WebKit-https.git
13 years ago Reviewed by Darin.
adele [Wed, 11 Jul 2007 19:09:39 +0000 (19:09 +0000)]
    Reviewed by Darin.

        Fix for <rdar://problem/5230188> REGRESSION: Active/inactive marked text is not distinguished in textfields and textareas

        We were previously trying to leave room for 1px of space between the text and the underline.
        This change removes that requirement so that if we don't have room for the extra space, we still draw the thick underline
        touching the text.

        * rendering/InlineTextBox.cpp: (WebCore::InlineTextBox::paintMarkedTextUnderline):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin and Geoff.
andersca [Wed, 11 Jul 2007 18:47:33 +0000 (18:47 +0000)]
    Reviewed by Darin and Geoff.

        <rdar://problem/5313508>
        REGRESSION (Leopard): http/tests/xmlhttprequest/basic-auth.html hangs

        Leopard Foundation now ignores credentials passed in as part of the URL so we need to pass them ourselves.

        For the asynchronous case, we simply keep track of the URL and if it has a user name and password, we
        pass those if we're asked to authenticate.

        For the synchronous case, implement a new connection delegate and run it in a separate mode so
        no other sources will fire.

        * platform/network/mac/ResourceHandleMac.mm:
        (WebCore::ResourceHandle::loadResourceSynchronously):
        (-[WebCoreResourceHandleAsDelegate dealloc]):
        (-[WebCoreResourceHandleAsDelegate connection:willSendRequest:redirectResponse:]):
        (-[WebCoreResourceHandleAsDelegate connection:didReceiveAuthenticationChallenge:]):
        (-[WebCoreSynchronousLoader _isDone]):
        (-[WebCoreSynchronousLoader dealloc]):
        (-[WebCoreSynchronousLoader connection:willSendRequest:redirectResponse:]):
        (-[WebCoreSynchronousLoader connection:didReceiveAuthenticationChallenge:]):
        (-[WebCoreSynchronousLoader connection:didReceiveResponse:]):
        (-[WebCoreSynchronousLoader connection:didReceiveData:]):
        (-[WebCoreSynchronousLoader connectionDidFinishLoading:]):
        (-[WebCoreSynchronousLoader connection:didFailWithError:]):
        (-[WebCoreSynchronousLoader _data]):
        (-[WebCoreSynchronousLoader _response]):
        (-[WebCoreSynchronousLoader _error]):
        (+[WebCoreSynchronousLoader loadRequest:returningResponse:error:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff.
andersca [Wed, 11 Jul 2007 18:21:09 +0000 (18:21 +0000)]
    Reviewed by Geoff.

        Create the lock file in a writable directory.

        * http/conf/apache2-httpd.conf:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam Weinig.
darin [Wed, 11 Jul 2007 18:11:32 +0000 (18:11 +0000)]
    Reviewed by Sam Weinig.

        - fix <rdar://problem/5315265> REGRESSION: Leopard launch time -
          there are about 190 stats of icudt36b files during launch

        * platform/TextEncodingRegistry.cpp: (WebCore::atomicCanonicalTextEncodingName):
        Add a special case for the empty string so we don't load the extended text
        encodings just so we can return 0 for that case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
andersca [Wed, 11 Jul 2007 17:12:22 +0000 (17:12 +0000)]
    Reviewed by Darin.

        Remove a bunch of unneeded accessors from ResourceHandle.

        * loader/icon/IconLoader.cpp:
        (WebCore::IconLoader::didReceiveResponse):
        (WebCore::IconLoader::didFail):
        (WebCore::IconLoader::didFinishLoading):
        * platform/network/ResourceHandle.cpp:
        * platform/network/ResourceHandle.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoReviewed by Mark.
zimmermann [Wed, 11 Jul 2007 12:56:05 +0000 (12:56 +0000)]
Reviewed by Mark.

Forwardport the hash table fix from CodeGeneratorJS.pm to create_hash_table.
Reran run-jsc-tests, couldn't find any regressions. Suggested by Darin.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
justing [Wed, 11 Jul 2007 10:21:09 +0000 (10:21 +0000)]
LayoutTests:

        Reviewed by Darin.

        <rdar://problem/5234383> REGRESSION: Sentence breaker method returning entire document length

        * editing/selection/5234383-1-expected.checksum: Added.
        * editing/selection/5234383-1-expected.png: Added.
        * editing/selection/5234383-1-expected.txt: Added.
        * editing/selection/5234383-1.html: Added.
        * editing/selection/5234383-2-expected.checksum: Added.
        * editing/selection/5234383-2-expected.png: Added.
        * editing/selection/5234383-2-expected.txt: Added.
        * editing/selection/5234383-2.html: Added.

WebCore:

        Reviewed by Darin.

        <rdar://problem/5234383> REGRESSION: Sentence breaker method returning entire document length

        The functions for moving the caret to the next/previous sentence boundary
        were in many cases moving to the beginning/end of the document.  The functions that
        should move the caret to an equivalent position in the next/previous sentence behaved
        like the aforementioned functions ought to.  Put the code from the second two functions
        in the first and made sure that the second two are properly FIXMEd.  They're rarely
        used (filed 5323691).

        * editing/SelectionController.cpp:
        (WebCore::SelectionController::modify): Added an option for sentenceBoundary granularity,
        so that we can test the functions that were broken from JS.
        * editing/visible_units.cpp:
        (WebCore::startOfSentence): Removed the FIXME about 5234383.
        (WebCore::endSentenceBoundary): Make this function do what nextSentencePositionBoundary
        current does.
        (WebCore::endOfSentence): Removed the FIXME about 5234383.
        (WebCore::previousSentencePosition): Removed an incorrect FIXME.
        (WebCore::nextSentencePositionBoundary): Tweaked FIXME.  This function is
        identical, not close to endSentenceBoundary.
        (WebCore::nextSentencePosition): Removed an incorrect FIXME.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24187 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by John.
antti [Wed, 11 Jul 2007 08:10:16 +0000 (08:10 +0000)]
    Reviewed by John.

        Fix <rdar://problem/4570550>
        Hang in layout/layoutBlock/layoutBlockChildren preparing to print certain Mail messages

        When printing from Mail, WebHTMLView is a subview of the view that is actually printed and does not
        receive calls that would set it to printing mode. Method adjustPageHeightNew is called repeatedly (for each page) during
        printing and it enables printing mode temporarily for each call. This triggers two full style recalcs and layouts
        each time making printing at least O(n^2).

        Instead of enabling printing mode and resetting it back immediatly do the resetting asynchronously, after
        all adjustPageHeightNew calls are done. Normal Safari printing is not affected as adjustPageHeightNew is only
        called in case WebHTMLView is embedded in the view that is being printed.

        No automatic test possible, requires printing and non-Safari client.

        * WebView/WebHTMLView.mm:
        (-[WebHTMLView adjustPageHeightNew:top:bottom:limit:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24186 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
rwlbuis [Wed, 11 Jul 2007 07:40:18 +0000 (07:40 +0000)]
    Reviewed by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=14453
        REGRESSION: www.nzherald.co.nz almost all the formating is gone

        Be less strict for the hex color definition.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24185 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Oliver.
rwlbuis [Wed, 11 Jul 2007 07:34:58 +0000 (07:34 +0000)]
    Reviewed by Oliver.

        Forgot to commit this as part of commit r24146. Now we are sure
        createElement creates elements in the xhtml namespace only.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24184 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej Stachowiak.
ggaren [Wed, 11 Jul 2007 06:20:28 +0000 (06:20 +0000)]
    Reviewed by Maciej Stachowiak.

        - Fixed <rdar://problem/5049509> REGRESSION(10.4.9-9A377a): REAP Suite
        installer shows empty modal alert window (hangs) if user cancels during
        "installing shared components" phase

        Added Adobe installers to the family of apps that need a data load
        loading quirk.

        Added a linked-on-or-after check because this code is no longer
        Tiger-only.

        I tested this code on Tiger and Leopard.

        * Misc/WebKitVersionChecks.h:
        * WebKit.xcodeproj/project.pbxproj:
        * WebView/WebDocumentLoaderMac.mm:
        (needsDataLoadWorkaround):
        (WebDocumentLoaderMac::setDataSource):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24183 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
weinig [Wed, 11 Jul 2007 05:10:24 +0000 (05:10 +0000)]
LayoutTests:

        Reviewed by Oliver.

        * fast/dom/Window/get-set-properties-expected.txt: Added.
        * fast/dom/Window/get-set-properties.html: Added.

WebCore:

        Reviewed by Oliver.

        Move more attributes from the pure JS bindings in KJS::Window
        to the implementation in DOMWindow and autogenerate the new bindings

        Test: fast/dom/Window/get-set-properties.html

        * bindings/js/JSDOMWindowCustom.cpp:
        (WebCore::JSDOMWindow::customGetOwnPropertySlot): Change to use propHashTable in
        JSDOMWindow instead of KJS::Window
        * bindings/js/kjs_window.cpp:
        (KJS::Window::getValueProperty):
        (KJS::Window::put):
        * bindings/js/kjs_window.h:
        (KJS::Window::):
        * bindings/scripts/CodeGeneratorJS.pm: Cleanup getValueProperty and put generation
        to reduce code duplication.
        * page/DOMWindow.cpp:
        (WebCore::DOMWindow::closed):
        (WebCore::DOMWindow::length):
        (WebCore::DOMWindow::name):
        (WebCore::DOMWindow::setName):
        (WebCore::DOMWindow::status):
        (WebCore::DOMWindow::setStatus):
        (WebCore::DOMWindow::defaultStatus):
        (WebCore::DOMWindow::setDefaultStatus):
        * page/DOMWindow.h:
        (WebCore::DOMWindow::defaultstatus):
        (WebCore::DOMWindow::setDefaultstatus):
        * page/DOMWindow.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
weinig [Wed, 11 Jul 2007 03:46:42 +0000 (03:46 +0000)]
LayoutTests:

        Reviewed by Geoff.

        Tests for <rdar://problem/5326791>

        * http/tests/security/cross-frame-access-expected.txt:
        * http/tests/security/cross-frame-access-put-expected.txt: Added.
        * http/tests/security/cross-frame-access-put.html: Added.
        * http/tests/security/resources/cross-frame-iframe-for-put-test.html: Added.

WebCore:

        Reviewed by Geoff.

        Fix for <rdar://problem/5326791> XSS vulnerability: ability to set window.defaultStatus,
        window.defaultstatus and window.status cross domains

        Test: http/tests/security/cross-frame-access-put.html

        * bindings/js/kjs_window.cpp:
        (KJS::Window::put): Adds isSafeScript check for defaultStatus,
        defaultstatus and status cases.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24181 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * Scripts/run-webkit-tests: Add an ignore item for a leak in Tiger's
darin [Wed, 11 Jul 2007 03:11:13 +0000 (03:11 +0000)]
    * Scripts/run-webkit-tests: Add an ignore item for a leak in Tiger's
        Foundation's multipart/mixed-replace support -- I already verified that
        the bug has been fixed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24180 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago rubber stamped by Maciej.
aliceli1 [Wed, 11 Jul 2007 02:13:05 +0000 (02:13 +0000)]
    rubber stamped by Maciej.

        fixed <rdar://5137972> editing/selection/editable-links.html fails on Windows

        * DumpRenderTree/DumpRenderTree.vcproj/DumpRenderTree/DumpRenderTree.cpp:
        (initializePreferences):
        set editable link behavior to match Mac DRT.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24179 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix build for BuildBot.
weinig [Wed, 11 Jul 2007 01:42:17 +0000 (01:42 +0000)]
    Fix build for BuildBot.

        * page/DOMWindow.cpp:
        (WebCore::DOMWindow::outerHeight):
        (WebCore::DOMWindow::outerWidth):
        (WebCore::DOMWindow::screenX):
        (WebCore::DOMWindow::screenY):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24178 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam
beidson [Wed, 11 Jul 2007 01:38:30 +0000 (01:38 +0000)]
    Reviewed by Sam

        <rdar://problem/5326653> - REGRESSION: [WebBackForwardList forwardListCount] might return incorrect result

        No layout test because this is an API issue that can't be tested in DRT or the browser

        * history/BackForwardList.cpp:
        (WebCore::BackForwardList::forwardListCount): If the index is NoCurrentItemIndex, return 0

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24177 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin.
weinig [Wed, 11 Jul 2007 00:28:45 +0000 (00:28 +0000)]
    Reviewed by Darin.

        Move more readonly attributes from pure JS bindings in KJS::Window
        to the implementation in DOMWindow and autogenerate the new bindings.

        * bindings/js/kjs_window.cpp:
        (KJS::Window::getValueProperty):  Sort the hashmap comment into categories
        to make further migration easier.
        * bindings/js/kjs_window.h: Sort the enum into categories to make further
        migration easier.
        (KJS::Window::):
        * page/DOMWindow.cpp: Add new methods.
        (WebCore::DOMWindow::focus):
        (WebCore::DOMWindow::blur):
        (WebCore::DOMWindow::offscreenBuffering):
        (WebCore::DOMWindow::outerHeight):
        (WebCore::DOMWindow::outerWidth):
        (WebCore::DOMWindow::innerHeight):
        (WebCore::DOMWindow::innerWidth):
        (WebCore::DOMWindow::screenX):
        (WebCore::DOMWindow::screenY):
        (WebCore::DOMWindow::scrollX):
        (WebCore::DOMWindow::scrollY):
        (WebCore::DOMWindow::getComputedStyle):
        (WebCore::DOMWindow::getMatchedCSSRules): Add null Document assertion.
        (WebCore::DOMWindow::devicePixelRatio): Add null Frame and Page checks.
        * page/DOMWindow.h:
        (WebCore::DOMWindow::screenLeft):
        (WebCore::DOMWindow::screenTop):
        (WebCore::DOMWindow::pageXOffset):
        (WebCore::DOMWindow::pageYOffset):
        * page/DOMWindow.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24176 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
oliver [Wed, 11 Jul 2007 00:23:21 +0000 (00:23 +0000)]
    Reviewed by Maciej.

        Fix for <rdar://problem/5326417> Editor::unmarkText needs to be crossplatform

        Missed this in earlier patch to make text marking platform independent.

        * editing/Editor.cpp:
        (WebCore::Editor::unmarkText):
           Add platform independent impl
        * editing/mac/EditorMac.mm:
           Remove the old mac path.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24175 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
ggaren [Tue, 10 Jul 2007 23:52:42 +0000 (23:52 +0000)]
WebCore:

        Reviewed by Tim Hatcher.

        Added a Page counter. This helped me while investigating
        <rdar://problem/5314463> 40MB of WebKit memory still allocated after
        closing large email in Mail

        * WebCore.xcodeproj/project.pbxproj:
        * page/Page.cpp:
        (WebCore::):
        (WebCore::PageCounter::~PageCounter):
        (WebCore::Page::Page):
        (WebCore::Page::~Page):

WebKitTools:

        Reviewed by Tim Hatcher.

        Ignore Page.o in debug builds because it now has a static PageCounter
        object.

        * Scripts/check-for-global-initializers:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24174 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff.
thatcher [Tue, 10 Jul 2007 23:44:17 +0000 (23:44 +0000)]
    Reviewed by Geoff.

        <rdar://problem/5326126> InspectorResource holds onto more data than it needs

        Reduces the fixed size of InspectorResource from 630 bytes to 224 bytes. Now
        selectively pick out parts of the ResourceRequest and ResourceResponse objects
        for the info needed and store that directly in InspectorResource.

        The ResourceRequest and ResourceResponse objects hold a reference to the original
        NSURL object, so almost double the data was being kept around. So the struct size
        reduction is just the tip of the ice burg on what this really saves.

        In a test of 100,000 XMLHTTPRequests using a 85 byte long data URL, I see ~21 MB less
        RSIZE compared to Safari running without this change.

        * page/InspectorController.cpp:
        (WebCore::InspectorResource::InspectorResource):
        (WebCore::InspectorResource::type):
        (WebCore::addSourceToFrame):
        (WebCore::scriptObjectForRequest):
        (WebCore::scriptObjectForResponse):
        (WebCore::InspectorController::addScriptResource):
        (WebCore::InspectorController::addAndUpdateScriptResource):
        (WebCore::updateResourceRequest):
        (WebCore::updateResourceResponse):
        (WebCore::InspectorController::updateScriptResourceRequest):
        (WebCore::InspectorController::updateScriptResourceResponse):
        (WebCore::InspectorController::didLoadResourceFromMemoryCache):
        (WebCore::InspectorController::identifierForInitialRequest):
        (WebCore::InspectorController::willSendRequest):
        (WebCore::InspectorController::didReceiveResponse):
        (WebCore::InspectorController::didFailLoading):
        * page/InspectorController.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24173 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Brady.
darin [Tue, 10 Jul 2007 22:37:44 +0000 (22:37 +0000)]
    Reviewed by Brady.

        - fix crash introduced by my patch for 5318732, only seen in release builds

        * page/Page.cpp: (WebCore::Page::~Page): Added a missing null check.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Andrea Cavalli to fix sunken state of text fields.
staikos [Tue, 10 Jul 2007 22:30:42 +0000 (22:30 +0000)]
Patch from Andrea Cavalli to fix sunken state of text fields.
Bug #14576

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24170 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Qing Zhao to draw focus rects on non-QStyle rendered nodes. Needs a
staikos [Tue, 10 Jul 2007 22:13:33 +0000 (22:13 +0000)]
Patch from Qing Zhao to draw focus rects on non-QStyle rendered nodes.  Needs a
bit more work once we figure out how to use QPainterPathStroker in a more
sophisticated manner.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24168 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-10 Ada Chan <adachan@apple.com>
adachan [Tue, 10 Jul 2007 21:53:19 +0000 (21:53 +0000)]
2007-07-10  Ada Chan  <adachan@apple.com>

        Fixed build.

        * WebView.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoReviewed by Geoff.
zimmermann [Tue, 10 Jul 2007 21:46:52 +0000 (21:46 +0000)]
Reviewed by Geoff.

Fix 'unsigned int' overflow in CodeGeneratorJS.pm.
This fixes hashing of "SVG_FEBLEND_MODE_NORMAL". Its hashvalue exceeded 2^32, which
is no problem in perl but in the cpp code. Emulate C/C++ behaviour by using "% 2^32" in
the perl code. Testcase available in feature-branch soon (js-svg-constructors.svg).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24165 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej Stachowiak.
ggaren [Tue, 10 Jul 2007 21:45:04 +0000 (21:45 +0000)]
    Reviewed by Maciej Stachowiak.

        Disabled Arena recycling in RenderArena. PLT shows no regression
        (possibly a small speedup), and this can reduce allocated memory in
        WebKit clients by about 120K (4K per Arena * 30 Arenas).

        I happened upon this optimization while investigating
        <rdar://problem/5314463> WebKit hangs on to 40MB RAM after closing
        large email in Mail (appears to be render tree arena allocator),
        although I don't think it will come close to fixing that bug
        completely.

        * rendering/RenderArena.cpp:
        (WebCore::RenderArena::~RenderArena): Use FinishArenaPool, which frees
        the ArenaPool's Arenas immediately.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24164 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
weinig [Tue, 10 Jul 2007 21:37:07 +0000 (21:37 +0000)]
LayoutTests:

        Reviewed by Hyatt.

        - test for http://bugs.webkit.org/show_bug.cgi?id=14449
          REGRESSION (r14345-r14375): Absolutely positioned image does not scale to containing element's height

        * fast/replaced/absolute-position-percentage-height-expected.checksum: Added.
        * fast/replaced/absolute-position-percentage-height-expected.png: Added.
        * fast/replaced/absolute-position-percentage-height-expected.txt: Added.
        * fast/replaced/absolute-position-percentage-height.html: Added.

WebCore:

        Reviewed by Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=14449
          REGRESSION (r14345-r14375): Absolutely positioned image does not scale to containing element's height

        Test: fast/replaced/absolute-position-percentage-height.html

        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::calcReplacedHeightUsing): For positioned elements, evaluate
        the percentage with respect to the container's final height.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24163 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoBuild fix.
thatcher [Tue, 10 Jul 2007 20:21:03 +0000 (20:21 +0000)]
Build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24162 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoCorrectness fix
beidson [Tue, 10 Jul 2007 20:13:26 +0000 (20:13 +0000)]
Correctness fix

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24161 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - fix build
darin [Tue, 10 Jul 2007 20:03:18 +0000 (20:03 +0000)]
    - fix build

        * WebView/WebHTMLView.mm: Add include of ContextMenu.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24159 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Mark Rowe
beidson [Tue, 10 Jul 2007 19:54:32 +0000 (19:54 +0000)]
    Reviewed by Mark Rowe

        Add 64-bit integer converters to our string classes for future work

        * platform/DeprecatedString.cpp:
        (WebCore::DeprecatedString::toInt64):
        (WebCore::DeprecatedString::toUInt64):
        * platform/DeprecatedString.h:

        * platform/PlatformString.h:
        * platform/String.cpp:
        (WebCore::String::toInt64):
        (WebCore::String::toUInt64):

        * platform/StringImpl.cpp:
        (WebCore::StringImpl::toInt64):
        (WebCore::StringImpl::toUInt64):
        * platform/StringImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24158 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * win/Skipped:
aliceli1 [Tue, 10 Jul 2007 19:45:38 +0000 (19:45 +0000)]
    * win/Skipped:
        removing tests that were blocked on lack of support for smart pasting and DOM Paste API

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24157 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Anders.
darin [Tue, 10 Jul 2007 19:21:23 +0000 (19:21 +0000)]
    Reviewed by Anders.

        - fix <rdar://problem/5318732> REGRESSION: In Mail, a crash occurs when attempting to send
          a HTML based message (nytimes.com, washingtonpost.com, latimes.com)

        Did an audit of calls to get the inspector and most other uses of Page* and added null
        checks since Frame* can outlast its Page*.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::createWindow): Added null check.
        (WebCore::FrameLoader::loadSubframe): Ditto.
        (WebCore::FrameLoader::load): Ditto.
        (WebCore::FrameLoader::transitionToCommitted): Ditto.
        (WebCore::FrameLoader::checkLoadCompleteForThisFrame): Ditto.
        (WebCore::FrameLoader::continueLoadAfterWillSubmitForm): Ditto.
        (WebCore::FrameLoader::addExtraFieldsToRequest): Ditto.
        (WebCore::FrameLoader::loadResourceSynchronously): Ditto.
        (WebCore::FrameLoader::continueLoadAfterNavigationPolicy): Ditto.
        (WebCore::FrameLoader::requestFromDelegate): Ditto.
        (WebCore::FrameLoader::canGoBackOrForward): Ditto.
        (WebCore::FrameLoader::getHistoryLength): Ditto.
        (WebCore::FrameLoader::historyURL): Ditto.
        (WebCore::FrameLoader::cachePageForHistoryItem): Ditto.
        (WebCore::FrameLoader::addBackForwardItemClippedAtTarget): Ditto.
        (WebCore::FrameLoader::goToItem): Ditto.
        (WebCore::FrameLoader::dispatchWindowObjectAvailable): Ditto.
        (WebCore::FrameLoader::dispatchDidCommitLoad): Ditto.
        (WebCore::FrameLoader::dispatchAssignIdentifierToInitialRequest): Ditto.
        (WebCore::FrameLoader::dispatchWillSendRequest): Ditto.
        (WebCore::FrameLoader::dispatchDidReceiveResponse): Ditto.
        (WebCore::FrameLoader::dispatchDidReceiveContentLength): Ditto.
        (WebCore::FrameLoader::dispatchDidFinishLoading): Ditto.
        (WebCore::FrameLoader::dispatchDidLoadResourceFromMemoryCache): Ditto.

        * page/ContextMenuController.h:
        * page/ContextMenuController.cpp:
        (WebCore::ContextMenuController::ContextMenuController): Removed unneeded and
        unused Page* parameter.
        (WebCore::openNewWindow): Added null check.
        (WebCore::ContextMenuController::contextMenuItemSelected): Ditto.

        * page/InspectorController.h:
        (WebCore::InspectorController::pageDestroyed): Added. Since this object is
        owned by the Page and has a back-pointer to it, it needs an explicit disconnect.
        There's already a higher-level one, but this is more reliable than that.
        * page/InspectorController.cpp:
        (WebCore::InspectorController::windowScriptObjectAvailable): Added null check.
        (WebCore::InspectorController::windowUnloading): Ditto.

        * page/Page.cpp:
        (WebCore::Page::Page): Updated for change in parameters of ContextMenuController
        constructor.
        (WebCore::Page::~Page): Added call to InspectorController pageDestroyed().

        * page/Chrome.cpp: Updated includes.
        * page/DragController.cpp: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24156 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by John Sullivan.
darin [Tue, 10 Jul 2007 18:47:49 +0000 (18:47 +0000)]
    Reviewed by John Sullivan.

        - <rdar://problem/5321953> remove workaround for fixed AppKit mouse moved bug

        * WebView/WebHTMLView.mm: Put the workaround for bug 3429631 inside an
        ifdef BUILDING_ON_TIGER.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24155 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Adam Treat to make the scroll wheel behave sanely
staikos [Tue, 10 Jul 2007 16:39:54 +0000 (16:39 +0000)]
Patch from Adam Treat to make the scroll wheel behave sanely

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24149 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
ggaren [Tue, 10 Jul 2007 16:25:40 +0000 (16:25 +0000)]
LayoutTests:

        Reviewed by Maciej Stachowiak.

        * fast/events/onload-after-document-close-no-subresource-expected.txt: Added.
        * fast/events/onload-after-document-close-no-subresource.html: Added.
        * fast/events/onload-after-document-close-with-subresource-expected.txt: Added.
        * fast/events/onload-after-document-close-with-subresource.html: Added.
        * fast/events/resources/onload-image.png: Added.

WebCore:

        Reviewed by Maciej Stachowiak.

        Fixed http://bugs.webkit.org/show_bug.cgi?id=13241 | <rdar://problem/5126396>
        onload fires before content is finished loading (woot.com)

        * dom/Document.cpp:
        (WebCore::Document::close): Try to go through the normal "Am I done loading?"
        machinery before dispatching the load event. As the comment indicates,
        we can't do this in all cases, so the fix isn't perfect. However, a perfect
        solution would require much riskier sweeping changes to the loader, which
        are best left for later, I think.

        * xml/XSLTProcessor.cpp:
        (WebCore::XSLTProcessor::createDocumentFromSource): Removed a slightly
        worse solution to the same problem.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24148 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-10 Mark Rowe <mrowe@apple.com>
bdash [Tue, 10 Jul 2007 10:25:57 +0000 (10:25 +0000)]
2007-07-10  Mark Rowe  <mrowe@apple.com>

        Qt build fix after r24126.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24147 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
rwlbuis [Tue, 10 Jul 2007 09:21:47 +0000 (09:21 +0000)]
    Reviewed by Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=8007
        createElement() and XHTML (A dynamically created table not rendered as such)

        Follow the HTML5 rules of making the name lowercase and then
        creating the element in the xhtml namespace.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24146 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Qing Zhao to add context menu support for Qt.
staikos [Tue, 10 Jul 2007 01:43:21 +0000 (01:43 +0000)]
Patch from Qing Zhao to add context menu support for Qt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24131 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * Scripts/run-webkit-tests: Moved leaks from Tiger that are marked as fixed...
darin [Mon, 9 Jul 2007 23:38:57 +0000 (23:38 +0000)]
    * Scripts/run-webkit-tests: Moved leaks from Tiger that are marked as fixed in Leopard to a Tiger-only section.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24130 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
aliceli1 [Mon, 9 Jul 2007 23:23:22 +0000 (23:23 +0000)]
LayoutTests:

        * win/Skipped:
        removed one test from skipped list to see whether it will pass with the new
        DOM Paste API change that is being added simultaneously.

WebKitTools:

        Reviewed by Adam Roben.

        setting DOM Paste Allowed Pref

        * DumpRenderTree/DumpRenderTree.vcproj/DumpRenderTree/DumpRenderTree.cpp:
        (initializePreferences):

win:

        Reviewed by Adam Roben.

        Adding DOM Paste Allowed preference, setters and getters
        for the purpose of layout tests.

        * Interfaces/IWebPreferences.idl:
        * WebPreferenceKeysPrivate.h:
        * WebPreferences.cpp:
        (WebPreferences::isDOMPasteAllowed):
        (WebPreferences::setDOMPasteAllowed):
        * WebPreferences.h:
        * WebView.cpp:
        (WebView::updateWebCoreSettingsFromPreferences):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24129 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build fix.
andersca [Mon, 9 Jul 2007 23:19:16 +0000 (23:19 +0000)]
    Build fix.

        * WebView.cpp:
        (WebView::stringByEvaluatingJavaScriptFromString):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24128 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build fix.
andersca [Mon, 9 Jul 2007 23:18:14 +0000 (23:18 +0000)]
    Build fix.

        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginViewWin::performRequest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24127 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff.
andersca [Mon, 9 Jul 2007 22:10:15 +0000 (22:10 +0000)]
    Reviewed by Geoff.

        Get rid of the node argument in our various script execution functions.

        * bindings/js/kjs_proxy.cpp:
        (WebCore::KJSProxy::evaluate):
        * bindings/js/kjs_proxy.h:
        * bindings/js/kjs_window.cpp:
        (KJS::ScheduledAction::execute):
        * dom/XMLTokenizer.cpp:
        (WebCore::XMLTokenizer::endElementNs):
        (WebCore::XMLTokenizer::notifyFinished):
        * html/HTMLScriptElement.cpp:
        (WebCore::HTMLScriptElement::evaluateScript):
        * html/HTMLTokenizer.cpp:
        (WebCore::HTMLTokenizer::scriptExecution):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::changeLocation):
        (WebCore::FrameLoader::urlSelected):
        (WebCore::FrameLoader::submitForm):
        (WebCore::FrameLoader::replaceContentsWithScriptResult):
        (WebCore::FrameLoader::executeScript):
        * loader/FrameLoader.h:
        * page/mac/WebCoreFrameBridge.mm:
        (-[WebCoreFrameBridge stringByEvaluatingJavaScriptFromString:forceUserGesture:]):
        (-[WebCoreFrameBridge aeDescByEvaluatingJavaScriptFromString:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24126 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam, Sam, and Ada.
kmccullo [Mon, 9 Jul 2007 22:06:54 +0000 (22:06 +0000)]
    Reviewed by Adam, Sam, and Ada.

        - Inital checkin of Drosera for Win.  This isn't pretty and doesn't connect to Safari but debug builds and runs.

        * Drosera/Images/Drosera.ico: Added.
        * Drosera/Images/small.ico: Added.
        * Drosera/win: Added.
        * Drosera/win/BaseDelegate.h: Added.
        (BaseDelegate::QueryInterface):
        (BaseDelegate::didFinishLoadForFrame):
        (BaseDelegate::windowScriptObjectAvailable):
        (BaseDelegate::didStartProvisionalLoadForFrame):
        (BaseDelegate::didReceiveServerRedirectForProvisionalLoadForFrame):
        (BaseDelegate::didFailProvisionalLoadWithError):
        (BaseDelegate::didCommitLoadForFrame):
        (BaseDelegate::didReceiveTitle):
        (BaseDelegate::didReceiveIcon):
        (BaseDelegate::didFailLoadWithError):
        (BaseDelegate::didChangeLocationWithinPageForFrame):
        (BaseDelegate::willPerformClientRedirectToURL):
        (BaseDelegate::didCancelClientRedirectForFrame):
        (BaseDelegate::willCloseFrame):
        (BaseDelegate::createWebViewWithRequest):
        (BaseDelegate::webViewShow):
        (BaseDelegate::webViewClose):
        (BaseDelegate::webViewFocus):
        (BaseDelegate::webViewUnfocus):
        (BaseDelegate::webViewFirstResponder):
        (BaseDelegate::makeFirstResponder):
        (BaseDelegate::setStatusText):
        (BaseDelegate::webViewStatusText):
        (BaseDelegate::webViewAreToolbarsVisible):
        (BaseDelegate::setToolbarsVisible):
        (BaseDelegate::webViewIsStatusBarVisible):
        (BaseDelegate::setStatusBarVisible):
        (BaseDelegate::webViewIsResizable):
        (BaseDelegate::setResizable):
        (BaseDelegate::setFrame):
        (BaseDelegate::webViewFrame):
        (BaseDelegate::setContentRect):
        (BaseDelegate::webViewContentRect):
        (BaseDelegate::runJavaScriptAlertPanelWithMessage):
        (BaseDelegate::runJavaScriptConfirmPanelWithMessage):
        (BaseDelegate::runJavaScriptTextInputPanelWithPrompt):
        (BaseDelegate::runBeforeUnloadConfirmPanelWithMessage):
        (BaseDelegate::runOpenPanelForFileButtonWithResultListener):
        (BaseDelegate::mouseDidMoveOverElement):
        (BaseDelegate::contextMenuItemsForElement):
        (BaseDelegate::validateUserInterfaceItem):
        (BaseDelegate::shouldPerformAction):
        (BaseDelegate::dragDestinationActionMaskForDraggingInfo):
        (BaseDelegate::willPerformDragDestinationAction):
        (BaseDelegate::dragSourceActionMaskForPoint):
        (BaseDelegate::willPerformDragSourceAction):
        (BaseDelegate::contextMenuItemSelected):
        (BaseDelegate::hasCustomMenuImplementation):
        (BaseDelegate::trackCustomPopupMenu):
        (BaseDelegate::measureCustomMenuItem):
        (BaseDelegate::drawCustomMenuItem):
        (BaseDelegate::addCustomMenuDrawingData):
        (BaseDelegate::cleanUpCustomMenuDrawingData):
        (BaseDelegate::canTakeFocus):
        (BaseDelegate::takeFocus):
        (BaseDelegate::registerUndoWithTarget):
        (BaseDelegate::removeAllActionsWithTarget):
        (BaseDelegate::setActionTitle):
        (BaseDelegate::undo):
        (BaseDelegate::redo):
        (BaseDelegate::canUndo):
        (BaseDelegate::canRedo):
        * Drosera/win/DebuggerApplication.cpp: Added.
        (DebuggerApplication::serverLoaded):
        (DebuggerApplication::serverUnloaded):
        (DebuggerApplication::attach):
        (DebuggerApplication::numberOfRowsInTableView):
        (DebuggerApplication::tableView):
        * Drosera/win/DebuggerApplication.h: Added.
        (DebuggerApplication::DebuggerApplication):
        (DebuggerApplication::knownServers):
        * Drosera/win/DebuggerObjectCallbacks.cpp: Added.
        (breakpointEditorHTMLCallback):
        (currentFunctionStackCallback):
        (doubleClickMillisecondsCallback):
        (evaluateScript_inCallFrame_Callback):
        (isPausedCallback):
        (localScopeVariableNamesForCallFrame_Callback):
        (pauseCallback):
        (resumeCallback):
        (stepIntoCallback):
        (valueForScopeVariableNamed_inCallFrame_Callback):
        (staticFunctions):
        * Drosera/win/DebuggerObjectCallbacks.h: Added.
        * Drosera/win/Drosera.cpp: Added.
        (_tWinMain):
        (RegisterDroseraClass):
        (DroseraWndProc):
        (About):
        (Attach):
        (Drosera::Drosera):
        (Drosera::initUI):
        (Drosera::QueryInterface):
        (Drosera::AddRef):
        (Drosera::Release):
        (Drosera::didFinishLoadForFrame):
        (getDroseraJSClass):
        (Drosera::windowScriptObjectAvailable):
        (Drosera::webViewClose):
        (Drosera::validateUserInterfaceItem):
        (Drosera::runJavaScriptAlertPanelWithMessage):
        (Drosera::onSize):
        * Drosera/win/Drosera.h: Added.
        (Drosera::webViewLoaded):
        * Drosera/win/Drosera.vcproj: Added.
        * Drosera/win/Drosera.vcproj/Drosera.rc: Added.
        * Drosera/win/Drosera.vcproj/Drosera.vcproj: Added.
        * Drosera/win/Drosera.vcproj/debug.vsprops: Added.
        * Drosera/win/Drosera.vcproj/release.vsprops: Added.
        * Drosera/win/HelperFunctions.h: Added.
        (CFStringToBSTR):
        * Drosera/win/Info.plist: Added.
        * Drosera/win/resource.h: Added.
        * Drosera/win/stdafx.cpp: Added.
        * Drosera/win/stdafx.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24125 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Eli Fidler to properly decode data: urls.
staikos [Mon, 9 Jul 2007 21:25:12 +0000 (21:25 +0000)]
Patch from Eli Fidler to properly decode data: urls.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24124 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
justing [Mon, 9 Jul 2007 21:23:28 +0000 (21:23 +0000)]
LayoutTests:

        Rolling the fix for 5310404 back in now that 5310404 has been fixed in Mail.

        * editing/deleting/4845371-expected.checksum:
        * editing/deleting/4845371-expected.txt:
        * editing/deleting/4845371.html:
        * editing/deleting/5126166-expected.txt:
        * editing/deleting/5290534.html: Added.
        * editing/deleting/5290534.html-disabled: Removed.

WebCore:

        Rolling the fix for 5290534 back in now that 5310404 has been fixed in Mail.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24123 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Anders Carlsson.
ggaren [Mon, 9 Jul 2007 21:08:10 +0000 (21:08 +0000)]
    Reviewed by Anders Carlsson.

        A little more FrameLoader cleanup while working on <rdar://problem/5126396>.

        Rename: "checkEmitLoadEvent" => "checkCallImplicitClose" because the function
        doesn't emit the load event; it just calls implicitClose.

        Rename: "m_wasLoadEventEmitted" => "m_didCallImplicitClose" because the variable
        tracks whether implicitClose was called.

        * dom/Document.cpp:
        (WebCore::Document::implicitClose): Turned 12 lines of code into 6, adding ASSERT.
        * loader/FrameLoader.h: Removed bogus declaration of non-existent emitLoadEvent().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24122 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
justing [Mon, 9 Jul 2007 20:27:05 +0000 (20:27 +0000)]
LayoutTests:

        Reviewed by John.

        <rdar://problem/5156801> REGRESSION: Crash when deleting table content in Leopard Mail

        * editing/deleting/5156801-expected.checksum: Added.
        * editing/deleting/5156801-expected.png: Added.
        * editing/deleting/5156801-expected.txt: Added.
        * editing/deleting/5156801.html: Added.

WebCore:

        Reviewed by John.

        REGRESSION: Crash when deleting table content in Leopard Mail

        * editing/DeleteSelectionCommand.cpp:
        (WebCore::DeleteSelectionCommand::handleGeneralDelete): Use a RefPtr
        for the next node to delete, in case it is destroyed when the
        current node is deleted.  In the crash reported in the radar, we
        crash because the next node is the div that holds the delete button
        and it's destroyed when the node that contains the selection is removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24120 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Adam Treat to convert QWebFrame from QFrame to QObject. Fixes many
staikos [Mon, 9 Jul 2007 19:41:36 +0000 (19:41 +0000)]
Patch from Adam Treat to convert QWebFrame from QFrame to QObject.  Fixes many
regressions from the previous patch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24119 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
andersca [Mon, 9 Jul 2007 18:28:49 +0000 (18:28 +0000)]
LayoutTests:

        Reviewed by Oliver.

        <rdar://problem/4954319>
        Acrobat 7 / Safari crash: CrashTracer: 99 crashes in Safari at
        com.apple.WebCore: WebCore::NetscapePlugInStreamLoader::isDone const + 0

        Add tests.

        * plugins/return-error-from-new-stream-callback-in-full-frame-plugin-expected.txt: Added.
        * plugins/return-error-from-new-stream-callback-in-full-frame-plugin.html: Added.

WebCore:

        Reviewed by Oliver.

        Add DocumentLoader::isLoadingMainResource.

        * WebCore.exp:

WebKit:

        Reviewed by Oliver.

        <rdar://problem/4954319>
        Acrobat 7 / Safari crash: CrashTracer: 99 crashes in Safari at
        com.apple.WebCore: WebCore::NetscapePlugInStreamLoader::isDone const + 0

        Add a new initWithFrameLoader: method to WebNetscapePluginStream which is to be used when
        the stream is a "fake" stream for full frame plug-ins.

        * Plugins/WebBaseNetscapePluginView.mm:
        (-[WebBaseNetscapePluginView pluginView:receivedResponse:]):
        * Plugins/WebNetscapePluginStream.h:
        * Plugins/WebNetscapePluginStream.mm:
        (-[WebNetscapePluginStream initWithFrameLoader:]):
        (-[WebNetscapePluginStream initWithRequest:plugin:notifyData:sendNotification:]):
        (-[WebNetscapePluginStream start]):
        (-[WebNetscapePluginStream cancelLoadWithError:]):
        (-[WebNetscapePluginStream stop]):
        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::createFrame):

WebKitTools:

        Reviewed by Oliver.

        <rdar://problem/4954319>
        Acrobat 7 / Safari crash: CrashTracer: 99 crashes in Safari at
        com.apple.WebCore: WebCore::NetscapePlugInStreamLoader::isDone const + 0

        If the src url is "data:application/x-webkit-test-netscape,returnerrorfromnewstream",
        return an error from NPP_NewStream.

        * DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.c:
        (pluginAllocate):
        * DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.h:
        * DumpRenderTree/TestNetscapePlugIn.subproj/main.c:
        (NPP_New):
        (NPP_NewStream):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24118 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin
sullivan [Mon, 9 Jul 2007 17:45:08 +0000 (17:45 +0000)]
    Reviewed by Darin

        - fixed <rdar://problem/5320208> WebKit should prevent Time Machine from backing up WebKit clients' icon databases

        * Misc/WebIconDatabase.mm:
        (-[WebIconDatabase init]):
        Re-added the code to exclude the icon database from backups. We now do this at the same time we (try to) import
        the old icon database format, which happens only once per icon database's lifetime.
        (-[WebIconDatabase _importToWebCoreFormat]):
        Assert that we haven't imported yet rather than bailing out. It's now the caller's responsibility to check
        whether we've imported yet.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24117 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago added missing Reviewer to previous checkin
sullivan [Mon, 9 Jul 2007 17:44:08 +0000 (17:44 +0000)]
    added missing Reviewer to previous checkin

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24116 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * WebCore.exp:
sullivan [Mon, 9 Jul 2007 17:43:20 +0000 (17:43 +0000)]
    * WebCore.exp:
        * WebCore.xcodeproj/project.pbxproj:
        exported String::createCFString, needed for <rdar://problem/5320208>

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24115 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix a minor memory leak in the loader
staikos [Mon, 9 Jul 2007 17:21:29 +0000 (17:21 +0000)]
Fix a minor memory leak in the loader

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24114 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix a massive memory leak in the loader.
staikos [Mon, 9 Jul 2007 16:52:40 +0000 (16:52 +0000)]
Fix a massive memory leak in the loader.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24113 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoRework much of WebCoreHttp to make it more stable and support Proxy and SSL.
staikos [Mon, 9 Jul 2007 14:50:22 +0000 (14:50 +0000)]
Rework much of WebCoreHttp to make it more stable and support Proxy and SSL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24108 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
mjs [Mon, 9 Jul 2007 07:57:10 +0000 (07:57 +0000)]
    Reviewed by Maciej.

        Test case for the bug below, which is http://bugs.webkit.org/show_bug.cgi?id=14207

        * manual-tests/flash-unload-tab.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24107 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
mjs [Mon, 9 Jul 2007 07:39:21 +0000 (07:39 +0000)]
JavaScriptCore:

        Reviewed by Oliver.

        - JavaScriptCore part of fix for: <rdar://problem/5295734> Repro crash closing tab/window @ maps.google.com in WTF::HashSet<KJS::RuntimeObjectImp*, WTF::PtrHash<KJS::RuntimeObjectImp*>, WTF::HashTraits<KJS::RuntimeObjectImp*> >::add + 11

        * JavaScriptCore.exp: Added needed export.

WebCore:

        Reviewed by Oliver.

        <rdar://problem/5295734> Repro crash closing tab/window @ maps.google.com in WTF::HashSet<KJS::RuntimeObjectImp*, WTF::PtrHash<KJS::RuntimeObjectImp*>, WTF::HashTraits<KJS::RuntimeObjectImp*> >::add + 11

        Automated test case is not possible. Did not bother with manual test this time.

        * bindings/js/kjs_dom.cpp:
        (WebCore::getRuntimeObject): Check that runtime root isn't null, as well as instance;
        this should cover the case where the plugin is already shut down and onunload time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24106 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam.
mjs [Mon, 9 Jul 2007 06:29:48 +0000 (06:29 +0000)]
    Reviewed by Sam.

        - fixed <rdar://problem/5220706> REGRESSION (TOT): repro crash in -[WebView(WebViewInternal) _addObject:forIdentifier:] [14425]
        http://bugs.webkit.org/show_bug.cgi?id=14425

        * bindings/js/kjs_window.cpp:
        (KJS::createWindow): No longer take an immediate argument - always do immediate loads
        on a newly created Window. Also, do a load of "" to make sure that the right info makes
        it to the app.
        (KJS::showModalDialog): Updated for above.
        (KJS::WindowFunc::callAsFunction): Updated for above.
        * dom/Document.cpp:
        (WebCore::Document::shouldBeAllowedToLoadLocalResources): If our URL is about:blank,
        we're allowed if our opener is (since the opener must have written the contents).
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::changeLocation): Add a variant which takes a KURL, which it
        expects to be pre-completed. This is to avoid completing "" to the opener URL.
        (WebCore::FrameLoader::urlSelected): Allow loading empty URLs.
        * loader/FrameLoader.h:

        Test case is manual only, since it takes particular app behavior to reproduce:

        * manual-tests/new-window-subresource-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24105 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-08 Mitz Pettel <mitz@webkit.org>
bdash [Sun, 8 Jul 2007 21:02:23 +0000 (21:02 +0000)]
2007-07-08  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Maciej.

        - fix leaks introduced in r24099

        * dom/Element.cpp:
        (WebCore::Element::~Element): Do not delete the cached computed style here.
        (WebCore::Element::detach): Delete the computed style here instead.
        (WebCore::Element::computedStyle): Removed a redundant ref() which was
        causing the leak. Changed to return 0 if the element is not attached,
        because in that case there is currently no way to tell when to delete the computed
        style.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24103 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-08 Mark Rowe <mrowe@apple.com>
bdash [Sun, 8 Jul 2007 19:36:31 +0000 (19:36 +0000)]
2007-07-08  Mark Rowe  <mrowe@apple.com>

        Change name from WebKit/Qt to WebKit.

        * CodeCoverage/regenerate-coverage-display:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24102 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * Misc/WebIconDatabase.mm:
sullivan [Sun, 8 Jul 2007 18:55:49 +0000 (18:55 +0000)]
    * Misc/WebIconDatabase.mm:
        (-[WebIconDatabase init]):
        Removed the code to exclude the icon database from backups. We still want to do this, but in a way
        that only runs once ever, instead of once per launch, due to performance concerns.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24101 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoMark code coverage scripts as executable.
bdash [Sun, 8 Jul 2007 18:47:47 +0000 (18:47 +0000)]
Mark code coverage scripts as executable.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24100 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
weinig [Sun, 8 Jul 2007 17:51:28 +0000 (17:51 +0000)]
LayoutTests:

        Reviewed by Darin.

        - test for http://bugs.webkit.org/show_bug.cgi?id=12384
          getComputedStyle with td.style.display='none' reports '' for all properties except display

        * fast/css/computed-style-without-renderer-expected.txt: Added.
        * fast/css/computed-style-without-renderer.html: Added.

WebCore:

        Reviewed by Darin.

        - fix http://bugs.webkit.org/show_bug.cgi?id=12384
          getComputedStyle with td.style.display='none' reports '' for all properties except display

        Test: fast/css/computed-style-without-renderer.html

        Implement rich computed style for elements that do not have a renderer.

        * css/CSSComputedStyleDeclaration.cpp:
        (WebCore::getPositionOffsetValue): Changed to take a RenderStyle instead of
        a RenderObject.
        (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue): Changed to
        use the computed style returned by the node instead of going through the
        renderer. Changed properties that compute by measuring the renderer to
        fall back on returning the length from the RenderStyle if a renderer does
        not exist.
        (WebCore::CSSComputedStyleDeclaration::length): Changed to return the number
        of properties we can compute even if the element has no renderer.
        (WebCore::CSSComputedStyleDeclaration::copyInheritableProperties): Changed
        to use the computed style returned by the node instead of going through the
        renderer.
        * dom/Element.cpp:
        (WebCore::ElementRareData::ElementRareData): Added a computed style member,
        used to cache computed style for elements that do not have a renderer.
        (WebCore::ElementRareData::resetComputedStyle): Added.
        (WebCore::Element::~Element): Delete the cached computed style.
        (WebCore::Element::recalcStyle): Added code to delete the cached computed
        style when necessary.
        (WebCore::Element::computedStyle): Added. Returns the style used in rendering
        the element, if available. Otherwise uses the style selector to compute style
        for the element.
        * dom/Element.h:
        * dom/Node.cpp:
        (WebCore::Node::computedStyle): Added. This method returns either the style
        used in rendering the node, or the style that would be used if the node
        had a renderer.
        * dom/Node.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24099 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-08 Mitz Pettel <mitz@webkit.org>
bdash [Sun, 8 Jul 2007 17:40:54 +0000 (17:40 +0000)]
2007-07-08  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Darin.

        - fix http://bugs.webkit.org/show_bug.cgi?id=14544
          Scroll wheel events are ignored when the cursor is over an iframe's border

        * page/mac/EventHandlerMac.mm:
        (WebCore::EventHandler::passWheelEventToWidget): Changed to return false if
        the event coordinates are not inside the widget's NSView hierarchy.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24098 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Mitz.
rwlbuis [Sun, 8 Jul 2007 15:37:19 +0000 (15:37 +0000)]
    Reviewed by Mitz.

        http://bugs.webkit.org/show_bug.cgi?id=10752
        The inspector should not respond to Make Text Smaller/Bigger commands

        Use the -webkit-text-size-adjust property to ignore text sizing
        and do not take the zoom factor into account for line height when
        -webkit-text-size-adjust is set to none.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24097 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
darin [Sun, 8 Jul 2007 11:49:45 +0000 (11:49 +0000)]
    Reviewed by Maciej.

        - fix <rdar://problem/5318994> Integer overflow in WebCore's Base64.cpp:base64Encode

        * platform/Base64.cpp:
        (WebCore::base64Encode): Check against a maximum size.
        (WebCore::base64Decode): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24096 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
darin [Sun, 8 Jul 2007 11:47:01 +0000 (11:47 +0000)]
WebCore:

        Reviewed by Maciej.

        - fix <rdar://problem/5124665> WebCore secondary-thread assertion should use linked-on-or-after check instead of building on Tiger check

        * platform/ThreadCheck.h: Renamed _WebCoreThreadViolationCheck to reportThreadViolation
        since it's in the WebCore namespace and also we shouldn't use leading underscores since
        those are reserved for the compiler and standard library. Added a new function named
        setDefaultThreadViolationBehavior and an enum for the parameter. Removed the
        threadViolationIsException parameter from WebCoreReportThreadViolation since we use
        a global variable for it now. Got rid of the unnecessary use of do while (0) in the
        WebCoreThreadViolationCheck macro -- since it's only a single function call it's already
        a statement and doesn't need to be turned into one with the do while (0) trick.

        * platform/mac/ThreadCheck.mm:
        (WebCore::readThreadViolationBehaviorFromUserDefaults): Moved the code to read from
        NSUserDefaults into this helper function. Changed it to use stringForKey instead of
        objectForKey so we don't get a "method not found" if someone puts something other than
        a string in there. Updated to use the new globals.
        (WebCore::setDefaultThreadViolationBehavior): Added.
        (WebCore::reportThreadViolation): Simplified, now that most of the work is in the
        readThreadViolationBehaviorFromUserDefaults function.
        (WebCoreReportThreadViolation): Changed to use a global variable instead of parameter
        to determine how to report a violation. Added LogOnFirstThreadViolation, which will
        log only once for each function that violates.

        * WebCore.exp: Updated for above changes.

WebKit:

        Reviewed by Maciej.

        - fix <rdar://problem/5124665> WebCore secondary-thread assertion should use linked-on-or-after check instead of building on Tiger check

        * Misc/WebKitVersionChecks.h: Added WEBKIT_FIRST_VERSION_WITH_MAIN_THREAD_EXCEPTIONS.

        * WebView/WebFrameView.mm: (-[WebFrameView initWithFrame:]): In first-time initialization block,
        check for binaries linked against older versions of WebKit, and set the default thread violation
        behavior to LogOnFirstThreadViolation.

        * WebView/WebView.mm: (+[WebView initialize]): Improved comments.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24095 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-08 Mitz Pettel <mitz@webkit.org>
bdash [Sun, 8 Jul 2007 09:54:42 +0000 (09:54 +0000)]
2007-07-08  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Darin.

        - fix http://bugs.webkit.org/show_bug.cgi?id=14221
          Repro crash (ASSERTION FAILED: oldText in RenderBlock::updateFirstLetter() during relayout of :before content with first-letter style)

        Test: fast/css-generated-content/before-with-first-letter.html

        * rendering/RenderContainer.cpp:
        (WebCore::RenderContainer::updateBeforeAfterContentForContainer): If a child
        of the existing 'before' container is not text, do not assume that it is an
        image, since it can also be an inline wrapping a first-letter. In that case,
        do not update its style (so in particular, keep it a FIRST_LETTER type style)
        and leave it to updateFirstLetter() to take care of.

2007-07-08  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Darin.

        - test for http://bugs.webkit.org/show_bug.cgi?id=14221
          Repro crash (ASSERTION FAILED: oldText in RenderBlock::updateFirstLetter() during relayout of :before content with first-letter style)

        * fast/css-generated-content/before-with-first-letter-expected.checksum: Added.
        * fast/css-generated-content/before-with-first-letter-expected.png: Added.
        * fast/css-generated-content/before-with-first-letter-expected.txt: Added.
        * fast/css-generated-content/before-with-first-letter.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24094 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
weinig [Sun, 8 Jul 2007 01:40:24 +0000 (01:40 +0000)]
LayoutTests:

        Reviewed by Oliver Hunt.

        - update results after removing extraneous calls to isSafeScript for
          readonly attributes.

        * http/tests/security/cross-frame-access-expected.txt:

WebCore:

        Reviewed by Oliver Hunt.

        Bailout early in Window::put() if attribute is readonly to avoid
        extra calls to isSafeScript().

        * bindings/js/kjs_window.cpp:
        (KJS::Window::put):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24093 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - updated results (I changed the test at the last minute, but didn't regenera...
darin [Sat, 7 Jul 2007 21:35:55 +0000 (21:35 +0000)]
    - updated results (I changed the test at the last minute, but didn't regenerate results)

        * fast/canvas/fill-stroke-clip-reset-path-expected.checksum: Updated.
        * fast/canvas/fill-stroke-clip-reset-path-expected.png: Updated.
        * fast/canvas/fill-stroke-clip-reset-path-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24092 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
darin [Sat, 7 Jul 2007 20:42:50 +0000 (20:42 +0000)]
LayoutTests:

        - test for <rdar://problem/5292364> REGRESSION: ZX Spectrum 3.0 widget - graphic anomalies in widget window frame

        * fast/canvas/fill-stroke-clip-reset-path-expected.checksum: Added.
        * fast/canvas/fill-stroke-clip-reset-path-expected.png: Added.
        * fast/canvas/fill-stroke-clip-reset-path-expected.txt: Added.
        * fast/canvas/fill-stroke-clip-reset-path.html: Added.

WebCore:

        Reviewed by Oliver Hunt.

        - fix <rdar://problem/5292364> REGRESSION: ZX Spectrum 3.0 widget - graphic anomalies in widget window frame

        Updated Dashboard quirk so it affects fill, stroke, and clip, rather than just stroke.

        Test: fast/canvas/fill-stroke-clip-reset-path.html

        * html/CanvasRenderingContext2D.h: Added clearPathForDashboardBackwardCompatibilityMode.
        * html/CanvasRenderingContext2D.cpp:
        (WebCore::CanvasRenderingContext2D::clearPathForDashboardBackwardCompatibilityMode): Added.
        (WebCore::CanvasRenderingContext2D::fill): Call the new function.
        (WebCore::CanvasRenderingContext2D::stroke): Replaced in-line code with a call to the new function.
        (WebCore::CanvasRenderingContext2D::clip): Call the new function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24091 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
darin [Sat, 7 Jul 2007 20:14:41 +0000 (20:14 +0000)]
LayoutTests:

        Reviewed by Darin.

        Testcase for:
        http://bugs.webkit.org/show_bug.cgi?id=8994
        Canvas doesn't reset on resize

        * fast/canvas/canvas-resize-reset-expected.checksum: Added.
        * fast/canvas/canvas-resize-reset-expected.png: Added.
        * fast/canvas/canvas-resize-reset-expected.txt: Added.
        * fast/canvas/canvas-resize-reset.html: Added.

WebCore:

        Reviewed and tweaked by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=8994
        Canvas doesn't reset on resize

        Also reset the rendering context when resetting the canvas.

        * html/HTMLCanvasElement.cpp:
        (WebCore::HTMLCanvasElement::reset):
        (WebCore::HTMLCanvasElement::drawingContext):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24090 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fixed a mismatch between test output and expected results (a typo that was...
ap [Sat, 7 Jul 2007 10:11:49 +0000 (10:11 +0000)]
    Fixed a mismatch between test output and expected results (a typo that was only fixed in the former).

        * fast/loader/early-load-cancel-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24089 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
ap [Sat, 7 Jul 2007 10:00:38 +0000 (10:00 +0000)]
    Reviewed by Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=14525
        Support exslt:node-set()

        Test: fast/xsl/exslt-node-set.xml

        * xml/XSLTExtensions.cpp: Added.
        (WebCore::exsltNodeSetFunction): A copy of exslt:node-set() implementation
        from libexslt 1.1.11 (shipped with Tiger).
        (WebCore::registerXSLTExtensions): This function registers any WebCore-provided
        extensions on a context. Currently, the only extension is exslt:node-set().
        * xml/XSLTExtensions.h: Added.

        * xml/XSLTProcessor.cpp:
        (WebCore::XSLTProcessor::transformToString):
        Register extensions when creating a context.

        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        Add XSLTExtensions.{h,cpp} to projects.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24088 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
mjs [Sat, 7 Jul 2007 06:29:08 +0000 (06:29 +0000)]
LayoutTests:

        Reviewed by Oliver.

        - test case for <rdar://problem/5133420> ASSERT in WebDocumentLoaderMac::decreaseLoadCount() un-discarding Gmail message

        * fast/loader/early-load-cancel-expected.txt: Added.
        * fast/loader/early-load-cancel.html: Added.
        * fast/loader/resources/early-load-cancel-inner.html: Added.

WebCore:

        Reviewed by Oliver.

        - fixed <rdar://problem/5133420> ASSERT in WebDocumentLoaderMac::decreaseLoadCount() un-discarding Gmail message

        test case: fast/loader/early-load-cancel.html

        * loader/ResourceLoader.cpp:
        (WebCore::ResourceLoader::didCancel): Don't send cancel-related client callbacks
        if we never even sent willSendRequest because we were deferred.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24087 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-07 Mark Rowe <mrowe@apple.com>
bdash [Sat, 7 Jul 2007 04:28:37 +0000 (04:28 +0000)]
2007-07-07  Mark Rowe  <mrowe@apple.com>

        Reviewed by Oliver.

        * DumpRenderTree/DumpRenderTree.m:
        (convertWebResourceResponseToDictionary): Fix leak of two NSMutableString's introduced in r24076.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24086 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
oliver [Sat, 7 Jul 2007 04:01:28 +0000 (04:01 +0000)]
WebCore:

        Reviewed by Maciej.

        Pulled a number of methods from WebKit and WebCoreFrameBridge into
        C++ objects.

        * WebCore.exp:
        * WebCore.xcodeproj/project.pbxproj:
        * editing/Editor.cpp:
        (WebCore::Editor::replaceMarkedText):
            From -[WebCoreFrameBridge replaceMarkedTextWithText:]
        * editing/Editor.h:
        * page/Frame.cpp:
        (WebCore::Frame::setMarkedTextRange):
            Pulled platform independent logic from Mac specific implementation
        (WebCore::Frame::selectRangeInMarkedText):
            Converted from -[WebHTMLView _selectRangeInMarkedText:]
        * page/Frame.h:
        * page/mac/FrameMac.mm:
        (WebCore::Frame::setMarkedTextRange):
            Now only does Mac specific logic when manipulating initial
            NSAttributedString
        * page/mac/WebCoreFrameBridge.h:
        * page/mac/WebCoreFrameBridge.mm:
            Removed -[WebCoreFrameBridge replaceMarkedTextWithText:]

WebKit:

        Reviewed by Maciej.

        <rdar://problem/5318756> Need to refactor IM/Marked text code to share logic with windows.

        Moved a number of methods from WebHTMLView into WebCore.
        Replaced bridge methods with calls directly into WebCore objects.

        * WebKit.xcodeproj/project.pbxproj:
        * WebView/WebHTMLView.mm:
        (-[WebHTMLView setMarkedText:selectedRange:]):
          Now calls WebCore object methods directly, rather than calling
          via the bridge.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24085 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-06 Mitz Pettel <mitz@webkit.org>
bdash [Sat, 7 Jul 2007 02:37:27 +0000 (02:37 +0000)]
2007-07-06  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Darin.

        - fix http://bugs.webkit.org/show_bug.cgi?id=14545
          REGRESSION (r21854-r21869): Repro crash in RenderBlock::updateFirstLetter @ nola.com/rose/

        Test: fast/css/first-letter-capitalized.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::updateFirstLetter): Moved the call to setFirstLetter
        further down so that the first letter is not destroyed if a text transform
        is applied to the remaining text fragment when it is add to the render tree.
        * rendering/RenderText.cpp:
        (WebCore::RenderText::previousCharacter): Factored out of setTextInternal.
        (WebCore::RenderText::setTextInternal): Changed to call previousCharacter().
        * rendering/RenderText.h:
        * rendering/RenderTextFragment.cpp:
        (WebCore::RenderTextFragment::previousCharacter): Override to return the
        character in the original string before the start of the fragment.
        * rendering/RenderTextFragment.h:

2007-07-06  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Darin.

        - test for http://bugs.webkit.org/show_bug.cgi?id=14545
          REGRESSION (r21854-r21869): Repro crash in RenderBlock::updateFirstLetter @ nola.com/rose/

        * fast/css/first-letter-capitalized-expected.checksum: Added.
        * fast/css/first-letter-capitalized-expected.png: Added.
        * fast/css/first-letter-capitalized-expected.txt: Added.
        * fast/css/first-letter-capitalized.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24084 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-07-06 Mitz Pettel <mitz@webkit.org>
bdash [Sat, 7 Jul 2007 02:35:07 +0000 (02:35 +0000)]
2007-07-06  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Beth.

        - fix http://bugs.webkit.org/show_bug.cgi?id=14536
          Unnecessary layout done when mousing down in text field

        No test possible because there is no change in functionality.

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::setStyle): Reset the height and the width in
        the old style to avoid getting a layout hint as a result of having mutated
        the old style during layout.
        (WebCore::RenderTextControl::layout): Update children's layout if an inner
        block's dimensions should change. This need for layout was previously masked
        by the bug.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24083 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * mac/leopard/Skipped: Updated bug numbers for some bugs that were duplicates.
darin [Sat, 7 Jul 2007 01:02:28 +0000 (01:02 +0000)]
    * mac/leopard/Skipped: Updated bug numbers for some bugs that were duplicates.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24082 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix builds that use IconDatabaseNone.cpp - better, this time
beidson [Sat, 7 Jul 2007 00:27:01 +0000 (00:27 +0000)]
    Fix builds that use IconDatabaseNone.cpp - better, this time

        * loader/icon/IconDatabaseNone.cpp:
        (WebCore::IconDatabase::isOpen):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24081 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix builds that use IconDatabaseNone.cpp
beidson [Sat, 7 Jul 2007 00:20:42 +0000 (00:20 +0000)]
    Fix builds that use IconDatabaseNone.cpp

        * loader/icon/IconDatabaseNone.cpp:
        (WebCore::IconDatabase::databasePath):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24080 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam.
thatcher [Fri, 6 Jul 2007 23:48:31 +0000 (23:48 +0000)]
    Reviewed by Adam.

        <rdar://problem/5313506> Some attributed string layout tests have an extra space character on Leopard

        Leopard serializes empty attribute dictionaries as "{ }", turn these into "{}" to match Tiger.

        * fast/text/attributed-substring-from-range-001.html:
        * mac/leopard/Skipped: Don't skip this test anymore.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24079 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix typo.
thatcher [Fri, 6 Jul 2007 23:44:28 +0000 (23:44 +0000)]
Fix typo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24078 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
sullivan [Fri, 6 Jul 2007 23:43:08 +0000 (23:43 +0000)]
WebCore:

        Reviewed by Brady

        - WebCore part of fix for:
          <rdar://problem/5310739> Time Machine shouldn't back up WebKit's icon database files

        * loader/icon/IconDatabase.h:
        * loader/icon/IconDatabase.cpp:
        (WebCore::IconDatabase::isOpen):
        made this method const (necessary to make IconDatabase::databasePath const)
        (WebCore::IconDatabase::databasePath):
        new method, returns SQLDatabase's path if open, otherwise empty string

        * loader/icon/SQLDatabase.cpp:
        (WebCore::SQLDatabase::open):
        don't append a null character to the m_path ivar. The appended null character was making the path()
        method return a String with a trailing null, which is bad. However, there were no callers to the
        path() method until now. Instead of appending a null character to the ivar, use charactersWithNullTermination()
        instead of characters() when opening the database.

        * loader/icon/SQLDatabase.h:
        (WebCore::SQLDatabase::isOpen):
        made this method const (necessary to make IconDatabase::databasePath const)
        (WebCore::SQLDatabase::path):
        made this method const (not required for these changes, but seemed worth doing at the same time)

        * WebCore.exp:
        added symbol for IconDatabase::databasePath

WebKit:

        Reviewed by Brady

        - WebKit part of fix for:
          <rdar://problem/5310739> Time Machine shouldn't back up WebKit's icon database files

        * Misc/WebIconDatabase.mm:
        (-[WebIconDatabase init]):
        Use CSBackupSetItemExcluded to tell Time Machine not to back up the icon database file

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24077 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
andersca [Fri, 6 Jul 2007 23:40:57 +0000 (23:40 +0000)]
LayoutTests:

        Reviewed by David Kilzer.

        <rdar://problem/5313502>
        Many webarchive tests fail due to different NSURLResponse serialization on Leopard

        * mac/leopard/Skipped:
        * webarchive/archive-with-unencoded-url-expected.txt:
        * webarchive/test-body-background-expected.txt:
        * webarchive/test-duplicate-resources-expected.txt:
        * webarchive/test-frameset-expected.txt:
        * webarchive/test-img-src-expected.txt:
        * webarchive/test-input-src-expected.txt:
        * webarchive/test-link-href-expected.txt:
        * webarchive/test-object-data-expected.txt:
        * webarchive/test-script-src-expected.txt:
        * webarchive/test-table-background-expected.txt:
        * webarchive/test-td-background-expected.txt:

WebKitTools:

        Reviewed by David Kilzer.

        <rdar://problem/5313502>
        Many webarchive tests fail due to different NSURLResponse serialization on Leopard

        Don't dump the serialized form of NSURLResponse. Instead, create a dictionary with the
        response's attributes. Also change the JavaScript MIME type to be "text/javascript".

        * DumpRenderTree/DumpRenderTree.m:
        (convertMIMEType):
        (convertWebResourceDataToString):
        (convertWebResourceResponseToDictionary):
        (serializeWebArchiveToXML):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24076 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
thatcher [Fri, 6 Jul 2007 23:22:49 +0000 (23:22 +0000)]
LayoutTests:

        Reviewed by Oliver.

        <rdar://problem/5313506> Some attributed string layout tests have an extra space character on Leopard

        Don't dump the attributed string's description, this isn't helpfull for the test. Just log
        the attribute names at index 0, which we expect to be nothing, to make sure the string
        has no attributes.

        * fast/forms/attributed-strings-expected.txt:
        * fast/forms/attributed-strings.html:

WebKitTools:

        Reviewed by Oliver.

        Initlize the NSMutableAttributedString directly instead of creating
        a NSAttributedString first.

        * DumpRenderTree/TextInputController.m:
        (-[TextInputController attributedStringWithString:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24075 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agomissed in previous checkin
staikos [Fri, 6 Jul 2007 22:44:11 +0000 (22:44 +0000)]
missed in previous checkin

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24074 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoMove locatizations into another file.
staikos [Fri, 6 Jul 2007 22:36:39 +0000 (22:36 +0000)]
Move locatizations into another file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24073 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Anders Carlsson.
ggaren [Fri, 6 Jul 2007 21:54:22 +0000 (21:54 +0000)]
    Reviewed by Anders Carlsson.

        A little more FrameLoader cleanup while working on <rdar://problem/5126396>.

        Moved url() and URL() methods next to each other to highlight their subtlety.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::url):
        * loader/FrameLoader.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24072 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
andersca [Fri, 6 Jul 2007 21:37:56 +0000 (21:37 +0000)]
LayoutTests:

        Reviewed by David Kilzer.

        Update tests.

        * webarchive/archive-empty-frame-dom-expected.txt:
        * webarchive/archive-empty-frame-source-expected.txt:
        * webarchive/archive-with-unencoded-url-expected.txt:
        * webarchive/test-body-background-expected.txt:
        * webarchive/test-duplicate-resources-expected.txt:
        * webarchive/test-frameset-expected.txt:
        * webarchive/test-img-src-expected.txt:
        * webarchive/test-input-src-expected.txt:
        * webarchive/test-link-href-expected.txt:
        * webarchive/test-object-data-expected.txt:
        * webarchive/test-script-src-expected.txt:
        * webarchive/test-table-background-expected.txt:
        * webarchive/test-td-background-expected.txt:

WebKitTools:

        Reviewed by David Kilzer.

        Replace "Apple Computer" with "Apple" in the DTD declaration.

        * DumpRenderTree/DumpRenderTree.m:
        (serializeWebArchiveToXML):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24071 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Anders Carlsson.
ggaren [Fri, 6 Jul 2007 21:30:12 +0000 (21:30 +0000)]
    Reviewed by Anders Carlsson.

        A little FrameLoader cleanup while working on <rdar://problem/5126396>.

        Renamed endIfNotLoading() to endIfNotLoadingMainResource() because "loading"
        and "loadingMainResource" have distinct meanings and this function
        honors the latter.

        (WebCore::FrameLoader::endIfNotLoadingMainResource): Protect after the
        early return to avoid ref count churn.
        * loader/FrameLoader.h:
        (WebCore::FrameLoader::isLoadingMainResource): Inlined this function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24070 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoPatch from Adam Treat to convert QWebFrame to a QFrame from a QAbstractScrollArea.
staikos [Fri, 6 Jul 2007 21:06:34 +0000 (21:06 +0000)]
Patch from Adam Treat to convert QWebFrame to a QFrame from a QAbstractScrollArea.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24069 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoLayoutTests:
andersca [Fri, 6 Jul 2007 20:52:42 +0000 (20:52 +0000)]
LayoutTests:

        Reviewed by Mitz.

        <rdar://problem/5280532>
        REGRESSION: Can't access nested embed by document.name if object is not immediate parent

        * plugins/embed-inside-object-expected.txt: Added.
        * plugins/embed-inside-object.html: Added.

WebCore:

        Reviewed by Mitz.

        <rdar://problem/5280532>
        REGRESSION: Can't access nested embed by document.name if object is not immediate parent

        Traverse up the tree looking for an object element.

        * html/HTMLEmbedElement.cpp:
        (WebCore::HTMLEmbedElement::getInstance):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24068 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoAdd another Leopard failure
aroben [Fri, 6 Jul 2007 20:29:41 +0000 (20:29 +0000)]
Add another Leopard failure

* mac/leopard/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24067 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoAdd another crasher to the Skipped file
aroben [Fri, 6 Jul 2007 18:11:16 +0000 (18:11 +0000)]
Add another crasher to the Skipped file

Also moved a test that had a bug filed about it out of the section of
unfiled failures.

* win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24066 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Alexey.
andersca [Fri, 6 Jul 2007 18:00:42 +0000 (18:00 +0000)]
    Reviewed by Alexey.

        <rdar://problem/5313518> fast/events/iframe-object-onload.html shows events out of order on Leopard

        Fix test so that it doesn't matter in what order the frames/objects are loaded. This is done by keeping
        the load logs in an array and sort it before dumping it.

        * fast/events/iframe-object-onload-expected.txt:
        * fast/events/iframe-object-onload.html:
        * mac/leopard/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@24065 268f45cc-cd09-0410-ab3c-d52691b4dbfc