): Put in some null checks to prevent crash experienced in bug.
1344
1345 2005-01-03  David Hyatt  <hyatt@apple.com>
1346
1347         Fix for 3936881, make sure positioned objects prooperly update y-position.
1348
1349         Reviewed by john
1350
1351         * khtml/rendering/render_block.cpp:
1352         (khtml::RenderBlock::layoutPositionedObjects):
1353
1354 2005-01-03  Ken Kocienda  <kocienda@apple.com>
1355
1356         Reviewed by Harrison
1357
1358         Fix for this bug:
1359
1360         <rdar://problem/3928250> REGRESSION (Mail): Typing style lost after hitting return key
1361
1362         * khtml/editing/htmlediting.cpp:
1363         (khtml::InsertLineBreakCommand::preservesTypingStyle): Now implemented, returning yes for this command.
1364         (khtml::InsertParagraphSeparatorCommand::InsertParagraphSeparatorCommand): Initialize new typing style member variable.
1365         (khtml::InsertParagraphSeparatorCommand::~InsertParagraphSeparatorCommand): Deref new typing style member variable.
1366         (khtml::InsertParagraphSeparatorCommand::preservesTypingStyle): Now implemented, returning yes for this command.
1367         (khtml::InsertParagraphSeparatorCommand::setFullTypingStyleBeforeInsertion): New function to set the typing style
1368         (khtml::InsertParagraphSeparatorCommand::calculateAndSetTypingStyleAfterInsertion): Function called after the <p>
1369         insertion is done. This function diffs the style created in setFullTypingStyleBeforeInsertion() with the style
1370         of the new <p> and only sets those styles needed to preserve the style in effect before the insertion.
1371         (khtml::InsertParagraphSeparatorCommand::doApply): Call new functions.
1372         (khtml::TypingCommand::preservesTypingStyle): Now yes for inserting line breaks and paragraphs.
1373          * khtml/editing/htmlediting.h: Declare new functions.
1374
1375 2004-12-25  Kevin Decker  <kdecker@apple.com>
1376
1377         Reviewed by Hyatt.
1378
1379         Fixed: <rdar://problem/3505072> hang in KHTMLParser::parseToken (consulting.soroos.net)
1380
1381         * khtml/html/dtd.cpp: <label> needed a much higher priority such that it will close block elements.  Malformed HTML (ugh) is the only reason why we need to do this.
1382
1383 2004-12-23  Darin Adler  <darin@apple.com>
1384
1385         Reviewed by Ken.
1386
1387         - re-fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
1388
1389         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
1390         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
1391
1392         * khtml/editing/htmlediting.cpp:
1393         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
1394         per-document tags and is just better all around for things like the document.
1395         (khtml::debugNode): Ditto.
1396         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
1397         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
1398         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
1399         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
1400
1401         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
1402         non-HTML elements to be nested as desired.
1403
1404         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
1405         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
1406         much larger number.
1407         * khtml/html/htmlparser.cpp:
1408         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
1409         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
1410         to eliminate code that used ID_CLOSE_TAG for an array size.
1411         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
1412         that manages isindex to use deref instead of delete.
1413         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
1414         mistake of using ID_CLOSE_TAG for the array size too.
1415         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
1416         there and it would prevent custom tags from working. Added range check before using the forbidden
1417         tag array with the token ID since custom tags will use index values past the end of the array.
1418         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
1419         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
1420         createElement call is still here. Last time I left out a few form element types from this switch;
1421         fixed now.
1422         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
1423         document if getTagID fails; this creates a unique per-document ID.
1424
1425         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
1426         * khtml/misc/htmlhashes.cpp:
1427         (khtml::getTagID): Changed return type to unsigned short.
1428         (khtml::getAttrID): Ditto.
1429
1430         * khtml/misc/htmltags.c: Regenerated.
1431         * khtml/misc/htmltags.h: Regenerated.
1432
1433         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
1434         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
1435         Also rewrote getTagName to work with the new scheme.
1436
1437         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
1438         * khtml/xml/dom_docimpl.cpp:
1439         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
1440         Also updated for a few tags that the parser handled but this did not.
1441
1442         * kwq/KWQRenderTreeDebug.cpp:
1443         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
1444         the tag ID directly, which only works for standard nodes.
1445         (operator<<): Update to call getTagName.
1446         (nodePositionRelativeToRoot): Ditto.
1447         (writeSelection): Ditto.
1448
1449 === Safari-177 ===
1450
1451 2004-12-22  Darin Adler  <darin@apple.com>
1452
1453         - rolled out my custom tag name change -- it broke amazon.com
1454
1455         * khtml/css/cssstyleselector.cpp:
1456         * khtml/editing/htmlediting.cpp:
1457         * khtml/editing/selection.cpp:
1458         * khtml/editing/visible_position.cpp:
1459         * khtml/html/dtd.cpp:
1460         * khtml/html/htmlparser.cpp:
1461         * khtml/html/htmlparser.h:
1462         * khtml/html/htmltokenizer.cpp:
1463         * khtml/misc/htmlhashes.cpp:
1464         * khtml/misc/htmlhashes.h:
1465         * khtml/misc/htmltags.c:
1466         * khtml/misc/htmltags.h:
1467         * khtml/misc/maketags:
1468         * khtml/xml/dom_docimpl.cpp:
1469         * khtml/xml/dom_docimpl.h:
1470         * khtml/xml/dom_nodeimpl.cpp:
1471         * khtml/xml/dom_position.cpp:
1472         * kwq/KWQRenderTreeDebug.cpp:
1473
1474 2004-12-22  David Harrison  <harrison@apple.com>
1475
1476         Reviewed by Darin Adler.
1477
1478         * khtml/editing/selection.cpp:
1479         (khtml::Selection::validate):
1480         The selecting/deselecting bad behavior is because the Selection code that expands by words
1481         had an inaccurate test for being at the end of the document (where double-clicking needs
1482         to select the last word).  Fixed that check.
1483
1484 2004-12-22  Adele Amchan  <adele@apple.com>
1485
1486         Reviewed by Chris.
1487
1488         Fix for <rdar://problem/3911650> tabs at safeway.com stop working after a while
1489
1490         * kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::scrollToAnchor): removed call to cancelRedirection
1491           so that we match Firefox and WinIE behavior.
1492
1493 2004-12-22  Darin Adler  <darin@apple.com>
1494
1495         Reviewed by Ken.
1496
1497         - fixed <rdar://problem/3760910> Request to include support for custom tag names in HTML (they already work in XML)
1498
1499         * khtml/css/cssstyleselector.cpp: (khtml::CSSStyleSelector::checkOneSelector): Changed
1500         some code that used ID_LAST_TAG in a slightly wrong, but harmless, way.
1501
1502         * khtml/editing/htmlediting.cpp:
1503         (khtml::debugPosition): Use nodeName rather than getTagName, since the latter works for
1504         per-document tags and is just better all around for things like the document.
1505         (khtml::debugNode): Ditto.
1506         * khtml/editing/selection.cpp: (khtml::Selection::debugPosition): Ditto.
1507         * khtml/editing/visible_position.cpp: (khtml::VisiblePosition::debugPosition): Ditto.
1508         * khtml/xml/dom_nodeimpl.cpp: (NodeImpl::displayNode): Ditto.
1509         * khtml/xml/dom_position.cpp: (DOM::Position::debugPosition): Ditto.
1510
1511         * khtml/html/dtd.cpp: (DOM::checkChild): Use ID_LAST_TAG rather than 1000 for the check that allows
1512         non-HTML elements to be nested as desired.
1513
1514         * khtml/html/htmlparser.h: Change forbidden tag array to be allocated in the object rather than on
1515         the heap. Also use ID_LAST_TAG for the array size; the old code used ID_CLOSE_TAG which is now a
1516         much larger number.
1517         * khtml/html/htmlparser.cpp:
1518         (KHTMLParser::KHTMLParser): Remove code that creates the forbidden tag array on the heap; makes more
1519         sense to just have the array be a member so we don't have to use new and delete on it. Also needed
1520         to eliminate code that used ID_CLOSE_TAG for an array size.
1521         (KHTMLParser::~KHTMLParser): Remove code that deletes the forbidden tag array. Also changed the code
1522         that manages isindex to use deref instead of delete.
1523         (KHTMLParser::reset): Use safer sizeof for clearing the forbidden tag array. Old code had the
1524         mistake of using ID_CLOSE_TAG for the array size too.
1525         (KHTMLParser::parseToken): Removed code that checks for bogus tags; it's not clear why it was ever
1526         there and it would prevent custom tags from working. Added range check before using the forbidden
1527         tag array with the token ID since custom tags will use index values past the end of the array.
1528         (KHTMLParser::getElement): Removed most of the code that makes an element by ID; now the work is done
1529         inside the DocumentImpl createHTMLElement. Any code that needs to be different than a dynamic
1530         createElement call is still here.
1531         * khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Call the tagId function in the
1532         document if getTagID fails; this creates a unique per-document ID.
1533
1534         * khtml/misc/htmlhashes.h: Changed return types to unsigned short.
1535         * khtml/misc/htmlhashes.cpp:
1536         (khtml::getTagID): Changed return type to unsigned short.
1537         (khtml::getAttrID): Ditto.
1538
1539         * khtml/misc/htmltags.c: Regenerated.
1540         * khtml/misc/htmltags.h: Regenerated.
1541
1542         * khtml/misc/maketags: Changed the generated constants to use "const unsigned short" instead of "#define".
1543         Changed ID_CLOSE_TAG to be a fixed value of 32000 rather than being just past the IDs of the HTML tags.
1544         Also rewrote getTagName to work with the new scheme.
1545
1546         * khtml/xml/dom_docimpl.h: Added overload of createHTMLElement, made it non-virtual since it's not overriden.
1547         * khtml/xml/dom_docimpl.cpp:
1548         (DocumentImpl::createHTMLElement): Refactored into two separate functions, one that takes the tag ID.
1549         Also updated for a few tags that the parser handled but this did not.
1550
1551         * kwq/KWQRenderTreeDebug.cpp:
1552         (getTagName): Added. Works for custom nodes, because it calls nodeName rather than using getTagName on
1553         the tag ID directly, which only works for standard nodes.
1554         (operator<<): Update to call getTagName.
1555         (nodePositionRelativeToRoot): Ditto.
1556         (writeSelection): Ditto.
1557
1558 2004-12-21  David Harrison  <harrison@apple.com>
1559
1560         Reviewed by Ken Kocienda.
1561
1562         <rdar://problem/3924934> REGRESSION: double click at end of line selects start of next line
1563
1564         Problem was the TextIterator was not handling exitNode() from a P block properly.
1565
1566         * khtml/editing/visible_text.cpp:
1567         (khtml::TextIterator::TextIterator):
1568         Add new param that specifies whether the iterator is for content or for searching.
1569         Search iterators do not prevent newlines at the beginning.
1570         (khtml::TextIterator::advance):
1571         Added some comments.
1572         (khtml::TextIterator::handleTextNode):
1573         Added some comments.
1574         (khtml::TextIterator::exitNode):
1575         Emit newline for P (and other) blocks with position following the block, instead of the m_lastTextNode.
1576         (khtml::TextIterator::emitCharacter):
1577         Added some comments.
1578         (khtml::TextIterator::range):
1579         Added some comments.
1580         (khtml::CharacterIterator::CharacterIterator):
1581         Specify search type TextIterator.
1582         (khtml::CharacterIterator::advance):
1583         * khtml/editing/visible_text.h:
1584         (khtml::):
1585         Add new TextIterator::TextIterator param that specifies whether the iterator is for content or for searching.
1586         * khtml/editing/visible_units.cpp:
1587         (khtml::nextWordBoundary):
1588         Specify search type TextIterator.
1589
1590 2004-12-21  David Harrison  <harrison@apple.com>
1591
1592         Reviewed by Ken Kocienda.
1593
1594         <rdar://problem/3924695> REGRESSION (Mail): double-clicking past end of line shows no selection, should select to EOL
1595
1596         Problem was that RenderText::setSelectionState did not handle the SelectionStart case where start and end are the end of the line.
1597         Fixed by pretending the start == end-1 in that situation, as long as end > 0.
1598
1599         * khtml/rendering/render_text.cpp:
1600         (RenderText::setSelectionState)
1601
1602 2004-12-21  Maciej Stachowiak  <mjs@apple.com>
1603
1604         Reviewed by John.
1605
1606         <rdar://problem/3929187> WebKit needs to restrict access to certain window operations by domain
1607
1608         * khtml/ecma/kjs_window.cpp:
1609         (Window::get): Change most window functions to be restricted by
1610         XSS domain check.
1611
1612 2004-12-21  Ken Kocienda  <kocienda@apple.com>
1613
1614         Reviewed by John
1615
1616         Fix for this bug:
1617
1618         <rdar://problem/3928305> selecting an entire line and typing over causes new inserted text at top of document
1619
1620         * khtml/editing/htmlediting.cpp:
1621         (khtml::DeleteSelectionCommand::insertPlaceholderForAncestorBlockContent): New function to detect case
1622         formerly undetected and unhandled. This is the crux of the bug fix.
1623         (khtml::DeleteSelectionCommand::doApply): Call insertPlaceholderForAncestorBlockContent() during
1624         execution of command.
1625         * khtml/editing/htmlediting.h: Declare new function.
1626         * layout-tests/editing/deleting/delete-3928305-fix-expected.txt: Added.
1627         * layout-tests/editing/deleting/delete-3928305-fix.html: Added.
1628
1629 2004-12-21  Ken Kocienda  <kocienda@apple.com>
1630
1631         Reviewed by me
1632
1633         * khtml/editing/htmlediting.cpp:
1634         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Note to self: Must compile code before
1635         checking in (aka must return false from function returning bool).
1636
1637 2004-12-21  Ken Kocienda  <kocienda@apple.com>
1638
1639         Reviewed by John
1640
1641         Fix for this bug:
1642
1643         <rdar://problem/3927752> Crash in khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded(DOM::NodeImpl*)
1644
1645         * khtml/editing/htmlediting.cpp:
1646         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Added some null checks.
1647         (khtml::CompositeEditCommand::removeBlockPlaceholderIfNeeded): Ditto.
1648
1649 2004-12-21  Ken Kocienda  <kocienda@apple.com>
1650
1651         Reviewed by Darin
1652
1653         * khtml/editing/htmlediting.cpp:
1654         (khtml::ReplacementFragment::mergeStartNode): Refine concept of how this node is found based on
1655         further experiements.
1656         (khtml::ReplaceSelectionCommand::doApply): Add a special case for determining merges that need to
1657         be done if the insertion point is in an empty block.
1658         * layout-tests/editing/pasteboard/paste-text-012-expected.txt: Added.
1659         * layout-tests/editing/pasteboard/paste-text-012.html: Added.
1660         * layout-tests/editing/pasteboard/paste-text-013-expected.txt: Added.
1661         * layout-tests/editing/pasteboard/paste-text-013.html: Added.
1662         * layout-tests/editing/pasteboard/paste-text-014-expected.txt: Added.
1663         * layout-tests/editing/pasteboard/paste-text-014.html: Added.
1664
1665 2004-12-21  Darin Adler  <darin@apple.com>
1666
1667         Reviewed by Ken.
1668
1669         - fixed <rdar://problem/3899133> text search in a Safari window takes a very long time on Tiger updates page (and some other pages)
1670
1671         * khtml/editing/visible_text.h: Add an "offset base node" parameter to emitCharacter, and also
1672         add a field to track it. Must make a few things mutable so we can update them in the range accessor.
1673         * khtml/editing/visible_text.cpp:
1674         (khtml::TextIterator::advance): Pass in base node and offsets rather than computing actual offsets
1675         using the node's index. We only compute the node index if actually asked for the range.
1676         (khtml::TextIterator::handleTextNode): Pass 0 for base node and set base node to 0 when setting
1677         up the offsets.
1678         (khtml::TextIterator::handleTextBox): Ditto.
1679         (khtml::TextIterator::handleReplacedElement): Pass base node and set base node instead of calling
1680         nodeIndex.
1681         (khtml::TextIterator::handleNonTextNode): Pass 0 for offset.
1682         (khtml::TextIterator::exitNode): More of the same.
1683         (khtml::TextIterator::emitCharacter): Ditto.
1684         (khtml::TextIterator::range): If an offset base node is stored, then get its node index, and then
1685         add that in to the offsets. Doing the work here guarantees it's done only once when doing a text search.
1686
1687         - another small fix
1688
1689         * khtml/khtml_part.cpp: Removed SPEED_DEBUG define. Not sure why it was on.
1690
1691 === Safari-176 ===
1692
1693 2004-12-20  Ken Kocienda  <kocienda@apple.com>
1694
1695         Reviewed by Darin
1696
1697         * khtml/editing/htmlediting.cpp:
1698         (khtml::DeleteSelectionCommand::moveNodesAfterNode): My one-liner for this bug introduced layout test
1699         regressions: <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
1700         Rolling out until I can develop a real fix.
1701
1702 2004-12-20  David Harrison  <harrison@apple.com>
1703
1704         Reviewed by Dave Hyatt.
1705
1706         Initial checkin of AXTextMarkerRef support.
1707
1708         * khtml/xml/dom_docimpl.cpp:
1709         (DocumentImpl::updateSelection):
1710         (DocumentImpl::setFocusNode):
1711         * kwq/KWQAccObject.h:
1712         * kwq/KWQAccObject.mm:
1713         (-[KWQAccObject detach]):
1714         (-[KWQAccObject anchorElement]):
1715         (-[KWQAccObject addChildrenToArray:]):
1716         (-[KWQAccObject accessibilityAttributeNames]):
1717         (-[KWQAccObject accessibilityActionDescription:]):
1718         (-[KWQAccObject accessibilityPerformAction:]):
1719         (-[KWQAccObject textMarkerRangeFromMarkers:andEndMarker:]):
1720         (-[KWQAccObject textMarkerForVisiblePosition:]):
1721         (-[KWQAccObject visiblePositionForTextMarker:]):
1722         (-[KWQAccObject AXTextMarkerRangeCopyStartMarkerWrapper:]):
1723         (-[KWQAccObject AXTextMarkerRangeCopyEndMarkerWrapper:]):
1724         (-[KWQAccObject visiblePositionForStartOfTextMarkerRange:]):
1725         (-[KWQAccObject visiblePositionForEndOfTextMarkerRange:]):
1726         (-[KWQAccObject accessibilityAttributeValue:]):
1727         (-[KWQAccObject accessibilityParameterizedAttributeNames]):
1728         (-[KWQAccObject textMarkerRangeFromVisiblePositions:andEndPos:]):
1729         (-[KWQAccObject getSelectedTextMarkerRange]):
1730         (-[KWQAccObject doAXLineForTextMarker:]):
1731         (-[KWQAccObject doAXTextMarkerRangeForLine:]):
1732         (-[KWQAccObject doAXStringForTextMarkerRange:]):
1733         (-[KWQAccObject doAXNextTextMarkerForTextMarker:]):
1734         (-[KWQAccObject doAXPreviousTextMarkerForTextMarker:]):
1735         (-[KWQAccObject doAXLeftWordTextMarkerRangeForTextMarker:]):
1736         (-[KWQAccObject doAXRightWordTextMarkerRangeForTextMarker:]):
1737         (-[KWQAccObject doAXLeftLineTextMarkerRangeForTextMarker:]):
1738         (-[KWQAccObject doAXRightLineTextMarkerRangeForTextMarker:]):
1739         (-[KWQAccObject accessibilityAttributeValue:forParameter:]):
1740         (-[KWQAccObject accessibilityFocusedUIElement]):
1741         (-[KWQAccObject clearChildren]):
1742         (-[KWQAccObject accObjectID]):
1743         (-[KWQAccObject setAccObjectID:]):
1744         (-[KWQAccObject removeAccObjectID]):
1745         * kwq/KWQAccObjectCache.h:
1746         * kwq/KWQAccObjectCache.mm:
1747         (KWQAccObjectCache::KWQAccObjectCache):
1748         (KWQAccObjectCache::~KWQAccObjectCache):
1749         (KWQAccObjectCache::getAccObjectID):
1750         (KWQAccObjectCache::removeAccObjectID):
1751         (KWQAccObjectCache::textMarkerForVisiblePosition):
1752         (KWQAccObjectCache::visiblePositionForTextMarker):
1753
1754 2004-12-19  Darin Adler  <darin@apple.com>
1755
1756         * kwq/KWQFoundationExtras.h: (KWQCFAutorelease): Added nil check.
1757
1758 2004-12-19  Darin Adler  <darin@apple.com>
1759
1760         Reviewed by Kevin.
1761
1762         - a garbage collection fix
1763
1764         * kwq/KWQFoundationExtras.h: (KWQCFAutorelease): Added a more-extensive comment and fixed the BUILDING_ON_PANTHER
1765         #ifndef to use #if (in practice, either is OK).
1766         * kwq/KWQKURL.mm: (KURL::getNSURL): Use KWQCFAutorelease instead of autorelease.
1767
1768 2004-12-17  David Hyatt  <hyatt@apple.com>
1769
1770         Fix for 3923255, specified percentage heights of divs with overflow auto inside tables not honored.
1771
1772         Reviewed by kocienda
1773
1774         * khtml/rendering/render_box.cpp:
1775         (RenderBox::calcPercentageHeight):
1776         * khtml/rendering/render_table.cpp:
1777         (RenderTableSection::layoutRows):
1778
1779 2004-12-17  David Harrison  <harrison@apple.com>
1780
1781         Reviewed by Ken Kocienda.
1782
1783         <rdar://problem/3924930> REGRESSION: triple click does not select to end of line
1784
1785         * khtml/editing/visible_units.cpp:
1786         (khtml::endOfParagraph):
1787         When includeLineBreak is true, allow traversal to next node after enclosingBlockFlowElement.
1788
1789 2004-12-17  Ken Kocienda  <kocienda@apple.com>
1790
1791         Reviewed by Vicki
1792
1793         * khtml/editing/htmlediting.cpp:
1794         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Missed deleting some now-obsolete expcetion
1795         codes when I added the createBreakElement() function.
1796
1797 2004-12-17  Richard Williamson   <rjw@apple.com>
1798
1799         Set the floor of max cacheable object size to 40K.  This restores
1800         the long standing floor.  Lower floor deleteriously impacts the PLT.
1801         Reviewed by Hyatt.
1802
1803         * khtml/khtml_part.cpp:
1804         (KHTMLPart::checkCompleted):
1805         * khtml/misc/loader.cpp:
1806         (Cache::setSize):
1807
1808 2004-12-17  Ken Kocienda  <kocienda@apple.com>
1809
1810         Reviewed by Hyatt
1811
1812         Fix for this bug:
1813
1814         <rdar://problem/3926142> REGRESSION (Mail): Deleting text decreases quote level
1815
1816         * khtml/editing/htmlediting.cpp:
1817         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Stop merging nodes when a <br> is hit.
1818         Formerly checked only for block boundary.
1819
1820 2004-12-17  David Harrison  <harrison@apple.com>
1821
1822         Reviewed by Darin.
1823
1824                 Add KWQCFAutorelease for autoreleasing CF objects.
1825         * kwq/KWQFoundationExtras.h:
1826         (KWQCFAutorelease):
1827         New.
1828
1829 2004-12-17  David Harrison  <harrison@apple.com>
1830
1831         Reviewed by Darin.
1832
1833                 Fix GC compatibility in getNSString.
1834
1835         * kwq/KWQString.mm:
1836         (QString::getNSString):
1837                 Use NSString allocator instead of CFString, so that autorelease works under GC.
1838
1839 2004-12-17  Ken Kocienda  <kocienda@apple.com>
1840
1841         Reviewed by Hyatt
1842
1843         Fix for this bug:
1844
1845         <rdar://problem/3890973> REGRESSION (Mail): Deleting reorders remaining text
1846
1847         * khtml/editing/htmlediting.cpp:
1848         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Added check to ensure that moving content will not
1849         move it after the <body> element.
1850
1851 2004-12-17  Ken Kocienda  <kocienda@apple.com>
1852
1853         Reviewed by me
1854
1855         Added new layout tests covering cases from recent bug fixes.
1856
1857         * layout-tests/editing/inserting/insert-div-018-expected.txt: Added.
1858         * layout-tests/editing/inserting/insert-div-018.html: Added.
1859         * layout-tests/editing/inserting/insert-div-019-expected.txt: Added.
1860         * layout-tests/editing/inserting/insert-div-019.html: Added.
1861         * layout-tests/editing/inserting/insert-div-020-expected.txt: Added.
1862         * layout-tests/editing/inserting/insert-div-020.html: Added.
1863         * layout-tests/editing/inserting/insert-div-021-expected.txt: Added.
1864         * layout-tests/editing/inserting/insert-div-021.html: Added.
1865
1866 2004-12-17  Ken Kocienda  <kocienda@apple.com>
1867
1868         Reviewed by me
1869
1870         Changes in layout resulting from giving <p> elements no margin rather than 0.1em margin.
1871
1872         * layout-tests/editing/inserting/insert-div-013-expected.txt
1873         * layout-tests/editing/inserting/insert-div-014-expected.txt
1874
1875 2004-12-17  Ken Kocienda  <kocienda@apple.com>
1876
1877         Reviewed by John
1878
1879         Fix for this bug:
1880
1881         <rdar://problem/3925317> Increase quote in email and with cursor below the quote the new line appears above the quote
1882
1883         * khtml/editing/htmlediting.cpp:
1884         (khtml::InsertParagraphSeparatorCommand::doApply):
1885         * khtml/editing/visible_position.cpp: Refine rules for whether to use the starting node or the starting
1886         block as the reference node for the insertion of the new block.
1887         (khtml::isLastVisiblePositionInBlock): Tweak rules again. Descendants of following blocks should answer true.
1888         I worked this all out on the whiteboard this time. This should be the last tweak.
1889
1890 2004-12-17  Kevin Decker  <kdecker@apple.com>
1891
1892         Reviewed by Ken.
1893
1894         Fixed <rdar://problem/3824438> Need a clean way for Dashboard to detect when an XML parsing error occurs
1895
1896         * khtml/xml/xml_tokenizer.cpp:
1897         (khtml::XMLTokenizer::insertErrorMessageBlock): In the xml error report, instead of a generic <div>, use <parsererror> to match Mozilla.
1898
1899 2004-12-16  Ken Kocienda  <kocienda@apple.com>
1900
1901         Reviewed by Chris
1902
1903         Fix for this bug:
1904
1905         <rdar://problem/3924888> REGRESSION (Mail): Hitting return key at end of line does not insert visible newline
1906
1907         * khtml/editing/htmlediting.cpp:
1908         (khtml::InsertParagraphSeparatorCommand::doApply): Tweaked code I added just before to fix 3924486, so that
1909         it checks the downstream node for the starting position.
1910
1911 2004-12-16  Ken Kocienda  <kocienda@apple.com>
1912
1913         Reviewed by Chris
1914
1915         Added new createBreakElement() function that makes creating a <br> element a one-liner.
1916         Converted all the code that used to call the DocumentImpl, and dealt with the exceptionCode,
1917         over to this new helper.
1918
1919         * khtml/editing/htmlediting.cpp:
1920         (khtml::DeleteSelectionCommand::moveNodesAfterNode): Now uses new helper function.
1921         (khtml::InsertLineBreakCommand::doApply): Ditto.
1922         (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
1923         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Ditto.
1924         (khtml::ReplaceSelectionCommand::doApply): Ditto.
1925         (khtml::createBreakElement): New helper.
1926         * khtml/editing/htmlediting.h: Ditto.
1927
1928 2004-12-16  David Hyatt  <hyatt@apple.com>
1929
1930         Fix a bug where vertical-align values that depended on their parent's value went crazy if the parent was
1931         vertical-align top or bottom.  The bug is 3771007, bankofamerica's ebills page.
1932
1933         Reviewed by kocienda
1934
1935         * khtml/rendering/render_object.cpp:
1936         (RenderObject::getVerticalPosition):
1937
1938 2004-12-16  Ken Kocienda  <kocienda@apple.com>
1939
1940         Reviewed by John
1941
1942         Fix for these bugs:
1943
1944         <rdar://problem/3924486> REGRESSION (Mail): Hitting return key does not insert visible newline
1945         <rdar://problem/3924579> REGRESSION (Mail): After deleting, hitting return key does not insert visible newline
1946
1947         * khtml/editing/htmlediting.cpp:
1948         (khtml::InsertParagraphSeparatorCommand::doApply): For the first bug, detect when inserting a <p> will
1949         make a visible <br> collapse. For the second bug, move the code that inserts the new <p> after
1950         a check which may move the node used as the reference node for the insertion. Also change this code to
1951         insert the <p> after the last sibling of the starting position, to move it past the nodes we're going
1952         to want to shift into the new <p>.
1953
1954 2004-12-16  Darin Adler  <darin@apple.com>
1955
1956         Reviewed by Ken.
1957
1958         - use <p> elements rather than <br> when pasting plain text; also use margin of 0 rather than 0.1 em
1959           after talking this over with Dave and Ken
1960
1961         * khtml/editing/htmlediting.h: Added createDefaultParagraphElement and createBlockPlaceholderElement.
1962         * khtml/editing/htmlediting.cpp:
1963         (khtml::CompositeEditCommand::insertBlockPlaceholderIfNeeded): Use new createBlockPlaceholderElement
1964         helper so this can share code with the plain-text conversion code.
1965         (khtml::InsertParagraphSeparatorCommand::createParagraphElement): Use new createDefaultParagraphElement
1966         helper so this can share code with the plain-text conversion code.
1967         (khtml::createDefaultParagraphElement): Added.
1968         (khtml::createBlockPlaceholderElement): Added.
1969
1970         * khtml/editing/markup.cpp: (khtml::createFragmentFromText): Use paragraphs rather than
1971         <br> elements for pasting plain text.
1972
1973         - regenerated these files with the newer gperf
1974
1975         * khtml/css/cssproperties.c: Regenerated.
1976         * khtml/css/cssvalues.c: Regenerated.
1977         * khtml/misc/htmlattrs.c: Regenerated.
1978         * khtml/misc/htmltags.c: Regenerated.
1979
1980 2004-12-16  Ken Kocienda  <kocienda@apple.com>
1981
1982         Reviewed by Hyatt
1983
1984         Fix for this bug:
1985
1986         <rdar://problem/3924291> REGRESSION (Mail): Crash deleting content following a <p> element
1987
1988         * khtml/xml/dom_nodeimpl.cpp:
1989         (NodeImpl::enclosingInlineElement): Fixed problem in this function where recursive search for
1990         parent that is not an inline would skip past previous siblings of nodes which were blocks. Now
1991         it stops looking in this situation. This problem eventually caused the delete code to try to
1992         merge a node under a descendent.
1993
1994 2004-12-16  Adele Amchan  <adele@apple.com>
1995
1996         Change by Richard, reviewed by me.
1997
1998         Fix for: <rdar://problem/3923983> background image decoding prevents page with 0Kb image from finishing to load
1999
2000         * khtml/misc/loader.cpp: (CachedImage::data): Added call to notifyFinished so the 0Kb image case finishes loading.
2001
2002 2004-12-16  Adele Amchan  <adele@apple.com>
2003
2004         Reviewed by Chris.
2005
2006         Fix for: <rdar://problem/3534824> VIP: some navigation links on safeway.com don't work due to a JavaScript quirk
2007
2008         Added a call to ObjectImp::get before Window::get just returns Undefined.
2009         This was preventing us from getting the toString function from a Window object.
2010
2011         * khtml/ecma/kjs_window.cpp: (Window::get):
2012
2013 2004-12-16  Ken Kocienda  <kocienda@apple.com>
2014
2015         Reviewed by John
2016
2017         Fix for this bug:
2018
2019         <rdar://problem/3918351> REGRESSION (Mail, 173-175+): Return before first char of line leaves insertion point in wrong place
2020
2021         * khtml/editing/htmlediting.cpp:
2022         (khtml::InsertParagraphSeparatorCommand::doApply): Basically, did a rewrite of this function
2023         to do a better job than it was doing before. Added several test cases to prove I am on a
2024         better track.
2025         * khtml/editing/visible_position.cpp:
2026         (khtml::isFirstVisiblePositionInBlock): Tweaked the rules a bit to fix an issue very similar to the
2027         leaving-the-bar-node case problem I just fixed in a recent checkin (relevant markup: <p>foo</p>bar).
2028         This function was returning true for the first position in "bar". Wrong. Also tightened up other
2029         rule: Should not report true when relationship between blocks cannot be determined.
2030         (khtml::isLastVisiblePositionInBlock): Tightened up rule as above: Should not report true
2031         when relationship between blocks cannot be determined.
2032         * layout-tests/editing/inserting/insert-div-010-expected.txt: Added.
2033         * layout-tests/editing/inserting/insert-div-010.html: Added.
2034         * layout-tests/editing/inserting/insert-div-011-expected.txt: Added.
2035         * layout-tests/editing/inserting/insert-div-011.html: Added.
2036         * layout-tests/editing/inserting/insert-div-012-expected.txt: Added.
2037         * layout-tests/editing/inserting/insert-div-012.html: Added.
2038         * layout-tests/editing/inserting/insert-div-013-expected.txt: Added.
2039         * layout-tests/editing/inserting/insert-div-013.html: Added.
2040         * layout-tests/editing/inserting/insert-div-014-expected.txt: Added.
2041         * layout-tests/editing/inserting/insert-div-014.html: Added.
2042         * layout-tests/editing/inserting/insert-div-015-expected.txt: Added.
2043         * layout-tests/editing/inserting/insert-div-015.html: Added.
2044         * layout-tests/editing/inserting/insert-div-016-expected.txt: Added.
2045         * layout-tests/editing/inserting/insert-div-016.html: Added.
2046         * layout-tests/editing/inserting/insert-div-017-expected.txt: Added.
2047         * layout-tests/editing/inserting/insert-div-017.html: Added.
2048
2049 2004-12-16  Ken Kocienda  <kocienda@apple.com>
2050
2051         Reviewed by me
2052
2053         Added a layout test based on my last checkin.
2054
2055         * layout-tests/editing/selection/move-by-word-001-expected.txt: Added.
2056         * layout-tests/editing/selection/move-by-word-001.html: Added.
2057
2058 2004-12-16  Ken Kocienda  <kocienda@apple.com>
2059
2060         Reviewed by John
2061
2062         Fix for this bug:
2063
2064         <rdar://problem/3917929> REGRESSION (Mail): Command-left-arrow leaves insertion point too high after specific steps
2065
2066         * khtml/editing/visible_text.cpp:
2067         (khtml::SimplifiedBackwardsTextIterator::advance): Add new check for leaving a text node and iterating
2068         backwards into a different block that is an descendent of the block containing the text node (as in leaving
2069         the "bar" node in this example: <p>foo</p>bar).
2070         (khtml::SimplifiedBackwardsTextIterator::handleNonTextNode): Call new emitNewlineForBROrText helper.
2071         (khtml::SimplifiedBackwardsTextIterator::emitNewlineForBROrText): Factor out code from handleNonTextNode, since
2072         it is called from there, and now also from advance().
2073
2074 2004-12-15  Darin Adler  <darin@apple.com>
2075
2076         Reviewed by Ken.
2077
2078         - fixed problem where plain-text would put a blank line between each <p> even when they have no margins
2079
2080         * khtml/editing/visible_text.cpp: (khtml::TextIterator::exitNode): Add more checks and only set the
2081         "add one more newline" flag if the margin is sufficient. A more complete fix would ignore the node
2082         type altogether and use the render tree instead.
2083
2084 2004-12-14  John Sullivan  <sullivan@apple.com>
2085
2086         Reviewed by Ken.
2087
2088         - rest of WebCore fix for <rdar://problem/3790011> undoable operations all say "Undo"
2089         in the menu, no specific action names
2090
2091         * khtml/editing/edit_actions.h: new header, contains EditAction enum
2092         (renamed from HTMLEditAction, formerly in htmlediting.h)
2093
2094         * ForwardingHeaders/editing/edit_actions.h: new forwarding header
2095
2096         * khtml/editing/htmlediting.h:
2097         removed HTMLEditAction enum in favor of including edit_actions.h; added khtml::
2098         namespace to EditAction usages; added m_editingAction ivar to ApplyStyleCommand
2099
2100         * khtml/editing/htmlediting.cpp:
2101         (khtml::EditCommandPtr::editingAction):
2102         update for enum name change
2103         (khtml::EditCommand::editingAction):
2104         ditto
2105         (khtml::ApplyStyleCommand::ApplyStyleCommand):
2106         added editingAction parameter to this constructor
2107         (khtml::ApplyStyleCommand::editingAction):
2108         return new ivar
2109         (khtml::DeleteSelectionCommand::editingAction):
2110         update for enum name change
2111         (khtml::MoveSelectionCommand::editingAction):
2112         ditto
2113         (khtml::TypingCommand::editingAction):
2114         ditto
2115         (khtml::ReplaceSelectionCommand::editingAction):
2116         ditto
2117
2118         * khtml/khtml_part.h:
2119         added EditAction parameter to applyStyle and computeAndSetTypingStyle
2120         * khtml/khtml_part.cpp:
2121         (KHTMLPart::computeAndSetTypingStyle):
2122         added EditAction parameter
2123         (KHTMLPart::applyStyle):
2124         ditto
2125
2126         * kwq/KWQKHTMLPart.mm:
2127         (KWQKHTMLPart::registerCommandForUndoOrRedo):
2128         do the cast from EditAction to WebUndoAction a different way to match other code
2129
2130         * kwq/WebCoreBridge.h:
2131         * kwq/WebCoreBridge.mm:
2132         (-[WebCoreBridge setTypingStyle:withUndoAction:]):
2133         added WebUndoAction parameter, passed into ApplyStyleCommand constructor
2134         (-[WebCoreBridge applyStyle:withUndoAction:]):
2135         ditto
2136
2137         * WebCore.pbproj/project.pbxproj:
2138         updated for new files
2139
2140 2004-12-14  David Hyatt  <hyatt@apple.com>
2141
2142         Fix for 3562458, rowspan and colspan converted to ints so that large values will work for them.  Remove
2143         the ridiculous 1024 limit on the span values.
2144
2145         * khtml/html/html_tableimpl.cpp:
2146         (HTMLTableCellElementImpl::parseHTMLAttribute):
2147         * khtml/rendering/render_table.cpp:
2148         (RenderTableCell::collapsedBottomBorder):
2149         * khtml/rendering/render_table.h:
2150         (khtml::RenderTableCell::colSpan):
2151         (khtml::RenderTableCell::setColSpan):
2152         (khtml::RenderTableCell::rowSpan):
2153         (khtml::RenderTableCell::setRowSpan):
2154         (khtml::RenderTableCol::span):
2155         (khtml::RenderTableCol::setSpan):
2156
2157 2004-12-14  David Hyatt  <hyatt@apple.com>
2158
2159         Make sure <col> and <colgroup> can have spans updated dynamically as well.
2160
2161         Reviewed by rjw
2162
2163         * khtml/html/html_tableimpl.cpp:
2164         (HTMLTableColElementImpl::parseHTMLAttribute):
2165         * khtml/rendering/render_table.cpp:
2166         (RenderTableCell::updateFromElement):
2167         (RenderTableCol::updateFromElement):
2168
2169 2004-12-14  David Hyatt  <hyatt@apple.com>
2170
2171         Fix for 3833123, setting a cell's colspan does not update rendering like it should.
2172
2173         Reviewed by rjw
2174
2175         * khtml/html/html_tableimpl.cpp:
2176         (HTMLTableCellElementImpl::parseHTMLAttribute):
2177         * khtml/rendering/render_table.cpp:
2178         (RenderTableCell::RenderTableCell):
2179         (RenderTableCell::updateFromElement):
2180
2181 2004-12-14  Chris Blumenberg  <cblu@apple.com>
2182
2183         Fixed: <rdar://problem/3864536> crash copying text from other apps and pasting in a sticky in dashboard
2184
2185         Reviewed by hyatt.
2186
2187         * khtml/editing/markup.cpp:
2188         (khtml::createFragmentFromText): ref and deref the fragment since calling appendChild can completely deref it
2189
2190 2004-12-14  David Hyatt  <hyatt@apple.com>
2191
2192         Fix for 3785211. Make sure to do a layout with the old position before doing a layout after changing to the new position.
2193         This is actually a regression from my positioned object DHTML optimization.
2194
2195         Reviewed by kocienda
2196
2197         * khtml/rendering/render_object.cpp:
2198         (RenderObject::setStyle):
2199
2200 2004-12-14  David Hyatt  <hyatt@apple.com>
2201
2202         Make sure the class attribute works when newlines are present in the attribute.
2203
2204         Reviewed by kocienda
2205
2206         * khtml/html/html_elementimpl.cpp:
2207         (HTMLNamedAttrMapImpl::parseClassAttribute):
2208
2209 2004-12-14  David Hyatt  <hyatt@apple.com>
2210
2211         Fix for 3724938, float element is duplicated and paints twice.  The logic for when to paint floats was
2212         actually fairly screwed up.  This patch simplifies the logic and makes addOverhangingFloats easier to
2213         read by splitting it into two separate functions.
2214
2215         Reviewed by kocienda
2216
2217         * khtml/rendering/render_block.cpp:
2218         (khtml::RenderBlock::layoutBlockChildren):
2219         (khtml::RenderBlock::clearFloats):
2220         (khtml::RenderBlock::addOverhangingFloats):
2221         (khtml::RenderBlock::addIntrudingFloats):
2222         * khtml/rendering/render_block.h:
2223
2224 2004-12-14  John Sullivan  <sullivan@apple.com>
2225
2226         Reviewed by Ken.
2227
2228         - added Undo action names for Cut, Paste, and Drag
2229
2230         * khtml/editing/htmlediting.h:
2231         * khtml/editing/htmlediting.cpp:
2232         (khtml::DeleteSelectionCommand::editingAction):
2233         overridden to return HTMLEditActionCut
2234         (khtml::MoveSelectionCommand::editingAction):
2235         overridden to return HTMLEditActionDrag
2236         (khtml::ReplaceSelectionCommand::editingAction):
2237         overridden to return HTMLEditActionPaste
2238
2239 2004-12-14  John Sullivan  <sullivan@apple.com>
2240
2241         Reviewed by Ken.
2242
2243         - architecture for WebCore part of fix for <rdar://problem/3790011> undoable operations all say "Undo" in the menu,
2244         no specific action names
2245
2246         The remaining step is to make each EditCommand subclass override editingAction() to return an
2247         appropriate value. (Unfortunately the mapping between subclass and user-distinguishable action
2248         is not completely straightforward, so this next step isn't trivial.)
2249
2250         * khtml/editing/htmlediting.h:
2251         new enum for HTMLEditAction
2252
2253         * khtml/editing/htmlediting.cpp:
2254         (khtml::EditCommandPtr::editingAction):
2255         new method, calls through to EditCommand
2256         (khtml::EditCommand::editingAction):
2257         new method for subclasses to override, returns HTMLEditActionUnspecified at this level
2258         (khtml::TypingCommand::editingAction):
2259         proof of concept override, returns HTMLEditActionTyping
2260
2261         * kwq/KWQKHTMLPart.h:
2262         declare new private bottleneck method registerCommandForUndoOrRedo
2263         * kwq/KWQKHTMLPart.mm:
2264         (KWQKHTMLPart::registerCommandForUndoOrRedo):
2265         new bottleneck method to reduce code duplication; now calls over the bridge
2266         to get the localized string to use for the Undo action name
2267         (KWQKHTMLPart::registerCommandForUndo):
2268         now calls new bottleneck method
2269         (KWQKHTMLPart::registerCommandForRedo):
2270         now calls new bottleneck method
2271
2272         * kwq/WebCoreBridge.h:
2273         new enum for WebUndoAction, maps directly to HTMLEditAction.
2274         Declaration of nameForUndoAction:
2275
2276 2004-12-14  Darin Adler  <darin@apple.com>
2277
2278         Reviewed by John.
2279
2280         - added a bunch of missing nil checks; our old version of inherits used to work for nil (by accident)
2281
2282         * khtml/rendering/render_frames.cpp:
2283         (RenderPart::~RenderPart): Check widget for nil.
2284         (RenderPart::setWidget): Ditto.
2285         (RenderFrame::slotViewCleared): Ditto.
2286         (RenderPartObject::slotViewCleared): Ditto.
2287
2288 2004-12-13  Darin Adler  <darin@apple.com>
2289
2290         Reviewed by Ken.
2291
2292         - moved markup-related functions into new sources files in the editing directory
2293         - removed all of the uses of dynamic_cast, preparing to turn off RTTI to make our code smaller and slightly faster
2294
2295         * ForwardingHeaders/editing/markup.h: Added.
2296         * khtml/editing/markup.h: Added.
2297         * khtml/editing/markup.cpp: Added.
2298
2299         * WebCore.pbproj/project.pbxproj: Added markup.h/cpp.
2300
2301         * khtml/dom/dom_node.cpp: (Node::toHTML): Call createMarkup since there's no toHTML in NodeImpl any more.
2302         * khtml/html/html_elementimpl.cpp:
2303         (HTMLElementImpl::innerHTML): Changed to call createMarkup.
2304         (HTMLElementImpl::outerHTML): Ditto.
2305
2306         * khtml/ecma/kjs_window.cpp:
2307         (Window::retrieveWindow): Comment out assert that uses dynamic_cast.
2308         (Window::retrieveActive): Ditto.
2309
2310         * khtml/editing/htmlediting.h: Added forward class declaration needed now that I removed one elsewhere.
2311         * khtml/xml/dom_docimpl.h: Ditto.
2312
2313         * khtml/khtml_part.cpp:
2314         (KHTMLPart::slotDebugDOMTree): Use createMarkup instead of toHTML.
2315         (KHTMLPart::processObjectRequest): Use inherits instead of dynamic_cast.
2316
2317         * khtml/rendering/render_image.cpp: (RenderImage::paint): Add an explicit QChar conversion so this code
2318         still works even with the additional replace overloads added to QString.
2319         * kwq/KWQTextCodec.mm: (QTextCodec::fromUnicode): Ditto.
2320
2321         * khtml/rendering/render_object.h: Removed the version of arenaDelete that does not take an object
2322         base pointer, because it used dynamic_cast in its implementation. Made the other version public.
2323         * khtml/rendering/render_object.cpp: Ditto.
2324         * khtml/rendering/render_replaced.cpp: (RenderWidget::deref): Pass object base pointer to arenaDelete.
2325
2326         * khtml/xml/dom2_rangeimpl.h: Removed extra parameters from toHTML, and unneeded includes and declarations.
2327         * khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::toHTML): Changed to call createMarkup, and moved all
2328         the support code into markup.cpp.
2329
2330         * khtml/xml/dom_nodeimpl.h: Moved toHTML and related functions into markup.cpp.
2331         * khtml/xml/dom_nodeimpl.cpp: Ditto.
2332
2333         * khtml/xml/dom_position.cpp:
2334         (DOM::startPosition): Implemented the version of this that takes a RangeImpl. Also added null checks
2335         so these return null positions rather than raising exceptions.
2336         (DOM::endPosition): Ditto.
2337
2338         * khtml/khtmlview.h: Added an APPLE_CHANGES function so inherits can detect this class without dynamic_cast.
2339         * kwq/KWQFrame.h: Ditto.
2340         * kwq/KWQFrame.mm: (QFrame::isQFrame): Ditto.
2341         * kwq/KWQKPartsPart.h: Ditto.
2342         * kwq/KWQKPartsPart.mm: (KParts::ReadOnlyPart::isKPartsReadOnlyPart): Ditto.
2343         * kwq/KWQScrollView.h: Ditto.
2344         * kwq/KWQScrollView.mm: (QScrollView::isQScrollView): Ditto.
2345         * kwq/KWQKHTMLPart.h: Ditto.
2346         * kwq/KWQKHTMLPart.mm:
2347         (KHTMLView::isKHTMLView): Ditto.
2348         (KWQKHTMLPart::setTitle): Added an explicit QChar conversion so this code still works even with the additional
2349         replace overloads added to QString.
2350         (KWQKHTMLPart::setStatusBarText): Ditto.
2351         (KWQKHTMLPart::runJavaScriptAlert): Ditto.
2352         (KWQKHTMLPart::runJavaScriptConfirm): Ditto.
2353         (KWQKHTMLPart::runJavaScriptPrompt): Ditto.
2354         (KWQKHTMLPart::attributedString): Ditto.
2355         (KWQKHTMLPart::isCharacterSmartReplaceExempt): Ditto.
2356         (KWQKHTMLPart::isKHTMLPart): That dynamic_cast thing (see above).
2357
2358         * kwq/KWQObject.h: Added virtual methods for the few cases where we need dynamic_cast-like behavior.
2359         * kwq/KWQObject.mm:
2360         (QObject::inherits): Changed to not use dynamic cast.
2361         (QObject::isKHTMLPart): Added. Returns false.
2362         (QObject::isKHTMLView): Ditto.
2363         (QObject::isKPartsReadOnlyPart): Ditto.
2364         (QObject::isQFrame): Ditto.
2365         (QObject::isQScrollView): Ditto.
2366
2367         * kwq/KWQRenderTreeDebug.cpp:
2368         (write): Changed to use inherits rather than dynamic_cast.
2369         (writeSelection): Ditto.
2370
2371         * kwq/KWQSlot.mm: (KWQSlot::call): Call through to the version with just a job pointer parameter rather
2372         than going straight on to the "no parameters at all" version.
2373
2374         * kwq/KWQString.h:
2375         * kwq/KWQString.mm: (QString::replace): Added overloads.
2376         * kwq/WebCoreBridge.mm:
2377         (-[WebCoreBridge markupStringFromNode:nodes:]): Changed to call functions in markup.h.
2378         (-[WebCoreBridge markupStringFromRange:nodes:]): Ditto.
2379         (-[WebCoreBridge selectedString]): Added an explicit QChar conversion so this code still works even with
2380         the additional replace overloads added to QString.
2381         (-[WebCoreBridge stringForRange:]): Ditto.
2382         (-[WebCoreBridge copyDOMNode:copier:]): Changed to call functions in markup.h.
2383         (-[WebCoreBridge elementAtPoint:]): QChar conversion.
2384         (-[WebCoreBridge documentFragmentWithMarkupString:baseURLString:]): Changed to call functions in markup.h.
2385         (-[WebCoreBridge documentFragmentWithText:]): Changed to call functions in markup.h.
2386
2387 2004-12-13  Ken Kocienda  <kocienda@apple.com>
2388
2389         Reviewed by John
2390
2391         Fix for this bug:
2392
2393         <rdar://problem/3917956> REGRESSION (Mail): pasting can leave insertion point inside pasted text
2394
2395         * khtml/editing/htmlediting.cpp:
2396         (khtml::ReplaceSelectionCommand::doApply): Fix coding mistake. Calculations of bool flag based on
2397         leading and trailing whitespace positions was reversed! I must have introduced this error recently
2398         when changing around this code.
2399
2400 2004-12-13  David Hyatt  <hyatt@apple.com>
2401
2402         Fix for 3915787, macobserver doesn't paint.  floatRect() needed to be const in the base class.  Also hit-testing
2403         and painting was using the wrong rect when setting up the x/y of the rect.
2404
2405         Reviewed by mjs
2406
2407         * khtml/rendering/render_block.cpp:
2408         (khtml::RenderBlock::paint):
2409         (khtml::RenderBlock::nodeAtPoint):
2410         * khtml/rendering/render_object.h:
2411         (khtml::RenderObject::floatRect):
2412
2413 2004-12-13  Ken Kocienda  <kocienda@apple.com>
2414
2415         Reviewed by John
2416
2417         Fix for this bug:
2418
2419         <rdar://problem/3917863> REGRESSION (Mail): pasting two lines of plain text copied from an RTF document results in two styles
2420
2421         Code to figuire out the end node to merge was missing the font tag in the second paragraph
2422         written out by AppKit convert-to-HTML function. I refined the algorithm to be smarter.
2423
2424         * khtml/editing/htmlediting.cpp:
2425         (khtml::ReplacementFragment::mergeEndNode): Refine algorithm used to walk through the fragment being pasted
2426         looking for the node that is the last inline in the last block of the fragment. The old algorithm was
2427         insufficiently powerful.
2428         (khtml::ReplacementFragment::enclosingBlock): New helper function.
2429         * khtml/editing/htmlediting.h: Add declaration for new helper function.
2430         * layout-tests/editing/pasteboard/paste-text-011-expected.txt: Added.
2431         * layout-tests/editing/pasteboard/paste-text-011.html: Added.
2432
2433 2004-12-13  Ken Kocienda  <kocienda@apple.com>
2434
2435         Reviewed by John
2436
2437         WebCore side of fix for this bug:
2438
2439         <rdar://problem/3768372> REGRESSION (Mail): paste of text ending in whitespace loses whitespace
2440
2441         Note that we are coordinating with Doug Davidson on the AppKit team to make a complete fix for this
2442         bug. This change involves our half of the needed changes.
2443
2444         Note that a lot of this change has to do with changing code to use a <br> element instead of
2445         a comment node as the mechanism to annotate HTML with information used to fix the bug. In some
2446         other places, code to handle comments in markup can be removed since we do not use comments for
2447         such annotations after this change.
2448
2449         * khtml/editing/htmlediting.cpp: Remove isComment() helper; no longer needed.
2450         (khtml::ReplacementFragment::ReplacementFragment): Change m_hasInterchangeNewlineComment name to m_hasInterchangeNewline.
2451         (khtml::ReplacementFragment::isInterchangeNewlineNode): Name changed from isInterchangeNewlineComment.
2452         (khtml::ReplacementFragment::isInterchangeConvertedSpaceSpan): Local variable name convertedSpaceSpanClass changed to
2453         convertedSpaceSpanClassString to match other uses of the idiom used here.
2454         (khtml::ReplaceSelectionCommand::doApply): Change hasInterchangeNewlineComment() name to hasInterchangeNewline().
2455         * khtml/editing/htmlediting.h: Change names as noted in .cpp file. Remove isComment() helper; no longer needed.
2456         (khtml::ReplacementFragment::hasInterchangeNewline):  Change hasInterchangeNewlineComment() name to hasInterchangeNewline().
2457         * khtml/html/html_elementimpl.cpp:
2458         (HTMLElementImpl::createContextualFragment): No longer has includeCommentsInDOM flag; no longer needed as we do not
2459         annotate fragments with comments any longer.
2460         * khtml/html/html_elementimpl.h: Ditto.
2461         * khtml/xml/dom2_rangeimpl.cpp: Remove addCommentToHTMLMarkup() helper. No longer needed.
2462         (DOM::interchangeNewlineMarkupString): New helper to return <br> element markup we use to annotate content for interchange.
2463         (DOM::RangeImpl::toHTML): No longer uses addCommentToHTMLMarkup; now calls interchangeNewlineMarkupString(). Remove
2464         spurious semi-colon.
2465         * khtml/xml/dom2_rangeimpl.h: Remove obsolete addCommentToHTMLMarkup() function and EAddToMarkup enum.
2466         * kwq/WebCoreBridge.mm:
2467         (-[WebCoreBridge documentFragmentWithMarkupString:baseURLString:]): No longer pass bool to ask for including comments
2468         in DOM when calling createContextualFragment().
2469
2470 2004-12-10  John Sullivan  <sullivan@apple.com>
2471
2472         fixed deployment build bustage that John Louch ran into
2473
2474         * kwq/KWQTextEdit.mm:
2475         (QTextEdit::setScrollBarModes):
2476         move bool declaration inside exception-handling block to avoid obscure
2477         compiler error
2478
2479 2004-12-10  Maciej Stachowiak  <mjs@apple.com>
2480
2481         Reviewed by Richard.
2482
2483         <rdar://problem/3907484> REGRESSION (125-173): crash when KWQTextField is dealloc'ed while setting focus (profoundlearning.com)
2484
2485         * kwq/KWQWidget.mm:
2486         (QWidget::setFocus): Handle the case where setting focus removed
2487         us from the superview - this can happen due to style changes on
2488         focus change.
2489
2490 2004-12-10  Ken Kocienda  <kocienda@apple.com>
2491
2492         Reviewed by Hyatt
2493
2494         Fix for this bug:
2495
2496         <rdar://problem/3915008> REGRESSION (Mail): Too much white space between lines separated by carriage returns
2497
2498         There are a number of interesting things we could do to fix this bug, including SPI and involving
2499         the WebKit delegate, etc., however it seems reasonable to start with a hard-coded default that
2500         will fix the bug in the general case until such time as we can come up with more specific
2501         solutions.
2502
2503         So, I added a helper method to create <p> elements with an inline style that sets top and bottom margins
2504         to 0.1em.
2505
2506         * khtml/editing/htmlediting.cpp:
2507         (khtml::InsertParagraphSeparatorCommand::createParagraphElement): New factory method to create
2508         paragraph elements to insert. Also adds style information to keep the <p> from having "too-big" margins.
2509         (khtml::InsertParagraphSeparatorCommand::doApply): Call new factory method.
2510         * khtml/editing/htmlediting.h: Add createParagraphElement() declaration.
2511
2512 2004-12-10  Darin Adler  <darin@apple.com>
2513
2514         Reviewed by Hyatt.
2515
2516         - fixed <rdar://problem/3910419> setting style={overflow:hidden} for <textarea> does not prevent appearance of scrollbars
2517
2518         * khtml/rendering/render_form.h: Remove now-unneeded wrap parameter.
2519         * khtml/rendering/render_form.cpp:
2520         (RenderSubmitButton::rawText): Convert to QChar explicitly.
2521         (RenderLineEdit::updateFromElement): Ditto.
2522         (RenderLineEdit::slotTextChanged): Ditto.
2523         (RenderSelect::updateFromElement): Ditto.
2524         (TextAreaWidget::TextAreaWidget): Moved out most of the initialization since it's not something
2525         that requires a derived class. Now we don't use this class at all for WebCore, but they still
2526         have it for KDE.
2527         (TextAreaWidget::event): Moved out the ifdefs.
2528         (RenderTextArea::RenderTextArea): Moved setting code from TextAreaWidget here. Put a bunch that
2529         we don't need at all inside !APPLE_CHANGES, and removed the setting for scroll bars, since that's
2530         now done in setStyle.
2531         (RenderTextArea::handleFocusOut): Use type QTextEdit instead of TextAreaWidget since that's all
2532         that's needed and WebCore no longer has TextAreaWidget.
2533         (RenderTextArea::calcMinMaxWidth): Ditto.
2534         (RenderTextArea::setStyle): Add code to set scroll bar modes based on wrap setting combined with
2535         overflow style.
2536         (RenderTextArea::updateFromElement): Use type QTextEdit.
2537         (RenderTextArea::text): Ditto.
2538         (RenderTextArea::select): Ditto.
2539
2540         * kwq/KWQTextArea.mm:
2541         (-[KWQTextArea _configureTextViewForWordWrapMode]): Don't set horizontal scroller visibility here,
2542         since it's now handled by QTextEdit.
2543         (-[KWQTextArea initWithFrame:]): Don't set vertical scroller visibility or scroller auto-hiding.
2544
2545         * kwq/KWQTextEdit.h: Add setScrollBarModes function to be used instead of separate setter for
2546         the horizontal and vertical mode; needed because AppKit switches "autohide" for both at once.
2547         * kwq/KWQTextEdit.mm: (QTextEdit::setScrollBarModes): Added.
2548
2549 2004-12-10  Ken Kocienda  <kocienda@apple.com>
2550
2551         Reviewed by Darin
2552
2553         Fix for this bug:
2554
2555         <rdar://problem/3915047> HItting return in empty document inserts <p> but
2556         insertion point does not move
2557
2558         * khtml/editing/htmlediting.cpp:
2559         (khtml::InsertParagraphSeparatorCommand::doApply): The issue is that the
2560         code to insert the <p> element for the return is not detecting the fact
2561         that the document is empty. Inserting a <p> into an empty body will not
2562         "add a new line" as the user expects. With this change, we'll add a second
2563         <p> when the root editable element has no rendered kids.
2564
2565 2004-12-10  Maciej Stachowiak  <mjs@apple.com>
2566
2567         Reviewed by John.
2568
2569         <rdar://problem/3912979> REGRESSION (125-173): repro crash in HTMLCollectionImpl code (www.clubtravel.ie)
2570
2571         * khtml/html/html_miscimpl.cpp:
2572         (HTMLCollectionImpl::traverseNextItem): Pass base when traversing
2573         the initial one step, otherwise we might inadvertantly step
2574         outside the collection base, thereby causing assertion failures or
2575         other badness later.
2576
2577 2004-12-10  Ken Kocienda  <kocienda@apple.com>
2578
2579         Reviewed by Chris
2580
2581         * khtml/editing/htmlediting.cpp:
2582         (khtml::InsertParagraphSeparatorCommand::doApply): There is a starting block which is supposed to
2583         act as the root node for this operation. However, a loop was incorrectly coded, and a parent node
2584         search could escape this node. Also, one other piece to code to move nodes to the new <p> element
2585         should do nothing if the starting point for the selection is itself the starting block.
2586         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Similar changes, in spirit, to the above
2587         function, though the names and concepts are slightly different.
2588
2589 2004-12-10  Maciej Stachowiak  <mjs@apple.com>
2590
2591         Reviewed by Ken.
2592
2593         <rdar://problem/3907705> REGRESSION (172-173): DHTML menus are broken at hrweb.apple.com
2594
2595         * khtml/dom/html_document.cpp:
2596         (HTMLDocument::nameableItems): New method, wrapper for HTMLCollection creation.
2597         * khtml/dom/html_document.h:
2598         * khtml/ecma/kjs_html.cpp:
2599         (KJS::HTMLDocument::tryGet): use doc.nameableItems(), not doc.all()!
2600         * khtml/html/html_miscimpl.cpp:
2601         (HTMLCollectionImpl::traverseNextItem): Added new DOC_NAMEABLE_ITEMS type, this represents
2602         the items that can be accessed directly as a document propery, in particular forms, images,
2603         objects, applets and embeds.
2604         (HTMLCollectionImpl::updateNameCache): Fix some nameCache/idCache confusion.
2605         (HTMLFormCollectionImpl::updateNameCache): Ditto.
2606         * khtml/html/html_miscimpl.h:
2607         (DOM::HTMLCollectionImpl::): Added new type.
2608
2609 2004-12-10  Ken Kocienda  <kocienda@apple.com>
2610
2611         Reviewed by John
2612
2613         Fix for this bug:
2614
2615         <rdar://problem/3914779> REGRESSION (Mail): Cannot arrow navigate to position before last character on text-wrapped line
2616
2617         * khtml/rendering/render_text.cpp:
2618         (RenderText::caretRect): Code was not detecting space at the end of a line correctly. Now it does.
2619
2620 2004-12-10  Ken Kocienda  <kocienda@apple.com>
2621
2622         Reviewed by John
2623
2624         Fix for this bug:
2625
2626         <rdar://problem/3914755> REGRESSION (Mail): Insertion point disappears after pasting paragraph
2627
2628         * khtml/editing/htmlediting.cpp:
2629         (khtml::ReplaceSelectionCommand::doApply): Selection could end up in a "placeholder" node
2630         that was removed from the document when pasting. This caused the disappearance. Now this
2631         is detected, and the selection is shifted to a node that is in the document.
2632
2633 2004-12-09  Richard Williamson   <rjw@apple.com>
2634
2635         Check to disable threaded decoding during
2636         layout tests wasn't invoking function, just checking address of
2637         function, which would always return true.
2638
2639         * kwq/WebCoreImageRendererFactory.m:
2640         (+[WebCoreImageRendererFactory shouldUseThreadedDecoding]):
2641
2642 2004-12-09  David Hyatt  <hyatt@apple.com>
2643
2644         Fix for 3892686, left/top overflow was not being propagated properly up to containing blocks because of a math
2645         error.
2646
2647         Also fix a bug I noticed on the same page where relative position offsets were not being added in properly for
2648         all inlines when repainting.
2649
2650         Reviewed by rjw
2651
2652         * khtml/rendering/render_block.cpp:
2653         (khtml::RenderBlock::layoutBlockChildren):
2654         * khtml/rendering/render_flow.cpp:
2655         (RenderFlow::getAbsoluteRepaintRect):
2656
2657 2004-12-09  David Hyatt  <hyatt@apple.com>
2658
2659         Fix for 3867545, finance.yahoo.com lays out incorrectly.  Add a quirk that will prevent tables from moving
2660         down below floats when there is insufficient space.  Instead we will match Gecko and just spill out of the
2661         containing block to the right.  This appears to be the more common desired behavior, despite being wrong.
2662         WinIE sometimes wraps and sometimes doesn't, but it's really hard for us to match its inconsistency.
2663
2664         Reviewed by john
2665
2666         * khtml/rendering/render_block.cpp:
2667         (khtml::RenderBlock::getClearDelta):
2668         * layout-tests/apple-only/base/www.cnn.com/index-expected.txt:
2669         * layout-tests/apple-only/base/www.ebay.com/index-expected.txt:
2670         * layout-tests/apple-only/base/www.excite.com/index-expected.txt:
2671         * layout-tests/fast/block/margin-collapse/102-expected.txt:
2672         * layout-tests/fast/block/margin-collapse/102.html:
2673
2674 2004-12-09  Richard Williamson   <rjw@apple.com>
2675
2676         Fixed <rdar://problem/3914078> worldclock crashing gc related
2677
2678         Use ProtectedValue for Context2D instance members.
2679
2680         Reviewed by Maciej.
2681
2682         * khtml/ecma/kjs_html.h:
2683
2684 2004-12-09  John Sullivan  <sullivan@apple.com>
2685
2686         Reviewed by Dave.
2687
2688         - fixed <rdar://problem/3731099> Move AXTitle string for image elements to AXDescription
2689
2690         * kwq/KWQAccObject.mm:
2691         (-[KWQAccObject title]): moved image alt tag code out of here
2692         (-[KWQAccObject accessibilityDescription]): moved image alt tag code into this new method
2693         (-[KWQAccObject accessibilityAttributeNames]): include AXDescription in the set of attributes
2694         that ordinary elements return; this means that ordinary elements that aren't images will return
2695         a nil description, which isn't ideal, but is in keeping with the way the rest of these attributes
2696         work here.
2697         (-[KWQAccObject accessibilityAttributeValue:]):
2698         call accessibilityDescription when asked for AXDescription
2699
2700 2004-12-09  Ken Kocienda  <kocienda@apple.com>
2701
2702         Reviewed by Harrison
2703
2704         Fix for this bug:
2705
2706         <rdar://problem/3910425> REGRESSION (Mail): Crash in ReplaceSelectionCommand; selection is empty, leading to null deref
2707
2708         * khtml/editing/htmlediting.cpp:
2709         (khtml::MoveSelectionCommand::doApply): The node representing the destination for the move may have
2710         been deleted. If this is the case, set the destination to the node the delete command provides in
2711         its ending selection.
2712
2713 === Safari-175 ===
2714
2715 2004-12-09  Ken Kocienda  <kocienda@apple.com>
2716
2717         Reviewed by John
2718
2719         Fix for this bug:
2720
2721         <rdar://problem/3912841> REGRESSION (173-TOT): Some images report 0x0 dimensions on layout tests, causes spurious test failures
2722
2723         The new threaded image decoding capability can throw off layout tests. The issue is that the decoding
2724         callback may not be delivered before the program asks for the dimensions of an image in order to
2725         wrote the layout dimensions. More generally, I think we need to ensure that there are no races in
2726         layout tests, so I have added a flag to the render tree debug code that we can set when debugging.
2727
2728         * kwq/KWQRenderTreeDebug.cpp:
2729         (debuggingRenderTree): New function. Returns flag which tells whether the program is debugging the render tree.
2730         (externalRepresentation): Sets debuggingRenderTree flag to true;
2731         * kwq/KWQRenderTreeDebug.h:
2732         * kwq/WebCoreImageRendererFactory.m:
2733         (+[WebCoreImageRendererFactory shouldUseThreadedDecoding]): Checks debuggingRenderTree flag and will not
2734         do threaded decoding in any case if the flag is set.
2735
2736 2004-12-09  Chris Blumenberg  <cblu@apple.com>
2737
2738         Fix for busting XMLHTTPRequest.
2739
2740         Reviewed by kocienda.
2741
2742         * khtml/misc/loader.cpp:
2743         (Loader::servePendingRequests): pass true for deliverAllData
2744         * kwq/KWQKJob.h:
2745         (KIO::get): take deliverAllData param
2746         (KIO::http_post): ditto
2747         * kwq/KWQKJobClasses.h:
2748         * kwq/KWQKJobClasses.mm:
2749         (KIO::TransferJob::TransferJob): if deliverAllData, create signal with data param
2750         (KIO::TransferJob::emitResult): if deliverAllData, call signal with data param
2751
2752 2004-12-09  Ken Kocienda  <kocienda@apple.com>
2753
2754         Reviewed by John
2755
2756         <rdar://problem/3911011> REGRESSION (Mail): Spaces at end of line causing word wrap lost when copied/pasted
2757
2758         * khtml/xml/dom_nodeimpl.cpp:
2759         (NodeImpl::renderedText): Fixed the code so that spaces at the end of lines are not skipped.
2760
2761 2004-12-07  Richard Williamson   <rjw@apple.com>
2762
2763         Support threaded image decoding on machines w/ > 2 CPUs.
2764
2765         Reviewed by Maciej and Chris.
2766
2767         * khtml/misc/loader.cpp:
2768         (CachedImageCallback::notifyUpdate):
2769         (CachedImageCallback::notifyFinished):
2770         (CachedImageCallback::notifyDecodingError):
2771         (CachedImageCallback::handleError):
2772         (CachedImageCallback::clear):
2773         (CachedImage::CachedImage):
2774         (CachedImage::clear):
2775         (CachedImage::data):
2776         (CachedImage::checkNotify):
2777         (Loader::servePendingRequests):
2778         (Loader::slotFinished):
2779         (Loader::numRequests):
2780         (Loader::cancelRequests):
2781         (Loader::removeBackgroundDecodingRequest):
2782         * khtml/misc/loader.h:
2783         (khtml::CachedImageCallback::CachedImageCallback):
2784         (khtml::CachedImageCallback::ref):
2785         (khtml::CachedImageCallback::deref):
2786         (khtml::CachedImage::decoderCallback):
2787         * khtml/rendering/render_object.cpp:
2788         (RenderObject::setPixmap):
2789         * kwq/KWQPixmap.h:
2790         * kwq/KWQPixmap.mm:
2791         (-[WebImageCallback initWithCallback:khtml::]):
2792         (-[WebImageCallback _commonTermination]):
2793         (-[WebImageCallback dealloc]):
2794         (-[WebImageCallback finalize]):
2795         (-[WebImageCallback notify]):
2796         (-[WebImageCallback setImageSourceStatus:]):
2797         (-[WebImageCallback status]):
2798         (QPixmap::shouldUseThreadedDecoding):
2799         (QPixmap::receivedData):
2800         * kwq/WebCoreImageRenderer.h:
2801         * kwq/WebCoreImageRendererFactory.h:
2802         * kwq/WebCoreImageRendererFactory.m:
2803         (+[WebCoreImageRendererFactory shouldUseThreadedDecoding]):
2804         (+[WebCoreImageRendererFactory setShouldUseThreadedDecoding:]):
2805
2806 2004-12-07  Ken Kocienda  <kocienda@apple.com>
2807
2808         Reviewed by John
2809
2810         * khtml/editing/htmlediting.cpp:
2811         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Fix a problem with my
2812         change in thei code yesterday. Call to insertBlockPlaceholderIfNeeded() must be done
2813         after new block has been inserted, otherwise a crash can result. Shuffle down call
2814         to insertBlockPlaceholderIfNeeded() a couple of lines (where the node is inserted),
2815         and all is good.
2816
2817 2004-12-07  Ken Kocienda  <kocienda@apple.com>
2818
2819         Reviewed by John
2820
2821         Fix for this bug:
2822
2823         <rdar://problem/3907422> REGRESSION (Mail): Pasting quoted content can place content after body element
2824
2825         * khtml/editing/htmlediting.cpp:
2826         (khtml::ReplaceSelectionCommand::doApply): Detect when the body element is the "reference block" used
2827         for determining the location for inserting content. Do not allow an insert before or after if the
2828         reference block is the body. Perform insertNodeAt(0) and appendNode, respectively, in the block-is-body case.
2829         * layout-tests/editing/inserting/insert-3907422-fix-expected.txt: Added.
2830         * layout-tests/editing/inserting/insert-3907422-fix.html: Added.
2831
2832 2004-12-07  Darin Adler  <darin@apple.com>
2833
2834         Reviewed by Don.
2835
2836         - fixed <rdar://problem/3908701> REGRESSION: Cursor does not change to "hand" over active links
2837
2838         * kwq/KWQEvent.mm: (positionForEvent): Add NSMouseMoved to list of events that have mouse location.
2839
2840 2004-12-07  Ken Kocienda  <kocienda@apple.com>
2841
2842         Reviewed by John
2843
2844         Changed name of constant from KHTMLInterchangeNewline to AppleInterchangeNewline.
2845         I discussed this with Hyatt and he agreed that going with "Apple" names was OK.
2846
2847         * khtml/editing/html_interchange.h: Name change, as described above.
2848         * khtml/editing/htmlediting.cpp:
2849         (khtml::ReplacementFragment::isInterchangeNewlineComment): Ditto.
2850         * khtml/xml/dom2_rangeimpl.cpp:
2851         (DOM::RangeImpl::toHTML): Ditto.
2852
2853 2004-12-07  Ken Kocienda  <kocienda@apple.com>
2854
2855         Reviewed by me
2856
2857         Added a couple more layout tests.
2858
2859         * layout-tests/editing/deleting/delete-line-013-expected.txt: Added.
2860         * layout-tests/editing/deleting/delete-line-013.html: Added.
2861         * layout-tests/editing/deleting/delete-line-014-expected.txt: Added.
2862         * layout-tests/editing/deleting/delete-line-014.html: Added.
2863
2864 2004-12-06  Maciej Stachowiak  <mjs@apple.com>
2865
2866         Reviewed by Hyatt.
2867
2868         - fixed <rdar://problem/3906974> assertion failure in QWidget::beforeMouseDown clicking on <select multiple>
2869
2870         * kwq/KWQListBox.mm:
2871         (QListBox::QListBox): Initialize KWQListBoxScrollView with this.
2872         (-[KWQListBoxScrollView initWithListBox:]): Make this class a KWQWidgetHolder.
2873         (-[KWQListBoxScrollView widget]): See above.
2874         (-[KWQTableView mouseDown:]): Pass outerView rather than self to beforeMouseDown and
2875         afterMouseDown, to avoid triggering an assertion failure.
2876
2877 2004-12-06  David Hyatt  <hyatt@apple.com>
2878
2879         Fix for 3615411, the linesAppended optimization was old and broken, and it's easier with the new code fixes
2880         made by me, kocienda and harrison to just remove it.
2881
2882         Reviewed by mjs
2883
2884         * khtml/rendering/bidi.cpp:
2885         (khtml::RenderBlock::layoutInlineChildren):
2886         * khtml/rendering/render_block.cpp:
2887         (khtml:::RenderFlow):
2888         * khtml/rendering/render_block.h:
2889         * khtml/rendering/render_flow.cpp:
2890         (RenderFlow::dirtyLinesFromChangedChild):
2891
2892 2004-12-06  David Hyatt  <hyatt@apple.com>
2893
2894         Fix for 3787133, some web pages print with many blank pages.  Make sure to use the real page print rect and
2895         not a damage rect that can be changed when intersected with the clip regions of the web page.
2896
2897         Reviewed by john
2898
2899         * khtml/rendering/render_canvas.h:
2900         (khtml::RenderCanvas::printRect):
2901         (khtml::RenderCanvas::setPrintRect):
2902         * khtml/rendering/render_flow.cpp:
2903         (RenderFlow::paintLines):
2904         * khtml/rendering/render_list.cpp:
2905         (RenderListMarker::paint):
2906         * kwq/KWQKHTMLPart.mm:
2907         (KWQKHTMLPart::adjustPageHeight):
2908
2909 2004-12-06  David Harrison  <harrison@apple.com>
2910
2911         Reviewed by Ken Kocienda and Dave Hyatt (OOPS!).
2912
2913                 <rdar://problem/3849947> Typing after pasting line does not appear until after window resize.
2914
2915
2916         * khtml/rendering/render_flow.cpp:
2917         (RenderFlow::dirtyLinesFromChangedChild):
2918                 Dirty the line above because new child can inval the cached line break position of previous line.
2919
2920 2004-12-06  David Hyatt  <hyatt@apple.com>
2921
2922         Fix for 3254464, radio buttons do not work for quiz on netscape.com.  Left/top overflow needed to be implemented.
2923         This also fixes 3106907, link hover color only partially set on rollover and the more general architecture bug
2924         3126929, handle top/left overflow.
2925
2926         This patch also fixes 3902891, scroll bar of position:fixed content moves when a page is scrolled.
2927
2928         Finally, the Emerson regression 3869718 (error involving computing the rightmost/lowest position of overflow:auto
2929         regions and web pages) has been fixed.
2930
2931         Reviewed by mjs
2932
2933         * khtml/rendering/bidi.cpp:
2934         (khtml::RenderBlock::computeHorizontalPositionsForLine):
2935         (khtml::RenderBlock::checkLinesForOverflow):
2936         * khtml/rendering/render_block.cpp:
2937         (khtml:::RenderFlow):
2938         (khtml::RenderBlock::overflowHeight):
2939         (khtml::RenderBlock::overflowWidth):
2940         (khtml::RenderBlock::overflowLeft):
2941         (khtml::RenderBlock::overflowTop):
2942         (khtml::RenderBlock::overflowRect):
2943         (khtml::RenderBlock::layoutBlock):
2944         (khtml::RenderBlock::layoutBlockChildren):
2945         (khtml::RenderBlock::paint):
2946         (khtml::RenderBlock::floatRect):
2947         (khtml::RenderBlock::lowestPosition):
2948         (khtml::RenderBlock::rightmostPosition):
2949         (khtml::RenderBlock::leftmostPosition):
2950         (khtml::RenderBlock::nodeAtPoint):
2951         * khtml/rendering/render_block.h:
2952         * khtml/rendering/render_box.h:
2953         (khtml::RenderBox::borderBox):
2954         (khtml::RenderBox::borderTopExtra):
2955         (khtml::RenderBox::borderBottomExtra):
2956         * khtml/rendering/render_layer.cpp:
2957         (RenderLayer::paintScrollbars):
2958         (mustExamineRenderer):
2959         (RenderLayer::intersectsDamageRect):
2960         (RenderLayer::containsPoint):
2961         * khtml/rendering/render_line.cpp:
2962         (khtml::InlineFlowBox::placeBoxesHorizontally):
2963         (khtml::InlineFlowBox::verticallyAlignBoxes):
2964         * khtml/rendering/render_line.h:
2965         (khtml::InlineBox::leftOverflow):
2966         (khtml::InlineBox::rightOverflow):
2967         (khtml::InlineFlowBox::setVerticalOverflowPositions):
2968         (khtml::RootInlineBox::RootInlineBox):
2969         (khtml::RootInlineBox::leftOverflow):
2970         (khtml::RootInlineBox::rightOverflow):
2971         (khtml::RootInlineBox::setVerticalOverflowPositions):
2972         (khtml::RootInlineBox::setHorizontalOverflowPositions):
2973         * khtml/rendering/render_object.h:
2974         (khtml::RenderObject::borderBox):
2975         (khtml::RenderObject::overflowLeft):
2976         (khtml::RenderObject::overflowTop):
2977         (khtml::RenderObject::overflowRect):
2978         (khtml::RenderObject::floatRect):
2979         * khtml/rendering/render_table.cpp:
2980         (RenderTable::layout):
2981         (RenderTable::paint):
2982         (RenderTable::paintBoxDecorations):
2983         (RenderTable::calcMinMaxWidth):
2984         * khtml/rendering/render_table.h:
2985         (khtml::RenderTableCell::borderTopExtra):
2986         (khtml::RenderTableCell::borderBottomExtra):
2987         * kwq/KWQRect.h:
2988         * kwq/KWQRect.mm:
2989         (QRect::inflate):
2990
2991 2004-12-06  Maciej Stachowiak  <mjs@apple.com>
2992
2993         Reviewed by John.
2994
2995         - fixed <rdar://problem/3903797> scripts can cause other frames/windows to execute arbitrary script using javascript: URLs
2996
2997         I changed all unprotected places that can navigate a different
2998         window or frame from script to check for a javascript: URL, and if
2999         found, to check for safety using cross-site-script rules.
3000
3001         I considered a few other possible exploits and made no change:
3002
3003         - document.location is already protected because the document
3004         object itself is protected
3005
3006         - frame.src, frame.location, iframe.src and targetted links are
3007         all safe because setting the URL of a frame to a javascript: URL
3008         executes the script in the context of the parent
3009
3010         * khtml/ecma/kjs_window.cpp:
3011         (WindowFunc::tryCall):
3012         (Location::put):
3013         (LocationFunc::tryCall):
3014
3015 2004-12-06  Ken Kocienda  <kocienda@apple.com>
3016
3017         Reviewed by Maciej
3018
3019         Fix for this bug:
3020
3021         <rdar://problem/3890955> 8A314: Forward delete sometimes fails to delete the selected quoted text
3022
3023         * khtml/editing/htmlediting.cpp:
3024         (khtml::DeleteSelectionCommand::handleSpecialCaseBRDelete): Fixed bonehead coding mistake in the
3025         check for one of the special cases being checked for in this function. The specific case
3026         intends to check for a selection that is only a <br> after a block ends (as in </div><br>). If it
3027         sees such markup, it deletes only the <br> and bails. However, this code would run in *any*
3028         case where a selection ended in a <br> after a block and would not delete any part of the
3029         selection preceding the <br>. Bad. I have tightened the check to see that only a <br> is
3030         selected.
3031
3032         Fixing the bug above was accomplished with an additional call to DOM::Position::downstream. This
3033         new use of the function exposed this bug:
3034
3035         <rdar://problem/3907666> Incorrectly coded loop in Position::downstream can lead to infinite loop
3036
3037         * khtml/xml/dom_position.cpp:
3038         (DOM::Position::downstream): I am ashamed of my first cut at this. Rewrote the loop so it does
3039         not have this fatal flaw. It is a much better design as well.
3040
3041         * layout-tests/editing/deleting/delete-3800834-fix-expected.txt: Changes made this test
3042         have what I consider to be a better result. Going with it.
3043
3044 2004-12-06  Chris Blumenberg  <cblu@apple.com>
3045
3046         Fixed: <rdar://problem/3871718> REGRESSION (125-168): text marked bold with font that does not have bold variant copies as non-bold
3047
3048         Reviewed by hyatt.
3049
3050         * kwq/DOM.mm:
3051         (-[DOMElement _font]): new SPI for AppKit
3052         * kwq/DOMPrivate.h:
3053
3054 2004-12-06  Darin Adler  <darin@apple.com>
3055
3056         Reviewed by Maciej.
3057
3058         - fixed <rdar://problem/3906327> Select All of a large document is slow (>15 secs on my machine for attached specimen)
3059
3060         * kwq/KWQScrollView.mm: (QScrollView::updateContents): Intersect with visibleRect before calling through
3061         to NSView to dirty; NSView could also be more efficient in this case (I filed 3906343).
3062
3063 2004-12-06  John Sullivan  <sullivan@apple.com>
3064
3065         Darin found what appears to be the real leak that we were falsely blaming
3066         on the 'leaks' tool (3880245). I made the change, and ran layout tests and PLT to make
3067         sure nothing barfed.
3068
3069         * khtml/css/cssparser.cpp:
3070         (CSSParser::parseValue):
3071         call clearProperties() instead of just setting numParsedProperties to 0
3072         (CSSParser::parseDeclaration):
3073         ditto
3074         (CSSParser::createStyleDeclaration):
3075         ditto
3076
3077 2004-12-06  Ken Kocienda  <kocienda@apple.com>
3078
3079         Reviewed by me
3080
3081         New layout tests.
3082
3083         * layout-tests/editing/inserting/insert-div-001-expected.txt: Added.
3084         * layout-tests/editing/inserting/insert-div-001.html: Added.
3085         * layout-tests/editing/inserting/insert-div-002-expected.txt: Added.
3086         * layout-tests/editing/inserting/insert-div-002.html: Added.
3087         * layout-tests/editing/inserting/insert-div-003-expected.txt: Added.
3088         * layout-tests/editing/inserting/insert-div-003.html: Added.
3089         * layout-tests/editing/inserting/insert-div-004-expected.txt: Added.
3090         * layout-tests/editing/inserting/insert-div-004.html: Added.
3091         * layout-tests/editing/inserting/insert-div-005-expected.txt: Added.
3092         * layout-tests/editing/inserting/insert-div-005.html: Added.
3093         * layout-tests/editing/inserting/insert-div-006-expected.txt: Added.
3094         * layout-tests/editing/inserting/insert-div-006.html: Added.
3095         * layout-tests/editing/inserting/insert-div-007-expected.txt: Added.
3096         * layout-tests/editing/inserting/insert-div-007.html: Added.
3097         * layout-tests/editing/inserting/insert-div-008-expected.txt: Added.
3098         * layout-tests/editing/inserting/insert-div-008.html: Added.
3099         * layout-tests/editing/inserting/insert-div-009-expected.txt: Added.
3100         * layout-tests/editing/inserting/insert-div-009.html: Added.
3101
3102 2004-12-06  Ken Kocienda  <kocienda@apple.com>
3103
3104         Reviewed by Harrison
3105
3106         Fix for this bug:
3107
3108         <rdar://problem/3906948> REGRESSION (Mail): Insert paragraph code can make the insertion point "stick" in place.
3109
3110         * khtml/editing/htmlediting.cpp:
3111         (khtml::InsertParagraphSeparatorCommand::doApply): Call insertBlockPlaceholderIfNeeded(), passing block
3112         being added to this function. This ensures that the added block has a height.
3113         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Ditto.
3114
3115 2004-12-06  Ken Kocienda  <kocienda@apple.com>
3116
3117         Reviewed by John
3118
3119         * khtml/dom/dom_string.cpp:
3120         (DOM::DOMString::substring): Expose method already on DOMStrimgImpl.
3121         * khtml/dom/dom_string.h: Ditto.
3122         * khtml/editing/htmlediting.cpp:
3123         (khtml::CompositeEditCommand::rebalanceWhitespace): New helper to create and execute a
3124         RebalanceWhitespaceCommand instance.
3125         (khtml::DeleteSelectionCommand::doApply): Call rebalanceWhitespace() after running command.
3126         (khtml::InsertLineBreakCommand::doApply): Ditto.
3127         (khtml::InsertParagraphSeparatorCommand::doApply): Ditto.
3128         (khtml::InsertParagraphSeparatorInQuotedContentCommand::doApply): Ditto.
3129         (khtml::InsertTextCommand::input): Ditto.
3130         (khtml::RebalanceWhitespaceCommand::RebalanceWhitespaceCommand): New command.
3131         (khtml::RebalanceWhitespaceCommand::~RebalanceWhitespaceCommand): Ditto.
3132         (khtml::RebalanceWhitespaceCommand::doApply): Ditto.
3133         (khtml::RebalanceWhitespaceCommand::doUnapply): Ditto.
3134         (khtml::RebalanceWhitespaceCommand::preservesTypingStyle): Ditto.
3135         (khtml::ReplaceSelectionCommand::completeHTMLReplacement): Ditto.
3136         * khtml/editing/htmlediting.h: Ditto.
3137         (khtml::RebalanceWhitespaceCommand::): Ditto.
3138
3139 2004-12-05  Darin Adler  <darin@apple.com>
3140
3141         - fixed small problem in my check-in from yesterday
3142
3143         * kwq/KWQEvent.mm:
3144         (positionForEvent): Get location from event without raising exception if it's the wrong type.
3145         (clickCountForEvent): Same, for clickCount.
3146         (QMouseEvent::QMouseEvent): Use the new helper functions so this can be constructed even with
3147         the wrong type of NSEvent. Required for cases where a keyboard event causes a "click" and we need
3148         to synthesize a QMouseEvent for KHTML internal use, using the key down NSEvent.
3149
3150 2004-12-04  Darin Adler  <darin@apple.com>
3151
3152         Reviewed by John.
3153
3154         - fixed <rdar://problem/3878329> REGRESSION (169-170): colors are wrong for my.yahoo.com due to CSS background parsing changes
3155
3156         * khtml/css/cssparser.cpp: (CSSParser::parseBackgroundShorthand): Changed function so it doesn't rely on the position
3157         attribute being at the end of the array and then moved position attribute before color attribute so it takes precedence.
3158         Since "0" can be both the X coordinate of a position and a legal color (meaning black), we need to do position first.
3159
3160         - fixed <rdar://problem/3760869> click events for input type=button or type=checkbox don't have flags like shiftKey set
3161
3162         * khtml/rendering/render_form.h: Remove unused RenderFormElement fields.
3163         * khtml/rendering/render_form.cpp:
3164         (RenderFormElement::RenderFormElement): Take out code to set a bunch of unused fields.
3165         (RenderFormElement::slotClicked): Change to create the QMouseEvent from the actual mouse event rather than
3166         creating it with all the flags set to 0, using the new QMouseEvent constructor that does so.
3167
3168         * kwq/KWQEvent.h: Added constructor that takes no parameters which uses the current event from AppKit.
3169         Made the click count getter const and added an isDoubleClick that matches the logic used elsewhere.
3170         Added a fixState helper method so the constructors can save code.
3171         * kwq/KWQEvent.mm:
3172         (QMouseEvent::QMouseEvent): Factored out the state-fixing code that was in the two existing constructors
3173         and added a third constructor that uses the "current event" from AppKit (used above).
3174         (QMouseEvent::fixState): Compute state and click count based on event type.
3175
3176         - fixed first symptom of <rdar://problem/3830936> REGRESSION (125-165): crash due to null font family, hang at changeforamerica.com
3177
3178         * kwq/KWQFontFamily.mm: (KWQFontFamily::getNSFamily): Handle empty strings specially so we don't run into trouble when
3179         the family name is a null string. This prevents the crash, but there are still other problems that may have the same
3180         underlying cause in CSS.
3181
3182         - fixed <rdar://problem/3829808> Safari crashes when adding a DOM node that was removed from an XMLHTTP request result
3183
3184         * khtml/html/html_baseimpl.cpp: (HTMLBodyElementImpl::insertedIntoDocument):
3185         Added nil check before calling scheduleRelayout. This is new code so the nil-dereference is a recent regression.
3186
3187 2004-12-03  Chris Blumenberg  <cblu@apple.com>
3188
3189         New fixes for:
3190         <rdar://problem/3685766> WebDataSource is missing subresources when they use cached WebCore data
3191         <rdar://problem/3722434> REGRESSION?: Assertion failure trying to drag image in iframe (itapema.sc.gov.br)
3192         <rdar://problem/3903173> REGRESSION (172-TOT): assertion failure and crash in slotAllData logging into hotmail account
3193         <rdar://problem/3902749> REGRESSION (Tiger): missing image symbol does not appear
3194
3195         Reviewed by darin, rjw, kocienda.
3196
3197         * khtml/misc/loader.cpp:
3198         (CachedObject::~CachedObject):
3199         (CachedCSSStyleSheet::checkNotify):
3200         (Loader::servePendingRequests):
3201         (Loader::slotFinished):
3202         (Loader::slotReceivedResponse):
3203         (Cache::requestImage):
3204         (Cache::requestScript):
3205         * khtml/misc/loader.h:
3206         (khtml::CachedObject::CachedObject):
3207         (khtml::CachedObject::response):
3208         (khtml::CachedObject::allData):
3209         * kwq/KWQKJobClasses.h:
3210         * kwq/KWQKJobClasses.mm:
3211         (KIO::TransferJobPrivate::TransferJobPrivate):
3212         (KIO::TransferJobPrivate::~TransferJobPrivate):
3213         (KIO::TransferJob::TransferJob):
3214         (KIO::TransferJob::assembleResponseHeaders):
3215         (KIO::TransferJob::retrieveCharset):
3216         (KIO::TransferJob::emitResult):
3217         (KIO::TransferJob::emitReceivedResponse):
3218         * kwq/KWQLoader.h:
3219         * kwq/KWQLoader.mm:
3220         (KWQHeaderStringFromDictionary):
3221         (KWQCheckCacheObjectStatus):
3222         (KWQIsResponseURLEqualToURL):
3223         (KWQResponseURL):
3224         (KWQResponseMIMEType):
3225         (KWQCacheObjectExpiresTime):
3226         (khtml::CachedObject::setResponse):
3227         (khtml::CachedObject::setAllData):
3228         * kwq/KWQPixmap.h:
3229         * kwq/KWQPixmap.mm:
3230         (QPixmap::QPixmap):
3231         * kwq/KWQResourceLoader.mm:
3232         (-[KWQResourceLoader finishJobAndHandle:]):
3233         (-[KWQResourceLoader cancel]):
3234         (-[KWQResourceLoader reportError]):
3235         (-[KWQResourceLoader finishWithData:]):
3236         * kwq/KWQSignal.h:
3237         * kwq/KWQSignal.mm:
3238         (KWQSignal::call):
3239         * kwq/KWQSlot.h:
3240         * kwq/KWQSlot.mm:
3241         (KWQSlot::KWQSlot):
3242         (KWQSlot::call):
3243         * kwq/WebCoreBridge.h:
3244         * kwq/WebCoreResourceLoader.h:
3245
3246 2004-12-04  Darin Adler  <darin@apple.com>
3247
3248         Reviewed by John.
3249
3250         - fixed <rdar://problem/3876093> REGRESSION (166-167): Setting slider control's value from JavaScript has no effect (breaks RSS)
3251
3252         * khtml/rendering/render_form.cpp:
3253         (RenderSlider::updateFromElement): Call setValue to update the value of the DOM element rather than
3254         modifying the m_value data member directly. We don't use m_value at all for sliders now, and in fact
3255         the code relies on the fact that m_value is null. Setting m_value to a non-null value was causing the bug.
3256         (RenderSlider::slotSliderValueChanged): Ditto.
3257
3258 2004-12-03  John Sullivan  <sullivan@apple.com>
3259
3260         Reviewed by Ken.
3261
3262         - fixed <rdar://problem/3889411> REGRESSION (125-172): repro crash in
3263         khtml::BackgroundLayer::cullEmptyLayers
3264
3265         * khtml/rendering/render_style.cpp:
3266         (BackgroundLayer::cullEmptyLayers):
3267         added missing nil check
3268
3269 === Safari-173 ===
3270
3271 2004-12-03  Ken Kocienda  <kocienda@apple.com>
3272
3273         Reviewed by me
3274
3275         Roll out some recent changes by Chris that caused a performance regression.
3276         Fix is in hand, but it is a little risky this close to a submission. So,
3277         we have decided to roll back the change with the regression and roll in
3278         the new code after we submit.
3279
3280         * khtml/css/cssproperties.c:
3281         (hash_prop):
3282         (findProp):
3283         * khtml/css/cssvalues.c:
3284         (hash_val):
3285         (findValue):
3286         * khtml/misc/htmlattrs.c:
3287         (hash_attr):
3288         (findAttr):
3289         * khtml/misc/htmltags.c:
3290         (hash_tag):
3291         (findTag):
3292         * khtml/misc/loader.cpp:
3293         (CachedObject::~CachedObject):
3294         (CachedObject::setResponse):
3295         (CachedCSSStyleSheet::checkNotify):
3296         (Loader::servePendingRequests):
3297         (Loader::slotFinished):
3298         (Loader::slotReceivedResponse):
3299         (Cache::requestImage):
3300         (Cache::requestScript):
3301         * khtml/misc/loader.h:
3302         (khtml::CachedObject::CachedObject):
3303         (khtml::CachedObject::response):
3304         * kwq/KWQKJobClasses.h:
3305         * kwq/KWQKJobClasses.mm:
3306         (KIO::TransferJobPrivate::TransferJobPrivate):
3307         (KIO::TransferJobPrivate::~TransferJobPrivate):
3308         (KIO::TransferJob::TransferJob):
3309         (KIO::TransferJob::assembleResponseHeaders):
3310         (KIO::TransferJob::retrieveCharset):
3311         (KIO::TransferJob::emitResult):
3312         (KIO::TransferJob::emitReceivedResponse):
3313         * kwq/KWQLoader.h:
3314         * kwq/KWQLoader.mm:
3315         (KWQHeaderStringFromDictionary):
3316         (KWQCheckCacheObjectStatus):
3317         (KWQRetainResponse):
3318         (KWQReleaseResponse):
3319         (KWQIsResponseURLEqualToURL):
3320         (KWQResponseURL):
3321         (KWQResponseMIMEType):
3322         (KWQResponseTextEncodingName):
3323         (KWQResponseHeaderString):
3324         (KWQCacheObjectExpiresTime):
3325         (KWQLoader::KWQLoader):
3326         * kwq/KWQPixmap.h:
3327         * kwq/KWQPixmap.mm:
3328         (QPixmap::QPixmap):
3329         * kwq/KWQResourceLoader.mm:
3330         (-[KWQResourceLoader finishJobAndHandle]):
3331         (-[KWQResourceLoader cancel]):
3332         (-[KWQResourceLoader reportError]):
3333         (-[KWQResourceLoader finish]):
3334         * kwq/KWQSignal.h:
3335         * kwq/KWQSignal.mm:
3336         (KWQSignal::call):