Updated WebKit from /home/shausman/src/webkit/trunk to origin/qtwebkit-4.6-staging ( b2abc0c271880b8135507861056af497f895adf5 )

Changes in WebKit since the last update:

++ b/WebKit/qt/ChangeLog
2009-07-13  Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Ariya Hidayat.

        Fix qdoc warnings for QWebPage::shouldInterruptJavaScript() and mention
        how to re-implement it.

        * Api/qwebpage.cpp:

2009-07-13  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Ariya Hidayat.

        Fix crash with plugins when the plugin stream is cancelled.

        Similar to r26667 handle the case where didReceiveResponse on the
        plugin view results in failure to set up the stream and
        setMainDocumentError being called instead. This will set the
        m_pluginView back to 0 and we need check for it before calling
        didReceiveData.

        This was triggered by consecutive execution of
        LayoutTests/plugins/return-error-from-new-stream-callback-in-full-frame-plugin.html
        followed by LayoutTests/scrollbars/scrollbar-crash-on-refresh.html

        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::committedLoad):

2009-07-13  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Ariya Hidayat.

        Added QWebDatabase::removeAllDatabases, as a way to delete all
        databases from the offline storage path.

        Used by the Qt DRT.

        * Api/qwebdatabase.cpp:
        (QWebDatabase::removeAllDatabases):
        * Api/qwebdatabase.h:

2009-07-13  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Ariya Hidayat.

        Added loadStarted() and loadFinished() signals to QWebFrame,
        to allow load tracking of individual frames, as opposed to
        QWebPage's loadStarted/loadFinished signals that are emitted
        whenever _any_ child frame loads/finishes.

        * Api/qwebframe.cpp: Document new signals.
        * Api/qwebframe.h: Add new signals.
        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::setFrame): Connect new signals.

2009-07-13  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Ariya Hidayat.

        Add hooks for the GCController JavaScript interface needed by the
        Qt DRT.

        Fixed sort order of includes in qwebframe.cpp.

        * Api/qwebframe.cpp:
        (qt_drt_javaScriptObjectsCount):
        (qt_drt_garbageCollector_collect):
        (qt_drt_garbageCollector_collectOnAlternateThread):

2009-07-13  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Ariya Hidayat.

        Add hooks for the GCController JavaScript interface needed by the
        Qt DRT.

        Fixed sort order of includes in qwebframe.cpp.

        * Api/qwebframe.cpp:
        (qt_drt_javaScriptObjectsCount):
        (qt_drt_garbageCollector_collect):
        (qt_drt_garbageCollector_collectOnAlternateThread):

2009-07-12  Brent Fulgham  <bfulgham@gmail.com>

        Speculative build fix after http://trac.webkit.org/changeset/45786.

        * WebCoreSupport/ChromeClientQt.cpp:
        (WebCore::ChromeClientQt::addMessageToConsole):
        * WebCoreSupport/ChromeClientQt.h:

2009-07-10  Yael Aharon  <yael.aharon@nokia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=27136

        Fix a bug where webkit hangs when executing infinite JavaScript loop.

        * Api/qwebpage.cpp:
        (QWebPage::shouldInterruptJavaScript):
        * Api/qwebpage.h:
        * WebCoreSupport/ChromeClientQt.cpp:
        (WebCore::ChromeClientQt::shouldInterruptJavaScript):
        * tests/qwebpage/tst_qwebpage.cpp:
        (JSTestPage::JSTestPage):
        (JSTestPage::shouldInterruptJavaScript):
        (tst_QWebPage::infiniteLoopJS):

2009-07-10  Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=27108

        Fix crash when in frame tree of a new frame before the new frame
        has been installed in the frame tree, similar to r35088.

        After calling Frame::init() the frame it may have been removed from the
        frame tree again through JavaScript. Detect this by checking the page()
        afterwards.

        To make this check safe the Frame::init() code was moved into
        QWebFrameData's constructor, where a RefPtr holds a reference to the frame.
        After the check back in FrameLoaderClientQt we would hold the single
        reference left and after release() the frame, its frame loader, its
        client as well as the QWebFrame should have disappeared then.

        * Api/qwebframe.cpp:
        (QWebFramePrivate::init): Only call Frame::init here, the rest is
        done in QWebFrameData's constructor.
        (QWebFrame::QWebFrame):
        * Api/qwebframe_p.h: Adjust declaration.
        (QWebFrameData::QWebFrameData): Create the Frame here.
        * Api/qwebpage.cpp:
        (QWebPagePrivate::createMainFrame): Adjust and simplify
        to new QWebFrame constructor.
        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::createFrame): Adjust to
        new QWebFrame construction using QWebFrameData and add the
        check like in r35088.

2009-07-09  Beth Dakin  <bdakin@apple.com>

        Reviewed by Dave Hyatt.

        Make Widget RefCounted to fix:

        <rdar://problem/7038831> REGRESSION (TOT): In Mail, a crash occurs
        at WebCore::Widget::afterMouseDown() after clicking To Do's close
        box
        <rdar://problem/6978804> WER #16: Repro Access Violation in
        WebCore::PluginView::bindingInstance (1310178023)
        -and-
        <rdar://problem/6991251> WER #13: Crash in WebKit!
        WebCore::PluginView::performRequest+203 (1311461169)

        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::createPlugin):
        (WebCore::FrameLoaderClientQt::createJavaAppletWidget):
        * WebCoreSupport/FrameLoaderClientQt.h:

2009-07-08  Pradeepto Bhattacharya  <pradeepto@kde.org>

        Reviewed by Ariya Hidayat.

        Build fix.

        * WebCoreSupport/FrameLoaderClientQt.h: Removed the slot slotCallPolicyFunction().

2009-07-08  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Tor Arne Vestbø.

        https://bugs.webkit.org/show_bug.cgi?id=27080

        Fix DRT instability issues with fast/loader/submit-form-while-parsing-2.html

        When the form is submitted we call the policy function in the frame
        loader delayed with a queued connection. That queued connection
        sometimes interferes with the javascript timeout set in the testcase.

        Eliminate the entire delayed policy function mechanism and instead always
        call back directly, like in the other ports. In most other places we called
        the slot directly anyway.

        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::FrameLoaderClientQt): Remove m_policyFunction.
        (WebCore::FrameLoaderClientQt::callPolicyFunction): Call the policy function directly instead
        of emitting the queued signal.
        (WebCore::FrameLoaderClientQt::cancelPolicyCheck): Call callPolicyFunction directly.
        (WebCore::FrameLoaderClientQt::dispatchWillSubmitForm): Ditto.
        (WebCore::FrameLoaderClientQt::dispatchDecidePolicyForMIMEType): Ditto.
        (WebCore::FrameLoaderClientQt::dispatchDecidePolicyForNewWindowAction): Ditto.
        (WebCore::FrameLoaderClientQt::dispatchDecidePolicyForNavigationAction): Ditto.
        * WebCoreSupport/FrameLoaderClientQt.h: Remove m_policyFunction as well as the associated
        signal.

2009-07-07  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Holger Freyther.

        Add Qt DRT hook for clearing the frame name.

        * Api/qwebframe.cpp:
        (qt_drt_clearFrameName):

2009-07-05  Simon Hausmann  <hausmann@webkit.org>

        Reviewed by Holger Freyther.

        Fix two qdoc warnings.

        Added missing \property for QWebFrame::hasFocus and added \a
        tag for pos of QWebPage::frameAt.

        * Api/qwebframe.cpp:
        * Api/qwebpage.cpp:

2009-07-04  Holger Hans Peter Freyther  <zecke@selfish.org>

        Reviewed by Simon Hausmann.

        Use the recently introduced FocusController::setFocused

        Use the recently introduced FocusController::setFocused
        in the Qt platform. The SelectionController will be updated
        from within the FocusController now.

        * Api/qwebpage.cpp:
        (QWebPagePrivate::focusInEvent):
        (QWebPagePrivate::focusOutEvent):

2009-07-02  Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Ariya Hidayat.

        Improve documentation of QWebFrame::setFocus and hasFocus()
        Added missing Q_PROPERTY for QWebFrame::hasFocus.

        * Api/qwebframe.cpp: Clarify the docs.
        * Api/qwebframe.h: add Q_PROPERTY(focus).

2009-07-02  Joe Ligman  <joseph.ligman@nokia.com>

        Reviewed by Simon Hausmann.

        Bug 26855: [Qt] New methods for QWebFrame to check and set focus.

        Added new public methods QWebFrame::hasFocus() and QWebFrame::setFocus()
        Added auto test.

        * Api/qwebframe.cpp:
        (QWebFrame::hasFocus):
        (QWebFrame::setFocus):
        * Api/qwebframe.h:
        * tests/qwebframe/tst_qwebframe.cpp:

2009-07-01  Robert Hogan <robert@roberthogan.net>

        Reviewed by NOBODY.

        Fix Qt segfault when javascript disabled.
        If clients call addToJavaScriptWindowObject even though JavascriptEnabled is false
        webkit will segfault on the assert:
          ASSERTION FAILED: _rootObject
          (../../../WebCore/bridge/runtime.cpp:52
          JSC::Bindings::Instance::Instance(WTF::PassRefPtr<JSC::Bindings::RootObject>))
        Fix is to ensure JavaScript is enabled when client calls addToJavaScriptWindowObject.

        https://bugs.webkit.org/show_bug.cgi?id=26906

        * Api/qwebframe.cpp:
        (QWebFrame::addToJavaScriptWindowObject):

2009-07-01  Jakub Wieczorek  <faw217@gmail.com>

        Reviewed by Simon Hausmann.

        [Qt] Move some API headers from WebCore.pro to headers.pri so that they
        get installed when running make install from the build directory.

        * Api/headers.pri:

2009-07-01  Balazs Kelemen  <kelemen.balazs.3@stud.u-szeged.hu>

        Reviewed by Simon Hausmann.

        Fixed robotized QtLauncher to work when there is no index.html in the user's home.

        * QtLauncher/main.cpp:
        (main):

2009-06-30  Brian Weinstein  <bweinstein@apple.com>

        Reviewed by Adam Roben.

        Renamed scrollbarUnderPoint to scrollbarAtPoint to follow conventions.

        * Api/qwebpage.cpp:
        (QWebPage::swallowContextMenuEvent):

2009-06-30  Joe Ligman  <joseph.ligman@nokia.com>

        Reviewed by Adam Treat.

        Bug 26422: [Qt] QWebPagePrivate::frameAt calculates wrong frame

        Added a public method QWebPage::frameAt
        Removed QWebPagePrivate::frameAt, which calcuated the wrong frame
        Modified QWebPage::swallowContextMenuEvent to use the new frameAt method
        New test case for frameAt added to tst_qwebpage.cpp

        * Api/qwebpage.cpp:
        (QWebPage::frameAt):
        (QWebPage::swallowContextMenuEvent):
        * Api/qwebpage.h:
        * Api/qwebpage_p.h:
        * tests/qwebpage/frametest/iframe.html: Added.
        * tests/qwebpage/frametest/iframe2.html: Added.
        * tests/qwebpage/frametest/iframe3.html: Added.
        * tests/qwebpage/tst_qwebpage.cpp:
        (frameAtHelper):
        (tst_QWebPage::frameAt):
        * tests/qwebpage/tst_qwebpage.qrc:

2009-06-30  Jakub Wieczorek  <faw217@gmail.com>

        Reviewed by Simon Hausmann.

        Add QWebFrame::baseUrl() function that exposes the base URL of a frame.
        Autotests included.

        * Api/qwebframe.cpp:
        (QWebFrame::baseUrl):
        * Api/qwebframe.h:
        * tests/qwebframe/tst_qwebframe.cpp:

2 files changed, 3 insertions, 11 deletions

diff --git a/src/3rdparty/webkit/JavaScriptCore/jit/JIT.cpp b/src/3rdparty/webkit/JavaScriptCore/jit/JIT.cpp
index 02cb09b..f1b22c0 100644
--- a/src/3rdparty/webkit/JavaScriptCore/jit/JIT.cpp
+++ b/src/3rdparty/webkit/JavaScriptCore/jit/JIT.cpp
@@ -275,7 +275,6 @@ void JIT::privateCompileMainPass()
         DEFINE_OP(op_throw)
         DEFINE_OP(op_to_jsnumber)
         DEFINE_OP(op_to_primitive)
-        DEFINE_OP(op_unexpected_load)
 
         case op_get_array_length:
         case op_get_by_id_chain:
diff --git a/src/3rdparty/webkit/JavaScriptCore/jit/JITOpcodes.cpp b/src/3rdparty/webkit/JavaScriptCore/jit/JITOpcodes.cpp
index dbcb34d..b669dfa 100644
--- a/src/3rdparty/webkit/JavaScriptCore/jit/JITOpcodes.cpp
+++ b/src/3rdparty/webkit/JavaScriptCore/jit/JITOpcodes.cpp
@@ -505,13 +505,6 @@ void JIT::emit_op_jneq_ptr(Instruction* currentInstruction)
     RECORD_JUMP_TARGET(target + 3);
 }
 
-void JIT::emit_op_unexpected_load(Instruction* currentInstruction)
-{
-    JSValue v = m_codeBlock->unexpectedConstant(currentInstruction[2].u.operand);
-    move(ImmPtr(JSValue::encode(v)), regT0);
-    emitPutVirtualRegister(currentInstruction[1].u.operand);
-}
-
 void JIT::emit_op_jsr(Instruction* currentInstruction)
 {
     int retAddrDst = currentInstruction[1].u.operand;
@@ -759,7 +752,7 @@ void JIT::emit_op_new_error(Instruction* currentInstruction)
 {
     JITStubCall stubCall(this, JITStubs::cti_op_new_error);
     stubCall.addArgument(Imm32(currentInstruction[2].u.operand));
-    stubCall.addArgument(ImmPtr(JSValue::encode(m_codeBlock->unexpectedConstant(currentInstruction[3].u.operand))));
+    stubCall.addArgument(ImmPtr(JSValue::encode(m_codeBlock->getConstant(currentInstruction[3].u.operand))));
     stubCall.addArgument(Imm32(m_bytecodeIndex));
     stubCall.call(currentInstruction[1].u.operand);
 }
@@ -828,7 +821,7 @@ void JIT::emit_op_enter(Instruction*)
     // Even though CTI doesn't use them, we initialize our constant
     // registers to zap stale pointers, to avoid unnecessarily prolonging
     // object lifetime and increasing GC pressure.
-    size_t count = m_codeBlock->m_numVars + m_codeBlock->numberOfConstantRegisters();
+    size_t count = m_codeBlock->m_numVars;
     for (size_t j = 0; j < count; ++j)
         emitInitRegister(j);
 
@@ -839,7 +832,7 @@ void JIT::emit_op_enter_with_activation(Instruction* currentInstruction)
     // Even though CTI doesn't use them, we initialize our constant
     // registers to zap stale pointers, to avoid unnecessarily prolonging
     // object lifetime and increasing GC pressure.
-    size_t count = m_codeBlock->m_numVars + m_codeBlock->numberOfConstantRegisters();
+    size_t count = m_codeBlock->m_numVars;
     for (size_t j = 0; j < count; ++j)
         emitInitRegister(j);