ap [Thu, 19 Oct 2006 16:37:48 +0000 (16:37 +0000)]
2006-10-19 Mitz Pettel <mitz@webkit.org>
Reviewed and landed by ap.
- fixed the inspector's tree popup
* WebInspector/webInspector/inspector.html:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17135
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
andersca [Thu, 19 Oct 2006 16:15:21 +0000 (16:15 +0000)]
2006-10-19 Anders Carlsson <acarlsson@apple.com>
* loader/Cache.cpp:
Get rid of #include "loader.h", Cache.h already includes this and it breaks on operating systems
with case-sensitive file systems.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17134
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
andersca [Thu, 19 Oct 2006 16:09:00 +0000 (16:09 +0000)]
2006-10-19 Anders Carlsson <acarlsson@apple.com>
Reviewed by Maciej, Brady.
Remove "charset" from the ResourceLoader metadata map and use
a getter function instead. Also, rename charset to encoding in a couple of places.
* loader/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::setEncoding):
* loader/CachedCSSStyleSheet.h:
* loader/CachedResource.h:
(WebCore::CachedResource::setEncoding):
* loader/CachedScript.cpp:
(WebCore::CachedScript::setEncoding):
* loader/CachedScript.h:
* loader/CachedXBLDocument.cpp:
(WebCore::CachedXBLDocument::setEncoding):
* loader/CachedXBLDocument.h:
* loader/CachedXSLStyleSheet.cpp:
(WebCore::CachedXSLStyleSheet::setEncoding):
* loader/CachedXSLStyleSheet.h:
* loader/loader.cpp:
(WebCore::Loader::receivedResponse):
* platform/ResourceLoader.cpp:
(WebCore::ResourceLoader::responseEncoding):
(WebCore::ResourceLoader::queryMetaData):
* platform/ResourceLoader.h:
* platform/ResourceLoaderInternal.h:
(WebCore::ResourceLoaderInternal::ResourceLoaderInternal):
* platform/mac/ResourceLoaderMac.mm:
(WebCore::ResourceLoader::retrieveResponseEncoding):
(WebCore::ResourceLoader::receivedResponse):
* xml/xmlhttprequest.cpp:
(WebCore::XMLHttpRequest::receivedData):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17133
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Thu, 19 Oct 2006 14:58:51 +0000 (14:58 +0000)]
Updated this test to avoid machine-specific results.
* fast/js/window-properties2-expected.txt:
* fast/js/window-properties2.html:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17132
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Thu, 19 Oct 2006 09:21:54 +0000 (09:21 +0000)]
Rewrite of the WebCore memory cache. The following changes have been made (along with many more):
(1) Referenced objects are now also cached.
(2) Pruning of the cache will occur once the cache is both full and in addition
more than half full with unreferenced objects.
(3) The prune is much more aggressive now, since referenced objects are counted when
pruning. The cache will therefore often prune out every unreferenced object if lots of
windows/tabs are open.
(4) Consolidated the request methods for various types of resources to go through a single
requestResource bottleneck function.
(5) deref on CachedResource is no longer virtual, since the implementation in every subclass was
identical.
(6) Changed "object" to "resource" everywhere.
(7) The Cache is now an actual object, obtained via cache().
(8) Fixed bugs with the autoload images preference so that it would really work properly when disabled
and enabled.
(9) Tweaked ImageDocuments to set up the status properly for the CachedResource that they use.
(10) Fixed the LRU-SP algorithm so that the LRU aspect really kicks in (by leaving the live objects
in the LRU lists). Also fixed a bug with the "-SP" part of the algorithm where the access count was
being incorrectly weighted too much due to a math error.
(11) Fixed a bug where the old cache could grow without bound because it didn't know anything about
the size of referenced resources.
Reviewed by Maciej, bradee-oh
* bridge/mac/WebCoreCache.mm:
(+[WebCoreCache statistics]):
(+[WebCoreCache empty]):
(+[WebCoreCache setDisabled:]):
* bridge/mac/WebCoreFrameBridge.mm:
(-[WebCoreFrameBridge initMainFrameWithPage:]):
(-[WebCoreFrameBridge getData:andResponse:forURL:]):
(-[WebCoreFrameBridge getAllResourceDatas:andResponses:]):
* css/CSSImageValue.cpp:
(WebCore::CSSImageValue::image):
* dom/XMLTokenizer.cpp:
(WebCore::openFunc):
* html/HTMLImageLoader.cpp:
(WebCore::HTMLImageLoader::updateFromElement):
* loader/Cache.cpp:
(WebCore::cache):
(WebCore::Cache::Cache):
(WebCore::createResource):
(WebCore::Cache::requestResource):
(WebCore::Cache::resourceForURL):
(WebCore::Cache::prune):
(WebCore::Cache::setMaximumSize):
(WebCore::Cache::remove):
(WebCore::Cache::addDocLoader):
(WebCore::Cache::removeDocLoader):
(WebCore::fastLog2):
(WebCore::Cache::lruListFor):
(WebCore::Cache::removeFromLRUList):
(WebCore::Cache::insertInLRUList):
(WebCore::Cache::resourceAccessed):
(WebCore::Cache::adjustSize):
(WebCore::Cache::getStatistics):
(WebCore::Cache::setDisabled):
* loader/Cache.h:
(WebCore::LRUList::LRUList):
(WebCore::Cache::loader):
(WebCore::Cache::maximumSize):
(WebCore::Cache::disabled):
(WebCore::Cache::addToLiveObjectSize):
(WebCore::Cache::removeFromLiveObjectSize):
(WebCore::Cache::TypeStatistic::TypeStatistic):
* loader/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::CachedCSSStyleSheet):
* loader/CachedCSSStyleSheet.h:
* loader/CachedImage.cpp:
(WebCore::CachedImage::CachedImage):
(WebCore::CachedImage::~CachedImage):
(WebCore::brokenImage):
(WebCore::nullImage):
(WebCore::CachedImage::image):
(WebCore::CachedImage::data):
* loader/CachedImage.h:
(WebCore::CachedImage::stillNeedsLoad):
* loader/CachedResource.cpp:
(WebCore::CachedResource::CachedResource):
(WebCore::CachedResource::~CachedResource):
(WebCore::CachedResource::finish):
(WebCore::CachedResource::setExpireDate):
(WebCore::CachedResource::isExpired):
(WebCore::CachedResource::setRequest):
(WebCore::CachedResource::ref):
(WebCore::CachedResource::deref):
(WebCore::CachedResource::setSize):
* loader/CachedResource.h:
(WebCore::CachedResource::):
(WebCore::CachedResource::setCharset):
(WebCore::CachedResource::url):
(WebCore::CachedResource::type):
(WebCore::CachedResource::referenced):
(WebCore::CachedResource::count):
(WebCore::CachedResource::status):
(WebCore::CachedResource::size):
(WebCore::CachedResource::isLoaded):
(WebCore::CachedResource::setLoading):
(WebCore::CachedResource::isImage):
(WebCore::CachedResource::accessCount):
(WebCore::CachedResource::increaseAccessCount):
(WebCore::CachedResource::setInCache):
(WebCore::CachedResource::inCache):
(WebCore::CachedResource::cachePolicy):
(WebCore::CachedResource::response):
(WebCore::CachedResource::allData):
(WebCore::CachedResource::canDelete):
(WebCore::CachedResource::schedule):
(WebCore::CachedResource::accept):
(WebCore::CachedResource::setAccept):
* loader/CachedScript.cpp:
(WebCore::CachedScript::CachedScript):
* loader/CachedScript.h:
* loader/CachedXBLDocument.cpp:
* loader/CachedXBLDocument.h:
* loader/CachedXSLStyleSheet.cpp:
(WebCore::CachedXSLStyleSheet::CachedXSLStyleSheet):
* loader/CachedXSLStyleSheet.h:
* loader/DocLoader.cpp:
(WebCore::DocLoader::DocLoader):
(WebCore::DocLoader::~DocLoader):
(WebCore::DocLoader::checkForReload):
(WebCore::DocLoader::requestImage):
(WebCore::DocLoader::requestCSSStyleSheet):
(WebCore::DocLoader::requestScript):
(WebCore::DocLoader::requestXSLStyleSheet):
(WebCore::DocLoader::requestXBLDocument):
(WebCore::DocLoader::requestResource):
(WebCore::DocLoader::setAutoLoadImages):
(WebCore::DocLoader::removeCachedResource):
* loader/DocLoader.h:
(WebCore::DocLoader::cachedResource):
(WebCore::DocLoader::allCachedResources):
(WebCore::DocLoader::autoLoadImages):
(WebCore::DocLoader::cachePolicy):
(WebCore::DocLoader::expireDate):
(WebCore::DocLoader::frame):
(WebCore::DocLoader::doc):
(WebCore::DocLoader::loadInProgress):
* loader/ImageDocument.cpp:
(WebCore::ImageTokenizer::finish):
* loader/loader.cpp:
(WebCore::Loader::receivedAllData):
(WebCore::Loader::cancelRequests):
* loader/mac/LoaderFunctionsMac.mm:
(WebCore::NumberOfPendingOrLoadingRequests):
(WebCore::CheckCacheObjectStatus):
* page/Frame.cpp:
(WebCore::UserStyleSheetLoader::UserStyleSheetLoader):
(WebCore::Frame::Frame):
(WebCore::Frame::stopLoading):
(WebCore::Frame::setAutoLoadImages):
(WebCore::Frame::autoLoadImages):
(WebCore::Frame::begin):
(WebCore::Frame::checkCompleted):
(WebCore::Frame::reparseConfiguration):
* page/Frame.h:
* xml/XSLTProcessor.cpp:
(WebCore::docLoaderFunc):
* xml/xmlhttprequest.cpp:
(WebCore::XMLHttpRequest::send):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17131
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Thu, 19 Oct 2006 06:17:15 +0000 (06:17 +0000)]
Test for some window properties that WebCore previously didn't enumerate.
* fast/js/window-properties2-expected.txt: Added.
* fast/js/window-properties2.html: Added.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17130
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
thatcher [Thu, 19 Oct 2006 05:09:44 +0000 (05:09 +0000)]
Reviewed by Tim H.
Improve the doubleclick behavior of breakpoints, and make breakpoints with no custom condition set appear blank instead of return [-1, 1] depending on enabled state.
* Drosera/DebuggerDocument.m:
(-[WebScriptObject getDoubleClickMillis]): Add an ObjC wrapper for GetDblTime() so that JS can call it
* Drosera/Drosera.xcodeproj/project.pbxproj: link Carbon for GetDblTime()
* Drosera/debugger.js:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17129
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
thatcher [Thu, 19 Oct 2006 05:02:54 +0000 (05:02 +0000)]
Reviewed by Anders.
Bug 10851: Crash with Drosera
http://bugs.webkit.org/show_bug.cgi?id=10851
This crash results in an assert in debug builds.
assert(implementsCall());
The __drosera_introspection propery was not callable. Now we just assign
this.__drosera_introspection in the evaluateWebScript call.
This change also removes one DO message.
* Drosera/DebuggerDocument.m:
(-[WebScriptObject webScriptAttributeKeysForScriptObject:]):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17128
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
mjs [Thu, 19 Oct 2006 02:42:55 +0000 (02:42 +0000)]
JavaScriptCore:
Reviewed by Geoff.
- remove vestiges of KXMLCore name (former name of WTF).
* wtf/Assertions.h:
* wtf/FastMalloc.h:
(operator new):
(operator delete):
(operator new[]):
(operator delete[]):
* wtf/FastMallocInternal.h:
* wtf/Forward.h:
* wtf/GetPtr.h:
* wtf/HashCountedSet.h:
* wtf/HashFunctions.h:
* wtf/HashMap.h:
* wtf/HashSet.h:
* wtf/HashTable.h:
* wtf/HashTraits.h:
* wtf/ListRefPtr.h:
* wtf/MathExtras.h:
* wtf/Noncopyable.h:
* wtf/OwnArrayPtr.h:
* wtf/OwnPtr.h:
* wtf/PassRefPtr.h:
* wtf/Platform.h:
* wtf/RefPtr.h:
* wtf/StringExtras.h:
(snprintf):
* wtf/UnusedParam.h:
* wtf/Vector.h:
* wtf/VectorTraits.h:
WebCore:
Reviewed by Geoff.
- remove vestiges of KXMLCore name (former name of WTF).
* config.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17127
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Thu, 19 Oct 2006 01:43:01 +0000 (01:43 +0000)]
LayoutTests:
Reviewed by Beth.
Updated some results now that some global JS objects repeatably report their
statically declared properties.
* fast/js/window-properties-expected.txt:
* fast/js/window-properties.html:
WebCore:
Reviewed by Beth.
Fixed http://bugs.webkit.org/show_bug.cgi?id=11222
Some global JS objects lacked references to their property hashtables,
so for-in enumeration didn't work. I added the references, and then removed
a few bogus properties from the tables. The bogus ones got in there through
bit rot.
We should come up with some solution to these hand-written bindings bugs.
Maybe we could use a script to autogenerate them or something.
* bindings/js/kjs_window.cpp:
(KJS::):
(KJS::History::getValueProperty):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17126
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Thu, 19 Oct 2006 01:05:16 +0000 (01:05 +0000)]
Reviewed by Maciej.
Fix: http://bugs.webkit.org/show_bug.cgi?id=11342
fast/repaint/
4774354 is failing
It seems that this test got checked in with incorrect results, as it
has been failing since its introduction in r17093.
* fast/repaint/
4774354-expected.txt:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17125
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Thu, 19 Oct 2006 00:11:57 +0000 (00:11 +0000)]
LayoutTests:
Reviewed by Brady.
New test to make sure we're generating the correct keycodes for Tab
and Shift-Tab keypresses.
* fast/events/js-keyboard-event-creation-expected.txt: Added.
* fast/events/js-keyboard-event-creation.html: Added.
WebCore:
Reviewed by Brady.
Change Shift-Tab to generate keyIdentifier U+000009 (Tab) instead of
U+000019 (Backtab). This matches our windowsKeyCode for this key
combination.
* platform/mac/KeyEventMac.mm:
(WebCore::keyIdentifierForKeyEvent):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17124
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
sfalken [Wed, 18 Oct 2006 23:28:27 +0000 (23:28 +0000)]
2006-10-18 Steve Falkenburg <sfalken@apple.com>
Reviewed by Adam, Brady.
Loader tweaks.
* platform/cf/ResourceLoaderCFNet.cpp:
(WebCore::didFinishLoading):
(WebCore::didFail):
(WebCore::didReceiveChallenge):
(WebCore::runLoaderThread):
(WebCore::ResourceLoader::start):
(WebCore::ResourceLoader::cancel):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17123
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Wed, 18 Oct 2006 23:15:54 +0000 (23:15 +0000)]
Reviewed by Brady.
Fix: http://bugs.webkit.org/show_bug.cgi?id=11344
REGRESSION (r16741): webViewDidChangeSelection delegate not firing in
2 layout tests
The behavior for display-none-in-onchange-keyboard is still correct
(we don't crash when the element about to receive focus gets
display:none set), so I've just generated new results for it.
The behavior of tabbing-input-iframe has changed a little. iframe
elements are now in the tab order, so I've changed the test to take
that into account and generated new results.
* fast/forms/display-none-in-onchange-keyboard-expected.txt:
* fast/forms/tabbing-input-iframe-expected.txt:
* fast/forms/tabbing-input-iframe.html:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17122
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
thatcher [Wed, 18 Oct 2006 22:56:13 +0000 (22:56 +0000)]
Reviewed by Darin.
<rdar://problem/
4752138> Manipulating popup in HTML page crashed Xcode
- Store the Mac popup button cell as a RetainPtr to prevent GC collection.
- Convert more ObjC local statics and member variables to use RetainPtr.
- Convert more CFRetain/CFRelease to HardRetain/HardRelease.
* bindings/objc/DOM.mm:
(ObjCNodeFilterCondition::ObjCNodeFilterCondition): use HardRetain and not CFRetain
(ObjCNodeFilterCondition::~ObjCNodeFilterCondition): use HardRelease and not CFRelease
* bridge/mac/AXObjectCacheMac.mm:
(WebCore::AXObjectCache::~AXObjectCache): use HardRelease and not CFRelease
(WebCore::AXObjectCache::get): use HardRetain and not CFRetain
(WebCore::AXObjectCache::remove): use HardRelease and not CFRelease
* bridge/mac/WebCoreFrameBridge.mm:
(+[WebCoreFrameBridge supportedImageResourceMIMETypes]): use RetainPtr
(+[WebCoreFrameBridge supportedImageMIMETypes]): ditto
* bridge/mac/WebCoreIconDatabaseBridge.mm:
(+[WebCoreIconDatabaseBridge sharedInstance]): use RetainPtr
* platform/PopupMenu.h: renamed popup to m_popup and switched to RetainPtr<NSPopUpButtonCell>
* platform/cf/RetainPtr.h:
(WTF::::operator): implement a missing operator= template function
* platform/mac/PopupMenuMac.mm:
(WebCore::PopupMenu::PopupMenu): remove nil initialization
(WebCore::PopupMenu::~PopupMenu): use .get(), remove release message and rename to m_popup
(WebCore::PopupMenu::clear): use .get() when sending a message and rename to m_popup
(WebCore::PopupMenu::populate): ditto
(WebCore::PopupMenu::show): ditto
(WebCore::PopupMenu::hide): ditto
(WebCore::PopupMenu::addSeparator): ditto
(WebCore::PopupMenu::addGroupLabel): ditto
(WebCore::PopupMenu::addOption): ditto
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17121
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Wed, 18 Oct 2006 22:45:56 +0000 (22:45 +0000)]
Reviewed by Adele.
Fixed <rdar://problem/
3673233> JavaScript timers do not fire if controls
are tracking or menus are down (run loop modes)
I spent a few hours trying to write an automated test, but DRT doesn't
seem to support it. Manual tests attached to related bugs.
* bridge/mac/FrameMac.mm:
(WebCore::FrameMac::passMouseDownEventToWidget):
* platform/mac/SharedTimerMac.cpp:
(WebCore::setSharedTimerFireTime):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17120
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
bdash [Wed, 18 Oct 2006 21:30:08 +0000 (21:30 +0000)]
2006-10-18 Mark Rowe <bdash@webkit.org>
Reviewed by Tim H.
http://bugs.webkit.org/show_bug.cgi?id=11304
Bug 11304: Drosera fails to link as universal binary on PowerPC machine
* Drosera/Drosera.xcodeproj/project.pbxproj: Use -weak_framework to link against JavaScriptCore and WebCore directly
when they are not part of the WebKit umbrella framework.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17119
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Wed, 18 Oct 2006 21:12:33 +0000 (21:12 +0000)]
WebCore:
Reviewed by Maciej.
More Win32 buildfix.
* bridge/win/FrameWin.h:
WebKit:
Reviewed by Maciej.
Win32 build fix.
* COM/WebFrame.cpp:
(WebFrame::createNewWindow):
* COM/WebFrame.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17118
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
andersca [Wed, 18 Oct 2006 20:51:49 +0000 (20:51 +0000)]
2006-10-18 Anders Carlsson <acarlsson@apple.com>
Reviewed by Adam.
http://bugs.webkit.org/show_bug.cgi?id=11000
REGRESSION (r16101): css2.1/t0801-c412-hz-box-00-b-a is failing because the QuickTime plugin is taking over but not rendering the png
* WebCoreSupport/WebFrameBridge.m:
(-[WebFrameBridge determineObjectFromMIMEType:URL:]):
Return ObjectElementFrame if the MIME type is one of the image ones we support.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17117
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Wed, 18 Oct 2006 20:17:22 +0000 (20:17 +0000)]
Reviewed by Mitz
Win32 build fix.
* bridge/win/BrowserExtensionWin.cpp:
(WebCore::BrowserExtensionWin::BrowserExtensionWin):
(WebCore::BrowserExtensionWin::createNewWindow):
* bridge/win/BrowserExtensionWin.h:
* bridge/win/FrameWin.cpp:
(WebCore::FrameWin::createNewWindow):
* bridge/win/FrameWin.h:
* platform/win/CursorWin.cpp:
* platform/win/TemporaryLinkStubs.cpp:
(Image::getHBITMAP):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17116
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Wed, 18 Oct 2006 19:35:46 +0000 (19:35 +0000)]
fixo el buildo II: Release's Pride.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17115
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Wed, 18 Oct 2006 19:27:19 +0000 (19:27 +0000)]
LayoutTests:
Reviewed by harrison
<rdar://problem/
4784823>
GMail Editor: Hang occurs when removing list styling on text in a rich text message
* editing/execCommand/remove-list-1-expected.checksum: Added.
* editing/execCommand/remove-list-1-expected.png: Added.
* editing/execCommand/remove-list-1-expected.txt: Added.
* editing/execCommand/remove-list-1.html: Added.
WebCore:
Reviewed by harrison
<rdar://problem/
4784823>
GMail Editor: Hang occurs when removing list styling on text in a rich text message
List removal moves the contents of every list item out of the list it's
in. When the code tried to move the contents of an empty list item (an li
with no child nodes, not even a placeholder br), moveParagraph didn't prune
the li, like it would if the li had a placeholder inside it. So the list
removal code went into an infinite loop, continually attempting to de-list
an empty list item.
* editing/CompositeEditCommand.cpp:
(WebCore::CompositeEditCommand::moveParagraphs): If the selection
to move was empty and in an empty block that doesn't require a
placeholder to prop itself open (like a bordered div or an li),
remove it during the move.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17114
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Wed, 18 Oct 2006 19:22:45 +0000 (19:22 +0000)]
LayoutTests:
Reviewed by harrison
<rdar://problem/
4765600>
REGRESSION: Mail.app: smart deletion of words does not work
* editing/deleting/smart-delete-003-expected.checksum: Added.
* editing/deleting/smart-delete-003-expected.png: Added.
* editing/deleting/smart-delete-003-expected.txt: Added.
* editing/deleting/smart-delete-003.html: Added.
* editing/deleting/smart-delete-004-expected.checksum: Added.
* editing/deleting/smart-delete-004-expected.png: Added.
* editing/deleting/smart-delete-004-expected.txt: Added.
* editing/deleting/smart-delete-004.html: Added.
WebKit:
Reviewed by harrison
<rdar://problem/
4765600>
REGRESSION: Mail.app: smart deletion of words does not work
Regressed when we pushed selecion expansion down into WebCore. It's OK
to try a smart delete from _deleteWithDirection:, which is called by
deleteFoward: and deleteBackward: if the current selection is a
range.
* WebView/WebHTMLView.m:
(-[NSArray _deleteWithDirection:granularity:killRing:isTypingAction:]):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17113
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Wed, 18 Oct 2006 18:59:12 +0000 (18:59 +0000)]
fixo el buildo.
Work around #import of <PDFKit/PDFView.h>.
* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17112
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Wed, 18 Oct 2006 18:47:25 +0000 (18:47 +0000)]
Reviewed by Beth.
- Fixed http://bugs.webkit.org/show_bug.cgi?id=11343
REGRESSION (r16975): fast/events/objc-event-api is failing
When I moved the autoscroll code from FrameMac to Frame, I mistakingly changed the order of
a call to invalidateClick. This change moves it back to the sequence that existed before. This allows
the click count to get properly reset during a selection.
* page/Frame.cpp: (WebCore::Frame::handleMouseMoveEvent):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17111
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
thatcher [Wed, 18 Oct 2006 17:42:29 +0000 (17:42 +0000)]
Removing SVGZoomEvent.idl from Copy Resources.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17109
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Wed, 18 Oct 2006 17:37:04 +0000 (17:37 +0000)]
Reviewed by Maciej.
Adjust include paths
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17108
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
thatcher [Wed, 18 Oct 2006 16:29:44 +0000 (16:29 +0000)]
Reviewed by Tim H.
Add Xcode 3 style inline breakpoint editor. Credit to xenon for the CSS wizardry to get the appearance working properly.
* Drosera/Drosera.xcodeproj/project.pbxproj:
* Drosera/Images/breakpointeditor.png: Added.
* Drosera/breakpointEditor.css: Removed.
* Drosera/breakpointEditor.html: Removed.
* Drosera/breakpointEditor.js: Removed.
* Drosera/debugger.js:
* Drosera/viewer.css:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17107
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
thatcher [Wed, 18 Oct 2006 15:25:48 +0000 (15:25 +0000)]
Reviewed by Tim H.
Bug 11341: REGRESSION (r16760): editing/selection/editable-links is failing
http://bugs.webkit.org/show_bug.cgi?id=11341
Link editing behavior became a preference. DumpRenderTree needs to specify the
non-default behavior it wants (WebKitEditableLinkOnlyLiveWithShiftKey).
* DumpRenderTree/DumpRenderTree.m:
(main):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17106
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Wed, 18 Oct 2006 07:45:54 +0000 (07:45 +0000)]
Reviewed by Maciej.
Fix: http://bugs.webkit.org/show_bug.cgi?id=11340
editing/pasteboard/
4641033 needs new results
Update results after r17071.
* editing/pasteboard/
4641033-expected.txt:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17105
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Wed, 18 Oct 2006 06:04:03 +0000 (06:04 +0000)]
Reviewed by Anders.
Factored common scaling code into a helper function.
* platform/mac/ScreenMac.mm:
(WebCore::scale):
(WebCore::scaleScreenRectToPageCoordinates):
(WebCore::scalePageRectToScreenCoordinates):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17104
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
sfalken [Wed, 18 Oct 2006 05:46:21 +0000 (05:46 +0000)]
2006-10-17 Steve Falkenburg <sfalken@apple.com>
Build fixes
* bridge/win/BrowserExtensionWin.cpp:
(WebCore::BrowserExtensionWin::createNewWindow):
* bridge/win/BrowserExtensionWin.h:
* bridge/win/FrameWin.cpp:
(WebCore::FrameWin::createNewWindow):
* bridge/win/FrameWin.h:
* platform/gdk/BrowserExtensionGdk.h:
* platform/gdk/TemporaryLinkStubs.cpp:
(BrowserExtensionGdk::createNewWindow):
* platform/qt/BrowserExtensionQt.cpp:
(WebCore::BrowserExtensionQt::createNewWindow):
* platform/qt/BrowserExtensionQt.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17103
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Wed, 18 Oct 2006 05:13:41 +0000 (05:13 +0000)]
Reviewed by Anders.
Renamed WindowArgs to WindowFeatures, and moved it into its own header.
("Features" is the term of art in IE documentation. "Args" is pirate
talk.)
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/kjs_window.cpp:
(KJS::createNewWindow):
(KJS::showModalDialog):
(KJS::setWindowFeature):
(KJS::parseWindowFeatures):
(KJS::constrainToVisible):
(KJS::WindowFunc::callAsFunction):
* bridge/BrowserExtension.h:
* bridge/WindowFeatures.h: Added.
* bridge/mac/BrowserExtensionMac.h:
* bridge/mac/BrowserExtensionMac.mm:
(WebCore::BrowserExtensionMac::createNewWindow):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17102
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
sfalken [Wed, 18 Oct 2006 04:09:51 +0000 (04:09 +0000)]
2006-10-17 Steve Falkenburg <sfalken@apple.com>
Reviewed by Maciej.
Adjust include paths
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17101
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Wed, 18 Oct 2006 03:04:58 +0000 (03:04 +0000)]
Reviewed by Tim H.
Patch for http://bugs.webkit.org/show_bug.cgi?id=11268
Implement the Objective-C DOM bindings for SVGNumber, SVGRect, and SVGPoint
- Adds implementation for DOMSVGNumber, DOMSVGRect, and DOMSVGPoint.
- Moves the contents of ksvg/bindings/js/ to bindings/js/.
- Moves SVGZoomEvent.idl from ksvg/svg/ to ksvg/events/.
- Adds isSVGColor() and isSVGPaint() methods to StyleBase, matching
the other CSSValue classing, and allowing the bindings to allocate
appropriately.
- Removes #ifdefs around the getBBox() method in SVGLocatable.idl allowing
generation for Objective-C.
- Assorted style cleanups.
* DerivedSources.make:
* WebCore.xcodeproj/project.pbxproj:
* bindings/objc/DOMCSS.mm:
(+[DOMCSSValue _CSSValueWith:WebCore::]):
* bindings/objc/DOMEvents.mm:
(+[DOMEvent _eventWith:WebCore::]):
* bindings/objc/DOMSVG.h:
* bindings/objc/DOMSVGNumber.mm:
(-[DOMSVGNumber dealloc]):
(-[DOMSVGNumber finalize]):
(-[DOMSVGNumber value]):
(-[DOMSVGNumber setValue:]):
(-[DOMSVGNumber _SVGNumber]):
(-[DOMSVGNumber _initWithFloat:]):
(+[DOMSVGNumber _SVGNumberWith:]):
* bindings/objc/DOMSVGPoint.mm:
(-[DOMSVGPoint dealloc]):
(-[DOMSVGPoint finalize]):
(-[DOMSVGPoint x]):
(-[DOMSVGPoint setX:]):
(-[DOMSVGPoint y]):
(-[DOMSVGPoint setY:]):
(-[DOMSVGPoint WebCore::]):
(-[DOMSVGPoint _initWithFloatPoint:WebCore::]):
(+[DOMSVGPoint _SVGPointWith:WebCore::]):
* bindings/objc/DOMSVGRect.mm:
(-[DOMSVGRect dealloc]):
(-[DOMSVGRect finalize]):
(-[DOMSVGRect x]):
(-[DOMSVGRect setX:]):
(-[DOMSVGRect y]):
(-[DOMSVGRect setY:]):
(-[DOMSVGRect width]):
(-[DOMSVGRect setWidth:]):
(-[DOMSVGRect height]):
(-[DOMSVGRect setHeight:]):
(-[DOMSVGRect WebCore::]):
(-[DOMSVGRect _initWithFloatRect:WebCore::]):
(+[DOMSVGRect _SVGRectWith:WebCore::]):
* bindings/scripts/CodeGeneratorObjC.pm:
* css/StyleBase.h:
(WebCore::StyleBase::isSVGColor):
(WebCore::StyleBase::isSVGPaint):
* ksvg2/bindings/js/JSSVGElementWrapperFactory.cpp: Removed.
* ksvg2/bindings/js/JSSVGElementWrapperFactory.h: Removed.
* ksvg2/bindings/js/JSSVGNumber.cpp: Removed.
* ksvg2/bindings/js/JSSVGNumber.h: Removed.
* ksvg2/bindings/js/JSSVGPoint.cpp: Removed.
* ksvg2/bindings/js/JSSVGPoint.h: Removed.
* ksvg2/bindings/js/JSSVGRect.cpp: Removed.
* ksvg2/bindings/js/JSSVGRect.h: Removed.
* ksvg2/svg/SVGColor.h:
(WebCore::SVGColor::isSVGColor):
* ksvg2/svg/SVGFECompositeElement.idl:
* ksvg2/svg/SVGFEFloodElement.idl:
* ksvg2/svg/SVGLocatable.idl:
* ksvg2/svg/SVGMarkerElement.idl:
* ksvg2/svg/SVGPaint.h:
(WebCore::SVGPaint::isSVGPaint):
* ksvg2/svg/SVGRectElement.idl:
* ksvg2/svg/SVGZoomEvent.idl: Removed.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17100
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Wed, 18 Oct 2006 02:12:14 +0000 (02:12 +0000)]
Fix changelog file list. It wasn't right.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17099
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Wed, 18 Oct 2006 02:10:17 +0000 (02:10 +0000)]
Prepare for the elimination of the RenderLayer for <html> elements. The layer is still present,
but this patch makes everything work without a root layer being around.
Reviewed by beth
* WebCore.xcodeproj/project.pbxproj:
* platform/KURL.cpp: Removed.
* platform/KURL.h: Removed.
* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::paint):
(WebCore::RenderBlock::paintChildren):
(WebCore::RenderBlock::nodeAtPoint):
* rendering/RenderLayer.cpp:
(WebCore::frameVisibleRect):
(WebCore::RenderLayer::hitTest):
(WebCore::RenderLayer::hitTestLayer):
(WebCore::RenderLayer::intersectsDamageRect):
* rendering/RenderObject.cpp:
(WebCore::RenderObject::setInnerNode):
* rendering/RenderView.cpp:
(WebCore::RenderView::paint):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17098
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Wed, 18 Oct 2006 01:56:21 +0000 (01:56 +0000)]
Back out my previous patch to the WebCore memory cache.
* loader/Cache.cpp:
(WebCore::Cache::flush):
(WebCore::Cache::setSize):
(WebCore::Cache::getLRUListFor):
(WebCore::Cache::removeFromLRUList):
(WebCore::Cache::insertInLRUList):
* loader/Cache.h:
* loader/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::deref):
* loader/CachedCSSStyleSheet.h:
* loader/CachedImage.cpp:
(WebCore::CachedImage::deref):
* loader/CachedImage.h:
* loader/CachedResource.cpp:
(WebCore::CachedResource::setRequest):
(WebCore::CachedResource::deref):
* loader/CachedResource.h:
(WebCore::CachedResource::allowInLRUList):
* loader/CachedScript.cpp:
(WebCore::CachedScript::deref):
* loader/CachedScript.h:
* loader/CachedXBLDocument.cpp:
(WebCore::CachedXBLDocument::deref):
* loader/CachedXBLDocument.h:
* loader/CachedXSLStyleSheet.cpp:
(WebCore::CachedXSLStyleSheet::deref):
* loader/CachedXSLStyleSheet.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17097
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
oliver [Wed, 18 Oct 2006 01:31:33 +0000 (01:31 +0000)]
2006-10-17 Oliver Hunt <oliver@apple.com>
Reviewed by Anders.
Make north and east resize cursors use correct image
* platform/win/CursorWin.cpp:
(WebCore::eastResizeCursor):
(WebCore::northResizeCursor):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17096
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
oliver [Wed, 18 Oct 2006 00:30:26 +0000 (00:30 +0000)]
2006-10-17 Oliver Hunt <oliver@apple.com>
Reviewed by Brady
Various cursor fixes
* page/FrameView.cpp:
(WebCore::FrameView::handleMouseMoveEvent):
* platform/Cursor.h:
* platform/mac/CursorMac.mm:
(WebCore::pointerCursor):
* platform/win/CursorWin.cpp:
(WebCore::supportsAlphaCursors):
(WebCore::Cursor::Cursor):
(WebCore::Cursor::~Cursor):
(WebCore::pointerCursor):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17095
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Wed, 18 Oct 2006 00:28:37 +0000 (00:28 +0000)]
LayoutTests:
Reviewed by harrison
<rdar://problem/
4776665>
REGRESSION: In Mail, caret is visible when the message body isn't focused
* editing/selection/
4776665-expected.checksum: Added.
* editing/selection/
4776665-expected.png: Added.
* editing/selection/
4776665-expected.txt: Added.
* editing/selection/
4776665.html: Added.
WebCore:
Reviewed by harrison
<rdar://problem/
4776665>
REGRESSION: In Mail, caret is visible when the message body isn't focused
* page/Frame.cpp:
(WebCore::Frame::paintCaret): If a caret blink timer is ever on in error
(like when the frame that contains the caret isn't focused) then we don't
want to paint the caret. We should eventually move to using a single bool
for caret painting.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17094
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Tue, 17 Oct 2006 23:52:15 +0000 (23:52 +0000)]
LayoutTests:
Reviewed by harrison
<rdar://problem/
4774354>
REGRESSION: In Gmail, caret reappears when dragging a selection over some text in the message body
* fast/repaint/
4774354-expected.checksum: Added.
* fast/repaint/
4774354-expected.png: Added.
* fast/repaint/
4774354-expected.txt: Added.
* fast/repaint/
4774354.html: Added.
* fast/repaint/resources/contenteditable-iframe-src.html: Added.
WebCore:
Reviewed by harrison
<rdar://problem/
4774354>
REGRESSION: In Gmail, caret reappears when dragging a selection over some text in the message body
Functions call recomputCaretRect to see if the caret rect has changed,
and, if it has, they invalidate (repaint) it. recomputeCaretRect was
returning false if the caret turned into a range selection or was blown
away.
* editing/SelectionController.cpp:
(WebCore::SelectionController::recomputeCaretRect): Don't early return if
!isCaret(). Obtain the FrameView* using the m_frame pointer, because m_sel
may now be null.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17093
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Tue, 17 Oct 2006 23:37:01 +0000 (23:37 +0000)]
Reviewed by Brady.
BrowserExtension cleanup.
- Removed unnecessary #include of BrowserExtension in Frame.h, added
#includes to compensate in other files.
- Removed unused versions of createNewWindow.
* bindings/js/kjs_window.cpp:
* bridge/BrowserExtension.h:
* bridge/mac/BrowserExtensionMac.h:
* bridge/mac/BrowserExtensionMac.mm:
(WebCore::BrowserExtensionMac::createNewWindow):
* bridge/mac/WebCoreFrameBridge.mm:
* html/HTMLFormElement.cpp:
* loader/mac/LoaderFunctionsMac.mm:
* page/Frame.h:
* page/FramePrivate.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17092
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Tue, 17 Oct 2006 21:46:36 +0000 (21:46 +0000)]
Reviewed by Brady.
Fixes a JavaScriptCore math issue on win.
* kjs/math_object.cpp:
(MathFuncImp::callAsFunction):
* wtf/MathExtras.h:
(wtf_atan2):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17091
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Tue, 17 Oct 2006 21:17:28 +0000 (21:17 +0000)]
Reviewed by Adam.
- Fixed "FIXME: Violates Liskov Substitution principle."
IFRAME was never a true subclass of FRAME, since they disagreed on a
number of behaviors, which, confusingly, IFRAME just overrode and contradicted.
The fix here is to factor out things common to IFRAME and FRAME into a
common base class, HTMLFrameElementBase.
I also changed some " *" to "* " and added #includes in files that implicitly
assumed that HTMLIFrameElement.h #included HTMLFrameElement.h.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17090
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Tue, 17 Oct 2006 19:47:55 +0000 (19:47 +0000)]
LayoutTests:
Reviewed by sullivan
<rdar://problem/
4776765>
REGRESSION: Caret's ghost left behind after inserting a paragraph separator (11237)
* fast/repaint/
4776765-expected.checksum: Added.
* fast/repaint/
4776765-expected.png: Added.
* fast/repaint/
4776765-expected.txt: Added.
* fast/repaint/
4776765.html: Added.
WebCore:
Reviewed by sullivan
<rdar://problem/
4776765>
REGRESSION: Caret's ghost left behind after inserting a paragraph separator (11237)
We set m_needsLayout to false and call caretRect() in the hopes that it will give us
the old caret rect. It in fact corrects the caret rect for an offset that it
believes is due to scrolling but which is actually due to a change in selection
without an accompanying layout. So it returns the new caret rect regardless of
what m_needsLayout is set to.
* editing/SelectionController.cpp:
(WebCore::repaintRectForCaret): Moved the code from caretRepaintRect that
adds a one pixel slop to this new function.
(WebCore::SelectionController::caretRepaintRect): Moved this code to
repaintRectForCaret.
(WebCore::SelectionController::recomputeCaretRect): Compare the old
caret rect to the new one that's computed with a fresh layout. If
they are different, invalidate both repaint rects.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17089
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aliceli1 [Tue, 17 Oct 2006 19:06:40 +0000 (19:06 +0000)]
Reviewed by Tim Hatcher.
When nodes are removed, selections are cleared, and when http://bugs.webkit.org/show_bug.cgi?id=6498 was fixed, we started sending didChangeSelection notifications. Updating the test results fixes some of the tests mentioned in http://bugs.webkit.org/show_bug.cgi?id=10924
* fast/dynamic/move-node-with-selection-expected.txt:
* fast/events/dblclick-addEventListener-expected.txt:
* fast/events/updateLayoutForHitTest-expected.txt:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17084
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
harrison [Tue, 17 Oct 2006 18:07:03 +0000 (18:07 +0000)]
Reviewed by Adele.
<rdar://problem/
4770453> VO not honoring secure edit fields in web pages
* bridge/mac/WebCoreAXObject.mm:
(-[WebCoreAXObject isPasswordField]):
New. Checks for input with inputType() of PASSWORD.
(-[WebCoreAXObject subrole]):
Returns NSAccessibilitySecureTextFieldSubrole if [self isPasswordField].
(-[WebCoreAXObject roleDescription]):
Pass [self subrole] instead of nil when calling NSAccessibilityRoleDescription.
(-[WebCoreAXObject value]):
Return nil if [self isPasswordField].
(-[WebCoreAXObject accessibilityAttributeNames]):
Return generic attributes if [self isPasswordField].
(-[WebCoreAXObject accessibilityParameterizedAttributeNames]):
Return empty array if [self isPasswordField].
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17083
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aliceli1 [Tue, 17 Oct 2006 16:49:32 +0000 (16:49 +0000)]
Reviewed by aroben.
updating test results after r17068 - http://bugs.webkit.org/show_bug.cgi?id=11320
* fast/block/float/013-expected.checksum:
* fast/block/float/013-expected.txt:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17080
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Tue, 17 Oct 2006 16:36:16 +0000 (16:36 +0000)]
Reviewed by Adele.
Fixes http://bugs.webkit.org/show_bug.cgi?id=11214
http/tests/xmlhttprequest/exceptions.html needs new results
Updating results after r16794.
* http/tests/xmlhttprequest/exceptions-expected.txt:
* http/tests/xmlhttprequest/exceptions.html: Add comments about why we
don't follow the XHR spec.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17078
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Tue, 17 Oct 2006 08:52:26 +0000 (08:52 +0000)]
Fix "flash to white' problem on yahoo.com as it finishes loading. Refine the FOUC suppression
code so that it will only run if the body still has no renderer.
* dom/Document.cpp:
(WebCore::Document::updateLayoutIgnorePendingStylesheets):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17076
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Tue, 17 Oct 2006 07:32:34 +0000 (07:32 +0000)]
Fix a bug where the WebCore memory cache could grow without bound during extended browsing
sessions.
Reviewed by mjs
* loader/Cache.cpp:
(WebCore::Cache::flush):
(WebCore::Cache::setSize):
(WebCore::Cache::getLRUListFor):
(WebCore::Cache::removeFromLRUList):
(WebCore::Cache::insertInLRUList):
* loader/Cache.h:
* loader/CachedCSSStyleSheet.cpp:
* loader/CachedCSSStyleSheet.h:
* loader/CachedImage.cpp:
* loader/CachedImage.h:
* loader/CachedResource.cpp:
(WebCore::CachedResource::setRequest):
(WebCore::CachedResource::deref):
* loader/CachedResource.h:
(WebCore::CachedResource::allowInLRUList):
* loader/CachedScript.cpp:
* loader/CachedScript.h:
* loader/CachedXBLDocument.cpp:
* loader/CachedXBLDocument.h:
* loader/CachedXSLStyleSheet.cpp:
* loader/CachedXSLStyleSheet.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17075
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Tue, 17 Oct 2006 00:48:03 +0000 (00:48 +0000)]
LayoutTests:
Reviewed by harrison
<rdar://problem/
3655385>
Editing: -indent: method unimplemented
* editing/execCommand/nsresponder-indent-expected.checksum: Added.
* editing/execCommand/nsresponder-indent-expected.png: Added.
* editing/execCommand/nsresponder-indent-expected.txt: Added.
* editing/execCommand/nsresponder-indent.html: Added.
* editing/execCommand/nsresponder-outdent-expected.checksum: Added.
* editing/execCommand/nsresponder-outdent-expected.png: Added.
* editing/execCommand/nsresponder-outdent-expected.txt: Added.
* editing/execCommand/nsresponder-outdent.html: Added.
WebCore:
Reviewed by harrison
<rdar://problem/
3655385>
Editing: -indent: method unimplemented
Added Frame and bridge methods.
* bridge/mac/WebCoreFrameBridge.h:
* bridge/mac/WebCoreFrameBridge.mm:
(-[WebCoreFrameBridge indent]):
(-[WebCoreFrameBridge outdent]):
* page/Frame.cpp:
(WebCore::Frame::indent):
(WebCore::Frame::outdent):
* page/Frame.h:
WebKit:
Reviewed by harrison
<rdar://problem/
3655385>
Editing: -indent: method unimplemented
* WebView/WebHTMLView.m:
(-[NSArray validateUserInterfaceItem:]): Only allow indent:/outdent: in
richly editable areas.
(-[NSArray indent:]):
(-[NSArray outdent:]):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17074
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Tue, 17 Oct 2006 00:35:59 +0000 (00:35 +0000)]
Reviewed by Geof.
Removed unecessary global specifiers.
* kjs/math_object.cpp:
(MathFuncImp::callAsFunction):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17073
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Mon, 16 Oct 2006 23:12:03 +0000 (23:12 +0000)]
Reviewed by John.
Fixes a compile order issue for testkjs on win.
* JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17072
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
aroben [Mon, 16 Oct 2006 21:44:00 +0000 (21:44 +0000)]
Reviewed by John.
More menulist cleanup/tweaking.
* html/HTMLSelectElement.cpp:
(WebCore::HTMLSelectElement::setSelectedIndex): Renamed 'index'
parameter to 'optionIndex' for clarity.
(WebCore::HTMLSelectElement::menuListDefaultEventHandler): Skip over
items that are not enabled <option> elements when using arrow keys, and
make sure we pass an option index to setSelectedIndex.
* platform/PopupMenu.h:
(WebCore::PopupMenu::setWasClicked): Specify default argument.
* rendering/RenderMenuList.cpp:
(WebCore::RenderMenuList::setTextFromOption): Strip whitespace from
option text before setting the control text so that options within
<optgroup>s don't appear indented in the actual control.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17071
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Mon, 16 Oct 2006 20:52:49 +0000 (20:52 +0000)]
Reviewed by Adam.
Win32 build fix.
* WebCore.vcproj/WebCore/WebCore.vcproj:
* bridge/win/BrowserExtensionWin.cpp:
(WebCore::BrowserExtensionWin::createNewWindow):
* bridge/win/FrameWin.cpp:
(WebCore::FrameWin::urlSelected):
(WebCore::FrameWin::submitForm):
(WebCore::FrameWin::createNewWindow):
* bridge/win/FrameWin.h:
* loader/loader.cpp:
(WebCore::Loader::receivedResponse):
* platform/win/GraphicsContextWin.cpp:
(WebCore::GraphicsContext::getWindowsContext):
(WebCore::GraphicsContext::releaseWindowsContext):
* platform/win/TemporaryLinkStubs.cpp:
(FrameWin::openURLRequest):
(WebCore::ServeSynchronousRequest):
(WebCore::CheckIfReloading):
(WebCore::IsResponseURLEqualToURL):
(WebCore::ResponseMIMEType):
(WebCore::ResponseIsMultipart):
(WebCore::CacheObjectExpiresTime):
(CachedResource::setResponse):
(CachedResource::setAllData):
* rendering/RenderThemeWin.cpp:
(WebCore::doneDrawing):
(WebCore::RenderThemeWin::paintButton):
(WebCore::RenderThemeWin::paintTextField):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17070
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Mon, 16 Oct 2006 08:29:41 +0000 (08:29 +0000)]
Fix for bug 9347, positioned elements have the wrong containing block. This is a change from CSS2 to
CSS2.1. In the absence of any other enclosing positioned elements, the containing block for a positioned
element is the initial containing block (our RenderView).
Reviewed by mjs
Test cases added to fast/block/positioning for quirks and strict modes.
* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::lowestPosition):
(WebCore::RenderBlock::rightmostPosition):
(WebCore::RenderBlock::leftmostPosition):
* rendering/RenderBox.cpp:
(WebCore::RenderBox::containingBlockHeightForPositioned):
(WebCore::RenderBox::calcAbsoluteVertical):
* rendering/RenderObject.cpp:
(WebCore::RenderObject::containingBlock):
(WebCore::RenderObject::container):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17068
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Mon, 16 Oct 2006 08:27:25 +0000 (08:27 +0000)]
Add test cases for bug 9347.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17067
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
bdash [Mon, 16 Oct 2006 01:33:04 +0000 (01:33 +0000)]
2006-10-15 Krzysztof Kowalczyk <kkowalczyk@gmail.com>
Reviewed by Anders.
Remove junk (as gcc calls it) after #else clause.
* wtf/FastMalloc.cpp:
(WTF::do_free):
2006-10-15 Krzysztof Kowalczyk <kkowalczyk@gmail.com>
Reviewed by Anders.
Misc Gdk/Linux build fixes.
* platform/gdk/BrowserExtensionGdk.h:
* platform/gdk/FrameGdk.cpp:
(WebCore::FrameGdk::submitForm):
(WebCore::FrameGdk::urlSelected):
* platform/gdk/FrameGdk.h:
* platform/gdk/RenderPopupMenuGdk.cpp:
(WebCore::PopupMenu::PopupMenu):
(WebCore::PopupMenu::~PopupMenu):
(WebCore::PopupMenu::clear):
(WebCore::PopupMenu::populate):
(WebCore::PopupMenu::show):
(WebCore::PopupMenu::hide):
(WebCore::PopupMenu::addSeparator):
(WebCore::PopupMenu::addGroupLabel):
(WebCore::PopupMenu::addOption):
* platform/gdk/RenderPopupMenuGdk.h:
* platform/gdk/RenderThemeGdk.cpp:
* platform/gdk/RenderThemeGdk.h:
(WebCore::RenderThemeGdk::RenderThemeGdk):
* platform/gdk/TemporaryLinkStubs.cpp:
(BrowserExtensionGdk::createNewWindow):
(BrowserExtensionGdk::BrowserExtensionGdk):
(BrowserExtensionGdk::setTypedIconURL):
(BrowserExtensionGdk::setIconURL):
(BrowserExtensionGdk::getHistoryLength):
(WebCore::CheckIfReloading):
(WebCore::CacheObjectExpiresTime):
(WebCore::ResponseURL):
(WebCore::IsResponseURLEqualToURL):
(WebCore::ResponseIsMultipart):
(WebCore::ResponseMIMEType):
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
(FileChooser::disconnectUploadControl):
* rendering/RenderTheme.h:
* rendering/RenderThemeMac.h:
2006-10-15 Krzysztof Kowalczyk <kkowalczyk@gmail.com>
Reviewed by Anders.
Make compiler not complain about unused gk. SpinneretWebHost was not setting
initial refcount upon creation, so it is completely bogus.
* GdkLauncher/main.cpp:
(main):
* Spinneret/Spinneret/Spinneret.h:
(SpinneretWebHost::SpinneretWebHost):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17065
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Sun, 15 Oct 2006 22:22:59 +0000 (22:22 +0000)]
Reviewed by Mitz.
Fix for http://bugs.webkit.org/show_bug.cgi?id=11294
Add getModifierState method to KeyboardEvent
* bindings/objc/DOMEvents.h:
* bindings/objc/DOMEvents.mm:
* bindings/objc/PublicDOMInterfaces.h:
* dom/KeyboardEvent.cpp:
(WebCore::KeyboardEvent::getModifierState):
* dom/KeyboardEvent.h:
* dom/KeyboardEvent.idl:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17064
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ap [Sun, 15 Oct 2006 19:20:57 +0000 (19:20 +0000)]
2006-10-15 Nikolas Zimmermann <zimmermann@kde.org>
Reviewed and landed by ap.
Changed PlatformResponse handling for Qt. Don't just store
a QString object containing the data but a new PlatformResponseQt struct
containing data and url, and store it as pointer. This is compatible
to the OS X handling and removes the Qt plaform specific hacks from
CachedCSSStyleSheet, CachedResource & loader.
Add new file LoaderFunctionsQt, similar to LoaderFunctionsMac, and move
all methods from TemporaryLinkStubs related to loading in there, and implement
some like ServeSynchronousRequest (yay for xslt handling!).
* CMakeLists.txt:
* loader/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::checkNotify):
* loader/CachedResource.h:
(WebCore::CachedResource::CachedResource):
* loader/loader.cpp:
(WebCore::Loader::receivedResponse):
* platform/ResourceLoaderClient.h:
* platform/ResourceLoaderInternal.h:
* platform/qt/FrameQt.cpp:
(WebCore::FrameQt::~FrameQt):
(WebCore::FrameQt::createEmptyDocument):
* platform/qt/FrameQt.h:
(WebCore::QtFrame):
* platform/qt/FrameQtClient.cpp:
(WebCore::FrameQtClientDefault::setFrame):
(WebCore::FrameQtClientDefault::openURL):
(WebCore::FrameQtClientDefault::loadFinished):
(WebCore::FrameQtClientDefault::receivedAllData):
* platform/qt/FrameQtClient.h:
* platform/qt/LoaderFunctionsQt.cpp: Added.
(WebCore::ServeSynchronousRequest):
(WebCore::NumberOfPendingOrLoadingRequests):
(WebCore::CheckIfReloading):
(WebCore::CheckCacheObjectStatus):
(WebCore::IsResponseURLEqualToURL):
(WebCore::ResponseURL):
(WebCore::ResponseMIMEType):
(WebCore::ResponseIsMultipart):
(WebCore::CacheObjectExpiresTime):
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
* platform/qt/ResourceLoaderManager.cpp:
(WebCore::ResourceLoaderManager::remove):
* platform/qt/ResourceLoaderQt.cpp:
(WebCore::ResourceLoader::~ResourceLoader):
(WebCore::ResourceLoader::start):
(WebCore::ResourceLoader::assembleResponseHeaders):
(WebCore::ResourceLoader::receivedResponse):
* platform/qt/TemporaryLinkStubs.cpp:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17063
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
rwlbuis [Sun, 15 Oct 2006 10:00:13 +0000 (10:00 +0000)]
Reviewed by Mitz.
http://bugs.webkit.org/show_bug.cgi?id=11217
Cleanup svg coding style
Cleanup coding style in svg basic datastructure classes.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17062
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
oliver [Sun, 15 Oct 2006 00:31:05 +0000 (00:31 +0000)]
2006-10-15 Nikolas Zimmermann <zimmermann@kde.org>
Reviewed by Oliver.
Implement some functions in FrameQt / ScrollViewQt.
Added comments what needs to be done in quite some places,
marked some functions as no-ops.
* CMakeLists.txt:
* platform/qt/FrameQt.cpp:
(WebCore::FrameQt::FrameQt):
(WebCore::FrameQt::runJavaScriptConfirm):
(WebCore::FrameQt::locationbarVisible):
(WebCore::FrameQt::createFrame):
(WebCore::FrameQt::passSubframeEventToSubframe):
(WebCore::FrameQt::menubarVisible):
(WebCore::FrameQt::personalbarVisible):
(WebCore::FrameQt::statusbarVisible):
(WebCore::FrameQt::toolbarVisible):
(WebCore::FrameQt::markedTextRange):
(WebCore::FrameQt::lastEventIsMouseUp):
(WebCore::FrameQt::openURLRequest):
(WebCore::FrameQt::scheduleClose):
(WebCore::FrameQt::unfocusWindow):
(WebCore::FrameQt::focusWindow):
(WebCore::FrameQt::overrideMediaType):
(WebCore::FrameQt::runJavaScriptPrompt):
(WebCore::FrameQt::bindingRootObject):
(WebCore::FrameQt::addPluginRootObject):
(WebCore::FrameQt::registerCommandForUndo):
(WebCore::FrameQt::registerCommandForRedo):
(WebCore::FrameQt::clearUndoRedoOperations):
(WebCore::FrameQt::respondToChangedSelection):
(WebCore::FrameQt::respondToChangedContents):
(WebCore::FrameQt::shouldChangeSelection):
(WebCore::FrameQt::canGoBackOrForward):
(WebCore::FrameQt::canPaste):
(WebCore::FrameQt::canRedo):
(WebCore::FrameQt::canUndo):
* platform/qt/FrameQt.h:
* platform/qt/FrameQtClient.cpp:
(WebCore::FrameQtClientDefault::runJavaScriptAlert):
(WebCore::FrameQtClientDefault::runJavaScriptConfirm):
(WebCore::FrameQtClientDefault::runJavaScriptPrompt):
(WebCore::FrameQtClientDefault::menubarVisible):
(WebCore::FrameQtClientDefault::toolbarVisible):
(WebCore::FrameQtClientDefault::statusbarVisible):
(WebCore::FrameQtClientDefault::personalbarVisible):
(WebCore::FrameQtClientDefault::locationbarVisible):
* platform/qt/FrameQtClient.h:
* platform/qt/ImageSourceQt.cpp:
* platform/qt/RenderThemeQt.cpp:
* platform/qt/ScrollViewQt.cpp:
(WebCore::ScrollView::scrollPointRecursively):
* platform/qt/TemporaryLinkStubs.cpp:
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17061
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
oliver [Sun, 15 Oct 2006 00:04:24 +0000 (00:04 +0000)]
2006-10-15 Nikolas Zimmermann <zimmermann@kde.org>
Reviewed by Oliver.
Add another hanging test, to the "to be skipped" list.
* DumpRenderTree/DumpRenderTree.qtproj/tests-skipped.txt:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17060
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
harrison [Sat, 14 Oct 2006 17:38:34 +0000 (17:38 +0000)]
Reviewed by Geoff.
<rdar://problem/
4775213> REGRESSION: Can't from tab from last link on the page to URL address field
<rdar://problem/
4783926> REGRESSION: Typing tab key enters no text
The problem is that EventTargetNode::defaultEventHandler() calls event->setDefaultHandled() regardless
of whether frame->view()->advanceFocus() actually changed the focus.
Trivial fix is to preventDefault() only if advanceFocus() returns true.
* No test is possible because DumpRenderTree sets the WebView preference to
tab to links, and that preference cannot be changed dynamically after
the webview is initialized. Also, the "tab from last link" is not testable
because DumpRenderTree uses the WebView for the whole window, so tabs can't
shift focus out of the WebView regardless.
* dom/EventTargetNode.cpp:
(WebCore::EventTargetNode::defaultEventHandler):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17059
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Sat, 14 Oct 2006 17:21:29 +0000 (17:21 +0000)]
Reviewed by Anders.
Some fixes to get the Qt BuildSlave to run the LayoutTests.
* Scripts/build-dumprendertree: No need to call cmake again.
* Scripts/run-webkit-tests: Expose LD_LIBRARY_PATH.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17058
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
weinig [Sat, 14 Oct 2006 12:53:53 +0000 (12:53 +0000)]
Reviewed by Mitz.
Fix Qt/Linux build & some little fixes:
- Adapt to Maciej's FrameLoadRequest changes.
- Fix response handling - don't try to compare with '0' - for us it's a non-ptr QString object.
- Actually call setAllData from loader/loader.cpp (David Carson forgot that in his last commit)
- Fix CacheObjectExpiresTime/ResponseIsMultipart signature.
* loader/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::checkNotify):
* loader/CachedResource.h:
(WebCore::CachedResource::CachedResource):
* loader/loader.cpp:
(WebCore::Loader::receivedAllData):
(WebCore::Loader::receivedResponse):
* platform/qt/BrowserExtensionQt.cpp:
(WebCore::BrowserExtensionQt::createNewWindow):
* platform/qt/BrowserExtensionQt.h:
* platform/qt/FrameQt.cpp:
(WebCore::FrameQt::submitForm):
(WebCore::FrameQt::urlSelected):
(WebCore::FrameQt::openURLRequest):
* platform/qt/FrameQt.h:
* platform/qt/GraphicsContextQt.cpp:
(WebCore::GraphicsContext::setURLForRect):
* platform/qt/TemporaryLinkStubs.cpp:
(WebCore::CacheObjectExpiresTime):
(WebCore::ResponseIsMultipart):
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17057
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
bdash [Sat, 14 Oct 2006 09:42:25 +0000 (09:42 +0000)]
2006-10-14 David Carson <dacarson@gmail.com>
Reviewed by Maciej.
Fix: http://bugs.webkit.org/show_bug.cgi?id=11257
Removed Mac platform dependance, changed WebCore functions to
use PlatformResponse/PlatformData rather than NSURLResponse/
NSData.
* loader/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::checkNotify):
* loader/CachedImage.cpp:
(WebCore::CachedImage::createImage):
* loader/CachedResource.cpp:
(WebCore::CachedResource::~CachedResource):
* loader/CachedResource.h:
(WebCore::CachedResource::CachedResource):
(WebCore::CachedResource::response):
(WebCore::CachedResource::allData):
* loader/LoaderFunctions.h:
* loader/loader.cpp:
(WebCore::Loader::receivedResponse):
* loader/mac/LoaderFunctionsMac.mm:
(WebCore::ResponseMIMEType):
(WebCore::ResponseIsMultipart):
(WebCore::CacheObjectExpiresTime):
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
* platform/qt/TemporaryLinkStubs.cpp:
(WebCore::CacheObjectExpiresTime):
(WebCore::ResponseIsMultipart):
(WebCore::ResponseMIMEType):
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
* platform/win/TemporaryLinkStubs.cpp:
(WebCore::CacheObjectExpiresTime):
(WebCore::ResponseIsMultipart):
(WebCore::ResponseMIMEType):
(WebCore::CachedResource::setResponse):
(WebCore::CachedResource::setAllData):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17056
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
bdash [Sat, 14 Oct 2006 09:39:55 +0000 (09:39 +0000)]
2006-10-14 Krzysztof Kowalczyk <kkowalczyk@gmail.com>
Reviewed by Maciej.
Define KXMLCORE_USE_CURL for platforms that wish to use CURL as
networking, and set it for GDK build
* wtf/Platform.h:
2006-10-14 Krzysztof Kowalczyk <kkowalczyk@gmail.com>
Reviewed by Maciej.
Fix GDK build by defining FrameGdkClient() constructor as non-virtual.
Pure virtual classes require non-pure destructor, who knew.
Make CURL networking dependent on USE(CURL) instead of PLATFORM(GDK),
so that other platforms can also use CURL, if they wish.
Silence compiler warning about unused variable in SharedTimerLinux.cpp
* platform/ResourceLoaderInternal.h:
(WebCore::ResourceLoaderInternal::ResourceLoaderInternal):
* platform/gdk/FrameGdk.h:
(WebCore::FrameGdkClient::~FrameGdkClient):
* platform/gdk/ResourceLoaderManager.cpp:
(WebCore::ResourceLoaderManager::downloadTimerCallback):
* platform/gdk/SharedTimerLinux.cpp:
(WebCore::stopSharedTimer):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17055
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
bdakin [Sat, 14 Oct 2006 02:44:11 +0000 (02:44 +0000)]
Reviewed by Geoff.
Build fix from my last check-in. Make hasOutlineAnnotation() non-
inline so that we don't have to include Document.h in
RenderObject.h
Doing so causes a build failure in WebKit, and it probably isn't a
good idea anyway.
* rendering/RenderObject.cpp:
(WebCore::RenderObject::hasOutlineAnnotation):
* rendering/RenderObject.h:
(WebCore::RenderObject::hasOutline):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17054
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Sat, 14 Oct 2006 01:37:15 +0000 (01:37 +0000)]
Reviewed by Beth.
Split build, run, and debug steps into separate pages. Darin recommended
this a while back.
* building/build.html:
* building/checkout.html:
* building/debug.html: Added.
* building/run.html: Added.
* nav.inc:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17053
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
bdakin [Sat, 14 Oct 2006 01:12:32 +0000 (01:12 +0000)]
Patch written mostly by Ken Kraisler, but also by me.
Reviewed by Hyatt.
Fix for http://bugs.webkit.org/show_bug.cgi?id=10216 and <rdar://
problem/
3391162> PDF created by printing should have live
hyperlinks
* platform/GraphicsContext.cpp:
(WebCore::GraphicsContext::focusRingBoundingRect):
* platform/GraphicsContext.h:
* platform/cairo/GraphicsContextCairo.cpp:
(WebCore::GraphicsContext::setURLForRect):
* platform/cg/GraphicsContextCG.cpp:
(WebCore::GraphicsContext::setURLForRect): Implement method to add
URL link to PDF document.
* rendering/InlineFlowBox.cpp:
(WebCore::InlineFlowBox::paint): Ask hasOutline() instead of
querying the outline width
* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::paintObject): Same as above.
* rendering/RenderFlow.cpp:
(WebCore::RenderFlow::paintLines): Call the new paintOutline()
(WebCore::RenderFlow::paintOutline): Take care of focus ring and
pdf url rects, and outline painting.
* rendering/RenderFlow.h:
* rendering/RenderObject.cpp:
(WebCore::RenderObject::addPDFURLRect): Declaration to apply a PDF
link to a rectanglular region.
(WebCore::RenderObject::paintOutline): Take care of pdf rects as
well as focus ring painting.
* rendering/RenderObject.h:
(WebCore::RenderObject::hasOutlineAnnotation): Returns true is the
element is a link and we are printing.
(WebCore::RenderObject::hasOutline): Returns true is the style has
an outline and hasOutlineAnnotation() is true.
* rendering/RenderStyle.h:
(WebCore::RenderStyle::hasOutline): Returns true if outlineWidth is
greater than 0 and outlineStyle is greater than BHIDDEN.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17052
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
justing [Fri, 13 Oct 2006 23:58:37 +0000 (23:58 +0000)]
LayoutTests:
Reviewed by ggaren and harrison
<rdar://problem/
4744008>
9A270: Mail crashes when I try to paste large items from Safari
* editing/pasteboard/
4744008-expected.txt: Added.
* editing/pasteboard/
4744008.html: Added.
WebCore:
Reviewed by ggaren and harrison
<rdar://problem/
4744008>
9A270: Mail crashes when I try to paste large items from Safari
* editing/ReplaceSelectionCommand.cpp:
(WebCore::ReplaceSelectionCommand::removeRedundantStyles):
Even though we put nodes in the hash map in pre-order, they don't necessary
come out of the iterator that way. If a node is a redundant style span
and one of its pruned ancestors comes off the hash map after it, and that
ancestor is a redundant style span or font tag, we will try to remove it
even though it has already been removed and we'll crash.
The test case added with the fix depends on our hash map implementation
because it requires that a particular redundant style span come out of
the iterator before its parent. So, it may eventually not provide
coverage for the fix.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17051
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
mjs [Fri, 13 Oct 2006 23:54:52 +0000 (23:54 +0000)]
Reviewed by Darin.
- split FrameLoadRequest off from ResourceRequest - it's a ResourceRequest plus frame name plus lockHistory bool
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/kjs_window.cpp:
(KJS::createNewWindow):
(KJS::WindowFunc::callAsFunction):
* bridge/BrowserExtension.h:
* bridge/mac/BrowserExtensionMac.h:
* bridge/mac/BrowserExtensionMac.mm:
(WebCore::BrowserExtensionMac::createNewWindow):
* bridge/mac/FrameMac.h:
* bridge/mac/FrameMac.mm:
(WebCore::FrameMac::submitForm):
(WebCore::FrameMac::openURLRequest):
(WebCore::FrameMac::urlSelected):
* bridge/win/BrowserExtensionWin.h:
* page/Frame.cpp:
(WebCore::Frame::changeLocation):
(WebCore::Frame::urlSelected):
(WebCore::Frame::requestFrame):
(WebCore::Frame::submitForm):
* page/Frame.h:
* page/FrameLoadRequest.h: Added.
(WebCore::FrameLoadRequest::FrameLoadRequest):
(WebCore::FrameLoadRequest::lockHistory):
(WebCore::FrameLoadRequest::setLockHistory):
* page/ResourceRequest.h:
(WebCore::ResourceRequest::ResourceRequest):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17050
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
sullivan [Fri, 13 Oct 2006 23:50:05 +0000 (23:50 +0000)]
Reviewed by Maciej
* loader/mac/WebFrameLoader.mm:
(-[WebFrameLoader loadDocumentLoader:]):
Replace ASSERT(!policyDocumentLoader) with code that handles this case. This can
happen when loading a page during a decidePolicyForNavigationAction callback.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17049
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
mjs [Fri, 13 Oct 2006 23:45:51 +0000 (23:45 +0000)]
Not reviewed, build fix.
* icu/unicode/putil.h: Added - needed for build if you don't have apple internal headers.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17048
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
mjs [Fri, 13 Oct 2006 23:16:22 +0000 (23:16 +0000)]
Not reviewed, build fix.
* icu/unicode/ustring.h: Added - needed for build if you don't have apple internal headers.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17047
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
hyatt [Fri, 13 Oct 2006 22:09:24 +0000 (22:09 +0000)]
Transparency layer work.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17046
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
mjs [Fri, 13 Oct 2006 21:42:36 +0000 (21:42 +0000)]
Reviewed by Anders.
- add a bunch of casts to get this compiling with older Xcode versions
(I used static_cast so it will be easier to find and remove these once we have completely moved
on to a new enough compiler version.)
* Plugins/WebBaseNetscapePluginView.m:
(+[WebBaseNetscapePluginView getCarbonEvent:]):
(-[WebBaseNetscapePluginView getCarbonEvent:withEvent:]):
(-[WebBaseNetscapePluginView fixWindowPort]):
(-[WebBaseNetscapePluginView saveAndSetNewPortStateForUpdate:]):
(-[WebBaseNetscapePluginView _createWindowlessAGLContext]):
(-[WebBaseNetscapePluginView _reshapeAGLWindow]):
(-[WebBaseNetscapePluginView _aglOffscreenImageForDrawingInRect:]):
* WebKit.xcodeproj/project.pbxproj:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17044
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
rwlbuis [Fri, 13 Oct 2006 21:04:55 +0000 (21:04 +0000)]
Reviewed by Mitz.
http://bugs.webkit.org/show_bug.cgi?id=11217
Cleanup svg coding style
Cleanup coding style in svg paint server classes.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17041
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
ggaren [Fri, 13 Oct 2006 20:22:08 +0000 (20:22 +0000)]
LayoutTests:
Tests for accessing renderer-dependent properties from a javascript: URL.
* fast/frames/frame-js-url-clientWidth-expected.txt: Added.
* fast/frames/frame-js-url-clientWidth.html: Added.
* fast/frames/iframe-js-url-clientWidth-expected.txt: Added.
* fast/frames/iframe-js-url-clientWidth.html: Added.
WebCore:
Reviewed by Darin, John.
- Merged more frame and iframe code
- Fixed a bug where iframes returned incorrect values for renderer-dependent
properties during javascript: loads because they didn't have renderers at
load time
PLT insists this is a small performance win. Don't believe its lies.
* bindings/js/kjs_html.cpp:
(KJS::JSHTMLElement::frameGetter): Renamed frameWidth and frameHeight to
width and height, for consistency with the rest of the DOM.
* dom/ContainerNode.cpp:
(WebCore::ContainerNode::queuePostAttachCallback): Added a post-attach callback
queue that gets drained after the render tree is fully constructed. Currently,
this is only used for frame loading.
(WebCore::ContainerNode::attach):
* dom/ContainerNode.h:
* html/HTMLFrameElement.cpp:
(WebCore::HTMLFrameElement::insertedIntoDocument): Removed frame loading from
attach() because loading iframes during attach() causes a crash.
Moved frame loading logic into insertedIntoDocument(). That's a more
logical place for it because document insertion is what
triggers frame loading. Made frame loading a post-attach callback, instead
of an immediate action, to fix the incorrect values bug.
(WebCore::HTMLFrameElement::attach): Added setWidget call that used to be
in HTMLIFrameElement::attach. IFRAME requires this call. FRAME will soon
require this call, once I remove the call from WebKit.
(WebCore::HTMLFrameElement::setLocation):
(WebCore::HTMLFrameElement::width):
(WebCore::HTMLFrameElement::height):
* html/HTMLFrameElement.h: Made openURL non-virtual to avoid the unnecessary
killing of puppies.
* html/HTMLIFrameElement.cpp: Merged code into HTMLFrameElement
(WebCore::HTMLIFrameElement::rendererIsNeeded):
(WebCore::HTMLIFrameElement::createRenderer):
(WebCore::HTMLIFrameElement::insertedIntoDocument):
(WebCore::HTMLIFrameElement::removedFromDocument):
(WebCore::HTMLIFrameElement::attach):
* page/FrameView.h: Removed unused method.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17040
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
rwlbuis [Fri, 13 Oct 2006 19:21:31 +0000 (19:21 +0000)]
Reviewed by Eric.
Force --no-http on Qt/Linux.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17039
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
rwlbuis [Fri, 13 Oct 2006 19:10:47 +0000 (19:10 +0000)]
Add PopupMenuQt.cpp, remove RenderPopupMenuQt.*, as per WildFox' request.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17038
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Fri, 13 Oct 2006 18:56:41 +0000 (18:56 +0000)]
Reviewed by Kevin McCullough.
Fixes http://bugs.webkit.org/show_bug.cgi?id=11283
Fixes Qt/Linux and Windows build
* kjs/DateMath.cpp:
* kjs/DateMath.h:
* kjs/date_object.cpp:
(KJS::DateProtoFunc::callAsFunction):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17037
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
harrison [Fri, 13 Oct 2006 18:17:52 +0000 (18:17 +0000)]
LayoutTests:
Reviewed by Justin.
<rdar://problem/
4746860> Mail crash in WebCore::CSSMutableStyleDeclaration::merge(WebCore::CSSMutableStyleDeclaration*, bool)
* editing/style/style-boundary-005.html: Added. Matches steps from bug.
* editing/execCommand/hilitecolor-expected.txt: Updated (no empty Apple style spans).
* editing/style/remove-underline-across-paragraph-in-bold-expected.txt: Updated (no empty Apple style spans).
* editing/style/remove-underline-in-bold-expected.txt: Updated (no empty Apple style spans).
* editing/style/typing-style-003-expected.txt: Updated (no empty Apple style spans).
WebCore:
Reviewed by Justin.
<rdar://problem/
4746860> Mail crash in WebCore::CSSMutableStyleDeclaration::merge(WebCore::CSSMutableStyleDeclaration*, bool)
Tests added or changed:
* editing/style/style-boundary-005.html: Added. Matches steps from bug.
* editing/execCommand/hilitecolor-expected.txt: Updated (no empty Apple style spans).
* editing/style/remove-underline-across-paragraph-in-bold-expected.txt: Updated (no empty Apple style spans).
* editing/style/remove-underline-in-bold-expected.txt: Updated (no empty Apple style spans).
* editing/style/typing-style-003-expected.txt: Updated (no empty Apple style spans).
* dom/Text.cpp:
(WebCore::Text::splitText):
Comment changes.
* editing/ApplyStyleCommand.cpp:
(WebCore::isUnstyledStyleSpan):
Renamed from isEmptyStyleSpan.
(WebCore::ApplyStyleCommand::applyRelativeFontStyleChange):
Rename emptySpans local to unstyledSpans.
(WebCore::dummySpanAncestorForNode):
New. To locate Apple style span created when splitting a text node.
(WebCore::ApplyStyleCommand::cleanupUnstyledAppleStyleSpans):
Renamed from cleanUpEmptyStyleSpans. Algorithm changed to scan children
of dummy span's parents, rather than traversing up from start and
end locations. Old algorithm would miss empty spans in the bug case.
(WebCore::ApplyStyleCommand::applyInlineStyle):
Use dummySpanAncestorForNode() and cleanupUnstyledAppleStyleSpans().
(WebCore::ApplyStyleCommand::removeCSSStyle):
Renamed isEmptyStyleSpan() to isUnstyledStyleSpan().
(WebCore::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
Comment changes.
* editing/ApplyStyleCommand.h:
Renamed cleanUpEmptyStyleSpans() to cleanupUnstyledAppleStyleSpans().
* editing/ReplaceSelectionCommand.cpp:
(WebCore::ReplaceSelectionCommand::removeRedundantStyles):
Allow for Apple style span to be unstyled, because old email may have this in the DOM.
Use !node->hasChildNodes() instead of !node->firstChild().
* editing/SplitTextNodeCommand.cpp:
(WebCore::SplitTextNodeCommand::doApply):
Comment changes.
* rendering/RenderTreeAsText.cpp:
(WebCore::isEmptyOrUnstyledAppleStyleSpan):
(WebCore::operator<<):
Flag empty or unstyled AppleStyleSpa nbecause we never want to leave them in the DOM
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17036
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
rwlbuis [Fri, 13 Oct 2006 18:14:35 +0000 (18:14 +0000)]
Adding a file that should have gone in in r17021, as requested by WildFox.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17035
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Fri, 13 Oct 2006 18:08:18 +0000 (18:08 +0000)]
Reviewed by Eric.
Fix Qt/Linux build after Adele's popup menu changes.
* CMakeLists.txt:
* platform/qt/FileChooserQt.cpp:
(WebCore::FileChooser::disconnectUploadControl):
* platform/qt/PopupMenuQt.cpp: Added.
(WebCore::PopupMenu::PopupMenu):
(WebCore::PopupMenu::~PopupMenu):
(WebCore::PopupMenu::clear):
(WebCore::PopupMenu::populate):
(WebCore::PopupMenu::show):
(WebCore::PopupMenu::hide):
(WebCore::PopupMenu::addSeparator):
(WebCore::PopupMenu::addGroupLabel):
(WebCore::PopupMenu::addOption):
* platform/qt/RenderPopupMenuQt.cpp: Removed.
* platform/qt/RenderPopupMenuQt.h: Removed.
* platform/qt/RenderThemeQt.cpp:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17034
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Fri, 13 Oct 2006 17:57:23 +0000 (17:57 +0000)]
Reviewed by Adam.
Gets JavaScripCore tests running on windows.
* Scripts/run-javascriptcore-tests:
* Scripts/webkitdirs.pm:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17033
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Fri, 13 Oct 2006 17:48:54 +0000 (17:48 +0000)]
Reviewed by Darin.
fixed build failure
* WebCore.xcodeproj/project.pbxproj:
* loader/mac/WebFrameLoader.mm:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17032
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Fri, 13 Oct 2006 17:31:11 +0000 (17:31 +0000)]
Reviewed by Adam, Geoff, Darin.
Fixed displaying the UTC offset and time zone string, as well as renamed the GregorianDateTime structure and clean up.
* ChangeLog:
* kjs/DateMath.cpp:
(KJS::getUTCOffset):
(KJS::getDSTOffsetSimple):
(KJS::gregorianDateTimeToMS):
(KJS::msToGregorianDateTime):
* kjs/DateMath.h:
(KJS::GregorianDateTime::GregorianDateTime):
(KJS::GregorianDateTime::~GregorianDateTime):
(KJS::GregorianDateTime::toTM):
* kjs/date_object.cpp:
(KJS::gmtoffset):
(KJS::formatDate):
(KJS::formatDateUTCVariant):
(KJS::formatTime):
(KJS::fillStructuresUsingTimeArgs):
(KJS::fillStructuresUsingDateArgs):
(KJS::DateInstance::getTime):
(KJS::DateInstance::getUTCTime):
(KJS::DateProtoFunc::callAsFunction):
(KJS::DateObjectImp::construct):
(KJS::DateObjectImp::callAsFunction):
(KJS::DateObjectFuncImp::callAsFunction):
(KJS::parseDate):
* kjs/date_object.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17031
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
kmccullo [Fri, 13 Oct 2006 17:22:26 +0000 (17:22 +0000)]
Changed by Darin, reviewed by me.
* Plugins/WebNetscapePluginStream.m: Fixed case of import so we can compile on
case-sensitive file system.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17030
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
darin [Fri, 13 Oct 2006 17:07:20 +0000 (17:07 +0000)]
WebCore:
Reviewed by Adele.
- converted WebFormState from Objective-C to C++
* WebCore.exp: Added a new symbol, changed one symbol from C to C++.
* WebCore.xcodeproj/project.pbxproj: Made lots of headers "private" instead of "project".
Added UChar.h and FrameLoaderTypes.h.
* loader/FrameLoaderTypes.h: Added. Has enums from WebFrameLoader.h.
* loader/mac/WebFormState.h: Converted class to C++.
* loader/mac/WebFormState.mm: Ditto.
* loader/mac/WebFrameLoader.h:
* loader/mac/WebFrameLoader.mm:
(-[WebFrameLoader loadURL:referrer:loadType:target:triggeringEvent:form:formValues:]):
(-[WebFrameLoader continueFragmentScrollAfterNavigationPolicy:formState:]):
(-[WebFrameLoader _loadRequest:triggeringAction:loadType:formState:]):
(-[WebFrameLoader invalidatePendingPolicyDecisionCallingDefaultAction:]):
(-[WebFrameLoader checkNewWindowPolicyForRequest:action:frameName:formState:andCall:withSelector:]):
(-[WebFrameLoader continueAfterNewWindowPolicy:]):
(-[WebFrameLoader checkNavigationPolicyForRequest:documentLoader:formState:andCall:withSelector:]):
(-[WebFrameLoader continueAfterNavigationPolicy:]):
(-[WebFrameLoader continueLoadRequestAfterNavigationPolicy:formState:]):
(-[WebFrameLoader loadDocumentLoader:withLoadType:formState:]):
(-[WebFrameLoader continueLoadRequestAfterNewWindowPolicy:frameName:formState:]):
(-[WebFrameLoader postWithURL:referrer:target:data:contentType:triggeringEvent:form:formValues:]):
Changed Objective-C WebFormState uses to C++ WebCore::FormState.
WebKit:
Reviewed by Adele.
- converted WebFormState from Objective-C to C++
* ForwardingHeaders: Added an entire copy of WebCore's forwarding headers here.
We should eventually come up with a more-elegant solution.
* WebKit.xcodeproj/project.pbxproj: Added ForwardingHeaders to the include paths.
Converted many files from Objective-C to Objective-C++. In a later check-in, I'll
rename them to .mm instead of .m. Removed C-only warning options for now. In a
later check-in I will add these back in a way that omits them for C++.
* Plugins/WebPluginContainerCheck.m: Updated for header changes.
* WebView/WebFrameInternal.h: Updated for header changes. Removed WebFrameLoaderClient category
so this file can still be used by Objective-C code (not just Objective-C++).
* WebView/WebFrame.m: Put WebFrameLoaderClient category in here.
(-[WebFrame _loadItem:withLoadType:]): Changed to use 0 instead of nil for FormState and fixed
enum code for C++ compatibility.
(-[WebFrame _initWithWebFrameView:webView:bridge:]): Added call to setFrameLoaderClient: here.
(-[WebFrame _updateHistoryForCommit]): Fixed enum code for C++ compatibility.
(-[WebFrame _updateHistoryForReload]): Ditto.
* WebCoreSupport/WebFrameBridge.m:
(-[WebFrameBridge initMainFrameWithPage:frameName:view:]): Removed call to setFrameLoaderClient:.
(-[WebFrameBridge initSubframeWithOwnerElement:frameName:view:]): Ditto.
(-[WebFrameBridge _retrieveKeyboardUIModeFromPreferences:]): Fixed enum code for C++ compatibility.
(-[WebFrameBridge runModal]): Changed code to not use "namespace" as a local variable name.
* WebView/WebPDFView.m: Added extern "C" so this can compile as Objective-C++.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17029
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Fri, 13 Oct 2006 14:36:26 +0000 (14:36 +0000)]
Reviewed by Maciej.
- Fix for <rdar://problem/
4768591> REGRESSION: List box selection color should just be a darker version of regular selection color
* rendering/RenderListBox.cpp:
(WebCore::RenderListBox::paintItemForeground):
* rendering/RenderTheme.cpp:
(WebCore::RenderTheme::activeListBoxSelectionBackgroundColor):
(WebCore::RenderTheme::activeListBoxSelectionForegroundColor):
(WebCore::RenderTheme::inactiveListBoxSelectionBackgroundColor):
(WebCore::RenderTheme::inactiveListBoxSelectionForegroundColor):
* rendering/RenderThemeMac.h:
* rendering/RenderThemeMac.mm: (WebCore::RenderThemeMac::activeListBoxSelectionBackgroundColor):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17028
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
rwlbuis [Fri, 13 Oct 2006 12:56:42 +0000 (12:56 +0000)]
Reviewed by Mitz.
http://bugs.webkit.org/show_bug.cgi?id=11217
Cleanup svg coding style
Cleanup coding style in svg structure classes.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17027
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Fri, 13 Oct 2006 06:33:48 +0000 (06:33 +0000)]
LayoutTests:
Reviewed by Anders.
- Test for http://bugs.webkit.org/show_bug.cgi?id=11172
REGRESSION (NativeListBox): OPTION text paints over the edge of the select element at http://jobs.apple.com/
* fast/forms/listbox-width-change-expected.txt: Added.
* fast/forms/listbox-width-change-expected.png: Added.
* fast/forms/listbox-width-change-expected.checksum: Added.
* fast/forms/listbox-width-change.html: Added.
WebCore:
Reviewed by Anders.
- Fix for http://bugs.webkit.org/show_bug.cgi?id=11172
REGRESSION (NativeListBox): OPTION text paints over the edge of the select element at http://jobs.apple.com/
Test: fast/forms/listbox-width-change.html
* rendering/RenderListBox.cpp: (WebCore::RenderListBox::updateFromElement):
When the options change, call setNeedsLayoutAndMinMaxRecalc.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17026
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
mjs [Fri, 13 Oct 2006 05:22:05 +0000 (05:22 +0000)]
WebCore:
Reviewed by Oliver.
- moved a bunch of WebFrameBridge code down to WebCoreFrameBridge, now that it no longer has any WebKit dependencies
* WebCore.xcodeproj/project.pbxproj:
* bridge/mac/WebCoreFrameBridge.h:
* bridge/mac/WebCoreFrameBridge.mm:
(stringByCollapsingNonPrintingCharacters):
(-[WebCoreFrameBridge setTitle:]):
(-[WebCoreFrameBridge didFirstLayout]):
(-[WebCoreFrameBridge notifyIconChanged:]):
(-[WebCoreFrameBridge originalRequestURL]):
(-[WebCoreFrameBridge isLoadTypeReload]):
(-[WebCoreFrameBridge frameDetached]):
(-[WebCoreFrameBridge tokenizerProcessedData]):
(-[WebCoreFrameBridge receivedData:textEncodingName:]):
(-[WebCoreFrameBridge startLoadingResource:withMethod:URL:customHeaders:]):
(-[WebCoreFrameBridge objectLoadedFromCacheWithURL:response:data:]):
(-[WebCoreFrameBridge startLoadingResource:withMethod:URL:customHeaders:postData:]):
(-[WebCoreFrameBridge reportClientRedirectToURL:delay:fireDate:lockHistory:isJavaScriptFormAction:]):
(-[WebCoreFrameBridge reportClientRedirectCancelled:]):
(-[WebCoreFrameBridge loadURL:referrer:reload:userGesture:target:triggeringEvent:form:formValues:]):
(-[WebCoreFrameBridge postWithURL:referrer:target:data:contentType:triggeringEvent:form:formValues:]):
(-[WebCoreFrameBridge syncLoadResourceWithMethod:URL:customHeaders:postData:finalURL:responseHeaders:statusCode:]):
(-[WebCoreFrameBridge incomingReferrer]):
(-[WebCoreFrameBridge isReloading]):
(-[WebCoreFrameBridge handledOnloadEvents]):
(-[WebCoreFrameBridge mainResourceURLResponse]):
(-[WebCoreFrameBridge loadEmptyDocumentSynchronously]):
* loader/mac/LoaderNSURLExtras.h:
* loader/mac/LoaderNSURLRequestExtras.h: Added.
* loader/mac/LoaderNSURLRequestExtras.m: Added.
(isConditionalRequest):
(setHTTPReferrer):
* loader/mac/WebFrameLoader.mm:
* loader/mac/WebSubresourceLoader.mm:
WebKit:
Reviewed by Oliver.
- liberate more WebKit code down to WebCore
* WebCoreSupport/WebFrameBridge.h:
* WebCoreSupport/WebFrameBridge.m:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17025
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Fri, 13 Oct 2006 02:28:20 +0000 (02:28 +0000)]
RS by Geoff.
Renamed menuListIsDetaching to disconnectMenuList,
uploadControlDetaching to disconnectUploadControl, and fileChooserDetachingSoon to disconnectFileChooser.
These names match similar methods in WebCore a little better.
* platform/FileChooser.h:
* platform/PopupMenu.h: (WebCore::PopupMenu::disconnectMenuList):
* platform/mac/FileChooserMac.mm: (-[OpenPanelController disconnectFileChooser]):
(WebCore::FileChooser::disconnectUploadControl):
* platform/win/TemporaryLinkStubs.cpp: (FileChooser::disconnectUploadControl):
* rendering/RenderFileUploadControl.cpp: (WebCore::RenderFileUploadControl::~RenderFileUploadControl):
* rendering/RenderMenuList.cpp: (WebCore::RenderMenuList::~RenderMenuList):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17024
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Fri, 13 Oct 2006 02:13:57 +0000 (02:13 +0000)]
WebCore:
Reviewed by Maciej.
WebCore part of fix for <rdar://problem/
4450613> need a means to attach user data to any menu that is popuped up in HTML
Added private delegate method for clients that want access to a PopupMenu's NSMenu.
* bridge/mac/FrameMac.h:
* bridge/mac/FrameMac.mm: (WebCore::FrameMac::willPopupMenu):
* bridge/mac/WebCoreFrameBridge.h:
* platform/mac/PopupMenuMac.mm: (WebCore::PopupMenu::show): Calls willPopupMenu with the NSMenu about to be popped up.
WebKit:
Reviewed by Maciej.
WebKit part of fix for <rdar://problem/
4450613> need a means to attach user data to any menu that is popuped up in HTML
Added private delegate method for clients that want access to a PopupMenu's NSMenu.
* DefaultDelegates/WebDefaultUIDelegate.m: (-[NSApplication webView:willPopupMenu:]):
* WebCoreSupport/WebFrameBridge.m: (-[WebFrameBridge willPopupMenu:]):
* WebView/WebUIDelegatePrivate.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17023
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
adele [Fri, 13 Oct 2006 01:39:14 +0000 (01:39 +0000)]
Reviewed by Alice.
- Fix for <rdar://problem/
4707372>
After timers fix, crash in -[NSMenu performActionForItemAtIndex:] after moving/detroying active popup menu
* platform/PopupMenu.h: (WebCore::PopupMenu::menuListIsDetaching): Added. Zeros out the pointer to the menu list.
* rendering/RenderMenuList.cpp: (WebCore::RenderMenuList::~RenderMenuList): Calls menuListIsDetaching.
* platform/mac/PopupMenuMac.mm: (WebCore::PopupMenu::show): Added a RefPtr to protect the PopupMenu while the NSMenu is up,
in case the RenderMenuList goes away. Also, checks the pointer to the menu list immediately afterwards.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@17022
268f45cc-cd09-0410-ab3c-
d52691b4dbfc