⚠ Archived content — this site is no longer maintained.   Current WebKit documentation is at docs.webkit.org.

Timeline



Nov 29, 2007:

11:01 PM Changeset in webkit [28228] by beidson@apple.com
  • 3 edits in trunk/WebCore

Reviewed by Tim

Part of <rdar://problem/5556377> - Need to accurately track and enforce database quota

SQLite has the ability to enforce the page usage for databases. This gives us bulletproof
enforcement of the database size limit with reasonable granularity (within 1023 bytes by
default on OS X with SQLite 3.4.0)

This also involved enhancing the ability to install/remove the authorizer to run the PRAGMA
statements involved

Note this patch does not actually use the new maximumSize() functionality of the
SQLiteDatabase class - that can be reviewed/landed seperately


  • platform/sql/SQLiteDatabase.cpp: (WebCore::SQLiteDatabase::SQLiteDatabase): (WebCore::SQLiteDatabase::maximumSize): Return the current maximum size in bytes (WebCore::SQLiteDatabase::setMaximumSize): Based on m_pageSize, set the maximum page count to enforce the maximum size in bytes (WebCore::SQLiteDatabase::pageSize): Fetch m_pageSize if it hasn't been fetched, or return the cached value (WebCore::SQLiteDatabase::setAuthorizer): Use enableAuthorizer for the sqlite_* level authorizer setup (WebCore::SQLiteDatabase::enableAuthorizer): Install or remove the sqlite authorizer handler
  • platform/sql/SQLiteDatabase.h: Add the m_pageSize member so we only have to fetch it once
9:50 PM Changeset in webkit [28227] by zecke@webkit.org
  • 4 edits in trunk
  • Fix fast/events/mousedown_in_scrollbar.html
  • This test used to crash in PlatformScrollbarQt::mouse*Event due parent() returning 0.
  • Extend the current check that the scrollbar in the mev can be go away. This happens if we destroy a RenderLayer as the result of a click. In the case of Qt we can have this scrollbar in m_lastScrollbarUnderMouse. The RenderLayer is the ScrollbarClient, m_lastScrollbarUnderMouse is a RefPtr and keeps the last reference. When calling into this scrollbar the m_client pointer is a dangling pointer. We extend the check to set m_lastScrollbarUnderMouse back to zero when it was the scrollbar in the old mev and is gone in the new one.
  • This happens only on the Qt port because we send more/different move events.
9:25 PM Changeset in webkit [28226] by Antti Koivisto
  • 5 edits
    5 adds in trunk

WebCore:

Reviewed by Mitz.

Fix aspect ratio calculation of video.


Video should always be treated as a replaced element in size calculation.


Test: media/video-aspect-ratio.html

  • rendering/RenderBox.cpp: (WebCore::RenderBox::calcWidth): (WebCore::RenderBox::calcHeight):
  • rendering/RenderBox.h: (WebCore::RenderBox::shouldCalculateSizeAsReplaced):
  • rendering/RenderMedia.h: (WebCore::RenderMedia::shouldCalculateSizeAsReplaced):

LayoutTests:

Reviewed by Mitz.


Test for video aspect ratio calculation.

  • media/video-aspect-ratio.html: Added.
  • platform/mac/media: Added.
  • platform/mac/media/video-aspect-ratio-expected.checksum: Added.
  • platform/mac/media/video-aspect-ratio-expected.png: Added.
  • platform/mac/media/video-aspect-ratio-expected.txt: Added.
9:20 PM Changeset in webkit [28225] by weinig@apple.com
  • 5 edits in trunk/WebCore

Reviewed by Oliver.

Additional fix for <rdar://problem/5592988> / ​http://bugs.webkit.org/show_bug.cgi?id=15936

  • More closely match IE's policy for frame navigation.
  • bindings/js/kjs_window.cpp: (KJS::WindowProtoFuncOpen::callAsFunction):
  • loader/FrameLoader.cpp: (WebCore::FrameLoader::shouldAllowNavigation):
  • page/FrameTree.cpp: (WebCore::FrameTree::top):
  • page/FrameTree.h:
8:34 PM Changeset in webkit [28224] by mitz@apple.com
  • 7 edits in trunk/WebCore

Reviewed by Adam Roben and John Sullivan.

  • make the recent searches popup extend only along the straight part of the search field (unless it needs to be wider).
  • platform/PopupMenuClient.h: Added clientInsetLeft() and clientInsetRight() for clients to specify desired horizontal insets for the menu.
  • platform/win/PopupMenuWin.cpp: (WebCore::PopupMenu::calculatePositionAndSize): (WebCore::PopupMenu::paint):
  • rendering/RenderMenuList.cpp: (WebCore::RenderMenuList::clientInsetLeft): Added. Returns 0. (WebCore::RenderMenuList::clientInsetRight): Ditto.
  • rendering/RenderMenuList.h:
  • rendering/RenderTextControl.cpp: (WebCore::RenderTextControl::clientInsetLeft): Added. Returns half the control's height, which is the radius of the cap on the left. (WebCore::RenderTextControl::clientInsetRight): Added.
  • rendering/RenderTextControl.h:
7:02 PM Changeset in webkit [28223] by oliver@apple.com
  • 2 edits in trunk/JavaScriptCore

Make the JS collector work with multiple threads

Reviewed by Maciej and Darin.

Under heavy contention it was possible the GC to suspend other
threads inside the pthread spinlock, which could lead to the GC
thread blocking on the pthread spinlock itself.

We now determine and store each thread's stack base when it is
registered, thus removing the need for any calls to pthread_get_stackaddr_np
that needed the pthread spinlock.

6:42 PM Changeset in webkit [28222] by andersca@apple.com
  • 6 edits in trunk

WebKitTools:

Reviewed by Sam.

<rdar://problem/5230478>
FrameLoadDelegate callbacks are not dumped in DRT.


  • DumpRenderTree/win/DumpRenderTree.cpp: (shouldLogFrameLoadDelegates): (runTest): (main):
  • DumpRenderTree/win/FrameLoaderDelegate.cpp: (BSTRtoString): (descriptionSuitableForTestResult): (FrameLoadDelegate::QueryInterface): (FrameLoadDelegate::didStartProvisionalLoadForFrame): (FrameLoadDelegate::didFailProvisionalLoadWithError): (FrameLoadDelegate::didCommitLoadForFrame): (FrameLoadDelegate::didFinishLoadForFrame): (FrameLoadDelegate::willCloseFrame): (FrameLoadDelegate::didClearWindowObject): (FrameLoadDelegate::didFinishDocumentLoadForFrame): (FrameLoadDelegate::didHandleOnloadEventsForFrame):
  • DumpRenderTree/win/FrameLoaderDelegate.h: (FrameLoadDelegate::windowScriptObjectAvailable): (FrameLoadDelegate::didFirstLayoutInFrame):

LayoutTests:

Reviewed by Sam.

<rdar://problem/5230478>
FrameLoadDelegate callbacks are not dumped in DRT.

Remove now passing tests.

  • platform/win/Skipped:
6:38 PM Changeset in webkit [28221] by andersca@apple.com
  • 3 edits in trunk/WebKit/win

Reviewed by Sam.

Add an IWebFrameLoadDelegate2 interface with a didClearWindowObject method
that passes in the frame whose window has been cleared. This matches a newly
added delegate method on the Mac.


  • Interfaces/IWebFrameLoadDelegate.idl:
  • WebFrame.cpp: (WebFrame::windowObjectCleared):
6:26 PM Changeset in webkit [28220] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Reviewed by Adam.

  • platform/win/Skipped: Removed two Mac-specific tests.
6:21 PM Changeset in webkit [28219] by alice.liu@apple.com
  • 2 edits
    1 copy
    1 add in trunk/LayoutTests

Move incorrect, Mac-specific results of fast/dom/Window/window-resize.html into platform/mac and replace them with non-Mac results

The incorrect results are tracked by <rdar://5563578>.

Reviewed by Adam.

  • fast/dom/Window/window-resize-expected.txt: Replaced.
  • platform/mac/fast/dom/Window/window-resize-expected.txt: Copied from fast/dom/Window/window-resize-expected.txt.
  • platform/win/Skipped: Removed window-resize tests.
6:19 PM Changeset in webkit [28218] by weinig@apple.com
  • 4 edits in trunk

WebKitTools:

Reviewed by Anders.

Add database quota UIDelegates methods to DRT UIDelegate.

  • DumpRenderTree/mac/UIDelegate.mm: (-[UIDelegate webView:quotaForSecurityOrigin:toCreateDatabase:withEstimatedSize:]): (-[UIDelegate webView:quotaForSecurityOrigin:fromProposedQuota:database:]):

LayoutTests:

Reviewed by Anders.

Revert r28213 which added an error callback that was not needed and changed
the results.

  • storage/transaction_callback_exception_crash.html:
6:07 PM Changeset in webkit [28217] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Reviewed by Adam.

  • platform/win/Skipped: Removed a passing test.
6:01 PM Changeset in webkit [28216] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Group failures related to dumpResourceLoadCallbacks being unimplemented

Reviewed by Adam.

  • platform/win/Skipped:
6:00 PM Changeset in webkit [28215] by alp@webkit.org
  • 16 edits in trunk/WebCore

2007-11-30 Alp Toker <​alp@atoker.com>

Reviewed by Maciej.

​http://bugs.webkit.org/show_bug.cgi?id=16173
Licensing change

Change license from BSD to LGPL.

This is being done to encourage vendors to contribute their fixes back
upstream.

5:44 PM Changeset in webkit [28214] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Group failures related to using Cygwin paths

Reviewed by Adam.

  • platform/win/Skipped:
5:02 PM Changeset in webkit [28213] by beidson@apple.com
  • 2 edits in trunk/LayoutTests

Not reviewed.

Fixed a hang in the layout tests.
My checkin earlier today broke things such that the transaction call errors out instead of succeeds.
Therefore LayoutTestController is never notified that the test is done.


This changes modifies the test such that the controller is notified of completion even if
there's a transaction error.

The net effect of this fix is to transform the hang to a single layout test failure, so people
and the bots don't have their run-webkit-tests hang while I try to fix the real regression.

  • storage/transaction_callback_exception_crash.html:
4:53 PM Changeset in webkit [28212] by Darin Adler
  • 4 edits
    2 adds in trunk

WebCore:

Reviewed by Adele.

Test: fast/forms/input-radio-checked-tab.html

  • html/HTMLFormElement.cpp: (WebCore::HTMLFormElement::CheckedRadioButtons::addButton): Fix this code so that it doesn't call setChecked(false) until after the map has been updated. Otherwise, we can end up deallocating the map before manipulating it. As long as I was changing the function, I decided to make it do only a single hash table lookup.
  • html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::setChecked): Remove the button from the radio buttons set before changing the checked state. This matches the idiom used elsewhere and fixes the problem where setting checked to false would not remove it from the set.

LayoutTests:

Reviewed by Adele.

  • fast/forms/input-radio-checked-tab-expected.txt: Added.
  • fast/forms/input-radio-checked-tab.html: Added.
4:50 PM Changeset in webkit [28211] by alice.liu@apple.com
  • 1 edit
    1 copy
    1 add in trunk/LayoutTests

Move Mac-specific results of fast/dom/wrapper-classes.html into platform/mac and replace them with non-Mac results

Reviewed by Adam.

  • fast/dom/wrapper-classes-expected.txt: Replaced.
  • platform/mac/fast/dom/wrapper-classes-expected.txt: Copied from fast/dom/wrapper-classes-expected.txt.
4:43 PM Changeset in webkit [28210] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Group drag-n-drop related failures under one bug

We don't know the specific causes of these failures, but we believe
they're due to bugs in DRT's drag-n-drop implementation.

Reviewed by Adam.

  • platform/win/Skipped:
4:36 PM Changeset in webkit [28209] by kmccullough@apple.com
  • 11 edits in trunk

WebKit/win:

Reviewed by Sam.

  • <rdar://5618976> Drosera: should listen for the WebScriptDebugServer dying and vice versa.
  • Interfaces/IWebScriptDebugListener.idl: Added the new function to the interface.
  • WebKit.vcproj/WebKit.def: Added the new function to the def file so Safari can tell WebKit when the server is dying.
  • WebKit.vcproj/WebKit_debug.def: Ditto.
  • WebKitDLL.cpp: Added the new function for Safari to call. (RunAsLocalServer): (LocalServerDidDie):
  • WebScriptDebugServer.cpp: (WebScriptDebugServer::WebScriptDebugServer): Removed unused member. (WebScriptDebugServer::sharedWebScriptDebugServer): Ditto, and added new member to prevent adding or removing listeners when the server is dying. (WebScriptDebugServer::addListener): Don't let new listeners be added when the server is dying. (WebScriptDebugServer::removeListener): Don't let listeners remove themselves when the server is dying. The server will remove them all anyways. (WebScriptDebugServer::serverDidDie): Notify all listeners that the server is dying and remove all listeners.
  • WebScriptDebugServer.h: Added the new method and removed an unused member.

WebKitTools:

Reviewed by Sam.

  • <rdar://5618976> Drosera: should listen for the WebScriptDebugServer dying and vice versa.
  • This fix will allow Drosera and Safari to reconnect if either of them closes correctly, but does not fix the case where one of them dies silently.
  • Drosera/win/ServerConnection.cpp: (ServerConnection::attemptToCreateServerConnection): Added a safety check, because it's possilbe to try to connect to a server that's dying. (ServerConnection::serverDidDie): Implemented. This resets Drosera when the server has died.
  • Drosera/win/ServerConnection.h: Added the new function and removed an unnecessary member.
4:29 PM Changeset in webkit [28208] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Reviewed by Adam.

  • platform/win/Skipped: Removed another passing test.
4:26 PM Changeset in webkit [28207] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Reviewed by Adam.

  • platform/win/Skipped: Recategorized one failing test and removed a passing test.
4:26 PM Changeset in webkit [28206] by justin.garcia@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Darin Adler.

<rdar://problem/5620093> Brief hang when Safari caches page text of XML document

  • editing/TextIterator.cpp: (WebCore::TextIterator::representNodeOffsetZero): When we haven't been emitting any characters, shouldRepresentNodeOffsetZero() can create VisiblePositions, which is expensive. So, we perform the inexpensive checks on m_node to see if it necessitates emitting a character first and will early return before encountering shouldRepresentNodeOffsetZero()s worse case behavior.
4:23 PM Changeset in webkit [28205] by andersca@apple.com
  • 2 edits in trunk/WebKitTools
  • DumpRenderTree/win/DumpRenderTree.vcproj: Add shlwapi.lib to all configurations.
3:30 PM Changeset in webkit [28204] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Removing <rdar://5507350> from skipped list because it's not failing.

  • platform/win/Skipped:
2:15 PM Changeset in webkit [28203] by andersca@apple.com
  • 1 edit
    4 moves in trunk/LayoutTests

Reviewed by Alexey.

Move Mac-specific tests to platform/mac.


  • fast/events/objc-event-api-expected.txt: Removed.
  • fast/events/objc-event-api.html: Removed.
  • fast/events/objc-keyboard-event-creation-expected.txt: Removed.
  • fast/events/objc-keyboard-event-creation.html: Removed.
  • platform/mac/fast/events/objc-event-api-expected.txt: Copied from fast/events/objc-event-api-expected.txt.
  • platform/mac/fast/events/objc-event-api.html: Copied from fast/events/objc-event-api.html.
  • platform/mac/fast/events/objc-keyboard-event-creation-expected.txt: Copied from fast/events/objc-keyboard-event-creation-expected.txt.
  • platform/mac/fast/events/objc-keyboard-event-creation.html: Copied from fast/events/objc-keyboard-event-creation.html.
2:10 PM Changeset in webkit [28202] by andersca@apple.com
  • 3 edits in trunk/WebKit/mac

Reviewed by John.

Rename WebKitShrinksStandaloneImagesToFitKey to WebKitShrinksStandaloneImagesToFit.


This is safe to do because the preference is off by default and Safari 3, which is the only client that turns it on,
is using the setter and not messing around with NSUserDefaults.


  • WebView/WebPreferenceKeysPrivate.h:
  • WebView/WebPreferences.m: (+[WebPreferences initialize]): (-[WebPreferences shrinksStandaloneImagesToFit]): (-[WebPreferences setShrinksStandaloneImagesToFit:]):
1:57 PM Changeset in webkit [28201] by Adam Roben
  • 3 edits in trunk/JavaScriptCore

Windows build fix

Removed some unreachable code (ironically, the code was some
ASSERT_NOT_REACHED()s).

  • pcre/pcre_compile.cpp: (compile_branch):
  • pcre/pcre_exec.cpp: (match):
1:53 PM Changeset in webkit [28200] by andersca@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Sam.

Don't install event handlers or set the zoom cursor if auto-shrink has been disabled.

  • loader/ImageDocument.cpp: (WebCore::ImageDocument::createDocumentStructure): (WebCore::ImageDocument::imageChanged):
1:51 PM Changeset in webkit [28199] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Removing some fixed tests

  • platform/win/Skipped:
1:49 PM Changeset in webkit [28198] by beidson@apple.com
  • 2 edits in trunk/WebKit/win

Build fix that works...?

  • WebChromeClient.h:
1:26 PM Changeset in webkit [28197] by kmccullough@apple.com
  • 2 edits in trunk/WebKitTools
  • Build fix. Added additional includes for VS Express to the Release project.
  • Drosera/win/Drosera.vcproj/Drosera.vcproj:
1:13 PM Changeset in webkit [28196] by beidson@apple.com
  • 2 edits in trunk/WebKit/gtk

Better build fix for Gtk

  • WebCoreSupport/ChromeClientGtk.h:
1:09 PM Changeset in webkit [28195] by beidson@apple.com
  • 2 edits in trunk/WebKit/win

Build fix

  • WebChromeClient.cpp: (WebChromeClient::requestQuotaIncreaseForNewDatabase): (WebChromeClient::requestQuotaIncreaseForDatabaseOperation):
11:53 AM Changeset in webkit [28194] by andersca@apple.com
  • 5 edits in trunk

WebKitTools:

Reviewed by Adam.

<rdar://problem/5132005>
setUserStyleSheetEnabled is not fully implemented in Windows DRT.


  • DumpRenderTree/win/DumpRenderTree.vcproj: Add shlwapi.lib


  • DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::setUserStyleSheetEnabled): Implement this.


(appendComponentToPath):
New method which wraps the Win32 API PathAppend.


(followShortcuts):
New method which checks if a file points to a shortcut and
follows the shortcut.


(resolveCygwinPath):
New method that takes a cygwin unix-style path and returns the Win32 path.


(cfStringRefToWString):


(LayoutTestController::setUserStyleSheetLocation):
Implement this.

LayoutTests:

Reviewed by Adam.

<rdar://problem/5132005>
setUserStyleSheetEnabled is not fully implemented in Windows DRT.


  • platform/win/Skipped:
11:38 AM Changeset in webkit [28193] by Beth Dakin
  • 2 edits in trunk/WebCore

Reviewed by Geoff.

Fix for <rdar://problem/5620151> Disable JavaScript for SVG as
image for <img> and CSS images

  • platform/graphics/svg/SVGImage.cpp: (WebCore::SVGImage::dataChanged): When we create the empty clients, explicitly disable JavaScript in the settings.
11:29 AM Changeset in webkit [28192] by beidson@apple.com
  • 31 edits in trunk

WebCore:

Reviewed by Anders

Support for <rdar://problem/5556381> and <rdar://problem/5556379>

The Database feature in the engine needs to support delegate calls for policy decisions and
also provide notifications when origins and databases change their state.

This patch also polishes off the missing features of the management API

  • page/Chrome.cpp: Implement the two UIDelegate methods (WebCore::Chrome::requestQuotaIncreaseForNewDatabase): Ask for more space to create a new database if it won't fit (WebCore::Chrome::requestQuotaIncreaseForDatabaseOperation): Ask for more space to complete an in-progress operation
  • page/Chrome.h:
  • page/ChromeClient.h:
  • platform/graphics/svg/SVGImageEmptyClients.h: (WebCore::SVGEmptyChromeClient::requestQuotaIncreaseForNewDatabase): (WebCore::SVGEmptyChromeClient::requestQuotaIncreaseForDatabaseOperation):
  • storage/Database.cpp: (WebCore::Database::openDatabase): Check to see if this database can be opened - the quota must be high enough and if it isn't, the UIDelegate should have a change to bump the quota. Also update the UI-relevant details (display name and estimated size) upon successful opening of the database
  • storage/DatabaseDetails.h: Remove the version parameter as it is a programatic detail of a site database and is not important to API clients (WebCore::DatabaseDetails::DatabaseDetails): (WebCore::DatabaseDetails::isValid): (WebCore::DatabaseDetails::name):
  • storage/DatabaseTracker.cpp: (WebCore::DatabaseTracker::openTrackerDatabase): Tweaked the schema here - there was a horrible bug with the old schema that would prevent you from having two databases of the same name from two different origins. Also simplify the origin-management schema (WebCore::DatabaseTracker::canEstablishDatabase): Added. Does some estimated size vs quota checks, and asks the UI delegate for more space if necessary (WebCore::DatabaseTracker::hasEntryForOrigin): (WebCore::DatabaseTracker::establishEntryForOrigin): Establishes a tracker entry for the given origin with the current default quota Also notifies the client of the new origin (WebCore::DatabaseTracker::setDatabaseDetails): Update the display name and estimated size for the given database (WebCore::DatabaseTracker::fullPathForDatabase): Tweak to add the ability to get the path without creating it - for management purposes (WebCore::DatabaseTracker::populateOrigins): Populate origins from the Origins table instead of the Databases table (WebCore::DatabaseTracker::origins): (WebCore::DatabaseTracker::detailsForNameAndOrigin): For API management (WebCore::DatabaseTracker::usageForDatabase): (WebCore::DatabaseTracker::usageForOrigin): (WebCore::DatabaseTracker::quotaForOrigin): (WebCore::DatabaseTracker::setQuota): Notify the client (WebCore::DatabaseTracker::addDatabase): Notify the client (WebCore::DatabaseTracker::deleteAllDatabases): (WebCore::DatabaseTracker::deleteDatabasesWithOrigin): (WebCore::DatabaseTracker::deleteDatabase): (WebCore::DatabaseTracker::deleteDatabaseFile):
  • storage/DatabaseTracker.h:

WebKit/gtk:

Keep it building with new client method

  • WebCoreSupport/ChromeClientGtk.cpp: (WebKit::ChromeClient::requestQuotaIncreaseForNewDatabase): (WebKit::ChromeClient::requestQuotaIncreaseForDatabaseOperation):
  • WebCoreSupport/ChromeClientGtk.h:

WebKit/mac:

Reviewed by Anders

Support for <rdar://problem/5556381> and <rdar://problem/5556379>

Hook up UI Delegate calls for the database engine feature and other small tweaks

  • Storage/WebDatabaseManager.mm: (-[WebDatabaseManager detailsForDatabase:withOrigin:]): Renamed databaseName parameter to databaseIdentifier for clarity (-[WebDatabaseManager deleteDatabase:withOrigin:]): Renamed databaseName parameter to databaseIdentifier for clarity
  • Storage/WebDatabaseManagerPrivate.h:
  • Storage/WebDatabaseTrackerClient.h:
  • Storage/WebDatabaseTrackerClient.mm: (WebDatabaseTrackerClient::dispatchDidModifyDatabase): Renamed databaseName parameter to databaseIdentifier for clarity
  • WebCoreSupport/WebChromeClient.h:
  • WebCoreSupport/WebChromeClient.mm: (WebChromeClient::requestQuotaIncreaseForNewDatabase): Call through to the UI Delegate (WebChromeClient::requestQuotaIncreaseForDatabaseOperation): Ditto
  • WebView/WebUIDelegatePrivate.h: Added the two UI Delegate methods
  • WebView/WebView.mm: (CallDelegateReturningUnsignedLongLong): (CallUIDelegateReturningUnsignedLongLong):
  • WebView/WebViewInternal.h:

WebKit/qt:

Keep it building with new client method

  • WebCoreSupport/ChromeClientQt.cpp: (WebCore::ChromeClientQt::requestQuotaIncreaseForNewDatabase): (WebCore::ChromeClientQt::requestQuotaIncreaseForDatabaseOperation):
  • WebCoreSupport/ChromeClientQt.h:

WebKit/win:

Keep it building with new client method

  • WebChromeClient.cpp: (ChromeClient::requestQuotaIncreaseForNewDatabase): (ChromeClient::requestQuotaIncreaseForDatabaseOperation):
  • WebChromeClient.h:

WebKit/wx:

Keep it building with new client method

  • WebKitSupport/ChromeClientWx.cpp: (WebCore::ChromeClient::requestQuotaIncreaseForNewDatabase): (WebCore::ChromeClient::requestQuotaIncreaseForDatabaseOperation):
  • WebKitSupport/ChromeClientWx.h:
11:21 AM Changeset in webkit [28191] by alice.liu@apple.com
  • 2 edits in trunk/WebKitTools

Reviewed by Adam.

Fixed <rdar://5133828> fast/frames/iframe-window-focus.html output is lowercase

  • DumpRenderTree/win/EventSender.cpp: (keyDownCallback): uppercase letters were being sent as lowercase letters without the shift key down.
10:56 AM Drosera edited by kmccullough@apple.com
Added new information about the design of Drosera, specifically to add … (diff)
10:48 AM Changeset in webkit [28190] by mitz@apple.com
  • 1 edit in trunk/WebKit/win/WebView.cpp

Fix typo

10:46 AM Changeset in webkit [28189] by kmccullough@apple.com
  • 2 edits in trunk/WebKitTools
  • Windows build fix. VS express needs some love.
  • Drosera/win/Drosera.vcproj/Drosera.vcproj:
10:45 AM Changeset in webkit [28188] by mitz@apple.com
  • 3 edits
    1 add in trunk

WebCore:

Reviewed by Beth Dakin and Darin Adler.

  • manual test for <rdar://problem/5346394> Contextmenu event doesn't fire on body element inside frame
  • manual-tests/context-click-unfocused-frame.html: Added.

WebKit/win:

Reviewed by Beth Dakin and Darin Adler.

  • fix <rdar://problem/5346394> Contextmenu event doesn't fire on body element inside frame
  • WebView.cpp: (WebView::handleContextMenuEvent): Send the context menu event to the frame containing the node that was hit.
10:36 AM Changeset in webkit [28187] by kmccullough@apple.com
  • 2 edits in trunk/WebKit/mac

Reviewed by Sam.

  • Added recursive runloop guards.
  • DefaultDelegates/WebScriptDebugServer.m: (-[WebScriptDebugServer suspendProcessIfPaused]):
6:48 AM BuildingGtk edited by hendry@iki.fi
let people know about ccache (diff)
6:38 AM UsingGitWithWebKit edited by hendry@iki.fi
initial setup (diff)
4:49 AM Changeset in webkit [28186] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-29 Eric Seidel <​eric@webkit.org>

Reviewed by Mark Rowe.

Fix for --guard crash of fast/js/regexp-charclass-crash introduced by r28151.

  • pcre/pcre_compile.cpp: (is_anchored):
3:57 AM Changeset in webkit [28185] by mrowe@apple.com
  • 2 edits in trunk/JavaScriptCore

Gtk build fix.

3:44 AM Changeset in webkit [28184] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-29 Eric Seidel <​eric@webkit.org>

Reviewed by Mark Rowe.


Attempt to fix non-mac builds after PCRE cleanup.

  • JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  • JavaScriptCoreSources.bkl:
  • pcre/pcre.pri:
3:39 AM Changeset in webkit [28183] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Give consistent naming to the RegExp options/compile flags

  • pcre/pcre_compile.cpp: (compile_branch): (is_anchored): (find_firstassertedchar): (printCompiledRegExp): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:38 AM Changeset in webkit [28182] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Give consistent naming to the RegExp options/compile flags

  • pcre/pcre_compile.cpp: (compile_branch): (is_anchored): (find_firstassertedchar): (printCompiledRegExp): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:37 AM Changeset in webkit [28181] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Pull first_byte and req_byte optimizations out into separate static funtions, SunSpider reported this as a win.

  • pcre/pcre_exec.cpp: (tryFirstByteOptimization): (tryRequiredByteOptimization): (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:37 AM Changeset in webkit [28180] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

give PCRE_MULTILINE a better name: OptionMatchAcrossMultipleLines

  • pcre/pcre_compile.cpp: (compile_branch): (is_anchored): (printCompiledRegExp): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:35 AM Changeset in webkit [28179] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Oliver.

Deprecate jsRegExpExecute's offset-vector fallback code

  • pcre/pcre_exec.cpp: (jsRegExpExecute):
3:34 AM Changeset in webkit [28178] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-28 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Centralize code for subjectPtr adjustments using inlines, only ever check for a single trailing surrogate (as UTF16

only allows one), possibly fix PCRE bugs involving char classes and garbled UTF16 strings.

  • pcre/pcre_exec.cpp: (match): (jsRegExpExecute):
  • pcre/pcre_internal.h: (getPreviousChar): (movePtrToPreviousChar): (movePtrToNextChar): (movePtrToStartOfCurrentChar):
3:33 AM Changeset in webkit [28177] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Make cur_is_word and prev_is_word locals, and change OP_ANY to OP_ANY_CHAR for clarity

  • pcre/pcre_compile.cpp: (find_fixedlength): (compile_branch): (canApplyFirstCharOptimization):
  • pcre/pcre_exec.cpp: (match):
  • pcre/pcre_internal.h:
3:32 AM Changeset in webkit [28176] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Mitz & Maciej.

Change _NC operators to use _IGNORING_CASE for clarity

  • pcre/pcre_compile.cpp: (find_fixedlength): (compile_branch): (find_firstassertedchar):
  • pcre/pcre_exec.cpp: (match):
  • pcre/pcre_internal.h:
3:32 AM Changeset in webkit [28175] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Mitz.

Remove branch from return

  • pcre/pcre_compile.cpp: (compile_branch):
  • pcre/pcre_exec.cpp: (match):
3:31 AM Changeset in webkit [28174] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Add repeatInformationFromInstructionOffset inline

  • pcre/pcre_exec.cpp: (repeatInformationFromInstructionOffset): (match):
3:30 AM Changeset in webkit [28173] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Remove no longer used error code JSRegExpErrorMatchLimit

  • kjs/regexp.cpp: (KJS::RegExp::match):
  • pcre/pcre.h:
  • pcre/pcre_internal.h:
3:30 AM Changeset in webkit [28172] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-28 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

change getChar* functions to return result and push 'c' into local scopes for clarity

  • pcre/pcre_compile.cpp: (compile_branch): (calculateCompiledPatternLengthAndFlags):
  • pcre/pcre_exec.cpp: (match):
  • pcre/pcre_internal.h: (getChar): (getCharAndAdvance): (getCharAndLength): (getCharAndAdvanceIfSurrogate):
3:29 AM Changeset in webkit [28171] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Make i locally scoped for better code clarity

  • pcre/pcre_exec.cpp: (match):
3:29 AM Changeset in webkit [28170] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-28 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Comment cleanup

  • pcre/pcre_exec.cpp: (match):
3:28 AM Changeset in webkit [28169] by eric@webkit.org
  • 5 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Give subjectPtr and instructionPtr sane names, reduce size of MatchFrame for a 0.2% speedup.

  • pcre/pcre_compile.cpp: (compile_branch): (calculateCompiledPatternLengthAndFlags):
  • pcre/pcre_exec.cpp: (match_ref): (MatchStack::pushNewFrame): (getUTF8CharAndIncrementLength): (match):
  • pcre/pcre_internal.h: (getChar): (getCharAndAdvance): (getCharAndLength): (getCharAndAdvanceIfSurrogate):
  • pcre/pcre_xclass.cpp: (getUTF8CharAndAdvancePointer):
3:27 AM Changeset in webkit [28168] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-26 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Small speedup (0.7%) by simplifying canUseStackBufferForNextFrame() check

  • pcre/pcre_exec.cpp: (MatchStack::MatchStack): (MatchStack::popCurrentFrame):
3:27 AM Changeset in webkit [28167] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-25 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Lower MATCH_LIMIT_RECURSION to more sane levels to prevent hangs on run-javascriptcore-tests

  • pcre/pcre_internal.h:
3:26 AM Changeset in webkit [28166] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-25 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Remove match_is_group variable for another 5% speedup

  • pcre/pcre_compile.cpp:
  • pcre/pcre_exec.cpp: (startNewGroup): (match):
3:25 AM Changeset in webkit [28165] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Remove redundant eptrblock struct

  • pcre/pcre_exec.cpp: (MatchStack::pushNewFrame): (match):
3:25 AM Changeset in webkit [28164] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Remove redundant match_call_count and move recursion check out of super-hot code path
SunSpider says this is at least an 8% speedup for regexp.

  • pcre/pcre_exec.cpp: (MatchStack::MatchStack): (MatchStack::pushNewFrame): (MatchStack::popCurrentFrame): (MatchStack::popAllFrames): (match): (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:24 AM Changeset in webkit [28163] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Get rid of GETCHAR* macros, replacing them with better named inlines

  • pcre/pcre_compile.cpp: (compile_branch): (calculateCompiledPatternLengthAndFlags):
  • pcre/pcre_exec.cpp: (match):
  • pcre/pcre_internal.h: (getCharAndAdvance): (getCharAndLength): (getCharAndAdvanceIfSurrogate):
3:24 AM Changeset in webkit [28162] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Further cleanup GET/PUT inlines

  • pcre/pcre_internal.h: (putOpcodeValueAtOffset): (getOpcodeValueAtOffset): (putOpcodeValueAtOffsetAndAdvance): (put2ByteOpcodeValueAtOffset): (get2ByteOpcodeValueAtOffset): (put2ByteOpcodeValueAtOffsetAndAdvance):
3:23 AM Changeset in webkit [28161] by eric@webkit.org
  • 5 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Give GET, PUT better names, and add (poor) moveOpcodePtrPastAnyAlternateBranches

  • pcre/pcre_compile.cpp: (firstSignificantOpCodeSkippingAssertions): (find_fixedlength): (complete_callout): (compile_branch): (compile_regex): (is_anchored): (canApplyFirstCharOptimization): (find_firstassertedchar):
  • pcre/pcre_exec.cpp: (match):
  • pcre/pcre_internal.h: (putOpcodeValueAtOffset): (getOpcodeValueAtOffset): (putOpcodeValueAtOffsetAndAdvance): (put2ByteOpcodeValueAtOffset): (get2ByteOpcodeValueAtOffset): (moveOpcodePtrPastAnyAlternateBranches):
  • pcre/pcre_ucp_searchfuncs.cpp: (_pcre_ucp_othercase):
3:22 AM Changeset in webkit [28160] by eric@webkit.org
  • 5 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Add inlines for toLowerCase, isWordChar, isSpaceChar for further regexp speedup

  • pcre/pcre_compile.cpp: (compile_branch): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (match): (jsRegExpExecute):
  • pcre/pcre_internal.h: (toLowerCase): (flipCase): (classBitmapForChar): (charTypeForChar): (isWordChar): (isSpaceChar): (CompileData::CompileData):
  • pcre/pcre_xclass.cpp: (_pcre_xclass):
3:21 AM Changeset in webkit [28159] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

cleanup _pcre_ucp_othercase

  • pcre/pcre_ucp_searchfuncs.cpp: (_pcre_ucp_othercase):
3:21 AM Changeset in webkit [28158] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-28 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Abstract frame variables into locals and args

  • pcre/pcre_compile.cpp: (compile_branch):
  • pcre/pcre_exec.cpp: (match):
  • pcre/pcre_internal.h:
3:20 AM Changeset in webkit [28157] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-28 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Section off MatchData arguments into args struct

  • pcre/pcre_exec.cpp: (MatchStack::pushNewFrame): (match):
3:19 AM Changeset in webkit [28156] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

split first_significant_code into two simpler functions

  • pcre/pcre_compile.cpp: (firstSignificantOpCode): (firstSignificantOpCodeSkippingAssertions): (is_anchored): (canApplyFirstCharOptimization): (find_firstassertedchar):
3:18 AM Changeset in webkit [28155] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

split first_significant_code into two simpler functions

  • pcre/pcre_compile.cpp: (firstSignificantOpCode): (firstSignificantOpCodeSkippingAssertions): (is_anchored): (canApplyFirstCharOptimization): (find_firstassertedchar):
3:18 AM Changeset in webkit [28154] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

clean up is_counted_repeat

  • pcre/pcre_compile.cpp: (is_counted_repeat):
3:17 AM Changeset in webkit [28153] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

clean up check_escape

  • pcre/pcre_compile.cpp: (check_escape):
3:16 AM Changeset in webkit [28152] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Reformat find_fixedlength

  • pcre/pcre_compile.cpp: (find_fixedlength):
3:16 AM Changeset in webkit [28151] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

reformat is_anchored

  • pcre/pcre_compile.cpp: (is_anchored):
3:15 AM Changeset in webkit [28150] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Remove unused function could_be_empty_branch

  • pcre/pcre_compile.cpp: (first_significant_code): (find_fixedlength): (compile_branch): (canApplyFirstCharOptimization):
3:14 AM Changeset in webkit [28149] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Pass around MatchData objects by reference

  • pcre/pcre_exec.cpp: (pchars): (match_ref): (match): (jsRegExpExecute):
3:13 AM Changeset in webkit [28148] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

give PCRE_STARTLINE a better name and rename match_data to MatchData

  • pcre/pcre_compile.cpp: (compile_branch): (canApplyFirstCharOptimization): (find_firstassertedchar): (printCompiledRegExp): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (pchars): (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:12 AM Changeset in webkit [28147] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Clean up find_firstassertedchar

  • pcre/pcre_compile.cpp: (get_othercase_range): (find_firstassertedchar): (calculateCompiledPatternLengthAndFlags):
3:11 AM Changeset in webkit [28146] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Tim Hatcher.

Pass around CompileData& instead of CompileData*

  • pcre/pcre_compile.cpp: (compile_branch): (jsRegExpCompile):
3:10 AM Changeset in webkit [28145] by eric@webkit.org
  • 4 edits
    1 delete in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Clean up compile_branch, move _pcre_ord2utf8, and rename CompileData

  • JavaScriptCore.xcodeproj/project.pbxproj:
  • pcre/pcre_compile.cpp: (_pcre_ord2utf8): (calculateCompiledPatternLengthAndFlags): (jsRegExpCompile):
  • pcre/pcre_internal.h:
  • pcre/pcre_ord2utf8.cpp: Removed.
3:09 AM Changeset in webkit [28144] by eric@webkit.org
  • 5 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

removing more macros

  • pcre/pcre_compile.cpp: (could_be_empty_branch): (compile_branch): (calculateCompiledPatternLengthAndFlags):
  • pcre/pcre_exec.cpp: (match): (jsRegExpExecute):
  • pcre/pcre_internal.h:
  • pcre/pcre_xclass.cpp:
3:08 AM Changeset in webkit [28143] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

clean up formating in compile_branch

  • pcre/pcre_compile.cpp: (compile_branch):
3:07 AM Changeset in webkit [28142] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Fix spacing for read_repeat_counts

  • pcre/pcre_compile.cpp: (read_repeat_counts):
3:06 AM Changeset in webkit [28141] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Get rid of PCRE custom char types

  • pcre/pcre_compile.cpp: (check_escape): (complete_callout): (compile_branch): (compile_regex): (calculateCompiledPatternLengthAndFlags): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (match_ref): (match): (jsRegExpExecute):
  • pcre/pcre_internal.h:
3:06 AM Changeset in webkit [28140] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

reformat get_othercase_range

  • pcre/pcre_compile.cpp: (get_othercase_range):
3:05 AM Changeset in webkit [28139] by eric@webkit.org
  • 6 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Remove register keyword and more cleanup

  • pcre/pcre_compile.cpp: (find_fixedlength): (compile_branch): (is_anchored): (is_startline): (find_firstassertedchar): (calculateCompiledPatternLengthAndFlags): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (MatchStack::canUseStackBufferForNextFrame): (MatchStack::allocateNextFrame): (MatchStack::pushNewFrame): (MatchStack::frameIsStackAllocated): (MatchStack::popCurrentFrame): (MatchStack::unrollAnyHeapAllocatedFrames): (getUTF8CharAndIncrementLength): (match): (jsRegExpExecute):
  • pcre/pcre_internal.h: (PUT2INC): (isLeadingSurrogate): (isTrailingSurrogate): (decodeSurrogatePair): (getChar):
  • pcre/pcre_ord2utf8.cpp: (_pcre_ord2utf8):
  • pcre/pcre_xclass.cpp: (getUTF8CharAndAdvancePointer): (_pcre_xclass):
3:04 AM Changeset in webkit [28138] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-24 Eric Seidel <​eric@webkit.org>

Reviewed by Maciej.

Clean up jsRegExpExecute

  • pcre/pcre_compile.cpp: (returnError): (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (jsRegExpExecute):
  • pcre/pcre_internal.h:
2:16 AM Changeset in webkit [28137] by mrowe@apple.com
  • 2 edits in trunk/JavaScriptGlue

Fix REGRESSION(r27885): Installer hits assertion failure in JavaScriptGlue.

Reviewed by Maciej.

  • JSValueWrapper.cpp:

(getThreadGlobalExecState): Set the global object on the Interpreter before using it.

1:05 AM Changeset in webkit [28136] by mrowe@apple.com
  • 2 edits in trunk/WebKit/mac

Fix an assertion failure seen on the layout tests, and when closing the window after visiting <​http://www.coudal.com/losalamos/>.

Reviewed by Oliver Hunt.

  • Plugins/WebBaseNetscapePluginStream.mm:

(-[WebBaseNetscapePluginStream _destroyStream]): Unlink the file and close the file
descriptor even when the stream is being destroyed without the load completing. This
avoids leaking the path and file descriptor, and leaving the temporary file on disk.

12:46 AM Changeset in webkit [28135] by oliver@apple.com
  • 5 edits in trunk/JavaScriptCore

Merging updated system alloc and spinlock code from r38 of TCMalloc.

Reviewed by Geoff

This is needed as a precursor to the merge of TCMalloc proper.

12:44 AM template.svg attached to IconContestIdeas by Adam Roben
Preliminary entry template

Nov 28, 2007:

11:57 PM IconContestIdeas edited by Adam Roben
(diff)
11:40 PM IconContestIdeas edited by Adam Roben
(diff)
11:39 PM IconContestIdeas edited by Adam Roben
Stole a judging criterion from the SVG logo contest (diff)
11:32 PM IconContestIdeas edited by Adam Roben
Change start of submissions to be the same as the announcement (diff)
10:20 PM Changeset in webkit [28134] by mrowe@apple.com
  • 1 edit
    4 moves in trunk/LayoutTests

Unreviewed layout test fix. Fix typos in the name of four files used by HTTP layout tests.

10:07 PM Changeset in webkit [28133] by Antti Koivisto
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Geoff.


Fallback content inside media elements should not be rendered.

Test: media/fallback.html

  • rendering/RenderMedia.h: (WebCore::RenderMedia::canHaveChildren):

LayoutTests:

Reviewed by Geoff.


Test that fallback content inside media elements is not be rendered.

  • media/fallback-expected.txt: Added.
  • media/fallback.html: Added.
10:04 PM Changeset in webkit [28132] by beidson@apple.com
  • 2 edits in trunk/JavaScriptCore

Reviewed by Geoff

Add copyKeysToVector utility, mirroring copyValuesToVector
Also change the copyValuesToVector implementation to be a little more attractive

  • wtf/HashMap.h: (WTF::copyKeysToVector): (WTF::copyValuesToVector):
9:33 PM Changeset in webkit [28131] by mrowe@apple.com
  • 2 edits in trunk/WebCore

Windows build fix.

9:29 PM Changeset in webkit [28130] by mrowe@apple.com
  • 2 edits in trunk/WebCore

Gtk build fix.

8:50 PM Changeset in webkit [28129] by Antti Koivisto
  • 2 edits in trunk/WebCore

Reviewed by Oliver.

Make QTMovieView fill color transparent. This behavior matches the specification
(and allows nice alpha blended video).

  • platform/graphics/mac/MoviePrivateQTKit.mm: (WebCore::MoviePrivate::createQTMovieView):
6:54 PM IconContestIdeas edited by Adam Roben
(diff)
6:48 PM IconContestIdeas edited by Adam Roben
(diff)
6:48 PM IconContestIdeas edited by Adam Roben
(diff)
6:25 PM Changeset in webkit [28128] by alice.liu@apple.com
  • 17 edits in trunk/LayoutTests

just some more typo fixes

  • http/tests/security/javascriptURL/xss-ALLOWED-from-javascript-url-sub-frame-2-level-expected.txt:
  • http/tests/security/javascriptURL/xss-ALLOWED-from-javascript-url-sub-frame-2-level.html:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-from-javscript-url-expected.txt:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-from-javscript-url.html:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-sub-frame-2-level-expected.txt:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-sub-frame-2-level.html:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-sub-frame-expected.txt:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-sub-frame.html:
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-subframe-expected.txt:
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-subframe.html:
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-window-open-expected.txt:
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-window-open.html:
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-subframe-expected.txt:
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-subframe.html:
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-window-open-expected.txt:
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-window-open.html:
6:02 PM Changeset in webkit [28127] by Antti Koivisto
  • 54 edits
    4 adds in trunk

WebCore:

Reviewed by Darin.

Test: media/video-controls.html


<rdar://problem/5605668>
Implement controls attribute for video element


  • Add RenderMedia and make RenderVideo its subclass. Controls code goes to RenderMedia while video specific (painting etc) code stays in RenderVideo. For example audio controls can be implemented later by just instantiating RenderMedia as renderer.
  • Make media renderer hierarchy inherit from RenderBlock instead of RenderReplaced so it can have child nodes (for controls).
  • Controls are implemented as a shadow DOM.
  • Current look is a placeholder. It is defined purely in CSS.
  • Some things like volume controls are not yet implemented.
  • Fade-in/out is done manually, CSS animations don't work well in shadow trees.


  • WebCore.xcodeproj/project.pbxproj:
  • css/html4.css:
  • html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::attributeChanged): (WebCore::HTMLMediaElement::updateMovie): (WebCore::HTMLMediaElement::defaultEventHandler):
  • html/HTMLMediaElement.h:
  • rendering/RenderMedia.cpp: Added. (WebCore::MediaControlShadowRootElement::MediaControlShadowRootElement): (WebCore::MediaControlShadowRootElement::isShadowNode): (WebCore::MediaControlShadowRootElement::shadowParentNode): (WebCore::MediaControlInputElement::MediaControlInputElement): (WebCore::MediaControlInputElement::attachToParent): (WebCore::MediaControlPlayButtonElement::MediaControlPlayButtonElement): (WebCore::MediaControlPlayButtonElement::inPausedState): (WebCore::MediaControlPlayButtonElement::defaultEventHandler): (WebCore::MediaControlPlayButtonElement::update): (WebCore::MediaControlTimelineElement::MediaControlTimelineElement): (WebCore::MediaControlTimelineElement::defaultEventHandler): (WebCore::MediaControlTimelineElement::update): (WebCore::RenderMedia::RenderMedia): (WebCore::RenderMedia::~RenderMedia): (WebCore::RenderMedia::mediaElement): (WebCore::RenderMedia::movie): (WebCore::RenderMedia::setStyle): (WebCore::RenderMedia::createControlsShadowRoot): (WebCore::RenderMedia::createPanel): (WebCore::RenderMedia::createPlayButton): (WebCore::RenderMedia::createTimeline): (WebCore::RenderMedia::createTimeDisplay): (WebCore::RenderMedia::updateFromElement): (WebCore::RenderMedia::updateControls): (WebCore::RenderMedia::timeUpdateTimerFired): (WebCore::RenderMedia::updateTimeDisplay): (WebCore::RenderMedia::updateControlVisibility): (WebCore::RenderMedia::changeOpacity): (WebCore::RenderMedia::opacityAnimationTimerFired): (WebCore::RenderMedia::forwardEvent):
  • rendering/RenderMedia.h: Added. (WebCore::RenderMedia::renderName): (WebCore::RenderMedia::isMedia): (WebCore::RenderMedia::intrinsicSize):
  • rendering/RenderObject.h: (WebCore::RenderObject::isMedia):
  • rendering/RenderVideo.cpp: (WebCore::RenderVideo::RenderVideo): (WebCore::RenderVideo::videoSizeChanged): (WebCore::RenderVideo::paintObject): (WebCore::RenderVideo::layout): (WebCore::RenderVideo::updateFromElement): (WebCore::RenderVideo::calcAspectRatioWidth): (WebCore::RenderVideo::calcAspectRatioHeight):
  • rendering/RenderVideo.h:

LayoutTests:

Reviewed by Darin.


<rdar://problem/5605668>
Implement controls attribute for video element


  • A new test for the controls attribute.
  • Add the controls attibute to all existing video tests. This will give the code some additional exercise without affecting test results.
  • media/progress-event.html:
  • media/video-append-source.html:
  • media/video-autoplay.html:
  • media/video-buffered.html:
  • media/video-controls-expected.txt: Added.
  • media/video-controls.html: Added.
  • media/video-currentTime-set.html:
  • media/video-currentTime-set2.html:
  • media/video-currentTime.html:
  • media/video-dom-autoplay.html:
  • media/video-dom-end.html:
  • media/video-dom-loopcount.html:
  • media/video-dom-loopend.html:
  • media/video-dom-loopstart.html:
  • media/video-dom-src.html:
  • media/video-dom-start.html:
  • media/video-end.html:
  • media/video-error-abort.html:
  • media/video-error-does-not-exist.html:
  • media/video-load-networkState.html:
  • media/video-load-readyState.html:
  • media/video-loopcount.html:
  • media/video-loopend.html:
  • media/video-loopstart.html:
  • media/video-muted.html:
  • media/video-no-autoplay.html:
  • media/video-pause-empty-events.html:
  • media/video-play-empty-events.html:
  • media/video-play-pause-events.html:
  • media/video-play-pause-exception.html:
  • media/video-poster.html:
  • media/video-seek-past-end-paused.html:
  • media/video-seek-past-end-playing.html:
  • media/video-seekable.html:
  • media/video-seeking.html:
  • media/video-size.html:
  • media/video-source-media.html:
  • media/video-source-type.html:
  • media/video-source.html:
  • media/video-src-change.html:
  • media/video-src-remove.html:
  • media/video-src-set.html:
  • media/video-src-source.html:
  • media/video-src.html:
  • media/video-start.html:
  • media/video-volume.html:
  • media/video-width-height.html:
5:33 PM Changeset in webkit [28126] by justin.garcia@apple.com
  • 15 edits
    6 adds in trunk

WebCore:

Reviewed by Darin Adler.

<rdar://problem/5573879> GMail Beta: Crash in when undoing Remove Formating


In the new beta, Google has added some custom code on top of our RemoveFormat
to work around a bug where fully selected lists aren't removed. This code corrupts
the undo stack and causes crashes. This change fixes two problems with RemoveFormat
so that Google doesn't have to have any custom code. After checking this in I'll
work on bulletproofing the undo stack

  • WebCore.xcodeproj/project.pbxproj:
  • editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::insertLineBreak): Added this convenience method. (WebCore::CompositeEditCommand::inputText): Added code to select all inserted text, not just the last paragraph. Added code to support an input string with '\n's.
  • editing/CompositeEditCommand.h:
  • editing/Editor.cpp: (WebCore::Editor::removeFormattingAndStyle): Moved code to its own EditCommand.
  • editing/RemoveFormatCommand.cpp: Added. Moved code from removeFormattingAndStyle here. (WebCore::RemoveFormatCommand::RemoveFormatCommand): (WebCore::RemoveFormatCommand::doApply): Added code to remove fully selected lists.
  • editing/RemoveFormatCommand.h: Added. (WebCore::RemoveFormatCommand::editingAction):

LayoutTests:

Reviewed by Darin Adler.


<rdar://problem/5573879> GMail Beta: Crash in when undoing Remove Formating

Added to demonstrate fixes:

  • editing/execCommand/5573879.html: Added.
  • platform/mac/editing/execCommand/5573879-expected.checksum: Added.
  • platform/mac/editing/execCommand/5573879-expected.png: Added.
  • platform/mac/editing/execCommand/5573879-expected.txt: Added.


Fixes (RemoveFormat now preserves the selection correctly):

  • platform/mac/editing/execCommand/4920742-1-expected.checksum:
  • platform/mac/editing/execCommand/4920742-1-expected.png:
  • platform/mac/editing/execCommand/4920742-1-expected.txt:
  • platform/mac/editing/execCommand/remove-formatting-2-expected.checksum:
  • platform/mac/editing/execCommand/remove-formatting-2-expected.png:
  • platform/mac/editing/execCommand/remove-formatting-2-expected.txt:
  • platform/mac/editing/execCommand/remove-formatting-expected.checksum:
  • platform/mac/editing/execCommand/remove-formatting-expected.png:
  • platform/mac/editing/execCommand/remove-formatting-expected.txt:
5:17 PM IconContestIdeas edited by Adam Roben
Little re-organziation (diff)
5:16 PM IconContestIdeas edited by Adam Roben
Added a submissions section (diff)
5:13 PM WebKit Team edited by Adam Roben
(diff)
5:06 PM IconContestIdeas edited by Adam Roben
(diff)
5:05 PM Changeset in webkit [28125] by timothy@apple.com
  • 3 edits in trunk/LayoutTests

Make this test not depend on the screen size.

  • fast/dom/Window/window-resize-expected.txt:
  • fast/dom/Window/window-resize.html:
5:04 PM IconContestIdeas edited by Adam Roben
(diff)
5:02 PM IconContestIdeas edited by Adam Roben
Added a disclaimer (diff)
4:59 PM IconContestIdeas edited by Adam Roben
Added a preliminary schedule (diff)
4:44 PM IconContestIdeas edited by Adam Roben
Added voting section (diff)
4:43 PM Changeset in webkit [28124] by Darin Adler
  • 3 edits in trunk/WebKitTools

Reviewed by Adam Roben.

  • DumpRenderTree/mac/EventSendingController.mm: (-[EventSendingController keyDown:withModifiers:]): Send capital letters through as lowercase letters with the shift key down rather than sending them as if they were highly unusual "capital letter keys".
  • Scripts/update-javascriptcore-test-results: Add a "--force" option for cases where you need to update results and more tests are failing than before.
4:33 PM Changeset in webkit [28123] by andersca@apple.com
  • 6 edits in trunk

WebKitTools:

Reviewed by Darin.

<rdar://problem/5132001>
contextClick is not implemented in DRT on Windows.


  • DumpRenderTree/win/EventSender.cpp: (contextClickCallback): Add a callback for contextClick, which sends a WM_RBUTTONDOWN message followed by a WM_RBUTTONUP message.


  • DumpRenderTree/win/UIDelegate.cpp: (UIDelegate::hasCustomMenuImplementation): (UIDelegate::trackCustomPopupMenu):
  • DumpRenderTree/win/UIDelegate.h: Add a no-op implementation of trackCustomPopupMenu, to prevent the default popup menu from being shown (and causing the DRT to hang).

LayoutTests:

Reviewed by Darin.

<rdar://problem/5132001>
contextClick is not implemented in DRT on Windows.

Remove tests that pass now that we have contextClick.


  • platform/win/Skipped:
3:05 PM Changeset in webkit [28122] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Removing some tests tracked by <rdar://5130762> because they don't fail.

  • platform/win/Skipped:
3:01 PM IconContestIdeas edited by timothy@apple.com
Rename resolution to size. (diff)
2:54 PM HackingGtk edited by alp@atoker.com
Add an API rationale (diff)
2:51 PM IconContestIdeas edited by Adam Roben
(diff)
2:45 PM Changeset in webkit [28121] by alice.liu@apple.com
  • 2 edits in trunk/LayoutTests

Removing fixed tests tracked by <rdar://5128229>

  • platform/win/Skipped:
2:41 PM IconContestIdeas edited by Adam Roben
(diff)
2:29 PM IconContestIdeas edited by Adam Roben
(diff)
2:29 PM IconContestIdeas edited by Adam Roben
(diff)
2:27 PM Changeset in webkit [28120] by alice.liu@apple.com
  • 2 edits
    8 moves in trunk/LayoutTests

Rubberstamped by Sam.

Fixing a typo in some filenames

  • http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-subframe-expected.txt: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-subframe.html: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-window-open-expected.txt: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-window-open.html: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-subframe-expected.txt: Copied from http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-subframe-expected.txt.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-subframe.html: Copied from http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-subframe.html.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-window-open-expected.txt: Copied from http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-window-open-expected.txt.
  • http/tests/security/javascriptURL/xss-DENIED-from-javascript-url-in-foreign-domain-window-open.html: Copied from http/tests/security/javascriptURL/xss-DENIED-from-javascipt-url-in-foreign-domain-window-open.html.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-subframe-expected.txt: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-subframe.html: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-window-open-expected.txt: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-window-open.html: Removed.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-subframe-expected.txt: Copied from http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-subframe-expected.txt.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-subframe.html: Copied from http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-subframe.html.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-window-open-expected.txt: Copied from http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-window-open-expected.txt.
  • http/tests/security/javascriptURL/xss-DENIED-to-javascript-url-in-foreign-domain-window-open.html: Copied from http/tests/security/javascriptURL/xss-DENIED-to-javascipt-url-in-foreign-domain-window-open.html.
  • platform/win/Skipped:
2:25 PM IconContestIdeas created by Adam Roben
Initial ideas for a WebKit icon contest
2:24 PM Changeset in webkit [28119] by kmccullough@apple.com
  • 2 edits in trunk/WebKit/win

Reviewed by Tim.

  • Added guards to the wait loop so execution would not hang and now the console window can process JavaScript.
  • WebScriptDebugServer.cpp: (WebScriptDebugServer::suspendProcessIfPaused):
2:04 PM Changeset in webkit [28118] by mitz@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Dave Hyatt.

  • fix <rdar://problem/5612459> CrashTracer: [REGRESSION] 61 crashes in Safari at com.apple.WebCore: WebCore::AutoTableLayout::layout + 2046

Test: fast/table/empty-auto-column-zero-divide.html

  • rendering/AutoTableLayout.cpp: (WebCore::AutoTableLayout::recalcColumn): (WebCore::AutoTableLayout::fullRecalc): (WebCore::AutoTableLayout::calcEffectiveWidth): (WebCore::AutoTableLayout::layout): When distributing the remaining width among columns, skip those whose effective (rather than specified) width is auto and contain only empty cells.

LayoutTests:

Reviewed by Dave Hyatt.

  • test for <rdar://problem/5612459> CrashTracer: [REGRESSION] 61 crashes in Safari at com.apple.WebCore: WebCore::AutoTableLayout::layout + 2046
  • fast/table/empty-auto-column-zero-divide-expected.txt: Added.
  • fast/table/empty-auto-column-zero-divide.html: Added.
2:03 PM Changeset in webkit [28117] by adele@apple.com
  • 2 edits in trunk/WebKit/mac

Reviewed by Darin.

Fix for <rdar://problem/5524216> CrashTracer: [USER] 496 crashes in Safari at com.apple.WebCore: WebCore::Frame::eventHandler const + 6

The CrashTracer shows a variety of crashes in different methods (including keyDown and keyUp). This change adds nil checks for the frame in
WebHTMLView to prevent future problems in other methods as well.

  • WebView/WebHTMLView.mm: (-[WebHTMLView validateUserInterfaceItemWithoutDelegate:]): (-[WebHTMLView keyDown:]): (-[WebHTMLView keyUp:]): (-[WebHTMLView flagsChanged:]): (-[WebHTMLView _selectionStartFontAttributesAsRTF]): (-[WebHTMLView _changeCSSColorUsingSelector:inRange:]): (-[WebHTMLView checkSpelling:]): (-[WebHTMLView showGuessPanel:]): (-[WebHTMLView indent:]): (-[WebHTMLView outdent:]): (-[WebHTMLView paste:]): (-[WebHTMLView _lookUpInDictionaryFromMenu:]): (-[WebHTMLView insertText:]): (-[WebHTMLView selectionTextRects]):
1:17 PM Changeset in webkit [28116] by andersca@apple.com
  • 4 edits in trunk/WebKit/win

Reviewed by Adam.

Implement the last bits of the WebDatabaseManager API.


  • Interfaces/IWebDatabaseManager.idl:
  • WebDatabaseManager.cpp: (isEqual): (WebDatabaseManager::sharedWebDatabaseManager): (WebDatabaseManager::dispatchDidModifyOrigin): (WebDatabaseManager::dispatchDidModifyDatabase):
  • WebDatabaseManager.h:
12:55 PM WebKit Team edited by zimmermann@kde.org
(diff)
12:12 PM Changeset in webkit [28115] by andersca@apple.com
  • 2 edits in trunk/WebKit/win

Get Drosera working again for real.


  • COMEnumVariant.h: (::Next):
11:57 AM Changeset in webkit [28114] by alp@webkit.org
  • 3 edits in trunk/WebCore

2007-11-28 Peter Kasting <​zerodpx@gmail.com>

Reviewed by Alp Toker.

​http://bugs.webkit.org/show_bug.cgi?id=16169
GIF decoder needs to set hasAlpha() correctly on subsequent frames.


This also removes the workaround for this problem in
ImageSourceCairo.cpp.

  • platform/graphics/cairo/ImageSourceCairo.cpp: (WebCore::ImageSource::frameHasAlphaAtIndex):
  • platform/image-decoders/gif/GIFImageDecoder.cpp: (WebCore::GIFImageDecoder::initFrameBuffer):
11:37 AM Changeset in webkit [28113] by jhoneycutt@apple.com
  • 2 edits in trunk/LayoutTests

2007-11-28 Jon Honeycutt <​jhoneycutt@apple.com>

Reviewed by Dan.

<rdar://problem/5507360> plugins/return-error-from-new-stream-callback-
in-full-frame-plugin.html is crashing

Alice fixed this in r28046

  • platform/win/Skipped:
10:08 AM HackingGtk edited by alp@atoker.com
Codify URL vs. URI public API policy (diff)
9:46 AM Changeset in webkit [28112] by kmccullough@apple.com
  • 2 edits in trunk/WebKit/win

Reviewed by Maciej and Adam.

  • Added Drosera to the WebKit project.
  • WebKit.vcproj/WebKit.sln:
8:47 AM Changeset in webkit [28111] by mitz@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Darin Adler and Adam Roben.

  • <rdar://problem/5057218> Reenable "recent searches" search field menu when menu is fully implemented
  • platform/win/SearchPopupMenuWin.cpp: (WebCore::SearchPopupMenu::enabled): Changed to return true. (WebCore::autosaveKey): Added. Returns a preferences key for the autosave name by prefixing it with "com.apple.WebKit.searchField:". (WebCore::SearchPopupMenu::saveRecentSearches): Implemented. (WebCore::SearchPopupMenu::loadRecentSearches): Implemented.
5:12 AM Changeset in webkit [28110] by mrowe@apple.com
  • 59 edits
    10 deletes in trunk

Roll out r28106 and r28108. These introduced a frequent assertion failure on page load and broke all non-Mac builds.

4:47 AM Changeset in webkit [28109] by alp@webkit.org
  • 2 edits in trunk/WebCore

2007-11-28 Alp Toker <​alp@atoker.com>

Reviewed by Mark Rowe.

​http://bugs.webkit.org/show_bug.cgi?id=15689
[GTK] Background of loading images is always black

frameHasAlphaAtIndex() exists only to allow optimization for cases we
are certain the image can be blitted rather than composited.

Thus we need to be conservative, returning false only when we are
absolutely certain there is no need for composited copying, and true
otherwise.

CG doesn't even bother with this optimization at all and always
returns true.

Patch includes a workaround for

​http://bugs.webkit.org/show_bug.cgi?id=16169
GIF ImageDecoder hasAlpha() return value incorrect

  • platform/graphics/cairo/ImageSourceCairo.cpp: (WebCore::ImageSource::frameHasAlphaAtIndex):
3:03 AM Changeset in webkit [28108] by mjs@apple.com
  • 2 edits
    10 adds in trunk

JavaScriptCore:

Add files missing from previous commit.

  • kjs/MarkStack.h: Added.

LayoutTests:

add files missing from previous commit.

  • fast/js/gc-breadth-2-expected.txt: Added.
  • fast/js/gc-breadth-2.html: Added.
  • fast/js/gc-breadth-expected.txt: Added.
  • fast/js/gc-breadth.html: Added.
  • fast/js/gc-depth-expected.txt: Added.
  • fast/js/gc-depth.html: Added.
  • fast/js/resources/gc-breadth-2.js: Added.
  • fast/js/resources/gc-breadth.js: Added.
  • fast/js/resources/gc-depth.js: Added.
2:42 AM Changeset in webkit [28107] by mrowe@apple.com
  • 2 edits in trunk

2007-11-26 Rodney Dawes <​dobey@wayofthemonkey.com>

Reviewed by Mark Rowe.

Set CONFIG option for the window system GTK+ is using

  • WebKit.pri:
2:08 AM Changeset in webkit [28106] by mjs@apple.com
  • 59 edits in trunk

JavaScriptCore:

Reviewed by Darin and Geoff.


Implement mark stack. This version is not suitable for prime time because it makes a
huge allocation on every collect, and potentially makes marking of detached subtrees
slow. But it is an 0.4% SunSpider speedup even without much tweaking.


The basic approach is to replace mark() methods with
markChildren(MarkStack&) methods. Reachable references are pushed
onto a mark stack (which encapsulates ignoring already-marked
references).


Objects are no longer responsible for actually setting their own
mark bits, the collector does that. This means that for objects on
the number heap we don't have to call markChildren() at all since
we know there aren't any.


The mark phase of collect pushes roots onto the mark stack
and drains it as often as possible.


To make this approach viable requires a constant-size mark stack
and a slow fallback approach for when the stack size is exceeded,
plus optimizations to make the required stack small in common
cases. This should be doable.

  • JavaScriptCore.exp: Export new symbols.
  • JavaScriptCore.xcodeproj/project.pbxproj: Add new file.
  • kjs/collector.cpp: (KJS::Collector::heapAllocate): (KJS::drainMarkStack): Helper for all of the below. (KJS::Collector::markStackObjectsConservatively): Use mark stack. (KJS::Collector::markCurrentThreadConservatively): ditto (KJS::Collector::markOtherThreadConservatively): ditto (KJS::Collector::markProtectedObjects): ditto (KJS::Collector::markMainThreadOnlyObjects): ditto (KJS::Collector::collect): ditto
  • kjs/collector.h: (KJS::Collector::cellMayHaveRefs): Helper for MarkStack.
  • kjs/MarkStack.h: Added. The actual mark stack implementation. (KJS::MarkStack::push): (KJS::MarkStack::pushAtom): (KJS::MarkStack::pop): (KJS::MarkStack::isEmpty): (KJS::MarkStack::reserveCapacity):

Changed mark() methods to markChildren() methods:


  • kjs/ExecState.cpp: (KJS::ExecState::markChildren):
  • kjs/ExecState.h:
  • kjs/JSWrapperObject.cpp: (KJS::JSWrapperObject::markChildren):
  • kjs/JSWrapperObject.h:
  • kjs/array_instance.cpp: (KJS::ArrayInstance::markChildren):
  • kjs/array_instance.h:
  • kjs/bool_object.cpp: (BooleanInstance::markChildren):
  • kjs/bool_object.h:
  • kjs/error_object.cpp:
  • kjs/error_object.h:
  • kjs/function.cpp: (KJS::FunctionImp::markChildren): (KJS::Arguments::Arguments): (KJS::Arguments::markChildren): (KJS::ActivationImp::markChildren):
  • kjs/function.h:
  • kjs/internal.cpp: (KJS::GetterSetterImp::markChildren):
  • kjs/interpreter.cpp: (KJS::Interpreter::markRoots):
  • kjs/interpreter.h:
  • kjs/list.cpp: (KJS::List::markProtectedListsSlowCase):
  • kjs/list.h: (KJS::List::markProtectedLists):
  • kjs/object.cpp: (KJS::JSObject::markChildren):
  • kjs/object.h: (KJS::ScopeChain::markChildren):
  • kjs/property_map.cpp: (KJS::PropertyMap::markChildren):
  • kjs/property_map.h:
  • kjs/scope_chain.h:
  • kjs/string_object.cpp: (KJS::StringInstance::markChildren):
  • kjs/string_object.h:

JavaScriptGlue:

Reviewed by Darin and Geoff.

Fixups for JavaScriptCore mark stack.

  • JSObject.cpp: (JSUserObject::Mark):
  • JSObject.h:
  • JSValueWrapper.cpp: (JSValueWrapper::JSObjectMark):
  • JSValueWrapper.h:
  • UserObjectImp.cpp:
  • UserObjectImp.h:

WebCore:

Reviewed by Darin and Geoff.

Implement mark stack. This version is not suitable for prime time because it makes a
huge allocation on every collect, and potentially makes marking of detached subtrees
slow. But it is a .2% - .4% speedup even without much tweaking.

I replaced mark() methods with markChildren() as usual. One
optimization that is lost is avoiding walking detached DOM
subtrees more than once to mark them; since marking is not
recursive there's no obvious way to bracket operation on the tree
any more.

  • bindings/js/JSDocumentCustom.cpp: (WebCore::JSDocument::markChildren):
  • bindings/js/JSNodeCustom.cpp: (WebCore::JSNode::markChildren):
  • bindings/js/JSNodeFilterCondition.cpp:
  • bindings/js/JSNodeFilterCondition.h:
  • bindings/js/JSNodeFilterCustom.cpp: (WebCore::JSNodeFilter::markChildren):
  • bindings/js/JSNodeIteratorCustom.cpp: (WebCore::JSNodeIterator::markChildren):
  • bindings/js/JSTreeWalkerCustom.cpp: (WebCore::JSTreeWalker::markChildren):
  • bindings/js/JSXMLHttpRequest.cpp: (KJS::JSXMLHttpRequest::markChildren):
  • bindings/js/JSXMLHttpRequest.h:
  • bindings/js/kjs_binding.cpp: (KJS::ScriptInterpreter::markDOMNodesForDocument):
  • bindings/js/kjs_binding.h:
  • bindings/js/kjs_events.cpp: (WebCore::JSUnprotectedEventListener::markChildren):
  • bindings/js/kjs_events.h:
  • bindings/js/kjs_window.cpp: (KJS::Window::markChildren):
  • bindings/js/kjs_window.h:
  • bindings/scripts/CodeGeneratorJS.pm:
  • dom/Node.cpp: (WebCore::Node::Node):
  • dom/Node.h:
  • dom/NodeFilter.h:
  • dom/NodeFilterCondition.h:

LayoutTests:

Not reviewed.


I have fixed this with the mark stack work.


  • fast/js/gc-breadth-2-expected.txt: Added.
  • fast/js/gc-breadth-2.html: Added.
  • fast/js/gc-breadth-expected.txt: Added.
  • fast/js/gc-breadth.html: Added.
  • fast/js/gc-depth-expected.txt: Added.
  • fast/js/gc-depth.html: Added.
  • fast/js/resources/gc-breadth-2.js: Added.
  • fast/js/resources/gc-breadth.js: Added.
  • fast/js/resources/gc-depth.js: Added.
1:24 AM Changeset in webkit [28105] by mjs@apple.com
  • 1 edit in trunk/WebKitTools/ChangeLog

Not reviewer for this patch.

1:04 AM Changeset in webkit [28104] by mitz@apple.com
  • 2 edits in trunk/WebKit/mac

Reviewed by Maciej Stachowiak.

  • fix <rdar://problem/5596160> fast/events/objc-event-api.html fails when run alone (or first)
  • WebView/WebHTMLView.mm: (-[WebHTMLView setDataSource:]): This method calls addMouseMovedObserver because addMouseMovedObserver returns early if the dataSource is not nil. But if the dataSource is already set (which happens when a WebHTMLView is being reused) then addMouseMovedObserver must not be called again.
12:27 AM Changeset in webkit [28103] by alp@webkit.org
  • 3 edits
    1 add in trunk

2007-11-27 Alp Toker <​alp@atoker.com>

Reviewed by Mark Rowe.

Install the JavaScriptCore headers as part of the GTK+ port.

Note: See TracTimeline for information about the timeline view.