Strip carriage returns from results

        Carriage returns were screwing up sunspider-compare-results on
        Windows.

        Rubberstamped by Oliver.

        * sunspider:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27075 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Dave Hyatt and Sam Weinig.

        Build fix

        * platform/graphics/cg/ImageCG.cpp:
        (WebCore::Image::drawPattern):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27074 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix the windows build.

        * win/lib/WebKitSystemInterface.lib:
        * win/lib/WebKitSystemInterface_debug.lib:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27073 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver Hunt.

        Rolled out my last patch. It turns out that I needed 2 words, not 1,
        so it didn't help.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27072 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Add WKSetPatternBaseCTM.

        * WebKitSystemInterface.h:
        * libWebKitSystemInterfaceLeopard.a:
        * libWebKitSystemInterfaceTiger.a:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27071 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove the empty WebKitQt folders.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27070 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-23  Jan Michael Alonzo  <jmalonzo@unpluggable.com>

        Reviewed by Alp.

        http://bugs.webkit.org/show_bug.cgi?id=15656
        [GTK] Implement WebCore::Widget::isEnabled/setEnabled

        * platform/gtk/WidgetGtk.cpp:
        (WebCore::Widget::isEnabled):
        (WebCore::Widget::setEnabled): Implemented.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27069 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver Hunt.

        Fixed http://bugs.webkit.org/show_bug.cgi?id=15694
        Shrink the size of an activation object by 1 word

        This is in preparation for adding a symbol table to the activation
        object.

        The basic strategy here is to rely on the mutual exclusion between
        the arguments object pointer and the function pointer (you only need
        the latter in order to create the former), and store them in the same
        place. The LazyArgumentsObject class encapsulates this strategy.

        Also inlined the ArgumentsImp constructor, for good measure.

        SunSpider reports no regression. Regression tests pass.

        * JavaScriptCore.xcodeproj/project.pbxproj:
        * kjs/Context.cpp:
        (KJS::Context::~Context):
        * kjs/function.cpp:
        (KJS::ActivationImp::LazyArgumentsObject::createArgumentsObject):
        (KJS::ActivationImp::LazyArgumentsObject::mark):
        (KJS::):
        (KJS::ActivationImp::argumentsGetter):
        (KJS::ActivationImp::mark):
        * kjs/function.h:
        (KJS::ActivationImp::LazyArgumentsObject::LazyArgumentsObject):
        (KJS::ActivationImp::LazyArgumentsObject::getOrCreate):
        (KJS::ActivationImp::LazyArgumentsObject::resetArguments):
        (KJS::ActivationImp::LazyArgumentsObject::setArgumentsObject):
        (KJS::ActivationImp::LazyArgumentsObject::argumentsObject):
        (KJS::ActivationImp::LazyArgumentsObject::setFunction):
        (KJS::ActivationImp::LazyArgumentsObject::function):
        (KJS::ActivationImp::LazyArgumentsObject::createdArgumentsObject):
        (KJS::ActivationImp::LazyArgumentsObject::):
        (KJS::ActivationImp::ActivationImp::ActivationImp):
        (KJS::ActivationImp::resetArguments):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27068 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Weinig also reviewed this patch, and wanted Changelog credit for it

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27067 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam.

        - Implemented server calls now the WebScriptDebugServer exists.
        - Also removed no longer needed call to initialize CG.

        * Drosera/win/Drosera.cpp: No longer initializes CG because this happens
        automatically now.
        (_tWinMain):
        * Drosera/win/ServerConnection.cpp: Now uses the COM class.
        (ServerConnection::ServerConnection):
        (ServerConnection::~ServerConnection):
        (ServerConnection::pause):
        (ServerConnection::resume):
        (ServerConnection::stepInto):
        (ServerConnection::applicationTerminating):
        (ServerConnection::serverConnectionDidDie):
        * Drosera/win/ServerConnection.h: Now uses the COM class.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27066 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add border image transforms test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27065 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add border image transforms test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27064 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add background transform test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27063 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add background transform test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27062 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix some JavaScriptCore build issues

JavaScriptCore:

        Change JavaScriptCore.vcproj to use DerivedSources.make

        We were trying to emulate the logic of make in
        build-generated-files.sh, but we got it wrong. We now use a
        build-generated-files very much like the one that WebCore uses to
        invoke make.

        We also now only have a Debug configuration of dftables which we build
        even when doing a Release build of JavaScriptCore. dftables also no
        longer has the "_debug" name suffix.

        Changes mostly made by Darin, reviewed by me.

        * DerivedSources.make: Add a variable to set the extension used for
        the dftables executable.
        * JavaScriptCore.vcproj/JavaScriptCore.sln: Updated to use Debug
        dftables in Release configurations.
        * JavaScriptCore.vcproj/JavaScriptCoreSubmit.sln: Ditto.
        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
           - Updated include path to point to the new location of the derived
             sources.
           - Modified pre-build event to pass the right arguments to
             build-generated-files.sh and not call dftables directly.
           - Added the derived source files to the project.
           - Removed grammarWrapper.cpp, which isn't needed now that we're
             compiling grammar.cpp directly.
        * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
        Slightly modified from the WebCore version.
        * JavaScriptCore.vcproj/JavaScriptCore/grammarWrapper.cpp: Removed.
        * JavaScriptCore.vcproj/dftables/dftables.vcproj:
            - Changed the output location to match Mac.
            - Removed the Release configuration.
            - Removed the _debug suffix.

win:

        Update WebKit.sln for the removal of Release dftables

        Reviewed by NOBODY.

        * WebKit.vcproj/WebKit.sln:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27061 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Fix for bug 15672, background images don't tile properly inside transforms.  This patch fixes background
        tiling to work in the presence of transforms and fixes bugs in both SVG and CSS transforms.

        Reviewed by aroben and mitz

        * WebCore.base.exp:
        * platform/graphics/Image.cpp:
        (WebCore::Image::setData):
        * platform/graphics/cg/ImageCG.cpp:
        (WebCore::ImageInfo::ImageInfo):
        (WebCore::Image::drawPatternCallback):
        (WebCore::Image::drawPattern):
        * platform/graphics/mac/GraphicsContextMac.mm:
        (WebCore::GraphicsContext::drawLineForMisspellingOrBadGrammar):
        * platform/mac/WebCoreSystemInterface.h:
        * platform/mac/WebCoreSystemInterface.mm:

WebKit:

        Fix for bug 15672, backgrounds don't tile properly inside transforms.  This patch fixes tiling
        of backgrounds inside CSS transforms and also of HTML content with background images inside SVG
        transforms.

        Reviewed by aroben and mmitz

        * WebCoreSupport/WebSystemInterface.m:
        (InitWebCoreSystemInterface):
        * WebKit.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27060 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Tim Hatcher

        Removed the support for toggling whether WebKit uses the 10.5 PDFKit improvements. Now it
        always does, when available.

        * WebView/WebPreferencesPrivate.h:
        removed _usePDFPreviewView and _setUsePDFPreviewView:. Note that these were guarded with
        a comment that says that they can be removed when no longer needed. That time is now.

        * WebView/WebPreferences.m:
        (+[WebPreferences initialize]):
        removed WebKitUsePDFPreviewViewPreferenceKey
        (-[WebPreferences _usePDFPreviewView]):
        removed
        (-[WebPreferences _setUsePDFPreviewView:]):
        removed

        * WebView/WebPDFView.mm:
        (-[WebPDFView initWithFrame:]):
        don't check _usePDFPreviewView

        * WebView/WebPreferenceKeysPrivate.h:
        removed WebKitUsePDFPreviewViewPreferenceKey

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27059 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Rubber stamped by Geoff.

        - Stubbed out the WebScriptDebugServer methods to give Drosera something
        to connect to and now the signature of the interface matches the mac.

        * Interfaces/IWebScriptDebugServer.idl:
        * WebScriptDebugServer.cpp:
        (WebScriptDebugServer::addListener):
        (WebScriptDebugServer::removeListener):
        (WebScriptDebugServer::step):
        (WebScriptDebugServer::pause):
        (WebScriptDebugServer::resume):
        (WebScriptDebugServer::isPaused):
        * WebScriptDebugServer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27058 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        Update the Leopard WebKitSystemInterface to be 4-way univeral to include 64-bit.

        * libWebKitSystemInterfaceLeopard.a:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27057 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Blind build fix attempt

        * WebCore.pro:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27056 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Added wrapper for getting the foundation cache directory.

        Reviewed by Adam Roben.

        * win/include/WebKitSystemInterface/WebKitSystemInterface.h:
        * win/lib/WebKitSystemInterface.lib:
        * win/lib/WebKitSystemInterface_debug.lib:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27055 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Anders

        Bulk rename of platform/sql/SQL* to platform/sql/SQLite*

        This is more accurate in that there is no realistic "SQL" abstraction, the classes are obviously
        tied extremely close to SQLite, and is necessitated by the introduction of "SQLTransaction" in
        the HTML5 database API which we are adopting.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27054 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin.

        - Renamed WebDebugProgram to WebScriptDebugServer to match the naming
        scheme on the mac.

        * Interfaces/IWebDebugProgram.idl: Removed.
        * Interfaces/IWebScriptDebugServer.idl: Copied from win/Interfaces/IWebDebugProgram.idl.
        * Interfaces/WebKit.idl:
        * WebDebugProgram.cpp: Removed.
        * WebDebugProgram.h: Removed.
        * WebKit.vcproj/Interfaces.vcproj:
        * WebKit.vcproj/WebKit.vcproj:
        * WebKit.vcproj/WebKitGUID.vcproj:
        * WebKitClassFactory.cpp:
        (WebKitClassFactory::CreateInstance):
        * WebKitDLL.cpp:
        (RunAsLocalServer):
        * WebScriptDebugServer.cpp: Copied from win/WebDebugProgram.cpp.
        (WebScriptDebugServer::WebScriptDebugServer):
        (WebScriptDebugServer::~WebScriptDebugServer):
        (WebScriptDebugServer::createInstance):
        (WebScriptDebugServer::QueryInterface):
        (WebScriptDebugServer::AddRef):
        (WebScriptDebugServer::Release):
        (WebScriptDebugServer::viewAdded):
        (WebScriptDebugServer::viewRemoved):
        (WebScriptDebugServer::attach):
        (WebScriptDebugServer::detach):
        (WebScriptDebugServer::statistics):
        (WebScriptDebugServer::webViews):
        * WebScriptDebugServer.h: Copied from win/WebDebugProgram.h.
        * WebView.cpp:
        (WebView::WebView):
        (WebView::~WebView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27053 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=15650
        XML attribute nodes are not importable

        Tests: fast/dom/import-attribute-node.html
               fast/dom/import-document-fragment.html

        * dom/Document.cpp:
        (WebCore::Document::importNode): Implemented for Attribute and DocumentFragment nodes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27052 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Dave Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15362
          <rdar://problem/5558715> Safari Crashes when opening a JS TreeGrid widget

        Test: fast/dynamic/insert-before-table-part-in-continuation.html

        * rendering/RenderFlow.cpp:
        (WebCore::RenderFlow::addChildWithContinuation): If beforeChild's
        parent is an anonymous table part, let the table figure out where to
        insert the new child.

LayoutTests:

        Reviewed by Dave Hyatt.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15362
          <rdar://problem/5558715> Safari Crashes when opening a JS TreeGrid widget

        * fast/dynamic/insert-before-table-part-in-continuation.html: Added.
        * platform/mac/fast/dynamic/insert-before-table-part-in-continuation-expected.checksum: Added.
        * platform/mac/fast/dynamic/insert-before-table-part-in-continuation-expected.png: Added.
        * platform/mac/fast/dynamic/insert-before-table-part-in-continuation-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27051 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Alp Toker  <alp@atoker.com>

        Reviewed by Brady.

        http://bugs.webkit.org/show_bug.cgi?id=15686
        GtkLauncher aborts on launch due to uninitialized threading subsystem

        Re-enable database support in the GTK+ port, with a fix.

        Initialize GLib threading as early as possible.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27050 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Steve.

        <rdar://5548217>: [NTS] Java 6 update 3 crashes Safari when loading a
        java page

        Previous patch erroneously compared a path and a
        path-including-filename. This corrects that. It also caches the result
        of safariPluginsPath() and uses shell API funcs for determining filename
        and parent directory from a full path.

        * ChangeLog:
        * plugins/win/PluginDatabaseWin.cpp:
        (WebCore::safariPluginsPath): Cache return value. Use API functions to
        build the path
        (WebCore::PluginDatabaseWin::pluginForMIMEType): Compare again plugin's
        parent directory, not full path
        (WebCore::PluginDatabaseWin::pluginForExtension): Same
        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackageWin::PluginPackageWin): Store parent directory
        * plugins/win/PluginPackageWin.h:
        (WebCore::PluginPackageWin::parentDirectory):
        * plugins/win/PluginStreamWin.cpp:
        (WebCore::PluginStreamWin::didReceiveData):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27048 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Alp Toker  <alp@atoker.com>

        Unreviewed fix to make the GTK+ port run.

        http://bugs.webkit.org/show_bug.cgi?id=15686
        GtkLauncher aborts on launch due to uninitialized threading subsystem

        http://bugs.webkit.org/show_bug.cgi?id=15688
        [GTK] Make it possible to disable database support

        Disable database support until #15686 is fixed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27047 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fixed a crash in the Qt 4.4 based text iterators when they're called with a null string.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27046 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* We need to set a != 0 status code for the fast/loader/xmlhttprequest-missing-file-exception.html
* libxml2 has the semantic that when writing an empty string and finishing it will report an error. For QXmlStreamReader this is valid.
* This is causing some regressions...

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27045 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Make fast/loader/xmlhttprequest-bad-mimetype.html pass. We use QHttp to download local files but we may not set the HTTP result code on the ResourceResponse.
* We can use the cross-platform result now. QWebNetworkInterface/Manager behaves the same as mac for local files.

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27044 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* fast/dom/onerror-img.html regressed due checking the JobStates because in case of error (e.g. not being able to connect) the job will no be started.
* Use the error message from Qt. It might or might not be translated.

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27043 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Use the JobStatus to make sure to not deliver finished/data before the job has started. This is the case with the fast/dom/onerror-img.html test case.
* We have no idea if any data will come so we can still finish and then get pending data. This luckily can't happen for the local file case.

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27042 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* No need to initialize values in the QWebNetworkJob c'tor
* Add a JobStatus to QWebNetworkJob and verify that the jobs are handled in the way we expect them to be handled. This means no data after the job has finished, not finishing a job before it has been started.

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27041 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Consistency: Always name the jobs job.

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27040 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Implement our own queuing of network jobs to allow special handling of synchronous jobs. This makes us pass the fast/dom/xmlhttprequest-html-response-encoding.html test without a crash. Sync jobs will get a special treatment over the normals ones and in theory more than one sync job is supported.
* This should be thread-safe besides QWebNetworkJob::{ref,deref}

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27039 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Do the percent replacement only when we are not base64. With base64 we should not have any % in it anyway.
* Have a custom decodePercentEncoding method that works without doing any charset conversion. With converting back to latin1() we lost some information.
* We pass the char-decoding.html test now

Signed-off-by: Simon Hausmann <hausmann@kde.org>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27038 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Alp Toker  <alp@atoker.com>

        Reviewed by Mark Rowe.

        Add support for list box theme colors and styled menu list buttons.

        Remove obsolete FIXMEs.

        * platform/gtk/RenderThemeGtk.cpp:
        (WebCore::RenderThemeGtk::RenderThemeGtk):
        (WebCore::RenderThemeGtk::paintCheckbox):
        (WebCore::RenderThemeGtk::paintRadio):
        (WebCore::RenderThemeGtk::paintButton):
        (WebCore::RenderThemeGtk::adjustMenuListStyle):
        (WebCore::RenderThemeGtk::paintMenuList):
        (WebCore::RenderThemeGtk::activeListBoxSelectionBackgroundColor):
        (WebCore::RenderThemeGtk::inactiveListBoxSelectionBackgroundColor):
        (WebCore::RenderThemeGtk::activeListBoxSelectionForegroundColor):
        (WebCore::RenderThemeGtk::inactiveListBoxSelectionForegroundColor):
        (WebCore::RenderThemeGtk::gtkTreeView):
        * platform/gtk/RenderThemeGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27037 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Corrected ChangeLog entry.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27036 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Rubber-stamped by Mark Rowe.

        * images/download.png: Updated to say "Nightly builds" instead of
       "Nightly (6 MB)" - the Mac version is ~17 MB now, and this isn't a direct
       download link anyway.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27035 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Roll out r27033 as it broke the JavaScriptCore tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27034 268f45cc-cd09-0410-ab3c-d52691b4dbfc

HackTop [30311:JavaScriptCore]% less ChangeLog                                        [/Stuff/Projects/WebKit/JavaScriptCore]
2007-10-25  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Roben.

        Start work on long-running-mode from the perl-script side of things.
        jsDriver.pl will likely be removed (and replaced by something better)
        But for now, I'm just hacking it down to something smaller.

        * tests/mozilla/jsDriver.pl: Remove lots of unused code.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27033 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Eric Seidel.

        Slightly elaborated the differences between declaration procesing in
        Function Code and Program Code.

        .3% speedup on SunSpider.

        * kjs/nodes.cpp:
        (KJS::FunctionBodyNode::processDeclarationsFunctionCode):
        (KJS::FunctionBodyNode::processDeclarationsProgramCode): Store a
        minimum set of attributes instead of recomputing all the time. Also,
        ignore m_parameters, since programs don't have arguments.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27032 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        More preparation work before adding long-running mode to testkjs.

        * kjs/testkjs.cpp:
        (TestFunctionImp::callAsFunction):
        (prettyPrintScript):
        (runWithScripts):
        (parseArguments):
        (kjsmain):
        (fillBufferWithContentsOfFile):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27031 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Forgot to check in this file in r27019

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27030 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Bring testkjs code out of the dark ages in preparation for more
        radical improvements (like long-running testing support!)

        * kjs/testkjs.cpp:
        (TestFunctionImp::callAsFunction):
        (setupInterpreter):
        (doIt):
        (fillBufferWithContentsOfFile):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27029 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Maciej Stachowiak.

        Make a fast path for declaration processing inside Function Code.

        Lifted declaration processing code up from individual declaration nodes
        and into processDeclarations.

        Broke out processDeclarations into two cases, depending on the type of
        code. This eliminates 2 branches, and facilitates more radical
        divergeance in the future.

        2.5% SunSpider speedup.

        * JavaScriptCore.xcodeproj/project.pbxproj:
        * kjs/nodes.cpp:
        (KJS::FunctionBodyNode::initializeDeclarationStacks):
        (KJS::FunctionBodyNode::processDeclarationsFunctionCode):
        (KJS::FunctionBodyNode::processDeclarationsProgramCode):
        (KJS::FunctionBodyNode::execute):
        (KJS::FuncDeclNode::makeFunction):
        * kjs/nodes.h:
        (KJS::):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27028 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        - add header includes needed on platforms that don't use AllInOneFile.cpp

        * API/JSCallbackObject.cpp:
        * kjs/Context.cpp:
        * kjs/ExecState.cpp:
        * kjs/array_instance.cpp:
        * kjs/function_object.cpp:
        * kjs/interpreter.cpp:
        * kjs/nodes.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27027 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Eric Seidel  <eric@webkit.org>

        Reviewed by Geoff.

        * JavaScriptCore.xcodeproj/project.pbxproj: re-mark JSGlobalObject.h as private

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27026 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Maciej Stachowiak.

        Fixed http://bugs.webkit.org/show_bug.cgi?id=15683
        Re-order declaration initialization to avoid calling hasProperty inside
        VarDeclNode::processDeclaration

        .7% speedup on SunSpider.

        * kjs/function.h:
        * kjs/function.cpp: Merged parameter processing into FunctionBodyNode's
        other processing of declared symbols, so the order of execution could
        change.

        * kjs/nodes.cpp:
        (KJS::VarDeclNode::getDeclarations): Added special case for the
        "arguments" property name, explained in the comment.

        (KJS::VarDeclNode::processDeclaration): Removed call to hasProperty
        in the case of function code, since we know the declared symbol
        management will resolve conflicts between symbols. Yay!

        (KJS::VarDeclListNode::getDeclarations): Now that VarDeclNode's
        implementation of getDeclarations is non-trivial, we can't take a
        short-cut here any longer -- we need to put the VarDecl node on the
        stack so it gets processed normally.

        (KJS::FunctionBodyNode::processDeclarations): Changed the order of
        processing to enforce mutual exclusion rules.

        * kjs/nodes.h:
        (KJS::DeclarationStacks::DeclarationStacks): Structure includes an
        ExecState now, for fast access to the "arguments" property name.

LayoutTests:

        Layout tests for bugs that might result from changes like
        http://bugs.webkit.org/show_bug.cgi?id=15683

        * fast/js/vardecl-preserve-parameters-expected.txt: Added.
        * fast/js/vardecl-preserve-parameters.html: Added.
        * fast/js/vardecl-preserve-vardecl-expected.txt: Added.
        * fast/js/vardecl-preserve-vardecl.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27025 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Eric Seidel  <eric@webkit.org>

        Reviewed by Geoff.

        * bindings/js/kjs_window.h: fix bogus comment, s/DOMNode/DOMObject/

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27024 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-25  Eric Seidel  <eric@webkit.org>

        Reviewed by Geoff.

        Make JSG build again after JSGlobalObject change.

        * JSRun.cpp:
        (JSRun::JSRun):
        (JSRun::GlobalObject):
        * JSRun.h:
        (JSInterpreter::JSInterpreter):
        * JSUtils.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27023 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Add a JSGlobalObject class and remove the InterpreterMap
        http://bugs.webkit.org/show_bug.cgi?id=15681

        This required making JSCallbackObject a template class to allow for
        JSGlobalObjects with JSCallbackObject functionality.

        SunSpider claims this was a 0.5% speedup.

        * API/JSCallbackObject.cpp:
        (KJS::):
        * API/JSCallbackObject.h:
        * API/JSCallbackObjectFunctions.h: Copied from API/JSCallbackObject.cpp.
        (KJS::::JSCallbackObject):
        (KJS::::init):
        (KJS::::~JSCallbackObject):
        (KJS::::initializeIfNeeded):
        (KJS::::className):
        (KJS::::getOwnPropertySlot):
        (KJS::::put):
        (KJS::::deleteProperty):
        (KJS::::implementsConstruct):
        (KJS::::construct):
        (KJS::::implementsHasInstance):
        (KJS::::hasInstance):
        (KJS::::implementsCall):
        (KJS::::callAsFunction):
        (KJS::::getPropertyNames):
        (KJS::::toNumber):
        (KJS::::toString):
        (KJS::::setPrivate):
        (KJS::::getPrivate):
        (KJS::::inherits):
        (KJS::::cachedValueGetter):
        (KJS::::staticValueGetter):
        (KJS::::staticFunctionGetter):
        (KJS::::callbackGetter):
        * API/JSClassRef.cpp:
        (OpaqueJSClass::prototype):
        * API/JSContextRef.cpp:
        (JSGlobalContextCreate):
        * API/JSObjectRef.cpp:
        (JSObjectMake):
        (JSObjectGetPrivate):
        (JSObjectSetPrivate):
        * API/JSValueRef.cpp:
        (JSValueIsObjectOfClass):
        * JavaScriptCore.exp:
        * JavaScriptCore.xcodeproj/project.pbxproj:
        * bindings/c/c_utility.cpp:
        (KJS::Bindings::convertValueToNPVariant):
        * bindings/jni/jni_jsobject.cpp:
        * bindings/objc/objc_utility.mm:
        (KJS::Bindings::convertValueToObjcValue):
        * kjs/Context.cpp:
        (KJS::Context::Context):
        * kjs/ExecState.cpp:
        (KJS::ExecState::lexicalInterpreter):
        * kjs/JSGlobalObject.h: Added.
        (KJS::JSGlobalObject::JSGlobalObject):
        (KJS::JSGlobalObject::isGlobalObject):
        (KJS::JSGlobalObject::interpreter):
        (KJS::JSGlobalObject::setInterpreter):
        * kjs/array_instance.cpp:
        * kjs/context.h:
        * kjs/function.cpp:
        (KJS::FunctionImp::callAsFunction):
        (KJS::GlobalFuncImp::callAsFunction):
        * kjs/interpreter.cpp:
        (KJS::Interpreter::Interpreter):
        (KJS::Interpreter::init):
        (KJS::Interpreter::~Interpreter):
        (KJS::Interpreter::globalObject):
        (KJS::Interpreter::initGlobalObject):
        (KJS::Interpreter::evaluate):
        * kjs/interpreter.h:
        * kjs/lookup.h:
        (KJS::cacheGlobalObject):
        * kjs/object.h:
        (KJS::JSObject::isGlobalObject):
        * kjs/testkjs.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27022 268f45cc-cd09-0410-ab3c-d52691b4dbfc

unbreak Qt build

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27021 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Fix <rdar://5549919> Add font database initialization code to WebCore and call it from WebKit

 WebCore:

         Add font database initialization code to WebCore

         Reviewed by Ada.

         * WebCore.vcproj/WebCore.vcproj: Added FontDatabase.{cpp,h}.
         * platform/win/FontDatabase.cpp: Added.
         (WebCore::systemFontsDirectory):
         (WebCore::fontsPlistPath):
         (WebCore::systemHasFontsNewerThanFontsPlist):
         (WebCore::readFontPlist):
         (WebCore::populateFontDatabaseFromPlist):
         (WebCore::populateFontDatabaseFromFileSystem):
         (WebCore::writeFontDatabaseToPlist):
         (WebCore::populateFontDatabase): This is the only function callable
         from outside this file. It populates the font database once, either
         from the fonts plist, or from the filesystem (and then saves a new
         fonts plist).
         * platform/win/FontDatabase.h: Added.

 WebKitTools:

         Remove now-unnecessary call to InitializeCoreGraphics

         WebKit takes care of this now.

         Reviewed by Ada.

         * DumpRenderTree/win/DumpRenderTree.cpp:
         (initialize):

 WebKit/win:

         Fix <rdar://5549919> Initialize the font database before any font code is invoked

         We initialize the database in two places:
            1. When instantiating WebKitClassFactory, which is guaranteed to
            happen before any WebView is instantiated.
            2. When making a WebCore::Font in WebKitGraphics.

         This ensures that the font database will be populated before any font
         code is invoked. We rely on WebCore to only populate the database
         once.

         Reviewed by Ada.

         * WebKitClassFactory.cpp:
         (WebKitClassFactory::WebKitClassFactory): Populate the font database.
         * WebKitGraphics.cpp:
         (makeFont): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27020 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Add [I]WebTextRenderer

 WebKitTools:

         Use WebTextRenderer in DRT

         Reviewed by Ada.

         * DumpRenderTree/win/DumpRenderTree.cpp:
         (initialize): Use WebTextRenderer instead of using
         InitializeCoreGraphics/AddFontResourceEx.

 WebKit/win:

         Add [I]WebTextRenderer

         Right now this class can only be used to add private fonts for use by
         the running process, but will eventually be the home of the
         WebKitGraphics functions.

         Reviewed by Ada.

         * ForEachCoClass.h: Added WebTextRenderer.
         * Interfaces/WebKit.idl: Ditto.
         * WebKit.vcproj/Interfaces.vcproj: Ditto.
         * WebKit.vcproj/WebKit.vcproj: Ditto.
         * WebKit.vcproj/WebKitGUID.vcproj: Ditto.
         * WebKitClassFactory.cpp: Ditto.
         * WebTextRenderer.cpp: Added.
         (WebTextRenderer::createInstance):
         (WebTextRenderer::WebTextRenderer):
         (WebTextRenderer::~WebTextRenderer):
         (WebTextRenderer::QueryInterface):
         (WebTextRenderer::AddRef):
         (WebTextRenderer::Release):
         (WebTextRenderer::registerPrivateFont):
         * WebTextRenderer.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27019 268f45cc-cd09-0410-ab3c-d52691b4dbfc

     Use FOR_EACH_COCLASS in WebKitClassFactory

         Reviewed by Ada.

         * WebError.cpp:
         (WebError::createInstance): Added an overload that takes no arguments
         to make the macro used in WebKitClassFactory work.
         * WebError.h:
         * WebKitClassFactory.cpp:
         (WebKitClassFactory::CreateInstance): Use FOR_EACH_COCLASS.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27018 268f45cc-cd09-0410-ab3c-d52691b4dbfc

     Put FOR_EACH_COCLASS macro into its own file and export it

         The macro used to be called FOR_EACH_CLASS and lived in WebKitDLL.cpp.
         This way we will be able to use the macro in more places that care
         about all WebKit's COM classes.

         Reviewed by Ada.

         * ForEachCoClass.h: Added.
         * WebKit.vcproj/WebKit.vcproj: Copy ForEachCoClass.h to
         WebKitOutputDir, and added it to the project.
         * WebKitDLL.cpp: Updated for macro rename, and changed to #undef the
         macros we pass to FOR_EACH_COCLASS after we're done with them.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27017 268f45cc-cd09-0410-ab3c-d52691b4dbfc

     Added some font-related functions needed for <rdar://5549919>

         Reviewed by Ada.

         * win/include/WebKitSystemInterface/WebKitSystemInterface.h:
         * win/lib/WebKitSystemInterface.lib:
         * win/lib/WebKitSystemInterface_debug.lib:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27016 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Sven Herzberg  <sven@imendio.com>

        Reviewed by Mark Rowe.

        Fixes http://bugs.webkit.org/show_bug.cgi?id=15614
        Bug 15614: [GTK] qmake based backends don't build on OS X

        * Scripts/build-webkit: set QMAKESPEC correctly if building the QT or
        GTK backend on a Mac
        * Scripts/webkitdirs.pm: extracted the darwin-test from isOSX() into
        isDarwin() to make it reusable in other places (like the workaround-
        for-prebuilt-qmake in build-webkit)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27015 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Refurbish update-webkit script.

        Reviewed by Adam.

        * Scripts/update-webkit: Add -h|--help switch and usage statement.  Check result of
        GetOptions() call.  Fix -q|--quiet switch to be passed to svn command properly.  Use
        multi-argument version of system() for flexibility and security.  Check for existence
        of Internal directory using -d test instead of -x.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27014 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Minor clean-up of prepare-ChangeLog script.

        Reviewed by Adam.

        * Scripts/prepare-ChangeLog: Keep list of updated ChangeLog files in an array instead
        of a string.  Use multi-argument versions of open() and system() for flexibility and
        security.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27013 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    <rdar://problem/5552221> REGRESSION(310A24-ToT): Shortcut key disable. (15604)

        Reviewed by Adam.

        * WebView.cpp:
        (WebView::handleEditingKeyboardEvent): don't handle system key events as text input
        (WebView::keyDown): only remove WM_SYSCHAR message from the queue if we handle it.
        For WM_SYSCHAR message that we don't handle, let it stay in the queue and return
        false to let windows handle it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27012 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Touch all files that include *.lut.h to fix Windows builds

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27011 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24 Mark Rowe <mrowe@apple.com>

        Gtk build fix.  Track WebCore changes in r27004.

        * Api/webkitgtkglobal.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27010 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        Correcting the fix for:
        <rdar://problem/5544856>
        REGRESSION: After typing 2-byte text, undo only undoes one keystroke at a time

        Made removal of the previous composition part of the current Undo step in the
        case where the new composition is the empty string, too.

        * editing/Editor.cpp:
        (WebCore::Editor::confirmComposition): Call the new TypingCommand::deleteSelection,
        which either has the currently open typing command delete the current selection, or
        opens a new typing command (of type DeleteSelection) if one is not already open.
        (WebCore::Editor::setComposition): Ditto.
        * editing/TypingCommand.cpp:
        (WebCore::TypingCommand::deleteSelection): Added.
        (WebCore::TypingCommand::doApply): Handle DeleteSelection.
        (WebCore::TypingCommand::deleteKeyPressed): Clarified which deleteSelection
        is called.
        (WebCore::TypingCommand::forwardDeleteKeyPressed): Ditto.
        (WebCore::TypingCommand::preservesTypingStyle): Handle DeleteSelection.
        * editing/TypingCommand.h:

LayoutTests:

        Reviewed by Darin.

        <rdar://problem/5544856>
        REGRESSION: After typing 2-byte text, undo only undoes one keystroke at a time

        * platform/mac/editing/input/text-input-controller-expected.txt: The delete
        we were using previously to remove the old composition makes fewer calls to
        the editing delegate.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27009 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Build fix.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27008 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27007 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit:

        Reviewed by Mark Rowe.

        <rdar://problem/5069711> OpenSource version of libWebKitSystemInterface.a is Tiger only, causes issues if used on Leopard

        Use the WebKitSystemInterface that matches the system version.

        * Configurations/DebugRelease.xcconfig:
        * WebKit.xcodeproj/project.pbxproj:

WebKitLibraries:

        Reviewed by Mark Rowe.

        <rdar://problem/5069711> OpenSource version of libWebKitSystemInterface.a is Tiger only, causes issues if used on Leopard

        Add system specific versions of WebKitSystemInterface.

        * libWebKitSystemInterface.a: Removed.
        * libWebKitSystemInterfaceLeopard.a: Added.
        * libWebKitSystemInterfaceTiger.a: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27006 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Oliver.

        Fix <rdar://5410959>  editing/selection/drag-to-contenteditable-iframe.html fails on Windows

        * page/win/EventHandlerWin.cpp:
        (WebCore::EventHandler::passMouseMoveEventToSubframe):
        Some mouse move events are actually drags, which on mac return early from this function.
        Adding the same logic to its Windows equivalent.

WebKitTools:

        Reviewed by Oliver.

        Fix <rdar://5410959>  editing/selection/drag-to-contenteditable-iframe.html fails on Windows

        Move the call to replaySavedEvents from doDragDrop to doMouseMove because we don't want to
        replay the saved events when we're still processing the mousedown that starts the drag
        * DumpRenderTree/win/EventSender.cpp:
        (doMouseMove):
        * DumpRenderTree/win/UIDelegate.cpp:
        (UIDelegate::doDragDrop):

LayoutTests:

        Removing fixed test

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27005 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:
        Reviewed by Anders

        <rdar://5554130> DatabaseTracker.o has a global initializer

        Since DatabaseTracker is a singleton, it makes much more sense to keep the database path as a member variable.
        Now constructing the shared DatabaseTracker no longer implicitly opens it - It is only opened when the databases
        path is set.

        * WebCore.vcproj/WebCore.vcproj: Copy WebCore/storage headers for WebKit build

        * storage/Database.cpp:
        (WebCore::Database::~Database): Remove bogus assertion

        * storage/DatabaseTracker.cpp:
        (WebCore::DatabaseTracker::DatabaseTracker): Does nothing now!  Move this code to openTrackerDatabase
        (WebCore::DatabaseTracker::setDatabasePath): Set the member variable, also closing/opening the database if needed
        (WebCore::DatabaseTracker::databasePath):
        (WebCore::DatabaseTracker::openTrackerDatabase): To open and validate the Databases db, moved from the c'tor
        (WebCore::DatabaseTracker::fullPathForDatabase): Return the member variable
        * storage/DatabaseTracker.h:

WebKit/win:
        Reviewed by Anders

        Windows portion of <rdar://5554130>

        Slowly introduce Windows WebKit portion of the Database API that sets the
        on-disk location for databases

        * WebDatabaseManager.cpp: Added.
        (WebKitSetWebDatabasesPathIfNecessary):
        * WebDatabaseManager.h: Added.

        * WebKit.vcproj/WebKit.vcproj:

        * WebView.cpp:
        (WebView::initWithFrame): Call WebKitSetWebDatabasesPathIfNecessary()

WebKit:
        Reviewed by Anders

        <rdar://problem/5554130> DatabaseTracker.o has a global initializer

        * Misc/WebDatabaseManager.mm:
        (WebKitSetWebDatabasesPathIfNecessary): Call the member function instead of a static one

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27004 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKitTools:

        Reviewed by Sam, Steve and Darin.

        - Stubbed out the WebScriptDebugListener functionality in the Server
        Connection class to prepare it for receiving those callbacks.
        - Also I changed the instantiation of DebuggerClient, DebuggerDocument
        and the ServerConnection to not need to be initialized with a server
        name, since that is not the way we connect to the WebKit server.

        * Drosera/DebuggerDocument.cpp: Fixed a bug where I was always logging
        no exception.  Now it only loggs when there is an exception
        (DebuggerDocument::willExecuteStatement):
        (DebuggerDocument::didEnterCallFrame):
        (DebuggerDocument::willLeaveCallFrame):
        (DebuggerDocument::windowScriptObjectAvailable):
        (DebuggerDocument::callFunctionOnObject):
        * Drosera/win/DebuggerClient.cpp: No longer needs the ServerConnection
        to be instantiated with a server's name.
        (DebuggerClient::DebuggerClient):
        (DebuggerClient::didFinishLoadForFrame):
        * Drosera/win/DebuggerClient.h: Removed unsued variable.
        * Drosera/win/Drosera.cpp: Client no longer needs to be initialized with
        a server name.
        (Drosera::init):
        (Drosera::initServer):
        * Drosera/win/Drosera.h: No longer need the ServerConnection to be
        instantiated with a server's name.
        * Drosera/win/ServerConnection.cpp:
        (ServerConnection::ServerConnection): No longer needs a server name
        (ServerConnection::~ServerConnection): Only release the global context
        if there is one.
        (ServerConnection::serverConnectionDidDie): Stub for
        IWebScriptDebugListener
        (ServerConnection::QueryInterface): ditto
        (ServerConnection::AddRef): ditto
        (ServerConnection::Release): ditto
        (ServerConnection::didLoadMainResourceForDataSource): ditto
        (ServerConnection::didParseSource): ditto
        (ServerConnection::failedToParseSource): ditto
        (ServerConnection::didEnterCallFrame): ditto
        (ServerConnection::willExecuteStatement): ditto
        (ServerConnection::willLeaveCallFrame): ditto
        (ServerConnection::exceptionWasRaised): ditto
        * Drosera/win/ServerConnection.h: Stubbed out the
        IWebScriptDebugListener functions, and this class no longer needs to be
        instantiated with a server name.

win:

        Reviewed by Sam, Steve and Darin.

        - Added stubs for what will be neede to let Drosera attach to the
        WebKit process and debug it.

        * Interfaces/IWebScriptCallFrame.idl: Added.
        * Interfaces/IWebScriptDebugListener.idl: Added.
        * Interfaces/WebKit.idl:
        * WebKit.vcproj/Interfaces.vcproj:
        * WebKit.vcproj/WebKitGUID.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27003 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Eric Seidel  <eric@webkit.org>

        Build fix for Gtk, no review.

        * kjs/collector.cpp: #include "context.h"

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27002 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Stop checking isOutOfMemory after every allocation, instead let the collector
        notify all ExecStates if we ever hit this rare condition.

        SunSpider claims this was a 2.2% speedup.

        * kjs/collector.cpp:
        (KJS::Collector::collect):
        (KJS::Collector::reportOutOfMemoryToAllInterpreters):
        * kjs/collector.h:
        * kjs/nodes.cpp:
        (KJS::TryNode::execute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27001 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Don't try to gap fill transformed selections.  Fix the invalidation of selection to use the clip to visible content
        code path so that it works with multi-column layouts and transforms.

        Reviewed by aroben

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::isSelectionRoot):
        (WebCore::RenderBlock::fillSelectionGaps):
        * rendering/RenderView.cpp:
        (WebCore::RenderView::setSelection):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@27000 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Eric Seidel  <eric@webkit.org>

        No review, I want to tickle the CIA bot to test changes,
        and took this as an opportunity to clean up some old change logs.

        * ChangeLog-2005-08-23: update email address
        * ChangeLog-2006-12-31: fix spelling mistakes

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26999 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Oliver Hunt.

        - add "(anonymous)" to the renderName of anonymous table cells, rows and sections

        * rendering/RenderTableCell.h:
        (WebCore::RenderTableCell::renderName):
        * rendering/RenderTableRow.h:
        (WebCore::RenderTableRow::renderName):
        * rendering/RenderTableSection.h:
        (WebCore::RenderTableSection::renderName):

LayoutTests:

        Reviewed by Oliver Hunt.

        - updated results to reflect which table parts are anonymous

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26998 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        Bug 15476: DOM tree fully expands some elements when arrowing up
        http://bugs.webkit.org/show_bug.cgi?id=15476

        Add a new expandTreeElementsWhenArrowing option to TreeOutline. This option
        is used only for the sidebar and search results.

        * page/inspector/inspector.js: Opt-in for the sidebar and search results.
        * page/inspector/treeoutline.js: Add expandTreeElementsWhenArrowing and pass
          this option to traverseNextTreeElement and traversePreviousTreeElement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26997 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Mark Rowe  <mrowe@apple.com>

        Gtk build fix.

        * kjs/identifier.h:  Remove extra qualification.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26996 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam Weinig.

        Disable ALWAYS_INLINE in debug builds, since it drives the debugger
        crazy.

        * wtf/AlwaysInline.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26995 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam Weinig.

        Inlined the fast path for creating an Identifier from an Identifier.

        This is a .4% speedup on SunSpider overall, but as big as a 2.5%
        speedup on certain individual tests. 65% of the Identifiers creating
        by SunSpider are already Identifiers.

        (The main reason I'm making this change is that it resolves a large
        regression in a patch I haven't checked in yet.)

        * JavaScriptCore.exp:
        * kjs/identifier.cpp:
        (KJS::Identifier::addSlowCase):
        * kjs/identifier.h:
        (KJS::Identifier::Identifier::add):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26994 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix ChangeLog.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26993 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff and Mitz.

        <rdar://problem/5493833>
        REGRESSION: ~5MB of image data leaked @ cuteoverload.com (often seen while browsing other sites, too)

        * WebCore.xcodeproj/project.pbxproj:
        * bindings/js/kjs_binding.cpp:
        (KJS::ScriptInterpreter::markDOMNodesForDocument):
        If an image element that is currently loading an image is not in the document,
        it should still be marked.

        * bindings/js/kjs_html.cpp:
        (WebCore::ImageConstructorImp::construct):
        Force the document wrapper to be created.

        * html/HTMLImageElement.h:
        (WebCore::HTMLImageElement::haveFiredLoadEvent):
        New method which calls down to the image loader.

        * html/HTMLImageLoader.cpp:
        (WebCore::HTMLImageLoader::HTMLImageLoader):
        (WebCore::HTMLImageLoader::~HTMLImageLoader):
        (WebCore::HTMLImageLoader::setLoadingImage):
        (WebCore::HTMLImageLoader::dispatchLoadEvent):
        Remove code that's not needed anymore.

        * html/HTMLImageLoader.h:
        (WebCore::HTMLImageLoader::haveFiredLoadEvent):
        Make this public.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26992 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Move Windows safe file creation code into WebCore from WebPreferences

 WebCore:

         Move Windows safe file creation code into WebCore from WebPreferences

         Reviewed by Brady.

         * platform/FileSystem.h:
         * platform/win/FileSystemWin.cpp:
         (WebCore::safeCreateFile): Moved here from WebKit/win/WebPreferences.cpp.

 WebKit/win:

         Move safe file creation code to WebCore

         Reviewed by Brady.

         * WebPreferences.cpp:
         (preferencesPath): Made into a static helper function.
         (WebPreferences::save): Now calls WebCore's safeCreateFile function.
         (WebPreferences::load): Uses String/CString to handle the UTF-8
         conversion.
         * WebPreferences.h: Removed preferencesPath and
         safeCreateFileWithData.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26991 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reduce code duplication by using WebCore's FileSystem functions

WebCore:

        Add methods to FileSystemWin to get some user profile directories

        These directories are used to hold things like preferences, caches,
        etc.

        Reviewed by Brady.

        * platform/FileSystem.h: Added new method declarations for Windows
        only.
        * platform/win/FileSystemWin.cpp:
        (WebCore::bundleName): Added.
        (WebCore::storageDirectory): Added.
        (WebCore::cachedStorageDirectory): Added.
        (WebCore::localUserSpecificStorageDirectory): Added. Returns the
        directory where WebKit should store any user-specific data that should
        stay local to the current machine (i.e., shouldn't be stored in a
        roaming profile).
        (WebCore::roamingUserSpecificStorageDirectory): Added. Returns the
        directory where WebKit should store any user-specific data that should
        move with the user from machine to machine (i.e., should be stored in
        a roaming profile).

WebKit/win:

        Reduce code duplication by using WebCore's FileSystem functions

        Reviewed by Brady.

        * WebIconDatabase.cpp: Removed a now-unused function and a fixed
        FIXME.
        (WebIconDatabase::init): Changed to use FileSystem functions.
        * WebPreferences.cpp:
        (WebPreferences::preferencesPath): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26990 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Mark Rowe  <mrowe@apple.com>

        Fix incorrect test results landed with r26986.

        * http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getAllRequestHeaders-expected.txt:
        * http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getRequestHeader-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26989 268f45cc-cd09-0410-ab3c-d52691b4dbfc

update some test results for the Qt port.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26988 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Alp Toker  <alp@atoker.com>

        Reviewed by Mark Rowe.

        http://bugs.webkit.org/show_bug.cgi?id=15659
        InlineTextBox does not setStrokeStyle() as needed

        GraphicsContextCairo workaround to support a GraphicsContextCG quirk:

        Save and restore the StrokeStyle manually.

        * platform/graphics/cairo/GraphicsContextCairo.cpp:
        (WebCore::GraphicsContext::drawLineForText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26987 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-10-24  Julien Chaffraix  <julien.chaffraix@gmail.com>

        Reviewed by Darin.

        Patch for http://bugs.webkit.org/show_bug.cgi?id=15356
        Bug 15356: getResponseHeader and getAllResponseHeaders do not throw exceptions

        - getAllResponseHeaders and getResponseHeader throws INVALID_STATE_ERR exception as specified in the specification

        - Add the check to field-name production in getResponseHeader (call to isValidToken) and returns an empty string if the header is not valid (Opera and IE behaviour)

        - Updated XMLHttpRequest::responseMIMEType() to keep its behaviour

        Tests: http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getAllRequestHeaders.html
               http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getRequestHeader.html
               http/tests/xmlhttprequest/xmlhttprequest-invalidHeader-getRequestHeader.html

        * bindings/js/JSXMLHttpRequest.cpp:
        (KJS::JSXMLHttpRequestPrototypeFunction::callAsFunction):
        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::getAllResponseHeaders):
        (WebCore::XMLHttpRequest::getResponseHeader):
        (WebCore::XMLHttpRequest::responseMIMEType):
        * xml/XMLHttpRequest.h:

2007-10-24  Julien Chaffraix  <julien.chaffraix@gmail.com>

        Reviewed by Darin.

        Tests for http://bugs.webkit.org/show_bug.cgi?id=15356
        getResponseHeader and getAllResponseHeaders do not throw exceptions

        * http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getAllRequestHeaders-expected.txt: Added.
        * http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getAllRequestHeaders.html: Added.
        * http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getRequestHeader-expected.txt: Added.
        * http/tests/xmlhttprequest/xmlhttprequest-InvalidStateException-getRequestHeader.html: Added.
        * http/tests/xmlhttprequest/xmlhttprequest-invalidHeader-getRequestHeader-expected.txt: Added.
        * http/tests/xmlhttprequest/xmlhttprequest-invalidHeader-getRequestHeader.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26986 268f45cc-cd09-0410-ab3c-d52691b4dbfc

these tests don't crash anymore :)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26985 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Stop crashing on fast/events/frame-tab-focus.html the keyEvent can be 0.

Signed-off-by: Lars Knoll <lars@trolltech.com>
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26984 268f45cc-cd09-0410-ab3c-d52691b4dbfc

implement most of DragData.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26983 268f45cc-cd09-0410-ab3c-d52691b4dbfc

add a layoutTestController.dumpSelectionRect() dummy. It only has an effect on mac pixel tests anyway. Makes us pass another 5 test cases.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26982 268f45cc-cd09-0410-ab3c-d52691b4dbfc

we're passing most editing tests now. Commit them.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26981 268f45cc-cd09-0410-ab3c-d52691b4dbfc

add another command.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26980 268f45cc-cd09-0410-ab3c-d52691b4dbfc

convert the text to be text only, and fix it up a little so it doesn't depend on specific font metrics anymore.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26979 268f45cc-cd09-0410-ab3c-d52691b4dbfc

remove some notImplemented() warnings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26978 268f45cc-cd09-0410-ab3c-d52691b4dbfc

add empty eventSender.clearKillRing(), and map some of the special keys used in eventSender.keyDown() to things that work on Qt/X11.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26977 268f45cc-cd09-0410-ab3c-d52691b4dbfc

remove a stupid notImplemented() warning and implement PasteBoad::clear().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26976 268f45cc-cd09-0410-ab3c-d52691b4dbfc

allow paste from DOM so we pass more test cases.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26975 268f45cc-cd09-0410-ab3c-d52691b4dbfc