14 years ago2008-01-11 Alp Toker <alp@atoker.com>
alp@webkit.org [Sat, 12 Jan 2008 07:40:52 +0000 (07:40 +0000)]
2008-01-11  Alp Toker  <alp@atoker.com>

        Reviewed by Oliver Hunt.

        Enable FastMalloc by default in GTK+/autotools and add a configure

        * GNUmakefile.am:
        * configure.ac:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29429 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
ggaren@apple.com [Sat, 12 Jan 2008 07:19:27 +0000 (07:19 +0000)]

        Reviewed by Oliver Hunt.

        Fixed <rdar://problem/5665251> REGRESSION (r28880-r28886): Global
        variable access (16644)

        This bug was caused by var declarations shadowing built-in properties of
        the global object.

        To match Firefox, we've decided that var declarations will never shadow
        built-in properties of the global object or its prototypes. We used to
        behave more like IE, which allows shadowing, but walking that line got
        us into trouble with websites that sent us down the Firefox codepath.

        * kjs/JSVariableObject.h:
        (KJS::JSVariableObject::symbolTableGet): New code to support calling
        hasProperty before the variable object is fully initialized (so you
        can call it during initialization).

        * kjs/nodes.cpp:.
        (KJS::ProgramNode::initializeSymbolTable): Always do a full hasProperty
        check when looking for duplicates, not getDirect, since it only checks
        the property map, and not hasOwnProperty, since it doesn't check
        (KJS::EvalNode::processDeclarations): ditto

        * kjs/property_slot.h:
        (KJS::PropertySlot::ungettableGetter): Best function name evar.


        Reviewed by Oliver Hunt.

        Fixed <rdar://problem/5665251> REGRESSION (r28880-r28886): Global
        variable access (16644)

        Removed the ReadOnly bit from some properties, to match Firefox. Also
        removed status-related setters, to allow using their names as variable

        * bindings/scripts/CodeGeneratorJS.pm: Added support for properties that
        are one-way across domain boundaries, to match Firefox.

        * bindings/js/kjs_window.cpp: Changed ReadOnly declarations to match FF.

        * bindings/scripts/CodeGeneratorJS.pm: Don't use JSObject:: because
        we don't know that JSObject is our base class.

        * page/DOMWindow.idl: Replaced lots of readonly declarations with
        [Replaceable] declarations.

        * page/DOMWindow.h: Removed interfaces for setting status text via the
        DOM. (They were getting in the way of, e.g., "var status"
        declarations.) By default, IE 7 and FF disable these interfaces in order
        to defend against phishing attacks that try to spoof domain names in the
        * page/DOMWindow.cpp:


        Reviewed by Oliver Hunt.

        Fixed <rdar://problem/5665251> REGRESSION (r28880-r28886): Global
        variable access (16644)

        Added a test. Updated other tests to match new behavior.

        * fast/js/var-declarations-shadowing-expected.txt: Added.
        * fast/js/var-declarations-shadowing.html: Added.

        * fast/dom/HTMLScriptElement/script-load-events.html: Changed this test
        a bit because the original design made it hard to understand why it was failing.
        * fast/dom/HTMLScriptElement/script-load-events-expected.txt:

        * fast/dom/Window/get-set-properties.html: Changed this test to expect
        our new behavior, which matches Firefox.
        * fast/dom/Window/get-set-properties-expected.txt:

        * fast/dom/Window/window-property-shadowing.html: Removed some cases
        that differed from Firefox.
        * fast/dom/Window/window-property-shadowing-expected.txt:

        * http/tests/security/cross-frame-access-put-expected.txt: This test emits
        more "Unsafe JavaScript attempt" messages now because property sets that
        used to be prohibited (somewhat accidentally) by the ReadOnly attribute
        are now prohibited by security checks.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29428 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
mitz@apple.com [Sat, 12 Jan 2008 05:50:36 +0000 (05:50 +0000)]

2008-01-11  Anyang Ren  <anyang.ren@gmail.com>

        Reviewed by Darin Adler.

        The view source mode should skip an empty attribute value only if
        the attribute name is not followed by an equal sign (=).

        Test: fast/frames/viewsource-empty-attribute-value.html

        * html/HTMLViewSourceDocument.cpp:


2008-01-11  Anyang Ren  <anyang.ren@gmail.com>

        Reviewed by Darin Adler.

        Add a new regression test for
        The view source mode should skip an empty attribute value only if
        the attribute name is not followed by an equal sign (=).

        * fast/frames/resources/viewsource-frame-1.html: Added.
        * fast/frames/viewsource-empty-attribute-value-expected.txt: Added.
        * fast/frames/viewsource-empty-attribute-value.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29427 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Sylvain Pasche <sylvain.pasche@gmail.com>
alp@webkit.org [Sat, 12 Jan 2008 02:22:48 +0000 (02:22 +0000)]
2008-01-11  Sylvain Pasche  <sylvain.pasche@gmail.com>

        Reviewed by Alp Toker.

        [Gtk] Uneven glyph spacing with subpixel antialiasing

        Use cairo font options from the default GDK screen when creating a
        scaled font.

        * platform/graphics/gtk/FontPlatformDataGtk.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29426 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
mjs@apple.com [Sat, 12 Jan 2008 02:08:50 +0000 (02:08 +0000)]

        Reviewed by Maciej.

        Optimized ActivationImp allocation, so that activation records are now
        first allocated on an explicitly managed stack and only heap allocated
        when necessary. Roughly a 5% improvement on SunSpider, and a larger
        improvement on benchmarks that use more function calls.

        * JavaScriptCore.xcodeproj/project.pbxproj:
        * kjs/Activation.h: Added.
        * kjs/ExecState.cpp:
        * kjs/ExecState.h:
        * kjs/JSGlobalObject.cpp:
        * kjs/JSGlobalObject.h:
        * kjs/JSVariableObject.h:
        * kjs/function.cpp:
        * kjs/function.h:
        * kjs/nodes.cpp:
        * kjs/object.h:
        * kjs/scope_chain.h:
        * kjs/scope_chain_mark.h: Added.


        Reviewed by Maciej.

        Added a new forwarding header, because Activation.h has been separated
        from function.h

        * ForwardingHeaders/kjs/Activation.h: Added.


        Reviewed by Maciej.

        Added a test case that came up when developing the ActivationImp tear-off.

        * fast/js/resources/vardecl-preserve-arguments.js:
        * fast/js/vardecl-preserve-arguments-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29425 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago <rdar://problem/5541054> editing/pasteboard/paste-RTFD.html failing on Tiger
justin.garcia@apple.com [Sat, 12 Jan 2008 01:52:54 +0000 (01:52 +0000)]
    <rdar://problem/5541054> editing/pasteboard/paste-RTFD.html failing on Tiger

        * editing/pasteboard/paste-RTFD.html: Mention the Tiger bug that this demonstrates.
        * platform/mac-leopard/editing/pasteboard/paste-RTFD-expected.checksum: Added.
        * platform/mac-leopard/editing/pasteboard/paste-RTFD-expected.png: Added.
        * platform/mac-leopard/editing/pasteboard/paste-RTFD-expected.txt: Updated expected results
        to reflect the new info included in the test.
        * platform/mac/editing/pasteboard/paste-RTFD-expected.checksum: Ditto.
        * platform/mac/editing/pasteboard/paste-RTFD-expected.png:
        * platform/mac/editing/pasteboard/paste-RTFD-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29424 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Luca Bruno <lethalman88@gmail.com>
alp@webkit.org [Sat, 12 Jan 2008 01:36:02 +0000 (01:36 +0000)]
2008-01-11  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        [cURL] Allow multiple files for upload

        * platform/network/ResourceHandleInternal.h:
        * platform/network/curl/ResourceHandleCurl.cpp:
        * platform/network/curl/ResourceHandleManager.cpp:
        (WebCore::readCallback): added
        (WebCore::ResourceHandleManager::setupPOST): setup for streaming the POST
        (WebCore::ResourceHandleManager::cancel): revert the previous patch for regression
        * platform/network/curl/ResourceHandleManager.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29423 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Christian Dywan <christian@imendio.com>
alp@webkit.org [Sat, 12 Jan 2008 00:25:41 +0000 (00:25 +0000)]
2008-01-11  Christian Dywan  <christian@imendio.com>

        Reviewed by Alp Toker.

        [Gtk] Menu items need underscores

        Add underscores to appropriate menu labels.
        Also adjust strings slightly.

        * platform/gtk/ContextMenuItemGtk.cpp:
        * platform/gtk/LocalizedStringsGtk.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29422 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago <rdar://problem/5681557> On Windows Safari, mouse events are ignored after...
adachan@apple.com [Fri, 11 Jan 2008 23:56:40 +0000 (23:56 +0000)]
    <rdar://problem/5681557> On Windows Safari, mouse events are ignored after clicking on link that triggers download
        Moved the call to cache page from provisionalLoadStarted() to commitProvisionalLoad(), since
        provisionalLoadStarted() can be called for cases that do not result in a page navigation, for example,
        when a link to download a file has been clicked.

        Reviewed by John and Anders.

        * loader/FrameLoader.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29421 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Jon Honeycutt <jhoneycutt@apple.com>
jhoneycutt@apple.com [Fri, 11 Jan 2008 22:42:12 +0000 (22:42 +0000)]
2008-01-11  Jon Honeycutt  <jhoneycutt@apple.com>

        * platform/win/Skipped: Remove a test that was fixed in r29419.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29420 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Jon Honeycutt <jhoneycutt@apple.com>
jhoneycutt@apple.com [Fri, 11 Jan 2008 22:38:31 +0000 (22:38 +0000)]
2008-01-11  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Anders.

        <rdar://problem/5683529> plugins/embed-inside-object.html is timing

        Revert to using FrameTree::find() so that frame aliases like "_self" and
        "_current" are interpreted properly

        * plugins/win/PluginViewWin.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29419 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Alp Toker <alp@atoker.com>
alp@webkit.org [Fri, 11 Jan 2008 21:58:11 +0000 (21:58 +0000)]
2008-01-11  Alp Toker  <alp@atoker.com>

        Reviewed by Dave Hyatt and Mark Rowe.

        [GTK] Support custom fonts, CachedFont::platformDataFromCustomData()

        Add support for CSS2 @font-face custom/downloadable fonts to the GTK+

        * GNUmakefile.am:
        * WebCore.pro:
        * loader/CachedFont.cpp:
        * platform/graphics/gtk/FontCustomPlatformData.cpp: Added.
        * platform/graphics/gtk/FontCustomPlatformData.h: Added.
        * platform/graphics/gtk/FontPlatformData.h:
        * platform/graphics/gtk/FontPlatformDataGtk.cpp:
        * platform/graphics/gtk/SimpleFontDataGtk.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29418 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben.
ap@webkit.org [Fri, 11 Jan 2008 21:04:36 +0000 (21:04 +0000)]
    Reviewed by Adam Roben.

        <rdar://problem/5667275> fast/dynamic/layer-hit-test-crash.html is failing

        * DumpRenderTree/win/DumpRenderTree.cpp: (runTest): Ignore WM_MOUSELEAVE events,
        as these are only posted because the test window is not a normal visible one, and
        they confuse drag&drop machinery.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29417 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Remove FrameLoader::committedFirstRealDocumentLoad
aroben@apple.com [Fri, 11 Jan 2008 20:51:19 +0000 (20:51 +0000)]
     Remove FrameLoader::committedFirstRealDocumentLoad

         This method no longer has any callers.

         Reviewed by Hyatt.

         * loader/FrameLoader.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29416 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix Bug 16828: WebView never paints if first page load is canceled after...
aroben@apple.com [Fri, 11 Jan 2008 20:50:52 +0000 (20:50 +0000)]
     Fix Bug 16828: WebView never paints if first page load is canceled after being committed but before layout


         Now that we have 1) a real loader and 2) multiple FrameViews per
         Frame, we no longer need the hackery we had in our WM_PAINT handler
         that decided whether or not to paint. We now always want to paint,
         letting the swapping of FrameViews that happens in
         transitionToCommittedForNewPage handle which document gets painted.

         Reviewed by Hyatt.

         * WebView.cpp:
         (WebViewWndProc): Removed hackery that decided sometimes not to paint.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29415 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Add a standalone version of the blog post video player as a manual test.
antti@apple.com [Fri, 11 Jan 2008 20:19:13 +0000 (20:19 +0000)]
    Add a standalone version of the blog post video player as a manual test.

        * manual-tests/resources/touch-poster.png: Added.
        * manual-tests/video-player.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29414 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agotest is failing on trunk-win-release, so putting it back to make it green
alice.liu@apple.com [Fri, 11 Jan 2008 20:07:54 +0000 (20:07 +0000)]
test is failing on trunk-win-release, so putting it back to make it green

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29413 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Add shared vsprops to help unify our Windows tools settings.
sfalken@apple.com [Fri, 11 Jan 2008 19:47:02 +0000 (19:47 +0000)]
    Add shared vsprops to help unify our Windows tools settings.

        Reviewed by Hyatt.

        * win/tools: Added.
        * win/tools/WinTools.make: Added.
        * win/tools/vsprops: Added.
        * win/tools/vsprops/common.vsprops: Added.
        * win/tools/vsprops/debug.vsprops: Added.
        * win/tools/vsprops/release.vsprops: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29412 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoremoving test that doesn't fail
alice.liu@apple.com [Fri, 11 Jan 2008 19:29:29 +0000 (19:29 +0000)]
removing test that doesn't fail

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29411 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam.
antti@apple.com [Fri, 11 Jan 2008 19:19:57 +0000 (19:19 +0000)]
    Reviewed by Adam.

        Fix <rdar://problem/5682767>
        Video does not show up in http://webkit.org/blog/140/html5-media-support/ on Windows

        Take care that GWorld is created and deletes when needed as size or visibility changes.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29410 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix for bug 11188, setting hspace on a table overrides align=center. Fix...
hyatt@apple.com [Fri, 11 Jan 2008 18:27:18 +0000 (18:27 +0000)]
    Fix for bug 11188, setting hspace on a table overrides align=center.  Fix align=center to be done using
        mapped attributes so that it does not get overridden by hspace all the time.

        Reviewed by mitz

        Added fast/table/table-hspace-align-center.html

        * css/html4.css:
        * html/HTMLTableElement.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29409 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdd layout test for table bug.
hyatt@apple.com [Fri, 11 Jan 2008 18:25:44 +0000 (18:25 +0000)]
Add layout test for table bug.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29408 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agomoved couple of tests to a better category
alice.liu@apple.com [Fri, 11 Jan 2008 18:15:32 +0000 (18:15 +0000)]
moved couple of tests to a better category

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29407 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix a crash when pathToLocalResource fails and a leak
aroben@apple.com [Fri, 11 Jan 2008 18:14:34 +0000 (18:14 +0000)]
     Fix a crash when pathToLocalResource fails and a leak

         Reviewed by Darin.

         * DumpRenderTree/LayoutTestController.cpp:
         (pathToLocalResourceCallback): Dont leak the JSStringRef, and make
         sure not to pass null to JSValueMakeString.
         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
         (LayoutTestController::pathToLocalResource): Print an error message if
         the conversion fails so it's clear what happened.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29406 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago * platform/win/Skipped: Added some new failures.
aroben@apple.com [Fri, 11 Jan 2008 16:32:29 +0000 (16:32 +0000)]
    * platform/win/Skipped: Added some new failures.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29405 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Luca Bruno <lethalman88@gmail.com>
alp@webkit.org [Fri, 11 Jan 2008 16:31:34 +0000 (16:31 +0000)]
2008-01-11  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        [GTK] Bad autotools debug and video options

        Fix some configure options.

        * configure.ac:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29404 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoQt build fix.
mrowe@apple.com [Fri, 11 Jan 2008 12:40:32 +0000 (12:40 +0000)]
Qt build fix.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29403 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoReviewed by Mark Rowe.
hausmann@webkit.org [Fri, 11 Jan 2008 12:26:21 +0000 (12:26 +0000)]
Reviewed by Mark Rowe.

Fix the (clean) qmake build. For generating chartables.c we don't
depend on a separate input source file anymore, the dftables perl
script is enough. So use that instead as value for the .input
variable, to ensure that qmake also generates a rule to call dftables.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29402 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWindows build fix. Update to match method name in WebCore.
mrowe@apple.com [Fri, 11 Jan 2008 11:56:58 +0000 (11:56 +0000)]
Windows build fix.  Update to match method name in WebCore.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29401 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoQt build fix. Add link stubs for the new FileSystem.h functions introduced in r29399.
mrowe@apple.com [Fri, 11 Jan 2008 11:53:48 +0000 (11:53 +0000)]
Qt build fix.  Add link stubs for the new FileSystem.h functions introduced in r29399.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29400 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-11 Rodney Dawes <dobey@wayofthemonkey.com>
mrowe@apple.com [Fri, 11 Jan 2008 11:32:46 +0000 (11:32 +0000)]
2008-01-11  Rodney Dawes  <dobey@wayofthemonkey.com>

        Reviewed by Anders Carlsson.

        Bug 16779: Make the PluginStream implementation be shared across platforms

        Add the new shared PluginStream files to the GTK+ and Win32 builds.
        Add PluginStreamClient class for the streamDidFinishLoading method.
        Add open/close/write methods to FileSystem for temporary file handling.
        Add PluginDebug.h and npfunctions.h for shared PluginStream.
        Add shared PluginStream.
        Remove PluginStreamWin.
        Update PluginViewWin to use shared PluginStream and PluginStreamClass.

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * platform/FileSystem.h:
        * platform/gtk/FileSystemGtk.cpp:
        * platform/win/FileSystemWin.cpp:
        * plugins/PluginDebug.h:
        * plugins/PluginStream.cpp:
        * plugins/PluginStream.h:
        * plugins/npfunctions.h:
        * plugins/win/PluginDebug: Moved to PluginDebug.h.
        * plugins/win/PluginStreamWin.cpp: Moved to PluginStream.cpp.
        * plugins/win/PluginStreamWin.h: Moved to PluginStream.h
        * plugins/win/PluginViewWin.cpp:
        * plugins/win/PluginViewWin.h:
        * plugins/win/npfunctions.h: Moved to npfunctions.h

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29399 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Try to fix Mac build: Edit the right .exp file.
ggaren@apple.com [Fri, 11 Jan 2008 10:09:51 +0000 (10:09 +0000)]
    Try to fix Mac build: Edit the right .exp file.

        * WebCore.base.exp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29398 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Try to fix Qt build: don't use pthreads if they're not available.
ggaren@apple.com [Fri, 11 Jan 2008 10:03:38 +0000 (10:03 +0000)]
    Try to fix Qt build: don't use pthreads if they're not available.

        * bindings/js/GCController.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29397 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
ggaren@apple.com [Fri, 11 Jan 2008 09:52:27 +0000 (09:52 +0000)]

        Reviewed by John Sullivan.

        Fixed some world leak reports:
        * <rdar://problem/5669436> PLT complains about world leak of 1 JavaScript
        Interpreter after running cvs-base suite

        * <rdar://problem/5669423> PLT complains about world leak if browser
        window is open when PLT starts

        * kjs/collector.h: Added the ability to distinguish between global
        objects and GC-protected global objects, since we only consider the
        latter to be world leaks.
        * kjs/collector.cpp:


        Reviewed by John Sullivan.

        Fixed some world leak reports:
        * <rdar://problem/5669436> PLT complains about world leak of 1 JavaScript
        Interpreter after running cvs-base suite

        * <rdar://problem/5669423> PLT complains about world leak if browser
        window is open when PLT starts

        These were both bugs in the reporting mechanism, so I took the
        opportunity to do some house cleaning there.

        Stupid class, I kill you:
        * bridge/JavaScriptStatistics.cpp: Removed.
        * bridge/JavaScriptStatistics.h: Removed.

        * bindings/js/GCController.h: Adopted the only useful features of
        JavaScriptStatistics, since they were GC-related.
        * bindings/js/GCController.cpp:


        Reviewed by John Sullivan.

        Fixed some world leak reports:
        * <rdar://problem/5669436> PLT complains about world leak of 1 JavaScript
        Interpreter after running cvs-base suite

        * <rdar://problem/5669423> PLT complains about world leak if browser
        window is open when PLT starts

        These were both bugs in the reporting mechanism, so I took the
        opportunity to do some house cleaning there.

        * Misc/WebCoreStatistics.h: Did a little renaming, to match
        JavaScriptCore better. I kept the methods with the old names around,
        though, because old versions of Safari need them.

        * Misc/WebCoreStatistics.mm: Removed dependence on
        WebCore::JavaScriptStatistics, which is gone now.

        These two methods are now distinct, for the sake of world leak reporting:
        (+[WebCoreStatistics javaScriptGlobalObjectsCount]):
        (+[WebCoreStatistics javaScriptProtectedGlobalObjectsCount]):


        Reviewed by John Sullivan.

        Updated for transition away from JavaScriptStatistics, which is gone now.

        * WebJavaScriptCollector.cpp:


        Reviewed by John Sullivan.

        * DumpRenderTree/mac/GCControllerMac.mm:
        (GCController::collectOnAlternateThread): Updated for name change.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29396 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoSilence qmake warning about ctgen lacking input. Rubber-stamped by Alp Toker.
mrowe@apple.com [Fri, 11 Jan 2008 08:22:47 +0000 (08:22 +0000)]
Silence qmake warning about ctgen lacking input.  Rubber-stamped by Alp Toker.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29395 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAttempt to fix the GTK+ and Qt builds.
mrowe@apple.com [Fri, 11 Jan 2008 08:04:21 +0000 (08:04 +0000)]
Attempt to fix the GTK+ and Qt builds.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29394 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Not reviewed, build fix.
mjs@apple.com [Fri, 11 Jan 2008 07:50:44 +0000 (07:50 +0000)]
    Not reviewed, build fix.

        - more build fixing

        * WebDatabaseManager.cpp:
        * WebSecurityOrigin.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29393 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Not reviewed, build fix.
mjs@apple.com [Fri, 11 Jan 2008 07:30:41 +0000 (07:30 +0000)]
    Not reviewed, build fix.

        - try to fix build

        * WebDatabaseManager.cpp:
        * WebSecurityOrigin.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29392 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agowx build fix. Don't build dftables any longer since it's now a Perl script.
kevino@webkit.org [Fri, 11 Jan 2008 05:29:08 +0000 (05:29 +0000)]
wx build fix. Don't build dftables any longer since it's now a Perl script.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29391 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdd FIXME for replacing YYPARSE_PARAM with %parse-param to CSSGrammar.y
eric@webkit.org [Fri, 11 Jan 2008 05:12:34 +0000 (05:12 +0000)]
Add FIXME for replacing YYPARSE_PARAM with %parse-param to CSSGrammar.y

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29390 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adele.
eric@webkit.org [Fri, 11 Jan 2008 04:42:35 +0000 (04:42 +0000)]
    Reviewed by Adele.

        No functional changes, only code cleanup.

        * css/MediaQueryEvaluator.cpp:
        (WebCore::compareValue): renamed from cmpvalue

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29389 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Anders Carlsson.
mitz@apple.com [Fri, 11 Jan 2008 04:19:48 +0000 (04:19 +0000)]
    Reviewed by Anders Carlsson.

        - fix a crash when calling alert() from a repeating timer

        On non-Mac platforms, the PageGroupLoadDeferrer pauses DOM timers during
        alert() and other similar functions, which deletes the actual
        DOMWindowTimer objects and replaces them with new objects when resuming.

        * bindings/js/kjs_window.cpp:
        (KJS::Window::timerFired): Re-fetch the timer object from the map in
        case it has been deleted or replaced.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29388 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Not reviewed. Build fix.
mjs@apple.com [Fri, 11 Jan 2008 03:51:00 +0000 (03:51 +0000)]
    Not reviewed. Build fix.

        - Attempt to fix mac build.

        * Storage/WebDatabaseManager.mm:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29387 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
mjs@apple.com [Fri, 11 Jan 2008 03:36:14 +0000 (03:36 +0000)]

        Reviewed by Sam.

        - remove SecurityOriginData and fold its functionality into SecurityOrigin

        * GNUmakefile.am:
        * WebCore.base.exp:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * bindings/js/JSDOMWindowCustom.cpp:
        * dom/Document.cpp:
        * page/Chrome.cpp:
        * page/Chrome.h:
        * page/ChromeClient.h:
        * platform/SecurityOrigin.cpp:
        * platform/SecurityOrigin.h:
        * platform/SecurityOriginData.cpp: Removed.
        * platform/SecurityOriginData.h: Removed.
        * storage/Database.cpp:
        * storage/Database.h:
        * storage/DatabaseTracker.cpp:
        * storage/DatabaseTracker.h:
        * storage/DatabaseTrackerClient.h:
        * storage/SQLTransaction.cpp:
        * svg/graphics/SVGImageEmptyClients.h:


        Reviewed by Sam.

        - remove SecurityOriginData and fold its functionality into SecurityOrigin

        * WebCoreSupport/ChromeClientGtk.cpp:
        * WebCoreSupport/ChromeClientGtk.h:


        Reviewed by Sam.

        - remove SecurityOriginData and fold its functionality into SecurityOrigin

        * Storage/WebDatabaseManager.mm:
        (-[WebDatabaseManager origins]):
        (-[WebDatabaseManager databasesWithOrigin:]):
        (-[WebDatabaseManager detailsForDatabase:withOrigin:]):
        (-[WebDatabaseManager deleteDatabasesWithOrigin:]):
        (-[WebDatabaseManager deleteDatabase:withOrigin:]):
        * Storage/WebDatabaseTrackerClient.h:
        * Storage/WebDatabaseTrackerClient.mm:
        * Storage/WebSecurityOrigin.mm:
        (-[WebSecurityOrigin initWithProtocol:domain:port:]):
        (-[WebSecurityOrigin protocol]):
        (-[WebSecurityOrigin domain]):
        (-[WebSecurityOrigin port]):
        (-[WebSecurityOrigin usage]):
        (-[WebSecurityOrigin quota]):
        (-[WebSecurityOrigin setQuota:]):
        (-[WebSecurityOrigin isEqual:]):
        (-[WebSecurityOrigin dealloc]):
        (-[WebSecurityOrigin finalize]):
        (-[WebSecurityOrigin _initWithWebCoreSecurityOrigin:]):
        (-[WebSecurityOrigin _core]):
        * Storage/WebSecurityOriginInternal.h:
        * WebCoreSupport/WebChromeClient.h:
        * WebCoreSupport/WebChromeClient.mm:


        Reviewed by Sam.

        - remove SecurityOriginData and fold its functionality into SecurityOrigin

        * WebCoreSupport/ChromeClientQt.cpp:
        * WebCoreSupport/ChromeClientQt.h:


        Reviewed by Sam.

        - remove SecurityOriginData and fold its functionality into SecurityOrigin

        * WebChromeClient.cpp:
        * WebChromeClient.h:
        * WebDatabaseManager.cpp:
        * WebDatabaseManager.h:
        * WebSecurityOrigin.cpp:
        * WebSecurityOrigin.h:


        Reviewed by Sam.

        - remove SecurityOriginData and fold its functionality into SecurityOrigin

        * WebKitSupport/ChromeClientWx.cpp:
        * WebKitSupport/ChromeClientWx.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29386 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-10 Stephanie <slewis@apple.com>
slewis@apple.com [Fri, 11 Jan 2008 02:48:06 +0000 (02:48 +0000)]
2008-01-10  Stephanie  <slewis@apple.com>


        Move svg/batik/text/smallfonts.svg to the right skip list.

        * platform/mac-tiger/Skipped:
        * platform/mac/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29385 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Sam.
andersca@apple.com [Fri, 11 Jan 2008 02:09:55 +0000 (02:09 +0000)]
    Reviewed by Sam.

        Use the correct frame loader load method. Using the old method would not cause a
        new window to be open if the frame navigation was not allowed.

        * plugins/win/PluginViewWin.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29384 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-10 Alp Toker <alp@atoker.com>
alp@webkit.org [Fri, 11 Jan 2008 00:55:03 +0000 (00:55 +0000)]
2008-01-10  Alp Toker  <alp@atoker.com>

        SVG font build fix for GTK+/autotools.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29383 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoBUILD FIX: Make dftables executable.
ddkilzer@apple.com [Fri, 11 Jan 2008 00:41:44 +0000 (00:41 +0000)]
BUILD FIX: Make dftables executable.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29382 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
ddkilzer@apple.com [Fri, 11 Jan 2008 00:35:14 +0000 (00:35 +0000)]

        dftables should be rewritten as a script


        Reviewed by Darin.

        Rewrote the dftables utility in Perl.  Attempted to switch all
        build systems to call the script directly instead of building
        a binary first.  Only the Xcode build was able to be tested.

        * DerivedSources.make: Added pcre directory to VPATH and changed
        to invoke dftables directly.
        * GNUmakefile.am: Removed build information and changed to invoke
        dftables directly.
        * JavaScriptCore.vcproj/JavaScriptCore.sln: Removed reference to
        dftables project.
        * JavaScriptCore.vcproj/JavaScriptCoreSubmit.sln: Ditto.
        * JavaScriptCore.vcproj/dftables: Removed.
        * JavaScriptCore.vcproj/dftables/dftables.vcproj: Removed.
        * JavaScriptCore.xcodeproj/project.pbxproj: Removed dftables target.
        * jscore.bkl: Removed dftables executable definition.
        * pcre/dftables: Copied from JavaScriptCore/pcre/dftables.cpp.
        * pcre/dftables.cpp: Removed.
        * pcre/dftables.pro: Removed.
        * pcre/pcre.pri: Removed references to dftables.cpp and changed to
        invoke dftables directly.


        dftables should be rewritten as a script


        Reviewed by Darin.

        * WebKit.vcproj/WebKit.sln: Removed reference to dftables project.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29381 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
weinig@apple.com [Fri, 11 Jan 2008 00:23:13 +0000 (00:23 +0000)]

        Reviewed by Sam Weinig and Anders Carlsson.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=16522

        This patch makes two changes:

        1) Java calls FrameLoader::load in a slightly different way than
           JavaScript, which previously let a malicious web site bypass the
           shouldAllowNavigation check.  This patch adds that check to that
           code path.

        2) FrameLoader now wraps calls to m_frame->tree()->find(name) with
           findFrameForNavigation, which calls shouldAllowNavigation.  This
           treats disallowed frame navigations as if the named frame did not
           exist, resulting in a popup window when appropriate.

        Tests: http/tests/security/frameNavigation/xss-DENIED-plugin-navigation.html

        * WebCore.base.exp:
        * bindings/js/kjs_window.cpp:
        * loader/FrameLoader.cpp:
        * loader/FrameLoader.h:


        Reviewed by Anders Carlsson.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=16522

        * Plugins/WebBaseNetscapePluginView.mm:
        (-[WebBaseNetscapePluginView loadPluginRequest:]): call findFrameForNavigation
        to ensure the shouldAllowNavigation check is made.


        Reviewed by Anders Carlsson.

        Make DRT track open windows instead of allocated windows so that
        we can avoid ASSERTION due to late deallocs out of our control.

        * DumpRenderTree/mac/DumpRenderTree.mm:
        * DumpRenderTree/mac/DumpRenderTreeMac.h:
        * DumpRenderTree/mac/DumpRenderTreeWindow.h:
        * DumpRenderTree/mac/DumpRenderTreeWindow.mm:
        (+[DumpRenderTreeWindow openWindows]):
        (-[DumpRenderTreeWindow initWithContentRect:styleMask:backing:defer:]):
        (-[DumpRenderTreeWindow close]):
        * DumpRenderTree/mac/LayoutTestControllerMac.mm:


        Reviewed by Anders Carlsson.

        Tests for http://bugs.webkit.org/show_bug.cgi?id=16522

        * http/tests/security/frameNavigation/resources/frame-with-link-to-navigate.html: Added.
        * http/tests/security/frameNavigation/resources/frame-with-plugin-to-navigate.html: Added.
        * http/tests/security/frameNavigation/resources/navigation-happened.html: Added.
        * http/tests/security/frameNavigation/xss-DENIED-plugin-navigation-expected.txt: Added.
        * http/tests/security/frameNavigation/xss-DENIED-plugin-navigation.html: Added.
        * http/tests/security/frameNavigation/xss-DENIED-targeted-link-navigation-expected.txt: Added.
        * http/tests/security/frameNavigation/xss-DENIED-targeted-link-navigation.html: Added.
        * platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29380 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Written by Hyatt, reviewed by me
sullivan@apple.com [Fri, 11 Jan 2008 00:10:05 +0000 (00:10 +0000)]
    Written by Hyatt, reviewed by me

        - fixed <rdar://problem/5654297> Mail crashes occurs at WebCore::FontFallbackList::fontDataAt() when attempting to display
          a <video> element that uses controls attribute

        * rendering/MediaControlElements.cpp:
        force the render style to inherit from the media element's style; Hyatt filed 5682383 to cover cleaning
        up this architecture, but this one-line fix will prevent the crash in the meantime

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29379 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWhoops, did not mean to commit that.
andersca@apple.com [Fri, 11 Jan 2008 00:02:16 +0000 (00:02 +0000)]
Whoops, did not mean to commit that.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29378 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix fast/forms/input-radio-checked-tab.html
andersca@apple.com [Fri, 11 Jan 2008 00:00:57 +0000 (00:00 +0000)]
    Fix fast/forms/input-radio-checked-tab.html
        Meta key is not the same as Alt key on windows.

        Reviewed by Darin.

        * platform/win/KeyEventWin.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29377 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Meta key is not the same as Alt key on windows.
adachan@apple.com [Thu, 10 Jan 2008 23:11:54 +0000 (23:11 +0000)]
    Meta key is not the same as Alt key on windows.

        Reviewed by Darin.

        * DumpRenderTree/win/EventSender.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29376 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix fast/forms/input-radio-checked-tab.html
adachan@apple.com [Thu, 10 Jan 2008 23:11:27 +0000 (23:11 +0000)]
    Fix fast/forms/input-radio-checked-tab.html
        Meta key is not the same as Alt key on windows.

        Reviewed by Darin.

        * platform/win/KeyEventWin.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29375 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago * platform/win/Skipped: Removed a test that was fixed in r26826.
aroben@apple.com [Thu, 10 Jan 2008 22:17:26 +0000 (22:17 +0000)]
    * platform/win/Skipped: Removed a test that was fixed in r26826.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29374 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago * platform/win/Skipped: Removed a test that was fixed in r28372.
aroben@apple.com [Thu, 10 Jan 2008 22:14:14 +0000 (22:14 +0000)]
    * platform/win/Skipped: Removed a test that was fixed in r28372.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29373 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by John Sullivan.
andersca@apple.com [Thu, 10 Jan 2008 21:51:40 +0000 (21:51 +0000)]
    Reviewed by John Sullivan.

        REGRESSION: BumperCar crashes when attempting to load a long invalid URL

        Make sure to call FrameLoader::receivedMainResourceError before calling
        FrameLoader::didFailToLoad. The call to receivedMainResourceError takes care of
        clearing out the provisional document loader so that we won't call the didFail
        ResourceLoadDelegate method twice. This also makes us call the FrameLoadDelegate method
        didFailPrivisionalLoad before calling the ResourceLoadDelegate method, which Safari 2.0 does.

        * loader/MainResourceLoader.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29372 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adele.
antti@apple.com [Thu, 10 Jan 2008 21:48:18 +0000 (21:48 +0000)]
    Reviewed by Adele.

        Fix <rdar://problem/5658048>
        After <video> has finished playing, dragging the scroller on different location of controller starts to play movie

        - Move to paused state if the playback had ended and the controller is used to seek to earlier time
        - Pause video playback during drag so the knob does not constantly try to escape from the mouse pointer

        * html/HTMLMediaElement.cpp:
        * html/HTMLMediaElement.h:
        * rendering/MediaControlElements.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29371 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben.
ap@webkit.org [Thu, 10 Jan 2008 21:33:19 +0000 (21:33 +0000)]
    Reviewed by Adam Roben.

        <rdar://problem/5667003> fast/dom/xmlhttprequest-html-response-encoding.html is failing

        * xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::send): Do not disable sniffing for file://
        requests, as CFNetwork doesn't perform extension to MIME type mapping then.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29370 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoFix many bugs by giving Windows one FrameView per page load
aroben@apple.com [Thu, 10 Jan 2008 21:23:34 +0000 (21:23 +0000)]
Fix many bugs by giving Windows one FrameView per page load


        Fixes to allow multiple FrameViews on Windows

        Reviewed by Hyatt.

        * page/FrameView.cpp:
        (WebCore::FrameView::FrameView): Added a new constructor that takes an
        IntSize to specify the FrameView's initial size.
        (WebCore::FrameView::scheduleRelayout): Added an assertion that our
        Document is not in the page cache.
        * page/FrameView.h:
        * platform/gtk/WidgetGtk.cpp:
        (WebCore::Widget::~Widget): Add a warm, fuzzy ASSERT.
        * platform/qt/WidgetQt.cpp:
        (WebCore::Widget::~Widget): Ditto.
        * rendering/RenderWidget.cpp:
        (WebCore::RenderWidget::setWidget): Make sure to remove any existing
        Widget from the Widget hierarchy before deleting it. One instance
        where this is needed is when setWidget is called during FrameView
        creation on Windows.


        Fix many bugs by giving Windows one FrameView per page load

        Bugs include:
                Windows back/forward cache causes crashes in the layout tests
                REGRESSION: PLT broken on Windows due to back/forward cache
                Random crashes in FrameView::scheduleRelayout while surfing

        On Windows until now we've only had one FrameView per Frame. Once the
        back/forward cache was turned on this started causing assertions to
        fail and crashes to occur due to a single FrameView being both in the
        back/forward cache (possibly multiple times!) and used by a live
        document. We now create a new FrameView for each page load, just as
        Mac does.

        This has the side-effect of plugging some of the world leaks seen at
        the end of the PLT.

        Reviewed by Hyatt.

        * WebFrame.cpp:
        (WebFrame::initWithWebFrameView): Don't create the FrameView right
        away -- it'll be created when the load is committed.
        (WebFrame::transitionToCommittedFromCachedPage): Match the Mac by no
        longer calling resetMultipleFormSubmissionProtection here.
        (WebFrame::transitionToCommittedForNewPage): Ported code from
        -[WebCoreFrameBridge installInFrame:], and moved code here from
        WebFrame::initWithWebFrameView and WebView::initWithFrame. WebCore
        takes care of resetMultipleFormSubmissionProtection, just like it does
        on the Mac.
        * WebView.cpp:
        (WebView::initWithFrame): Moved FrameView initialization code to

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29369 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-10 Alp Toker <alp@atoker.com>
alp@webkit.org [Thu, 10 Jan 2008 21:10:31 +0000 (21:10 +0000)]
2008-01-10  Alp Toker  <alp@atoker.com>

        Include math.h to get ceilf(). Part of the SVG font GTK+ build fix.

        * svg/SVGFontFaceElement.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29368 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Alice Liu.
justin.garcia@apple.com [Thu, 10 Jan 2008 20:40:34 +0000 (20:40 +0000)]
    Reviewed by Alice Liu.

        <rdar://problem/5658603> Crash in InsertNodeBefore::doUnapply() on Undo in Yahoo Mail
        <rdar://problem/5658709> Crash in RenderView::setSelection on Undo in Yahoo Mail

        Make sure we have an updated layout before we perform any editing work.

        * editing/EditCommand.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29367 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-10 Luca Bruno <lethalman88@gmail.com>
alp@webkit.org [Thu, 10 Jan 2008 20:29:39 +0000 (20:29 +0000)]
2008-01-10  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        Back out r29206 which was causing regressions in curl http job

        * platform/network/curl/ResourceHandleManager.cpp

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29366 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
kmccullough@apple.com [Thu, 10 Jan 2008 19:33:44 +0000 (19:33 +0000)]

        Reviewed by Darin, Sam, and Adam.

        - <rdar://problem/5654486> REGRESSION (Safari 3.0.4-TOT): clicking on
        - link in gmail message displays JavaScript alert falsely complaining
        about pop-up blocking
        - When trying to open a new window, we now see if the user gesture
        occurred in the global dynamic object instead of the frame since that is
        the only place an event can occur.

        * bindings/js/kjs_window.cpp: - Check the dynamic global object instead
        of the frame.


        Reviewed by Darin, Sam and Adam.

        -<rdar://problem/5654486> REGRESSION (Safari 3.0.4-TOT): clicking on
        link in gmail message displays JavaScript alert falsely complaining
        about pop-up blocking.
        - Added the ability to enable the pop-up blocker via the

        * DumpRenderTree/LayoutTestController.cpp:
        * DumpRenderTree/LayoutTestController.h:
        * DumpRenderTree/mac/DumpRenderTree.mm:
        * DumpRenderTree/mac/LayoutTestControllerMac.mm:


        Reviewed by Darin, Sam, and Adam.

        -<rdar://problem/5654486> REGRESSION (Safari 3.0.4-TOT): clicking on
        link in gmail message displays JavaScript alert falsely complaining
        about pop-up blocking
        - When trying to open a new window, we now see if the user gesture
        occurred in the global dynamic object instead of the frame since that is
        the only place an event can occur.

        * fast/events/open-window-from-another-frame-expected.txt: Added.
        * fast/events/open-window-from-another-frame.html: Added.
        * fast/events/resources/open-window-from-another-frame-otherFrame.html: Added.
        * fast/events/resources/window-opened.html: Added.
        * platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29365 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix for bug 16247, visibility:hidden not honored when hit testing inline...
hyatt@apple.com [Thu, 10 Jan 2008 19:29:32 +0000 (19:29 +0000)]
    Fix for bug 16247, visibility:hidden not honored when hit testing inline replaced elements.

        Reviewed by adele

        Added fast/css/visibility-hit-test.html test case.

        * rendering/RenderBox.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29364 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdd new layout test for visibility bug.
hyatt@apple.com [Thu, 10 Jan 2008 19:28:14 +0000 (19:28 +0000)]
Add new layout test for visibility bug.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29363 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
mitz@apple.com [Thu, 10 Jan 2008 17:39:03 +0000 (17:39 +0000)]

        Reviewed by Darin Adler.

        - fix http://bugs.webkit.org/show_bug.cgi?id=16782
          <rdar://problem/5675331> REGRESSION(r29266): Reproducible crash in fast/replaced/image-map.html

        The crash resulted from a native object (DumpRenderTree's
        EventSender) causing its wrapper to be invalidated (by clicking a
        link that replaced the document in the window) and consequently
        deallocated. The fix is to use RefPtrs to protect the native object
        from deletion by self-invalidation.

        * bindings/runtime_method.cpp:
        * bindings/runtime_object.cpp:


        Reviewed by Darin Adler.

        - re-enable crashing test after fixing http://bugs.webkit.org/show_bug.cgi?id=16782
          <rdar://problem/5675331> REGRESSION(r29266): Reproducible crash in fast/replaced/image-map.html

        * fast/replaced/image-map-expected.txt: Updated results for the new
        behavior, which is to replace the document with the success message.
        * fast/replaced/image-map.html: Copied from LayoutTests/fast/replaced/image-map.html-disabled.
        * fast/replaced/image-map.html-disabled: Removed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29362 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agodocument QWebHistory
hausmann@webkit.org [Thu, 10 Jan 2008 16:47:59 +0000 (16:47 +0000)]
document QWebHistory

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29361 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMake the reset() functions not do anything on the default QWebSettings object.
hausmann@webkit.org [Thu, 10 Jan 2008 16:47:42 +0000 (16:47 +0000)]
Make the reset() functions not do anything on the default QWebSettings object.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29360 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agorename QWebPageHistory to QWebHistory.
hausmann@webkit.org [Thu, 10 Jan 2008 16:47:24 +0000 (16:47 +0000)]
rename QWebPageHistory to QWebHistory.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29359 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoDocumentation for QWebSettings
hausmann@webkit.org [Thu, 10 Jan 2008 16:39:06 +0000 (16:39 +0000)]
Documentation for QWebSettings

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29358 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoDocument QWebHistoryInterface.
hausmann@webkit.org [Thu, 10 Jan 2008 16:38:51 +0000 (16:38 +0000)]
Document QWebHistoryInterface.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29357 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoDocumentation for QWebFrame.
hausmann@webkit.org [Thu, 10 Jan 2008 16:38:33 +0000 (16:38 +0000)]
Documentation for QWebFrame.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29356 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agofix the drawing errors that where introduced due to refactoring.
hausmann@webkit.org [Thu, 10 Jan 2008 16:19:55 +0000 (16:19 +0000)]
fix the drawing errors that where introduced due to refactoring.

Correctly clip to the rectangle we want to draw in ScrollView::paint().

Signed-off-by: Simon
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29355 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdded a whole bunch of docs for QWebPage and fixed some minor doc glitches in QWebView.
hausmann@webkit.org [Thu, 10 Jan 2008 16:19:23 +0000 (16:19 +0000)]
Added a whole bunch of docs for QWebPage and fixed some minor doc glitches in QWebView.

Signed-off-by: Lars
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29354 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoTemporarily disable gzip decompression in qhttp due to a bug.
hausmann@webkit.org [Thu, 10 Jan 2008 16:19:04 +0000 (16:19 +0000)]
Temporarily disable gzip decompression in qhttp due to a bug.

Signed-off-by: Lars
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29353 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agohttp://bugs.webkit.org/show_bug.cgi?id=16588
hausmann@webkit.org [Thu, 10 Jan 2008 15:28:59 +0000 (15:28 +0000)]

Added a (last) Frame argument to ResourceHandle::loadResourceSynchronously.
This allows implementing the synchronous loading correctly for the Qt port where the networking
backend is bound to the page.

Signed-off-by: Lars
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29352 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by John Sullivan.
mitz@apple.com [Thu, 10 Jan 2008 15:22:07 +0000 (15:22 +0000)]
    Reviewed by John Sullivan.

        - make more tests pass on Windows by forcing font fallback to match
          Mac OS X.

        * fast/css/beforeSelectorOnCodeElement.html:
        * fast/css/rtl-ordering.html:
        * fast/css/text-security.html:
        * platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29351 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agowx build fix for changes in r29328
kevino@webkit.org [Thu, 10 Jan 2008 08:49:42 +0000 (08:49 +0000)]
wx build fix for changes in r29328

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29350 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Sam.
mjs@apple.com [Thu, 10 Jan 2008 03:48:41 +0000 (03:48 +0000)]
    Reviewed by Sam.

        - refactor SecurityOrigin in preparation for merging with SecurityOriginData

        * platform/SecurityOrigin.cpp:
        * platform/SecurityOrigin.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29349 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
adachan@apple.com [Thu, 10 Jan 2008 03:03:11 +0000 (03:03 +0000)]
        Tabs with ctrl, meta, or altgraph modifier key down should not advance focus.

        Reviewed by Darin.

        Test: fast/forms/tabs-with-modifiers.html

        * page/EventHandler.cpp:
        (WebCore::EventHandler::defaultTabEventHandler): bail if ctrl, meta, or altgraph key is down.
        Clean up the code a bit.
        * page/EventHandler.h:
        * page/FocusController.cpp: Remove the advanceFocus() that takes in a KeyboardEvent.  It was
        only called in EventHandler::defaultTabEventHandler() but we have cleaned up the code there and
        no longer needs it.
        * page/FocusController.h:

        Tests tabbing with different combinations of modifiers keys, where some
        should advance focus and some should not.

        Reviewed by Darin.

        * fast/forms/tabs-with-modifiers-expected.txt: Added.
        * fast/forms/tabs-with-modifiers.html: Copied from fast/forms/focus2.html.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29348 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix crash that could happen if the key we are passing to the HashMap is 0...
adachan@apple.com [Thu, 10 Jan 2008 02:57:42 +0000 (02:57 +0000)]
    Fix crash that could happen if the key we are passing to the HashMap is 0 in WebView::interpretKeyEvent().

        Reviewed by Darin.

        * WebView.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29347 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Darin.
mjs@apple.com [Thu, 10 Jan 2008 02:51:41 +0000 (02:51 +0000)]
    Reviewed by Darin.

        - add some more renames, adjust some

        * Scripts/do-webcore-rename:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29346 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Mitz.
antti@apple.com [Thu, 10 Jan 2008 01:41:13 +0000 (01:41 +0000)]
    Reviewed by Mitz.

        Fix http://bugs.webkit.org/show_bug.cgi?id=16376
        <video> element fails to play frames when navigating back to page (16376)

        Don't make MediaPlayer visible when it is actually in page cache.

        * rendering/RenderVideo.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29345 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Darin.
andersca@apple.com [Thu, 10 Jan 2008 01:38:51 +0000 (01:38 +0000)]
    Reviewed by Darin.

        CrashTracer: 9840 crashes in Safari at com.apple.JavaScriptCore: KJS::Bindings::CInstance::~CInstance [in-charge deleting] + 35

        Clear the frame's plugin root objects so that they don't outlive the plugin bundle.

        * page/Frame.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29344 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Darin Adler.
mitz@apple.com [Thu, 10 Jan 2008 01:37:32 +0000 (01:37 +0000)]
    Reviewed by Darin Adler.

        - move a WebKit test which was inadvertently added to a test suite
          directory into a custom test directory and change it to give the same
          results on Windows.

        * css1/text_properties/text-transCapitalize.html: Removed.
        * fast/text/capitalize-boundaries.html: Copied from LayoutTests/css1/text_properties/text-transCapitalize.html,
        updated the Bugzilla URL and added @font-face rules to make Windows
        results match Mac OS X.
        * platform/mac/css1/text_properties/text-transCapitalize-expected.checksum: Removed.
        * platform/mac/css1/text_properties/text-transCapitalize-expected.png: Removed.
        * platform/mac/css1/text_properties/text-transCapitalize-expected.txt: Removed.
        * platform/mac/fast/text/capitalize-boundaries-expected.checksum: Copied from LayoutTests/platform/mac/css1/text_properties/text-transCapitalize-expected.checksum.
        * platform/mac/fast/text/capitalize-boundaries-expected.png: Copied from LayoutTests/platform/mac/css1/text_properties/text-transCapitalize-expected.png.
        * platform/mac/fast/text/capitalize-boundaries-expected.txt: Copied from LayoutTests/platform/mac/css1/text_properties/text-transCapitalize-expected.txt.
        * platform/qt/css1/text_properties/text-transCapitalize-expected.txt: Removed.
        * platform/win/Skipped: Removed css1/text_properties/text-transCapitalize.html.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29343 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben and Anders Carlsson
sullivan@apple.com [Wed, 9 Jan 2008 23:02:07 +0000 (23:02 +0000)]
    Reviewed by Adam Roben and Anders Carlsson

        - fixed <rdar://problem/5469398> Repro assertion failure in context menu code due to
          missing-but-expected Reload item

        * platform/ContextMenu.cpp:
        use isLoadingInAPISense when deciding whether to include Stop or Reload in context
        menu, to match the WebKit API

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29342 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-09 Luca Bruno <lethalman88@gmail.com>
alp@webkit.org [Wed, 9 Jan 2008 20:26:47 +0000 (20:26 +0000)]
2008-01-09  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        [GTK] Missing gtk properties

        Add missing properties to WebKitViewFrame and WebKitWebView.

        * WebView/webkitprivate.h: add some useful defines for param specs
        * WebView/webkitwebframe.cpp:
        (webkit_web_frame_get_property): added
        (webkit_web_frame_class_init): add name, title and uri read-only properties
        * WebView/webkitwebview.cpp:
        (webkit_web_view_get_property): add editable property read
        (webkit_web_view_set_property): added for editable property write
        (webkit_web_view_class_init): add read-write editable property

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29341 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben.
ap@webkit.org [Wed, 9 Jan 2008 20:13:07 +0000 (20:13 +0000)]
    Reviewed by Adam Roben.

        <rdar://problem/5673716> fast/loader/opaque-base-url.html is failing
        <rdar://problem/5670275> fast/loader/onunload-form-submit-crash.html is failing
        <rdar://problem/5667277> fast/loader/onunload-form-submit-crash-2.html is failing

        * fast/loader/meta-refresh-vs-open.html: Changed to be more DRT-friendly.
        * platform/win/Skipped: Removed tests that were victims of the above.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29340 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben.
ap@webkit.org [Wed, 9 Jan 2008 20:10:33 +0000 (20:10 +0000)]
    Reviewed by Adam Roben.

        Fix fast/encoding/frame-default-enc.html, which is failing on buildbot.

        Internal frame names for this test are changing for no apparent reason (see r25269).
        Make the test independent of the names.

        * fast/encoding/frame-default-enc-expected.txt:
        * fast/encoding/frame-default-enc.html:
        * fast/encoding/resources/frame-default-enc-frame.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29339 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoFix Windows debug build for opensource developers.
mrowe@apple.com [Wed, 9 Jan 2008 18:03:06 +0000 (18:03 +0000)]
Fix Windows debug build for opensource developers.

* WebCore.vcproj/QTMovieWin.vcproj: Use the correct library suffix.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29338 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Update some test results
aroben@apple.com [Wed, 9 Jan 2008 18:01:20 +0000 (18:01 +0000)]
     Update some test results

         Reviewed by Anders.

         * fast/dom/Window/window-properties-expected.txt: Updated after
         * platform/win/media/audio-controls-rendering-expected.txt:
         Regenerated results with the correct fonts.
         * platform/win/media/video-aspect-ratio-expected.txt: Ditto.
         * platform/win/media/video-controls-rendering-expected.txt: Ditto.
         * platform/win/media/video-transformed-expected.txt: Ditto.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29337 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reworked by Eric, Reviewed by Hyatt.
eric@webkit.org [Wed, 9 Jan 2008 09:23:16 +0000 (09:23 +0000)]
    Reworked by Eric, Reviewed by Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=9454
        Add support for :lang inheritance and xml:lang support.

        Tests: fast/selectors/lang-inheritance.html

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::canShareStyleWithElement): Do not share style between elements with
        different LANG-attribute.
        (WebCore::CSSStyleSelector::checkOneSelector): Change :lang() to recursively check the LANG attribute
        for all the elements parents and the content-language of the document.
        * dom/Document.cpp:
        (WebCore::Document::processHttpEquiv): Parse MIME Content-Language
        * dom/Document.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29336 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Brady.
timothy@apple.com [Wed, 9 Jan 2008 04:49:23 +0000 (04:49 +0000)]
    Reviewed by Brady.

        Bug 16678: Unreproducible crash in KJS::JSObject::inherits() after using Web Inspector

        Add a NULL check for controller before calling JSObjectSetPrivate.

        * page/InspectorController.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29335 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-08 Xan Lopez <xan@gnome.org>
alp@webkit.org [Wed, 9 Jan 2008 04:45:32 +0000 (04:45 +0000)]
2008-01-08  Xan Lopez  <xan@gnome.org>

        Reviewed by Alp Toker.

        [GTK] Text rendering using Pango

        Use Pango to render Complex path text.

        * platform/graphics/gtk/FontGtk.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29334 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Darin Adler.
timothy@apple.com [Wed, 9 Jan 2008 04:35:32 +0000 (04:35 +0000)]
    Reviewed by Darin Adler.

        <rdar://problem/5665860> With the web inspector displayed, a crash occurs
        at WebCore::Frame::document() when navigating back to previous page

        This fixes the crash, but the inspector was totally broken with back/forward.
        So this also fixes back/forward navigation so the right main resource shows
        up in the inspector.

        * page/InspectorController.cpp:
        (WebCore::addSourceToFrame): Add some null checks for the frame when
        getting the textEncoding. This was the crash.
        (WebCore::InspectorController::addScriptResource): Create a script object
        only if needed, and always add it by calling addResource.
        (WebCore::InspectorController::didCommitLoad): Check if the loader is
        loading from the page cache, and clear m_mainResource. If the load is
        normal, then call addAndUpdateScriptResource with the main resource.
        (WebCore::InspectorController::identifierForInitialRequest): If the load
        is from the page cache and the resource is the main resource call
        addAndUpdateScriptResource since didCommitLoad did not do it.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29333 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-08 Alp Toker <alp@atoker.com>
alp@webkit.org [Wed, 9 Jan 2008 04:15:09 +0000 (04:15 +0000)]
2008-01-08  Alp Toker  <alp@atoker.com>

        Back out VIDEO by default in the GTK+ qmake build. The build bot
        doesn't have the necessary libraries installed.

        * WebCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29332 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-08 Alp Toker <alp@atoker.com>
alp@webkit.org [Wed, 9 Jan 2008 03:41:13 +0000 (03:41 +0000)]
2008-01-08  Alp Toker  <alp@atoker.com>

        Win build fix for breakage introduced in r29328.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29331 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-08 Alp Toker <alp@atoker.com>
alp@webkit.org [Wed, 9 Jan 2008 03:37:45 +0000 (03:37 +0000)]
2008-01-08  Alp Toker  <alp@atoker.com>

        Rubber-stamped by Mark Rowe.

        Enable VIDEO by default in the GTK+ qmake build.

        * WebCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@29330 268f45cc-cd09-0410-ab3c-d52691b4dbfc