Timeline
Sep 23, 2004:
- 7:20 PM Changeset in webkit [7664] by
-
- 2 edits in trunk/WebCore
Reviewed by Maciej.
- fixed <rdar://problem/3551850> hang caused by interpreting bad javascript guarded by a deliberately bogus "language" attribute (www.riibe.com)
- khtml/html/htmltokenizer.cpp: (khtml::HTMLTokenizer::parseTag): Check for language attribute of <script> tag in a way that matches WinIE. Previously we were far too permissive.
- 5:27 PM Changeset in webkit [7663] by
-
- 2 edits in trunk/WebCore
Fix for 3685234 and 3548444, the x-offset for frame borders was off by 1 pixel, causing mojibake to occur when
repainting happened.
- khtml/rendering/render_canvas.cpp: (RenderCanvas::repaintViewRectangle):
- 5:17 PM Changeset in webkit [7662] by
-
- 12 edits in trunk/WebCore
Fixed <rdar://problem/3813271> dashboard-region-circle and dashboard-region-rectangle should be collapsed into dashboard-region
Fixed <rdar://problem/3813289> dashboard regions need to correctly account for overflow/scrolling
Reviewed by Chris.
- khtml/css/css_valueimpl.cpp: (CSSPrimitiveValueImpl::cssText):
- khtml/css/css_valueimpl.h:
- khtml/css/cssparser.cpp: (CSSParser::parseValue): (skipCommaInDashboardRegion): (CSSParser::parseDashboardRegions):
- khtml/khtmlview.cpp: (KHTMLView::layout): (KHTMLView::updateDashboardRegions):
- khtml/khtmlview.h:
- khtml/rendering/render_layer.cpp: (RenderLayer::scrollToOffset):
- khtml/rendering/render_object.cpp: (RenderObject::addDashboardRegions):
- khtml/rendering/render_object.h:
- kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::dashboardRegionsChanged):
- kwq/WebDashboardRegion.h:
- kwq/WebDashboardRegion.m: (-[WebDashboardRegion initWithRect:clip:type:]): (-[WebDashboardRegion dashboardRegionClip]): (-[WebDashboardRegion description]):
- 5:05 PM Changeset in webkit [7661] by
-
- 2 edits in trunk/WebCore
Reviewed by Richard
- khtml/xml/dom_position.cpp: (DOM::Position::inRenderedContent): Make the "empty block" check the same as the one use in visible position. This fixes a recent regression which broke up and down arrowing between blocks with an empty block in between.
- 2:25 PM Changeset in webkit [7660] by
-
- 4 edits in trunk
WebCore:
Reviewed by Darin.
<rdar://problem/3685235> REGRESSION (Mail): links are not properly editable
- khtml/html/html_inlineimpl.cpp: (HTMLAnchorElementImpl::isFocusable): If this element is editable, then follow the normal focus rules so the link does not swallow focus when you arrow key or drag-select into it.
WebKit:
Reviewed by Darin.
<rdar://problem/3685235> REGRESSION (Mail): links are not properly editable
- WebView.subproj/WebDefaultUIDelegate.m: By default, don't allow link dragging if the element under the mouse pointer is editable. This way, you can drag-select starting inside a link.
- 10:56 AM Changeset in webkit [7659] by
-
- 3 edits in trunk/WebKit
WebKit:
Reviewed by Chris.
- WebKit part of fix for <rdar://problem/3415264> Default encoding should initially be set to current system encoding
- WebView.subproj/WebPreferences.m: (-[WebPreferences _setInitialDefaultTextEncodingToSystemEncoding]): new SPI that sets the initial value of the default text encoding to be the system encoding, with a special-case conversion of MacRoman->Latin1. This is not done automatically for WebKit clients for fear of breaking them.
- WebView.subproj/WebPreferencesPrivate.h: declare new SPI
WebBrowser:
Reviewed by Chris.
- WebBrowser part of fix for <rdar://problem/3415264> Default encoding should initially be set to current system encoding
- AppController.m: (-[AppController applicationDidFinishLaunching:]): call new SPI -[WebPreferences _setInitialDefaultTextEncodingToSystemEncoding]
- 10:47 AM Changeset in webkit [7658] by
-
- 1 edit in trunk/WebCore/khtml/rendering/render_block.cpp
* empty log message *
- 10:29 AM Changeset in webkit [7657] by
-
- 1 edit2 adds in trunk
- added test for the DOM::Range bug fixed recently
- layout-tests/fast/dom/clone-contents-0-end-offset-expected.txt: Added.
- layout-tests/fast/dom/clone-contents-0-end-offset.html: Added.
- 10:22 AM Changeset in webkit [7656] by
-
- 4 edits in trunk/WebCore
Fix for hitlist bug, crash when deleting.
Also fix repaint bug when inlines are inside overflow:auto/scroll regions.
Reviewed by kocienda
- khtml/editing/htmlediting_impl.cpp: (khtml::DeleteSelectionCommandImpl::DeleteSelectionCommandImpl): (khtml::DeleteSelectionCommandImpl::doApply):
- khtml/rendering/render_block.cpp: (khtml::RenderBlock::removeChild):
- khtml/rendering/render_flow.cpp: (RenderFlow::getAbsoluteRepaintRect):
- 9:52 AM Changeset in webkit [7655] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3527840> reproducible crash at johnbrown.v32.qikker.com (nil-deref in NodeImpl::dispatchEvent)
- khtml/xml/dom_nodeimpl.cpp: (NodeImpl::dispatchEvent): guard against document or document->document() being nil
- 9:10 AM Changeset in webkit [7654] by
-
- 6 edits2 adds in trunk/WebCore
Reviewed by Ken.
- added new VisibleRange class; not used yet
- WebCore.pbproj/project.pbxproj: Added visible_range.h and visible_range.cpp.
- khtml/editing/visible_range.cpp: Added.
- khtml/editing/visible_range.h: Added.
- tweaks
- kwq/WebCoreBridge.mm: (-[WebCoreBridge alterCurrentSelection:direction:granularity:]): Use switch statement instead of if statement so we get a warning if we ever add a new granularity.
- khtml/editing/selection.h: khtml, not DOM, namespace
- khtml/editing/visible_position.h: Ditto.
- 9:08 AM Changeset in webkit [7653] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3812758> 8S266: attaching mutation events to an editable div crashes Safari
- khtml/xml/dom_nodeimpl.cpp: (NodeImpl::dispatchEvent): Added ref/deref so this works if passed a "floating" event object. (NodeImpl::dispatchGenericEvent): Ditto. (NodeImpl::dispatchHTMLEvent): Removed unneeded ref/deref now that dispatchEvent takes care of it. (NodeImpl::dispatchMouseEvent): Ditto. (NodeImpl::dispatchUIEvent): Ditto.
- 9:05 AM Changeset in webkit [7652] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed some minor mistakes discovered by code inspection
- khtml/khtml_part.cpp: (KHTMLPart::setSelection): Changed to not grab focus if clearing the selection. Grabbing focus when we get the selection is also probably something that should be conditional, since you can use a WebView in a mode where it can have selection even when not first responder, ala NSTextView. (KHTMLPart::setCaretVisible): Changed to not grab focus if making the caret invisible. Not sure if it's ever important to grab focus when making the caret visible, but by code inspection it seemed that was unlikely to do harm.
- 9:03 AM Changeset in webkit [7651] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3790595> "ERROR: unimplemented propertyID: 97" logged to console when copying selection (table-layout)
- khtml/css/css_computedstyle.cpp: (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): Added CSS_PROP_TABLE_LAYOUT case.
- 9:02 AM Changeset in webkit [7650] by
-
- 3 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3809600> REGRESSION: text in search field doesn't disappear when clicked at developer.apple.com
Reversing the order of scope caused us to get and set too many properties in the window
object; in the case of this bug setting value ended up setting a window.value property
instead of the value of the <input> element.
- khtml/ecma/kjs_window.h: Removed bogus hasProperty function that always returns true. Testing shows that MacIE and Gecko match our behavior when we remove this, despite the "need this to match IE behavior" comment in the file, which I believe is incorrect.
- khtml/ecma/kjs_window.cpp: Ditto.
- 8:58 AM Changeset in webkit [7649] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3812471> -[DOMRange cloneContents] clones too much in some cases where endOffset is 0
- khtml/xml/dom2_rangeimpl.cpp: (DOM::RangeImpl::processContents): Added some nil checks to prevent this function from crashing if offsets are greater than the number of child nodes. Added a special case for offset 0 in one case that has a loop that won't work correctly for that case.
- 8:50 AM Changeset in webkit [7648] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3811890> when selecting and moving the caret, some words aren't spell-checked (test case included)
This was mostly fixed by changes I made recently, but using the test case in the bug
I discovered one regression I introduced and another problem that wasn't fixed yet.
This change fixes both.
- kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::markMisspellingsInSelection): Removed unneeded inRenderedContent() check. It was returning false in cases involving a <br> at the end of line, and the check isn't all that helpful now that we use VisiblePosition, which takes care of that sort of thing for you. Changed code to determine the range to check to use the two different flavors of word boundary to expand to words we touch in both directions. (KWQKHTMLPart::updateSpellChecking): Changed to use the same logic about word boundaries. By being consistent, the anomalies reported in the bug report go away.
- 8:22 AM Changeset in webkit [7647] by
-
- 3 edits in trunk/WebKit
Reviewed by Ken.
- fixed <rdar://problem/3811584> REGRESSION (85-125): iframe.document undefined in function called from button onclick; works from img onclick
The fix is to not let "defers callbacks" have any effect on loading "about:blank".
I also had to fix one bug in WebCore that could then be reproduced by going to "about:blank"
while using a button.
- WebView.subproj/WebMainResourceClient.m: (-[WebMainResourceClient connection:didReceiveResponse:]): Loosen asserts to allow this callback for the specific case of "about:blank" even if the defers callbacks flag is true. (-[WebMainResourceClient connectionDidFinishLoading:]): Ditto. (-[WebMainResourceClient loadWithRequestNow:]): Added NSURLRequest return value. Loosened asserts as above. Changed code to return a new request if we get a new request back that is not empty when the defers callbacks flag is true. (-[WebMainResourceClient loadWithRequest:]): If the defers callbacks flag is set, but the URL is one that gives us an empty document, then do the work right away, don't defer it.
- 8:21 AM Changeset in webkit [7646] by
-
- 2 edits in trunk/WebCore
Reviewed by Ken.
- fixed <rdar://problem/3811584> REGRESSION (85-125): iframe.document undefined in function called from button onclick; works from img onclick
Since we are doing a bit less "defers callbacks" work, when testing I ran into this code path.
- khtml/xml/dom2_eventsimpl.cpp: (MouseEventImpl::computeLayerPos): Check for document without renderer.
- 12:10 AM Changeset in webkit [7645] by
-
- 1 edit1 add in trunk/WebCore
- checked in a new file I forgot
- khtml/editing/text_granularity.h: Added.
- 12:09 AM Changeset in webkit [7644] by
-
- 1 edit1 add in trunk/WebKit
- fixed B&I builds by checking in generated file
- WebCoreSupport.subproj/WebDashboardRegion.h: Added.
- 12:05 AM Changeset in webkit [7643] by
-
- 29 edits1 add in trunk/WebCore
- renamed
DOM::CaretPosition -> khtml::VisibleRange
DOM::Selection -> khtml::Selection
- moved all the functions in visible_units.h from DOM to khtml namespace
- moved the one thing from Selection that KHTMLPart uses into its own header
- WebCore.pbproj/project.pbxproj: Added text_granularity.h.
- ForwardingHeaders/editing/selection.h: Added.
- WebCore-combined.exp: Regenerated.
- WebCore-tests.exp: Updated symbol for DOM::Selection debugging.
- khtml/ecma/kjs_window.cpp: Update names and namespaces.
- khtml/editing/htmlediting.cpp: Ditto.
- khtml/editing/htmlediting.h: Ditto.
- khtml/editing/htmlediting_impl.cpp: Ditto.
- khtml/editing/htmlediting_impl.h: Ditto.
- khtml/editing/jsediting.cpp: Ditto.
- khtml/editing/selection.cpp: Ditto.
- khtml/editing/selection.h: Ditto.
- khtml/editing/visible_position.cpp: Ditto.
- khtml/editing/visible_position.h: Ditto.
- khtml/editing/visible_text.h: Ditto.
- khtml/editing/visible_units.cpp: Ditto.
- khtml/editing/visible_units.h: Ditto.
- khtml/khtml_part.cpp: Ditto.
- khtml/khtml_part.h: Ditto.
- khtml/khtmlpart_p.h: Ditto.
- khtml/rendering/render_block.cpp: Ditto.
- khtml/xml/dom_docimpl.h: Ditto.
- khtml/xml/dom_nodeimpl.cpp: Ditto.
- kwq/KWQKHTMLPart.h: Ditto.
- kwq/KWQKHTMLPart.mm: Ditto.
- kwq/KWQRenderTreeDebug.cpp: Ditto.
- kwq/WebCoreBridge.mm: Ditto.
Sep 22, 2004:
- 11:21 PM Changeset in webkit [7642] by
-
- 217 edits in trunk/WebCore
- ForwardingHeaders/*: Use import instead of include.
- 10:23 PM Changeset in webkit [7641] by
-
- 34 edits4 adds9 deletes in trunk/WebCore
- renamed these files:
misc/khtml_text_operations.h -> editing/visible_text.h
misc/khtml_text_operations.cpp -> editing/visible_text.cpp
xml/dom_caretposition.cpp -> editing/visible_position.cpp
xml/dom_caretposition.h -> editing/visible_position.h
xml/dom_selection.cpp -> editing/selection.cpp
xml/dom_selection.h -> editing/selection.h
- broke out the word/line/paragraph part of visible_position.h into visible_units.h
- removed some unnecessary includes from some header files to reduce the number of files that trigger "building the world"
- WebCore.pbproj/project.pbxproj: Added the files under new names, removed the old ones.
- ForwardingHeaders/editing/jsediting.h: Added.
- ForwardingHeaders/editing/visible_text.h: Added.
- ForwardingHeaders/misc/khtml_text_operations.h: Removed.
- ForwardingHeaders/xml/dom_caretposition.h: Removed.
- ForwardingHeaders/xml/dom_selection.h: Removed.
- khtml/ecma/kjs_window.cpp: Updated or removed #include lines.
- khtml/editing/htmlediting.cpp: Ditto.
- khtml/editing/htmlediting.h: Ditto.
- khtml/editing/htmlediting_impl.cpp: Ditto.
- khtml/editing/htmlediting_impl.h: Ditto.
- khtml/editing/jsediting.cpp: Ditto.
- khtml/editing/selection.cpp: Ditto.
- khtml/editing/selection.h: Ditto.
- khtml/editing/visible_position.cpp: Ditto.
- khtml/editing/visible_position.h: Ditto.
- khtml/editing/visible_text.cpp: Ditto.
- khtml/editing/visible_text.h: Ditto.
- khtml/editing/visible_units.cpp: Added.
- khtml/editing/visible_units.h: Added.
- khtml/html/html_elementimpl.cpp: Updated or removed #include lines.
- khtml/khtml_part.cpp: Ditto.
- khtml/khtml_part.h: Ditto.
- khtml/khtmlpart_p.h: Ditto.
- khtml/khtmlview.cpp: Ditto.
- khtml/misc/khtml_text_operations.cpp: Removed.
- khtml/misc/khtml_text_operations.h: Removed.
- khtml/rendering/render_block.cpp: Updated or removed #include lines.
- khtml/xml/dom2_rangeimpl.cpp: Ditto.
- khtml/xml/dom_caretposition.cpp: Removed.
- khtml/xml/dom_caretposition.h: Removed.
- khtml/xml/dom_docimpl.cpp: Updated or removed #include lines.
- khtml/xml/dom_docimpl.h: Ditto.
- khtml/xml/dom_elementimpl.cpp: Ditto.
- khtml/xml/dom_nodeimpl.cpp: Ditto.
- khtml/xml/dom_position.cpp: Ditto.
- khtml/xml/dom_selection.cpp: Removed.
- khtml/xml/dom_selection.h: Removed.
- kwq/KWQKHTMLPart.mm: Updated or removed #include lines.
- kwq/KWQKHTMLPart.h: Ditto.
- kwq/KWQRenderTreeDebug.cpp: Ditto.
- kwq/WebCoreBridge.mm: Ditto.
- 6:50 PM Changeset in webkit [7640] by
-
- 13 edits2 adds in trunk
WebCore:
Pass dashboard regions up to WebKit.
Don't collect regions from RenderTexts.
Made more args and return types references to avoid
copying value lists.
Reviewed by Hyatt.
- WebCore.pbproj/project.pbxproj:
- khtml/khtmlview.cpp: (KHTMLView::layout):
- khtml/rendering/render_object.cpp: (RenderObject::collectDashboardRegions):
- khtml/xml/dom_docimpl.cpp: (DocumentImpl::dashboardRegions): (DocumentImpl::setDashboardRegions):
- khtml/xml/dom_docimpl.h:
- kwq/KWQKHTMLPart.h:
- kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::didFirstLayout): (KWQKHTMLPart::dashboardRegionsChanged):
- kwq/WebCoreBridge.h:
- kwq/WebDashboardRegion.h: Added.
- kwq/WebDashboardRegion.m: Added. (-[WebDashboardRegion initWithRect:type:]): (-[WebDashboardRegion copyWithZone:]): (-[WebDashboardRegion dashboardRegionRect]): (-[WebDashboardRegion dashboardRegionType]): (-[WebDashboardRegion description]):
WebKit:
Pass dashboard regions to UI delegate.
Reviewed by Hyatt.
- WebCoreSupport.subproj/WebBridge.m: (-[WebBridge dashboardRegionsChanged:]):
- WebView.subproj/WebUIDelegatePrivate.h:
- copy-webcore-files-to-webkit:
WebBrowser:
Debug only code to display dashboard regions.
Reviewed by Hyatt.
- BrowserWebController.m: (-[BrowserWebView webView:dashboardRegionsChanged:]):
- 4:57 PM Changeset in webkit [7639] by
-
- 5 edits in trunk
Reviewed by Hyatt
Fix for this hitlist bug:
<rdar://problem/3805486> REGRESSION (Mail): Empty block elements can cause the caret to get "stuck" in one spot
- khtml/xml/dom_caretposition.cpp: (DOM::CaretPosition::isCandidate): You can no longer caret into empty blocks with no height. Also did a little clean up in this function.
- layout-tests/editing/deleting/delete-3800834-fix-expected.txt: Regenerated expected results. This test was actually failing and had bogus results checked in!
- layout-tests/editing/selection/select-all-003-expected.txt: Regenerated expected results.
- 4:45 PM Changeset in webkit [7638] by
-
- 2 edits in trunk/WebKit
Fixed build that I just broke.
- WebView.subproj/WebHTMLView.m: (-[WebHTMLView concludeDragForDraggingInfo:actionMask:]):
- 4:42 PM Changeset in webkit [7637] by
-
- 9 edits in trunk
WebCore:
Fixed: <rdar://problem/3812091> REGRESSION (Mail): double-clicked word is not smart inserted on drag
Reviewed by john.
- khtml/editing/htmlediting.cpp: (khtml::MoveSelectionCommand::MoveSelectionCommand): take smartMove arg
- khtml/editing/htmlediting.h:
- khtml/editing/htmlediting_impl.cpp: (khtml::CompositeEditCommandImpl::deleteSelection): take smartDelete arg (khtml::MoveSelectionCommandImpl::MoveSelectionCommandImpl): take smartMove arg (khtml::MoveSelectionCommandImpl::doApply): pass smartMove for smartDelete and smartReplace
- khtml/editing/htmlediting_impl.h:
- kwq/WebCoreBridge.h:
- kwq/WebCoreBridge.mm: (-[WebCoreBridge moveSelectionToDragCaret:smartMove:]): take smartMove arg
WebKit:
<rdar://problem/3812091> REGRESSION (Mail): double-clicked word is not smart inserted on drag
Reviewed by john.
- WebView.subproj/WebHTMLView.m: (-[WebHTMLView concludeDragForDraggingInfo:actionMask:]): pass value for smartMove
- 3:17 PM Changeset in webkit [7636] by
-
- 5 edits in trunk/WebCore
Reviewed by Hyatt
- khtml/css/css_computedstyle.cpp: (DOM::CSSComputedStyleDeclarationImpl::getPropertyCSSValue): New overloaded version of this function, one that takes a flag to determine whether to perform a document updateLayout() call before querying the style system.
- khtml/css/css_computedstyle.h: (DOM::EUpdateLayout): Give a symbolic constant to true/false for the purpose of the new call to getPropertyCSSValue.
- khtml/editing/htmlediting_impl.cpp: (khtml::StyleChange::currentlyHasStyle): Pass DoNotUpdateLayout to call to getPropertyCSSValue. (khtml::CompositeEditCommandImpl::applyTypingStyle): Add in top-level calls to updateLayout before doing style changes that now do not update styles themselves. (khtml::ApplyStyleCommandImpl::doApply): Ditto. (khtml::ApplyStyleCommandImpl::removeCSSStyle): Pass DoNotUpdateLayout to call to getPropertyCSSValue. (khtml::DeleteSelectionCommandImpl::computeTypingStyle): Remove this dead code.
- khtml/editing/htmlediting_impl.h: Ditto.
- 2:25 PM Changeset in webkit [7635] by
-
- 5 edits in trunk/WebKit
WebKit:
Fixed:
<rdar://problem/3667301> Frequent crashes in Mail when viewing HTML messages (CFURLGetByteRangeForComponent)
<rdar://problem/3810354> WebResourceLoadDelegate can't refuse requests by returning nil; code asserts/crashes instead
Reviewed by rjw.
- WebCoreSupport.subproj/WebSubresourceClient.m: (+[WebSubresourceClient startLoadingResource:withRequest:customHeaders:referrer:forDataSource:]): removed broken code that handled loadWithRequest returning NO
- WebView.subproj/WebBaseResourceHandleDelegate.m: (-[WebBaseResourceHandleDelegate loadWithRequest:]): if nil is returned from the client for willSendRequest, report the cancelled error and return
WebBrowser:
Reviewed by rjw.
- LoadProgressMonitor.m: (-[LoadProgressMonitor webView:resource:willSendRequest:redirectResponse:fromDataSource:]): updated to use "new" API, this wasn't being called!
- 2:12 PM Changeset in webkit [7634] by
-
- 9 edits2 adds in trunk
WebCore:
Reviewed by Darin
- WebCore.pbproj/project.pbxproj: New files added.
- khtml/khtml_part.cpp: (KHTMLPart::clearSelection): No longer calls setFocusIfNeeded. (KHTMLPart::slotClearSelection): Merged old implementation from this function to clearSelection. Now just calls clearSelection. Also put in !APPLE_CHANGES since we do not call it.
- khtml/xml/dom_docimpl.cpp: (DocumentImpl::setFocusNode): Add some code to determine when setting the focus should clear the selection.
- kwq/KWQWidget.mm: (QWidget::hasFocus): Call _webcore_effectiveFirstResponder to get view to check for focus. (QWidget::setFocus): Ditto.
- kwq/WebCoreBridge.mm: (-[WebCoreBridge deselectText]): Called slotClearSelection and should have been. Now calls clearSelection.
- kwq/WebCoreView.h: Added.
- kwq/WebCoreView.m: Added. (-[NSView _webcore_effectiveFirstResponder]): New method to yield the correct responder to check for firstResponder-ness before calling makeFirstResonder. This helps to prevent unwanted firstResponder switching. (-[NSClipView _webcore_effectiveFirstResponder]): Ditto. (-[NSScrollView _webcore_effectiveFirstResponder]): Ditto.
WebKit:
Reviewed by Darin
- WebView.subproj/WebFrameView.m: (-[WebFrameView _webcore_effectiveFirstResponder]): New function to yield the correct responder to check for firstResponder-ness before calling makeFirstResonder. This helps to prevent unwanted firstResponder switching.
- WebView.subproj/WebView.m: (-[WebView _webcore_effectiveFirstResponder]): Ditto.
- 1:36 PM Changeset in webkit [7633] by
-
- 3 edits in trunk/WebCore
Reviewed by Darin.
<rdar://problem/3805137> REGRESSION: Cannot search on pricetool.com
The problem here is that residual style handling can cause an
element associated with a misnested form to lose the association,
because it can become detached and then reattached in this
case. So we need to maintain the association.
- khtml/html/html_formimpl.h: Added list of dormant elements to form, and dormant bit to generic form element.
- khtml/html/html_formimpl.cpp: (DOM::HTMLFormElementImpl::~HTMLFormElementImpl): Clear m_form field for dormant elements too. (DOM::HTMLFormElementImpl::registerFormElement): Remove from dormant list in addition to adding to main list. (DOM::HTMLFormElementImpl::removeFormElement): Remove from both lists. (DOM::HTMLFormElementImpl::makeFormElementDormant): New method, remove from main list, add to dormant list. (DOM::HTMLGenericFormElementImpl::HTMLGenericFormElementImpl): Initialize dormant bit to false. (DOM::HTMLGenericFormElementImpl::insertedIntoDocument): If the element is dormant and has a form, re-register it and clear the dormant bit. (DOM::HTMLGenericFormElementImpl::removedFromDocument): If the document has a form, tell the form it is dormant and set the dormant bit.
- 1:26 PM Changeset in webkit [7632] by
-
- 4 edits in trunk/WebCore
Fixed: <rdar://problem/3811187> REGRESSION (Mail): Control-click past end of document does not spell check last word
Reviewed by kocienda.
- khtml/xml/dom_selection.cpp: (DOM::Selection::validate): if at the end of the document, expand to the left.
- 1:15 PM Changeset in webkit [7631] by
-
- 1 edit in trunk/WebCore/ChangeLog-2005-08-23
Remove stray non-ascii characters in ChangeLog.
- 1:14 PM Changeset in webkit [7630] by
-
- 3 edits in trunk/WebCore
Reviewed by Ken and John.
<rdar://problem/3759228> REGRESSION (Mail): stray characters when entering text via input method right after newline
- khtml/editing/htmlediting_impl.cpp: (khtml::CompositeEditCommandImpl::inputText): Add an optional selectInsertedText parameter. (khtml::ReplaceSelectionCommandImpl::doApply): Let the inputText operation select the text when inserting plain text, because it already knows all the right information to do so, and this function doesn't (sometimes text is inserted before the start of the old selection, not after!)
- khtml/editing/htmlediting_impl.h:
- 1:03 PM Changeset in webkit [7629] by
-
- 3 edits in trunk/WebCore
More tweaks to dashboard regions style.
Move dashboard regions list into css3NonInheritedData so it will be shared
by all styles that don't modify the empty region list.
Make the initial value for dashboard region list a static to minimize allocations
of empty lists.
Unrelated to regions: make marquee, flex box, and opacity style accessor functions const.
Reviewed by Hyatt.
- khtml/rendering/render_style.cpp: (RenderStyle::diff):
- khtml/rendering/render_style.h: (khtml::RenderStyle::opacity): (khtml::RenderStyle::boxAlign): (khtml::RenderStyle::boxDirection): (khtml::RenderStyle::boxFlexGroup): (khtml::RenderStyle::boxOrdinalGroup): (khtml::RenderStyle::boxOrient): (khtml::RenderStyle::boxPack): (khtml::RenderStyle::marqueeIncrement): (khtml::RenderStyle::marqueeSpeed): (khtml::RenderStyle::marqueeLoopCount): (khtml::RenderStyle::marqueeBehavior): (khtml::RenderStyle::marqueeDirection): (khtml::RenderStyle::dashboardRegions): (khtml::RenderStyle::setDashboardRegions): (khtml::RenderStyle::setDashboardRegion): (khtml::RenderStyle::initialDashboardRegions):
- 11:42 AM Changeset in webkit [7628] by
-
- 2 edits in trunk/WebCore
Fix for 3802766, entire view redrawn in Blot while typing. Don't needlessly create text nodes just to find
out the font for the caret position.
Reviewed by darin
- kwq/KWQKHTMLPart.mm: (KWQKHTMLPart::fontForSelection):