WebCore/ChangeLog

   1 2009-03-25  Darin Adler  <darin@apple.com>
   2
   3         Reviewed by John Sullivan.
   4
   5         Tidy up LegacyWebArchive a bit. And don't include favicons in web archives when
   6         they are being used only for selections, not an entire document.
   7
   8         * WebCore.base.exp: Remove exported function that's not used in WebKit.
   9
  10         * loader/archive/cf/LegacyWebArchive.cpp:
  11         (WebCore::LegacyWebArchive::createPropertyListRepresentation): Made a static member
  12         function so it can share private things with other member functions. Added some FIXME
  13         comments to a branch of code I think is dead. Changed boolean argument into a named
  14         one. Renamed to match other similar functions.
  15         (WebCore::LegacyWebArchive::createResourceResponseFromPropertyListData): Ditto.
  16         (WebCore::LegacyWebArchive::createResource): Ditto.
  17         (WebCore::LegacyWebArchive::create): Merged the create and init functions.
  18         (WebCore::LegacyWebArchive::rawDataRepresentation): Added some assertions because
  19         we should never fail to write (we can fail to read).
  20         (WebCore::LegacyWebArchive::createResourceResponseFromMacArchivedData): Made a static
  21         member function as above.
  22         (WebCore::LegacyWebArchive::createFromSelection): Ditto.
  23         (WebCore::LegacyWebArchive::create): Iterate the vector with indices rather than
  24         iterators. Only include the favicon if the first node is the document since we don't
  25         want to include the favicon when copying and pasting.
  26
  27         * loader/archive/cf/LegacyWebArchive.h: Made one of the create functions private.
  28         Made the Mac-specific functions be static member functions. Made other helpers be
  29         static member functions. Removed the nit function.
  30
  31         * loader/archive/cf/LegacyWebArchiveMac.mm:
  32         (WebCore::LegacyWebArchive::createResourceResponseFromMacArchivedData): Updated
  33         comment and made this a static member function.
  34         (WebCore::LegacyWebArchive::createPropertyListRepresentation): Ditto. Also made
  35         it so we only type cast in one place.
  36
  37 2009-03-25  Darin Adler  <darin@apple.com>
  38
  39         Reviewed by David Hyatt.
  40
  41         Bug 24740: crash in RenderSlider::setPositionFromValue when calling pause() after setting <video> to display: none
  42         https://bugs.webkit.org/show_bug.cgi?id=24740
  43         rdar://problem/6679873
  44
  45         Bug 12104: Native Slider: When the thumb's height is specified as a percentage, it is not centered properly
  46         https://bugs.webkit.org/show_bug.cgi?id=12104
  47
  48         Test: media/video-display-none-crash.html
  49
  50         The problem here was that RenderSlider was trying to position its thumb in a way that
  51         requires it to call updateLayout inside rendering code. The right way to position a
  52         child renderer is to do layout, so I had to write a layout function. And then fix a few
  53         other small problems shown by the same test case.
  54
  55         * rendering/RenderSlider.cpp: Made defaultTrackLength have internal linkage.
  56         Renamed HTMLSliderThumbElement to SliderThumbElement because we only use the HTML
  57         prefix for public DOM classes, not elements used as implementation details.
  58         Made SliderThumbElement function members private and got rid of unneeded default
  59         argument value for shadowParent.
  60         (WebCore::SliderRange::SliderRange): Added. Parses precision, max, and min attributes.
  61         (WebCore::SliderRange::clampValue): Added. Does standard clamping based on the above.
  62         (WebCore::SliderRange::valueFromElement): Added. Reads the value from the element in
  63         a way that clamps to the range.
  64         (WebCore::sliderPosition): Added. Computes the slider position: a double.
  65         (WebCore::SliderThumbElement::SliderThumbElement): Removed unneeded explicit
  66         initialization of m_initialClickPoint.
  67         (WebCore::SliderThumbElement::defaultEventHandler): Call setValueForPosition instead
  68         of calling setCurrentPosition and valueChanged.
  69         (WebCore::RenderSlider::RenderSlider): Remove unneeded explicit initialization of m_thumb.
  70         (WebCore::RenderSlider::styleDidChange): Remove unneeded second argument to createThumbStyle.
  71         (WebCore::RenderSlider::createThumbStyle): Remove unneeded second argument. Get rid of code
  72         setting the position to relative and setting the left and top. We now handle positioning
  73         in a custom layout function.
  74         (WebCore::RenderSlider::layout): Rewrote to handle positioning of the thumb as layout.
  75         (WebCore::RenderSlider::updateFromElement): Added code to immediately update the value
  76         in the element if it's out of range. This clamping used to be done as a side effect of
  77         setPositionFromValue. Also, this has nothing to do with the renderer, so at some point
  78         it could be moved into HTMLInputElement. Removed call to setPositionFromValue
  79         and instead just rely on the call to setNeedsLayout. Fix the setNeedsLayout call to be
  80         a normal setNeedsLayout(true), not a setNeedsLayout(true, false), because we do want
  81         this to be propagated to the parent -- it's not called during layout.
  82         (WebCore::RenderSlider::setValueForPosition): Refactor to use the new SliderRange
  83         class. Also don't call setCurrentPosition; instead just call setNeedsLayout.
  84         (WebCore::RenderSlider::currentPosition): Use the actual position of the renderer rather
  85         than the style to find the position; that means this needs to be done after layout is done.
  86         Also removed unneeded runtime checks and replaced them with assertions, after checking
  87         all callers to see they already guarantee this.
  88         (WebCore::RenderSlider::trackSize): Removed unneeded runtime checks and replaced them
  89         with assertions, after checking all callers to see they already guarantee this.
  90         (WebCore::RenderSlider::inDragMode): Added a null check for m_thumb so this won't
  91         crash if called early on a brand new RenderSlider.
  92
  93         * rendering/RenderSlider.h: Made all functions private except for forwardEvent and inDragMode.
  94         Renamed HTMLSliderThumbElement to SliderThumbElement because we only use the HTML
  95         prefix for public DOM classes, not elements used as implementation details. Made the
  96         mouseEventIsInThumb function non-virtual. Removed the return value and argument from
  97         setPositionFromValue. Removed valueChanged and setCurrentPosition. Removed the oldStyle
  98         argument to createThumbStyle (see above). Made SliderThumbElement a friend so it can use some
  99         private member functions.
 100
 101 2009-03-25  Eli Fidler  <eli.fidler@torchmobile.com>
 102
 103         Reviewed by George Staikos.
 104
 105         Fix ambiguous else cases by adding braces to nested if()s with elses.
 106
 107         * loader/MainResourceLoader.cpp:
 108         (WebCore::MainResourceLoader::continueAfterContentPolicy):
 109         * page/EventHandler.cpp:
 110         (WebCore::EventHandler::updateDragAndDrop):
 111         * rendering/RenderListMarker.cpp:
 112         (WebCore::toArmenianUnder10000):
 113         * rendering/TextControlInnerElements.cpp:
 114         (WebCore::TextControlInnerTextElement::defaultEventHandler):
 115
 116 2009-03-24  Eli Fidler  <eli.fidler@torchmobile.com>
 117
 118         Reviewed by George Staikos.
 119
 120         Move variable into proper platform block to quiet warning.
 121
 122         * html/CanvasRenderingContext2D.cpp:
 123         (WebCore::CanvasRenderingContext2D::drawTextInternal):
 124
 125 2009-03-24  Eric Carlson  <eric.carlson@apple.com>
 126
 127         Reviewed by Antti Koivisto.
 128
 129         Fix for <rdar://problem/6719375> Deal with QTKit loadstate changes when playing streaming movies
 130
 131         *  WebCore/html/HTMLMediaElement.cpp:
 132         (WebCore::HTMLMediaElement::setNetworkState): Deal with media engine reviving after having
 133         network state of NETWORK_NO_SOURCE.
 134
 135         *  WebCore/platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
 136         (WebCore::MediaPlayerPrivate::updateStates): Reset m_readyState at each state change since
 137         some types of movies currently cause QTKit's load state to bounce around. QTMovieLoadStatePlaythroughOK
 138         corresponds to HaveFutureData, not HaveEnoughData.
 139
 140 2009-03-24  Adele Peterson  <adele@apple.com>
 141
 142         Reviewed by Darin Adler.
 143
 144         Fix for https://bugs.webkit.org/show_bug.cgi?id=24707
 145         <rdar://problem/6593021> Deleting with a caret in a blank, quoted line decreases the quoting, but does not delete the line
 146
 147         * editing/TypingCommand.cpp: (WebCore::TypingCommand::deleteKeyPressed):
 148         After breaking out of an empty mail blockquote, we still want continue with the deletion
 149         so actual content will get deleted, and not just the quote style.
 150
 151 2009-03-24  Darin Adler  <darin@apple.com>
 152
 153         Reviewed by Brady Eidson.
 154
 155         <rdar://problem/6624662> REGRESSION (r39904): can't save certain web pages as web archive (ones without favicons!)
 156
 157         * loader/archive/cf/LegacyWebArchive.cpp:
 158         (WebCore::LegacyWebArchive::create): Added a null check.
 159
 160 2009-03-24  Adam Langley  <agl@google.com>
 161
 162         Reviewed by Darin Fisher.
 163
 164         https://bugs.webkit.org/show_bug.cgi?id=24685
 165
 166         The Skia drawLine function takes subpixel values and attempts to draw
 167         a line with subpixel precision. This is complex and slow for drawing
 168         scrollbars which use only vertical and horizontal lines.
 169
 170         This changes the Chromium Linux scrollbar code to use drawIRect.
 171
 172         * platform/chromium/ScrollbarThemeChromiumLinux.cpp:
 173         (WebCore::ScrollbarThemeChromium::invalidateOnMouseEnterExit):
 174         (WebCore::drawVertLine):
 175         (WebCore::drawHorizLine):
 176         (WebCore::drawBox):
 177         (WebCore::ScrollbarThemeChromium::paintTrackPiece):
 178         (WebCore::ScrollbarThemeChromium::paintThumb):
 179
 180 2009-03-24  David Kilzer  <ddkilzer@apple.com>
 181
 182         Bug 23310: Setting an absolute path (/abs) on an <iframe> with no src doesn't resolve the URL properly
 183
 184         <https://bugs.webkit.org/show_bug.cgi?id=23310>
 185
 186         Reviewed by Darin Adler.
 187
 188         Test: fast/frames/iframe-no-src-set-location.html
 189
 190         * dom/Document.cpp:
 191         (WebCore::Document::completeURL): If m_baseURL is empty or is
 192         about:blank and we have a parent document, use the parent
 193         document's URL for the base when completing a new URL.
 194
 195 2009-03-24  Simon Fraser  <simon.fraser@apple.com>
 196
 197         Reviewed by Dave Hyatt
 198
 199         https://bugs.webkit.org/show_bug.cgi?id=24659
 200
 201         When hit-testing 3d-transformed layers, it doesn't make sense to project the hitTestRect
 202         into the coordinate space of the layer, and doing so can result in pathalogical quads
 203         that break hit testing. In that case, simply use the same bounds as used for painting,
 204         which are the composited bounds for this layer.
 205
 206         Test: transforms/3d/hit-testing/rotated-hit-test.html
 207
 208         * rendering/RenderLayer.cpp:
 209         (WebCore::RenderLayer::hitTestLayer):
 210
 211 2009-03-24  Simon Fraser  <simon.fraser@apple.com>
 212
 213         Reviewed by Dave Hyatt
 214
 215         https://bugs.webkit.org/show_bug.cgi?id=24436
 216
 217         When compositing is enabled, painting and hit testing end up using different
 218         clipping roots, because, for painting, every composited layer is a clipping root,
 219         but for hit testing only layers with transforms are. To fix this, we use
 220         temporary clip rects for hit testing, if the page is in compositing mode.
 221
 222         * rendering/RenderLayer.cpp:
 223         (WebCore::RenderLayer::paintLayer):
 224         (WebCore::RenderLayer::hitTestLayer):
 225         (WebCore::RenderLayer::parentClipRects):
 226         (WebCore::RenderLayer::calculateRects):
 227         * rendering/RenderLayer.h:
 228
 229 2009-03-24  Simon Fraser  <simon.fraser@apple.com>
 230
 231         Reviewed by Dan Bernstein
 232
 233         https://bugs.webkit.org/show_bug.cgi?id=24782
 234
 235         Fix regression with CSS clip rects with non-length values, by reverting to
 236         the original code, but passing the zoom multiplier through convertToLength().
 237
 238         Also make stylistic change in createTransformOperations(), renaming 'inStyle'
 239         to 'style.
 240
 241         Tested by existing tests.
 242
 243         * css/CSSStyleSelector.cpp:
 244         (WebCore::convertToLength):
 245         (WebCore::CSSStyleSelector::applyProperty):
 246         (WebCore::CSSStyleSelector::createTransformOperations):
 247
 248 2009-03-24  David Hyatt  <hyatt@apple.com>
 249
 250         Reviewed by Simon Fraser
 251
 252         https://bugs.webkit.org/show_bug.cgi?id=21789, overflow:hidden elements should clip their foreground contents
 253         to a border-radius.
 254
 255         This patch makes non-self-painting overflow layers and control clips work with border-radius.  RenderLayers
 256         that should be clipped by overflow ancestors with border-radius are still broken.
 257
 258         Added fast/clip/overflow-border-radius-clip.html
 259
 260         * rendering/RenderBox.cpp:
 261         (WebCore::RenderBox::pushContentsClip):
 262
 263 2009-03-24  David Hyatt  <hyatt@apple.com>
 264
 265         Reviewed by Simon Fraser
 266
 267         https://bugs.webkit.org/show_bug.cgi?id=21789, overflow:hidden elements should clip their foreground contents.
 268
 269         This first patch makes overflow:hidden properly clip the foreground contents of overflow:hidden replaced elements.  Common
 270         replaced elements now default to overflow:hidden in the UA stylesheet (this is what the spec specifically recommends be
 271         done).
 272
 273         Added fast/replaced/border-radius-clip.html
 274
 275         * css/html4.css:
 276         * rendering/RenderReplaced.cpp:
 277         (WebCore::RenderReplaced::paint):
 278         * rendering/RenderWidget.cpp:
 279         (WebCore::RenderWidget::paint):
 280
 281 2009-03-24  Adele Peterson  <adele@apple.com>
 282
 283         RS by Mark Rowe.
 284
 285         Only build these Mail quirks checks in on the Mac.
 286
 287         * dom/Document.cpp: (WebCore::disableRangeMutation):
 288         * html/HTMLElement.cpp: (WebCore::HTMLElement::inEitherTagList):
 289
 290 2009-03-24  Brent Fulgham  <bfulgham@webkit.org>
 291
 292         Build fix, no review.
 293
 294         Remove ResourceLoaderCFNet.cpp from build list for Cairo Releas
 295         and Debug targets.
 296
 297         * WebCore.vcproj/WebCore.vcproj:
 298
 299 2009-03-24  Eric Carlson  <eric.carlson@apple.com>
 300
 301         Fix layout tests broken by r41907.
 302
 303         * html/HTMLMediaElement.cpp:
 304         (WebCore::HTMLMediaElement::setReadyState): Don't set "was playing" based on the new ready state
 305
 306 2009-03-24  Dmitry Titov  <dimich@chromium.org>
 307
 308         Reviewed by Dimitri Glazkov.
 309
 310         https://bugs.webkit.org/show_bug.cgi?id=24689
 311         Fix Chromium compilation errors.
 312
 313         * bindings/v8/WorkerContextExecutionProxy.cpp:
 314         (WebCore::WorkerContextExecutionProxy::FindOrCreateEventListener):
 315         (WebCore::WorkerContextExecutionProxy::RemoveEventListener):
 316
 317 2009-03-24  Yury Semikhatsky  <yurys@chromium.org>
 318
 319         Reviewed by Darin Adler.
 320
 321         https://bugs.webkit.org/show_bug.cgi?id=24759
 322         Add missing methods Element.prototype.removeMatchingStyleClasses and
 323         Node.prototype.enclosingNodeOrSelfWithNodeNameInArray to SourceFrame content
 324         iframe. These methods are called by Element.prototype.removeStyleClass and
 325         Node.prototype.enclosingNodeOrSelfWithNodeName.
 326
 327         * inspector/front-end/SourceFrame.js:
 328         (WebInspector.SourceFrame.prototype._loaded):
 329
 330 2009-03-24  Mark Mentovai  <mark@chromium.org>
 331
 332         Reviewed by Mark Rowe.
 333
 334         https://bugs.webkit.org/show_bug.cgi?id=24653
 335         WebKit should be buildable without prefix header injection.
 336         Adds missing #includes and forward declarations as needed.
 337
 338         * editing/SmartReplaceCF.cpp:
 339         * platform/graphics/mac/ColorMac.mm:
 340         * platform/graphics/mac/FontCacheMac.mm:
 341         * platform/graphics/mac/FontCustomPlatformData.h:
 342         * platform/graphics/mac/FontMac.mm:
 343         * platform/graphics/mac/FontMacATSUI.mm:
 344         * platform/graphics/mac/FontPlatformData.h:
 345         * platform/graphics/mac/FontPlatformDataMac.mm:
 346         * platform/graphics/mac/GraphicsContextMac.mm:
 347         * platform/graphics/mac/SimpleFontDataMac.mm:
 348         * platform/mac/FoundationExtras.h:
 349         * platform/mac/LocalCurrentGraphicsContext.h:
 350         * platform/mac/WebCoreSystemInterface.h:
 351         * platform/mac/WebCoreSystemInterface.mm:
 352         * platform/mac/WebCoreTextRenderer.h:
 353         * platform/mac/WebCoreTextRenderer.mm:
 354         * platform/mac/WebFontCache.h:
 355         * platform/mac/WebFontCache.mm:
 356         * platform/text/PlatformString.h:
 357         * platform/text/mac/ShapeArabic.c:
 358         * platform/text/mac/StringMac.mm:
 359         * rendering/RenderThemeChromiumMac.h:
 360
 361 2009-03-24  Dan Bernstein  <mitz@apple.com>
 362
 363         Reviewed by Darin Adler.
 364
 365         - fix <rdar://problem/6107874> by capping the nesting depth of
 366           "block-level" elements generated by the parser
 367
 368         Test: fast/parser/block-nesting-cap.html
 369
 370         * html/HTMLParser.cpp:
 371         (WebCore::HTMLParser::HTMLParser):
 372         (WebCore::HTMLParser::insertNode):
 373         (WebCore::HTMLParser::pushBlock):
 374         (WebCore::HTMLParser::popOneBlockCommon):
 375         (WebCore::HTMLParser::freeBlock):
 376         * html/HTMLParser.h:
 377
 378 2009-03-23  Greg Bolsinga  <bolsinga@apple.com>
 379
 380         Reviewed by David Kilzer.
 381
 382         https://bugs.webkit.org/show_bug.cgi?id=24771
 383
 384         DOMTimeStamps are based upon 1970.
 385
 386         * platform/mac/GeolocationServiceMac.mm:
 387         (-[WebCoreCoreLocationObserver locationManager:didUpdateToLocation:fromLocation:]):
 388
 389 2009-03-23  Eric Seidel  <eric@webkit.org>
 390
 391         Build fix, no review.
 392
 393         * rendering/RenderObject.cpp:
 394         (WebCore::RenderObject::createVisiblePosition): change .container to .isNotNull()
 395
 396 2009-03-23  Eric Seidel  <eric@webkit.org>
 397
 398         Reviewed by Darin Adler.
 399
 400         Rename Position::container to m_anchorNode and make it private
 401         https://bugs.webkit.org/show_bug.cgi?id=24760
 402
 403         More code cleanup for Position.
 404
 405         Change all uses of m_container to node()
 406         Eventually most uses of node() should change to anchorNode() to designate
 407         that it's the node the Position is anchored to, but not necessarily the
 408         container of the position (it could be the before/after neighbor).
 409
 410         Remove any code which sets m_container, and change it to use a new
 411         Position::moveToPosition function which takes a node and offset.
 412         It never makes sense to change the node and leave the offset.
 413
 414         * dom/Position.h:
 415         (WebCore::Position::Position):
 416         (WebCore::Position::clear):
 417         (WebCore::Position::anchorNode):
 418         (WebCore::Position::node):
 419         (WebCore::Position::moveToPosition):
 420         (WebCore::Position::moveToOffset):
 421         (WebCore::Position::isNull):
 422         (WebCore::Position::isNotNull):
 423         (WebCore::operator==):
 424         * dom/Range.cpp:
 425         (WebCore::Range::create):
 426         (WebCore::Range::compareBoundaryPoints):
 427         * dom/RangeBoundaryPoint.h:
 428         (WebCore::RangeBoundaryPoint::container):
 429         (WebCore::RangeBoundaryPoint::set):
 430         (WebCore::RangeBoundaryPoint::setOffset):
 431         (WebCore::RangeBoundaryPoint::setToChild):
 432         (WebCore::RangeBoundaryPoint::setToStart):
 433         (WebCore::RangeBoundaryPoint::setToEnd):
 434
 435 2009-03-17  Eric Seidel  <eric@webkit.org>
 436
 437         Reviewed by David Hyatt.
 438
 439         document.write() should be able to make a document strict mode
 440         https://bugs.webkit.org/show_bug.cgi?id=24336
 441
 442         Remove an implicit write of "<html>" on the first document.write call
 443         this was added as part of a KDE import http://trac.webkit.org/changeset/798
 444         with no layout test or explanation.  I can't think of any reason why
 445         an implicit <html> write is necessary (or correct), so I'm removing it and
 446         adding a test for the correct behavior.  The parser will add any necessary
 447         HTMLHTMLElements during the write() anyway.
 448
 449         Our behavior is now tested by fast/dom/Document/document-write-doctype
 450         and matches IE, FF fails this new test.  Mozilla bug filed:
 451         https://bugzilla.mozilla.org/show_bug.cgi?id=483908
 452
 453         * dom/Document.cpp:
 454         (WebCore::Document::write):
 455
 456 2009-03-23  Sam Weinig  <sam@webkit.org>
 457
 458         Reviewed by Dan Bernstein.
 459
 460         Fix for <rdar://problem/6140966>
 461         Empty Caches does not clear the Cross-site XMLHttpRequest preflight cache
 462
 463         * WebCore.base.exp:
 464         * WebCore.xcodeproj/project.pbxproj:
 465         * loader/CrossOriginPreflightResultCache.cpp:
 466         (WebCore::CrossOriginPreflightResultCache::empty):
 467         * loader/CrossOriginPreflightResultCache.h:
 468
 469 2009-03-23  Darin Adler  <darin@apple.com>
 470
 471         Reviewed by Adele Peterson.
 472
 473         Bug 24726: hit testing doesn't work right when the click is on anonymous content
 474         https://bugs.webkit.org/show_bug.cgi?id=24726
 475         rdar://problem/6696992
 476
 477         Test: editing/selection/hit-test-anonymous.html
 478
 479         * rendering/RenderBR.cpp:
 480         (WebCore::RenderBR::positionForPoint): Call createVisiblePosition instead of
 481         creating a VisiblePosition directly. It will handle finding non-anonymous
 482         content nearby if node() is 0.
 483         * rendering/RenderBlock.cpp:
 484         (WebCore::positionForPointRespectingEditingBoundaries): Ditto.
 485         (WebCore::positionForPointWithInlineChildren): Ditto.
 486         (WebCore::RenderBlock::positionForPoint): Ditto.
 487         * rendering/RenderBox.cpp:
 488         (WebCore::RenderBox::positionForPoint): Ditto.
 489         * rendering/RenderObject.cpp:
 490         (WebCore::RenderObject::positionForPoint): Ditto.
 491         (WebCore::RenderObject::createVisiblePosition): Added.
 492         * rendering/RenderObject.h: Added createVisiblePosition.
 493         * rendering/RenderReplaced.cpp:
 494         (WebCore::RenderReplaced::positionForPoint): Call createVisiblePosition.
 495         * rendering/RenderSVGInlineText.cpp:
 496         (WebCore::RenderSVGInlineText::positionForPoint): Ditto.
 497         * rendering/RenderText.cpp:
 498         (WebCore::RenderText::positionForPoint): Ditto.
 499
 500 2009-03-23  Adele Peterson  <adele@apple.com>
 501
 502         Reviewed by Darin Adler & Dave Hyatt.
 503
 504         Fix for <rdar://problem/6621310> REGRESSION(35185): Apple Travel HTML emails missing some style after Safari 4 upgrade
 505
 506         Leopard Mail doesn't expect <style> to be in the body.  This change reverts back to the old behavior of
 507         moving <style> to <head> for that version of Mail.
 508
 509         * html/HTMLElement.cpp: (WebCore::HTMLElement::inEitherTagList):
 510
 511 2009-03-23  Adele Peterson  <adele@apple.com>
 512
 513         Reviewed by Mark Rowe & Dave Hyatt.
 514
 515         Merge some of the individual Mail quirks into two settings that we can check for future quirks.
 516
 517         * WebCore.base.exp:
 518         * dom/Document.cpp:
 519         (WebCore::disableRangeMutation):
 520         (WebCore::Document::nodeChildrenChanged):
 521         (WebCore::Document::nodeWillBeRemoved):
 522         (WebCore::Document::textInserted):
 523         (WebCore::Document::textRemoved):
 524         (WebCore::Document::textNodesMerged):
 525         (WebCore::Document::textNodeSplit):
 526         * page/Settings.cpp:
 527         (WebCore::Settings::Settings):
 528         (WebCore::Settings::setNeedsLeopardMailQuirks):
 529         (WebCore::Settings::setNeedsTigerMailQuirks):
 530         * page/Settings.h:
 531         (WebCore::Settings::needsLeopardMailQuirks):
 532         (WebCore::Settings::needsTigerMailQuirks):
 533
 534 2009-03-23  Ada Chan  <adachan@apple.com>
 535
 536         https://bugs.webkit.org/show_bug.cgi?id=24762
 537         Support text-indent in <option> elements on windows platform.
 538
 539         Reviewed by Adele Peterson.
 540
 541         * css/themeWin.css:
 542         * platform/PopupMenuStyle.h: We don't honor font specified on <option> elements right now.  Make this
 543         explicit via windows themed default stylesheet.
 544         (WebCore::PopupMenuStyle::PopupMenuStyle): Also store text-indent and text-direction.
 545         (WebCore::PopupMenuStyle::textIndent):
 546         (WebCore::PopupMenuStyle::textDirection):
 547         * platform/win/PopupMenuWin.cpp:
 548         (WebCore::PopupMenu::paint): Adjust the text's x-coordinate if text-indent is supported for options and
 549         text-indent is specified with LTR direction.
 550         * rendering/RenderMenuList.cpp:
 551         (WebCore::RenderMenuList::updateOptionsWidth): Take text-indent into account if theme supports text-indent for options.
 552         (WebCore::RenderMenuList::itemStyle): Use new PopupMenuStyle constructor on windows.
 553         (WebCore::RenderMenuList::menuStyle): Ditto.
 554         * rendering/RenderTextControlSingleLine.cpp:
 555         (WebCore::RenderTextControlSingleLine::menuStyle): Ditto.
 556         * rendering/RenderTheme.h:
 557         (WebCore::RenderTheme::popupOptionSupportsTextIndent): Added.  Default is false since we are only supporting it in windows for now.
 558         * rendering/RenderThemeWin.h:
 559         (WebCore::RenderThemeWin::popupOptionSupportsTextIndent): Returns true for windows.
 560
 561 2009-03-23  Sam Weinig  <sam@webkit.org>
 562
 563         Reviewed by Anders Carlsson.
 564
 565         Fix for https://bugs.webkit.org/show_bug.cgi?id=24699
 566         REGRESSION: Java Applets broken
 567         <rdar://problem/6707494>
 568
 569         Fix loading Java applets without a codeBase. Only pass the base (up to the
 570         the last path component) of the baseURL to the plug-in.
 571
 572         * html/HTMLAppletElement.cpp:
 573         (WebCore::HTMLAppletElement::createRenderer):
 574         * platform/KURL.cpp:
 575         (WebCore::KURL::baseAsString):
 576         * platform/KURL.h:
 577
 578 2009-03-23  Darin Adler  <darin@apple.com>
 579
 580         Reviewed by Sam Weinig.
 581
 582         * platform/KURL.h: Removed now-incorrect comments.
 583         None of the parts include the separator characters any more, now that
 584         query doesn't include the "?", so the comments explaining which do and do
 585         not are no-longer helpful.
 586
 587 2009-03-23  David Levin  <levin@chromium.org>
 588
 589         Reviewed by Dimitri Glazkov.
 590
 591         https://bugs.webkit.org/show_bug.cgi?id=24764
 592
 593         Renamed files V8NodeFilter -> V8NodeFilterCondition to reflect
 594         class names.  Also fixed some headers that got mangled in search/replace
 595         operations.
 596
 597         No change in behavior, so no test.
 598
 599         * bindings/v8/V8NodeFilter.h: Removed.
 600         * bindings/v8/V8NodeFilterCondition.cpp: Renamed from WebCore/bindings/v8/V8NodeFilter.cpp.
 601         * bindings/v8/V8NodeFilterCondition.h: Added.
 602         * bindings/v8/custom/V8ClipboardCustom.cpp:
 603         * bindings/v8/custom/V8DocumentCustom.cpp:
 604         * bindings/v8/custom/V8ElementCustom.cpp:
 605         * bindings/v8/custom/V8HTMLCanvasElementCustom.cpp:
 606         * bindings/v8/custom/V8NavigatorCustom.cpp:
 607
 608 2009-03-23  Darin Fisher  <darin@chromium.org>
 609
 610         Reviewed by Antti Koivisto.
 611
 612         https://bugs.webkit.org/show_bug.cgi?id=24741
 613
 614         Adds a unique across-browser-sessions identifier to FormData, which may
 615         be used by ResourceHandle as a secondary cache key to enable cached
 616         form submissions.
 617
 618         At issue: two otherwise identical form submissions may result in
 619         completely independent responses, which may each be appropriate to
 620         store and reuse from cache.
 621
 622         * html/HTMLFormElement.cpp: Added call to FormData::setIdentifier so that
 623         we only enable cached form submissions for those generated by HTML.  This
 624         way we do not bother with POSTs generated by XMLHttpRequest.
 625         (WebCore::generateFormDataIdentifier):
 626         (WebCore::HTMLFormElement::createFormData):
 627         * platform/network/FormData.cpp: Initialize m_identifier to 0, which means
 628         the unspecified identifier.  So by default there is no identifier and
 629         nothing changes.
 630         (WebCore::FormData::FormData):
 631         * platform/network/FormData.h: Added m_identifier with setter and getter.
 632         (WebCore::FormData::setIdentifier):
 633         (WebCore::FormData::identifier):
 634
 635 2009-03-23  Simon Fraser  <simon.fraser@apple.com>
 636
 637         Reviewed by Antti Koivisto
 638
 639         https://bugs.webkit.org/show_bug.cgi?id=24733
 640
 641         Fix media controller with full-page zoom. Previously, the media controller
 642         shadow nodes never saw style changes on the RenderMedia, so did not respond
 643         to zooming at all. Now, we update the style on the shadow renderers
 644         whenever RenderMedia gets a style change. Also fix the video thumb in
 645         the theme to be scaled properly.
 646
 647         Test: media/video-controls-zoomed.html
 648
 649         * rendering/MediaControlElements.cpp:
 650         (WebCore::MediaControlShadowRootElement::updateStyle):
 651         (WebCore::MediaTextDisplayElement::MediaTextDisplayElement):
 652         (WebCore::MediaTextDisplayElement::updateStyle):
 653         (WebCore::MediaControlInputElement::MediaControlInputElement):
 654         (WebCore::MediaControlInputElement::updateStyle):
 655         * rendering/MediaControlElements.h:
 656         * rendering/RenderMedia.cpp:
 657         (WebCore::RenderMedia::styleDidChange):
 658         * rendering/RenderMedia.h:
 659         * rendering/RenderThemeMac.mm:
 660         (WebCore::RenderThemeMac::adjustSliderThumbSize):
 661
 662 2009-03-23  Mike Belshe  <mike@belshe.com>
 663
 664         Reviewed by Darin Adler.
 665
 666         https://bugs.webkit.org/show_bug.cgi?id=24739
 667
 668         Rework StringImpl::create methods to try to allocate a single buffer
 669         rather than allocating both the StringImpl class and a separate data
 670         buffer.
 671
 672         * platform/text/StringImpl.cpp:
 673         * platform/text/StringImpl.h:
 674
 675 2009-03-23  Darin Adler  <darin@apple.com>
 676
 677         Reviewed by Antti Koivisto.
 678
 679         Based on a patch by Nico Weber <nicolasweber@gmx.de>
 680
 681         Bug 24755: LayoutTests/http/tests/misc/url-in-utf16le.html regression
 682         https://bugs.webkit.org/show_bug.cgi?id=24755
 683
 684         * platform/text/TextEncoding.cpp:
 685         (WebCore::TextEncoding::isUTF7Encoding): Added. Checks if the current encoding
 686         is UTF7 without loading extended codecs.
 687         (WebCore::TextEncoding::encodingForFormSubmission): Use isUTF7Encoding() instead
 688         of comparing with UTF7Encoding(). Eliminate the explicit check of
 689         noExtendedTextEncodingNameUsed() because that's now handled by the functions
 690         that this function calls instead.
 691
 692         * platform/text/TextEncoding.h: Added isUTF7Encoding function.
 693         Also tweaked formatting a bit.
 694
 695 2009-03-20  Peter Kasting  <pkasting@google.com>
 696
 697         Reviewed by Darin Fisher.
 698
 699         https://bugs.webkit.org/show_bug.cgi?id=24720
 700         RenderThemeChromium should draw something for Slider parts instead of
 701         dropping them on the floor.
 702
 703         * platform/chromium/ChromiumBridge.h:
 704         * rendering/RenderThemeChromiumWin.cpp:
 705         (WebCore::RenderThemeChromiumWin::adjustSliderThumbSize):
 706         (WebCore::RenderThemeChromiumWin::paintSliderTrack):
 707         (WebCore::RenderThemeChromiumWin::determineSliderThumbState):
 708         (WebCore::RenderThemeChromiumWin::getThemeData):
 709         * rendering/RenderThemeChromiumWin.h:
 710         (WebCore::RenderThemeChromiumWin::paintSliderThumb):
 711
 712 2009-03-23  Glen Murphy  <glen@chromium.org>
 713
 714         Reviewed by Darin Fisher.
 715
 716         https://bugs.webkit.org/show_bug.cgi?id=24657
 717
 718         Fix Skia drawing of highly scaled bitmaps; the conversion to
 719         IntRect produced visible layout test failures in highly scaled
 720         coordinate systems.
 721
 722         Test: svg/custom/image-small-width-height.svg
 723
 724         * WebCore\platform\graphics\skia\ImageSkia.cpp:
 725
 726 2009-03-23  David Kilzer  <ddkilzer@apple.com>
 727
 728         Provide JavaScript exception information after slow script timeout
 729
 730         Reviewed by Oliver Hunt.
 731
 732         * bindings/js/ScriptController.cpp:
 733         (WebCore::ScriptController::evaluate): Changed to report
 734         exceptions for the Interrupted completion type as well.
 735
 736 2009-03-23  Simon Fraser  <simon.fraser@apple.com>
 737
 738         Reviewed by Darin Adler
 739
 740         https://bugs.webkit.org/show_bug.cgi?id=24736
 741
 742         Fix three mostly-unrelated problems with full-page zoom:
 743
 744         * dom/Document.cpp:
 745         (WebCore::Document::elementFromPoint):
 746         Document::elementFromPoint() needs to take full-page zoom into account.
 747
 748         * dom/MouseRelatedEvent.cpp:
 749         (WebCore::MouseRelatedEvent::receivedTarget):
 750         Take full-page zoom into account when computing offsetX/offsetY.
 751
 752         * html/HTMLSelectElement.cpp:
 753         (WebCore::HTMLSelectElement::listBoxDefaultEventHandler):
 754         Don't use offsetX/offsetY when hit testing list boxes; offsets were broken
 755         with full-page zoom, and using pageX/pageY is easier because we don't
 756         have to worry about the event target, and we already have a point in
 757         absolute coordinates.
 758
 759         Tests: fast/forms/listbox-hit-test-zoomed.html
 760                fast/forms/search-zoomed.html
 761                fast/forms/slider-zoomed.html
 762
 763 2009-03-23  Simon Fraser  <simon.fraser@apple.com>
 764
 765         Reviewed by Dan Bernstein
 766
 767         https://bugs.webkit.org/show_bug.cgi?id=24753
 768
 769         The rect for CSS 'clip' needs to have zooming applied to it.
 770
 771         Test: fast/css/clip-zooming.html
 772
 773         * css/CSSStyleSelector.cpp:
 774         (WebCore::CSSStyleSelector::applyProperty):
 775
 776 2009-03-23  Eric Carlson  <eric.carlson@apple.com>
 777
 778         Reviewed by Adam Roben.
 779
 780         <rdar://problem/6704282>
 781         https://bugs.webkit.org/show_bug.cgi?id=24719
 782         QTMovieWinTimer logic inversion
 783
 784         Fix logic inversion in the Win32 timer used by QTMovieWin that caused it to always
 785         use SetTimer, even when the intervals was below USER_TIMER_MINIMUM. A side effect of
 786         this was that a movie timer would sometimes be blocked for significant amounts of time
 787         because WM_TIMER messages are not processed when the thread's message queue has any
 788         higher priority messages, and WebCore/Win's timer uses PostMessage for low interval
 789         timers. Also change SetTimer call to use HWND and custom message instead of
 790         timer function since the timer already has an HWND for processing PostMessage.
 791
 792         Not possible to make a test for this because it is so timing dependant.
 793
 794         * platform/graphics/win/QTMovieWinTimer.cpp:
 795         (TimerWindowWndProc):
 796         (setSharedTimerFireDelay):
 797
 798 2009-03-23  Eric Carlson  <eric.carlson@apple.com>
 799
 800         Reviewed by Adele Peterson.
 801
 802         https://bugs.webkit.org/show_bug.cgi?id=24588
 803
 804         Update media element implementation to current HTML5 spec
 805
 806         New tests:
 807                media/media-constants.html
 808                media/video-seek-no-src-exception.html
 809                media/video-source-add-src.html
 810                media/video-src-invalid-remove.html
 811                media/video-src-plus-source.html
 812                media/video-timeupdate-during-playback.html
 813
 814         * dom/EventNames.h: Remove obsolute events, add new ones.
 815         * html/HTMLMediaElement.cpp:
 816         (WebCore::HTMLMediaElement::HTMLMediaElement): Initialize new member vars.
 817         (WebCore::HTMLMediaElement::attributeChanged): Trigger load() only when we don't
 818         have a source.
 819         (WebCore::HTMLMediaElement::removedFromDocument): Deal with state name changes.
 820         (WebCore::HTMLMediaElement::scheduleProgressEvent): New, create a progress event and
 821         add it to the event queue to be dispatch when the timer fires.
 822         (WebCore::HTMLMediaElement::scheduleEvent): New, create a generic event and add
 823         it to the event queue to be dispatch when the timer fires.
 824         (WebCore::HTMLMediaElement::enqueueEvent): Add an event to the queue and ticke the
 825         asynch event timer.
 826         (WebCore::HTMLMediaElement::asyncEventTimerFired): Dispatch all pending events.
 827         (WebCore::HTMLMediaElement::loadTimerFired): Either trigger the initial load or
 828         try to load the next <source> url.
 829         (WebCore::HTMLMediaElement::load): Minor style change.
 830         (WebCore::HTMLMediaElement::loadInternal): The first part of the spec load algorithm, cleanup
 831         the current load (if any) and set up state for a new load.
 832         (WebCore::HTMLMediaElement::selectMediaResource): Deal with no 'src' or <source>, post 'loadstart'
 833         event, and initiate load from 'src' if present.
 834         (WebCore::HTMLMediaElement::loadNextSourceChild): Initiate load from next <source> url, or trigger
 835         noneSupported() if no more to consider.
 836         (WebCore::HTMLMediaElement::loadResource): Instantiate a new MediaPlayer and ask it to load a url.
 837         (WebCore::HTMLMediaElement::startProgressEventTimer): Start the repeating progress event timer.
 838         (WebCore::HTMLMediaElement::noneSupported): Post error event and set up state when no valid
 839         media url was found.
 840         (WebCore::HTMLMediaElement::mediaEngineError): Post error event and set up state when no valid
 841         media engine failed with a decode error or a network error.
 842         (WebCore::HTMLMediaElement::mediaPlayerNetworkStateChanged):
 843         (WebCore::HTMLMediaElement::setNetworkState): Updated for new spec network states.
 844         (WebCore::HTMLMediaElement::mediaPlayerReadyStateChanged):
 845         (WebCore::HTMLMediaElement::setReadyState): Updated for new spec ready state.
 846         (WebCore::HTMLMediaElement::progressEventTimerFired): Bail if the network is not active.
 847         (WebCore::HTMLMediaElement::seek): Return INVALID_STATE_ERR exception if state is too low or
 848         if player hasn't been set up yet. This is necessary becase load() is async. Clear the flag
 849         we use to guard against sending 'ended' more than once.
 850         (WebCore::HTMLMediaElement::duration): Don't bother calling media engine before it has metadata.
 851         (WebCore::HTMLMediaElement::setDefaultPlaybackRate): Remove exception param, 0 is no longer an
 852         invalid rate.
 853         (WebCore::HTMLMediaElement::setPlaybackRate): Remove exception param, 0 is no longer an
 854         invalid rate. Cache rate being set so we can use it later if media engine isn't ready now.
 855         (WebCore::HTMLMediaElement::play): Remove exception param, play() before load() now just
 856         starts loading asynchronously.
 857         (WebCore::HTMLMediaElement::playInternal): Remove exception param. Fire 'waiting' or 'playing'
 858         event depending on current state.
 859         (WebCore::HTMLMediaElement::pause): Remove exception param, pause() before load() now just
 860         starts loading asynchronously.
 861         (WebCore::HTMLMediaElement::pauseInternal): Remove exception param.
 862         (WebCore::HTMLMediaElement::setVolume): dispatchEventAsync -> scheduleEvent
 863         (WebCore::HTMLMediaElement::setMuted): dispatchEventAsync -> scheduleEvent
 864         (WebCore::HTMLMediaElement::togglePlayState): Remove exception param.
 865         (WebCore::HTMLMediaElement::beginScrubbing): pause() doesn't take an exception param.
 866         (WebCore::HTMLMediaElement::startPlaybackProgressTimer): New, starts timer that fires 4 times per
 867         second when the movie is playing to timeupdate so we can post 'timeupdate' events.
 868         (WebCore::HTMLMediaElement::playbackProgressTimerFired): Timer proc.
 869         (WebCore::HTMLMediaElement::scheduleTimeupdateEvent): Bottleneck around scheduling a 'timeupdate'
 870         event because we both fire them them when the spec says we should and when the media engine
 871         says that time has jumped, but we don't want to fire more than one at a given movie time. We also
 872         use this bottleneck to keep track of the last time one was posted so we won't fire too often
 873         during playback.
 874         (WebCore::HTMLMediaElement::canPlay): readyState now tracks whether or not we have metadata.
 875         (WebCore::HTMLMediaElement::havePotentialSourceChild): New, checks to see if there are a <source>
 876         element with a 'src' attribute that we have not tried to load yet.
 877         (WebCore::HTMLMediaElement::nextSourceChild): New, returns the url and content type of the next
 878         <source> element that we haven't tried to load.
 879         (WebCore::HTMLMediaElement::mediaPlayerTimeChanged): Schedule 'seeked' event when seeking completes.
 880         Set a flag when we post the 'ended' event, clear it when time changed and we aren't at the end since
 881         some media engines call this proc more than once when playback reaches the end and stops, but we
 882         don't want to post 'ended' more than once.
 883         (WebCore::HTMLMediaElement::mediaPlayerDurationChanged): New, added so media engine can inform
 884         when the movie duration changes and we can post 'durationchanged' event.
 885         (WebCore::HTMLMediaElement::mediaPlayerRateChanged): New, added so media engine can inform when
 886         the rate changed and we can updated our cached rate. This is useful because we only want to know
 887         post periodic 'timeupdate' events when the movie is actually playing, and because we want to know
 888         the actual playback rate when it differs from what we tried to set.
 889         (WebCore::HTMLMediaElement::mediaPlayerSizeChanged): New, added so media engine can inform when
 890         a movie's intrinsic size changes and we can inform the renderer.
 891         (WebCore::HTMLMediaElement::potentiallyPlaying): Renamed from activelyPlaying since the spec now
 892         uses "actively playing" for this concept. Update logic for new state names and un-comment calls
 893         to stoppedDueToErrors() and pausedForUserInteraction() since the spec says those condiditons
 894         are part of the answer.
 895         (WebCore::HTMLMediaElement::endedPlayback): Update logic for new state names.
 896         (WebCore::HTMLMediaElement::stoppedDueToErrors): New, spec says this logic should be part of
 897         the determination of "potentially playing".
 898         (WebCore::HTMLMediaElement::pausedForUserInteraction): New, placeholder for when (if) user
 899         agent supports this spec concept.
 900         (WebCore::HTMLMediaElement::updatePlayState): Stop timer used to fire periodic 'timeupdate'
 901         events when we pauses the movie. Set the media engine rate before calling play() in case it
 902         wasn't set up when the rate was changed.
 903         (WebCore::HTMLMediaElement::stopPeriodicTimers): New, stop the progress event and 'timeupate'
 904         event timers.
 905         (WebCore::HTMLMediaElement::userCancelledLoad): New, logic pulled out of documentWillBecomeInactive
 906         and updated for the current spec.
 907         (WebCore::HTMLMediaElement::documentWillBecomeInactive): Moved some logic to userCancelledLoad.
 908         (WebCore::HTMLMediaElement::documentDidBecomeActive): Update comments.
 909         (WebCore::HTMLMediaElement::initialURL): Update for refactoring of code that determines the
 910         initial url.
 911         * html/HTMLMediaElement.h: Change ReadyState and NetworkState enums to match names in the spec,
 912         update for changes in .cpp.
 913         (WebCore::HTMLMediaElement::):
 914
 915         * html/HTMLMediaElement.idl: Update ready state and network state constants for spec changes.
 916         defaultPlaybackRate, playbackRate, play(), and pause() no longer raise exceptions.
 917
 918         * html/HTMLSourceElement.cpp:
 919         (WebCore::HTMLSourceElement::insertedIntoDocument): Update for network state name changes.
 920
 921         * html/HTMLVideoElement.cpp:
 922         (WebCore::HTMLVideoElement::updatePosterImage): Update for ready state name changes.
 923
 924         * html/MediaError.h:
 925         (WebCore::MediaError::): add MEDIA_ERR_NONE_SUPPORTED.
 926
 927         * html/MediaError.idl: add MEDIA_ERR_NONE_SUPPORTED.
 928
 929         * loader/MediaDocument.cpp:
 930         (WebCore::MediaDocument::defaultEventHandler): play() and pause() don't take an exception.
 931
 932         * platform/graphics/MediaPlayer.cpp:
 933         (WebCore::NullMediaPlayerPrivate::readyState): Update for newtork state name changes.
 934         (WebCore::MediaPlayer::sizeChanged): New, so engine can report intrinsic size changes.
 935         (WebCore::MediaPlayer::rateChanged): New, so engine can report rate changes.
 936         (WebCore::MediaPlayer::durationChanged): New, so engine can report duration changes.
 937         * platform/graphics/MediaPlayer.h: Update NetworkState and ReadyState enum names to match spec
 938         states.
 939         (WebCore::MediaPlayerClient::mediaPlayerDurationChanged): New.
 940         (WebCore::MediaPlayerClient::mediaPlayerRateChanged): New.
 941         (WebCore::MediaPlayerClient::mediaPlayerSizeChanged): New.
 942         (WebCore::MediaPlayer::):
 943
 944         * platform/graphics/gtk/MediaPlayerPrivateGStreamer.cpp:
 945         (WebCore::MediaPlayerPrivate::MediaPlayerPrivate): Update for network/ready state name changes.
 946         (WebCore::MediaPlayerPrivate::load): Ditto.
 947         (WebCore::MediaPlayerPrivate::updateStates): Ditto.
 948         (WebCore::MediaPlayerPrivate::loadingFailed): Ditto.
 949
 950         * platform/graphics/mac/MediaPlayerPrivateQTKit.h: Update for network/ready state name changes.
 951         Remove endPointTimer, it is no longer necessary. Add m_enabledTrackCount and m_duration.
 952         (WebCore::MediaPlayerPrivate::metaDataAvailable):
 953         * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
 954         (WebCore::MediaPlayerPrivate::MediaPlayerPrivate): No more m_endPointTimer or m_endTime. Initialize
 955         m_enabledTrackCount and m_duration. Update for network/ready state name changes.
 956         (WebCore::MediaPlayerPrivate::load): Update for network/ready state name changes.
 957         (WebCore::MediaPlayerPrivate::play): No more m_endPointTimer.
 958         (WebCore::MediaPlayerPrivate::pause): Ditto.
 959         (WebCore::MediaPlayerPrivate::currentTime): No more m_endTime.
 960         (WebCore::MediaPlayerPrivate::seek): Ditto.
 961         (WebCore::MediaPlayerPrivate::doSeek): Ditto, plus don't call setRate(0) when the rate is
 962         already zero.
 963         (WebCore::MediaPlayerPrivate::setEndTime): No more m_endTime.
 964         (WebCore::MediaPlayerPrivate::updateStates): Update for network/ready state name changes. Return
 965         different errors depending on what causes a failure. Watch for and report duration changes.
 966         (WebCore::MediaPlayerPrivate::rateChanged): Report rate changes.
 967         (WebCore::MediaPlayerPrivate::sizeChanged): Report size changes.
 968         (WebCore::MediaPlayerPrivate::didEnd): No more endpoint timer.
 969         (WebCore::MediaPlayerPrivate::setVisible): Update for network/ready state name changes.
 970         (WebCore::MediaPlayerPrivate::disableUnsupportedTracks): Don't return number of unsupported
 971         tracks, store in m_enabledTrackCount so we can use it to help determine causes of failure.
 972
 973         * platform/graphics/qt/MediaPlayerPrivatePhonon.cpp:
 974         (WebCore::MediaPlayerPrivate::MediaPlayerPrivate): Update for network/ready state name changes.
 975         (WebCore::MediaPlayerPrivate::load): Ditto.
 976         (WebCore::MediaPlayerPrivate::duration): Ditto.
 977         (WebCore::MediaPlayerPrivate::updateStates): Ditto.
 978         (WebCore::MediaPlayerPrivate::naturalSize): Ditto.
 979
 980         * platform/graphics/win/MediaPlayerPrivateQuickTimeWin.cpp:
 981         (WebCore::MediaPlayerPrivate::MediaPlayerPrivate): No more m_endPointTimer. Update for
 982         network/ready state name changes.
 983         (WebCore::MediaPlayerPrivate::load): Update for network/ready state name changes. No more
 984         m_endPointTimer.
 985         (WebCore::MediaPlayerPrivate::play): No more m_endPointTimer.
 986         (WebCore::MediaPlayerPrivate::pause): Ditto.
 987         (WebCore::MediaPlayerPrivate::setEndTime): Ditto.
 988         (WebCore::MediaPlayerPrivate::updateStates): Update for network/ready state name changes.
 989         (WebCore::MediaPlayerPrivate::didEnd): No more m_endPointTimer.
 990         * platform/graphics/win/MediaPlayerPrivateQuickTimeWin.h:
 991
 992         * rendering/MediaControlElements.cpp:
 993         (WebCore::MediaControlPlayButtonElement::defaultEventHandler): Update for network/ready state
 994         name changes.
 995         (WebCore::MediaControlSeekButtonElement::defaultEventHandler): Ditto.
 996
 997 2009-03-22  Kevin Ollivier  <kevino@theolliviers.com>
 998
 999         wxGTK build fix. Add missing header.
1000
1001         * platform/wx/wxcode/gtk/non-kerned-drawing.cpp:
1002
1003 2009-03-22  Simon Fraser  <simon.fraser@apple.com>
1004
1005         Reviewed by Dan Bernstein
1006
1007         https://bugs.webkit.org/show_bug.cgi?id=24665
1008
1009         Image-map code in RenderImage could result in RenderImage::nodeAtPoint()
1010         setting HitTestResult::innerNode(), but returning false, which violates
1011         hit testing rules. Use a temporary HitTestResult so that we only fill in
1012         result when we know we've hit.
1013
1014         * rendering/RenderImage.cpp:
1015         (WebCore::RenderImage::nodeAtPoint):
1016
1017 2009-03-22  Simon Fraser  <simon.fraser@apple.com>
1018
1019         Reviewed by Dan Bernstein
1020
1021         https://bugs.webkit.org/show_bug.cgi?id=24743
1022
1023         Fix hit testing regression from r41840. We need to pass the temporary
1024         HitTestResult when testing sublayers, then only copy to 'result' when
1025         the layer is known to have been hit.
1026
1027         Test: fast/layers/zindex-hit-test.html
1028
1029         * rendering/RenderLayer.cpp:
1030         (WebCore::RenderLayer::hitTestLayer):
1031
1032 2009-03-20  Simon Fraser  <simon.fraser@apple.com>
1033
1034         Reviewed by Darin Adler
1035
1036         https://bugs.webkit.org/show_bug.cgi?id=24733
1037
1038         Fix hit testing on video controls after full page zoom by fixing wider issue
1039         with event->pageX(), pageY() with zooming. pageX and pageY are "fixed" to be
1040         invariant under zooming (for JavaScript), so we keep an actual page point around
1041         in MouseEvent::absoluteLocation() to avoid the need to factor in zooming everywhere.
1042
1043         * dom/MouseRelatedEvent.cpp:
1044         (WebCore::MouseRelatedEvent::initCoordinates):
1045         (WebCore::MouseRelatedEvent::computePageLocation):
1046         * dom/MouseRelatedEvent.h:
1047         (WebCore::MouseRelatedEvent::absoluteLocation):
1048         (WebCore::MouseRelatedEvent::setAbsoluteLocation):
1049         Member var, and getter and setter for absoluteLocation.
1050         New method, computePageLocation(), to compute the actual page point,
1051         and call it when creating and initting mouse-related events.
1052
1053         * dom/Node.cpp:
1054         (WebCore::Node::dispatchMouseEvent):
1055         (WebCore::Node::dispatchWheelEvent):
1056         Keep non-adjusted pageX and pageY around, and call setAbsoluteLocation()
1057         on the event to replace a potentially rounded point.
1058
1059         * html/HTMLInputElement.cpp:
1060         (WebCore::HTMLInputElement::defaultEventHandler):
1061         Clean up slider handling code.
1062
1063         * html/HTMLSelectElement.cpp:
1064         (WebCore::HTMLSelectElement::listBoxDefaultEventHandler):
1065         Add FIXME comment for use of offsetX/offsetY.
1066
1067         * page/ContextMenuController.cpp:
1068         (WebCore::ContextMenuController::handleContextMenuEvent):
1069         Use absoluteLocation() when hit testing for context menus.
1070
1071         * rendering/RenderFrameSet.cpp:
1072         (WebCore::RenderFrameSet::userResize):
1073         Use absoluteLocation() when resizing frames.
1074
1075         * rendering/RenderMedia.cpp:
1076         (WebCore::RenderMedia::forwardEvent):
1077         Use absoluteLocation() when hit testing media controls.
1078
1079         * rendering/RenderSlider.cpp:
1080         (WebCore::HTMLSliderThumbElement::defaultEventHandler):
1081         (WebCore::RenderSlider::mouseEventIsInThumb):
1082         Use absoluteLocation() when handling slider events.
1083
1084         (WebCore::RenderSlider::forwardEvent):
1085         Factor some code out of HTMLInputElement::defaultEventHandler().
1086
1087         * rendering/RenderTextControlSingleLine.cpp:
1088         (WebCore::RenderTextControlSingleLine::forwardEvent):
1089         Use absoluteLocation() when hit testing search field buttons, which fixees
1090         bugs in the search field with zooming.
1091
1092 2009-03-21  David Levin  <levin@chromium.org>
1093
1094         Reviewed by Dimitri Glazkov.
1095
1096         https://bugs.webkit.org/show_bug.cgi?id=24727
1097         Add V8XMLHttpRequest*.
1098
1099         * bindings/v8/V8XMLHttpRequestUtilities.cpp: Added.
1100         * bindings/v8/V8XMLHttpRequestUtilities.h: Added.
1101         * bindings/v8/custom/V8XMLHttpRequestConstructor.cpp: Added.
1102         * bindings/v8/custom/V8XMLHttpRequestCustom.cpp: Added.
1103         * bindings/v8/custom/V8XMLHttpRequestUploadCustom.cpp: Added.
1104
1105 2009-03-21  David Levin  <levin@chromium.org>
1106
1107         Reviewed by Dimitri Glazkov.
1108
1109         https://bugs.webkit.org/show_bug.cgi?id=24725
1110         Add V8NodeFilter.
1111
1112         * bindings/v8/V8NodeFilter.cpp: Added.
1113         * bindings/v8/V8NodeFilter.h: Added.
1114
1115 2009-03-21  Dan Bernstein  <mitz@apple.com>
1116
1117         Rubber-stamped by Kevin Decker
1118
1119         - remove some redundant #include statements
1120
1121         * bindings/js/JSDOMWindowBase.cpp:
1122
1123 2009-03-20  Dan Bernstein  <mitz@apple.com>
1124
1125         Reviewed by Mark Rowe.
1126
1127         - fix <rdar://problem/6574185> REGRESSION (3.2.2-TOT): hang in text drawing code
1128
1129         * platform/graphics/win/FontCGWin.cpp:
1130         (WebCore::drawGDIGlyphs): Changed glyph stroking to fill and stroke each
1131         glyph as a separate path, instead of all glyphs as a single path. This
1132         matches what CGContextShowGlyphsWithAdvances() does, and has comparable
1133         performance.
1134
1135 2009-03-20  Dean Jackson  <dino@apple.com>
1136
1137         Reviewed by Simon Fraser
1138
1139         Build fix for ENABLE(3D_RENDERING)
1140
1141         * rendering/RenderObject.h:
1142         (WebCore::makeMatrixRenderable):
1143
1144 2009-03-20  Dave Moore  <davemoore@google.com>
1145
1146         Reviewed by Dimitri Glazkov.
1147
1148         https://bugs.webkit.org/show_bug.cgi?id=24705
1149
1150         A bug in the V8 bindings is preventing chromium from setting the href on the
1151         location object...any attempt throws a security error, not just for javascript
1152         protocol
1153
1154         * page/Location.idl:
1155
1156 2009-03-20  Craig Schlenter  <craig.schlenter@gmail.com>
1157
1158         Reviewed by Dimitri Glazkov.
1159
1160         https://bugs.webkit.org/show_bug.cgi?id=24608
1161
1162         Include stdio.h needed for printf on gcc 4.4.0
1163
1164         * platform/KURLGoogle.cpp:
1165
1166 2009-03-20  Mike Belshe <mike@belshe.com>
1167
1168         Reviewed by Darin Fisher.
1169
1170         https://bugs.webkit.org/show_bug.cgi?id=24577
1171
1172         Don't let comments at the end of an event handler
1173         break the event handler.
1174
1175         No change in behavior, so no test.
1176
1177         * bindings/v8/V8LazyEventListener.cpp:
1178         (WebCore::V8LazyEventListener::getWrappedListenerFunction):
1179
1180 2009-03-20  Norbert Leser  <norbert.leser@nokia.com>
1181
1182         Reviewed by Darin Adler.
1183
1184         https://bugs.webkit.org/show_bug.cgi?id=24535
1185
1186         Fixes missing line terminator character (;) after macro call.
1187         It is common practice to add the trailing ";" where macros are substituted
1188         and not where they are defined with #define.
1189         This change is consistent with other macro declarations across webkit,
1190         and it also solves compilation failure with symbian compilers.
1191
1192         No change in behavior, so no test.
1193
1194         * bindings/js/JSDOMWindowShell.cpp:
1195         * bindings/js/JSEventListener.cpp:
1196         * bindings/js/JSImageConstructor.cpp:
1197         * bindings/js/JSInspectedObjectWrapper.cpp:
1198         * bindings/js/JSInspectorCallbackWrapper.cpp:
1199         * bindings/js/JSNamedNodesCollection.cpp:
1200         * bindings/js/JSNodeFilterCondition.cpp:
1201         * bindings/js/JSOptionConstructor.cpp:
1202         * bindings/js/JSQuarantinedObjectWrapper.cpp:
1203         * bindings/js/JSRGBColor.cpp:
1204         * bindings/js/JSWorkerContextBase.cpp:
1205         * bindings/js/JSXMLHttpRequestConstructor.cpp:
1206         * bindings/js/JSXSLTProcessorConstructor.cpp:
1207         * bindings/scripts/CodeGeneratorJS.pm:
1208
1209 2009-03-20  Mike Belshe  <mike@belshe.com>
1210
1211         Reviewed by Dimitri Glazkov and Dave Hyatt.
1212
1213         https://bugs.webkit.org/show_bug.cgi?id=24324
1214
1215         Make the minimum timer configurable for different platforms.
1216
1217         * page/DOMTimer.cpp:
1218         (WebCore::DOMTimer::DOMTimer):
1219         (WebCore::DOMTimer::fired):
1220         * page/DOMTimer.h:
1221         (WebCore::DOMTimer::minTimerInterval):
1222         (WebCore::DOMTimer::setMinTimerInterval):
1223
1224 2009-03-20  Dean McNamee  <deanm@chromium.org>
1225
1226         Reviewed by Darin Adler.
1227
1228         https://bugs.webkit.org/show_bug.cgi?id=22834
1229
1230         Make sure to consistently match new/delete and fastMalloc/fastFree.
1231
1232         * css/CSSSelectorList.cpp:
1233         (WebCore::CSSSelectorList::adoptSelectorVector):
1234         (WebCore::CSSSelectorList::deleteSelectors):
1235
1236 2009-03-20  Dan Bernstein  <mitz@apple.com>
1237
1238         Reviewed by Dave Hyatt.
1239
1240         - fix https://bugs.webkit.org/show_bug.cgi?id=23739
1241           <rdar://problem/6556371> REGRESSION (r36513): iframe isn't sized properly upon load
1242
1243         * rendering/RenderBlock.cpp:
1244         (WebCore::RenderBlock::percentHeightDescendants): Added this accessor.
1245         * rendering/RenderBlock.h:
1246         * rendering/RenderTableSection.cpp:
1247         (WebCore::RenderTableSection::layoutRows): Extended the check for
1248         children that flex to include other descendants with percent height
1249         which is relative to the cell.
1250
1251 2009-03-20  Dmitry Titov  <dimich@chromium.org>
1252
1253         Reviewed by Alexey Proskuryakov.
1254
1255         https://bugs.webkit.org/show_bug.cgi?id=24706
1256         Remove ScriptExecutionContext::encoding() since Workers do not need it.
1257         WorkerContext::encoding() is simply removed, while Document::encoding()
1258         made non-virtual and private. Workers use UTF-8 now except when instructed
1259         otherwise by http header. Also updated test.
1260
1261         * dom/Document.h: Made encoding() non-virtual and private.
1262         * dom/ScriptExecutionContext.h: removed encoding().
1263
1264         * workers/Worker.cpp:
1265         (WebCore::Worker::Worker):
1266         (WebCore::Worker::notifyFinished):
1267         * workers/WorkerContext.cpp: removed encoding() implementation.
1268         (WebCore::WorkerContext::WorkerContext):
1269         (WebCore::WorkerContext::completeURL):
1270         * workers/WorkerContext.h: removed encoding()
1271         (WebCore::WorkerContext::create):
1272         * workers/WorkerContextProxy.h:
1273         * workers/WorkerImportScriptsClient.cpp:
1274         (WebCore::WorkerImportScriptsClient::didReceiveData):
1275         * workers/WorkerMessagingProxy.cpp:
1276         (WebCore::WorkerMessagingProxy::startWorkerContext):
1277         * workers/WorkerMessagingProxy.h:
1278         * workers/WorkerThread.cpp:
1279         (WebCore::WorkerThreadStartupData::create):
1280         (WebCore::WorkerThreadStartupData::WorkerThreadStartupData):
1281         (WebCore::WorkerThread::create):
1282         (WebCore::WorkerThread::WorkerThread):
1283         (WebCore::WorkerThread::workerThread):
1284         * workers/WorkerThread.h:
1285         In all these, removed storing encoding and 'inheriting' it from the parent.
1286         Instead, they are all using UTF-8 now.
1287
1288 2009-03-20  Timothy Hatcher  <timothy@apple.com>
1289
1290         Change how threading exceptions are checked so they are reported
1291         by what round they were added. That way WebKit can decided the
1292         behavior per-round based on linked-on-or-after checks.
1293
1294         <rdar://problem/6626741&6648478&6635474&6674079>
1295
1296         Reviewed by Darin Adler.
1297
1298         * WebCore.base.exp: Export the new symbols.
1299         * bindings/objc/DOMAbstractView.mm: Use the new WebCoreThreadViolationCheckRoundOne macro.
1300         * bindings/scripts/CodeGeneratorObjC.pm: Ditto.
1301         * platform/ThreadCheck.h:
1302         * platform/mac/ThreadCheck.mm:
1303         (WebCore::readThreadViolationBehaviorFromUserDefaults): Refactor how the default is read.
1304         (WebCore::setDefaultThreadViolationBehavior): Take a round argument.
1305         (WebCore::reportThreadViolation): Ditto.
1306         (WebCoreReportThreadViolation): Ditto.
1307
1308 2009-03-20  Geoffrey Garen  <ggaren@apple.com>
1309
1310         Reviewed by Sam Weinig.
1311
1312         Fixed up an out-of-date comment.
1313
1314         * bindings/js/JSDOMWindowCustom.h:
1315         (WebCore::JSDOMWindow::customPut):
1316
1317 2009-03-20  Beth Dakin  <bdakin@apple.com>
1318
1319         Reviewed by Dave Hyatt.
1320
1321         Fix for https://bugs.webkit.org/show_bug.cgi?id=20909 REGRESSION
1322         (r35318): A press release at pfizer.com does not display correctly
1323         - and corresponding -
1324         <rdar://problem/6680073>
1325
1326         * rendering/RenderObject.cpp:
1327         (WebCore::RenderObject::invalidateContainerPrefWidths):
1328         * rendering/RenderObject.h:
1329         (WebCore::RenderObject::markContainingBlocksForLayout):
1330
1331 2009-03-20  Eric Seidel  <eric@webkit.org>
1332
1333         Reviewed by Justin Garcia.
1334
1335         maxDeepOffset is confusing and should be removed
1336         https://bugs.webkit.org/show_bug.cgi?id=24586
1337
1338         Abstract some hard-to-read (but shared) logic into a new renderedAsNonInlineTableOrHR function.
1339         Add first/lastDeepEditingPositionForNode Position creation functions
1340         and deploy them to places we used to call maxDeepOffset.
1341
1342         Rename Position::atStart and atEnd to atStartOfTree atEndOfTree
1343         Add a new Position::atFirst/atLastEditingPositionForNode() and use these
1344         to replace a few more callers for maxDeepOffset()
1345
1346         Rename maxDeepOffset to lastEditingOffsetForNode (so that we mere mortals have some clue what it does)
1347
1348         "Editing positions" are confusing because they have one
1349         of two behaviors, depending on if the container node is ignored
1350         by editing (if editingIgnoresContent(node) returns true) or not.
1351         Positions referring to nodes ignored by editing are
1352         neighbor-relative (they are before or after the node) where as
1353         positions reffering to other nodes are container-relative
1354         (they are between two child nodes of the container, identified
1355         by the offset() member).  I will be fixing this confusion in
1356         future patches.  These renames hopefully make the current behavior clearer.
1357
1358         * dom/Position.cpp:
1359         (WebCore::Position::previous):
1360         (WebCore::Position::next):
1361         (WebCore::Position::atFirstEditingPositionForNode):
1362         (WebCore::Position::atLastEditingPositionForNode):
1363         (WebCore::Position::atStartOfTree):
1364         (WebCore::Position::atEndOfTree):
1365         (WebCore::Position::previousCharacterPosition):
1366         (WebCore::Position::nextCharacterPosition):
1367         (WebCore::Position::upstream):
1368         (WebCore::Position::isCandidate):
1369         (WebCore::firstDeepEditingPositionForNode):
1370         (WebCore::lastDeepEditingPositionForNode):
1371         * dom/Position.h:
1372         * dom/PositionIterator.cpp:
1373         (WebCore::PositionIterator::operator Position):
1374         (WebCore::PositionIterator::increment):
1375         (WebCore::PositionIterator::decrement):
1376         (WebCore::PositionIterator::atEnd):
1377         (WebCore::PositionIterator::atEndOfNode):
1378         * editing/CompositeEditCommand.cpp:
1379         (WebCore::CompositeEditCommand::positionAvoidingSpecialElementBoundary):
1380         * editing/DeleteSelectionCommand.cpp:
1381         (WebCore::isTableCellEmpty):
1382         (WebCore::DeleteSelectionCommand::removeNode):
1383         (WebCore::DeleteSelectionCommand::handleGeneralDelete):
1384         * editing/Editor.cpp:
1385         (WebCore::Editor::advanceToNextMisspelling):
1386         * editing/InsertLineBreakCommand.cpp:
1387         (WebCore::InsertLineBreakCommand::doApply):
1388         * editing/InsertListCommand.cpp:
1389         (WebCore::InsertListCommand::doApply):
1390         * editing/ReplaceSelectionCommand.cpp:
1391         (WebCore::ReplaceSelectionCommand::positionAtEndOfInsertedContent):
1392         * editing/TypingCommand.cpp:
1393         (WebCore::TypingCommand::forwardDeleteKeyPressed):
1394         * editing/VisiblePosition.cpp:
1395         (WebCore::VisiblePosition::previous):
1396         (WebCore::VisiblePosition::leftVisuallyDistinctCandidate):
1397         (WebCore::VisiblePosition::left):
1398         (WebCore::VisiblePosition::rightVisuallyDistinctCandidate):
1399         (WebCore::VisiblePosition::right):
1400         * editing/VisibleSelection.cpp:
1401         (WebCore::VisibleSelection::selectionFromContentsOfNode):
1402         (WebCore::VisibleSelection::adjustSelectionToAvoidCrossingEditingBoundaries):
1403         * editing/htmlediting.cpp:
1404         (WebCore::nextVisuallyDistinctCandidate):
1405         (WebCore::previousVisuallyDistinctCandidate):
1406         (WebCore::firstEditablePositionAfterPositionInRoot):
1407         (WebCore::lastEditablePositionBeforePositionInRoot):
1408         (WebCore::lastOffsetForEditing):
1409         (WebCore::isFirstPositionAfterTable):
1410         (WebCore::isLastPositionBeforeTable):
1411         (WebCore::positionBeforeNode):
1412         (WebCore::positionAfterNode):
1413         (WebCore::enclosingEmptyListItem):
1414         (WebCore::caretMaxOffset):
1415         * editing/htmlediting.h:
1416         * editing/visible_units.cpp:
1417         (WebCore::renderedAsNonInlineTableOrHR):
1418         (WebCore::startOfParagraph):
1419         (WebCore::endOfParagraph):
1420         (WebCore::startOfEditableContent):
1421         (WebCore::endOfEditableContent):
1422         * page/AccessibilityObject.cpp:
1423         (WebCore::endOfStyleRange):
1424         * page/AccessibilityRenderObject.cpp:
1425         (WebCore::AccessibilityRenderObject::visiblePositionRange):
1426         * rendering/RenderBox.cpp:
1427         (WebCore::RenderBox::positionForPoint):
1428
1429 2009-03-20  Dmitry Titov  <dimich@chromium.org>
1430
1431         Reviewed by Dimitri Glazkov.
1432
1433         https://bugs.webkit.org/show_bug.cgi?id=24689
1434         Add (upstream) V8 bindings for Workers. Mostly style cleaning.
1435
1436         * bindings/v8/WorkerContextExecutionProxy.cpp: Added.
1437         * bindings/v8/WorkerContextExecutionProxy.h: Added.
1438         * bindings/v8/WorkerScriptController.cpp: Added.
1439         * bindings/v8/WorkerScriptController.h: Added.
1440         * bindings/v8/V8Index.h: Added. This is just a wrapper for v8_index.h, like V8Proxy.h
1441         * bindings/v8/V8Proxy.h: Added domObjectMap() function that wraps GetDOMObjectMap().
1442         * bindings/v8/V8WorkerContextEventListener.cpp:
1443         (WebCore::V8WorkerContextEventListener::callListenerFunction): TrackEvent() renamed trackEvent()
1444
1445 2009-03-20  Stephen White  <senorblanco@chromium.org>
1446
1447         Reviewed by Eric Seidel.
1448
1449         Fix for LayoutTests/fast/canvas/canvas-text-alignment.html
1450         on chromium/skia.  The problem was that the gradient matrix
1451         for text was being applied twice.  Fixed by reverting some of
1452         https://bugs.webkit.org/show_bug.cgi?id=23957, so that skiaDrawText
1453         is no longer responsible for measuring the text and scaling up
1454         the gradient matrix.  Instead, the text bounding box is passed
1455         in from SVGPaintServerGradient.  I didn't make this change for CG,
1456         since it uses a different method (the gradient is drawn using the
1457         text as a pre-rendered mask).
1458         https://bugs.webkit.org/show_bug.cgi?id=24687
1459
1460         * platform/graphics/skia/SkiaFontWin.cpp:
1461         (WebCore::skiaDrawText):
1462         * svg/graphics/SVGPaintServerGradient.cpp:
1463         (WebCore::SVGPaintServerGradient::setup):
1464
1465 2009-03-20  Xan Lopez  <xlopez@igalia.com>
1466
1467         Rubber-stamped by Holger Freyther.
1468
1469         There seems to be some rounding error in cairo (or in how we use
1470         cairo) with some fonts, like DejaVu Sans Mono, which makes cairo
1471         report a height smaller than ascent + descent, which is wrong and
1472         confuses WebCore's layout system. Workaround this while we figure
1473         out what's going on.
1474
1475         * platform/graphics/gtk/SimpleFontDataGtk.cpp:
1476         (WebCore::SimpleFontData::platformInit):
1477         * platform/graphics/gtk/SimpleFontDataPango.cpp:
1478         (WebCore::SimpleFontData::platformInit):
1479
1480 2009-03-19  Alexey Proskuryakov  <ap@webkit.org>
1481
1482         Reviewed by Darin Adler.
1483
1484         https://bugs.webkit.org/show_bug.cgi?id=24122
1485         <rdar://problem/6674179>
1486         REGRESSION: DOM Range extractContents/deleteContents failures seen on Moxiecode tests
1487
1488         Test: fast/dom/Range/deleted-range-endpoints.html
1489
1490         * dom/Range.cpp: (WebCore::Range::processContents): Set the final range in accordance
1491         to the specification.
1492
1493 2009-03-19  Anders Carlsson  <andersca@apple.com>
1494
1495         Reviewed by Dan Bernstein.
1496
1497         <rdar://problem/6682554> Flash content not being rendered (Shockwave Flash 10.0 r22)
1498
1499         If calling updateWidget for some reason resulted in another widget being added to m_widgetUpdateSet, then
1500         that object would never be updated.
1501
1502         * page/FrameView.cpp:
1503         (WebCore::FrameView::updateWidgets):
1504         Factor the widget updating code out into this method. Return true if the update set is empty.
1505
1506         (WebCore::FrameView::performPostLayoutTasks):
1507         Loop over the update set multiple times until all widgets have been updated or until we reach the cap.
1508
1509         * page/FrameView.h:
1510
1511 2009-03-18  Timothy Hatcher  <timothy@apple.com>
1512
1513         Make the defered data loading timer honor the Page's scheduled runloop pairs.
1514         Introduces a new RunLoopTimer class that has an API mimicking Timer but
1515         allows it to be scheduled with one or more SchedulePairs.
1516
1517         <rdar://problem/6687342> -[WebView scheduleInRunLoop:forMode:] has no affect on timers
1518
1519         Reviewed by Darin Adler.
1520
1521         * WebCore.xcodeproj/project.pbxproj: Adds the new RunLoopTimer.{cpp,h} files.
1522         * loader/MainResourceLoader.cpp:
1523         (WebCore::MainResourceLoader::handleDataLoadNow): Use the MainResourceLoaderTimer typedef.
1524         (WebCore::MainResourceLoader::startDataLoadTimer): Added. Start the timer and on
1525         Mac platforms also schedule with the Page's SchedulePairs.
1526         (WebCore::MainResourceLoader::handleDataLoadSoon): Call startDataLoadTimer().
1527         (WebCore::MainResourceLoader::setDefersLoading): Ditto.
1528         * loader/MainResourceLoader.h:
1529         * platform/cf/RunLoopTimerCF.cpp: Added.
1530         (WebCore::RunLoopTimerBase::~RunLoopTimerBase):
1531         (WebCore::timerFired):
1532         (WebCore::RunLoopTimerBase::start):
1533         (WebCore::RunLoopTimerBase::schedule):
1534         (WebCore::RunLoopTimerBase::stop):
1535         (WebCore::RunLoopTimerBase::isActive):
1536         * platform/RunLoopTimer.h: Added.
1537         (WebCore::RunLoopTimerBase::RunLoopTimerBase):
1538         (WebCore::RunLoopTimerBase::startRepeating):
1539         (WebCore::RunLoopTimerBase::startOneShot):
1540         (WebCore::RunLoopTimer::RunLoopTimer):
1541         (WebCore::RunLoopTimer::fired):
1542
1543 2009-03-19  Dimitri Glazkov  <dglazkov@chromium.org>
1544
1545         Reviewed by Darin Fisher.
1546
1547         https://bugs.webkit.org/show_bug.cgi?id=24702
1548         Upstream miscellaneous bindings changes.
1549
1550         * bindings/v8/ScheduledAction.cpp:
1551         (WebCore::ScheduledAction::execute): Changed to call lower-case evaluate.
1552         * bindings/v8/ScriptCallStack.h: Added an extra include.
1553
1554 2009-03-19  Simon Fraser  <simon.fraser@apple.com>
1555
1556         Reviewed by Dave Hyatt
1557
1558         https://bugs.webkit.org/show_bug.cgi?id=24686
1559
1560         When hit testing a RenderLayer whose parent lives in a preserves-3D hierarchy,
1561         we need to compare the computed z-offset with the depth-test z-offset before
1562         deciding that such a RenderLayer was hit. This fixes an issue, tested by the
1563         3d-point-mapping-overlapping.html test, where the child of a transformed element
1564         is found by hit testing, even when some other element with greater Z overlaps
1565         them both.
1566
1567         Improved the code by adding a utility method, isHitCandidate(), which computes and tests
1568         z-depth when necessary.
1569
1570         Tests: transforms/3d/point-mapping/3d-point-mapping-coplanar.html
1571                transforms/3d/point-mapping/3d-point-mapping-overlapping.html
1572
1573         * rendering/RenderLayer.cpp:
1574         (WebCore::isHitCandidate):
1575         (WebCore::RenderLayer::hitTestLayer):
1576
1577 2009-03-19  Jeremy Moskovich  <jeremy@chromium.org>
1578
1579         Reviewed by Dimitri Glazkov.
1580
1581         https://bugs.webkit.org/show_bug.cgi?id=24456
1582         Split ColorChromium.cpp into Mac & Windows variants.
1583         Remove Chromium Dependency on platform/graphics/mac/ColorMac.mm since we
1584         ultimately need to take a different approach.  For now, createCGColor()
1585         is copied from ColorMac.mm.
1586
1587         No observable change in behavior, so no test.
1588
1589         * platform/graphics/chromium/ColorChromium.cpp:
1590         (WebCore::focusRingColor):
1591         * platform/graphics/chromium/ColorChromiumMac.mm: Added.
1592
1593 2009-03-19  Pavel Feldman  <pfeldman@chromium.org>
1594
1595         Reviewed by Dimitri Glazkov.
1596
1597         https://bugs.webkit.org/show_bug.cgi?id=24675
1598         Unforking frontend: add custom InspectorController methods
1599         implementation.
1600
1601         * bindings/v8/custom/V8InspectorControllerCustom.cpp: Added.
1602
1603 2009-03-19  Jay Campan  <jcampan@google.com>
1604
1605         Reviewed by Dimitri Glazkov.
1606
1607         https://bugs.webkit.org/show_bug.cgi?id=24625
1608         Adding an accessor to the currently selected index in the PopupMenuChromium.
1609         This is required for implementing the deletion of an autocomplete entry in Chromium.
1610
1611         * platform/chromium/PopupMenuChromium.cpp:
1612         (WebCore::PopupContainer::selectedIndex):
1613         * platform/chromium/PopupMenuChromium.h:
1614
1615 2009-03-19  Evan Stade  <estade@chromium.org>
1616
1617         Reviewed by Dimitri Glazkov.
1618
1619         https://bugs.webkit.org/show_bug.cgi?id=24526
1620         Improve windows skia text stroking.
1621
1622         Test: LayoutTests/svg/custom/struct-use-09-b.svg
1623
1624         * platform/graphics/skia/SkiaFontWin.cpp: Close the path representing
1625         each font glyph polygon, rather than only closing the path once per
1626         letter. This fixes stroking for letters with multiple polygons, such
1627         as 'A' or 'D'.
1628         (WebCore::getPathForGlyph):
1629         (WebCore::skiaDrawText):
1630
1631 2009-03-19  Simon Hausmann  <simon.hausmann@nokia.com>
1632
1633         Reviewed by Tor Arne Vestbø.
1634
1635         Fixed support for doing calls from JavaScript into NPAPI Plugins for the Qt port on Windows.
1636
1637         Removed dead code for distinguishing between Widget and PluginView in the Qt port.
1638
1639         * bindings/js/ScriptControllerQt.cpp:
1640         (WebCore::ScriptController::createScriptInstanceForWidget): Removed incorrect isNPAPI check.
1641         * plugins/PluginView.cpp:
1642         (WebCore::PluginView::PluginView): Removed m_isNPAPIPlugin variable.
1643         * plugins/PluginView.h: Removed setter/getter.
1644         * plugins/mac/PluginViewMac.cpp:
1645         (WebCore::PluginView::init): Removed call to setIsNPAPIPlugin.
1646         * plugins/qt/PluginViewQt.cpp:
1647         (WebCore::PluginView::init): Ditto.
1648
1649 2009-03-19  Cameron Zwarich  <cwzwarich@uwaterloo.ca>
1650
1651         Reviewed by Oliver Hunt.
1652
1653         Bug 24596: ASSERT in JSC::PropertySlot::slotBase @ iGoogle homepage
1654         <https://bugs.webkit.org/show_bug.cgi?id=24596>
1655         <rdar://problem/6686493>
1656
1657         JSDOMWindow::customGetOwnPropertySlot() does an access check after calling
1658         JSGlobalObject::getOwnPropertySlot(). This causes the PropertySlot to be
1659         set twice, once to the value that is illegal to access, and then to undefined
1660         This causes an assertion failure in property access caching code.
1661
1662         The fix is to do the access check before calling JSGlobalObject::getOwnPropertySlot().
1663
1664         * bindings/js/JSDOMWindowCustom.h:
1665         (WebCore::JSDOMWindow::customGetOwnPropertySlot):
1666
1667 2009-03-18  Alexey Proskuryakov  <ap@webkit.org>
1668
1669         Reviewed by Sam Weinig.
1670
1671         https://bugs.webkit.org/show_bug.cgi?id=24676
1672         Simple cross-origin requests shouldn't dispatch upload progress events
1673
1674         Test: http/tests/xmlhttprequest/simple-cross-origin-progress-events.html
1675
1676         * loader/CrossOriginAccessControl.cpp: (WebCore::passesAccessControlCheck): Added a comment
1677         explaining the somewhat unexpected behavior of this function.
1678
1679         * xml/XMLHttpRequestUpload.cpp: (WebCore::XMLHttpRequestUpload::hasListeners):
1680         * xml/XMLHttpRequestUpload.h:
1681         Report whether there are any event listeners registered.
1682
1683         * xml/XMLHttpRequest.h: Added m_uploadEventsAllowed.
1684
1685         * xml/XMLHttpRequest.cpp:
1686         (WebCore::XMLHttpRequest::createRequest): Set m_uploadEventsAllowed flag.
1687         (WebCore::XMLHttpRequest::makeSameOriginRequest): Ditto.
1688         (WebCore::XMLHttpRequest::makeCrossOriginAccessRequest): Ditto.
1689         (WebCore::XMLHttpRequest::makeSimpleCrossOriginAccessRequest): Set request body - it can be
1690         non-empty for POST requests.
1691         (WebCore::XMLHttpRequest::makeCrossOriginAccessRequestWithPreflight): Set m_uploadEventsAllowed flag.
1692         (WebCore::XMLHttpRequest::handleAsynchronousPreflightResult): Ditto.
1693         (WebCore::XMLHttpRequest::abort): Only dispatch upload progress events if allowed.
1694         (WebCore::XMLHttpRequest::networkError): Ditto.
1695         (WebCore::XMLHttpRequest::abortError): Ditto.
1696         (WebCore::XMLHttpRequest::didSendData): Ditto.
1697
1698 2009-03-18  Marc-Antoine Ruel  <maruel@chromium.org>
1699
1700         Reviewed by Darin Fisher.
1701
1702         https://bugs.webkit.org/show_bug.cgi?id=24398
1703         Fix a crash when loading a svg file in Chromium's test_shell and
1704         then reloading the page.
1705
1706         * history/BackForwardListChromium.cpp:
1707         (WebCore::BackForwardList::BackForwardList):
1708         (WebCore::BackForwardList::close):
1709
1710 2009-03-18  Eric Carlson  <eric.carlson@apple.com>
1711
1712         Reviewed by Simon Fraser.
1713
1714         Fix for <rdar://problem/6685235>
1715         <video> element poster cannot be set dynamically if not originally set up in HTML
1716
1717         Allocate the media engine immediately so the plug-in proxy is available for
1718         scripting right away.
1719
1720         * platform/graphics/MediaPlayer.cpp:
1721         (WebCore::MediaPlayer::MediaPlayer):
1722
1723 2009-03-18  Beth Dakin  <bdakin@apple.com>
1724
1725         Reviewed by Dave Hyatt.
1726
1727         Fix for <rdar://problem/6636747> REGRESSION (Safari 4 PB-r41326):
1728         Popup menu appears at the wrong location on page at http://
1729         www.signonsandiego.com/
1730
1731         This was a regression from http://trac.webkit.org/changeset/40769,
1732         which changed the base class of RenderInline to
1733         RenderBoxModelObject rather than RenderBox.
1734
1735         * rendering/RenderObject.cpp:
1736         (WebCore::RenderObject::offsetParent): offsetParent should return a
1737         RenderBoxModelObject rather than just a RenderBox, which is more
1738         restrictive.
1739
1740 2009-03-18  Anders Carlsson  <andersca@apple.com>
1741
1742         Reviewed by Darin Adler.
1743
1744         <rdar://problem/6504776>
1745         CrashTracer: [USER] 188 crashes in Safari at com.apple.WebCore • WTF::HashTableIterator<WTF::RefPtr<WebCore::ResourceLoader>, ...
1746
1747         Return early in case calling the client ends up spinning the run loop and completing/cancelling the load.
1748
1749         * loader/NetscapePlugInStreamLoader.cpp:
1750         (WebCore::NetscapePlugInStreamLoader::didCancel):
1751
1752 2009-03-18  David Levin  <levin@chromium.org>
1753
1754         Reviewed by Dimitri Glazkov.
1755
1756         https://bugs.webkit.org/show_bug.cgi?id=24664
1757         Upstreaming v8 collection.h
1758
1759         No change in behavior, so no test.
1760
1761         * bindings/v8/V8Collection.h: Added.
1762
1763 2009-03-18  Sam Weinig  <sam@webkit.org>
1764
1765         Reviewed by Simon Fraser.
1766
1767         Fix for https://bugs.webkit.org/show_bug.cgi?id=23966
1768         REGRESSION: Custom arrow navigation functionality doesn't work at dropular.net
1769         <rdar://problem/6589657>
1770
1771         Match the CSSOM spec for getClientRects and getBoundingClientRect by
1772         returning a 0x0 rect at the correct top/left position for empty inline
1773         boxes.
1774
1775         * rendering/RenderInline.cpp:
1776         (WebCore::RenderInline::absoluteRects):
1777         (WebCore::RenderInline::absoluteQuads):
1778
1779 2009-03-18  David Levin  <levin@chromium.org>
1780
1781         Reviewed by Eric Seidel.
1782
1783         Match the changes done for windows in the bug
1784         https://bugs.webkit.org/show_bug.cgi?id=24530.
1785
1786         No change in behavior, so no test.
1787
1788         * rendering/RenderThemeChromiumLinux.cpp:
1789         (WebCore::RenderThemeChromiumLinux::extraDefaultStyleSheet):
1790
1791 2009-03-18  Dirk Schulze  <krit@webkit.org>
1792
1793         Reviewed by Holger Freyther.
1794
1795         Added support for SVG's stroked texts to Qt.
1796
1797         [Qt] SVGs stroked text support missing
1798         https://bugs.webkit.org/show_bug.cgi?id=24262
1799
1800         * platform/graphics/qt/FontQt.cpp:
1801         (WebCore::Font::drawComplexText):
1802
1803 2009-03-18  Mark Rowe  <mrowe@apple.com>
1804
1805         Reviewed by Alexey Proskuryakov.
1806
1807         Fix fast/encoding/char-decoding.html with recent ICU.
1808
1809         * platform/text/TextCodecICU.cpp:
1810         (WebCore::TextCodecICU::decode): Look for gb18030 case-insensitively,
1811         as newer versions of ICU use GB18030 as the canonical name.
1812
1813 2009-03-17  David Levin  <levin@chromium.org>
1814
1815         Reviewed by Dimitri Glazkov.
1816
1817         https://bugs.webkit.org/show_bug.cgi?id=24662
1818         Chromium build fixes.
1819
1820         Bug https://bugs.webkit.org/show_bug.cgi?id=24663 tracks
1821         the invalid implementation of ScriptCallStack.
1822
1823         No change in behavior, so no test.
1824
1825         * bindings/v8/ScriptCallStack.h:
1826         (WebCore::ScriptCallStack::state):
1827         * platform/graphics/skia/PlatformContextSkia.cpp:
1828         (PlatformContextSkia::drawRect):
1829
1830 2009-03-17  David Carson  <dacarson@apple.com>
1831
1832         Reviewed by David Hyatt.
1833
1834         Add anchor elements to the simple style sheet so that applications that
1835         are using WebKit just for simple text and links will also take advantage
1836         of the fast path.
1837
1838         * css/CSSStyleSelector.cpp:
1839         (WebCore::elementCanUseSimpleDefaultStyle):
1840
1841 2009-03-17  Darin Adler  <darin@apple.com>
1842
1843         Reviewed by Adele Peterson.
1844
1845         <rdar://problem/6687005> Need support for new move-left/right selectors.
1846
1847         * editing/EditorCommand.cpp:
1848         (WebCore::executeMoveToLeftEndOfLine): Added.
1849         (WebCore::executeMoveToLeftEndOfLineAndModifySelection): Added.
1850         (WebCore::executeMoveToRightEndOfLine): Added.
1851         (WebCore::executeMoveToRightEndOfLineAndModifySelection): Added.
1852         Added command entries for the functions above.
1853
1854 2009-03-17  Ojan Vafai  <ojan@chromium.org>
1855
1856         Reviewed by Adele Peterson.
1857
1858         Put the padding inside scrollbars on textareas.
1859         https://bugs.webkit.org/show_bug.cgi?id=24370
1860
1861         Put the padding inside the scrollbars on textareas. This is done by moving the
1862         overflow from the shadowNode to the RenderTextControlMultiline. As a result,
1863         all of the scroll handing methods that RenderTextControl overrides can be moved
1864         down into RenderTextControlSingleLine since RenderTextControlMultiline can now
1865         just use RenderBlock's versions. This also allows RenderTextControlMultiLine to
1866         no longer need a custom layout method since the shadowNode can now just size
1867         like a regular DIV.
1868
1869         * css/CSSStyleSelector.cpp:
1870         (WebCore::CSSStyleSelector::adjustRenderStyle):
1871         * rendering/RenderTextControl.cpp:
1872         (WebCore::RenderTextControl::styleDidChange):
1873         (WebCore::RenderTextControl::selection):
1874         (WebCore::RenderTextControl::calcHeight):
1875         (WebCore::RenderTextControl::hitInnerTextElement):
1876         * rendering/RenderTextControl.h:
1877         * rendering/RenderTextControlMultiLine.cpp:
1878         (WebCore::RenderTextControlMultiLine::nodeAtPoint):
1879         (WebCore::RenderTextControlMultiLine::createInnerTextStyle):
1880         * rendering/RenderTextControlMultiLine.h:
1881         * rendering/RenderTextControlSingleLine.cpp:
1882         (WebCore::RenderTextControlSingleLine::styleDidChange):
1883         (WebCore::RenderTextControlSingleLine::autoscroll):
1884         (WebCore::RenderTextControlSingleLine::scrollWidth):
1885         (WebCore::RenderTextControlSingleLine::scrollHeight):
1886         (WebCore::RenderTextControlSingleLine::scrollLeft):
1887         (WebCore::RenderTextControlSingleLine::scrollTop):
1888         (WebCore::RenderTextControlSingleLine::setScrollLeft):
1889         (WebCore::RenderTextControlSingleLine::setScrollTop):
1890         (WebCore::RenderTextControlSingleLine::scroll):
1891         * rendering/RenderTextControlSingleLine.h:
1892         * rendering/TextControlInnerElements.cpp:
1893         (WebCore::RenderTextControlInnerBlock::RenderTextControlInnerBlock):
1894         (WebCore::RenderTextControlInnerBlock::positionForPoint):
1895         (WebCore::TextControlInnerTextElement::createRenderer):
1896
1897 2009-03-17  Darin Adler  <darin@apple.com>
1898
1899         Reviewed by David Hyatt.
1900
1901         Bug 24517: REGRESSION (r41552): innerHTML does an updateLayout -- unneeded and can be slow
1902         https://bugs.webkit.org/show_bug.cgi?id=24517
1903
1904         * editing/DeleteButtonController.cpp:
1905         (WebCore::DeleteButtonController::enable): Added a call to updateRendering, since
1906         determining whether to display the delete button involves style and updateRendering
1907         also updates style (should probably be named updateStyle, in fact). Not needed to fix
1908         this bug, but would have prevented the crash that led to this bug in the first place.
1909
1910         * editing/EditCommand.cpp:
1911         (WebCore::EditCommand::EditCommand): Get rid of unneeded null check. All frames have
1912         delete button controllers.
1913         * editing/Editor.cpp:
1914         (WebCore::Editor::rangeForPoint): Ditto.
1915
1916         * editing/markup.cpp:
1917         (WebCore::appendStartMarkup): Changed a "&" to a "&&" so that generating markup
1918         doesn't depend on renderers at all when the convertBlocksToInlines boolean is false.
1919         This allows us to omit the call to updateLayoutIgnorePendingStylesheets in the
1920         createMarkup function that's called by innerHTML.
1921         (WebCore::MarkupAccumulator::appendMarkup): Turned this into a class with a member
1922         function. Added a feature where the accumulator will skip a node. Moved arguments
1923         that don't change during recursion into an object. This function still is a bit
1924         inefficient, since it creates a new HashMap at every level as it recurses, but for now
1925         I did not tackle that. Also replaced the onlyIncludeChildren boolean with EChildrenOnly
1926         for consistency and clarity.
1927         (WebCore::createMarkup): Removed the call to updateLayoutIgnorePendingStylesheets.
1928         Instead of calling disable/enable on the delete button controller's container element,
1929         pass it in to the markup accumulator as a node to skip.
1930
1931 2009-03-17  Scott Violet  <sky@google.com>
1932
1933         Reviewed by Eric Seidel.
1934
1935         https://bugs.webkit.org/show_bug.cgi?id=24651
1936         Skia does not always render text fill/stroke pattern/gradient/color correctly
1937
1938         Changes Skia's font rendering to only render gradient/pattern if current
1939         color space indicates the gradient/pattern should be used.
1940         This is covered by LayoutTests/fast/canvas/canvas-text-alignment.html .
1941
1942         * platform/graphics/GraphicsContext.cpp:
1943         (WebCore::GraphicsContext::strokeColorSpace):
1944         (WebCore::GraphicsContext::fillColorSpace):
1945         * platform/graphics/GraphicsContext.h:
1946         (WebCore::):
1947         * platform/graphics/GraphicsContextPrivate.h:
1948         * platform/graphics/skia/SkiaFontWin.cpp:
1949         (WebCore::paintSkiaText):
1950
1951 2009-03-17  Adele Peterson  <adele@apple.com>
1952
1953         Reviewed by Darin Adler.
1954
1955         Fix for https://bugs.webkit.org/show_bug.cgi?id=24655
1956         <rdar://problem/6633727> Hitting return at the end of a line with an anchor jumps me to the bottom of the message
1957
1958         Test: editing/inserting/6633727.html
1959
1960         This changes does a few things:
1961         1) Renames pos to insertionPosition.
1962         2) Eliminates "startNode".  It doesn't work well to consider the node separately from the insertionPosition.
1963            The insertionPosition gets updated at various times, and it seems likely that startNode can get out of sync.
1964         3) Before building up a list of ancestors to move around when we insert the new block, make sure to use the deepest
1965            representation of the insertionPosition, so all ancestor nodes are correctly included.
1966
1967         * editing/InsertParagraphSeparatorCommand.cpp: (WebCore::InsertParagraphSeparatorCommand::doApply):
1968
1969 2009-03-17  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
1970
1971         Reviewed by Kevin Ollivier.
1972
1973         https://bugs.webkit.org/show_bug.cgi?id=24115
1974         Introduce platform independent stubs for plugins.
1975
1976         * plugins/PluginDataNone.cpp: Copied from WebCore/plugins/wx/PluginDataWx.cpp.
1977         * plugins/PluginPackageNone.cpp: Copied from WebCore/plugins/wx/PluginPackageWx.cpp.
1978         * plugins/PluginViewNone.cpp: Copied from WebCore/plugins/wx/PluginViewWx.cpp.
1979         * plugins/wx/PluginDataWx.cpp: Removed.
1980         * plugins/wx/PluginPackageWx.cpp: Removed.
1981         * plugins/wx/PluginViewWx.cpp: Removed.
1982         * webcore-wx.bkl:
1983
1984 2009-03-17  Darin Adler  <darin@apple.com>
1985
1986         Earlier version reviewed by Adele Peterson.
1987
1988         Bug 24304: REGRESSION (r39864): Hitting the space bar to select an <input type=radio>
1989         or push an <input type=button> or <button> causes the page to scroll down.
1990
1991         Would be best to add a regression test for Windows eventually; tested that this has
1992         no effect on the Mac OS X platform.
1993
1994         * html/HTMLInputElement.cpp:
1995         (WebCore::HTMLInputElement::defaultEventHandler): Added FIXMEs and tweaked formatting.
1996         Use the code that calls the base class's defaultEventHandler early only in the cases
1997         where it's needed: keydown and keypress events in text fields. In other cases, do the
1998         more typical thing and call the default handler only at the end of the function.
1999         This function already had code to make sure the keypress event for space never gets
2000         through, but it was running too late since the scrolling code was moved into the
2001         base class default event handler.
2002
2003 2009-03-17  Simon Fraser  <simon.fraser@apple.com>
2004
2005         Reviewed by Dave Hyatt
2006
2007         https://bugs.webkit.org/show_bug.cgi?id=24632
2008
2009         Fix repaint issues when composited layers come and go (only applies
2010         when ACCELERATED_COMPOSITING is turned on).
2011
2012         * rendering/RenderLayer.cpp:
2013         (WebCore::RenderLayer::RenderLayer):
2014         * rendering/RenderLayer.h:
2015         (WebCore::RenderLayer::mustOverlayCompositedLayers):
2016         (WebCore::RenderLayer::setMustOverlayCompositedLayers):
2017         * rendering/RenderLayerBacking.cpp:
2018         (WebCore::RenderLayerBacking::RenderLayerBacking):
2019         * rendering/RenderLayerBacking.h:
2020
2021         Move what used to be the 'forceCompositingLayer' flag from RenderLayerBacking
2022         to RenderLayer, because we don't want the side-effects of creating RenderLayerBacking
2023         when setting this flag.
2024
2025         * rendering/RenderLayerCompositor.cpp:
2026         (WebCore::RenderLayerCompositor::updateLayerCompositingState):
2027         When a RenderLayer flips into or out of compositing mode, compute a repaint
2028         rect relative to the containerForRepaint, and repaint it.
2029
2030         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
2031         Call layer->setMustOverlayCompositedLayers() rather than setForcedCompositingLayer().
2032
2033         (WebCore::RenderLayerCompositor::needsToBeComposited):
2034         (WebCore::RenderLayerCompositor::requiresCompositingLayer):
2035         (WebCore::RenderLayerCompositor::requiresCompositingForTransform):
2036         (WebCore::RenderLayerCompositor::requiresCompositingForAnimation):
2037         * rendering/RenderLayerCompositor.h:
2038
2039         Rename requiresCompositingLayerForTransform() to requiresCompositingForTransform()
2040         and make it a class static method to match requiresCompositingForAnimation(). Both
2041         now take RenderObjects, rathern than RenderLayers.
2042
2043         * rendering/style/RenderStyle.h:
2044         (WebCore::InheritedFlags::hasTransformRelatedProperty):
2045         Minor tidyup using convenience methods added in an earlier commit.
2046
2047 2009-03-17  Simon Fraser  <simon.fraser@apple.com>
2048
2049         Reviewed by Darin Adler
2050
2051         https://bugs.webkit.org/show_bug.cgi?id=24396
2052
2053         * config.h:
2054         Add WTF_USE_ACCELERATED_COMPOSITING, defined to 0 for now, and add some
2055         comments to make the #ifdefs more readable.
2056
2057         * css/CSSComputedStyleDeclaration.cpp:
2058         (WebCore::computedTransform):
2059         Add a comment to mention that we don't flatten the matrix.
2060
2061         * css/MediaQueryEvaluator.cpp:
2062         (WebCore::transform_3dMediaFeatureEval):
2063         Have the 'transform-3d' media query evaluate to 'true' if 3d-rendering
2064         is supported.
2065
2066         * platform/graphics/mac/GraphicsLayerCA.mm:
2067         (WebCore::GraphicsLayerCA::animateTransform):
2068         No need for the #ifdef here. If we don't support 3d, we will have already flattened
2069         the matrix.
2070
2071         * platform/graphics/transforms/TransformationMatrix.cpp:
2072         (WebCore::TransformationMatrix::makeAffine):
2073         * platform/graphics/transforms/TransformationMatrix.h:
2074         New method to convert the matrix to an affine matrix by throwing a way the non-affine
2075         parts.
2076
2077         * rendering/RenderLayer.cpp:
2078         (WebCore::RenderLayer::updateTransform):
2079         (WebCore::RenderLayer::currentTransform):
2080         * rendering/RenderLayerBacking.cpp:
2081         (WebCore::RenderLayerBacking::updateLayerTransform):
2082         If 3d rendering is not supported, convert the matrix to an affine matrix
2083         which can be rendered, and used for hit testing.
2084
2085         * rendering/RenderLayerCompositor.cpp:
2086         Change the name of the exported symbol that webkitdirs.pm uses to know if
2087         3d rendering is supported. There is no other 3d-rendering-specific symbol we can sniff.
2088
2089         * rendering/RenderObject.cpp:
2090         (WebCore::RenderObject::transformFromContainer):
2091         Only take perspective into account if 3d rendering is supported.
2092
2093         * rendering/RenderObject.h:
2094         (WebCore::makeMatrixRenderable):
2095         Utility method that flattens a matrix if 3d rendering is not supported.
2096
2097 2009-03-17  Kevin Ollivier  <kevino@theolliviers.com>
2098
2099         wx build fix. Fix typo after mouse wheel changes.
2100
2101         * platform/wx/MouseWheelEventWx.cpp:
2102         (WebCore::PlatformWheelEvent::PlatformWheelEvent):
2103
2104 2009-03-17  Darin Adler  <darin@apple.com>
2105
2106         * inspector/ConsoleMessage.cpp:
2107         (WebCore::ConsoleMessage::isEqual): Fix build, remove stray parenthesis.
2108
2109 2009-03-17  Darin Adler  <darin@apple.com>
2110
2111         Fix crash seen right away when running run-webkit-tests.
2112
2113         * inspector/ConsoleMessage.cpp:
2114         (WebCore::ConsoleMessage::isEqual): Restore assertion to its behavior pre-refactoring.
2115         Also tweaked formatting a bit.
2116
2117 2009-03-17  Alexey Proskuryakov  <ap@webkit.org>
2118
2119         Reviewed by Darin Adler.
2120
2121         HTMLSelectElement::add() doesn't look at exception code returned from insertBefore(), so
2122         it doesn't need to zero it out before calling.
2123
2124         * html/HTMLSelectElement.cpp: (WebCore::HTMLSelectElement::add): Removed "ec = 0" line.
2125
2126 2009-03-17  Dan Bernstein  <mitz@apple.com>
2127
2128         Reviewed by Adam Roben.
2129
2130         - WebCore part of adding a mechanism for controlling the caching of
2131           responses through WebFrameLoaderClient
2132
2133         Mac already has such a mechanism, and this adds one for CFNetwork ports.
2134
2135         * WebCore.vcproj/WebCore.vcproj: Added EmptyClients.h
2136         and ResourceLoaderCFNet.cpp.
2137
2138         * loader/EmptyClients.h:
2139         (WebCore::EmptyFrameLoaderClient::shouldCacheResponse): Added an
2140         implementation that always returns true.
2141
2142         * loader/FrameLoaderClient.h: Declared shouldCacheResponse().
2143
2144         * loader/ResourceLoader.h: Ditto.
2145
2146         * loader/cf/ResourceLoaderCFNet.cpp: Added.
2147         (WebCore::ResourceLoader::shouldCacheResponse): Added. Calls through to
2148         FrameLoaderClient::shouldCacheResponse().
2149
2150         * platform/network/ResourceHandleClient.h:
2151         (WebCore::ResourceHandleClient::shouldCacheResponse): Added an
2152         implementation that always returns true.
2153
2154         * platform/network/cf/ResourceHandleCFNet.cpp:
2155         (WebCore::willCacheResponse): Added a call to
2156         ResourceHandleClient::shouldCacheResponse(). If the client returns
2157         false, return 0, which will prevent CFNetwork from caching the response.
2158
2159 2009-03-17  Alexey Proskuryakov  <ap@webkit.org>
2160
2161         Reviewed by Darin Adler.
2162
2163         https://bugs.webkit.org/show_bug.cgi?id=13287
2164         Cannot change SELECT to a dynamically created option
2165
2166         Tests: fast/forms/add-and-remove-option.html
2167                fast/forms/add-remove-option-modification-event.html
2168                fast/forms/add-selected-option.html
2169                fast/forms/select-cache-desynchronization.html
2170
2171         * dom/ContainerNode.cpp:
2172         (WebCore::dispatchChildInsertionEvents): Increment DOM tree version. This will happen when
2173         dispatching DOMSubtreeModified again, but the version should be incremented for event
2174         listeners to have an up to date view of the DOM.
2175         (WebCore::dispatchChildRemovalEvents): Ditto.
2176
2177         * html/HTMLOptionElement.cpp: (WebCore::HTMLOptionElement::insertedIntoTree):
2178         Make sure that the select element knows about its new selected option.
2179
2180         * html/HTMLOptionElement.h: Use insertedIntoTree() instead of insertedIntoDocument(),
2181         because DOM also needs to be updated for forms that are not in document yet. Similar
2182         problems exist for node removing, but removedFromTree() is called at a wrong time, so
2183         those problems cannot be fixed without deeper refactoring.
2184
2185         * html/HTMLSelectElement.cpp:
2186         (WebCore::HTMLSelectElement::setRecalcListItems): Reset m_activeSelectionAnchorIndex - it
2187         doesn't make sense to keep the anchor after programmatically changing the selection, and
2188         keeping it was causing a failure in fast/forms/listbox-selection.html.
2189
2190         * html/HTMLSelectElement.h: Removed overrides for ContainerNode methods that only called
2191         base class versions.
2192
2193 2009-03-17  Steve Falkenburg  <sfalken@apple.com>
2194
2195         <rdar://problem/6690324> Accessing FTP sites reads unallocated memory, can result in garbled entries or crashes
2196
2197         Reviewed by Darin Adler.
2198
2199         * loader/FTPDirectoryDocument.cpp:
2200         (WebCore::FTPDirectoryTokenizer::parseAndAppendOneLine): Assign CString to a local while we hold pointers into it.
2201
2202 2009-03-16  David Hyatt  <hyatt@apple.com>
2203
2204         <rdar://problem/6648411> REGRESSION: Layout of page is wrong at http://www.popcap.com/
2205
2206         Make sure that the initial shouldPaint check that looks at enclosingLayers properly skips over
2207         layers that don't paint themselves.  This is done by adding a new enclosingSelfPaintingLayer method
2208         so that RenderObjects can walk up the enclosing layer chain and skip any layers that don't paint
2209         themselves.
2210
2211         Reviewed by Darin Adler.
2212
2213         Added fast/block/float/overlapping-floats-with-overflow-hidden.html
2214
2215         * WebCore.base.exp:
2216         * rendering/RenderBlock.cpp:
2217         (WebCore::RenderBlock::addOverhangingFloats):
2218         * rendering/RenderObject.cpp:
2219         (WebCore::RenderObject::enclosingSelfPaintingLayer):
2220         * rendering/RenderObject.h:
2221
2222 2009-03-17  Xan Lopez  <xlopez@igalia.com>
2223
2224         Reviewed by Holger Freyther.
2225
2226         https://bugs.webkit.org/show_bug.cgi?id=24592
2227         [GTK] Crash in FcPatternHash
2228
2229         Style fixes.
2230
2231         * platform/graphics/gtk/FontPlatformDataGtk.cpp:
2232         (WebCore::FontPlatformData::operator=):
2233         * platform/graphics/gtk/FontPlatformDataPango.cpp:
2234
2235 2009-03-17  Xan Lopez  <xlopez@igalia.com>
2236
2237         Reviewed by Holger Freyther.
2238
2239         https://bugs.webkit.org/show_bug.cgi?id=24592
2240         [GTK] Crash in FcPatternHash
2241
2242         Sanitize memory management in pango fonts.
2243
2244         Release memory allocated by FontPlatformDataPango in its own
2245         destructor instead of doing it from other classes, and add copy
2246         constructor and '=' operator to be able to track referenced
2247         objects properly.
2248
2249         * platform/graphics/gtk/FontPlatformDataPango.cpp:
2250         (WebCore::FontPlatformData::~FontPlatformData):
2251         (WebCore::FontPlatformData::operator=):
2252         (WebCore::FontPlatformData::FontPlatformData):
2253         * platform/graphics/gtk/SimpleFontDataPango.cpp:
2254
2255 2009-03-17  Darin Adler  <darin@apple.com>
2256
2257         Reviewed by Alexey Proskuryakov.
2258
2259         Bug 24624: Crash in imageLoadEventTimerFired after adoptNode used on <img>,
2260         seen with inspector, which uses adoptNode
2261         https://bugs.webkit.org/show_bug.cgi?id=24624
2262         rdar://problem/6422850
2263
2264         Test: fast/dom/HTMLImageElement/image-load-cross-document.html
2265
2266         * dom/Document.cpp:
2267         (WebCore::Document::Document): Removed m_imageLoadEventTimer.
2268         (WebCore::Document::detach): Removed m_imageLoadEventDispatchSoonList and
2269         m_imageLoadEventDispatchingList.
2270         (WebCore::Document::implicitClose): Called ImageLoader::dispatchPendingLoadEvents
2271         instead of dispatchImageLoadEventsNow.
2272
2273         * dom/Document.h: Removed ImageLoader, dispatchImageLoadEventSoon,
2274         dispatchImageLoadEventsNow, removeImage, m_imageLoadEventDispatchSoonList,
2275         m_imageLoadEventDispatchingList, m_imageLoadEventTimer, and imageLoadEventTimerFired.
2276
2277         * loader/ImageLoader.cpp:
2278         (WebCore::loadEventSender): Added. Returns the single global ImageLoadEventSender
2279         object used privately as the target of the load event timer.
2280         (WebCore::ImageLoader::~ImageLoader): Call ImageLoadEventSender::cancelLoadEvent
2281         rather than Document::removeImage.
2282         (WebCore::ImageLoader::setImage): Use m_element directly, not element().
2283         (WebCore::ImageLoader::updateFromElement): Ditto. Also name the local variable
2284         document instead of doc.
2285         (WebCore::ImageLoader::notifyFinished): Call ImageLoadEventSender::dispatchLoadEventSoon
2286         rather than Document::dispatchImageLoadEventSoon.
2287         (WebCore::ImageLoader::dispatchPendingLoadEvent): Added. Handles the common logic
2288         about when load events can be dispatched so that dispatchLoadEvent only has to
2289         have the specific part for each derived class. This includes a check that the
2290         document is attached, which used to be handled by having documents empty out the
2291         image load event vectors in the detach function.
2292         (WebCore::ImageLoader::dispatchPendingLoadEvents): Added. Calls the appropriate
2293         function on the ImageLoadEventSender, which avoids the need to have that class be
2294         public in the ImageLoader header.
2295         (WebCore::ImageLoadEventSender::ImageLoadEventSender): Added. Has the code that
2296         was previously in the Document constructor.
2297         (WebCore::ImageLoadEventSender::dispatchLoadEventSoon): Added. Has the code that
2298         was previously in Document::dispatchImageLoadEventSoon.
2299         (WebCore::ImageLoadEventSender::cancelLoadEvent): Added. Has the code that was
2300         previously in Document::removeImage.
2301         (WebCore::ImageLoadEventSender::dispatchPendingLoadEvents): Added. Has the code
2302         that was previously in Document::dispatchImageLoadEventsNow.
2303         (WebCore::ImageLoadEventSender::timerFired): Added. Calls dispatchPendingLoadEvents.
2304
2305         * loader/ImageLoader.h: Improved comments. Made the virtual functions private
2306         or protected rather than public. Added static dispatchPendingLoadEvents function
2307         for use by Document and private dispatchPendingLoadEvent function for use by
2308         ImageLoadEventSender. Made setLoadingImage private and eliminated
2309         setHaveFiredLoadEvent since that can all be done inside the class without any
2310         member functions.
2311
2312         * html/HTMLImageLoader.cpp:
2313         (WebCore::HTMLImageLoader::dispatchLoadEvent): Removed logic to check whether a
2314         load event already fired and whether image() is 0. These are now both base class
2315         responsibilities.
2316         * svg/SVGImageLoader.cpp:
2317         (WebCore::SVGImageLoader::dispatchLoadEvent): Ditto.
2318         * wml/WMLImageLoader.cpp:
2319         (WebCore::WMLImageLoader::dispatchLoadEvent): Ditto.
2320
2321 2009-03-17  Dimitri Glazkov  <dglazkov@chromium.org>
2322
2323         Reviewed by Timothy Hatcher.
2324
2325         https://bugs.webkit.org/show_bug.cgi?id=24623
2326         Refactor ConsoleMessage to use ScriptFuncitonCall and eliminate JSC
2327         dependencies.
2328
2329         * bindings/js/ScriptFunctionCall.cpp:
2330         (WebCore::ScriptFunctionCall::appendArgument): Added uint and ScriptString-taking methods.
2331         * bindings/js/ScriptFunctionCall.h:
2332         * bindings/js/ScriptObjectQuarantine.cpp:
2333         (WebCore::quarantineValue): Added generic ScriptValue quarantine helper.
2334         * bindings/js/ScriptObjectQuarantine.h:
2335         * bindings/js/ScriptValue.cpp:
2336         (WebCore::ScriptValue::isEqual): Added.
2337         * bindings/js/ScriptValue.h:
2338         * inspector/ConsoleMessage.cpp:
2339         (WebCore::ConsoleMessage::ConsoleMessage):
2340         (WebCore::ConsoleMessage::addToConsole): Added.
2341         (WebCore::ConsoleMessage::isEqual): Changed to use ScriptValue::isEqual.
2342         * inspector/ConsoleMessage.h:
2343         (WebCore::ConsoleMessage::incrementCount): Added.
2344         * inspector/InspectorController.cpp:
2345         (WebCore::InspectorController::addConsoleMessage): Changed to use ConsoleMessage::addToConsole.
2346         (WebCore::InspectorController::populateScriptObjects): Ditto.
2347         * inspector/InspectorController.h:
2348
2349 2009-03-17  Kevin Ollivier  <kevino@theolliviers.com>
2350
2351         Reviewed by Mark Rowe.
2352
2353         Get BUILDING_ON_* defines from Platform.h.
2354
2355         https://bugs.webkit.org/show_bug.cgi?id=24630
2356
2357         * WebCorePrefix.h:
2358
2359 2009-03-16  Xan Lopez  <xlopez@igalia.com>
2360
2361         Reviewed by Holger Freyther.
2362
2363         https://bugs.webkit.org/show_bug.cgi?id=24592
2364         [GTK] Crash in FcPatternHash
2365
2366         Sanitize memory management in gtk fonts.
2367
2368         Release memory allocated by FontPlatformDataGtk in its own
2369         destructor instead of doing it from other classes, and add copy
2370         constructor and '=' operator to be able to track referenced
2371         objects properly.
2372
2373         * platform/graphics/gtk/FontPlatformData.h:
2374         * platform/graphics/gtk/FontPlatformDataGtk.cpp:
2375         (WebCore::FontPlatformData::operator=):
2376         (WebCore::FontPlatformData::FontPlatformData):
2377         (WebCore::FontPlatformData::~FontPlatformData):
2378         * platform/graphics/gtk/SimpleFontDataGtk.cpp:
2379         (WebCore::SimpleFontData::platformDestroy):
2380
2381 2009-03-17  Ariya Hidayat  <ariya.hidayat@nokia.com>
2382
2383         Build fix for Qt < 4.5.
2384
2385         As reported by Yael Aharon  <yael.aharon@nokia.com>
2386
2387         * platform/graphics/qt/GraphicsContextQt.cpp:
2388         (WebCore::GraphicsContext::drawLine):
2389
2390 2009-03-17  Alexey Proskuryakov  <ap@webkit.org>
2391
2392         Reviewed by Sam Weinig.
2393
2394         https://bugs.webkit.org/show_bug.cgi?id=24614
2395         Access control checks are different in cached and uncached cases
2396
2397         Test: http/tests/xmlhttprequest/access-control-basic-non-simple-deny-cached.html
2398
2399         * loader/CrossOriginAccessControl.cpp:
2400         (WebCore::isOnAccessControlSimpleRequestMethodWhitelist): Factored out simple method
2401         check for use in both cached and uncached cases. In cached case, an old definition that
2402         omitted HEAD was still used.
2403         (WebCore::isOnAccessControlSimpleRequestHeaderWhitelist): Check that content type has an
2404         allowed value. This is needed in all call sites. Also changed to compare MIME type, not
2405         content type.
2406         (WebCore::isSimpleCrossOriginAccessRequest): Use the above methods.
2407
2408         * loader/CrossOriginAccessControl.h: Expose isOnAccessControlSimpleRequestMethodWhitelist.
2409
2410         * loader/CrossOriginPreflightResultCache.cpp:
2411         (WebCore::CrossOriginPreflightResultCacheItem::allowsCrossOriginMethod):
2412         (WebCore::CrossOriginPreflightResultCacheItem::allowsCrossOriginHeaders):
2413         Use the new checks for simple method and header.
2414
2415 2009-03-16  Gustavo Noronha Silva  <gns@gnome.org> and Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
2416
2417         Reviewed by Mark Rowe.
2418
2419         https://bugs.webkit.org/show_bug.cgi?id=24638
2420         [GTK] HTML5 media tags do not work
2421
2422         Add a repaint-requested signal to the video sink, and use it to
2423         call MediaPlayerPrivate::repaint, so that the video actually
2424         plays.
2425
2426         * platform/graphics/gtk/MediaPlayerPrivateGStreamer.cpp:
2427         (WebCore::mediaPlayerPrivateRepaintCallback):
2428         (WebCore::MediaPlayerPrivate::createGSTPlayBin):
2429         * platform/graphics/gtk/VideoSinkGStreamer.cpp:
2430         (webkit_video_sink_idle_func):
2431         (webkit_video_sink_render):
2432         (webkit_video_sink_class_init):
2433
2434 2009-03-16  Gustavo Noronha Silva  <gns@gnome.org> and Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
2435
2436         Reviewed by Holger Freyther.
2437
2438         https://bugs.webkit.org/show_bug.cgi?id=24638
2439         [GTK] HTML5 media tags do not work
2440
2441         Work-around the fact that gst_element_query_duration returns true even
2442         though it is unable to figure out the duration when in stream (push)
2443         mode.
2444
2445         * platform/graphics/gtk/MediaPlayerPrivateGStreamer.cpp:
2446         (WebCore::MediaPlayerPrivate::duration):
2447
2448 2009-03-16  Darin Adler  <darin@apple.com>
2449
2450         Reviewed by Kevin Decker.
2451
2452         <rdar://problem/6642742> Top Sites malfunction when switching text zoom mode
2453
2454         * page/Frame.cpp:
2455         (WebCore::Frame::setNeedsReapplyStyles): Don’t do anything if the frame is
2456         currently showing a non-HTML view.
2457
2458 2009-03-16  Darin Adler  <darin@apple.com>
2459
2460         Reviewed by Adele Peterson.
2461
2462         Bug 24629: moving forward or backward a paragraph fails at edge of document
2463         https://bugs.webkit.org/show_bug.cgi?id=24629
2464         rdar://problem/6544413
2465
2466         Test: editing/selection/move-paragraph-document-edges.html
2467
2468         * editing/visible_units.cpp:
2469         (WebCore::previousParagraphPosition): Use the last result from
2470         previousLinePosition rather than going all the way back to what was originally
2471         passed in when we hit exception cases like null or not moving. This correctly
2472         inherits the behavior of previousLinePosition when we are in a paragraph at the
2473         edge of a document.
2474         (WebCore::nextParagraphPosition): Ditto.
2475
2476 2009-03-16  Darin Adler  <darin@apple.com>
2477
2478         Reviewed by Adele Peterson.
2479
2480         Bug 24619: RenderObject::selectionStartEnd does not need to be a virtual function
2481         https://bugs.webkit.org/show_bug.cgi?id=24619
2482
2483         * rendering/RenderObject.h: Remove virtual keyword from selectionStartEnd declaration.
2484         * rendering/RenderView.h: Ditto.
2485
2486 2009-03-16  Peter Kasting  <pkasting@google.com>
2487
2488         Reviewed by David Hyatt.
2489
2490         https://bugs.webkit.org/show_bug.cgi?id=24368
2491         DOM scroll events should be based off the actual number of wheel
2492         ticks, not off the number of lines scrolled.  This matches IE.
2493
2494         * dom/Node.cpp:
2495         (WebCore::Node::dispatchWheelEvent):
2496         * dom/WheelEvent.cpp:
2497         (WebCore::WheelEvent::WheelEvent):
2498         * dom/WheelEvent.h:
2499         (WebCore::WheelEvent::create):
2500         * platform/PlatformWheelEvent.h:
2501         (WebCore::PlatformWheelEvent::wheelTicksX):
2502         (WebCore::PlatformWheelEvent::wheelTicksY):
2503         * platform/gtk/WheelEventGtk.cpp:
2504         (WebCore::PlatformWheelEvent::PlatformWheelEvent):
2505         * platform/mac/WheelEventMac.mm:
2506         (WebCore::PlatformWheelEvent::PlatformWheelEvent):
2507         * platform/qt/WheelEventQt.cpp:
2508         * platform/win/WheelEventWin.cpp:
2509         (WebCore::PlatformWheelEvent::PlatformWheelEvent):
2510         * platform/wx/MouseWheelEventWx.cpp:
2511         (WebCore::PlatformWheelEvent::PlatformWheelEvent):
2512
2513 2009-03-16  Simon Fraser  <simon.fraser@apple.com>
2514
2515         Reviewed by Eric Seidel
2516
2517         Clean up a few issues in the Animation code:
2518
2519         * page/animation/AnimationBase.cpp:
2520         (WebCore::AnimationBase::updateStateMachine):
2521         Whitespace
2522
2523         (WebCore::AnimationBase::willNeedService):
2524         Don't round to float, use std::max
2525
2526         * page/animation/AnimationController.cpp:
2527         (WebCore::AnimationControllerPrivate::startTimeResponse):
2528         Fix erroneously copied line to null out m_lastResponseWaiter.
2529
2530 2009-03-12  David Hyatt  <hyatt@apple.com>
2531
2532         Reviewed by Eric Seidel
2533
2534         https://bugs.webkit.org/show_bug.cgi?id=13632
2535
2536         Overflow scrolling needs to account for the bottom/right padding on the object itself as well
2537         as for bottom/right margins on children.
2538
2539         Existing tests cover this.
2540
2541         * rendering/RenderBlock.cpp:
2542         (WebCore::RenderBlock::lowestPosition):
2543         (WebCore::RenderBlock::rightmostPosition):
2544
2545 2009-03-16  Sam Weinig  <sam@webkit.org>
2546
2547         Reviewed by Anders Carlsson.
2548
2549         Fix for <rdar://problem/6320555>
2550         Add an upper limit for setting HTMLSelectElement.length.
2551
2552         Test: fast/forms/select-max-length.html
2553
2554         * html/HTMLSelectElement.cpp:
2555         (WebCore::HTMLSelectElement::setOption):
2556         (WebCore::HTMLSelectElement::setLength):
2557
2558 2009-03-16  Eric Carlson  <eric.carlson@apple.com>
2559
2560         Reviewed by Simon Fraser.
2561
2562         <rdar://problem/6686721> Media document crash in 64-bit WebKit
2563
2564         * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
2565         (WebCore::MediaPlayerPrivate::createQTMovieView): QTMovieContentViewClass is only used when
2566         rendering inline with old versions of QuickTime, so don't look for it when we are in a
2567         media document.
2568
2569 2009-03-16  Dimitri Glazkov  <dglazkov@chromium.org>
2570
2571         Reviewed by Timothy Hatcher.
2572
2573         https://bugs.webkit.org/show_bug.cgi?id=24590
2574         Refactor InspectorDOMStorageResource to use ScriptFunctionCall.
2575
2576         * bindings/js/ScriptFunctionCall.cpp:
2577         (WebCore::ScriptFunctionCall::appendArgument): Added method for bool argument.
2578         * bindings/js/ScriptFunctionCall.h: Ditto, also cleaned up.
2579         * bindings/js/ScriptObjectQuarantine.cpp:
2580         (WebCore::getQuarantinedScriptObject): Added Storage helper.
2581         * bindings/js/ScriptObjectQuarantine.h: Ditto.
2582         * inspector/InspectorController.cpp:
2583         (WebCore::InspectorController::populateScriptObjects): Changed to use bind method.
2584         (WebCore::InspectorController::resetScriptObjects): Changed to use unbind method.
2585         (WebCore::InspectorController::didUseDOMStorage): Changed to use isSameHostAndType and bind methods.
2586         * inspector/InspectorController.h: Removed add/remove methods for DOM storage.
2587         * inspector/InspectorDOMStorageResource.cpp:
2588         (WebCore::InspectorDOMStorageResource::InspectorDOMStorageResource):
2589         (WebCore::InspectorDOMStorageResource::isSameHostAndType): Added.
2590         (WebCore::InspectorDOMStorageResource::bind): Added.
2591         (WebCore::InspectorDOMStorageResource::unbind): Added.
2592         * inspector/InspectorDOMStorageResource.h:
2593
2594 2009-03-16   Mike Belshe  <mike@belse.com>
2595
2596         Reviewed by Dimitri Glazkov.
2597
2598         https://bugs.webkit.org/show_bug.cgi?id=24580
2599         Fix query() to match KURL behavior, this time with the code that
2600         compiles.
2601
2602         * platform/KURLGoogle.cpp:
2603         (WebCore::KURL::query): Fix copy/paste mistake.
2604
2605 2009-03-16  Alexey Proskuryakov  <ap@webkit.org>
2606
2607         Reviewed by Darin Adler.
2608
2609         https://bugs.webkit.org/show_bug.cgi?id=21752
2610         REGRESSION: referencing XHR constructor for a not yet loaded frame permanently breaks it
2611
2612         Test: fast/dom/Window/window-early-properties-xhr.html
2613
2614         For some transitions, the Window object is not replaced, but Document is. When this happened,
2615         window.document property was updated, but references to Document kept in cached constructors
2616         were not.
2617
2618         * bindings/js/JSAudioConstructor.cpp:
2619         (WebCore::JSAudioConstructor::JSAudioConstructor):
2620         (WebCore::JSAudioConstructor::document):
2621         (WebCore::JSAudioConstructor::mark):
2622         * bindings/js/JSAudioConstructor.h:
2623         * bindings/js/JSImageConstructor.cpp:
2624         (WebCore::JSImageConstructor::JSImageConstructor):
2625         (WebCore::JSImageConstructor::document):
2626         (WebCore::JSImageConstructor::mark):
2627         * bindings/js/JSImageConstructor.h:
2628         * bindings/js/JSMessageChannelConstructor.cpp:
2629         (WebCore::JSMessageChannelConstructor::JSMessageChannelConstructor):
2630         (WebCore::JSMessageChannelConstructor::scriptExecutionContext):
2631         (WebCore::JSMessageChannelConstructor::mark):
2632         * bindings/js/JSMessageChannelConstructor.h:
2633         * bindings/js/JSOptionConstructor.cpp:
2634         (WebCore::JSOptionConstructor::JSOptionConstructor):
2635         (WebCore::JSOptionConstructor::document):
2636         (WebCore::JSOptionConstructor::mark):
2637         * bindings/js/JSOptionConstructor.h:
2638         * bindings/js/JSXMLHttpRequestConstructor.cpp:
2639         (WebCore::JSXMLHttpRequestConstructor::JSXMLHttpRequestConstructor):
2640         (WebCore::JSXMLHttpRequestConstructor::scriptExecutionContext):
2641         (WebCore::JSXMLHttpRequestConstructor::mark):
2642         * bindings/js/JSXMLHttpRequestConstructor.h:
2643         Changed cached constructors to keep a reference to Window, not Document.
2644
2645 2009-03-15  Alexey Proskuryakov  <ap@webkit.org>
2646
2647         Reviewed by Darin Adler.
2648
2649         https://bugs.webkit.org/show_bug.cgi?id=24549
2650         Impose a limit on Access-Control-Max-Age value
2651
2652         * loader/CrossOriginPreflightResultCache.cpp:
2653         (WebCore::CrossOriginPreflightResultCacheItem::parse):
2654
2655 2009-03-15  Greg Bolsinga  <bolsinga@apple.com>
2656
2657         Reviewed by David Kilzer.
2658
2659         <rdar://problem/6668875> Normalize Geolocation results
2660
2661         * platform/mac/GeolocationServiceMac.mm:
2662         (-[WebCoreCoreLocationObserver locationManager:didUpdateToLocation:fromLocation:]):
2663
2664 2009-03-15  Greg Bolsinga  <bolsinga@apple.com>
2665
2666         Reviewed by David Kilzer.
2667
2668         Update ::toString format as suggested by Darin Adler.
2669
2670         * page/Geoposition.cpp:
2671         (WebCore::Geoposition::toString):
2672
2673 2009-03-15  David Kilzer  <ddkilzer@apple.com>
2674
2675         <rdar://problem/6668238> WebCore is registering text encodings needlessly from KURL constructor.
2676
2677         Reviewed by Darin Adler.
2678
2679         Yet another case where we would trigger extended encoding loading needlessly.
2680
2681         * platform/text/TextEncoding.cpp:
2682         (WebCore::TextEncoding::encodingForFormSubmission):
2683
2684 2009-03-15  Simon Fraser  <simon.fraser@apple.com>
2685
2686         Build fix: no review.
2687
2688         * rendering/style/ContentData.h:
2689
2690 2009-03-15  David Kilzer  <ddkilzer@apple.com>
2691
2692         Bug 24542: Improve ContentData encapsulation
2693
2694         <https://bugs.webkit.org/show_bug.cgi?id=24542>
2695
2696         Reviewed by Simon Fraser.
2697
2698         No tests since there is no change in behavior.
2699
2700         * rendering/RenderObject.cpp:
2701         (WebCore::RenderObject::createObject): Used getter methods
2702         instead of data members on ContentData class.  Used isImage()
2703         convenience method.
2704         * rendering/RenderObjectChildList.cpp:
2705         (WebCore::RenderObjectChildList::updateBeforeAfterContent): Ditto.
2706
2707         * rendering/style/ContentData.cpp:
2708         (WebCore::ContentData::clear): Extracted code into
2709         deleteContent() method.
2710         (WebCore::ContentData::dataEquivalent): Added.  Extracted code
2711         from StyleRareNonInheritedData::contentDataEquivalent().
2712         (WebCore::ContentData::deleteContent): Added.  Used by setter
2713         methods.
2714         * rendering/style/ContentData.h: Made m_type, m_content and
2715         m_next private.
2716         (WebCore::ContentData::isCounter): Added.
2717         (WebCore::ContentData::isImage): Added.
2718         (WebCore::ContentData::isNone): Added.
2719         (WebCore::ContentData::isText): Added.
2720         (WebCore::ContentData::type): Added.
2721         (WebCore::ContentData::dataEquivalent): Added.
2722         (WebCore::ContentData::image): Added.
2723         (WebCore::ContentData::setImage): Added.
2724         (WebCore::ContentData::text): Added.
2725         (WebCore::ContentData::setText): Added.
2726         (WebCore::ContentData::counter): Added.
2727         (WebCore::ContentData::setCounter): Added.
2728         (WebCore::ContentData::next): Added.
2729         (WebCore::ContentData::setNext): Added.
2730
2731         * rendering/style/CounterContent.h:
2732         (WebCore::operator!=): Removed.
2733         (WebCore::operator==): Renamed operator!=() and reversed its
2734         logic after extracting code from
2735         StyleRareNonInheritedData::contentDataEquivalent() to create
2736         ContentData::dataEquivalent().
2737
2738         * rendering/style/RenderStyle.cpp:
2739         (WebCore::RenderStyle::setContent): Used new getters and setters
2740         on ContentData class.  Changed first argument from a StringImpl*
2741         to a PassRefPtr<StrimgImpl>.  Used isText() convenience method.
2742         * rendering/style/RenderStyle.h:
2743         (WebCore::RenderStyle::setContent): Updated declaration.
2744         * rendering/style/StyleRareNonInheritedData.cpp:
2745         (WebCore::StyleRareNonInheritedData::contentDataEquivalent):
2746         Extracted most logic in while() loop into
2747         ContentData::dataEquivalent().
2748
2749 2009-03-15  Gustavo Noronha Silva  <gns@gnome.org>
2750
2751         Reviewed by Anders Carlsson.
2752
2753         https://bugs.webkit.org/show_bug.cgi?id=24602
2754         [Gtk] Searching in thepiratebay.org doesn't work with more than 1 word
2755
2756         Reintroduce the URI into the soup message after having set it in
2757         the KURL, on redirects, to make sure it is properly encoded. This
2758         fixes bad request problems when servers give bad URIs on their
2759         response's Location header.
2760
2761         * platform/network/soup/ResourceHandleSoup.cpp:
2762         (WebCore::restartedCallback):
2763
2764 2009-03-15  Alexey Proskuryakov  <ap@webkit.org>
2765
2766         Reviewed by Darin Adler.
2767
2768         https://bugs.webkit.org/show_bug.cgi?id=19737
2769         No cursor and paste not enabled right clicking text field/area
2770
2771         Test: fast/events/right-click-focus.html
2772
2773         * page/EventHandler.cpp:
2774         (WebCore::EventHandler::handleMousePressEventSingleClick):
2775         (WebCore::EventHandler::handleMousePressEvent):
2776         Take normal code path for right clicks (we were taking it for Ctrl-clicks anyway).
2777
2778 2009-03-15  Alexey Proskuryakov  <ap@webkit.org>
2779
2780         Reviewed by Adele Peterson.
2781
2782         https://bugs.webkit.org/show_bug.cgi?id=23949
2783         HTMLSelectElement is in inconsistent state when handling mutation events
2784
2785         Test: fast/forms/mutation-event-recalc.html
2786
2787         * dom/ContainerNode.cpp:
2788         (WebCore::ContainerNode::insertBefore):
2789         (WebCore::ContainerNode::appendChild):
2790         Call childrenChanged() before dispatching modification events, not after.
2791
2792         * html/HTMLOptGroupElement.cpp:
2793         (WebCore::HTMLOptGroupElement::insertBefore):
2794         (WebCore::HTMLOptGroupElement::replaceChild):
2795         (WebCore::HTMLOptGroupElement::removeChild):
2796         (WebCore::HTMLOptGroupElement::appendChild):
2797         (WebCore::HTMLOptGroupElement::removeChildren):
2798         * html/HTMLSelectElement.cpp:
2799         (WebCore::HTMLSelectElement::add):
2800         (WebCore::HTMLSelectElement::remove):
2801         (WebCore::HTMLSelectElement::insertBefore):
2802         (WebCore::HTMLSelectElement::replaceChild):
2803         (WebCore::HTMLSelectElement::removeChild):
2804         (WebCore::HTMLSelectElement::appendChild):
2805         (WebCore::HTMLSelectElement::removeChildren):
2806         Remove calls to recalcSelectOptions(). It is too late to recalc now, after mutation events
2807         were already dispatched.
2808
2809 2009-03-14  Greg Bolsinga  <bolsinga@apple.com>
2810
2811         <rdar://problem/6683465>
2812
2813         Geolocation has to be able to handle NULL Frames.
2814
2815         Reviewed by Mark Rowe
2816
2817 2009-03-14  Greg Bolsinga  <bolsinga@apple.com>
2818
2819         Reviewed by David Kilzer.
2820
2821         Geoposition::toString was missing the Coordinates after the update.
2822
2823         * page/Geoposition.cpp:
2824         (WebCore::Geoposition::toString):
2825
2826 2009-03-14  Jan Michael Alonzo  <jmalonzo@webkit.org>
2827
2828         Reviewed by Mark Rowe.
2829
2830         [Gtk] Build fix - pass a Coordinate to Geoposition::create
2831         https://bugs.webkit.org/show_bug.cgi?id=24603
2832
2833         Gtk build fix per r41650
2834         https://bugs.webkit.org/show_bug.cgi?id=24506
2835         Geolocation in Safari differs from the spec, no Coordinates attribute on Position
2836
2837         Pass a Coordinate to Geoposition::create.
2838
2839         * platform/gtk/GeolocationServiceGtk.cpp:
2840         (WebCore::GeolocationServiceGtk::updatePosition):
2841
2842 2009-03-14  Alexey Proskuryakov  <ap@webkit.org>
2843
2844         Reviewed by Darin Adler.
2845
2846         https://bugs.webkit.org/show_bug.cgi?id=24545
2847         Make cross-site XHR simple request definition match current spec draft
2848
2849         Tests: http/tests/xmlhttprequest/access-control-basic-post-fail-non-simple-content-type.html
2850                http/tests/xmlhttprequest/access-control-basic-whitelist-request-headers.html
2851
2852         * loader/CrossOriginAccessControl.cpp:
2853         (WebCore::isOnAccessControlSimpleRequestHeaderWhitelist): Added Content-Language.
2854         (WebCore::isSimpleCrossOriginAccessRequest): Added HEAD. Restricted content types to those
2855         that could be sent via form submission.
2856
2857 2009-03-14  Alexey Proskuryakov  <ap@webkit.org>
2858
2859         Reviewed by Darin Adler.
2860
2861         https://bugs.webkit.org/show_bug.cgi?id=15172
2862         object fallback - empty string for first argument of setAttributeNS does not work like null
2863
2864         https://bugs.webkit.org/show_bug.cgi?id=24548
2865         createElementNS("", name) should create an element in null namespace
2866
2867         Tests: fast/dom/createElementNS-empty-namespace.html
2868                fast/dom/setAttributeNS-empty-namespace.html
2869
2870         * dom/QualifiedName.h: (WebCore::QualifiedName::QualifiedNameImpl::QualifiedNameImpl):
2871         Convert empty namespaces to null.
2872
2873 2009-03-13  Mark Rowe  <mrowe@apple.com>
2874
2875         Rubber-stamped by Dan Bernstein.
2876
2877         Take advantage of the ability of recent versions of Xcode to easily switch the active
2878         architecture.
2879
2880         * Configurations/DebugRelease.xcconfig:
2881
2882 2009-03-13  John Abd-El-Malek  <jam@google.com>
2883
2884         Reviewed by Darin Fisher.
2885
2886         https://bugs.webkit.org/show_bug.cgi?id=24593
2887         Added requestorID so we can track the request to its WebView without using frame (which was removed).
2888
2889         * platform/network/chromium/ResourceRequest.h:
2890         (WebCore::ResourceRequest::ResourceRequest):
2891         (WebCore::ResourceRequest::requestorID):
2892         (WebCore::ResourceRequest::setRequestorID):
2893         (WebCore::ResourceRequest::setTargetType):
2894         (WebCore::ResourceRequest::policyURL):
2895         (WebCore::ResourceRequest::setPolicyURL):
2896         (WebCore::ResourceRequest::requestorPid):
2897         (WebCore::ResourceRequest::setRequestorPid):
2898
2899 2009-03-13  Mike Belshe  <mike@belshe.com>
2900
2901         Reviewed by Dimitri Glazkov.
2902
2903         https://bugs.webkit.org/show_bug.cgi?id=24561
2904         Add custom V8 bindings for HTMLElementCanvas, Location.
2905
2906         * bindings/v8/custom/V8HTMLCanvasElementCustom.cpp: Added.
2907         * bindings/v8/custom/V8LocationCustom.cpp: Added.
2908
2909 2009-03-13  Stephen White  <senorblanco@chromium.org>
2910
2911         Reviewed by Eric Seidel.
2912
2913         https://bugs.webkit.org/show_bug.cgi?id=24584
2914
2915         Fix transparent text rendering on Chromium.  FontChromiumWin
2916         was calling beginTransparencyLayer()/endTransparencyLayer(), with
2917         a TransparencyWin inside to do GDI ClearType rendering over an
2918         opaque background.  TransparencyWin does its special sauce
2919         in the destructor, but it was being called too late to be used
2920         correctly in the layer.  Put the special sauce into a new function,
2921         composite(), and call that explicitly instead.
2922
2923         * platform/graphics/chromium/FontChromiumWin.cpp:
2924         (WebCore::TransparencyAwareFontPainter::TransparencyAwareFontPainter::~TransparencyAwareFontPainter):
2925         * platform/graphics/chromium/TransparencyWin.cpp:
2926         (WebCore::TransparencyWin::~TransparencyWin):
2927         (WebCore::TransparencyWin::composite):
2928         (WebCore::TransparencyWin::init):
2929         * platform/graphics/chromium/TransparencyWin.h:
2930         * rendering/RenderThemeChromiumWin.cpp:
2931         (WebCore::):
2932
2933 2009-03-13  Mike Belshe  <mike@belshe.com>
2934
2935         Reviewed by Dimitri Glazkov.
2936
2937         https://bugs.webkit.org/show_bug.cgi?id=24562
2938         Add custom implementation for getCSSCanvasContext to V8 bindings.
2939
2940         * bindings/v8/custom/V8DocumentCustom.cpp: Added new method.
2941
2942 2009-03-13  Mike Belshe  <mike@belse.com>
2943
2944         Reviewed by Dimitri Glazkov.
2945
2946         https://bugs.webkit.org/show_bug.cgi?id=24580
2947         Fix query() to match KURL behavior.
2948
2949         * platform/KURLGoogle.cpp:
2950         (WebCore::KURL::query): remove extra logic around question mark.
2951
2952 2009-03-13  Chris Fleizach  <cfleizach@apple.com>
2953
2954         Reviewed by Beth Dakin.
2955
2956         Bug 24474: AX: in multi-body tables, asking for a cell at a specific coordinate can return nil
2957         https://bugs.webkit.org/show_bug.cgi?id=24474
2958
2959         Test: platform/mac-snowleopard/accessibility/table-multi-bodies.html
2960
2961         * page/AccessibilityTable.cpp:
2962         (WebCore::AccessibilityTable::cellForColumnAndRow):
2963
2964 2009-03-13  Jian Li  <jianli@chromium.org>
2965
2966         Reviewed by Dimitri Glazkov.
2967
2968         https://bugs.webkit.org/show_bug.cgi?id=24589
2969         Upstream changes to V8 event listeners (Chromium r11133) in order to
2970         fix worker functionality break in Chromium.
2971
2972         * bindings/v8/V8AbstractEventListener.cpp:
2973         (WebCore::V8AbstractEventListener::invokeEventHandler):
2974         (WebCore::V8AbstractEventListener::handleEvent):
2975         * bindings/v8/V8AbstractEventListener.h:
2976         * bindings/v8/V8WorkerContextEventListener.cpp:
2977         (WebCore::V8WorkerContextEventListener::handleEvent):
2978
2979 2009-03-13  Dimitri Glazkov  <dglazkov@chromium.org>
2980
2981         Reviewed by Timothy Hatcher.
2982
2983         https://bugs.webkit.org/show_bug.cgi?id=24524
2984         Introduce ScriptObject and ScriptFunctionCall abstractions.
2985
2986         * GNUmakefile.am: Added ScriptObject and ScriptFunctionCall to project.
2987         * WebCore.pro: Ditto.
2988         * WebCore.vcproj/WebCore.vcproj: Ditto.
2989         * WebCore.xcodeproj/project.pbxproj: Ditto.
2990         * WebCoreSources.bkl: Ditto.
2991         * bindings/js/ScriptFunctionCall.cpp: Added.
2992         * bindings/js/ScriptFunctionCall.h: Added.
2993         * bindings/js/ScriptObject.cpp: Added.
2994         * bindings/js/ScriptObject.h: Added.
2995         * bindings/js/ScriptObjectQuarantine.cpp: Added.
2996         * bindings/js/ScriptObjectQuarantine.h: Added.
2997         * inspector/InspectorController.cpp:
2998         (WebCore::InspectorController::populateScriptObjects): Changed to use InspectorDatabaseResource::bind.
2999         (WebCore::InspectorController::resetScriptObjects): Changed to use InspectorDatabaseResource::unbind.
3000         (WebCore::InspectorController::didOpenDatabase): Changed to use InspectorDatabaseResource::unbind.
3001         * inspector/InspectorController.h: Removed addScriptDatabaseResource and
3002         removeScriptDatabaseResource declarations.
3003         * inspector/InspectorDatabaseResource.cpp:
3004         (WebCore::InspectorDatabaseResource::InspectorDatabaseResource):
3005         (WebCore::InspectorDatabaseResource::bind): Added.
3006         (WebCore::InspectorDatabaseResource::unbind): Added.
3007         * inspector/InspectorDatabaseResource.h: Added bind and unbind declarations.
3008
3009 2009-03-13  Peter Kasting  <pkasting@google.com>
3010
3011         Reviewed by Eric Seidel.
3012
3013         https://bugs.webkit.org/show_bug.cgi?id=24467
3014         Make Skia drawRect() and stroke behavior match CG.
3015
3016         This makes drawRect() ignore the stroke width (like CG does), and
3017         adds a warning comment about that to the appropriate header.
3018
3019         It also eliminates some hacky code in Skia's stroke preparation,
3020         which tried to adjust odd-width strokes to fall on pixel boundaries.
3021         Not only did this not match CG, it wouldn't necessarily work right,
3022         because there could be other transforms (e.g. full-page zoom) that
3023         would affect the stroke before it reached the device pixel level.
3024
3025         * platform/graphics/GraphicsContext.h:
3026         * platform/graphics/skia/GraphicsContextSkia.cpp:
3027         (WebCore::GraphicsContext::drawRect):
3028         (WebCore::GraphicsContext::fillRect):
3029         * platform/graphics/skia/PlatformContextSkia.cpp:
3030         (PlatformContextSkia::drawRect):
3031         (PlatformContextSkia::setupPaintForStroking):
3032
3033 2009-03-13  Jian Li  <jianli@chromium.org>
3034
3035         Reviewed by Dimitri Glazkov.
3036
3037         https://bugs.webkit.org/show_bug.cgi?id=24583
3038         Need to change scope of constructor and destructor of V8ObjectEventListener from private
3039         to protected so that they could be used in its derived class V8WorkerContextEventListener.
3040
3041         * bindings/v8/V8ObjectEventListener.h:
3042
3043 2009-03-13  David Levin  <levin@chromium.org>
3044
3045         Reviewed by Dimitri Glazkov.
3046
3047         https://bugs.webkit.org/show_bug.cgi?id=24579
3048         Many V8*Event* files are misplaced.
3049
3050         This code is infrastructure for supporting the bindings and thus should
3051         be in the bindings/v8 directory.
3052
3053         * bindings/v8/V8AbstractEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8AbstractEventListener.cpp.
3054         * bindings/v8/V8AbstractEventListener.h: Renamed from WebCore/bindings/v8/custom/V8AbstractEventListener.h.
3055         * bindings/v8/V8LazyEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8LazyEventListener.cpp.
3056         * bindings/v8/V8LazyEventListener.h: Renamed from WebCore/bindings/v8/custom/V8LazyEventListener.h.
3057         * bindings/v8/V8ObjectEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8ObjectEventListener.cpp.
3058         * bindings/v8/V8ObjectEventListener.h: Renamed from WebCore/bindings/v8/custom/V8ObjectEventListener.h.
3059         * bindings/v8/V8WorkerContextEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8WorkerContextEventListener.cpp.
3060         * bindings/v8/V8WorkerContextEventListener.h: Renamed from WebCore/bindings/v8/custom/V8WorkerContextEventListener.h.
3061
3062 2009-03-13  Adele Peterson  <adele@apple.com>
3063
3064         Reviewed by Justin Garcia.
3065
3066         Fix for <rdar://problem/5089327> Color of quoted content is wrong when pasted inside other quoted content
3067
3068         Test: editing/pasteboard/5089327.html
3069
3070         Handle spans being pasted within a quoted region in the same way we handle "paste as quotation" content.
3071
3072         * editing/ReplaceSelectionCommand.cpp:
3073         (WebCore::handleStyleSpansBeforeInsertion):
3074         (WebCore::ReplaceSelectionCommand::handleStyleSpans):
3075
3076 2009-03-13  Greg Bolsinga  <bolsinga@apple.com>
3077
3078         Reviewed by Simon Fraser.
3079
3080         Update Geolocation perimission dialogs to be asynchronous.
3081         https://bugs.webkit.org/show_bug.cgi?id=24505
3082
3083         Geolocation now requests permission from the Chrome asynchronously.
3084         The Chrome is passed the Geolocation object, and the Chrome sets the permission
3085         on the Geolocation. Geolocation also tracks if the Chrome should clear its cache
3086         of SecurityOrigins with geolocation permission. This is so that the GeolocationService
3087         can inform the Chrome of its request, and the Chrome is also free to implement its
3088         own policy.
3089
3090         * WebCore.base.exp:
3091         * WebCore.xcodeproj/project.pbxproj:
3092         * page/Chrome.cpp:
3093         (WebCore::Chrome::requestGeolocationPermissionForFrame):
3094         * page/Chrome.h:
3095         * page/ChromeClient.h:
3096         (WebCore::ChromeClient::requestGeolocationPermissionForFrame):
3097         * page/Geolocation.cpp:
3098         (WebCore::Geolocation::Geolocation):
3099         (WebCore::Geolocation::getCurrentPosition):
3100         (WebCore::Geolocation::watchPosition):
3101         (WebCore::Geolocation::setIsAllowed):
3102         (WebCore::Geolocation::displayChallengeIfNecessary):
3103         (WebCore::Geolocation::geolocationServicePositionChanged):
3104         * page/Geolocation.h:
3105         (WebCore::Geolocation::isAllowed):
3106         (WebCore::Geolocation::setShouldClearCache):
3107         (WebCore::Geolocation::shouldClearCache):
3108         (WebCore::Geolocation::):
3109
3110 2009-03-13  Anders Carlsson  <andersca@apple.com>
3111
3112         Reviewed by Dan Bernstein.
3113
3114         <rdar://problem/6610666> Revise the Cocoa event model text API
3115
3116         Add a NPCocoaEventTextInput event type. Remove the text input variables.
3117
3118         * bridge/npapi.h:
3119         (_NPCocoaEvent::):
3120
3121 2009-03-13  Kevin Decker  <kdecker@apple.com>
3122
3123         Reviewed by Anders.
3124
3125         <rdar://problem/6630340> REGRESSION (39114-39115): Unity Web Player no longer works if Flip4Mac is also installed
3126
3127         The code assumed if we have a plug-in that supports "application/x-oleobject" we should always prefer the object tag
3128         over of an embed tag. That assumption can cause the Mac platform to load the wrong plug-in, as Flip4Mac claims supports
3129         for x-oleobject.
3130
3131         * rendering/RenderPartObject.cpp:
3132         (WebCore::shouldUseEmbedDescendant): Made the Mac platform always return true here.
3133
3134 2009-03-13  Jian Li  <jianli@chromium.org>
3135
3136         Reviewed by Dimitri Glazkov.
3137
3138         https://bugs.webkit.org/show_bug.cgi?id=24559
3139         Need to port JSC fix (r41565) to V8 in order to fix layout test onload-single-line-comment.html.
3140
3141         * bindings/v8/custom/V8LazyEventListener.cpp:
3142         (WebCore::V8LazyEventListener::getListenerFunction):
3143         (WebCore::V8LazyEventListener::getWrappedListenerFunction):
3144
3145 2009-03-13  Jian Li  <jianli@chromium.org>
3146
3147         Reviewed by Dimitri Glazkov.
3148
3149         https://bugs.webkit.org/show_bug.cgi?id=24557
3150         This is to support running multiple workers in a single worker process in chromium.
3151
3152         * bindings/v8/custom/V8WorkerContextEventListener.cpp:
3153         (WebCore::V8WorkerContextEventListener::handleEvent):
3154         Add locker to v8 event listener of worker context.
3155
3156 2009-03-13  Alexey Proskuryakov  <ap@webkit.org>
3157
3158         Reviewed by Sam Weinig.
3159
3160         https://bugs.webkit.org/show_bug.cgi?id=24575
3161         XHR response is incorrect after a network error.
3162
3163         Tests: http/tests/xmlhttprequest/cross-site-denied-response-sync-2.html
3164                http/tests/xmlhttprequest/cross-site-denied-response-sync.html
3165                http/tests/xmlhttprequest/cross-site-denied-response.html
3166
3167         * xml/XMLHttpRequest.cpp:
3168         (WebCore::XMLHttpRequest::networkError): Call internalAbort() to fully reset the request.
3169         (WebCore::XMLHttpRequest::didFailRedirectCheck): No need to call internalAbort() here. Note
3170         that since internalAbort() can drop GC protection, it is not safe to use the object after
3171         this call.
3172         (WebCore::XMLHttpRequest::didReceiveData): Check that the request wasn't aborted, and
3173         return early if it was. This can happen during sync requests, as the loader does not know
3174         that it was aborted, and just synthesizes all callbacks.
3175
3176 2009-03-13  Adam Bergkvist <adam.bergkvist@ericsson.com>
3177
3178         Reviewed by Alexey Proskuryakov.
3179
3180         https://bugs.webkit.org/show_bug.cgi?id=3547
3181         XMLHttpRequest.statusText returns always "OK"
3182
3183         Covered by existing tests (which now pass on all platforms but Mac).
3184
3185         * xml/XMLHttpRequest.cpp:
3186         (WebCore::XMLHttpRequest::statusText): Return ResourceResponse status text. It is now up
3187         to each platform to correctly set the status text or set it to "OK" to retain current
3188         behavior.
3189
3190 2009-03-13  Adam Bergkvist <adam.bergkvist@ericsson.com>
3191
3192         Reviewed by Alexey Proskuryakov.
3193
3194         https://bugs.webkit.org/show_bug.cgi?id=24349
3195         [QT] HTTP status text is never set
3196
3197         Set HTTP status text to the reason phrase attribute of QNetworkReply.
3198
3199         * platform/network/qt/QNetworkReplyHandler.cpp:
3200         (WebCore::QNetworkReplyHandler::sendResponseIfNeeded):
3201
3202 2009-03-12  Simon Fraser  <simon.fraser@apple.com>
3203
3204         Reviewed by Mark Rowe
3205
3206         <rdar://problem/6622300>: Reproducible crash on
3207         <http://www.editgrid.com/explore/tnc/dave/FusionChart%3A_Candlestick>
3208
3209         Prevent CSSStyleSheet::checkLoaded() writing to freed memory when it gets
3210         deleted from under itself. The sheetLoaded() notification can allow scripts
3211         to run via HTMLTokenizer::executeScriptsWaitingForStylesheets(),
3212         which can cause the last ref to the CSSStyleSheet to be released.
3213
3214         * css/CSSStyleSheet.cpp:
3215         (WebCore::CSSStyleSheet::checkLoaded):
3216
3217 2009-03-12  Kevin Ollivier  <kevino@theolliviers.com>
3218
3219         wx build fix. Allow make-generated-sources.sh to pass args down to DerivedSources.make
3220
3221         * make-generated-sources.sh:
3222
3223 2009-03-12  Adam Treat  <adam.treat@torchmobile.com>
3224
3225         Reviewed by Oliver Hunt.
3226
3227         https://bugs.webkit.org/show_bug.cgi?id=24498
3228         Fix the Qt port to use the same algorithm for drawing dashed and dotted
3229         borders as the other ports.  This makes the Qt port pixel-for-pixel perfect
3230         compared to border drawing with Apple's canonical mac port and much closer
3231         to konqueror and firefox behavior.
3232
3233         * platform/graphics/qt/GraphicsContextQt.cpp:
3234         (WebCore::GraphicsContext::drawLine):
3235
3236 2009-02-26  Eric Seidel  <eric@webkit.org>
3237
3238         Reviewed by Justin Garcia.
3239
3240         Clean up DOMSelection to use some helper functions
3241         making the code smaller, and less error-prone.
3242         https://bugs.webkit.org/show_bug.cgi?id=19221
3243
3244         I tried moving DOMSelection off of rangeCompliantEquivalent
3245         but failed.  VisibleSelection holds positions like (table, 1) to mean
3246         "after the table".
3247
3248         * page/DOMSelection.cpp:
3249         (WebCore::DOMSelection::visibleSelection):
3250         (WebCore::anchorPosition):
3251         (WebCore::focusPosition):
3252         (WebCore::basePosition):
3253         (WebCore::extentPosition):
3254         (WebCore::DOMSelection::anchorNode):
3255         (WebCore::DOMSelection::anchorOffset):
3256         (WebCore::DOMSelection::focusNode):
3257         (WebCore::DOMSelection::focusOffset):
3258         (WebCore::DOMSelection::baseNode):
3259         (WebCore::DOMSelection::baseOffset):
3260         (WebCore::DOMSelection::extentNode):
3261         (WebCore::DOMSelection::extentOffset):
3262         * page/DOMSelection.h:
3263
3264 2009-03-12  Dmitry Titov  <dimich@chromium.org>
3265
3266         Reviewed by Dimitri Glazkov.
3267
3268         https://bugs.webkit.org/show_bug.cgi?id=24563
3269         Change ResourceRequest to be a struct rather then a class to match other declarations.
3270
3271         * platform/network/chromium/ResourceRequest.h:
3272
3273 2009-03-12  David Levin  <levin@chromium.org>
3274
3275         Reviewed by Darin Fisher.
3276
3277         Bug 24530: width100percent-searchfield.html should be fixed for chromium.
3278         <https://bugs.webkit.org/show_bug.cgi?id=24530>
3279
3280         Compensate for r39924 in chromium, which broke LayoutTests/fast/replaced/width100percent-searchfield.html
3281
3282         * css/themeChromiumWin.css: Added.
3283         Overrides the changes to start supporting input[type="search"] on Windows (since
3284         Chromium doesn't do anything special for this).
3285
3286         * rendering/RenderThemeChromiumWin.cpp:
3287         (WebCore::supportsFocus):
3288         (WebCore::RenderThemeChromiumWin::extraDefaultStyleSheet):
3289         (WebCore::RenderThemeChromiumWin::determineState):
3290         (WebCore::RenderThemeChromiumWin::getThemeData):
3291
3292 2009-03-12  Greg Bolsinga  <bolsinga@apple.com>
3293
3294         Reviewed by Antti Koivisto.
3295
3296         https://bugs.webkit.org/show_bug.cgi?id=24506
3297         Geolocation in Safari differs from the spec, no Coordinates attribute on Position
3298
3299         Get up to date.
3300
3301         * DerivedSources.make:
3302         * GNUmakefile.am:
3303         * WebCore.pro:
3304         * WebCore.scons:
3305         * WebCore.vcproj/WebCore.vcproj:
3306         * WebCore.xcodeproj/project.pbxproj:
3307         * WebCoreSources.bkl:
3308         * bindings/js/JSGeolocationCustom.cpp:
3309         (WebCore::createPositionOptions):
3310         * page/Coordinates.cpp: Added.
3311         (WebCore::Coordinates::toString):
3312         * page/Coordinates.h: Added.
3313         (WebCore::Coordinates::create):
3314         (WebCore::Coordinates::latitude):
3315         (WebCore::Coordinates::longitude):
3316         (WebCore::Coordinates::altitude):
3317         (WebCore::Coordinates::accuracy):
3318         (WebCore::Coordinates::altitudeAccuracy):
3319         (WebCore::Coordinates::heading):
3320         (WebCore::Coordinates::speed):
3321         (WebCore::Coordinates::Coordinates):
3322         * page/Coordinates.idl: Added.
3323         * page/Geoposition.cpp:
3324         (WebCore::Geoposition::toString):
3325         * page/Geoposition.h:
3326         (WebCore::Geoposition::create):
3327         (WebCore::Geoposition::coords):
3328         (WebCore::Geoposition::Geoposition):
3329         * page/Geoposition.idl:
3330         * page/PositionOptions.h:
3331         (WebCore::PositionOptions::create):
3332         (WebCore::PositionOptions::maximumAge):
3333         (WebCore::PositionOptions::setMaximumAge):
3334         (WebCore::PositionOptions::PositionOptions):
3335         * platform/mac/GeolocationServiceMac.mm:
3336         (-[WebCoreCoreLocationObserver locationManager:didUpdateToLocation:fromLocation:]):
3337
3338 2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>
3339
3340         Reviewed by Darin Fisher.
3341
3342         https://bugs.webkit.org/show_bug.cgi?id=24496
3343         Fix console logging of non-string values by coercing the argument to
3344         a string, for V8 bindings.
3345
3346         Test: fast/js/console-non-string-values.html
3347
3348         * bindings/v8/ScriptValue.cpp: Added toString method.
3349         (WebCore::ScriptValue::toString):
3350         * bindings/v8/ScriptValue.h: Added PlatformString, ScriptState includes
3351             and toString declaration.
3352
3353 2009-03-12  Darin Fisher  <darin@chromium.org>
3354
3355         Reviewed by Dimitri Glazkov.
3356
3357         Eliminate ChromiumBridge::uiResourceProtocol
3358         https://bugs.webkit.org/show_bug.cgi?id=24558
3359
3360         * platform/chromium/ChromiumBridge.h:
3361
3362 2009-03-12  Greg Bolsinga  <bolsinga@apple.com>
3363
3364         Reviewed by Simon Fraser.
3365
3366         Missing breaks in switch statement
3367         https://bugs.webkit.org/show_bug.cgi?id=24556
3368
3369         * platform/mac/GeolocationServiceMac.mm:
3370         (-[WebCoreCoreLocationObserver locationManager:didFailWithError:]):
3371
3372 2009-03-12  Darin Fisher  <darin@chromium.org>
3373
3374         Reviewed by Dimitri Glazkov.
3375
3376         https://bugs.webkit.org/show_bug.cgi?id=24554
3377         Remove some unused functions from ChromiumBridge.
3378
3379         * platform/chromium/ChromiumBridge.h:
3380
3381 2009-03-02  Eric Seidel  <eric@webkit.org>
3382
3383         Reviewed by Darin Fisher and Justin Garcia.
3384
3385         Safari crashes during drag and drop in Google presentations
3386         due to mutation event handlers removing DOM content during insertNode
3387         https://bugs.webkit.org/show_bug.cgi?id=22634
3388
3389         Added a bunch of "null" checks to make sure nodes are still
3390         in the document before we operate on them.  This is an
3391         inelegant solution, but it's the best we have for now.
3392
3393         Test: editing/selection/crash-on-drag-with-mutation-events.html
3394
3395         * editing/CompositeEditCommand.cpp:
3396         (WebCore::CompositeEditCommand::insertNodeAt):
3397         * editing/ReplaceSelectionCommand.cpp:
3398         (WebCore::ReplaceSelectionCommand::doApply):
3399
3400 2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>
3401
3402         Reviewed by Adam Treat.
3403
3404         https://bugs.webkit.org/show_bug.cgi?id=24525
3405         REGRESSION: Inspector window doesn't close when inspected page is
3406         destroyed. This is a revert of r41158, which became unnecessary when
3407         InspectorController became ref-counted in r41462.
3408
3409         * inspector/InspectorController.cpp:
3410         (WebCore::InspectorController::inspectedPageDestroyed): Reset m_inspectedPage
3411             after calling close().
3412         (WebCore::InspectorController::stopUserInitiatedProfiling): Remove
3413             m_inspectedPage check guard around profile logic.
3414
3415 2009-03-12  Peter Kasting  <pkasting@google.com>
3416
3417         Reviewed by Darin Fisher.
3418
3419         https://bugs.webkit.org/show_bug.cgi?id=24502
3420         Make horizontal scrolling on Windows always go the correct direction.
3421
3422         * platform/PlatformWheelEvent.h:
3423         * platform/win/WheelEventWin.cpp:
3424         (WebCore::PlatformWheelEvent::PlatformWheelEvent):
3425
3426 2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>
3427
3428         Reviewed by Simon Fraser.
3429
3430         https://bugs.webkit.org/show_bug.cgi?id=24496
3431         Fix console logging of non-string values by coercing the argument to
3432         a string.
3433
3434         Test: fast/js/console-non-string-values.html
3435
3436         * bindings/js/ScriptValue.cpp: Removed PlatformString include.
3437         * bindings/js/ScriptValue.h: Added toString method.
3438         (WebCore::ScriptValue::toString):
3439         * page/Console.cpp:
3440         (WebCore::getFirstArgumentAsString): Changed firstArgumentAsString method to use
3441             ScriptValue::toString.
3442         (WebCore::Console::addMessage): added extra ScriptState argument to callsite.
3443         (WebCore::Console::count): Ditto.
3444
3445 2009-03-12  Simon Fraser  <simon.fraser@apple.com>
3446
3447         Reviewed by Adele Peterson.
3448
3449         https://bugs.webkit.org/show_bug.cgi?id=24508
3450
3451         Fix updating of text field with placeholder text when value is set
3452         by ensuring that updatePlaceholderVisibility() is called after
3453         the value has been updated, not before.
3454
3455         Test: fast/forms/placeholder-set-value.html
3456
3457         * html/HTMLInputElement.cpp:
3458         (WebCore::HTMLInputElement::setValue):
3459
3460 2009-03-12  Simon Fraser  <simon.fraser@apple.com>
3461
3462         Reviewed by Dave Hyatt
3463
3464         https://bugs.webkit.org/show_bug.cgi?id=24552
3465
3466         Hit testing involving normal flow, self-painting layers (e.g. reflections)
3467         was broken because it was not null-testing the result of hitTestLayer(),
3468         so bailing early with a nil hit layer.
3469
3470         Test: fast/layers/normal-flow-hit-test.html
3471
3472         * rendering/RenderLayer.cpp:
3473         (WebCore::RenderLayer::hitTestLayer):
3474
3475 2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>
3476
3477         Reviewed by Cameron Zwarich.
3478
3479         Removed zero-sized files, left over from improperly applied patch.
3480
3481         * platform/graphics/chromium/ThemeHelperChromiumWin.cpp: Removed.
3482         * platform/graphics/chromium/ThemeHelperChromiumWin.h: Removed.
3483
3484 2009-03-12  Adam Treat  <adam.treat@torchmobile.com>
3485
3486         Reviewed by George Staikos.
3487
3488         Do not ignore alpha color in optimization.  Hard to believe, but there
3489         are websites that fill the entire page with a tiled image consisting of
3490         nothing but a 100% transparent 1x1 image. <cough>orbitz.com</cough>
3491
3492         * platform/graphics/qt/ImageQt.cpp:
3493         (WebCore::BitmapImage::checkForSolidColor):
3494
3495 2009-03-12  Julien Chaffraix  <jchaffraix@webkit.org>
3496
3497         Reviewed by Darin Adler.
3498
3499         Bug 24110: cloneNode should call cloneElement and not the reverse
3500
3501         - Splitted the code from cloneNode into cloneElementWithChildren and cloneElementWithChildren.
3502           Now cloneNode calls one of the 2 previous methods.
3503
3504         - Renamed cloneElement to cloneElementWithoutChildren as it was the previous behaviour.
3505
3506         - Moved cloneNode to the Element private section so that WebCore callers cannot use it.
3507
3508         - Removed Element::cloneNode usage through WebCore.
3509
3510         * dom/Element.cpp:
3511         (WebCore::Element::cloneNode): Moved to Element's private section and it
3512         now calls the two next methods.
3513         (WebCore::Element::cloneElementWithChildren): Added.
3514         (WebCore::Element::cloneElementWithoutChildren): Renamed from cloneElement
3515         to avoid ambiguity.
3516         * dom/Element.h:
3517
3518         * editing/ApplyStyleCommand.cpp:
3519         (WebCore::ApplyStyleCommand::addInlineStyleIfNeeded): Changed call to cloneElement
3520         to call to cloneElementWithoutChildren.
3521         * editing/BreakBlockquoteCommand.cpp:
3522         (WebCore::BreakBlockquoteCommand::doApply): Ditto.
3523         * editing/IndentOutdentCommand.cpp:
3524         (WebCore::IndentOutdentCommand::indentRegion): Ditto.
3525         * editing/InsertParagraphSeparatorCommand.cpp:
3526         (WebCore::InsertParagraphSeparatorCommand::doApply): Ditto.
3527         * editing/ModifySelectionListLevel.cpp:
3528         (WebCore::IncreaseSelectionListLevelCommand::doApply): Ditto.
3529         * editing/SplitElementCommand.cpp:
3530         (WebCore::SplitElementCommand::doApply): Ditto.
3531         * editing/markup.cpp:
3532         (WebCore::createFragmentFromText): Ditto.
3533         * svg/SVGUseElement.cpp:
3534         (WebCore::SVGUseElement::buildShadowTree): Ditto.
3535         (WebCore::SVGUseElement::expandUseElementsInShadowTree): Ditto.
3536
3537 2009-03-12  Dirk Schulze  <krit@webkit.org>
3538
3539         Reviewed by Oliver Hunt.
3540
3541         Fixed a crash on the 1x1 pixel check for background image in gtk.
3542         We need to load the image first and need to be sure that the image
3543         is not null before checking the type.
3544
3545         * platform/graphics/cairo/ImageCairo.cpp:
3546         (WebCore::BitmapImage::checkForSolidColor):
3547
3548 2009-03-12  Xan Lopez  <xlopez@igalia.com>
3549
3550         Reviewed by Alexey Proskuryakov.
3551