AX: [ATK] Convert the get_text atktest.c unit tests to layout tests
[WebKit-https.git] / Source / WebKit / gtk / ChangeLog
index de5dba7..cbdfe28 100644 (file)
+2013-12-10  Joanmarie Diggs  <jdiggs@igalia.com>
+
+        AX: [ATK] Convert the get_text atktest.c unit tests to layout tests
+        https://bugs.webkit.org/show_bug.cgi?id=125497
+
+        Reviewed by Mario Sanchez Prada.
+
+        * tests/testatk.c: Remove the portions of the remaining tests which now exist as
+        layout tests.
+        (testWebkitAtkCaretOffsets):
+        (testWebkitAtkCaretOffsetsAndExtranousWhiteSpaces):
+        (testWebkitAtkComboBox):
+        (testWebkitAtkEmbeddedObjects):
+        (testWebkitAtkTextSelections):
+        (testWebkitAtkListsOfItems):
+        (main):
+
+2013-12-09  Joanmarie Diggs  <jdiggs@igalia.com>
+
+        AX: [ATK] Convert the get_{string,text}_at_offset atktest.c unit tests to layout tests
+        https://bugs.webkit.org/show_bug.cgi?id=125451
+
+        Reviewed by Mario Sanchez Prada.
+
+        * tests/testatk.c: Remove the tests which now exist as layout tests. Note that the
+        tests for atk_text_get_text_{before,after}_offset were removed without equivalents
+        added to the layout tests. The same is true for the END AtkTextBoundary types. Both
+        have been deprecated in ATK and are not being used by AT-SPI2 assistive technologies.
+        (testGetTextFunction):
+        (main):
+
+2013-12-08  Martin Robinson  <mrobinson@igalia.com>
+
+        [WK2][Soup] Use didReceiveBuffer instead of didReceiveData
+        https://bugs.webkit.org/show_bug.cgi?id=118598
+
+        Reviewed by Gustavo Noronha Silva.
+
+        Original patch by Kwang Yul Seo  <skyul@company100.net> and Csaba Osztrogonác  <ossy@webkit.org>.
+
+        Switch from using didReceiveData to didReceiveBuffer for the Soup backend and
+        let SharedBuffer wrap a SoupBuffer. This is necessary because the NetworkProcess
+        only supports getting data via SharedBuffer.
+
+        * webkit/webkitdownload.cpp:
+        (DownloadClient::didReceiveData): Replace with ASSERT_NOT_REACHED.
+        (DownloadClient::didReceiveBuffer): Use this to process incoming data.
+
+2013-12-07  Gustavo Noronha Silva  <gns@gnome.org>
+
+        Fix API test expectation following 160220.
+
+        Rubber-stamped by Martin Robinson.
+
+        * tests/testatkroles.c:
+        (finish_loading): rename variable documentFrame -> document.
+        (test_webkit_atk_get_role_document_frame): check for ATK_ROLE_DOCUMENT_WEB instead of
+        ATK_ROLE_DOCUMENT_FRAME.
+        (test_webkit_atk_get_role_heading): rename variable documentFrame -> document.
+        (test_webkit_atk_get_role_image): ditto.
+        (test_webkit_atk_get_role_link): ditto.
+        (test_webkit_atk_get_role_list_and_item): ditto.
+        (test_webkit_atk_get_role_paragraph): ditto.
+        (test_webkit_atk_get_role_section): ditto.
+        (test_webkit_atk_get_role_table): ditto.
+        (test_webkit_atk_get_role_separator): ditto.
+        (test_webkit_atk_get_role_combobox): ditto.
+        (test_webkit_atk_get_role_form): ditto.
+        (test_webkit_atk_get_role_check_box): ditto.
+        (test_webkit_atk_get_role_entry): ditto.
+        (test_webkit_atk_get_role_label): ditto.
+        (test_webkit_atk_get_role_listbox): ditto.
+        (test_webkit_atk_get_role_password_text): ditto.
+        (test_webkit_atk_get_role_push_button): ditto.
+        (test_webkit_atk_get_role_radio_button): ditto.
+
+2013-12-05  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] Fix GObject introspection warnings in webkitspellchecker
+        https://bugs.webkit.org/show_bug.cgi?id=125299
+
+        Reviewed by Philippe Normand.
+
+        * webkit/webkitspellchecker.cpp: Add missing ':' after some
+        gobject-introspection annotations.
+
+2013-12-03  Joseph Pecoraro  <pecoraro@apple.com>
+
+        Web Inspector: Push Remote Inspector debugging connection management into JavaScriptCore
+        https://bugs.webkit.org/show_bug.cgi?id=124613
+
+        Reviewed by Timothy Hatcher.
+
+        * WebCoreSupport/InspectorClientGtk.h:
+
+2013-12-02  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
+
+        [GTK] GTK2 paint code path does not render AC layers
+        https://bugs.webkit.org/show_bug.cgi?id=124967
+
+        Reviewed by Carlos Garcia Campos.
+
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_expose_event): also paint AC layers when painting the widget,
+        when AC is on.
+
+2013-11-28  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
+
+        REGRESSION(r154658): webkit_web_view_get_view_source_mode always returns false
+        https://bugs.webkit.org/show_bug.cgi?id=124954
+
+        Reviewed by Carlos Garcia Campos.
+
+        * tests/testwebview.c: new test to ensure setting and getting source mode work as intended.
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_get_view_source_mode): actually return the value we query from WebCore.
+
+2013-11-18  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        Unreviewed. Update NEWS and Versions.m4 for 2.3.2 release.
+
+        * NEWS: Added release notes for 2.3.2.
+
+2013-11-05  Xabier Rodriguez Calvar  <calvaris@igalia.com>
+
+        [GStreamer] Remove NATIVE_FULLSCREEN_VIDEO support
+        https://bugs.webkit.org/show_bug.cgi?id=123437
+
+        Reviewed by Philippe Normand.
+
+        Removed some dead code as no GStreamer port is using the native
+        fullscreen video support.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::enterFullScreenForElement):
+        (WebKit::ChromeClient::exitFullScreenForElement):
+        * WebCoreSupport/ChromeClientGtk.h: Removed code related to
+        GStreamer and NATIVE_FULLSCREEN_VIDEO.
+
+2013-11-04  Manuel Rego Casasnovas  <rego@igalia.com>
+
+        [GTK][WK1] Enable CSS Regions if feature flag is set to TRUE
+        https://bugs.webkit.org/show_bug.cgi?id=123739
+
+        Reviewed by Martin Robinson.
+
+        In WK1 CSS Regions were always disabled even if you set the feature flag
+        ENABLE_CSS_REGIONS to TRUE. Enabling CSS Regions support in that case.
+
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_update_settings): Enable CSS Regions if feautre flag is
+        set to TRUE.
+
+2013-10-29  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        Unreviewed. Update NEWS and Versions.m4 for 2.3.1 release.
+
+        * NEWS: Added release notes for 2.3.1.
+
+2013-10-28  Bastien Nocera <hadess@hadess.net>
+
+        Name all the GLib timeout sources
+        https://bugs.webkit.org/show_bug.cgi?id=123229
+
+        Reviewed by Anders Carlsson.
+
+        Give a name to GLib timeout sources, this is helpful when
+        profiling WebKitGTK applications.
+
+2013-10-28  Bastien Nocera <hadess@hadess.net>
+
+        Replace 0 timeouts g_timeout_add() by g_idle_add()
+        https://bugs.webkit.org/show_bug.cgi?id=123260
+
+        Reviewed by Carlos Garcia Campos.
+
+        A zero timeout should be equivalent to using g_idle_add_full(G_PRIORITY_DEFAULT, ...)
+        without the nagging feeling that the wrong API was used.
+
+        * WebCoreSupport/ChromeClientGtk.cpp: Use g_idle_add() instead
+        of 0-timer.
+        (WebKit::ChromeClient::closeWindowSoon):
+        (WebKit::ChromeClient::widgetSizeChanged):
+        * WebCoreSupport/GtkAdjustmentWatcher.cpp: Ditto.
+        (WebKit::GtkAdjustmentWatcher::updateAdjustmentsFromScrollbarsLater):
+        * webkit/webkitwebview.cpp: Ditto.
+        (webkit_web_view_get_subresources):
+
+2013-10-21  Zan Dobersek  <zdobersek@igalia.com>
+
+        [GTK] g-ir-scanner should use the configured C compiler
+        https://bugs.webkit.org/show_bug.cgi?id=123088
+
+        Reviewed by Carlos Garcia Campos.
+
+        * GNUmakefile.am: Pass the C compiler that was detected at configuration to the g-ir-scanner command
+        via the CC environment variable. This solves the issue of the scanner running the GCC compiler with
+        the Clang-specific CFLAGS.
+
+2013-10-17  Andreas Kling  <akling@apple.com>
+
+        Massage the Gtk build.
+
+        * WebCoreSupport/EditorClientGtk.cpp:
+        (WebKit::EditorClient::shouldApplyStyle):
+
+2013-10-16  Jochen Eisinger  <jochen@chromium.org>
+
+        A page should exit fullscreen mode if it opens a new popup
+        https://bugs.webkit.org/show_bug.cgi?id=122865
+
+        Reviewed by Jer Noble.
+
+        If a fullscreen page opens a popup, the popup would be hidden and
+        therefore invisible to the user. To avoid this, exit fullscreen mode
+        before opening a new window.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::createWindow):
+
+2013-10-09  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [ATK] Implement new API in AtkText: atk_text_get_string_at_offset()
+        https://bugs.webkit.org/show_bug.cgi?id=120638
+
+        Reviewed by Gustavo Noronha Silva.
+
+        Update ATK unit test to check the new API as well if the version
+        of ATK is new enough.
+
+        * tests/testatk.c:
+        (testGetStringFunction): New helper function, to check the new API.
+        (runGetStringTests): Ditto.
+        (testWebkitAtkGetStringAtOffset): New unit test, similar to the
+        one already present to check the old API but focused in the new one.
+        (testWebkitAtkGetStringAtOffsetNewlines): Ditto.
+        (testWebkitAtkGetStringAtOffsetTextarea): Ditto.
+        (testWebkitAtkGetStringAtOffsetTextInput): Ditto.
+        (testWebkitAtkGetStringAtOffsetWithPreformattedText): Ditto.
+        (testWebkitAtkGetStringAtOffsetWithSpecialCharacters): Ditto.
+        (testWebkitAtkGetStringAtOffsetWithWrappedLines): Ditto.
+        (testWebkitAtkGetStringAtOffsetWithEmbeddedObjects): Ditto.
+        (testWebkitAtkGetExtents): Add checks for the new API, if possible.
+        (testWebkitAtkLinksWithInlineImages): Ditto.
+        (main): Add the new unit tests to check the new API, if possible.
+
+2013-10-07  Sam Weinig  <sam@webkit.org>
+
+        Consolidate findString functions
+        https://bugs.webkit.org/show_bug.cgi?id=122480
+
+        Reviewed by Darin Adler.
+
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_search_text):
+
+2013-10-07  Sam Weinig  <sam@webkit.org>
+
+        CTTE: Use references in and around DragController
+        https://bugs.webkit.org/show_bug.cgi?id=122427
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/DragClientGtk.cpp:
+        (WebKit::DragClient::willPerformDragDestinationAction):
+        (WebKit::DragClient::willPerformDragSourceAction):
+        (WebKit::DragClient::actionMaskForDrag):
+        (WebKit::DragClient::startDrag):
+        * WebCoreSupport/DragClientGtk.h:
+        * webkit/webkitwebview.cpp:
+        (dragExitedCallback):
+        (webkit_web_view_drag_motion):
+        (webkit_web_view_drag_data_received):
+        (webkit_web_view_drag_drop):
+
+2013-10-04  Philippe Normand  <pnormand@igalia.com>
+
+        [GTK] UserMediaClientGtk needs API update
+        https://bugs.webkit.org/show_bug.cgi?id=122324
+
+        Reviewed by Martin Robinson.
+
+        Update UserMediaClientGtk API after changes made in the parent
+        class in r156473.
+
+        * WebCoreSupport/UserMediaClientGtk.cpp:
+        (WebKit::UserMediaClientGtk::requestPermission):
+        (WebKit::UserMediaClientGtk::cancelRequest):
+        * WebCoreSupport/UserMediaClientGtk.h:
+
+2013-10-05  Zan Dobersek  <zdobersek@igalia.com>
+
+        Unreviewed GTK build fix after r156946. The framePolicyFunction is an attribute
+        on the WebKitWebPolicyDecisionPrivate object.
+
+        * webkit/webkitwebpolicydecision.cpp:
+        (webkit_web_policy_decision_use):
+        (webkit_web_policy_decision_ignore):
+        (webkit_web_policy_decision_download):
+
+2013-10-04  Darin Adler  <darin@apple.com>
+
+        text-transform: lowercase is not lang-dependent (Turkish languages : tr,az)
+        https://bugs.webkit.org/show_bug.cgi?id=21312
+
+        Reviewed by Ryosuke Niwa.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp:
+        (DumpRenderTreeSupportGtk::doCommand): Use StringImpl::upper instead of
+        StringImpl::makeUpper.
+
+2013-10-04  Anders Carlsson  <andersca@apple.com>
+
+        FramePolicyFunction should be an std::function
+        https://bugs.webkit.org/show_bug.cgi?id=122362
+
+        Reviewed by Darin Adler.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::dispatchWillSubmitForm):
+        (WebKit::FrameLoaderClient::dispatchDecidePolicyForResponse):
+        (WebKit::FrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
+        * WebCoreSupport/FrameLoaderClientGtk.h:
+        * webkit/webkitwebpolicydecision.cpp:
+        (webkit_web_policy_decision_use):
+        (webkit_web_policy_decision_ignore):
+        (webkit_web_policy_decision_download):
+
+2013-10-03  Sam Weinig  <sam@webkit.org>
+
+        Remove shouldRubberBandInDirection from the WKBundlePageUIClient
+        https://bugs.webkit.org/show_bug.cgi?id=122309
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/ChromeClientGtk.h:
+
+2013-10-01  Sam Weinig  <sam@webkit.org>
+
+        CTTE: DOMWrapperWorlds should be passed around by reference
+        https://bugs.webkit.org/show_bug.cgi?id=122206
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::dispatchDidClearWindowObjectInWorld):
+        * WebCoreSupport/FrameLoaderClientGtk.h:
+
+2013-09-30  Benjamin Poulain  <benjamin@webkit.org>
+
+        Remove the code guarded by STYLE_SCOPED
+        https://bugs.webkit.org/show_bug.cgi?id=122123
+
+        Reviewed by Anders Carlsson.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp:
+        * WebCoreSupport/DumpRenderTreeSupportGtk.h:
+
+2013-09-28  Darin Adler  <darin@apple.com>
+
+        Add Frame::mainFrame and Frame::isMainFrame
+        https://bugs.webkit.org/show_bug.cgi?id=122064
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::assignIdentifierToInitialRequest):
+        Call isMainFrame.
+
+2013-09-28  Philippe Normand  <pnormand@igalia.com>
+
+        Unreviewed GTK build fix after r156560.
+
+        * WebCoreSupport/AcceleratedCompositingContext.h:
+
+2013-09-26  Darin Adler  <darin@apple.com>
+
+        rename KURL to URL
+        https://bugs.webkit.org/show_bug.cgi?id=16214
+
+        Reviewed by Andreas Kling.
+
+        * many files: Renamed, using script.
+
+2013-09-27  Darin Adler  <darin@apple.com>
+
+        Add empty MainFrame class
+        https://bugs.webkit.org/show_bug.cgi?id=121770
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/AcceleratedCompositingContextGL.cpp:
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        * WebCoreSupport/GtkAdjustmentWatcher.cpp:
+        * webkit/webkitwebframe.cpp:
+        * webkit/webkitwebview.cpp:
+        * webkit/webkitviewportattributes.cpp:
+        Include MainFrame.h instead of Frame.h as needed.
+
+2013-09-25  Brent Fulgham  <bfulgham@apple.com>
+
+        Refactor RuntimeEnabledFeatures as a Singleton.
+        https://bugs.webkit.org/show_bug.cgi?id=121883.
+
+        Reviewed by Jer Noble.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp:
+        (DumpRenderTreeSupportGtk::setCSSRegionsEnabled): Use singleton.
+        (DumpRenderTreeSupportGtk::setExperimentalContentSecurityPolicyFeaturesEnabled):
+        (DumpRenderTreeSupportGtk::setSeamlessIFramesEnabled):
+        (DumpRenderTreeSupportGtk::setShadowDOMEnabled):
+        (DumpRenderTreeSupportGtk::setStyleScopedEnabled):
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_update_settings): Ditto.
+
+2013-09-24  Lorenzo Tilve  <ltilve@igalia.com>
+
+        [GTK] Fix compilation problems when setting ENABLE_DRAG_SUPPORT = FALSE
+        https://bugs.webkit.org/show_bug.cgi?id=121782
+
+        Reviewed by Martin Robinson.
+
+        Disabled drag functions.
+
+        * WebCoreSupport/DragClientGtk.cpp:
+        * WebCoreSupport/DragClientGtk.h:
+
+2013-09-24  Zan Dobersek  <zdobersek@igalia.com>
+
+        [GTK] Fix accelerated compositing disabling under Wayland displays
+        https://bugs.webkit.org/show_bug.cgi?id=121788
+
+        Reviewed by Gustavo Noronha Silva.
+
+        * webkit/webkitwebview.cpp:
+        (updateAcceleratedCompositingSetting): Fix a ridiculous early return in the case of disabling the accelerated
+        compositing setting. There's no need to re-disable the setting if it was already disabled, but it should
+        definitely be disabled if the setting was previously enabled.
+
+2013-09-22  Zan Dobersek  <zdobersek@igalia.com>
+
+        Unreviewed, GTK build fix when building with GTK+ 2.0.
+        Patch provided by Dominique Leuenberger <dimstar@opensuse.org>.
+
+        * WebCoreSupport/AcceleratedCompositingContextGL.cpp: Include GtkVersioning.h to set up
+        the mock GDK_IS_X11_DISPLAY macro if it's not provided by the GTK+ headers.
+
+2013-09-20  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [ATK] Do not expose aria-help in ATK based platforms
+        https://bugs.webkit.org/show_bug.cgi?id=121675
+
+        Reviewed by Chris Fleizach.
+
+        Removed accessibilityHelpText() helper method from
+        DumpRenderSupportGtk's API, since it's not longer needed.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp: Removed method.
+        * WebCoreSupport/DumpRenderTreeSupportGtk.h: Ditto.
+
+2013-09-16  Enrica Casucci  <enrica@apple.com>
+
+        Remove unused function didSetSelectionTypesForPasteboard from EditorClient.
+        https://bugs.webkit.org/show_bug.cgi?id=121464
+
+        Reviewed by Darin Adler.
+
+        * WebCoreSupport/EditorClientGtk.cpp:
+        * WebCoreSupport/EditorClientGtk.h:
+
+2013-09-16  Gustavo Noronha Silva  <gns@gnome.org>
+
+        [GTK] Make symbol export filter more strict, and disable for dev/test builds
+        https://bugs.webkit.org/show_bug.cgi?id=120586
+
+        Reviewed by Martin Robinson.
+
+        * GNUmakefile.am: only use the version script when in developer mode..
+
+2013-09-14  Alberto Garcia  <berto@igalia.com>
+
+        [GTK] WebKitGTK+ is linking against libxslt in too many places
+        https://bugs.webkit.org/show_bug.cgi?id=121356
+
+        Reviewed by Martin Robinson.
+
+        Don't use LIBXSLT_CFLAGS when compiling libwebkitgtk, no source
+        files use libxslt there.
+
+        The webkit tests don't use libxslt, so we shouldn't link against
+        it.
+
+        * GNUmakefile.am:
+
+2013-09-13  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
+
+        [GTK] Move to the new web inspector
+        https://bugs.webkit.org/show_bug.cgi?id=120647
+
+        Reviewed by Carlos Garcia Campos.
+
+        * WebCoreSupport/InspectorClientGtk.cpp:
+        (WebKit::InspectorClient::openInspectorFrontend): update paths to the internal resource ones.
+        (WebKit::InspectorClient::inspectorFilesPath): ditto.
+        * tests/testwebinspector.c:
+        (consoleMessageCallback): the new inspector currently emits an error when loading, so add it to
+        the list of messages which are not considered failures.
+        * webkit/webkitglobals.cpp:
+        (webkitInit): register resource as a local scheme; currently used by the inspector.
+
+2013-09-13  Alberto Garcia  <berto@igalia.com>
+
+        download-requested arg should be WEBKIT_TYPE_DOWNLOAD not G_TYPE_OBJECT
+        https://bugs.webkit.org/show_bug.cgi?id=57634
+
+        Reviewed by Darin Adler.
+
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_class_init):
+
+2013-09-12  Zan Dobersek  <zdobersek@igalia.com>
+
+        [GTK][WK1] Block accelerated compositing under non-X11 displays
+        https://bugs.webkit.org/show_bug.cgi?id=121165
+
+        Reviewed by Martin Robinson.
+
+        This is analogous to the changes in r154728 and r154729 that enforce disabling accelerated compositing
+        under Wayland displays and made the RedirectedXCompositeWindow use in WebKit2 limited only to the X11 displays.
+
+        * WebCoreSupport/AcceleratedCompositingContextGL.cpp: Remove two unnecessary header includes that also
+        introduce symbols that conflict with the symbols defined in the X headers. Include the <gdk/gdkx.h> header
+        if the GTK+ dependency can provide it.
+        (WebKit::AcceleratedCompositingContext::initialize): Only set up the RedirectedXCompositeWindow instance
+        if running under an X11 display.
+        (WebKit::AcceleratedCompositingContext::renderLayersToWindow): The removal of the unnecessary header
+        inclusions also removed the std::max() injection into the global scope. Using std::max directly instead.
+        (WebKit::AcceleratedCompositingContext::scheduleLayerFlush): Ditto.
+        * webkit/webkitwebview.cpp:
+        (updateAcceleratedCompositingSetting): A helper function that ensures the accelerated compositing feature
+        stays disabled under Wayland displays. It also prints out a warning message the first time the user tries
+        to enable accelerated compositing under Wayland displays, telling that the feature is not supported and
+        will remain disabled.
+        (webkit_web_view_update_settings): Call updateAcceleratedCompositingSetting() to enable the feature if possible.
+        (webkit_web_view_settings_notify): Ditto.
+
+2013-09-11  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [GTK] Remove Gail dependency from build system for GTK3
+        https://bugs.webkit.org/show_bug.cgi?id=119673
+
+        Reviewed by Gustavo Noronha Silva.
+
+        * GNUmakefile.am: Removed GAIL_CFLAGS and GAIL_LIBS.
+
+2013-09-10  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [GTK] Reimplement atk_text_get_text_*_offset for LINE boundaries
+        https://bugs.webkit.org/show_bug.cgi?id=114872
+
+        Reviewed by Gustavo Noronha Silva.
+
+        Fixed wrong unit test.
+
+        * tests/testatk.c:
+        (testWebkitAtkGetTextAtOffsetWithPreformattedText): This test was
+        reporting a trailing '\n' for some reason for a <pre> block, which
+        is plainly wrong since, in order to return that, there should be
+        at least a trailing empty space after that and before the </pre>
+        closing tag. This is fixed now.
+        (testWebkitAtkGetTextAtOffsetWithWrappedLines): Uncommented tests
+        that were previously not passing due to a bug in GailTextUtil.
+
+2013-09-11  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
+
+        Generate more HTML type checks and casting
+        https://bugs.webkit.org/show_bug.cgi?id=121080
+
+        Reviewed by Andreas Kling.
+
+        Clean-up remained functions of HTML elements using auto-generated isFooElement(),
+        and replace toFooElement() with ELEMENT_TYPE_CASTS() macro.
+
+        Besides this patch clean-up unnecessary checks which are being supported by
+        auto-generated isFooElement().
+
+        No new tests, no behavior change.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::supportsFullscreenForNode):
+
+2013-09-04  Zan Dobersek  <zdobersek@igalia.com>
+
+        [GTK] Add support for the Wayland build target
+        https://bugs.webkit.org/show_bug.cgi?id=120627
+
+        Reviewed by Gustavo Noronha Silva.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::createBackingStore): Additionally guard the <gdk/gdkx.h> inclusion and the inclusion and use of
+        WidgetBackingStoreGtkX11 with PLATFORM(X11), ensuring this code is built when also building with X11 target
+        enabled. GDK_WINDOWING_X11 macro can be defined even if the X11 target is disabled.
+
+2013-09-02  Brian Holt  <brian.holt@samsung.com>
+
+        [ATK] Leak: Leaks in testatk.c
+        https://bugs.webkit.org/show_bug.cgi?id=118675
+
+        Reviewed by Mario Sanchez Prada.
+
+        Fixed memory leaks by matching ref calls with unrefs.
+
+        * tests/testatk.c:
+        (testWebkitAtkCaretOffsets):
+        (testWebkitAtkCaretOffsetsAndExtranousWhiteSpaces):
+        (testWebkitAtkGetTextAtOffset):
+        (testWebkitAtkGetTextAtOffsetNewlines):
+        (testWebkitAtkGetTextAtOffsetTextarea):
+        (testWebkitAtkGetTextAtOffsetTextInput):
+        (testWebkitAtkGetTextInParagraphAndBodySimple):
+        (testWebkitAtkGetTextInParagraphAndBodyModerate):
+        (testWebkitAtkTextAttributes):
+        (testWebkitAtkTextSelections):
+        (testWebkitAtkListsOfItems):
+
+2013-08-29  Zan Dobersek  <zdobersek@igalia.com>
+
+        [Automake] libWebCoreDOM.la could use a better name
+        https://bugs.webkit.org/show_bug.cgi?id=120232
+
+        Reviewed by Martin Robinson.
+
+        * GNUmakefile.am: libWebCoreDOM has a new name - libGObjectDOMBindings.
+
+2013-08-29  Arnaud Renevier  <a.renevier@samsung.com>
+
+        [cairo] canvas drawing on itself doesn't work with accelerated canvas
+        https://bugs.webkit.org/show_bug.cgi?id=118808
+
+        Reviewed by Martin Robinson.
+
+        Change cairoImageSurfaceToGdkPixbuf to cairoSurfaceToGdkPixbuf.
+
+        * webkit/webkitfavicondatabase.cpp:
+        (getIconPixbufSynchronously):
+
+2013-08-26  Zan Dobersek  <zdobersek@igalia.com>
+
+        Unreviewed GTK build fix.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::FrameLoaderClient): Remove an assertion that was not removed in r154658.
+
+2013-08-26  Ryosuke Niwa  <rniwa@webkit.org>
+
+        Another GTK+ build fix.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+
+2013-08-26  Ryosuke Niwa  <rniwa@webkit.org>
+
+        GTK+ build fix. Like r154620.
+
+        * webkit/webkitwebframe.cpp:
+        (webkit_web_frame_new):
+
+2013-08-26  Pratik Solanki  <psolanki@apple.com>
+
+        PageGroup::groupSettings() should return a reference
+        https://bugs.webkit.org/show_bug.cgi?id=120319
+
+        Reviewed by Andreas Kling.
+
+        PageGroup::m_groupSettings is never NULL so we can just return a reference from groupSettings().
+
+        * webkit/webkitwebdatabase.cpp:
+        (webkit_set_web_database_directory_path):
+
+2013-08-26  Andreas Kling  <akling@apple.com>
+
+        WebCore: Let Page create the main Frame.
+        <https://webkit.org/b/119964>
+
+        Tweak WebKit1/GTK for changes in WebCore.
+
+        Patch by Zan Dobersek <zdobersek@igalia.com>
+        Reviewed by Gustavo Noronha Silva.
+
+        * WebCoreSupport/FrameLoaderClientGtk.h:
+        (WebKit::FrameLoaderClient::setWebFrame):
+        * webkit/webkitwebframe.cpp:
+        (webkit_web_frame_new):
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_init):
+
+2013-08-24  Darin Adler  <darin@apple.com>
+
+        Frame::tree should return a reference instead of a pointer
+        https://bugs.webkit.org/show_bug.cgi?id=120259
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp:
+        (DumpRenderTreeSupportGtk::getFrameChildren):
+        (DumpRenderTreeSupportGtk::clearMainFrameName):
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::createFrame):
+        * webkit/webkitwebframe.cpp:
+        (webkit_web_frame_get_name):
+        (webkit_web_frame_get_parent):
+        (webkit_web_frame_find_frame):
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_set_highlight_text_matches):
+        Use tree(). instead of tree()->.
+
+2013-08-21  Zan Dobersek  <zdobersek@igalia.com>
+
+        <https://webkit.org/b/119836> [GTK] ChromeClient::paint is susceptible to system time changes
+
+        Reviewed by Martin Robinson.
+
+        Instead of using WTF::currentTime() as the value of the last display occurrence,
+        WTF::monotonicallyIncreasingTime() should be used. The latter is not affected by
+        the changes is the system's time, which can cause a halt in the display updating.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::paint):
+
+2013-08-19  Pratik Solanki  <psolanki@apple.com>
+
+        <https://webkit.org/b/119918> Frame::selection() should return a reference
+
+        Reviewed by Darin Adler.
+
+        m_selection is never NULL so return a reference from Frame::selection(). Also removed some
+        unnecessary null checks and assert diff ts exposed as a result.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::paint):
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp:
+        (DumpRenderTreeSupportGtk::firstRectForCharacterRange):
+        (DumpRenderTreeSupportGtk::selectedRange):
+        (DumpRenderTreeSupportGtk::rectangleForSelection):
+        * WebCoreSupport/EditorClientGtk.cpp:
+        (WebKit::collapseSelection):
+        (WebKit::setSelectionPrimaryClipboardIfNeeded):
+        * webkit/webkitwebframe.cpp:
+        (webkit_web_frame_get_range_for_word_around_caret):
+        * webkit/webkitwebview.cpp:
+        (getLocationForKeyboardGeneratedContextMenu):
+
+2013-08-17  Darin Adler  <darin@apple.com>
+
+        <https://webkit.org/b/119941> Make Page::dragController return a reference
+
+        Reviewed by Andreas Kling.
+
+        * webkit/webkitwebview.cpp:
+        (dragExitedCallback):
+        (webkit_web_view_drag_motion):
+        (webkit_web_view_drag_data_received):
+        (webkit_web_view_drag_drop):
+        Updated call sites.
+
+2013-08-16  Pratik Solanki  <psolanki@apple.com>
+
+        <https://webkit.org/b/119852> Frame::scriptController() should return a reference
+
+        Reviewed by Andreas Kling.
+
+        m_script is never NULL so we can just return a reference. Also remove some pointless null
+        checks as a result of doing this.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::dispatchDidClearWindowObjectInWorld):
+        * webkit/webkitwebframe.cpp:
+        (webkit_web_frame_get_global_context):
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_execute_script):
+
+2013-08-15  Zan Dobersek  <zdobersek@igalia.com>
+
+        Unreviewed GTK build fix for debug configurations.
+
+        * webkit/webkitwebframe.cpp:
+        (WebKit::kit): Remove the bogus assertion for Frame::loader() that now returns a reference.
+
+2013-08-15  Anders Carlsson  <andersca@apple.com>
+
+        <https://webkit.org/b/119859> Frame::loader() should return a reference
+
+        Reviewed by Andreas Kling.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.cpp:
+        (DumpRenderTreeSupportGtk::clearOpener):
+        (DumpRenderTreeSupportGtk::shouldClose):
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::dispatchWillSubmitForm):
+        (WebKit::FrameLoaderClient::committedLoad):
+        (WebKit::FrameLoaderClient::dispatchDecidePolicyForResponse):
+        (WebKit::FrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
+        (WebKit::FrameLoaderClient::dispatchDecidePolicyForNavigationAction):
+        (WebKit::FrameLoaderClient::createFrame):
+        (WebKit::FrameLoaderClient::dispatchDidCommitLoad):
+        * webkit/webkitwebframe.cpp:
+        (webkit_web_frame_finalize):
+        (webkit_web_frame_load_uri):
+        (webkit_web_frame_load_data):
+        (webkit_web_frame_load_request):
+        (webkit_web_frame_stop_loading):
+        (webkit_web_frame_reload):
+        (webkit_web_frame_get_data_source):
+        (webkit_web_frame_get_provisional_data_source):
+        (webkit_web_frame_get_response_mime_type):
+        (webkit_web_frame_get_network_response):
+        (WebKit::kit):
+        * webkit/webkitwebpolicydecision.cpp:
+        (webkit_web_policy_decision_use):
+        (webkit_web_policy_decision_ignore):
+        (webkit_web_policy_decision_download):
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_dispose):
+        (webkit_web_view_reload):
+        (webkit_web_view_reload_bypass_cache):
+        (webkit_web_view_stop_loading):
+        (webkit_web_view_can_show_mime_type):
+        (webkit_web_view_set_custom_encoding):
+        (webkit_web_view_get_custom_encoding):
+        (webkit_web_view_get_subresources):
+
+2013-08-15  Simon Pena  <simon.pena@samsung.com>
+
+        <https://webkit.org/b/119584> [Gtk] URL printing code in DumpRenderTree doesn't match WTR or Mac DRT
+
+        Reviewed by Gustavo Noronha Silva.
+
+        Following a similar approach as in r153977, return a path string
+        that is relative to main frame URL or just file name if the
+        resource is not in the same directory subtree, and replace empty
+        strings with "(null)".
+
+        Update the AuthenticationCallback used in DumpRenderTree so that
+        it receives a WebKitWebResource, and update
+        dispatchDidReceiveAuthenticationChallenge so that it retrieves the
+        WebKitWebResource and passes it to the callback.
+
+        * WebCoreSupport/DumpRenderTreeSupportGtk.h: Update the
+        AuthenticationCallback adding a WebKitWebResource parameter.
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::toString): Define this function earlier so we can use it
+        to get the WebKitWebResource from the identifier in the
+        AuthenticationChallenge.
+        (WebKit::FrameLoaderClient::dispatchDidReceiveAuthenticationChallenge):
+        Use toString to retrieve a WebKitWebResource from the identifier
+        in the authentication challenge, and pass that WebKitWebResource
+        to the AuthenticationCallback.
+
+2013-08-13  Xabier Rodriguez Calvar  <calvaris@igalia.com>
+
+        [GTK] WK does not link properly against libxslt
+        https://bugs.webkit.org/show_bug.cgi?id=119688
+
+        Reviewed by Gustavo Noronha Silva.
+
+        * GNUmakefile.am: Added LIBXSLT_LIBS to the unit tests.
+
+2013-08-13  Alberto Garcia  <berto@igalia.com>
+
+        [GTK] Closing inspector window crashes wk
+        https://bugs.webkit.org/show_bug.cgi?id=110865
+
+        Reviewed by Carlos Garcia Campos.
+
+        The previous fix for the inspector window crash breaks some unit
+        tests. This one goes back to the original code and only moves the
+        actual deletion of priv->corePage to the end of the function.
+
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_dispose):
+
+2013-08-13  Alberto Garcia  <berto@igalia.com>
+
+        [GTK] Closing inspector window crashes wk
+        https://bugs.webkit.org/show_bug.cgi?id=110865
+
+        Reviewed by Carlos Garcia Campos.
+
+        Deleting priv->corePage during the destruction of a webView will
+        trigger the deletion of InspectorFrontendClient. However that
+        object is supposed to handle the webView's destroy signal first in
+        order to do the necessary cleanup.
+
+        The solution is to wait until webkit_web_view_dispose finishes
+        before deleting priv->corePage.
+
+        * webkit/webkitwebview.cpp:
+        (webkit_web_view_dispose):
+
+2013-08-12  Anton Obzhirov  <a.obzhirov@samsung.com>
+
+        [GTK] Don't load GAIL when using GTK 3.2 or greater
+        https://bugs.webkit.org/show_bug.cgi?id=72735
+
+        Reviewed by Carlos Garcia Campos.
+
+        * tests/testatk.c:
+        (initializeTestingFramework):
+        (main):
+
+2013-08-11  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        Unreviewed. Update NEWS and Versions.m4 for 2.1.4 release.
+
+        * NEWS: Update release notes.
+
+2013-08-09  Alexey Proskuryakov  <ap@apple.com>
+
+        REGRESSION (r142755): window.open creates an invisible window when width and height are 0
+        https://bugs.webkit.org/show_bug.cgi?id=119633
+
+        Reviewed by Darin Adler.
+
+        * WebCoreSupport/ChromeClientGtk.cpp: (WebKit::ChromeClient::setWindowRect):
+        Once again, the passed rect cannot be empty.
+
+2013-07-27  Mark Rowe  <mrowe@apple.com>
+
+        Logging should be configurable using human-readable channel names rather than crazy bitmasks
+        <http://webkit.org/b/119031>
+
+        Implement shared logic for initializing logging channels based on human-readable channel names in WTF,
+        and rework the WebCore, WebKit and WebKit2 logging initialization on top of it.
+
+        Logging channels may now be enabled by providing a comma-separated list of channel names, with the special
+        "all" name enabling all channels. Channel names prefixed with a leading "-" will result in the named channel
+        being disabled. For instance, specifying "all,-history,-loading" will result in all logging channels except
+        for history and loading being enabled.
+
+        For OS X developers, this also changes the name of the user defaults used to enable logging. This is done to allow
+        the old user defaults to remain set for those people that need to switch between version of WebKit before and
+        after this change. Where the old user default keys were WebCoreLogLevel, WebKitLogLevel and WebKit2LogLevel,
+        the new user default keys are WebCoreLogging, WebKitLogging and WebKit2Logging.
+
+        For GTK developers, this changes the separator used in the WEBKIT_DEBUG environment variable to a comma for
+        consistency with the other platforms and to enable more code sharing.
+
+        While doing this work I've also taken the opportunity to eliminate the need to touch multiple files when
+        adding a new logging channel. Now only the header in the relevant project needs to be updated.
+
+        Reviewed by Sam Weinig.
+
+        * webkit/webkitglobals.cpp: Switch from WebCore's InitializeLogging.h to Logging.h.
+
+2013-08-02  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        Implement atk_text_get_text_*_offset for WORD
+        https://bugs.webkit.org/show_bug.cgi?id=114871
+
+        Reviewed by Martin Robinson.
+
+        Updated current unit tests and add a new one specific for embedded
+        objects, to ensure we are covering even more cases than before.
+
+        * tests/testatk.c:
+        (testWebkitAtkGetTextAtOffsetWithEmbeddedObjects): New.
+        (main): Added new test to the test suite.
+
+2013-07-30  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        Unreviewed. Update NEWS and Versions.m4 for 2.1.4 release.
+
+        * NEWS: Added release notes for 2.1.4.
+
+2013-07-29  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [ATK] Issues with edge cases when getting offsets for a text range in AtkText
+        https://bugs.webkit.org/show_bug.cgi?id=118908
+
+        Reviewed by Martin Robinson.
+
+        * tests/testatk.c:
+        (runGetTextTests): Updated unit tests to check more cases of
+        calling the atk_text_get_text_*_offset() functions for WORD.
+
+2013-07-25  Andreas Kling  <akling@apple.com>
+
+        ChromeClient::focusedNodeChanged() should be focusedElementChanged().
+        <http://webkit.org/b/119110>
+
+        Reviewed by Anders Carlsson.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::focusedElementChanged):
+        * WebCoreSupport/ChromeClientGtk.h:
+
+2013-07-23  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] GtkAdjustmentWatcher idle source not correctly handled
+        https://bugs.webkit.org/show_bug.cgi?id=119003
+
+        Reviewed by Martin Robinson.
+
+        GtkAdjustmentWatcher::updateAdjustmentsFromScrollbarsLater()
+        creates a new idle source and initializes
+        m_updateAdjustmentCallbackId, which is used everywhere and
+        specially in the destructor to cancel the source when the
+        GtkAdjustmentWatcher is deleted. However,
+        m_updateAdjustmentCallbackId is reset to 0 in
+        GtkAdjustmentWatcher::updateAdjustmentsFromScrollbars() only when
+        not returning early. This method is public and not only called by
+        the updateAdjustmentCallback, which means that a call to
+        updateAdjustmentsFromScrollbars could reset the
+        m_updateAdjustmentCallbackId without actually destroying the source.
+
+        * WebCoreSupport/GtkAdjustmentWatcher.cpp:
+        (WebKit::GtkAdjustmentWatcher::updateAdjustmentsFromScrollbars):
+        Cancel the adjustment later idle if it has been scheduled.
+        (WebKit::GtkAdjustmentWatcher::updateAdjustmentCallback): Reset
+        m_updateAdjustmentCallbackId.
+        * WebCoreSupport/GtkAdjustmentWatcher.h:
+
+2013-07-17  Kangil Han  <kangil.han@samsung.com>
+
+        Use toHTMLMediaElement
+        https://bugs.webkit.org/show_bug.cgi?id=118727
+
+        Reviewed by Ryosuke Niwa.
+
+        To avoid direct use of static_cast, this patch uses toHTMLMediaElement for code cleanup.
+
+        * WebCoreSupport/ChromeClientGtk.cpp:
+        (WebKit::ChromeClient::enterFullscreenForNode):
+        (WebKit::ChromeClient::exitFullscreenForNode):
+        (WebKit::ChromeClient::enterFullScreenForElement):
+        (WebKit::ChromeClient::exitFullScreenForElement):
+
+2013-07-09  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        Unreviewed. Update NEWS and Versions.m4 for 2.1.3 release.
+
+        * NEWS: Added release notes for 2.1.3.
+
+2013-07-08  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] webkit_set_cache_model is not turning the cache off
+        https://bugs.webkit.org/show_bug.cgi?id=118345
+
+        Reviewed by Gustavo Noronha Silva.
+
+        Our documentation says that WEBKIT_CACHE_MODEL_DOCUMENT_VIEWER
+        disables the cache completely, but we are only setting the cache
+        capacities to 0. Cache capacities are not checked when resources
+        are added to the cache, but when it's pruned. To disable the cache
+        we need to call MemoryCache::setDisabled explicitly.
+
+        * webkit/webkitglobals.cpp:
+        (webkit_set_cache_model): Disable the cache when min dead, max
+        dead and total capacities are all set to 0.
+
+2013-07-08  Brian Holt  <brian.holt@samsung.com>
+
+        [GTK] FrameLoaderClient: Refactor naked pointers to use smart pointers
+        https://bugs.webkit.org/show_bug.cgi?id=118417
+
+        Reviewed by Carlos Garcia Campos.
+
+        Use GOwnPtr and GRefPtr where possible.
+
+        * WebCoreSupport/FrameLoaderClientGtk.cpp:
+        (WebKit::FrameLoaderClient::dispatchDidFailLoad):
+
+2013-07-04  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [ATK] Do not expose '\n' for wrapped lines with ATK_TEXT_BOUNDARY_CHAR
+        https://bugs.webkit.org/show_bug.cgi?id=118359
+
+        Reviewed by Carlos Garcia Campos.
+
+        Added new unit test to make sure we exposed line breaks properly.
+
+        * tests/testatk.c:
+        (testWebkitAtkGetTextAtOffsetWithWrappedLines): New unit test.
+        (main): Added new test to the test suite.
+
+2013-07-01  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
+
+        [GTK] Remove unsupported AC backends
+        https://bugs.webkit.org/show_bug.cgi?id=117362
+
+        Reviewed by Martin Robinson.
+
+        * GNUmakefile.am:
+        * WebCoreSupport/AcceleratedCompositingContext.h: removed clutter and cairo
+        additions.
+        * WebCoreSupport/AcceleratedCompositingContextCairo.cpp: Removed.
+        * WebCoreSupport/AcceleratedCompositingContextClutter.cpp: Removed.
+        * webkit/webkitglobals.cpp:
+        (webkitInit): removed clutter-gtk initialization.
+
 2013-07-01  ChangSeok Oh  <changseok.oh@collabora.com>
 
         [GTK] Use PassOwnPtr for the returned value of createBackingStore