WebKit-https.git
11 years ago2009-04-24 Diego Escalante Urrelo <diegoe@gnome.org>
kov@webkit.org [Fri, 24 Apr 2009 13:16:57 +0000 (13:16 +0000)]
2009-04-24  Diego Escalante Urrelo  <diegoe@gnome.org>

        Reviewed by Gustavo Noronha.

        https://bugs.webkit.org/show_bug.cgi?id=15616
        [GTK] Add spell checking

        Implement EditorClient::ignoreWordInSpellDocument,
        EditorClient::learnWord and EditorClient::getGuessesForWord. This
        enables the corresponding user actions.

        * WebCoreSupport/EditorClientGtk.cpp:
        (WebKit::EditorClient::ignoreWordInSpellDocument):
        (WebKit::EditorClient::learnWord):
        (WebKit::EditorClient::getGuessesForWord):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-24 Diego Escalante Urrelo <diegoe@gnome.org>
kov@webkit.org [Fri, 24 Apr 2009 13:16:40 +0000 (13:16 +0000)]
2009-04-24  Diego Escalante Urrelo  <diegoe@gnome.org>

        Reviewed by Gustavo Noronha.

        https://bugs.webkit.org/show_bug.cgi?id=15616
        [GTK] Add spell checking

        Implement EditorClient::checkSpellingOfString; enabling spell checking
        to actually happen, this is the basis for other spelling functions.

        * WebCoreSupport/EditorClientGtk.cpp:
        (WebKit::EditorClient::checkSpellingOfString):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42819 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-24 Diego Escalante Urrelo <diegoe@gnome.org>
kov@webkit.org [Fri, 24 Apr 2009 13:16:23 +0000 (13:16 +0000)]
2009-04-24  Diego Escalante Urrelo  <diegoe@gnome.org>

        Reviewed by Gustavo Noronha.

        https://bugs.webkit.org/show_bug.cgi?id=15616
        [GTK] Add spell checking

        Add enchant support for spell-checking-languages property to work
        properly.

        * GNUmakefile.am:
        * configure.ac:

WebKit/gtk

2009-04-24  Diego Escalante Urrelo  <diegoe@gnome.org>

        Reviewed by Gustavo Noronha.

        https://bugs.webkit.org/show_bug.cgi?id=15616
        [GTK] Add spell checking

        Add spell-checking-languages property to WebKitWebSettings.

        * webkit/webkitprivate.h:
        * webkit/webkitwebsettings.cpp:
        (_WebKitWebSettingsPrivate::):
        (_WebKitWebSettingsPrivate::webkit_web_settings_class_init):
        (_WebKitWebSettingsPrivate::webkit_web_settings_init):
        (_WebKitWebSettingsPrivate::webkit_web_settings_finalize):
        (_WebKitWebSettingsPrivate::webkit_web_settings_set_property):
        (_WebKitWebSettingsPrivate::webkit_web_settings_get_property):
        (_WebKitWebSettingsPrivate::webkit_web_settings_copy):
        (_WebKitWebSettingsPrivate::webkit_web_settings_get_spell_languages):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42818 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-24 Diego Escalante Urrelo <diegoe@gnome.org>
kov@webkit.org [Fri, 24 Apr 2009 13:16:00 +0000 (13:16 +0000)]
2009-04-24  Diego Escalante Urrelo  <diegoe@gnome.org>

        Reviewed by Gustavo Noronha.

        https://bugs.webkit.org/show_bug.cgi?id=15616
        [GTK] Add spell checking

        Add enable-spell-checking property to WebKitWebSettings and implement
        EditorClient::isContinuousSpellCheckingEnabled and
        EditorClient::toggleContinuousSpellChecking to allow the UI to
        enable check as you type spell checking.

        * WebCoreSupport/EditorClientGtk.cpp:
        (WebKit::EditorClient::isContinuousSpellCheckingEnabled):
        (WebKit::EditorClient::toggleContinuousSpellChecking):
        * webkit/webkitwebsettings.cpp:
        (_WebKitWebSettingsPrivate::):
        (_WebKitWebSettingsPrivate::webkit_web_settings_class_init):
        (_WebKitWebSettingsPrivate::webkit_web_settings_set_property):
        (_WebKitWebSettingsPrivate::webkit_web_settings_get_property):
        (_WebKitWebSettingsPrivate::webkit_web_settings_copy):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-24 Ariya Hidayat <ariya.hidayat@nokia.com>
ariya@webkit.org [Fri, 24 Apr 2009 11:12:51 +0000 (11:12 +0000)]
2009-04-24  Ariya Hidayat  <ariya.hidayat@nokia.com>

        Reviewed by Simon Hausmann.

        Added Qt-specific expected result for color conversion.

        This is to compensate the lack of color profile in Qt to do
        color conversion, hence CMYK 0,0,0,1 always give pure black (#000)
        instead of very dark black.

        * platform/qt/fast/canvas/set-colors-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42816 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-24 Simon Hausmann <simon.hausmann@nokia.com>
hausmann@webkit.org [Fri, 24 Apr 2009 11:08:55 +0000 (11:08 +0000)]
2009-04-24  Simon Hausmann  <simon.hausmann@nokia.com>

        Rubber-stamped by Ariya Hidayat.

        Fix qdoc warning about link to QAction::isEnabled.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42815 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-24 Simon Hausmann <simon.hausmann@nokia.com>
hausmann@webkit.org [Fri, 24 Apr 2009 10:30:07 +0000 (10:30 +0000)]
2009-04-24  Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Ariya Hidayat.

        Added support for generating API docs in the Qt build using "make docs"

        Added code snippets and overview from the Qt sources references in the API docs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42814 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Adele Peterson <adele@apple.com>
adele@apple.com [Fri, 24 Apr 2009 06:38:48 +0000 (06:38 +0000)]
2009-04-23  Adele Peterson  <adele@apple.com>

        Reviewed by Dan Bernstein.

        Updating test to make it easier to see the blockquote in the results.

        * editing/pasteboard/paste-blockquote-into-blockquote-3.html:
        * platform/mac/editing/pasteboard/paste-blockquote-into-blockquote-3-expected.checksum:
        * platform/mac/editing/pasteboard/paste-blockquote-into-blockquote-3-expected.png:
        * platform/mac/editing/pasteboard/paste-blockquote-into-blockquote-3-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42813 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoMove deletionUI-helpers.js in to the correct location so that the
mrowe@apple.com [Fri, 24 Apr 2009 06:03:06 +0000 (06:03 +0000)]
Move deletionUI-helpers.js in to the correct location so that the
tests that use it can find it.

* platform/mac/editing/deleting/resources/deletionUI-helpers.js: Renamed from LayoutTests/editing/deleting/resources/deletionUI-helpers.js.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42812 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWith great sadness and a heavy heart I switch us back from YARR to WREC in
mrowe@apple.com [Fri, 24 Apr 2009 05:12:10 +0000 (05:12 +0000)]
With great sadness and a heavy heart I switch us back from YARR to WREC in
order to restore greenness to the world once more.

* wtf/Platform.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42811 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoMore Windows build fixage.
mrowe@apple.com [Fri, 24 Apr 2009 05:04:11 +0000 (05:04 +0000)]
More Windows build fixage.

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42810 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoAttempt to fix the Windows build.
mrowe@apple.com [Fri, 24 Apr 2009 04:16:30 +0000 (04:16 +0000)]
Attempt to fix the Windows build.

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:  Remove a symbol that no longer exists.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42809 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoBUG 24604: WebKit profiler reports incorrect total times
timothy@apple.com [Fri, 24 Apr 2009 03:47:50 +0000 (03:47 +0000)]
BUG 24604: WebKit profiler reports incorrect total times

JavaScriptCore:

2009-04-23  Francisco Tolmasky  <francisco@280north.com>

        BUG 24604: WebKit profiler reports incorrect total times
        <https://bugs.webkit.org/show_bug.cgi?id=24604>

        Reviewed by Timothy Hatcher and Kevin McCullough.

        * JavaScriptCore.exp:
        * JavaScriptCore.xcodeproj/project.pbxproj:
        * profiler/CallIdentifier.h:
        (JSC::CallIdentifier::Hash::hash):
        (JSC::CallIdentifier::Hash::equal):
        (JSC::CallIdentifier::hash):
        (WTF::):
        * profiler/HeavyProfile.cpp: Removed.
        * profiler/HeavyProfile.h: Removed.
        * profiler/Profile.cpp: No more need for TreeProfile/HeavyProfile
        (JSC::Profile::create):
        * profiler/Profile.h:
        * profiler/ProfileNode.cpp:
        * profiler/ProfileNode.h:
        * profiler/TreeProfile.cpp: Removed.
        * profiler/TreeProfile.h: Removed.

WebCore:

2009-04-23  Francisco Tolmasky  <francisco@280north.com>

        BUG 24604: WebKit profiler reports incorrect total times
        <https://bugs.webkit.org/show_bug.cgi?id=24604>

        Reviewed by Timothy Hatcher and Kevin McCullough.

        Made it so that most of the profiler functions now match the behavior of Shark. Most notably, in the
        heavy view, child nodes now represent the statistics of the root node. Each root node of heavy view
        displays flattened statistics for a particular function that ran during the profile, and each child
        of these root nodes represents a callpath that lead to it. Thus, the statistics for each of these child
        nodes should show how much of the root nodes values came from it. For example, if you had the following to
        stacks take place during the profile:

        A ->calls 1 times-> B ->calls 2 times-> C
        D ->calls 4 times-> C

        The tree for the C root node would look like this:

        C -> B -> A
          -> D

        The number of calls values would look like this:

        C (6) -> B (2) -> A(2)
              -> D (4)

        What this means is that "2 of the total 6 C calls came from B", "2 of the total C calls came from A", and
        "4 of the total C calls came from D".  Notice that the "A ->calls 2 time->" is completely ignored. This becomes
        particularly tricky during recursive calls, because each child note can represent multiple possible paths. This
        is the reason that we would get things like 40000% previously with recursion.

        This is also the way gprof works, and as close as we can get to Shark's behavior (Shark is not instrumented so it
        can't know exactly how many calls came from where, etc).

        * English.lproj/localizedStrings.js: Added "Average" for average times in the profile.
        * inspector/JavaScriptProfile.cpp:
        (WebCore::ProfileClass):
        * inspector/JavaScriptProfileNode.cpp:
        (WebCore::getParent):
        (WebCore::getHead):
        (WebCore::getCallUID):
        (WebCore::ProfileNodeClass):
        * inspector/front-end/BottomUpProfileDataGridTree.js: Added.
        (WebInspector.BottomUpProfileDataGridTree):
        (WebInspector.BottomUpProfileDataGridTree.prototype.focus):
        (WebInspector.BottomUpProfileDataGridNode):
        (WebInspector.BottomUpProfileDataGridNode.prototype._takePropertiesFromProfileDataGridNode):
        (WebInspector.BottomUpProfileDataGridNode.prototype._keepOnlyChild):
        (WebInspector.BottomUpProfileDataGridNode.prototype._exclude):
        (WebInspector.BottomUpProfileDataGridNode.prototype._merge):
        (WebInspector.BottomUpProfileDataGridNode.prototype._populate):
        * inspector/front-end/DataGrid.js:
        (WebInspector.DataGrid.prototype.insertChild):
        (WebInspector.DataGrid.prototype.removeChild):
        (WebInspector.DataGrid.prototype.removeChildren):
        (WebInspector.DataGridNode.prototype.set hasChildren):
        (WebInspector.DataGridNode.prototype.get hasChildren):
        * inspector/front-end/ProfileDataGridTree.js: Added.
        (WebInspector.ProfileDataGridNode):
        (WebInspector.ProfileDataGridNode.prototype.get data.formatMilliseconds):
        (WebInspector.ProfileDataGridNode.prototype.get data):
        (WebInspector.ProfileDataGridNode.prototype.createCell):
        (WebInspector.ProfileDataGridNode.prototype.select):
        (WebInspector.ProfileDataGridNode.prototype.deselect):
        (WebInspector.ProfileDataGridNode.prototype.expand):
        (WebInspector.ProfileDataGridNode.prototype.insertChild):
        (WebInspector.ProfileDataGridNode.prototype.removeChild):
        (WebInspector.ProfileDataGridNode.prototype.removeChildren):
        (WebInspector.ProfileDataGridNode.prototype.findChild):
        (WebInspector.ProfileDataGridNode.prototype.get averageTime):
        (WebInspector.ProfileDataGridNode.prototype.get averagePercent):
        (WebInspector.ProfileDataGridNode.prototype.get selfPercent):
        (WebInspector.ProfileDataGridNode.prototype.get totalPercent):
        (WebInspector.ProfileDataGridNode.prototype._save):
        (WebInspector.ProfileDataGridNode.prototype._restore):
        (WebInspector.ProfileDataGridNode.prototype._merge):
        (WebInspector.ProfileDataGridTree):
        (WebInspector.ProfileDataGridTree.prototype.get expanded):
        (WebInspector.ProfileDataGridTree.prototype.appendChild):
        (WebInspector.ProfileDataGridTree.prototype.insertChild):
        (WebInspector.ProfileDataGridTree.prototype.removeChildren):
        (WebInspector.ProfileDataGridTree.prototype.findChild.WebInspector.ProfileDataGridNode.prototype.findChild.sort.WebInspector.ProfileDataGridNode.prototype.sort._save):
        (WebInspector.ProfileDataGridTree.propertyComparator.comparator):
        (WebInspector.ProfileDataGridTree.propertyComparator.else.comparator):
        (WebInspector.ProfileDataGridTree.propertyComparator):
        * inspector/front-end/ProfileView.js:
        (WebInspector.ProfileView):
        (WebInspector.ProfileView.prototype.set profile):
        (WebInspector.ProfileView.prototype.get bottomUpProfileDataGridTree):
        (WebInspector.ProfileView.prototype.get topDownProfileDataGridTree):
        (WebInspector.ProfileView.prototype.get currentTree):
        (WebInspector.ProfileView.prototype.set currentTree):
        (WebInspector.ProfileView.prototype.get topDownTree):
        (WebInspector.ProfileView.prototype.get bottomUpTree):
        (WebInspector.ProfileView.prototype.refresh):
        (WebInspector.ProfileView.prototype.refreshVisibleData):
        (WebInspector.ProfileView.prototype.refreshShowAsPercents):
        (WebInspector.ProfileView.prototype.performSearch.matchesQuery):
        (WebInspector.ProfileView.prototype.performSearch):
        (WebInspector.ProfileView.prototype._changeView):
        (WebInspector.ProfileView.prototype._focusClicked):
        (WebInspector.ProfileView.prototype._excludeClicked):
        (WebInspector.ProfileView.prototype._resetClicked):
        (WebInspector.ProfileView.prototype._sortProfile):
        * inspector/front-end/ProfilesPanel.js:
        (WebInspector.ProfilesPanel.prototype.showProfile):
        (WebInspector.ProfilesPanel.prototype.showView):
        (WebInspector.ProfilesPanel.prototype.searchMatchFound):
        * inspector/front-end/TopDownProfileDataGridTree.js: Added.
        (WebInspector.TopDownProfileDataGridNode):
        (WebInspector.TopDownProfileDataGridNode.prototype._populate):
        (WebInspector.TopDownProfileDataGridNode.prototype._exclude):
        (WebInspector.TopDownProfileDataGridTree):
        (WebInspector.TopDownProfileDataGridTree.prototype.focus):
        (WebInspector.TopDownProfileDataGridTree.prototype.exclude):
        * inspector/front-end/WebKit.qrc:
        * inspector/front-end/inspector.css:
        * inspector/front-end/inspector.html:

LayoutTests:

2009-04-23  Francisco Tolmasky  <francisco@280north.com>

        BUG 24604: WebKit profiler reports incorrect total times
        <https://bugs.webkit.org/show_bug.cgi?id=24604>

        Reviewed by Timothy Hatcher and Kevin McCullough.

        Changed profile.treeProfile to just profile, since these aren't generated in C++ anymore.
        Removed heavy-view test since heavy-view isn't an actual tree that is generated in C++ land anymore,
        but rather just a different display of the normal treeProfile in the JS data grid.

        * fast/profiler/heavy-view-expected.txt: Removed.
        * fast/profiler/heavy-view.html: Removed.
        * fast/profiler/resources/profiler-test-JS-resources.js: profiles[i].treeProfile -> profiles[i].treeProfile
        (printProfilesDataWithoutTime):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42808 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Reviewed by Sam Weinig.
eric@webkit.org [Fri, 24 Apr 2009 01:50:58 +0000 (01:50 +0000)]
    Reviewed by Sam Weinig.

        Use static functions (and a couple templates)
        to further reduce the amount of copy/paste code in SVGRenderTreeAsText

        No test changes, only code cleanup.

        * rendering/SVGRenderTreeAsText.cpp:
        (WebCore::writeNameValuePair):
        (WebCore::writeNameAndQuotedValue):
        (WebCore::writeIfNotEmpty):
        (WebCore::writeIfNotDefault):
        (WebCore::writeStyle):
        (WebCore::writePositionAndStyle):
        (WebCore::operator<<):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42807 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Kevin McCullough <kmccullough@apple.com>
kmccullough@apple.com [Fri, 24 Apr 2009 01:42:15 +0000 (01:42 +0000)]
2009-04-23  Kevin McCullough  <kmccullough@apple.com>

        - Windows build fix.  Mac only tests go in platform/mac

        * editing/deleting/deletionUI-borders.html: Removed.
        * editing/deleting/deletionUI-differing-background.html: Removed.
        * editing/deleting/deletionUI-minimum-size.html: Removed.
        * platform/mac/editing/deleting/deletionUI-borders.html: Copied from editing/deleting/deletionUI-borders.html.
        * platform/mac/editing/deleting/deletionUI-differing-background.html: Copied from editing/deleting/deletionUI-differing-background.html.
        * platform/mac/editing/deleting/deletionUI-minimum-size.html: Copied from editing/deleting/deletionUI-minimum-size.html.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42806 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Gavin Barraclough <barraclough@apple.com>
barraclough@apple.com [Fri, 24 Apr 2009 01:09:43 +0000 (01:09 +0000)]
2009-04-23  Gavin Barraclough  <barraclough@apple.com>

        Not Reviewed.

        Speculative Windows build fix II.

        * yarr/RegexInterpreter.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42805 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Gavin Barraclough <barraclough@apple.com>
barraclough@apple.com [Fri, 24 Apr 2009 00:53:14 +0000 (00:53 +0000)]
2009-04-23  Gavin Barraclough  <barraclough@apple.com>

        Not Reviewed.

        Speculative Windows build fix.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
        * runtime/RegExp.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42804 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Reviewed by Simon Fraser.
eric@webkit.org [Fri, 24 Apr 2009 00:50:34 +0000 (00:50 +0000)]
    Reviewed by Simon Fraser.

        SVGRenderTreeAsText cleanup/code sharing.

        No test changes, only code cleanup.

        * rendering/RenderText.cpp:
        (WebCore::RenderText::firstRunOrigin):
        * rendering/RenderText.h:
        * rendering/SVGRenderTreeAsText.cpp:
        (WebCore::writeStandardPrefix):
        (WebCore::writeChildren):
        (WebCore::write):
        (WebCore::writeRenderResources):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42803 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Gavin Barraclough <barraclough@apple.com>
barraclough@apple.com [Fri, 24 Apr 2009 00:39:47 +0000 (00:39 +0000)]
2009-04-23  Gavin Barraclough  <barraclough@apple.com>

        Rubber stamped by salty sea dogs Sam & Geoff.

        Enable YARR_JIT by default (where supported), replacing WREC.

        * wtf/Platform.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42802 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Darin Adler <darin@apple.com>
darin@apple.com [Fri, 24 Apr 2009 00:37:11 +0000 (00:37 +0000)]
2009-04-23  Darin Adler  <darin@apple.com>

        Removed some stray files that Geoff missed when rolling out changes.

        * svg/custom/resources/svg-fonts-in-text-controls.js: Removed.
        * svg/custom/svg-fonts-in-text-controls-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42801 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoJavaScriptCore:
barraclough@apple.com [Fri, 24 Apr 2009 00:32:31 +0000 (00:32 +0000)]
JavaScriptCore:

2009-04-23  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by Geoff "Dread Pirate Roberts" Garen.

        Various small fixes to YARR JIT, in preparation for enabling it by default.

        * Correctly index into the callframe when storing restart addresses for
          nested alternatives.
        * Allow backtracking back into matched alternatives of parentheses.
        * Fix callframe offset calculation for parenthetical assertions.
        * When a set of parenthese are quantified with a fixed and variable portion,
          and the variable portion is quantified once, this should not reset the
          pattern match on failure to match (the last match from the firxed portion
          should be preserved).
        * Up the pattern size limit to match PCRE's new limit.
        * Unlclosed parentheses should be reported with the message "missing )".

        * wtf/Platform.h:
        * yarr/RegexCompiler.cpp:
        (JSC::Yarr::RegexPatternConstructor::quantifyAtom):
        (JSC::Yarr::RegexPatternConstructor::setupAlternativeOffsets):
        * yarr/RegexInterpreter.cpp:
        (JSC::Yarr::Interpreter::matchParentheses):
        (JSC::Yarr::Interpreter::backtrackParentheses):
        (JSC::Yarr::ByteCompiler::emitDisjunction):
        * yarr/RegexJIT.cpp:
        (JSC::Yarr::RegexGenerator::loadFromFrameAndJump):
        (JSC::Yarr::RegexGenerator::generateParenthesesDisjunction):
        (JSC::Yarr::RegexGenerator::generateParentheticalAssertion):
        (JSC::Yarr::RegexGenerator::generateTerm):
        (JSC::Yarr::executeRegex):
        * yarr/RegexParser.h:
        (JSC::Yarr::Parser::):
        (JSC::Yarr::Parser::parseTokens):
        (JSC::Yarr::Parser::parse):
        * yarr/RegexPattern.h:
        (JSC::Yarr::PatternTerm::):
        (JSC::Yarr::PatternTerm::PatternTerm):

LayoutTests:

2009-04-23  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by Geoff "Dread Pirate Roberts" Garen.

        This test tries to force itself into PCRE; modify the
        test so that as well as dodging WREC it can also avoid
        YARR!

        * fast/js/resources/regexp-overflow-too-big.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42800 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
bdakin@apple.com [Fri, 24 Apr 2009 00:25:08 +0000 (00:25 +0000)]
WebCore:

2009-04-23  Beth Dakin  <bdakin@apple.com>

        Reviewed by Darin Adler.

        WebCore part of fix for <rdar://problem/6333461> REGRESSION
        (r36864-r36869): Dragging stocks widget scrollbar drags the whole
        widget

        * platform/Scrollbar.h:
        (WebCore::Scrollbar::isScrollbar):
        * platform/Widget.h:
        (WebCore::Widget::isScrollbar):

WebKit/mac:

2009-04-23  Beth Dakin  <bdakin@apple.com>

        Reviewed by Darin Adler.

        Fix for <rdar://problem/6333461> REGRESSION (r36864-r36869):
        Dragging stocks widget scrollbar drags the whole widget

        Look for our new WebCore scrollbars in the WebHTMLView and add
        proper Dashboard regions for them.

        * WebView/WebView.mm:
        (-[WebView _addScrollerDashboardRegionsForFrameView:dashboardRegions:]):
        (-[WebView _addScrollerDashboardRegions:from:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42799 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
kmccullough@apple.com [Thu, 23 Apr 2009 23:59:37 +0000 (23:59 +0000)]
WebCore:

2009-04-23  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Adam Roben.

        <rdar://problem/6808109> "Deletion UI" is not available for many
        portions of HTML content

        This patch makes the deletion UI show up in some new situations:
        1) If a block's background color is different from its parent's
        2) If a block has a background image.
        3) If a block has a single visible border.
        However the block must now not only be at least a minimum width and
        height, but also exceed a minimum area.  In practice this has led to
        much better element selection.

        * editing/DeleteButtonController.cpp:
        (WebCore::isDeletableElement):

LayoutTests:

2009-04-23  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Adam Roben.

        <rdar://problem/6808109> "Deletion UI" is not available for many
        portions of HTML content

        This patch makes the deletion UI show up in some new situations:
        1) If a block's background color is different from its parent's
        2) If a block has a background image.
        3) If a block has a single visible border.
        However the block must now not only be at least a minimum width and
        height, but also exceed a minimum area.  In practice this has led to
        much better element selection.

        * editing/deleting/deletionUI-borders.html: Added.
        * editing/deleting/deletionUI-differing-background.html: Added.
        * editing/deleting/deletionUI-minimum-size.html: Added.
        * editing/deleting/resources: Added.
        * editing/deleting/resources/deletionUI-helpers.js: Added.
        (debug):
        (determineDeletionUIExistence):
        * platform/mac/editing/deleting/deletionUI-borders-expected.txt: Added.
        * platform/mac/editing/deleting/deletionUI-differing-background-expected.txt: Added.
        * platform/mac/editing/deleting/deletionUI-minimum-size-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42798 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agowx build fix. Adding XMLTokenizer.cpp to the build.
kevino@webkit.org [Thu, 23 Apr 2009 23:53:33 +0000 (23:53 +0000)]
wx build fix. Adding XMLTokenizer.cpp to the build.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42797 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 John Sullivan <sullivan@apple.com>
sullivan@apple.com [Thu, 23 Apr 2009 23:32:01 +0000 (23:32 +0000)]
2009-04-23  John Sullivan  <sullivan@apple.com>

        fixed <rdar://problem/6822479> Assertion failure after Reset Safari in new history-writing code

        Reviewed by Oliver Hunt

        * History/WebHistory.mm:
        (-[WebHistoryPrivate data]):
        Return nil immediately if there are no entries; this matches a recent Windows-platform fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42796 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Adele Peterson <adele@apple.com>
adele@apple.com [Thu, 23 Apr 2009 22:46:40 +0000 (22:46 +0000)]
2009-04-23  Adele Peterson  <adele@apple.com>

        Reviewed by Darin Adler.

        Fix for <rdar://problem/6788905> REGRESSION (41928?): hang in Position::upstream
        https://bugs.webkit.org/show_bug.cgi?id=25312

        I was unable to reproduce the problem, but I'm pretty sure this will fix it.

        * rendering/RenderObject.cpp: (WebCore::RenderObject::createVisiblePosition):
        Since VisiblePosition doesn't ensure the offset is good, we shouldn't pass max int as an offset.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoVersioning.
mrowe@apple.com [Thu, 23 Apr 2009 22:38:57 +0000 (22:38 +0000)]
Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42793 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
dglazkov@chromium.org [Thu, 23 Apr 2009 22:33:25 +0000 (22:33 +0000)]
WebCore:

2009-04-23  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=25313
        Missing scroll bars in GMail.

        Test: fast/overflow/overflow-y-scroll.html

        * platform/ScrollView.cpp:
        (WebCore::ScrollView::updateScrollbars): Added check for the ScrollbarAlwaysOn
        scroll mode.

LayoutTests:

2009-04-23  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.
        https://bugs.webkit.org/show_bug.cgi?id=25313
        Missing scroll bars in GMail.

        * fast/overflow/overflow-y-scroll-expected.txt: Added.
        * fast/overflow/overflow-y-scroll.html: Added.

WebKit/mac:

2009-04-23  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=25313
        Missing scroll bars in GMail.

        * WebView/WebDynamicScrollBarsView.m:
        (-[WebDynamicScrollBarsView updateScrollers]): Added check for the ScrollbarAlwaysOn
            scroll mode.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42791 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoFix the Tiger build
kdecker@apple.com [Thu, 23 Apr 2009 22:17:46 +0000 (22:17 +0000)]
Fix the Tiger build

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Anders Carlsson <andersca@apple.com>
andersca@apple.com [Thu, 23 Apr 2009 22:03:17 +0000 (22:03 +0000)]
2009-04-23  Anders Carlsson  <andersca@apple.com>

        Reviewed by Geoffrey Garen.

        Fix <rdar://problem/6821992>

        Add a new m_inDestroy member variable. Set it to true when in destroy, and have all NPRuntime functions return false
        when m_inDestroy is true.

        * Plugins/Hosted/NetscapePluginInstanceProxy.h:
        * Plugins/Hosted/NetscapePluginInstanceProxy.mm:
        (WebKit::NetscapePluginInstanceProxy::NetscapePluginInstanceProxy):
        (WebKit::NetscapePluginInstanceProxy::destroy):
        (WebKit::NetscapePluginInstanceProxy::invoke):
        (WebKit::NetscapePluginInstanceProxy::invokeDefault):
        (WebKit::NetscapePluginInstanceProxy::construct):
        (WebKit::NetscapePluginInstanceProxy::getProperty):
        (WebKit::NetscapePluginInstanceProxy::setProperty):
        (WebKit::NetscapePluginInstanceProxy::removeProperty):
        (WebKit::NetscapePluginInstanceProxy::hasProperty):
        (WebKit::NetscapePluginInstanceProxy::hasMethod):
        (WebKit::NetscapePluginInstanceProxy::enumerate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42789 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
mitz@apple.com [Thu, 23 Apr 2009 21:22:30 +0000 (21:22 +0000)]
WebCore:

        Reviewed by Justin Garcia.

        - fix <rdar://problem/6081309> Mail crash when pressing down arrow in
          some messages in WebCore::canHaveChildrenForEditing

        Test: editing/selection/extend-by-line-anonymous-content-crash.html

        * editing/visible_units.cpp:
        (WebCore::previousLinePosition): Null-check node. If p is not an
        editable position, then closestLeafChildForXPos() may have returned a
        non-editable box, and in particular one belonging to anonymous content.
        If node is 0, fall back on RenderObject::positionForPoint, which
        finds the closest position in non-anonymous content.
        (WebCore::nextLinePosition): Ditto.
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::createVisiblePosition): Fixed a typo.

LayoutTests:

        Reviewed by Justin Garcia.

        - test for <rdar://problem/6081309> Mail crash when pressing down arrow
          in some messages in WebCore::canHaveChildrenForEditing

        * editing/selection/extend-by-line-anonymous-content-crash-expected.txt: Added.
        * editing/selection/extend-by-line-anonymous-content-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42788 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 David Hyatt <hyatt@apple.com>
hyatt@apple.com [Thu, 23 Apr 2009 20:49:38 +0000 (20:49 +0000)]
2009-04-23  David Hyatt  <hyatt@apple.com>

        Reviewed by Maciej.

        Fix for <rdar://problem/6789879> REGRESSION (42464): Hitting assertion when loading message in Mail + TOT WebKit

        Make the Mac platform the same as all the other platforms.  Instead of (incorrectly) marking a FrameView for layout
        when its underlying document view changes, just mark the outermost frame view for layout when the WebView's size changes.

        * WebView/WebHTMLView.mm:
        (-[WebHTMLView _frameOrBoundsChanged]):
        (-[WebHTMLView addSuperviewObservers]):
        * WebView/WebView.mm:
        (-[WebView _boundsChanged]):
        (-[WebView removeSizeObservers]):
        (-[WebView addSizeObservers]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Reviewed by Tim Hatcher.
kdecker@apple.com [Thu, 23 Apr 2009 20:08:12 +0000 (20:08 +0000)]
    Reviewed by Tim Hatcher.

        Second part of the fix for <rdar://problem/6352982>

        * Plugins/WebBaseNetscapePluginView.h: Make this class conform to WebPluginContainerCheckController
        * Plugins/WebBaseNetscapePluginView.mm: Likewise.
        (-[WebBaseNetscapePluginView _webPluginContainerCancelCheckIfAllowedToLoadRequest:]): Added skeleton method.
        * Plugins/WebPluginContainerCheck.h: Added protocol for <WebPluginContainerCheckController>
        * Plugins/WebPluginContainerCheck.mm:
        (-[WebPluginContainerCheck initWithRequest:target:resultObject:selector:controller:]):
        * Plugins/WebPluginController.h: Make this class conform to WebPluginContainerCheckController

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Reviewed by Adele Peterson.
ap@webkit.org [Thu, 23 Apr 2009 19:58:08 +0000 (19:58 +0000)]
    Reviewed by Adele Peterson.

        <rdar://problem/6791439> REGRESSION: Get an error page instead of login page navigating back in gmail

        Test: http/tests/history/back-to-post.php

        * platform/network/cf/ResourceRequestCFNet.cpp: (WebCore::ResourceRequest::doUpdatePlatformRequest):
        Apply a new cache policy if it changed after the platform request was first created.
        While at it, also made sure to update timeout.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>
kov@webkit.org [Thu, 23 Apr 2009 19:02:34 +0000 (19:02 +0000)]
2009-04-23  Gustavo Noronha Silva  <gustavo.noronha@collabora.co.uk>

        Rubber-stamped by Xan Lopez.

        * pt_BR.po: translation update.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42784 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Zan Dobersek <zandobersek@gmail.com>
kov@webkit.org [Thu, 23 Apr 2009 18:45:15 +0000 (18:45 +0000)]
2009-04-23  Zan Dobersek  <zandobersek@gmail.com>

        Reviewed by Gustavo Noronha.

        When creating a new drawable object, collect system's
        information about the best depth it can provide and use it
        in a new colormap that the new drawable requires.

        * platform/graphics/gtk/ImageGtk.cpp:
        (WebCore::BitmapImage::getGdkPixbuf):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Reviewed by Anders Carlsson.
kdecker@apple.com [Thu, 23 Apr 2009 18:06:06 +0000 (18:06 +0000)]
    Reviewed by Anders Carlsson.

        First part of <rdar://problem/6352982>

        * Plugins/WebBaseNetscapePluginView.h: Imported #WebNetscapeContainerCheckPrivate.h; Added two new method:
        checkIfAllowedToLoadURL:url:frame:callbackFunc: and cancelCheckIfAllowedToLoadURL:
        * Plugins/WebBaseNetscapePluginView.mm:
        (-[WebBaseNetscapePluginView checkIfAllowedToLoadURL:frame:callbackFunc:]): Added skeleton method, does nothing interesting
        yet.
        (-[WebBaseNetscapePluginView cancelCheckIfAllowedToLoadURL:]): Likewise.
        * Plugins/WebNetscapeContainerCheckPrivate.h: Added.
        * Plugins/WebNetscapeContainerCheckPrivate.mm: Added.
        (browserContainerCheckFuncs): Added.
        * Plugins/WebNetscapePluginView.h: Imported #WebNetscapeContainerCheckPrivate.h;
        * Plugins/WebNetscapePluginView.mm: Imported #WebNetscapeContainerCheckPrivate.h; added WKN_CheckIfAllowedToLoadURL
        and WKN_CancelCheckIfAllowedToLoadURL functions.
        (-[WebNetscapePluginView getVariable:value:]): Return vtable for container check functions.
        * Plugins/npapi.mm:
        (WKN_CheckIfAllowedToLoadURL): Added new private function.
        (WKN_CancelCheckIfAllowedToLoadURL): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Skip a newly-added test that fails on Windows
aroben@apple.com [Thu, 23 Apr 2009 16:42:59 +0000 (16:42 +0000)]
    Skip a newly-added test that fails on Windows

        See Bug 25344: http/tests/misc/redirect-to-external-url.html fails on
        Windows
        <https://bugs.webkit.org/show_bug.cgi?id=25344>

        Reviewed by Brady Eidson.

        * platform/win/Skipped: Added
        http/tests/misc/redirect-to-external-url.html.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Zan Dobersek <zandobersek@gmail.com>
kov@webkit.org [Thu, 23 Apr 2009 15:39:29 +0000 (15:39 +0000)]
2009-04-23  Zan Dobersek  <zandobersek@gmail.com>

        Reviewed by Gustavo Noronha.

        https://bugs.webkit.org/show_bug.cgi?id=15654
        GdkPixbuf support for ImageCairo

        Add support for converting a Cairo surface to a GdkPixbuf.

        * platform/graphics/BitmapImage.h:
        * platform/graphics/Image.h:
        (WebCore::Image::getGdkPixbuf):
        * platform/graphics/gtk/ImageGtk.cpp:
        (WebCore::BitmapImage::getGdkPixbuf):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoFix Qt build warning
vestbo@webkit.org [Thu, 23 Apr 2009 15:07:11 +0000 (15:07 +0000)]
Fix Qt build warning

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42779 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>
ariya@webkit.org [Thu, 23 Apr 2009 14:55:30 +0000 (14:55 +0000)]
2009-04-23  Kenneth Rohde Christiansen  <kenneth.christiansen@openbossa.org>

        Reviewed by Simon Hausmann.

        Make windowed plugins (on X11, Qt) behave nicer with wheel scrolling.

        Basically, what we want is to send wheel events to the plugin *only*
        when it has focus, or else send it to the parent frame.

        This is a work around, and basically what we do, is to steal the
        wheel events from the plugin, modify the window ID, and put the
        event back into the Qt event loop.

        * plugins/qt/PluginContainerQt.cpp:
        (PluginClientWrapper::PluginClientWrapper):
        (PluginClientWrapper::~PluginClientWrapper):
        (PluginClientWrapper::x11Event):
        (PluginContainerQt::PluginContainerQt):
        (PluginContainerQt::~PluginContainerQt):
        (PluginContainerQt::on_clientClosed):
        (PluginContainerQt::on_clientIsEmbedded):
        (PluginContainerQt::redirectWheelEventsToParent):
        (PluginContainerQt::x11Event):
        (PluginContainerQt::focusInEvent):
        * plugins/qt/PluginContainerQt.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42778 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Antonio Gomes <antonio.gomes@openbossa.org>
ariya@webkit.org [Thu, 23 Apr 2009 13:53:10 +0000 (13:53 +0000)]
2009-04-23  Antonio Gomes   <antonio.gomes@openbossa.org>

        Reviewed by Ariya Hidayat.

        [Qt] Added QWebElement::computedStyleProperty method.

        * Api/qwebelement.cpp:
        (QWebElement::computedStyleProperty):
        * Api/qwebelement.h:
        * tests/qwebelement/tst_qwebelement.cpp:
        (tst_QWebElement::computedStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42777 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
vestbo@webkit.org [Thu, 23 Apr 2009 11:30:50 +0000 (11:30 +0000)]
2009-04-23  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>

        Rubber-stamped by Ariya Hidayat.

        [Qt] Rename QWebElement arguments from html to markup

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42776 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoFix the Windows DRT build.
mrowe@apple.com [Thu, 23 Apr 2009 07:25:06 +0000 (07:25 +0000)]
Fix the Windows DRT build.

* DumpRenderTree/win/PolicyDelegate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42775 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Brady Eidson <beidson@apple.com>
beidson@apple.com [Thu, 23 Apr 2009 07:10:51 +0000 (07:10 +0000)]
2009-04-23  Brady Eidson  <beidson@apple.com>

        Fix Windows DRT build.

        * DumpRenderTree/win/PolicyDelegate.cpp:
        (PolicyDelegate::unableToImplementPolicyWithError):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42774 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-23 Brady Eidson <beidson@apple.com>
beidson@apple.com [Thu, 23 Apr 2009 07:09:42 +0000 (07:09 +0000)]
2009-04-23  Brady Eidson  <beidson@apple.com>

        Update my just-landed layout test to agree with Tiger's CFNetwork

        * http/tests/misc/resources/redirect-to-external-url.php:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42773 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
beidson@apple.com [Thu, 23 Apr 2009 06:42:52 +0000 (06:42 +0000)]
WebCore:

2009-04-22  Brady Eidson  <beidson@apple.com>

        Reviewed by Maciej Stachowiak

        <rdar://problem/6786961> - "Are you sure you want to resubmit this form?" nag displays twice.

        For most loads, we were consulting the navigation policy delegate twice.  Once from FrameLoader before
        the load started and once from MainResourceLoader in its willSendRequest callback.

        In the past we tried removing MainResourceLoader's call altogether.  This caused a regression where urls
        that redirect to a url handled by an external application would no longer work in Safari.  It probably
        also broke other WebKit apps in subtle ways.

        Changing MainResourceLoader to make the check only on redirects fixes both bugs.  We now only call the
        policy delegate once for most standard loads, but we do correctly call it a second time for redirects.

        Tests: http/tests/misc/policy-delegate-called-twice.html
               http/tests/misc/redirect-to-external-url.html

        * loader/MainResourceLoader.cpp:
        (WebCore::MainResourceLoader::willSendRequest): Call the navigation policy delegate only for redirects.

WebKitTools:

2009-04-22  Brady Eidson  <beidson@apple.com>

        Reviewed by Maciej Stachowiak

        Add additional support for test:
        <rdar://problem/5689748> - Cannot redirect to protocols handled by external applications.

        * DumpRenderTree/mac/PolicyDelegate.mm:
        (-[PolicyDelegate webView:unableToImplementPolicyWithError:frame:]): Dump formatted information when
          a policy cannot be implemented.

        * DumpRenderTree/win/PolicyDelegate.cpp:
        (PolicyDelegate::unableToImplementPolicyWithError): Ditto.
        * DumpRenderTree/win/PolicyDelegate.h:

LayoutTests:

2009-04-22  Brady Eidson  <beidson@apple.com>

        Reviewed by Maciej Stachowiak

        With the fix for:
        <rdar://problem/6786961> - "Are you sure you want to resubmit this form?" nag displays twice.
        And also to make sure we never rebreak:
        <rdar://problem/5689748> - Cannot redirect to protocols handled by external applications.

        * http/tests/misc/redirect-to-external-url-expected.txt: Added.
        * http/tests/misc/redirect-to-external-url.html: Added.
        * http/tests/misc/resources/redirect-to-external-url-iframe.html: Added.
        * http/tests/misc/resources/redirect-to-external-url.php: Added.

        * http/tests/misc/policy-delegate-called-twice-expected.txt: Added.
        * http/tests/misc/policy-delegate-called-twice.html: Added.
        * http/tests/misc/resources/iframe-policy-1.html: Added.
        * http/tests/misc/resources/iframe-policy-2.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42772 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoAdd the m_ prefix on FixedVMPoolAllocator's member variables, and fix typos in a...
mrowe@apple.com [Thu, 23 Apr 2009 04:06:26 +0000 (04:06 +0000)]
Add the m_ prefix on FixedVMPoolAllocator's member variables, and fix typos in a few comments.

Rubber-stamped by Gavin Barraclough.

* jit/ExecutableAllocatorFixedVMPool.cpp:
(JSC::FixedVMPoolAllocator::addToFreeList):
(JSC::FixedVMPoolAllocator::coalesceFreeSpace):
(JSC::FixedVMPoolAllocator::FixedVMPoolAllocator):
(JSC::FixedVMPoolAllocator::alloc):
(JSC::FixedVMPoolAllocator::free):
(JSC::FixedVMPoolAllocator::isWithinVMPool):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42771 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoAdd some assertions to FixedVMPoolAllocator to guard against cases where we
mrowe@apple.com [Thu, 23 Apr 2009 04:06:22 +0000 (04:06 +0000)]
Add some assertions to FixedVMPoolAllocator to guard against cases where we
attempt to free memory that didn't originate from the pool, or we attempt to
hand out a bogus address from alloc.

Rubber-stamped by Gavin Barraclough.

* jit/ExecutableAllocatorFixedVMPool.cpp:
(JSC::FixedVMPoolAllocator::release):
(JSC::FixedVMPoolAllocator::FixedVMPoolAllocator):
(JSC::FixedVMPoolAllocator::alloc):
(JSC::FixedVMPoolAllocator::free):
(JSC::FixedVMPoolAllocator::isWithinVMPool):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42770 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Sam Weinig <sam@webkit.org>
weinig@apple.com [Thu, 23 Apr 2009 02:41:22 +0000 (02:41 +0000)]
2009-04-22  Sam Weinig  <sam@webkit.org>

        Reviewed by Mark Rowe.

        Fix for <rdar://problem/6800695>

        Add a hack to only allow navigating (via a link from a http page)
        to feed: urls (same for feeds: and feedsearch:) that map to http:
        or https: via their nested protocol. This includes both feed://example.com
        and feed:http://example.com.

        * loader/FrameLoader.cpp:
        (WebCore::isFeedWithNestedProtocolInHTTPFamily):
        (WebCore::FrameLoader::loadFrameRequest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42769 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Sam Weinig <sam@webkit.org>
weinig@apple.com [Thu, 23 Apr 2009 02:34:40 +0000 (02:34 +0000)]
2009-04-22  Sam Weinig  <sam@webkit.org>

        Reviewed by Anders Carlsson.

        Autogenerate the XMLHttpRequest constructor getter/setter for the
        WorkerContext and remove the now no longer needed JSWorkerContextBase.lut.h
        file.

        * DerivedSources.make: Remove JSWorkerContextBase.lut.h
        * WebCore.xcodeproj/project.pbxproj: Ditto.

        * bindings/js/JSWorkerContextBase.cpp:
        * bindings/js/JSWorkerContextBase.h: Remove code to support hand written
        code for XMLHttpRequest constructor lookup and shadowing.

        * bindings/js/JSWorkerContextCustom.cpp:
        (WebCore::JSWorkerContext::customGetOwnPropertySlot): Remove JSC prefixes.
        (WebCore::JSWorkerContext::xmlHttpRequest): Added. Custom implementation of
        cached constructor lookup.
        * workers/WorkerContext.idl: Make the XMLHttpRequestConstructor accessor not
        V8 specific.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebKitLibraries:
adachan@apple.com [Thu, 23 Apr 2009 01:43:07 +0000 (01:43 +0000)]
WebKitLibraries:
        Update WebKitSystemInterface with new method that maps CFNetwork error code to localized description.

        Reviewed by Darin Adler.

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

WebKit/win:
        <rdar://problem/5115298> Don't get error msg when download is interrupted and cannot resume due to server unavailability
        If we fail to get an error description back from CFNetwork, try to get the localized description
        based on the error code.

        Reviewed by Darin Adler.

        * WebError.cpp:
        (WebError::localizedDescription):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Darin Adler <darin@apple.com>
darin@apple.com [Thu, 23 Apr 2009 01:07:51 +0000 (01:07 +0000)]
2009-04-22  Darin Adler  <darin@apple.com>

        * page/Navigator.idl: Touch this file to fix build since Geolocation
        configuration was recently turned off.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoInclude updated results
oliver@apple.com [Thu, 23 Apr 2009 01:02:01 +0000 (01:02 +0000)]
Include updated results

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoModify test to only record the important mouse events
oliver@apple.com [Thu, 23 Apr 2009 00:59:39 +0000 (00:59 +0000)]
Modify test to only record the important mouse events

Reviewed by Geoff Garen.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
justin.garcia@apple.com [Thu, 23 Apr 2009 00:54:42 +0000 (00:54 +0000)]
WebCore:

2009-04-22  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=25322
        Can't delete empty paragraph after a block image

        Empty paragraph removal would have been handled by mergeParagraphs, but we stopped short
        because of some code that avoided problems that would arise from Positions that would go
        bad during deletion.

        We have several checks that prevent us from using bad Positions, this one in particular
        was added to avoid bad Positions that result from text removal.  So, I removed the check
        and started correctly updating m_downstreamEnd during text removal so that it doesn't go
        bad.  m_upstreamStart doesn't need to be updated during text removal, since only text
        after it will ever be removed.

        * editing/DeleteSelectionCommand.cpp:
        (WebCore::DeleteSelectionCommand::deleteTextFromNode):
        (WebCore::DeleteSelectionCommand::handleGeneralDelete):
        (WebCore::DeleteSelectionCommand::mergeParagraphs):

LayoutTests:

2009-04-22  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=25322
        Can't delete empty paragraph after a block image

        * editing/deleting/25322-1-expected.txt: Added.
        * editing/deleting/25322-1.html: Added.
        * editing/deleting/25322-2-expected.txt: Added.
        * editing/deleting/25322-2.html: Added.
        * editing/deleting/25322-3-expected.txt: Added.
        * editing/deleting/25322-3.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Darin Adler <darin@apple.com>
darin@apple.com [Thu, 23 Apr 2009 00:49:26 +0000 (00:49 +0000)]
2009-04-22  Darin Adler  <darin@apple.com>

        * platform/mac/svg/W3C-SVG-1.1/animate-elem-36-t-expected.txt: Fixed one
        case Eric seems to have missed earlier today of RenderImage -> RenderSVGImage.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42762 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago<rdar://problem/6757346> SAP: Prevent default on mouseDown does not stop iframe from...
oliver@apple.com [Thu, 23 Apr 2009 00:37:51 +0000 (00:37 +0000)]
<rdar://problem/6757346> SAP: Prevent default on mouseDown does not stop iframe from capturing subsequent mouse moves

Reviewed by Darin Adler.

Make mouseUp target the correct frame when the original
mousedown made the drag non-capturing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42761 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago<rdar://problem/6757346> SAP: Prevent default on mouseDown does not stop iframe from...
oliver@apple.com [Thu, 23 Apr 2009 00:29:04 +0000 (00:29 +0000)]
<rdar://problem/6757346> SAP: Prevent default on mouseDown does not stop iframe from capturing subsequent mouse moves

Reviewed by Darin Adler.

This is the first step of allowing drag events to match the behaviour
of mouse move events when the default action of the initial mouse down
is prevented.  Remaining issue is that the final mouse up event still
targets the original root frame.

Test: fast/events/mouse-drag-from-frame-to-other-frame.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42760 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Simon Fraser <simon.fraser@apple.com>
simon.fraser@apple.com [Thu, 23 Apr 2009 00:01:18 +0000 (00:01 +0000)]
2009-04-22  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Adele Peterson

        <rdar://problem/6702882> QT movie's video fails to be displayed after switching between tabs
        <rdar://problem/6754957> Resizing Web browser window and hitting play will cause video blocks and artifacting

        When in a media document, MediaPlayerPrivateQTKit uses a QTMovieView which may get
        layer backing under some circumstances. However, drawing the view via
        displayRectIgnoringOpacity:inContext: bypasses any layer setup that AppKit normally performs.
        So when in the media document, we draw via displayRectIgnoringOpacity:.

        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        (WebCore::MediaPlayerPrivate::paint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42759 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoUpdate test results after Blackbeard's change in r42752 to disable geolocation.
mrowe@apple.com [Wed, 22 Apr 2009 23:56:00 +0000 (23:56 +0000)]
Update test results after Blackbeard's change in r42752 to disable geolocation.

* fast/dom/Window/window-properties-expected.txt:
* fast/dom/navigator-detached-no-crash-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42758 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Pierre d'Herbemont <pdherbemont@apple.com>
simon.fraser@apple.com [Wed, 22 Apr 2009 22:59:23 +0000 (22:59 +0000)]
2009-04-22  Pierre d'Herbemont  <pdherbemont@apple.com>

        Reviewed by Simon Fraser.

        https://bugs.webkit.org/show_bug.cgi?id=24849

        Add tests for media 'played' attribute (missed in r42619).

        * media/video-played-expected.txt: Added.
        * media/video-played.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42757 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Anders Carlsson <andersca@apple.com>
andersca@apple.com [Wed, 22 Apr 2009 22:51:42 +0000 (22:51 +0000)]
2009-04-22  Anders Carlsson  <andersca@apple.com>

        Reviewed by Oliver Hunt.

        Fix <rdar://problem/6792694>

        When we're trying to instantiate a plug-in and the plug-in host has died, we need to invalidate the
        instance so that it doesn't stick around and do bad things.

        * Plugins/Hosted/NetscapePluginHostManager.mm:
        (WebKit::NetscapePluginHostManager::instantiatePlugin):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42756 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Gavin Barraclough <barraclough@apple.com>
barraclough@apple.com [Wed, 22 Apr 2009 21:52:08 +0000 (21:52 +0000)]
2009-04-22  Gavin Barraclough  <barraclough@apple.com>

        Rubber stamped by Sam "Blackbeard" Weinig.

        Although pirates do spell the word 'generate' as 'genertate',
        webkit developers do not.  Fixerate.

        * yarr/RegexJIT.cpp:
        (JSC::Yarr::RegexGenerator::generateAssertionBOL):
        (JSC::Yarr::RegexGenerator::generateAssertionEOL):
        (JSC::Yarr::RegexGenerator::generateAssertionWordBoundary):
        (JSC::Yarr::RegexGenerator::generatePatternCharacterSingle):
        (JSC::Yarr::RegexGenerator::generatePatternCharacterPair):
        (JSC::Yarr::RegexGenerator::generatePatternCharacterFixed):
        (JSC::Yarr::RegexGenerator::generatePatternCharacterGreedy):
        (JSC::Yarr::RegexGenerator::generatePatternCharacterNonGreedy):
        (JSC::Yarr::RegexGenerator::generateCharacterClassSingle):
        (JSC::Yarr::RegexGenerator::generateCharacterClassFixed):
        (JSC::Yarr::RegexGenerator::generateCharacterClassGreedy):
        (JSC::Yarr::RegexGenerator::generateCharacterClassNonGreedy):
        (JSC::Yarr::RegexGenerator::generateTerm):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42755 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Gavin Barraclough <barraclough@apple.com>
barraclough@apple.com [Wed, 22 Apr 2009 21:44:49 +0000 (21:44 +0000)]
2009-04-22  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by Sam "Blackbeard" Weinig.

        Improvements to YARR JIT.  This patch expands support in three key areas:
            * Add (temporary) support for falling back to PCRE for expressions not supported.
            * Add support for x86_64 and Windows.
            * Add support for singly quantified parentheses (? and ??), alternatives within
              parentheses, and parenthetical assertions.

        * runtime/RegExp.cpp:
        (JSC::RegExp::match):
        * yarr/RegexJIT.cpp:
        (JSC::Yarr::RegexGenerator::storeToFrame):
        (JSC::Yarr::RegexGenerator::storeToFrameWithPatch):
        (JSC::Yarr::RegexGenerator::loadFromFrameAndJump):
        (JSC::Yarr::RegexGenerator::AlternativeBacktrackRecord::AlternativeBacktrackRecord):
        (JSC::Yarr::RegexGenerator::TermGenerationState::resetAlternative):
        (JSC::Yarr::RegexGenerator::TermGenerationState::resetTerm):
        (JSC::Yarr::RegexGenerator::TermGenerationState::jumpToBacktrack):
        (JSC::Yarr::RegexGenerator::TermGenerationState::plantJumpToBacktrackIfExists):
        (JSC::Yarr::RegexGenerator::TermGenerationState::addBacktrackJump):
        (JSC::Yarr::RegexGenerator::TermGenerationState::linkAlternativeBacktracks):
        (JSC::Yarr::RegexGenerator::TermGenerationState::propagateBacktrackingFrom):
        (JSC::Yarr::RegexGenerator::genertateAssertionBOL):
        (JSC::Yarr::RegexGenerator::genertateAssertionEOL):
        (JSC::Yarr::RegexGenerator::matchAssertionWordchar):
        (JSC::Yarr::RegexGenerator::genertateAssertionWordBoundary):
        (JSC::Yarr::RegexGenerator::genertatePatternCharacterSingle):
        (JSC::Yarr::RegexGenerator::genertatePatternCharacterPair):
        (JSC::Yarr::RegexGenerator::genertatePatternCharacterFixed):
        (JSC::Yarr::RegexGenerator::genertatePatternCharacterGreedy):
        (JSC::Yarr::RegexGenerator::genertatePatternCharacterNonGreedy):
        (JSC::Yarr::RegexGenerator::genertateCharacterClassSingle):
        (JSC::Yarr::RegexGenerator::genertateCharacterClassFixed):
        (JSC::Yarr::RegexGenerator::genertateCharacterClassGreedy):
        (JSC::Yarr::RegexGenerator::genertateCharacterClassNonGreedy):
        (JSC::Yarr::RegexGenerator::generateParenthesesDisjunction):
        (JSC::Yarr::RegexGenerator::generateParenthesesSingle):
        (JSC::Yarr::RegexGenerator::generateParentheticalAssertion):
        (JSC::Yarr::RegexGenerator::generateTerm):
        (JSC::Yarr::RegexGenerator::generateDisjunction):
        (JSC::Yarr::RegexGenerator::generateEnter):
        (JSC::Yarr::RegexGenerator::generateReturn):
        (JSC::Yarr::RegexGenerator::RegexGenerator):
        (JSC::Yarr::RegexGenerator::generate):
        (JSC::Yarr::RegexGenerator::compile):
        (JSC::Yarr::RegexGenerator::generationFailed):
        (JSC::Yarr::jitCompileRegex):
        (JSC::Yarr::executeRegex):
        * yarr/RegexJIT.h:
        (JSC::Yarr::RegexCodeBlock::RegexCodeBlock):
        (JSC::Yarr::RegexCodeBlock::~RegexCodeBlock):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42754 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Fix a crash in Mail when forwarding a specific HTML email.
timothy@apple.com [Wed, 22 Apr 2009 21:24:05 +0000 (21:24 +0000)]
    Fix a crash in Mail when forwarding a specific HTML email.
        Caused by the fix for <rdar://problem/6512520> (r42096).
        Unable to create a test case that works outside of Mail.

        <rdar://problem/6796069>

        Reviewed by Ada Chan.

        * editing/markup.cpp:
        (WebCore::createMarkup): Null check pastEnd before trying to
        call Range::compareBoundaryPoints.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42753 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoJavaScriptCore:
weinig@apple.com [Wed, 22 Apr 2009 19:09:14 +0000 (19:09 +0000)]
JavaScriptCore:

2009-04-22  Sam Weinig  <sam@webkit.org>

        Rubber-stamped by Darin Adler.

        Fix for <rdar://problem/6816957>
        Turn off Geolocation by default

        * Configurations/FeatureDefines.xcconfig:

WebCore:

2009-04-22  Sam Weinig  <sam@webkit.org>

        Rubber-stamped by Darin Adler.

        Fix for <rdar://problem/6816957>
        Turn off Geolocation by default

        * Configurations/FeatureDefines.xcconfig:

WebKit/mac:

2009-04-22  Sam Weinig  <sam@webkit.org>

        Rubber-stamped by Darin Adler.

        Fix for <rdar://problem/6816957>
        Turn off Geolocation by default

        * Configurations/FeatureDefines.xcconfig:

WebKitTools:

2009-04-22  Sam Weinig  <sam@webkit.org>

        Rubber-stamped by Darin Adler.

        Fix for <rdar://problem/6816957>
        Turn off Geolocation by default

        * Scripts/build-webkit:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42752 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Rubber-stamped by David Hyatt.
eric@webkit.org [Wed, 22 Apr 2009 18:03:48 +0000 (18:03 +0000)]
    Rubber-stamped by David Hyatt.

        Change RenderSVGImage to dump as RenderSVGImage instead of RenderImage

        * rendering/RenderSVGImage.h:
        (WebCore::RenderSVGImage::renderName):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42751 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Fix a crash on Gmail when they remove a "before unload" event listener when
timothy@apple.com [Wed, 22 Apr 2009 15:48:38 +0000 (15:48 +0000)]
    Fix a crash on Gmail when they remove a "before unload" event listener when
        it was never added to the pending map in the first place.

        <rdar://problem/6814144>

        Reviewed by Darin Adler.

        * page/DOMWindow.cpp:
        (WebCore::allowsPendingBeforeUnloadListeners): Renamed from shouldAddPendingBeforeUnloadListener.
        (WebCore::DOMWindow::addEventListener): Call allowsPendingBeforeUnloadListeners now.
        (WebCore::DOMWindow::removeEventListener): Call allowsPendingBeforeUnloadListeners before
        removing the before unload listener.
        (WebCore::DOMWindow::clearAttributeEventListener): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42750 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>
ariya@webkit.org [Wed, 22 Apr 2009 13:39:50 +0000 (13:39 +0000)]
2009-04-22  Kenneth Rohde Christiansen  <kenneth.christiansen@openbossa.org>

        Reviewed by Ariya Hidayat.

        Made windowed plugins move/resize synchronized with the painting, so
        that windowed plugins are not moved before the rest of the parent
        frame during scrolling.

        * plugins/qt/PluginContainerQt.cpp:
        (PluginContainerQt::PluginContainerQt):
        (PluginContainerQt::requestGeometry):
        (PluginContainerQt::adjustGeometry):
        * plugins/qt/PluginContainerQt.h:
        * plugins/qt/PluginViewQt.cpp:
        (WebCore::PluginView::updatePluginWidget):
        (WebCore::PluginView::paint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42749 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Simon Hausmann <simon.hausmann@nokia.com>
hausmann@webkit.org [Wed, 22 Apr 2009 12:34:56 +0000 (12:34 +0000)]
2009-04-22  Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Ariya Hidayat.

        Fix various bugs in the X11 Qt plugin focus handling.

        Notify the focus controller on activation/deactivation of the plugin and
        set focus on the plugin's element when it receives the window system
        focus.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42748 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-22 Tamas Szirbucz <szirbucz.tamas@stud.u-szeged.hu>
ariya@webkit.org [Wed, 22 Apr 2009 12:08:55 +0000 (12:08 +0000)]
2009-04-22  Tamas Szirbucz  <szirbucz.tamas@stud.u-szeged.hu>

        Reviewed by Ariya Hidayat.

        https://bugs.webkit.org/show_bug.cgi?id=25023
        Delete reply in QNetworkReplyHandler::abort() to avoid leak.

        * platform/network/qt/QNetworkReplyHandler.cpp:
        (WebCore::QNetworkReplyHandler::abort):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42747 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-21 Jon Honeycutt <jhoneycutt@apple.com>
jhoneycutt@apple.com [Wed, 22 Apr 2009 08:52:18 +0000 (08:52 +0000)]
2009-04-21  Jon Honeycutt  <jhoneycutt@apple.com>

        Allow the UI delegate to control cursor-setting.

        Reviewed by Ada Chan.

        WebCore/:
        * page/Chrome.cpp:
        (WebCore::Chrome::setCursor):
        Pass the call to the client.

        * page/Chrome.h:

        * page/ChromeClient.h:
        (WebCore::ChromeClient::setCursor):
        Stubbed; returns false.

        * platform/Cursor.h:
        Define PlatformCursorHandle for each platform.

        * platform/win/WidgetWin.cpp:
        Reordered some includes. Added an include of Page.h.
        (WebCore::Widget::setCursor):
        Fixed a typo in the comment about ignoreNextSetCursor. Get the Page, and
        call its Chrome's setCursor() method with the passed cursor. Fall back
        to SetCursor() if the Page is not available.

        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginView::handleMouseEvent):
        Remove an unused variable. Cleaned up some whitespace.

        WebKit/win/:
        * Interfaces/IWebUIDelegatePrivate.idl:
        Added a function to the IWebUIDelegatePrivate5 interface.

        * WebCoreSupport/WebChromeClient.cpp:
        (WebChromeClient::setCursor):
        Try to call the UI delegate's webViewSetCursor() method. If it fails,
        fall back to calling SetCursor() with the passed cursor.

        * WebCoreSupport/WebChromeClient.h:

        * WebView.cpp:
        (WebViewWndProc):
        Call the Chrome's setCursor() method with the last set cursor.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42746 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoBuild fix
oliver@apple.com [Wed, 22 Apr 2009 07:22:19 +0000 (07:22 +0000)]
Build fix

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42745 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoBuild fix
oliver@apple.com [Wed, 22 Apr 2009 05:42:44 +0000 (05:42 +0000)]
Build fix

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42744 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoImprove String.replace performance slightly
oliver@apple.com [Wed, 22 Apr 2009 05:36:11 +0000 (05:36 +0000)]
Improve String.replace performance slightly

Reviewed by Maciej Stachowiak

Apply our vm reentry caching logic to String.replace with global
regexes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42743 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago - updated another result with action element information
mitz@apple.com [Wed, 22 Apr 2009 05:22:57 +0000 (05:22 +0000)]
    - updated another result with action element information

        * fast/encoding/mailto-always-utf-8-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42742 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebKit/mac:
mitz@apple.com [Wed, 22 Apr 2009 05:05:28 +0000 (05:05 +0000)]
WebKit/mac:

        Reviewed by Jon Honeycutt.

        - Mac part of fixing for <rdar://problem/6755137> Action dictionary for
          policy decision is missing keys when full-page zoom is used

        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::actionDictionary): Use absoluteLocation() instead
        of pageX() and pageY(), which are adjusted for zoom.

WebKit/win:

        Reviewed by Jon Honeycutt.

        - Windows part of fixing <rdar://problem/6755137> Action dictionary for
          policy decision is missing keys when full-page zoom is used

        * WebActionPropertyBag.cpp:
        (WebActionPropertyBag::Read): Use absoluteLocation() instead of
        clientX() and clientY(), which are adjusted for zoom.

LayoutTests:

        - test for <rdar://problem/6755137> Action dictionary for policy
          decision is missing keys when full-page zoom is used

        * fast/loader/policy-delegate-action-hit-test-zoomed-expected.txt: Added.
        * fast/loader/policy-delegate-action-hit-test-zoomed.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42741 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebKit/win:
mitz@apple.com [Wed, 22 Apr 2009 05:04:05 +0000 (05:04 +0000)]
WebKit/win:

        Reviewed by Jon Honeycutt.

        - implement DOMNode::nodeName()

        * DOMCoreClasses.cpp:
        (DOMNode::nodeName):

WebKitTools:

        Reviewed by Jon Honeycutt.

        - dump the node in the element dictionary in the action information
          passed to the policy delegate

        * DumpRenderTree/mac/PolicyDelegate.mm:
        (-[PolicyDelegate webView:decidePolicyForNavigationAction:request:frame:decisionListener:]):
        * DumpRenderTree/win/PolicyDelegate.cpp:
        (dumpPath):
        (PolicyDelegate::decidePolicyForNavigationAction):

LayoutTests:

        Reviewed by Jon Honeycutt.

        - updated result that shows the node in the element dictionary in the
          action information passed to the policy delegate

        * http/tests/security/feed-urls-from-remote-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42740 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-21 John Abd-El-Malek <jam@google.com>
dglazkov@chromium.org [Wed, 22 Apr 2009 04:20:05 +0000 (04:20 +0000)]
2009-04-21  John Abd-El-Malek  <jam@google.com>

        Reviewed by Dimitri Glazkov.

        Add a method to purge the Chromium plugin list cache.
        https://bugs.webkit.org/show_bug.cgi?id=25318

        * plugins/chromium/PluginDataChromium.cpp:
        (WebCore::resetChromiumPluginCache):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42739 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoFix <https://bugs.webkit.org/show_bug.cgi?id=25164> / <rdar://problem/6786319>.
mrowe@apple.com [Wed, 22 Apr 2009 04:03:52 +0000 (04:03 +0000)]
Fix <https://bugs.webkit.org/show_bug.cgi?id=25164> / <rdar://problem/6786319>.

Reviewed by Darin Adler.

We need to ensure that XMLTokenizer sets the current DocLoader before calling in to
any libxml2 methods that may trigger a load.  The presence of a DocLoader indicates
that the load was originated by WebCore's use of libxml2 and that we should enforce
the same-origin policy on it.  XMLTokenizer::initializeParserContext,
XMLTokenizer::doWrite and XMLTokenizer::doEnd were three methods that were not setting
the current DocLoader when they should have.

The XMLTokenizerScope class is introduced to simplify the pattern of saving, setting and
restoring the current DocLoader and libxml2 error handlers.  The DocLoader and error handlers
are saved and set when the scope is allocated, and restored to their previous values when
the scope is exited.

Test: http/tests/security/xss-DENIED-xml-external-entity.xhtml

* GNUmakefile.am:
* WebCore.pro:
* WebCore.scons:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* dom/XMLTokenizer.cpp:
* dom/XMLTokenizer.h:
* dom/XMLTokenizerLibxml2.cpp:
(WebCore::matchFunc):
(WebCore::shouldAllowExternalLoad):
(WebCore::openFunc):
(WebCore::XMLTokenizer::doWrite):
(WebCore::XMLTokenizer::initializeParserContext):
(WebCore::XMLTokenizer::doEnd):
(WebCore::xmlDocPtrForString):
* dom/XMLTokenizerScope.cpp:
(WebCore::XMLTokenizerScope::XMLTokenizerScope):
(WebCore::XMLTokenizerScope::~XMLTokenizerScope):
* dom/XMLTokenizerScope.h:
* xml/XSLStyleSheet.cpp:
(WebCore::XSLStyleSheet::parseString):
* xml/XSLTProcessor.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42738 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
weinig@apple.com [Wed, 22 Apr 2009 03:42:39 +0000 (03:42 +0000)]
WebCore:

2009-04-21  Sam Weinig  <sam@webkit.org>

        Reviewed by Dan Bernstein.

        Fix for <rdar://problem/5874009>
        Add port 3659 (apple-sasl / PasswordServer) to the list of blocked ports.

        * platform/network/ResourceHandle.cpp:
        (WebCore::portAllowed):

LayoutTests:

2009-04-21  Sam Weinig  <sam@webkit.org>

        Reviewed by Dan Bernstein.

        Update test and results for <rdar://problem/5874009>
        Add port 3659 (apple-sasl / PasswordServer) to the list of blocked ports.

        * platform/mac-tiger/security/block-test-expected.txt:
        * platform/mac/security/block-test-expected.txt:
        * security/block-test.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42737 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoCorrected a mistake in my last ChangeLog entry and updated some expected results.
justin.garcia@apple.com [Wed, 22 Apr 2009 02:18:06 +0000 (02:18 +0000)]
Corrected a mistake in my last ChangeLog entry and updated some expected results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42736 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
justin.garcia@apple.com [Wed, 22 Apr 2009 01:53:03 +0000 (01:53 +0000)]
WebCore:

2009-04-21  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Dan Bernstein.

        https://bugs.webkit.org/show_bug.cgi?id=25320
        Crash when changing block styles in a root that contains no visible content (but contains nodes with renderers)

        * editing/CompositeEditCommand.cpp:
        (WebCore::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): If the enclosing block is the
        root editable element and it contains no visible content, create a new block but don't try and move
        content into it, since there's nothing for moveParagraphs to move.

LayoutTests:

2009-04-21  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Dan Bernstein.

        https://bugs.webkit.org/show_bug.cgi?id=25320
        Crash when changing block styles in a root that contains no visible content (but contains nodes with renderers)

        * editing/execCommand/25320-expected.txt: Added.
        * editing/execCommand/25320.html: Added.
        * platform/mac/fast/forms/caret-rtl-expected.checksum: Replaced.
        * platform/mac/fast/forms/caret-rtl-expected.txt: Replaced.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42735 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-21 Geoffrey Garen <ggaren@apple.com>
ggaren@apple.com [Wed, 22 Apr 2009 00:18:33 +0000 (00:18 +0000)]
2009-04-21  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Cameron Zwarich and Oliver Hunt.

        Re-Fixed <rdar://problem/6406045> REGRESSION: Stack overflow on PowerPC on
        fast/workers/use-machine-stack.html (22531)

        SunSpider reports no change.

        Use a larger recursion limit on the main thread (because we can, and
        there's some evidence that it may improve compatibility), and a smaller
        recursion limit on secondary threads (because they tend to have smaller
        stacks).

        * interpreter/Interpreter.cpp:
        (JSC::Interpreter::execute):
        (JSC::Interpreter::prepareForRepeatCall):
        * interpreter/Interpreter.h:
        (JSC::): Ditto. I wrote the recursion test slightly funny, so that the
        common case remains a simple compare to constant.

        * runtime/ArrayPrototype.cpp:
        (JSC::arrayProtoFuncToString):
        (JSC::arrayProtoFuncToLocaleString):
        (JSC::arrayProtoFuncJoin): Conservatively, set the array recursion limits
        to the lower, secondary thread limit. We can do something fancier if
        compatibility moves us, but this seems sufficient for now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42734 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-21 Steve Falkenburg <sfalken@apple.com>
sfalken@apple.com [Wed, 22 Apr 2009 00:11:59 +0000 (00:11 +0000)]
2009-04-21  Steve Falkenburg  <sfalken@apple.com>

        Minor clean-ups to WebKitCreateInstance change.

        Reviewed by Jon Honeycutt.

        * WebKitCOMAPI.cpp:
        (classFactory):
        (WebKitCreateInstance):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42733 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
darin@apple.com [Tue, 21 Apr 2009 23:30:05 +0000 (23:30 +0000)]
WebCore:

2009-04-21  Darin Adler  <darin@apple.com>

        Reviewed by Dan Bernstein.

        Bug 25162: REGRESSION (r41176-r41242): Drag-selection above/below the line does not select to beginning/end
        https://bugs.webkit.org/show_bug.cgi?id=25162
        rdar://problem/6764354

        * page/Settings.cpp:
        (WebCore::Settings::Settings): Initialize editing behavior to Mac-style for PLATFORM(MAC).
        Later we may want to change this default to encompass other versions running on Mac.

        * page/Settings.h: Added EditingBehavior, setEditingBehavior, editingBehavior, and
        m_editingBehavior. Also moved m_maximumDecodedImageSize out of the middle of all
        the bit fields.

        * platform/graphics/IntPoint.h: Added a constructor to turn an IntSize into an IntPoint.
        I'm not sure the distinction here is serving us well at the moment. When converting from
        global to local coordinates you want to do IntPoint - IntPoint and have the result be
        another IntPoint, not an IntSize. And so on.

        * rendering/RenderBlock.cpp:
        (WebCore::positionForPointRespectingEditingBoundaries): Changed to use pointers more and
        separate coordinates less.
        (WebCore::RenderBlock::positionForPointWithInlineChildren): Turned into a member function.
        Added separate cases to support the Mac and Windows behavior.
        (WebCore::RenderBlock::positionForPoint): Updated for the change above. Also moved the
        computation of pointInContents closer to the place it's used.

        * rendering/RenderBlock.h: Updated for the new private member function.

LayoutTests:

2009-04-21  Darin Adler  <darin@apple.com>

        Reviewed by Dan Bernstein.

        Bug 25162: REGRESSION (r41176-r41242): Drag-selection above/below the line does not select to beginning/end
        https://bugs.webkit.org/show_bug.cgi?id=25162
        rdar://problem/6764354

        * editing/selection/resources/click-in-margins-inside-editable-div.js:
        Updated to expect different results on the Mac.
        * editing/selection/resources/click-in-padding-with-multiple-line-boxes.js: Ditto.

        * fast/css-generated-content/positioned-background-hit-test-crash.html: Clear the selection at
        the end of the test so the result won't include a caret.

        * platform/mac/editing/selection/click-in-margins-inside-editable-div-expected.txt:
        * platform/mac/editing/selection/click-in-padding-with-multiple-line-boxes-expected.txt:
        Copied from LayoutTests/editing/selection. Then changed to expect the different Mac results.

        * platform/mac/editing/selection/select-from-textfield-outwards-expected.checksum:
        * platform/mac/editing/selection/select-from-textfield-outwards-expected.png:
        * platform/mac/editing/selection/select-from-textfield-outwards-expected.txt:
        Changed to expect the different Mac results.

        * platform/mac/fast/css-generated-content/positioned-background-hit-test-crash-expected.txt:
        Changed to expect no caret.

        * platform/win/editing/selection/click-in-margins-inside-editable-div-expected.txt:
        * platform/win/editing/selection/click-in-padding-with-multiple-line-boxes-expected.txt:
        Copied from editing/selection.

        * platform/win/editing/selection/select-from-textfield-outwards-expected.checksum:
        * platform/win/editing/selection/select-from-textfield-outwards-expected.png:
        * platform/win/editing/selection/select-from-textfield-outwards-expected.txt:
        Copied from platform/mac/editing/selection.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42732 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Update StringsNotToBeLocalized.txt for recent changes
aroben@apple.com [Tue, 21 Apr 2009 21:58:42 +0000 (21:58 +0000)]
    Update StringsNotToBeLocalized.txt for recent changes

        Rubber-stamped by John Sullivan.

        * StringsNotToBeLocalized.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42731 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-21 Anders Carlsson <andersca@apple.com>
andersca@apple.com [Tue, 21 Apr 2009 21:35:00 +0000 (21:35 +0000)]
2009-04-21  Anders Carlsson  <andersca@apple.com>

        Reviewed by Darin Adler and Kevin Decker.

        WebKit side of <rdar://problem/6781642>.

        When we call resize with an actual changed size, block until the plug-in host is done.

        * Plugins/Hosted/NetscapePluginInstanceProxy.h:
        * Plugins/Hosted/NetscapePluginInstanceProxy.mm:
        (WebKit::NetscapePluginInstanceProxy::resize):
        * Plugins/Hosted/WebHostedNetscapePluginView.h:
        * Plugins/Hosted/WebHostedNetscapePluginView.mm:
        (-[WebHostedNetscapePluginView updateAndSetWindow]):
        * Plugins/Hosted/WebKitPluginHost.defs:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42730 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Windows build fix after r42726
aroben@apple.com [Tue, 21 Apr 2009 21:10:13 +0000 (21:10 +0000)]
    Windows build fix after r42726

        * DumpRenderTree/win/DumpRenderTree.cpp: Touched to force a rebuild.
        * DumpRenderTree/win/DumpRenderTree.vcproj: Delete any pre-r42726
        manifest files during the pre-build event.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42729 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoAnother attempt to fix the Windows and QT builds.
timothy@apple.com [Tue, 21 Apr 2009 20:52:22 +0000 (20:52 +0000)]
Another attempt to fix the Windows and QT builds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42728 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoAttempt to fix the Windows and QT builds.
timothy@apple.com [Tue, 21 Apr 2009 20:48:01 +0000 (20:48 +0000)]
Attempt to fix the Windows and QT builds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebKit/win:
sfalken@apple.com [Tue, 21 Apr 2009 20:42:20 +0000 (20:42 +0000)]
WebKit/win:

2009-04-21  Steve Falkenburg  <sfalken@apple.com>

        Add registry-free CoCreateInstance API for WebKit clients: WebKitCreateInstance.

        Manifest-based registry-free COM continues to work for clients that have embedded
        manifests with isolated COM information.

        Reviewed by Adam Roben.

        * WebKit.vcproj/WebKit.def: Added WebKitCreateInstance.
        * WebKit.vcproj/WebKit.vcproj: Link against rpcrt4.dll for UUID hashing.
        * WebKit.vcproj/WebKit_debug.def: Added WebKitCreateInstance.
        * WebKitCOMAPI.cpp: Added.
        (CLSIDHash::hash):
        (CLSIDHash::equal):
        (CLSIDHashTraits::constructDeletedValue):
        (CLSIDHashTraits::isDeletedValue):
        (classFactory):
        (WebKitCreateInstance): New API.
        * WebKitCOMAPI.h: Added.

WebKitTools:

2009-04-21  Steve Falkenburg  <sfalken@apple.com>

        Use new WebKitCreateInstance API instead of CoCreateInstance
        or COMPtr's createInstance to instantiate WebKit COM objects.

        Reviewed by Adam Roben.

        * DumpRenderTree/win/DumpRenderTree.cpp:
        (initialize):
        (runTest):
        (createWebViewAndOffscreenWindow):
        (main):
        * DumpRenderTree/win/DumpRenderTree.vcproj:
        * DumpRenderTree/win/GCControllerWin.cpp:
        (GCController::collect):
        (GCController::collectOnAlternateThread):
        (GCController::getJSObjectCount):
        * DumpRenderTree/win/LayoutTestControllerWin.cpp:
        (LayoutTestController::keepWebHistory):
        (LayoutTestController::webHistoryItemCount):
        (LayoutTestController::setIconDatabaseEnabled):
        * DumpRenderTree/win/WorkQueueItemWin.cpp:
        (LoadItem::invoke):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42726 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoChange pending unload and beforeunload counts to store the listeners in a single...
timothy@apple.com [Tue, 21 Apr 2009 20:21:50 +0000 (20:21 +0000)]
Change pending unload and beforeunload counts to store the listeners in a single data structure that can be quickly iterated for dispatch.

WebCore:

2009-04-17  Timothy Hatcher  <timothy@apple.com>

        Change pending unload and beforeunload counts to store the listeners
        in a single data structure that can be quickly iterated for dispatch.

        <rdar://problem/6383352&6383379&6383940>

        Reviewed by Darin Adler.

        * WebCore.base.exp: Change what symbols are exported.
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::stopLoading): Remove count updating code
        and move an if inside a block already checking m_frame->document().
        * page/Chrome.cpp: Remove disableSuddenTermination/enableSuddenTermination.
        * page/Chrome.h: Ditto.
        * page/ChromeClient.h: Ditto.
        * page/DOMWindow.cpp:
        (WebCore::pendingUnloadEventListenerMap): Create a static map.
        (WebCore::pendingBeforeUnloadEventListenerMap): Ditto.
        (WebCore::addPendingEventListener): Add the event listener to the map.
        (WebCore::removePendingEventListener): Remove the event listener from the map.
        (WebCore::removePendingEventListeners): Remove all listeners for the window.
        (WebCore::dispatchWindowEventToListeners): Dispatch the event to the given listeners.
        (WebCore::DOMWindow::dispatchAllPendingBeforeUnloadEvents): Send the beforeunload event.
        (WebCore::DOMWindow::pendingUnloadEventListeners): Return a count of pending listeners.
        (WebCore::DOMWindow::dispatchAllPendingUnloadEvents): Send the unload event.
        (WebCore::DOMWindow::~DOMWindow): Remove the window from both event listener maps.
        (WebCore::DOMWindow::addEventListener): Call addPendingEventListener when needed.
        (WebCore::DOMWindow::removeEventListener): Call removePendingEventListener when needed.
        (WebCore::DOMWindow::removeAllEventListeners): Call removePendingEventListeners.
        (WebCore::DOMWindow::removeInlineEventListenerForType): Call removePendingEventListener when needed.
        * page/DOMWindow.h:
        (WebCore::DOMWindow::frame): Changed to be const.
        * page/EventHandler.cpp:
        (WebCore::EventHandler::EventHandler): Remove count tracking code.
        * page/EventHandler.h: Ditto.
        * page/Page.cpp:
        (WebCore::Page::Page): Ditto.
        * page/Page.h: Ditto.

WebKit/mac:

2009-04-17  Timothy Hatcher  <timothy@apple.com>

        Change how sudden termination works with WebView teardown.

        <rdar://problem/6383352&6383379&6383940>

        Reviewed by Darin Adler.

        * WebCoreSupport/WebChromeClient.h: Remove disableSuddenTermination/enableSuddenTermination.
        * WebCoreSupport/WebChromeClient.mm: Ditto.
        * WebView/WebFrame.mm:
        (-[WebFrame _pendingFrameUnloadEventCount]): Ask the DOMWindow.
        * WebView/WebView.mm:
        (+[WebView canCloseAllWebViews]): Call DOMWindow::dispatchAllPendingBeforeUnloadEvents.
        (+[WebView closeAllWebViews]): Call DOMWindow::dispatchAllPendingUnloadEvents and
        call close on all the WebViews.
        (-[WebView _closeWithFastTeardown]): Remove code for unload event dispatch.
        (-[WebView _close]): Correct a comment.
        (+[WebView _applicationWillTerminate]): Call closeAllWebViews.
        * WebView/WebViewPrivate.h: Add canCloseAllWebViews and closeAllWebViews.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42725 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoJavaScriptCore:
ggaren@apple.com [Tue, 21 Apr 2009 19:58:00 +0000 (19:58 +0000)]
JavaScriptCore:

2009-04-21  Geoffrey Garen  <ggaren@apple.com>

        Rubber-stamped by Adam Roben.

        Disabled one more Mozilla JS test because it fails intermittently on Windows.
        (See https://bugs.webkit.org/show_bug.cgi?id=25160.)

        * tests/mozilla/expected.html:

WebKitTools:

2009-04-21  Geoffrey Garen  <ggaren@apple.com>

        Rubber-stamped by Adam Roben.

        Disabled one more Mozilla JS test because it fails intermittently on Windows.
        (See https://bugs.webkit.org/show_bug.cgi?id=25160.)

        * Scripts/run-javascriptcore-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42724 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-21 Darin Adler <darin@apple.com>
darin@apple.com [Tue, 21 Apr 2009 19:38:01 +0000 (19:38 +0000)]
2009-04-21  Darin Adler  <darin@apple.com>

        Reviewed by Adele Peterson.

        Fix some tests that are depending on Mac vs. Windows selection quirks accidentally.
        Added failing tests to the skipped list for SnowLeopard.

        * editing/resources/select-all-iframe-src.html: Moved.
        * editing/selection/resources/select-all-iframe-src.html: Moved from
        editing/resources/select-all-iframe-src.html, since this is used by a test in the
        selection directory. Changed mouse coordinates to be farther to the right so we
        don't depend on the rule for what's selected when you click below text on the last line.
        * editing/selection/select-all-iframe.html: Updated for new location of file above.

        * fast/events/standalone-image-drag-to-editable.html: Changed mouse coordinates to be
        farther to the right so we don't depend on the rule for what's selected when you click
        below text on the last line.

        * platform/mac-snowleopard/Skipped: Added some tests to the Skipped list.

        * svg/custom/pointer-events-path.svg: Clear the selection at the end of the test so it's
        not in the test results. Also removed unneeded use of the layout test controller.

        * platform/mac/editing/selection/select-all-iframe-expected.checksum: Generated new results.
        * platform/mac/editing/selection/select-all-iframe-expected.png: Ditto.
        * platform/mac/editing/selection/select-all-iframe-expected.txt: Ditto.
        * platform/mac/fast/events/standalone-image-drag-to-editable-expected.checksum: Ditto.
        * platform/mac/fast/events/standalone-image-drag-to-editable-expected.png: Ditto.
        * platform/mac/fast/events/standalone-image-drag-to-editable-expected.txt: Ditto.
        * platform/mac/svg/custom/pointer-events-path-expected.checksum: Ditto.
        * platform/mac/svg/custom/pointer-events-path-expected.png: Ditto.
        * platform/mac/svg/custom/pointer-events-path-expected.txt: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42723 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoWebCore:
justin.garcia@apple.com [Tue, 21 Apr 2009 19:20:27 +0000 (19:20 +0000)]
WebCore:

2009-04-21  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=19644
        Text copied with Select All pastes with a indent but shouldn't
        <rdar://problem/6102483>
        Pasting the content of an HTML message in Mail causes addition nested <div> elements to be added

        To improve selectall/copy/paste fidelity of certain pages, we began wrapping copied content with
        a div that held properties and attributes from the fully selected body.  To fix the above issues,
        only do this if if the body has certain properties or attributes.  We'll begin adding to this list as
        necessary.  For now it's just background colors and images.  Tested copy/paste of nytimes, wired,
        arstechnica, and several others.

        Massive nesting can still happen, it will just be much less common.

        * editing/markup.cpp:
        (WebCore::createMarkup):

LayoutTests:

2009-04-21  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=19644
        Text copied with Select All pastes with a indent but shouldn't
        <rdar://problem/6102483>
        Pasting the content of an HTML message in Mail causes addition nested <div> elements to be added

        * editing/pasteboard/19644-1-expected.txt: Added.
        * editing/pasteboard/19644-1.html: Added.
        * editing/pasteboard/19644-2-expected.txt: Added.
        * editing/pasteboard/19644-2.html: Added.
        * platform/mac/editing/inserting/insert-3907422-fix-expected.txt:
        * platform/mac/editing/pasteboard/paste-text-015-expected.checksum:
        * platform/mac/editing/pasteboard/paste-text-015-expected.png:
        * platform/mac/editing/pasteboard/paste-text-015-expected.txt:
        * platform/mac/editing/style/font-family-with-space-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42722 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2009-04-16 Peter Kasting <pkasting@google.com>
pkasting@chromium.org [Tue, 21 Apr 2009 18:34:35 +0000 (18:34 +0000)]
2009-04-16  Peter Kasting  <pkasting@google.com>

        Reviewed by David Hyatt.

        https://bugs.webkit.org/show_bug.cgi?id=25250
        Allow platforms to snap the scroll thumb back to the drag origin
        during a drag.  Implement functions for Safari/Win and Chromium/Win
        to do this snapping at distances approximating the native ones.

        * platform/Scrollbar.cpp:
        (WebCore::Scrollbar::Scrollbar):
        (WebCore::Scrollbar::setValue):
        (WebCore::Scrollbar::scroll):
        (WebCore::Scrollbar::moveThumb):
        (WebCore::Scrollbar::setCurrentPos):
        (WebCore::Scrollbar::mouseMoved):
        (WebCore::Scrollbar::mouseDown):
        * platform/Scrollbar.h:
        * platform/ScrollbarTheme.h:
        (WebCore::ScrollbarTheme::shouldSnapBackToDragOrigin):
        * platform/chromium/ScrollbarThemeChromium.h:
        * platform/chromium/ScrollbarThemeChromiumLinux.cpp:
        (WebCore::ScrollbarThemeChromium::shouldSnapBackToDragOrigin):
        * platform/chromium/ScrollbarThemeChromiumWin.cpp:
        (WebCore::ScrollbarThemeChromium::shouldSnapBackToDragOrigin):
        * platform/win/ScrollbarThemeWin.cpp:
        (WebCore::ScrollbarThemeWin::shouldCenterOnThumb):
        (WebCore::ScrollbarThemeWin::shouldSnapBackToDragOrigin):
        * platform/win/ScrollbarThemeWin.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42721 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Windows build fix
aroben@apple.com [Tue, 21 Apr 2009 18:15:28 +0000 (18:15 +0000)]
    Windows build fix

        * bindings/js/JSLazyEventListener.cpp:
        (WebCore::eventParameterName): Remove unreachable code when SVG is
        enabled.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42720 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Fix running run-safari and run-webkit-tests from the paths recommended on...
aroben@apple.com [Tue, 21 Apr 2009 18:09:24 +0000 (18:09 +0000)]
    Fix running run-safari and run-webkit-tests from the paths recommended on webkit.org

        Reviewed by Steve Falkenburg and Sam Weinig.

        * Scripts/run-webkit-nightly.cmd: Moved quotes from variable
        definitions to variable uses. Otherwise cmd.exe can get confused about
        the quoting.
        * Scripts/run-webkit-tests: Changed to use an absolute path to
        execAppWithEnv so that run-webkit-tests will work when run from
        outside the source tree.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42719 268f45cc-cd09-0410-ab3c-d52691b4dbfc