1. 10 Sep, 2010 40 commits
    • simon.fraser@apple.com's avatar
      2010-09-10 Simon Fraser <simon.fraser@apple.com> · c38d6941
      simon.fraser@apple.com authored
              Re-enable some Windows compositing tests that no longer assert.
      
              * platform/win/Skipped:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67237 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c38d6941
    • jschuh@chromium.org's avatar
      2010-09-10 Justin Schuh <jschuh@chromium.org> · fc23e5fd
      jschuh@chromium.org authored
              Reviewed by Darin Adler.
      
              SVGGElement::rendererIsNeeded should return false when parent isn't SVG
              https://bugs.webkit.org/show_bug.cgi?id=45562
      
              Test: svg/custom/g-outside-svg.html
      
              * svg/SVGGElement.cpp:
              (WebCore::SVGGElement::rendererIsNeeded):
              * svg/SVGGElement.h:
      2010-09-10  Justin Schuh  <jschuh@chromium.org>
      
              Reviewed by Darin Adler.
      
              SVGGElement::rendererIsNeeded should return false when parent isn't SVG
              https://bugs.webkit.org/show_bug.cgi?id=45562
      
              * svg/custom/g-outside-svg-expected.txt: Added.
              * svg/custom/g-outside-svg.html: Added.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67236 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fc23e5fd
    • jianli@chromium.org's avatar
      Skip file api tests for workers that could crash occasionally before · 6c76ca92
      jianli@chromium.org authored
      we figure out how to fix them.
      
      * platform/mac/Skipped:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67235 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      6c76ca92
    • abarth@webkit.org's avatar
      2010-09-10 Adam Barth <abarth@webkit.org> · f5c3ebb0
      abarth@webkit.org authored
              Chromium build fix (one hopes!)
      
              * src/WebFrameImpl.cpp:
              (WebKit::WebFrameImpl::commitDocumentData):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67234 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f5c3ebb0
    • abarth@webkit.org's avatar
      2010-09-10 Adam Barth <abarth@webkit.org> · 586c745b
      abarth@webkit.org authored
              Reviewed by Darin Adler.
      
              REGRESSION (HTML Parser): CNN's Money site is not formatted correctly with AdBlock installed
              https://bugs.webkit.org/show_bug.cgi?id=41371
      
              Drop support for closing comments with "-- >".  This was an attempt to
              be more IE-like in comment parsing, but it turns out to cause problems
              for some high-profile sites.  Firefox 4 is dropping support as well.
              We expect the spec to change soon too.
      
              * html/parser/HTMLTokenizer.cpp:
              (WebCore::HTMLTokenizer::nextToken):
              * html/parser/HTMLTokenizer.h:
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Darin Adler.
      
              REGRESSION (HTML Parser): CNN's Money site is not formatted correctly with AdBlock installed
              https://bugs.webkit.org/show_bug.cgi?id=41371
      
              Update test results for new behavior.
      
              * fast/parser/comments-expected.txt:
              * fast/parser/comments.html:
                - Removed the redundant in-test expectations.
              * html5lib/runner-expected.txt:
                - We now "fail" these two tests.  We'll pass them again once the spec
                  changes and HTML5lib updates its expectations.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67233 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      586c745b
    • simon.fraser@apple.com's avatar
      2010-09-10 Simon Fraser <simon.fraser@apple.com> · f94d5fa1
      simon.fraser@apple.com authored
              Reviewed by Adam Roben.
      
              compositing/iframes/iframe-src-change.html test fails on Windows
              https://bugs.webkit.org/show_bug.cgi?id=45554
      
              This test had the wrong behavior. It never actually entered compositing mode.
              Now give it a chance to do so (and test for it), and then test that the layers
              disappeared.
      
              * compositing/iframes/iframe-src-change-expected.txt:
              * compositing/iframes/iframe-src-change.html:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67232 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f94d5fa1
    • cmarrin@apple.com's avatar
      Restored a Changelog entry for a change that was reverted. I mistakenly... · 8b83ea02
      cmarrin@apple.com authored
      Restored a Changelog entry for a change that was reverted. I mistakenly thought reverted changes should be removed.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67231 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      8b83ea02
    • aroben@apple.com's avatar
      Give expected results files the right names · 12240863
      aroben@apple.com authored
      * platform/win/fast/text/hyphenate-first-word-expected.txt: Renamed from LayoutTests/platform/win/fast/text/hyphenate-first-word-actual.txt.
      * platform/win/fast/text/hyphenate-locale-expected.txt: Renamed from LayoutTests/platform/win/fast/text/hyphenate-locale-actual.txt.
      * platform/win/html5lib/runner-expected.txt: Renamed from LayoutTests/platform/win/html5lib/runner-actual.txt.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67230 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      12240863
    • cmarrin@apple.com's avatar
      2010-09-10 Chris Marrin <cmarrin@apple.com> · 49b8ffc5
      cmarrin@apple.com authored
              Unreviewed.
      
              Adding file that should have been in http://trac.webkit.org/changeset/67226 but 
              I forgot to svn add.
              
              * platform/graphics/opengl: Added.
              * platform/graphics/opengl/GraphicsContext3DOpenGL.cpp: Copied from platform/graphics/mac/GraphicsContext3DMac.mm.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67229 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      49b8ffc5
    • cmarrin@apple.com's avatar
      one more typo fix in Changelog · 97fc64aa
      cmarrin@apple.com authored
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67228 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      97fc64aa
    • cmarrin@apple.com's avatar
      fixed typo in Changelog · 54933c28
      cmarrin@apple.com authored
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67227 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      54933c28
    • cmarrin@apple.com's avatar
      2010-09-10 Paul Sawaya <psawaya@apple.com> · 020eee1b
      cmarrin@apple.com authored
              Reviewed by Chris Marr.
      
              Refactored non-Mac specific code into GraphicsContext3DOpenGL.
              
              Moving this file into platform/graphics/mac for now, and only building on Mac.
              
              https://bugs.webkit.org/show_bug.cgi?id=30625
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67226 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      020eee1b
    • oliver@apple.com's avatar
      Qt build fix. · cf4cc358
      oliver@apple.com authored
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67225 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      cf4cc358
    • abarth@webkit.org's avatar
      2010-09-10 Adam Barth <abarth@webkit.org> · bd3c1561
      abarth@webkit.org authored
              Reviewed by Darin Adler.
      
              Remove useless FrameLoader method
              https://bugs.webkit.org/show_bug.cgi?id=45492
      
              This method has one caller, and that caller already has a more direct
              way of accessing the correct DocumentLoader.
      
              * loader/FrameLoader.cpp:
              * loader/FrameLoader.h:
              * loader/MainResourceLoader.cpp:
              (WebCore::MainResourceLoader::addData):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67224 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bd3c1561
    • abarth@webkit.org's avatar
      2010-09-10 Adam Barth <abarth@webkit.org> · 9b003a98
      abarth@webkit.org authored
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              After the MainResourceLoader hands the bytes for the main resource to
              the DocumentLoader, the DocumentLoader hands the bytes to the
              FrameLoaderClient.  The FrameLoaderClient, in turn, returns the bytes
              to WebCore.  Prior to this patch, the FrameLoaderClient returned the
              bytes to the FrameLoader.  However, the FrameLoader was actually adding
              any value.
      
              In this patch, we change the FrameLoaderClient to return the bytes to
              the DocumentLoader directly.  In a future patch, we'll cut out the
              FrameLoaderClient middleman and handle the bytes internally in the
              DocumentLoader.
      
              * WebCore.exp.in:
              * loader/DocumentLoader.cpp:
              (WebCore::DocumentLoader::addData):
              * loader/DocumentLoader.h:
              * loader/FrameLoader.cpp:
              (WebCore::FrameLoader::finishedLoadingDocument):
              * loader/FrameLoader.h:
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              Also, removed comment about wrong code in the Mac port.
      
              * src/WebFrameImpl.cpp:
              (WebKit::WebFrameImpl::commitDocumentData):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebCoreSupport/FrameLoaderClientEfl.cpp:
              (WebCore::FrameLoaderClientEfl::committedLoad):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebCoreSupport/FrameLoaderClientGtk.cpp:
              (WebKit::FrameLoaderClient::committedLoad):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebCoreSupport/FrameLoaderClientHaiku.cpp:
              (WebCore::FrameLoaderClientHaiku::committedLoad):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              Previously, we were checking the document for null.  However, Frame can
              never have a null document, so this check is no longer needed.
      
              * WebView/WebFrame.mm:
              (-[WebFrame _addData:]):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebCoreSupport/FrameLoaderClientQt.cpp:
              (WebCore::FrameLoaderClientQt::committedLoad):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebCoreSupport/WebFrameLoaderClient.cpp:
              (WebFrameLoaderClient::receivedData):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebKitSupport/FrameLoaderClientWx.cpp:
              (WebCore::FrameLoaderClientWx::committedLoad):
      2010-09-10  Adam Barth  <abarth@webkit.org>
      
              Reviewed by Eric Seidel.
      
              Main resource bytes shouldn't bounce through FrameLoader
              https://bugs.webkit.org/show_bug.cgi?id=45496
      
              Now return the bytes to the DocumentLoader.
      
              * WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
              (WebKit::WebFrameLoaderClient::receivedData):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67223 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9b003a98
    • oliver@apple.com's avatar
      2010-09-10 Oliver Hunt <oliver@apple.com> · af02ea3b
      oliver@apple.com authored
              Reviewed by Darin Adler.
      
              Speed up deserialisation of strings
              https://bugs.webkit.org/show_bug.cgi?id=45555
      
              Rather than building a list of Identifiers for the string pool
              we now build a list of a tuple of UString and JSString*.  This
              doesn't hurt the property name case as ustring->identifier conversion
              is essentially free if the ustring has already been converted to
              an Identifier, but saves an unnecessary Identifier creation for
              strings we only ever use to create JSStrings.  We also reduce
              GC pressure for duplicate strings by caching the JSStrings.
      
              * bindings/js/SerializedScriptValue.cpp:
              (WebCore::CloneDeserializer::CachedString::CachedString):
              (WebCore::CloneDeserializer::CachedString::jsString):
              (WebCore::CloneDeserializer::CachedString::ustring):
              (WebCore::CloneDeserializer::readStringData):
              (WebCore::CloneDeserializer::putProperty):
              (WebCore::CloneDeserializer::readFile):
              (WebCore::CloneDeserializer::readTerminal):
              (WebCore::CloneDeserializer::deserialize):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67222 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      af02ea3b
    • rniwa@webkit.org's avatar
      2010-09-10 Ryosuke Niwa <rniwa@webkit.org> · 064213e0
      rniwa@webkit.org authored
              Reviewed by Antonio Gomes.
      
              REGRESSION(r67176): editing/selection/doubleclick-inline-first-last-contenteditable.html crashes
              https://bugs.webkit.org/show_bug.cgi?id=45525
      
              The crash was caused by ancestorToRetainStructureAndAppearance not taking into consideration
              the case where there is no common block ancestor. Fixed this by adding a null check.
      
              Test: editing/pasteboard/copy-without-common-block-crash.html
      
              * editing/markup.cpp:
              (WebCore::ancestorToRetainStructureAndAppearance): Added a null check.
      2010-09-10  Ryosuke Niwa  <rniwa@webkit.org>
      
              Reviewed by Antonio Gomes.
      
              REGRESSION(r67176): editing/selection/doubleclick-inline-first-last-contenteditable.html crashes
              https://bugs.webkit.org/show_bug.cgi?id=45525
      
              Added a test to ensure WebKit does not crash when copying a region that doesn't have any common block ancestor.
              Also removed doubleclick-inline-first-last-contenteditable.html from the qt's skipped list since
              this test should not crash anymore.
      
              * editing/pasteboard/copy-without-common-block-crash-expected.txt: Added.
              * editing/pasteboard/copy-without-common-block-crash.html: Added.
              * platform/qt/Skipped: Removed doubleclick-inline-first-last-contenteditable.html from the list.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67221 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      064213e0
    • kinuko@chromium.org's avatar
      2010-09-09 Kinuko Yasuda <kinuko@chromium.org> · 35a4a326
      kinuko@chromium.org authored
              Reviewed by Dumitru Daniliuc.
      
              [FileSystem] Do not call EntriesCallback more than once if there're no entries.
              https://bugs.webkit.org/show_bug.cgi?id=45498
      
              No new tests, layout-tests that confirm the behavior will be added in later patches.
      
              * fileapi/FileSystemCallbacks.cpp:
              (WebCore::EntriesCallbacks::EntriesCallbacks):
              (WebCore::EntriesCallbacks::didReadDirectoryEntry):
              (WebCore::EntriesCallbacks::didReadDirectoryEntries):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67220 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      35a4a326
    • cfleizach@apple.com's avatar
      WebCore: Use String::fromUTF8 instead of UTF8Encoding().decode · 01ee2ed9
      cfleizach@apple.com authored
      in the libxml XMLDocumentParser
      https://bugs.webkit.org/show_bug.cgi?id=45488
      
      Patch by Patrick Gansterer <paroga@paroga.com> on 2010-09-10
      Reviewed by Darin Adler.
      
      * dom/XMLDocumentParserLibxml2.cpp:
      (WebCore::toString):
      
      LayoutTests: Land new expected results for some tests on Windows
      
      Patch by Adam Roben <aroben@apple.com> on 2010-09-10
      Rubber-stamped in advance by Simon Fraser.
      
      * platform/win/fast/text/hyphenate-first-word-actual.txt: Added.
      * platform/win/fast/text/hyphenate-locale-actual.txt: Added.
      Windows doesn't support hyphenation.
      
      * platform/win/html5lib/runner-actual.txt: Added.
      We don't implement <keygen> on Windows, so we pass some tests that
      expect <keygen> to have no child elements (on Mac, the child elements
      are real DOM elements, rather than living in the shadow DOM tree like
      HTML5 mandates, so the tests fail).
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67219 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      01ee2ed9
    • tony@chromium.org's avatar
      2010-09-10 Tony Chang <tony@chromium.org> · efa63b90
      tony@chromium.org authored
              Unreviewed, rolling out r67216.
              http://trac.webkit.org/changeset/67216
              https://bugs.webkit.org/show_bug.cgi?id=44709
      
              Broke
      
              * Scripts/webkitpy/layout_tests/deduplicate_tests.py:
              * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67218 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      efa63b90
    • aroben@apple.com's avatar
      Land new expected results for some tests on Windows · 435069eb
      aroben@apple.com authored
      Rubber-stamped in advance by Simon Fraser.
      
      * platform/win/fast/text/hyphenate-first-word-actual.txt: Added.
      * platform/win/fast/text/hyphenate-locale-actual.txt: Added.
      Windows doesn't support hyphenation.
      
      * platform/win/html5lib/runner-actual.txt: Added.
      We don't implement <keygen> on Windows, so we pass some tests that
      expect <keygen> to have no child elements (on Mac, the child elements
      are real DOM elements, rather than living in the shadow DOM tree like
      HTML5 mandates, so the tests fail).
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67217 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      435069eb
    • tony@chromium.org's avatar
      2010-09-10 Tony Chang <tony@chromium.org> · 57170687
      tony@chromium.org authored
              Reviewed by Ojan Vafai.
      
              deduplicate-tests should be runnable from any WebKit directory
              https://bugs.webkit.org/show_bug.cgi?id=44709
      
              * Scripts/webkitpy/layout_tests/deduplicate_tests.py:
              * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67216 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      57170687
    • mrobinson@webkit.org's avatar
      2010-09-10 Martin Robinson <mrobinson@igalia.com> · 7aa9c774
      mrobinson@webkit.org authored
              Reviewed by Xan Lopez.
      
              [GTK] Placement new / manual destructor invocation should be used on private GObject memory
              https://bugs.webkit.org/show_bug.cgi?id=45550
      
              GLib allocates and deallocates GObject private data structs itself. When
              those structs contain C++ members, their constructors and destructors are not
              called. This is not only dangerous, it makes RefPtr-type smart pointers much
              less useful. We can fix this problem by calling placement new on the private
              data struct during instance initialization and calling the destructor during
              finalization.
      
              This patch takes that approach and switches plain char* members of
              WebKitWebView (with manual memory allocation) to use CString.
      
              * webkit/webkitprivate.h: Switch char* members to CString.
              * webkit/webkitwebview.cpp:
              (webkit_web_view_finalize): Manually call the destructor on the private
              data. Remove manual deallocation of members which are now CString.
              (webkit_web_view_query_tooltip): Update to reflect CString change.
              (webkit_web_view_init): Use placement new to initialize C++ members of
              the private data section.
              (webkit_web_view_get_encoding): Update to reflect CString change.
              (webkit_web_view_get_custom_encoding): Ditto.
              (webkit_web_view_add_resource): Ditto.
              (webkit_web_view_get_resource): Ditto.
              (webkit_web_view_clear_resources): Ditto.
              (webkit_web_view_set_tooltip_text): Ditto.
              (webkit_web_view_get_icon_uri): Ditto.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67215 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7aa9c774
    • commit-queue@webkit.org's avatar
      2010-09-10 Patrick Gansterer <paroga@paroga.com> · 181d1ee5
      commit-queue@webkit.org authored
              Reviewed by Darin Adler.
      
              Use String::fromUTF8 instead of UTF8Encoding().decode
              in the libxml XMLDocumentParser
              https://bugs.webkit.org/show_bug.cgi?id=45488
      
              * dom/XMLDocumentParserLibxml2.cpp:
              (WebCore::toString):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67214 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      181d1ee5
    • joepeck@webkit.org's avatar
      2010-09-09 Joseph Pecoraro <joepeck@webkit.org> · 90e48393
      joepeck@webkit.org authored
              Reviewed by Andreas Kling.
      
              Clean Up Some Expected Results for Missing Tests
              https://bugs.webkit.org/show_bug.cgi?id=45489
      
                Test was removed in r35456 but expected results were missed.
      
              * http/tests/xmlhttprequest/access-control-basic-exclude-expected.txt: Removed.
      
                Results were added in r19036 but the test was never added. No record of the test.
      
              * editing/pasteboard/4944770-expected.txt: Removed.
              * platform/mac/editing/pasteboard/4944770-expected.checksum: Removed.
              * platform/mac/editing/pasteboard/4944770-expected.png: Removed.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67213 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      90e48393
    • simon.fraser@apple.com's avatar
      2010-09-10 Simon Fraser <simon.fraser@apple.com> · 933e48fd
      simon.fraser@apple.com authored
              Reviewed by Chris Marrin.
      
              Update Windows LayoutTest/compositing results
              https://bugs.webkit.org/show_bug.cgi?id=45323
      
              Fix up some windows compositing test results.
      
              * platform/win/compositing/iframes/composited-parent-iframe-expected.txt: Removed.
              * platform/win/compositing/iframes/iframe-resize-expected.txt: Removed.
              * platform/win/compositing/iframes/leave-compositing-iframe-expected.txt: Removed.
              * platform/win/compositing/iframes/overlapped-iframe-expected.txt: Removed.
              * platform/win/compositing/plugins/composited-plugin-expected.txt: Added.
              * platform/win/compositing/plugins/small-to-large-composited-plugin-expected.txt: Copied from LayoutTests/compositing/iframes/leave-compositing-iframe-expected.txt.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67212 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      933e48fd
    • kov@webkit.org's avatar
      2010-09-10 Gustavo Noronha Silva <gns@gnome.org> · 2b9514d7
      kov@webkit.org authored
              Reviewed by Martin Robinson.
      
              [GTK] Google sites do not like WebKitGTK+
              https://bugs.webkit.org/show_bug.cgi?id=39617
      
              Special-case Google domains, and spoof User-Agent when talking to
              them, to stop being treated as a second-class citizen.
      
              * WebCoreSupport/FrameLoaderClientGtk.cpp:
              (WebKit::initializeDomainsList):
              (WebKit::isGoogleDomain):
              (WebKit::FrameLoaderClient::userAgent): If enable-site-specific-quirks
              is enabled, send the standard WebKit User-Agent string, disregarding the
              custom one set by the browser for Google domains.
              * webkit/webkitprivate.h:
              * webkit/webkitwebsettings.cpp:
              (webkitPlatform): Fix style, and simplify.
              (webkitOSVersion): Ditto.
              (webkitUserAgent): Add the Version/x.y string Safari has been
              using since Safari 2.2.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67211 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      2b9514d7
    • yurys@chromium.org's avatar
      2010-09-10 Pavel Podivilov <podivilov@chromium.org> · 948cf1a2
      yurys@chromium.org authored
              Reviewed by Yury Semikhatsky.
      
              [Chromium] rebaseline SVGFEDiffuseLightingElement-lighting-color-css-prop.html
              https://bugs.webkit.org/show_bug.cgi?id=45540
      
              * platform/chromium-linux/svg/dynamic-updates/SVGFEDiffuseLightingElement-lighting-color-css-prop-expected.checksum:
              * platform/chromium-linux/svg/dynamic-updates/SVGFEDiffuseLightingElement-lighting-color-css-prop-expected.png:
              * platform/chromium-win/svg/dynamic-updates/SVGFEDiffuseLightingElement-lighting-color-css-prop-expected.checksum:
              * platform/chromium-win/svg/dynamic-updates/SVGFEDiffuseLightingElement-lighting-color-css-prop-expected.png:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67210 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      948cf1a2
    • cfleizach@apple.com's avatar
      Mail crashes with searching for next misspelled word with VoiceOver · fbca8294
      cfleizach@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=45501
      
      Reviewed by David Kilzer.
      
      WebCore: 
      
      Test: platform/mac/accessibility/crash-in-element-for-text-marker.html
      
      * accessibility/mac/AccessibilityObjectWrapper.mm:
      (-[AccessibilityObjectWrapper accessibilityAttributeValue:forParameter:]):
          Check that the AX object is not nil before asking for the wrapper().
      
      LayoutTests: 
      
      * platform/mac/accessibility/crash-in-element-for-text-marker-expected.txt: Added.
      * platform/mac/accessibility/crash-in-element-for-text-marker.html: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67209 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fbca8294
    • commit-queue@webkit.org's avatar
      2010-09-10 Michael Nordman <michaeln@google.com> · fa990e91
      commit-queue@webkit.org authored
              Reviewed by Dumitru Daniliuc.
      
              Decouple Blob from ScriptExecutionContext.
              https://bugs.webkit.org/show_bug.cgi?id=45410
      
              - Removed ScriptExecutionContext pointers from Blob and File constructors.
              - Removed ScriptExecutionContext's collection of Blobs instance in that context.
              - Refactored BlobURL to make a distinction between 'public' blob urls and 'internal'
                urls that are used solely as identifiers. The former requires a SecurityOrigin to
                to manufacture, the latter does not.
              - Added a constant and accessor for the blob protocol scheme on the BlobURL class.
      
              No new tests are added since this is a cleanup.
      
              * bindings/js/SerializedScriptValue.cpp:
              (WebCore::CloneDeserializer::readFile):
              (WebCore::CloneDeserializer::readTerminal):
              * bindings/v8/SerializedScriptValue.cpp:
              (WebCore::ZigZag::Reader::readBlob):
              (WebCore::ZigZag::Reader::readFile):
              (WebCore::ZigZag::Reader::readFileList):
              * dom/ScriptExecutionContext.cpp:
              (WebCore::ScriptExecutionContext::~ScriptExecutionContext):
              (WebCore::ScriptExecutionContext::createPublicBlobURL):
              (WebCore::ScriptExecutionContext::revokePublicBlobURL):
              * dom/ScriptExecutionContext.h:
              * fileapi/Blob.cpp:
              (WebCore::Blob::Blob):
              (WebCore::Blob::~Blob):
              (WebCore::Blob::slice):
              * fileapi/Blob.h:
              (WebCore::Blob::create):
              (WebCore::Blob::url):
              * fileapi/Blob.idl:
              * fileapi/BlobBuilder.cpp:
              (WebCore::BlobBuilder::getBlob):
              * fileapi/BlobBuilder.h:
              * fileapi/BlobBuilder.idl:
              * fileapi/BlobURL.cpp:
              (WebCore::BlobURL::createPublicURL):
              (WebCore::BlobURL::createInternalURL):
              (WebCore::BlobURL::getOrigin):
              (WebCore::BlobURL::getIdentifier):
              (WebCore::BlobURL::createBlobURL):
              * fileapi/BlobURL.h:
              (WebCore::BlobURL::blobProtocol):
              (WebCore::BlobURL::BlobURL):
              * fileapi/File.cpp:
              (WebCore::File::File):
              * fileapi/File.h:
              (WebCore::File::create):
              * fileapi/FileReader.cpp: Register a public url to perform the read.
              (WebCore::FileReader::cleanup):
              (WebCore::FileReader::start):
              (WebCore::FileReader::didFinishLoading):
              (WebCore::FileReader::failed):
              * fileapi/FileReader.h: Added m_urlForReading data member.
              * fileapi/FileReaderSync.cpp:
              (WebCore::FileReaderSync::read): Register a public url to perform the read.
              * fileapi/ThreadableBlobRegistry.cpp:
              (WebCore::ThreadableBlobRegistry::registerBlobURL):
              (WebCore::ThreadableBlobRegistry::unregisterBlobURL):
              * fileapi/ThreadableBlobRegistry.h:
              * html/HTMLInputElement.cpp:
              (WebCore::HTMLInputElement::appendFormData):
              (WebCore::HTMLInputElement::setFileListFromRenderer):
              * page/SecurityOrigin.cpp: Use BlobURL.blobProtocol().
              (WebCore::SecurityOrigin::create):
              (WebCore::SecurityOrigin::canRequest):
              (WebCore::SecurityOrigin::canDisplay):
              * platform/chromium/ClipboardChromium.cpp:
              (WebCore::ClipboardChromium::files):
              * platform/gtk/ClipboardGtk.cpp:
              (WebCore::ClipboardGtk::files):
              * platform/mac/ClipboardMac.mm:
              (WebCore::ClipboardMac::files):
              * platform/win/ClipboardWin.cpp:
              (WebCore::ClipboardWin::files):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67208 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fa990e91
    • xan@webkit.org's avatar
      2010-09-10 Xan Lopez <xlopez@igalia.com> · da6ca9d9
      xan@webkit.org authored
              Reviewed by Martin Robinson.
      
              GDK_DISPLAY was removed in GTK+ 3.x, re-declare it when built
              against that.
      
              * platform/gtk/GtkVersioning.h:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67207 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      da6ca9d9
    • xan@webkit.org's avatar
      WebCore: · ae0e2c64
      xan@webkit.org authored
      2010-09-10  Xan Lopez  <xlopez@igalia.com>
      
              Reviewed by Martin Robinson.
      
              Fix compilation with GTK+ 3.x.
      
              GDK key macros were renamed from GDK_FOO to GDK_KEY_FOO, we need
              to include the compat header provided if we want to keep using the
              old names.
      
              * platform/gtk/GtkVersioning.h: add the compatibility header.
              * platform/gtk/KeyEventGtk.cpp: include GtkVersioning.h
      
      WebKit/gtk:
      
      2010-09-10  Xan Lopez  <xlopez@igalia.com>
      
              Reviewed by Martin Robinson.
      
              Fix compilation with GTK+ 3.x.
      
              GDK key macros were renamed from GDK_FOO to GDK_KEY_FOO, we need
              to include the compat header provided if we want to keep using the
              old names.
      
              * WebCoreSupport/FullscreenVideoController.cpp: include GtkVersioning.h
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67206 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      ae0e2c64
    • vestbo@webkit.org's avatar
      Skip a few media tests that started randomly failing after r66961 · 4796b841
      vestbo@webkit.org authored
      Reviewed by Eric Carlson.
      
      * platform/mac-leopard/Skipped:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67205 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4796b841
    • tonyg@chromium.org's avatar
      2010-09-09 Tony Gentilcore <tonyg@chromium.org> · 9fa4859d
      tonyg@chromium.org authored
              Reviewed by Eric Seidel.
      
              Factor out the NestingLevelIncrementer
              https://bugs.webkit.org/show_bug.cgi?id=45471
      
              No new tests because no new functionality.
      
              * GNUmakefile.am:
              * WebCore.gypi:
              * WebCore.vcproj/WebCore.vcproj:
              * WebCore.xcodeproj/project.pbxproj:
              * html/parser/HTMLDocumentParser.cpp:
              * html/parser/HTMLDocumentParser.h:
              * html/parser/HTMLScriptRunner.cpp:
              * html/parser/NestingLevelIncrementer.h: Added.
              (WebCore::NestingLevelIncrementer::NestingLevelIncrementer):
              (WebCore::NestingLevelIncrementer::~NestingLevelIncrementer):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67204 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9fa4859d
    • zimmermann@webkit.org's avatar
      2010-09-10 Nikolas Zimmermann <nzimmermann@rim.com> · 04f4596d
      zimmermann@webkit.org authored
              Not reviewed. Fix wrong gtk baseline for select-x-list-1.svg.
      
              * platform/gtk/svg/text/select-x-list-1-expected.txt:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67203 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      04f4596d
    • aroben@apple.com's avatar
      Don't crash when a frame is destroyed after the UI process has disconnected · bf92439f
      aroben@apple.com authored
      Fixes <http://webkit.org/b/45535> <rdar://problem/8412928> Crash in
      WebProcess::removeWebFrame on exit with WebKit2 (Release builds only)
      
      Reviewed by Darin Adler.
      
      * WebProcess/WebProcess.cpp:
      (WebKit::WebProcess::removeWebFrame): Null-check m_connection before
      dereferencing it.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67202 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bf92439f
    • mrobinson@webkit.org's avatar
      2010-09-10 Martin Robinson <mrobinson@igalia.com> · 4f59c84d
      mrobinson@webkit.org authored
              Slight skipped list reorganization.
      
              * platform/gtk/Skipped: Group one more test in the beginDragWithFiles section.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67201 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4f59c84d
    • zimmermann@webkit.org's avatar
      2010-09-10 Nikolas Zimmermann <nzimmermann@rim.com> · 6e96afd5
      zimmermann@webkit.org authored
              Reviewed by Dirk Schulze.
      
              SVG text chunk concept needs to be integrated in the InlineBox structure
              https://bugs.webkit.org/show_bug.cgi?id=45533
      
              Begin rewriting the SVG text layout code. As first step move the text chunk concept right into the inline box structure.
              This obsoletes the whole text chunk part concept (see SVGTextChunkLayoutInfo.h for details), and will soon simplify the
              whole layout code a lot. Previously we hacked around the fact that a single InlineTextBox could contain multiple text chunks
              For example: <text x="20 30 40">ABC</text> only led to the creation of one InlineTextBox containing three chunks. When
              painting such an InlineTextBox, we had to do nasty hacks (same for text selection etc.)
              
              This is finally fixed by incorporating the knownledge about text chunks into the line layout algorithm.
              Note, this has no impact on HTML at all, no worries.
      
              It's even required to do it this way as BiDi reordering shouldn't happen across text chunks.
              These are only the first bits, and currently slows down text layouting. Follow-up patches will remove the old layout
              algorithm, and will rely on the stored SVGTextLayoutAttributes in the RenerSVGInlineText renderers (instead of computing again).
      
              * Android.mk: Add rendering/svg/SVGTextLayoutAttributes.* & rendering/svg/SVGTextLayoutBuilder.* to build.
              * CMakeLists.txt: Ditto.
              * GNUmakefile.am: Ditto.
              * WebCore.gyp/WebCore.gyp: Ditto.
              * WebCore.gypi: Ditto.
              * WebCore.pro: Ditto.
              * WebCore.vcproj/WebCore.vcproj: Ditto.
              * WebCore.vcproj/WebCoreCommon.vsprops: Ditto. (Recognize new rendering/svg directory.)
              * WebCore.vcproj/copyForwardingHeaders.cmd: Ditto. (Copy forwarding headers from rendering/svg directory.)
              * WebCore.xcodeproj/project.pbxproj: Ditto.
              * rendering/RenderBlockLineLayout.cpp: This is the key of the new concept, take text chunks into account when building the InlineBox structure.
              (WebCore::RenderBlock::findNextLineBreak): Split midpointState, if we encounter the beginning of a new text chunk.
              * rendering/RenderObject.h:
              (WebCore::RenderObject::isSVGInline): New method, which will be used to identify RenderSVGInline objects.
              * rendering/RenderSVGInline.h:
              (WebCore::RenderSVGInline::isSVGInline): Return true.
              * rendering/RenderSVGInlineText.cpp:
              (WebCore::RenderSVGInlineText::linesBoundingBox): Overriden, to simplify the calculations for SVG.
              (WebCore::RenderSVGInlineText::characterStartsNewTextChunk): Determines wheter the given position starts a new text chunk.
              * rendering/RenderSVGInlineText.h:
              (WebCore::RenderSVGInlineText::storeLayoutAttributes): Stores the layout attributes calculated in the text layout process (will be used everywhere in follow-up patches).
              * rendering/RenderSVGText.cpp:
              (WebCore::RenderSVGText::layout): Invoke SVGTextLayoutBuilder before creating the inline box tree, so we can decide which characters start new text chunks.
              * rendering/RenderText.h: Made linesBoundingBox() virtual so RenderSVGInlineText can override it.
              * rendering/svg: Added.
              * rendering/svg/SVGTextLayoutAttributes.cpp: Added. Holds x/y/dx/dy/rotate values and all kerning/ligature information used during layout.
              * rendering/svg/SVGTextLayoutAttributes.h: Added.
              * rendering/svg/SVGTextLayoutBuilder.cpp: Added.
              * rendering/svg/SVGTextLayoutBuilder.h: Added.
      2010-09-10  Nikolas Zimmermann  <nzimmermann@rim.com>
      
              Reviewed by Dirk Schulze.
      
              SVG text chunk concept needs to be integrated in the InlineBox structure
              https://bugs.webkit.org/show_bug.cgi?id=45533
      
              Updated all tests containing text chunks, to reflect the fact that more SVGInlineTextBoxes are created now.
              All 'startOffset' dumps now begin at 0, indicating that there's only one chunk contained in a single SVGInlineTextBox.
              These dumps will be removed in follow-up patches, as they don't make any sense anymore in the new concept.
      
              * platform/gtk/svg/text/select-x-list-1-expected.txt:
              * platform/gtk/svg/text/select-x-list-2-expected.txt:
              * platform/gtk/svg/text/select-x-list-3-expected.txt:
              * platform/gtk/svg/text/select-x-list-4-expected.txt:
              * platform/gtk/svg/text/select-x-list-with-tspans-1-expected.txt:
              * platform/gtk/svg/text/select-x-list-with-tspans-2-expected.txt:
              * platform/gtk/svg/text/select-x-list-with-tspans-3-expected.txt:
              * platform/gtk/svg/text/select-x-list-with-tspans-4-expected.txt: 
              * platform/mac/svg/W3C-SVG-1.1/animate-elem-40-t-expected.txt:
              * platform/mac/svg/W3C-SVG-1.1/text-text-04-t-expected.txt:
              * platform/mac/svg/W3C-SVG-1.1/text-text-05-t-expected.txt:
              * platform/mac/svg/W3C-SVG-1.1/text-text-06-t-expected.txt:
              * platform/mac/svg/W3C-SVG-1.1/text-text-07-t-expected.txt:
              * platform/mac/svg/W3C-SVG-1.1/text-tspan-01-b-expected.txt:
              * platform/mac/svg/batik/text/textAnchor2-expected.txt:
              * platform/mac/svg/batik/text/textAnchor3-expected.txt:
              * platform/mac/svg/batik/text/textLayout-expected.txt:
              * platform/mac/svg/batik/text/textLayout2-expected.txt:
              * platform/mac/svg/batik/text/textOnPathSpaces-expected.txt:
              * platform/mac/svg/batik/text/textPosition-expected.txt:
              * platform/mac/svg/batik/text/textPosition2-expected.txt:
              * platform/mac/svg/custom/text-letter-spacing-expected.txt:
              * platform/mac/svg/custom/text-x-dx-lists-expected.txt:
              * platform/mac/svg/custom/text-x-override-in-tspan-child-expected.txt:
              * platform/mac/svg/text/select-x-list-1-expected.txt:
              * platform/mac/svg/text/select-x-list-2-expected.txt:
              * platform/mac/svg/text/select-x-list-3-expected.txt:
              * platform/mac/svg/text/select-x-list-4-expected.txt:
              * platform/mac/svg/text/select-x-list-with-tspans-1-expected.txt:
              * platform/mac/svg/text/select-x-list-with-tspans-2-expected.txt:
              * platform/mac/svg/text/select-x-list-with-tspans-3-expected.txt:
              * platform/mac/svg/text/select-x-list-with-tspans-4-expected.txt:
              * platform/mac/svg/text/text-text-04-t-expected.txt:
              * platform/mac/svg/text/text-text-05-t-expected.txt:
              * platform/mac/svg/text/text-text-06-t-expected.txt:
              * platform/mac/svg/text/text-text-07-t-expected.txt:
              * platform/mac/svg/text/text-tspan-01-b-expected.txt:
              * svg/custom/glyph-transformation-with-hkern-expected.txt:
              * platform/qt/svg/text/select-x-list-1-expected.txt:
              * platform/qt/svg/text/select-x-list-2-expected.txt:
              * platform/qt/svg/text/select-x-list-3-expected.txt:
              * platform/qt/svg/text/select-x-list-4-expected.txt:
              * platform/qt/svg/text/select-x-list-with-tspans-1-expected.txt:
              * platform/qt/svg/text/select-x-list-with-tspans-2-expected.txt:
              * platform/qt/svg/text/select-x-list-with-tspans-3-expected.txt:
              * platform/qt/svg/text/select-x-list-with-tspans-4-expected.txt:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67200 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      6e96afd5
    • ossy@webkit.org's avatar
      Unreviewed fix after r67085. · 152d2aa9
      ossy@webkit.org authored
      * WebCore.pro: Fix previously moved filenames.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67199 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      152d2aa9
    • abarth@webkit.org's avatar
      2010-09-10 Adam Barth <abarth@webkit.org> · 02d22791
      abarth@webkit.org authored
              Reviewed by Eric Seidel.
      
              Inline FrameLoader::committedLoad into its one caller
              https://bugs.webkit.org/show_bug.cgi?id=45481
      
              This saves one round-trip through the FrameLoader when loading
              documents.
      
              * loader/DocumentLoader.cpp:
              (WebCore::DocumentLoader::commitLoad):
              * loader/FrameLoader.cpp:
              * loader/FrameLoader.h:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@67198 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      02d22791