Initial checkin of Planet WebKit

        Rubberstamped by Sam.

        * README: Added.
        * config.ini: Added.
        * planet/AUTHORS: Added.
        * planet/INSTALL: Added.
        * planet/LICENCE: Added.
        * planet/NEWS: Added.
        * planet/PKG-INFO: Added.
        * planet/README: Added.
        * planet/THANKS: Added.
        * planet/TODO: Added.
        * planet/examples/atom.xml.tmpl: Added.
        * planet/examples/basic/config.ini: Added.
        * planet/examples/basic/index.html.tmpl: Added.
        * planet/examples/fancy/config.ini: Added.
        * planet/examples/fancy/index.html.tmpl: Added.
        * planet/examples/foafroll.xml.tmpl: Added.
        * planet/examples/opml.xml.tmpl: Added.
        * planet/examples/output/images/edd.png: Added.
        * planet/examples/output/images/evolution.png: Added.
        * planet/examples/output/images/feed-icon-10x10.png: Added.
        * planet/examples/output/images/jdub.png: Added.
        * planet/examples/output/images/keybuk.png: Added.
        * planet/examples/output/images/logo.png: Added.
        * planet/examples/output/images/opml.png: Added.
        * planet/examples/output/images/planet.png: Added.
        * planet/examples/output/images/thom.png: Added.
        * planet/examples/output/planet.css: Added.
        * planet/examples/rss10.xml.tmpl: Added.
        * planet/examples/rss20.xml.tmpl: Added.
        * planet/planet-cache.py: Added.
        * planet/planet.py: Added.
        * planet/planet/__init__.py: Added.
        * planet/planet/atomstyler.py: Added.
        * planet/planet/cache.py: Added.
        * planet/planet/compat_logging/__init__.py: Added.
        * planet/planet/compat_logging/config.py: Added.
        * planet/planet/compat_logging/handlers.py: Added.
        * planet/planet/feedparser.py: Added.
        * planet/planet/htmltmpl.py: Added.
        * planet/planet/sanitize.py: Added.
        * planet/planet/tests/__init__.py: Added.
        * planet/planet/tests/data/simple.tmpl: Added.
        * planet/planet/tests/data/simple2.tmpl: Added.
        * planet/planet/tests/test_channel.py: Added.
        * planet/planet/tests/test_main.py: Added.
        * planet/planet/tests/test_sanitize.py: Added.
        * planet/planet/tests/test_sub.py: Added.
        * planet/planet/timeoutsocket.py: Added.
        * planet/runtests.py: Added.
        * planet/setup.py: Added.
        * templates/atom.xml.tmpl: Added.
        * templates/foafroll.xml.tmpl: Added.
        * templates/index.html.tmpl: Added.
        * templates/opml.xml.tmpl: Added.
        * templates/rss10.xml.tmpl: Added.
        * templates/rss20.xml.tmpl: Added.
        * wwwroot/images/feed-icon-10x10.png: Added.
        * wwwroot/images/planet.png: Added.
        * wwwroot/planet.css: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28268 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Maciej.

        Clean up a couple of typos from Maciej's last patch.

        * coding/commit-review-policy.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28267 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    <rdar://problem/5621373> Added a way for the client to clear undo/redo
        operations from a WebView.

        Reviewed by Steve.

        * Interfaces/IWebViewPrivate.idl:
        * WebView.cpp:
        (WebView::clearUndoRedoOperations):
        * WebView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28266 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Bertrand.

        New commit and review policy.

        * coding/commit-review-policy.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28265 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Add platform/text project group.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28264 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-30  Eric Seidel  <eric@webkit.org>

        No review, build fix only.

        Fix uninitialized var warnings in release build.

        * JavaScriptCore.xcodeproj/project.pbxproj:
        * pcre/pcre_compile.cpp:
        (compile_regex):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28263 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Remove extra </Filter> tag which prevented the vcproj from being parsed correctly.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28262 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff Garen

        Added another symbol for WebDatabaseManager clients

        * WebKit.exp:
        added .objc_class_name_WebSecurityOrigin

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28261 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-30  Eric Seidel  <eric@webkit.org>

        Reviewed by darin.

        PCRE crashes under GuardMalloc
        http://bugs.webkit.org/show_bug.cgi?id=16127
        check against patternEnd to make sure we don't walk off the end of the string

        * pcre/pcre_compile.cpp:
        (compile_branch):
        (calculateCompiledPatternLengthAndFlags):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28260 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed. Try to fix qt/gtk build.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28259 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Eric.

Final integration of ksvg2 in WebKit. Moving ksvg2/ to svg/ and killing all it's subdirectories,
by moving the files into approriate locations (css, rendering, ...) - as dicsussed on the
mailing list and on IRC with David/Darin & Eric.

Updated all build files - though I can't guarantee it builds on non-mac. Need buildbot to verify.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28258 268f45cc-cd09-0410-ab3c-d52691b4dbfc

wx build fix. Add WebCore/platform/graphics/wx to includes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28257 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff

        Further hookup of Database API
        With this change, any database operation that writes to the database (UPDATE, INSERT, CREATE, etc) will
        result in a didModifyDatabase notification when the transaction is successfully committed

        * storage/DatabaseAuthorizer.cpp: Global rename of "m_lastActionIncreasedSize" to "m_lastActionChangedDatabase"
        (WebCore::DatabaseAuthorizer::reset):
        (WebCore::DatabaseAuthorizer::createTable):
        (WebCore::DatabaseAuthorizer::allowAlterTable):
        (WebCore::DatabaseAuthorizer::createIndex):
        (WebCore::DatabaseAuthorizer::createTrigger):
        (WebCore::DatabaseAuthorizer::createVTable):
        (WebCore::DatabaseAuthorizer::allowInsert):
        (WebCore::DatabaseAuthorizer::allowUpdate):
        * storage/DatabaseAuthorizer.h:
        (WebCore::DatabaseAuthorizer::lastActionChangedDatabase):

        * storage/DatabaseTracker.cpp:
        (WebCore::DatabaseTracker::notifyDatabaseChanged): Added.  Allow external entities (SQLTransaction) to be able
          to request a didModifyDatabase notification to the delegate
        * storage/DatabaseTracker.h:

        * storage/SQLTransaction.cpp: Added m_modifiedDatabase flag which tracks if any statement over the course of
          a transaction resulted in a change to the database.
        (WebCore::SQLTransaction::SQLTransaction):
        (WebCore::SQLTransaction::runCurrentStatement): Set m_modifiedDatabase to true if the authorizer reports a mutation
          statement occurred
        (WebCore::SQLTransaction::postflightAndCommit): If the transaction successfully commits, call DatabaseTracker::notifyDatabaseChanged
        (WebCore::SQLTransaction::cleanupAfterTransactionErrorCallback): Ditto
        * storage/SQLTransaction.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28256 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff

        Add isEqual operator to WebSecurityOrigin

        * Storage/WebSecurityOrigin.mm:
        (-[WebSecurityOrigin isEqual:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28255 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Hopefully the final build fix

        * Drosera/win/Drosera.vcproj/Drosera.vcproj: Link against WTF.lib to
        pull in WTF's assertion/logging functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28254 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Another build fix

        * Drosera/win/Drosera.vcproj/Drosera.vcproj: Use the right suffix for
        WebKit.lib.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28253 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Debug build fix

        * Drosera/win/Drosera.vcproj/Drosera.vcproj: Added a Debug_Internal
        configuration.
        * Drosera/win/Drosera.vcproj/debug.vsprops: Updated to match other
        projects.
        * Drosera/win/Drosera.vcproj/release.vsprops: Ditto.
        * Drosera/win/Drosera.vcproj/debug_internal.vsprops: Copied from
        WebKitTools/DumpRenderTree/win/debug_internal.vsprops.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28252 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by David.

Next round of moves - platform reorganization.
Move text/font/glyph related files from the individual port directories
into platform/text/<port> or platform/graphics/<port>.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28251 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Other wx port fixes needed to restore the build bot.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28250 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        Fix drawSelectionRect to actually draw the selection rect.

        * DumpRenderTree/cg/PixelDumpSupportCG.cpp:
        (drawSelectionRect):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28249 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the build with DATABASE=0.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28248 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Copy some cross-platform results into platform/win to avoid picking up the platform/mac versions

        * platform/win/fast/dom/Window/window-resize-expected.txt: Copied from LayoutTests/fast/dom/Window/window-resize-expected.txt.
        * platform/win/fast/dom/wrapper-classes-expected.txt: Copied from LayoutTests/fast/dom/wrapper-classes-expected.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28247 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Darin.

Fix misplaced file. The name "FontStyle" is not correct though - it probably
needs yet another move - but this time inside platform/graphics - as soon
as we agree on a better name.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28246 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin

        Tweaks to newly-declared NSString * constants to make them usable from clients

        * Storage/WebDatabaseManagerPrivate.h:
        * Storage/WebDatabaseManager.mm:
        removed "const" from new NSNotification names and userInfo keys; these generate compiler warnings when used

        * WebKit.exp:
        export new NSNotification names and userInfo keys so clients can use them

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28245 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Adam Roben.

        - fix http://bugs.webkit.org/show_bug.cgi?id=16207
          JavaScript regular expressions should match UTF-16 code units rather than characters

        SunSpider says this is 5.5% faster on the regexp test, 0.4% faste overall.

        Test: fast/js/regexp-non-bmp.html

        Renamed ANY_CHAR to NOT_NEWLINE to more-accurately reflect its meaning.

        * pcre/pcre_compile.cpp:
        (compile_branch): Removed calls to the UTF-16 character accessor functions, replacing
        them with simple pointer dereferences in some cases, and no code at all in others.
        (calculateCompiledPatternLengthAndFlags): Ditto.

        * pcre/pcre_exec.cpp:
        (match): Fixed indentation of some case labels (including all the BEGIN_OPCODE).
        Removed calls to the UTF-16 character accessor functions, replacing them with simple
        pointer dereferences in some cases, and no code at all in others. Also removed some
        explicit UTF-16 support code in a few cases. Removed the unneeded "UTF-8" code path
        in the ANY_CHAR repeat code, and in another case, eliminated the code to check against
        end_subject in because it is already done outside the loop.
        (jsRegExpExecute):

        * pcre/pcre_internal.h: Removed all the UTF-16 helper functions.

LayoutTests:

        Reviewed by Adam Roben.

        - test for http://bugs.webkit.org/show_bug.cgi?id=16207
          JavaScript regular expressions should match UTF-16 code units rather than characters

        * fast/js/regexp-non-bmp-expected.txt: Added.
        * fast/js/regexp-non-bmp.html: Added.
        * fast/js/resources/regexp-non-bmp.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28243 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * Scripts/do-webcore-rename: WildFox already did the TextStyle -> FontStyle one.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28242 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix after platform/text changes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28241 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Sam.

Rename "FontStyle FontStyle(...)" to "FontStyle fontStyle(...)".
I scripted a bit too much while doing the TextStyle -> FontStyle change.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28240 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-30  Alp Toker  <alp@atoker.com>

        GTK+ build fix following the recent text reorganization.

        * WebCore.pro:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28239 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-29  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Adam.

        Part of <rdar://problem/5506522> Seed: cannot type into text box in
        Facebook, don't see plugin content (Adobe Flash)

        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginViewWin::handleMouseEvent): Focus the plugin's parent
        frame

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28238 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * Scripts/do-webcore-rename: Get ready for some future renaming.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28237 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        - Removed some unnecessary functions and changed an unused return type.

        * Drosera/win/DebuggerClient.cpp:
        (registerConsoleClass):
        * Drosera/win/DebuggerClient.h:
        * Drosera/win/Drosera.cpp:
        (registerDroseraClass):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28236 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed. Trying to fix gtk/qt/wx build. (I expected problems because of the moves)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28235 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Brady.

Started platform reorganization (platform/text addition).
Moving string/text/font related classes into new places - as discussed on webkit-dev.

In a follow-up commit I'll move classes out of platform/{mac,win,...} into
platform/text/{mac,win,...}, and also do the ksvg2/ -> svg/ move.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28234 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin.

        Fix for http://bugs.webkit.org/show_bug.cgi?id=16072
        <rdar://problem/5463821> REGRESSION: Mousemove event isn't fired when mouse moves over scrollbar with pressed button (16072)

        Test: fast/events/mousemove-after-drag-over-scrollbar.html

        * page/EventHandler.cpp:
        (WebCore::EventHandler::handleMouseMoveEvent): If the mouse is pressed, and we've moved over a new scrollbar, then we can clear m_lastScrollbarUnderMouse.
         We should only need to show the hover effect when the mouse isn't pressed.

LayoutTests:

        Reviewed by Darin.

        Test for http://bugs.webkit.org/show_bug.cgi?id=16072
        <rdar://problem/5463821> REGRESSION: Mousemove event isn't fired when mouse moves over scrollbar with pressed button

        * fast/events/mousemove-after-drag-over-scrollbar-expected.txt: Added.
        * fast/events/mousemove-after-drag-over-scrollbar.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28233 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * platform/win/Skipped: Added editing/selection/5241148.html, which crashes for me.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28232 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-28  Alp Toker  <alp@atoker.com>

        Reviewed by Timothy Hatcher.

        http://bugs.webkit.org/show_bug.cgi?id=16174
        [GTK] Use "URI" not "URL" in public API

        Replace use of the term "URL" with "URI" in public headers,
        documentation and some internal code to match GLib/GTK+ convention.

        This is now mentioned in the API guidelines:
          http://trac.webkit.org/projects/webkit/wiki/HackingGtk

        API/ABI-breaking change.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28231 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Self destruct the clients like the ohter ports do

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28230 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-30  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Fix layout test regressions caused by r28186
        http://bugs.webkit.org/show_bug.cgi?id=16195
        change first_byte and req_byte back to shorts instead of chars
        (I think PCRE stuffs information in the high bits)

        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28229 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Tim

        Part of <rdar://problem/5556377> - Need to accurately track and enforce database quota

        SQLite has the ability to enforce the page usage for databases.  This gives us bulletproof
        enforcement of the database size limit with reasonable granularity (within 1023 bytes by
        default on OS X with SQLite 3.4.0)

        This also involved enhancing the ability to install/remove the authorizer to run the PRAGMA
        statements involved

        Note this patch does not actually use the new maximumSize() functionality of the
        SQLiteDatabase class - that can be reviewed/landed seperately

        * platform/sql/SQLiteDatabase.cpp:
        (WebCore::SQLiteDatabase::SQLiteDatabase):
        (WebCore::SQLiteDatabase::maximumSize): Return the current maximum size in bytes
        (WebCore::SQLiteDatabase::setMaximumSize): Based on m_pageSize, set the maximum page count
          to enforce the maximum size in bytes
        (WebCore::SQLiteDatabase::pageSize): Fetch m_pageSize if it hasn't been fetched, or return
          the cached value
        (WebCore::SQLiteDatabase::setAuthorizer): Use enableAuthorizer for the sqlite_* level
          authorizer setup
        (WebCore::SQLiteDatabase::enableAuthorizer): Install or remove the sqlite authorizer handler
        * platform/sql/SQLiteDatabase.h: Add the m_pageSize member so we only have to fetch it once

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28228 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Fix fast/events/mousedown_in_scrollbar.html
* This test used to crash in PlatformScrollbarQt::mouse*Event due parent() returning 0.
* Extend the current check that the scrollbar in the mev can be go away.
  This happens if we destroy a RenderLayer as the result of a click. In the
  case of Qt we can have this scrollbar in m_lastScrollbarUnderMouse. The RenderLayer
  is the ScrollbarClient, m_lastScrollbarUnderMouse is a RefPtr and keeps the
  last reference. When calling into this scrollbar the m_client pointer is a dangling pointer.
  We extend the check to set m_lastScrollbarUnderMouse back to zero when it was the
  scrollbar in the old mev and is gone in the new one.
* This happens only on the Qt port because we send more/different move events.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28227 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Mitz.

        Fix aspect ratio calculation of video.

        Video should always be treated as a replaced element in size calculation.

        Test: media/video-aspect-ratio.html

        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::calcWidth):
        (WebCore::RenderBox::calcHeight):
        * rendering/RenderBox.h:
        (WebCore::RenderBox::shouldCalculateSizeAsReplaced):
        * rendering/RenderMedia.h:
        (WebCore::RenderMedia::shouldCalculateSizeAsReplaced):

LayoutTests:

        Reviewed by Mitz.

        Test for video aspect ratio calculation.

        * media/video-aspect-ratio.html: Added.
        * platform/mac/media: Added.
        * platform/mac/media/video-aspect-ratio-expected.checksum: Added.
        * platform/mac/media/video-aspect-ratio-expected.png: Added.
        * platform/mac/media/video-aspect-ratio-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28226 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver.

        Additional fix for <rdar://problem/5592988> / http://bugs.webkit.org/show_bug.cgi?id=15936
        - More closely match IE's policy for frame navigation.

        * bindings/js/kjs_window.cpp:
        (KJS::WindowProtoFuncOpen::callAsFunction):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::shouldAllowNavigation):
        * page/FrameTree.cpp:
        (WebCore::FrameTree::top):
        * page/FrameTree.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28225 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben and John Sullivan.

        - make the recent searches popup extend only along the straight part of
          the search field (unless it needs to be wider).

        * platform/PopupMenuClient.h:
          Added clientInsetLeft() and clientInsetRight() for clients to
          specify desired horizontal insets for the menu.
        * platform/win/PopupMenuWin.cpp:
        (WebCore::PopupMenu::calculatePositionAndSize):
        (WebCore::PopupMenu::paint):
        * rendering/RenderMenuList.cpp:
        (WebCore::RenderMenuList::clientInsetLeft): Added. Returns 0.
        (WebCore::RenderMenuList::clientInsetRight): Ditto.
        * rendering/RenderMenuList.h:
        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::clientInsetLeft): Added. Returns half the
        control's height, which is the radius of the cap on the left.
        (WebCore::RenderTextControl::clientInsetRight): Added.
        * rendering/RenderTextControl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28224 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make the JS collector work with multiple threads

Reviewed by Maciej and Darin.

Under heavy contention it was possible the GC to suspend other
threads inside the pthread spinlock, which could lead to the GC
thread blocking on the pthread spinlock itself.

We now determine and store each thread's stack base when it is
registered, thus removing the need for any calls to pthread_get_stackaddr_np
that needed the pthread spinlock.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28223 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKitTools:

        Reviewed by Sam.

        <rdar://problem/5230478>
        FrameLoadDelegate callbacks are not dumped in DRT.

        * DumpRenderTree/win/DumpRenderTree.cpp:
        (shouldLogFrameLoadDelegates):
        (runTest):
        (main):
        * DumpRenderTree/win/FrameLoaderDelegate.cpp:
        (BSTRtoString):
        (descriptionSuitableForTestResult):
        (FrameLoadDelegate::QueryInterface):
        (FrameLoadDelegate::didStartProvisionalLoadForFrame):
        (FrameLoadDelegate::didFailProvisionalLoadWithError):
        (FrameLoadDelegate::didCommitLoadForFrame):
        (FrameLoadDelegate::didFinishLoadForFrame):
        (FrameLoadDelegate::willCloseFrame):
        (FrameLoadDelegate::didClearWindowObject):
        (FrameLoadDelegate::didFinishDocumentLoadForFrame):
        (FrameLoadDelegate::didHandleOnloadEventsForFrame):
        * DumpRenderTree/win/FrameLoaderDelegate.h:
        (FrameLoadDelegate::windowScriptObjectAvailable):
        (FrameLoadDelegate::didFirstLayoutInFrame):

LayoutTests:

        Reviewed by Sam.

        <rdar://problem/5230478>
        FrameLoadDelegate callbacks are not dumped in DRT.

        Remove now passing tests.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28222 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam.

        Add an IWebFrameLoadDelegate2 interface with a didClearWindowObject method
        that passes in the frame whose window has been cleared. This matches a newly
        added delegate method on the Mac.

        * Interfaces/IWebFrameLoadDelegate.idl:
        * WebFrame.cpp:
        (WebFrame::windowObjectCleared):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28221 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        * platform/win/Skipped: Removed two Mac-specific tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28220 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Move incorrect, Mac-specific results of fast/dom/Window/window-resize.html into platform/mac and replace them with non-Mac results

        The incorrect results are tracked by <rdar://5563578>.

        Reviewed by Adam.

        * fast/dom/Window/window-resize-expected.txt: Replaced.
        * platform/mac/fast/dom/Window/window-resize-expected.txt: Copied from fast/dom/Window/window-resize-expected.txt.
        * platform/win/Skipped: Removed window-resize tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28219 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKitTools:

        Reviewed by Anders.

        Add database quota UIDelegates methods to DRT UIDelegate.

        * DumpRenderTree/mac/UIDelegate.mm:
        (-[UIDelegate webView:quotaForSecurityOrigin:toCreateDatabase:withEstimatedSize:]):
        (-[UIDelegate webView:quotaForSecurityOrigin:fromProposedQuota:database:]):

LayoutTests:

        Reviewed by Anders.

        Revert r28213 which added an error callback that was not needed and changed
        the results.

        * storage/transaction_callback_exception_crash.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28218 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        * platform/win/Skipped: Removed a passing test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28217 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Group failures related to dumpResourceLoadCallbacks being unimplemented

        Reviewed by Adam.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28216 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-30  Alp Toker  <alp@atoker.com>

        Reviewed by Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=16173
        Licensing change

        Change license from BSD to LGPL.

        This is being done to encourage vendors to contribute their fixes back
        upstream.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28215 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Group failures related to using Cygwin paths

        Reviewed by Adam.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28214 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Not reviewed.

        Fixed a hang in the layout tests.
        My checkin earlier today broke things such that the transaction call errors out instead of succeeds.
        Therefore LayoutTestController is never notified that the test is done.

        This changes modifies the test such that the controller is notified of completion even if
        there's a transaction error.

        The net effect of this fix is to transform the hang to a single layout test failure, so people
        and the bots don't have their run-webkit-tests hang while I try to fix the real regression.

        * storage/transaction_callback_exception_crash.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28213 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Adele.

        - fix http://bugs.webkit.org/show_bug.cgi?id=16191
          REGRESSION: cannot tab to radio input after setting checked to false

        Test: fast/forms/input-radio-checked-tab.html

        * html/HTMLFormElement.cpp:
        (WebCore::HTMLFormElement::CheckedRadioButtons::addButton): Fix this code
        so that it doesn't call setChecked(false) until after the map has been updated.
        Otherwise, we can end up deallocating the map before manipulating it. As long
        as I was changing the function, I decided to make it do only a single hash
        table lookup.

        * html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::setChecked):
        Remove the button from the radio buttons set before changing the checked
        state. This matches the idiom used elsewhere and fixes the problem where
        setting checked to false would not remove it from the set.

LayoutTests:

        Reviewed by Adele.

        - test for http://bugs.webkit.org/show_bug.cgi?id=16191
          REGRESSION: cannot tab to radio input after setting checked to false

        * fast/forms/input-radio-checked-tab-expected.txt: Added.
        * fast/forms/input-radio-checked-tab.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28212 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Move Mac-specific results of fast/dom/wrapper-classes.html into platform/mac and replace them with non-Mac results

        Reviewed by Adam.

        * fast/dom/wrapper-classes-expected.txt: Replaced.
        * platform/mac/fast/dom/wrapper-classes-expected.txt: Copied from fast/dom/wrapper-classes-expected.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28211 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Group drag-n-drop related failures under one bug

        We don't know the specific causes of these failures, but we believe
        they're due to bugs in DRT's drag-n-drop implementation.

        Reviewed by Adam.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28210 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit/win:

        Reviewed by Sam.

        - <rdar://5618976> Drosera: should listen for the WebScriptDebugServer
        dying and vice versa.

        * Interfaces/IWebScriptDebugListener.idl: Added the new function to the
        interface.
        * WebKit.vcproj/WebKit.def: Added the new function to the def file so
        Safari can tell WebKit when the server is dying.
        * WebKit.vcproj/WebKit_debug.def: Ditto.
        * WebKitDLL.cpp: Added the new function for Safari to call.
        (RunAsLocalServer):
        (LocalServerDidDie):
        * WebScriptDebugServer.cpp:
        (WebScriptDebugServer::WebScriptDebugServer): Removed unused member.
        (WebScriptDebugServer::sharedWebScriptDebugServer): Ditto, and added new
        member to prevent adding or removing listeners when the server is dying.
        (WebScriptDebugServer::addListener): Don't let new listeners be added
        when the server is dying.
        (WebScriptDebugServer::removeListener): Don't let listeners remove
        themselves when the server is dying.  The server will remove them all
        anyways.
        (WebScriptDebugServer::serverDidDie): Notify all listeners that the
        server is dying and remove all listeners.
        * WebScriptDebugServer.h: Added the new method and removed an unused
        member.

WebKitTools:

        Reviewed by Sam.

        - <rdar://5618976> Drosera: should listen for the WebScriptDebugServer
        dying and vice versa.
        - This fix will allow Drosera and Safari to reconnect if either of them
        closes correctly, but does not fix the case where one of them dies
        silently.

        * Drosera/win/ServerConnection.cpp:
        (ServerConnection::attemptToCreateServerConnection): Added a safety
        check, because it's possilbe to try to connect to a server that's dying.
        (ServerConnection::serverDidDie): Implemented. This resets Drosera when
        the server has died.
        * Drosera/win/ServerConnection.h: Added the new function and removed
        an unnecessary member.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28209 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        * platform/win/Skipped: Removed another passing test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28208 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        * platform/win/Skipped: Recategorized one failing test and removed a
        passing test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28207 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        <rdar://problem/5620093> Brief hang when Safari caches page text of XML document

        * editing/TextIterator.cpp:
        (WebCore::TextIterator::representNodeOffsetZero): When we haven't been emitting any
        characters, shouldRepresentNodeOffsetZero() can create VisiblePositions, which is
        expensive.  So, we perform the inexpensive checks on m_node to see if it necessitates
        emitting a character first and will early return before encountering
        shouldRepresentNodeOffsetZero()s worse case behavior.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * DumpRenderTree/win/DumpRenderTree.vcproj:
        Add shlwapi.lib to all configurations.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Removing <rdar://5507350> from skipped list because it's not failing.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Alexey.

        Move Mac-specific tests to platform/mac.

        * fast/events/objc-event-api-expected.txt: Removed.
        * fast/events/objc-event-api.html: Removed.
        * fast/events/objc-keyboard-event-creation-expected.txt: Removed.
        * fast/events/objc-keyboard-event-creation.html: Removed.
        * platform/mac/fast/events/objc-event-api-expected.txt: Copied from fast/events/objc-event-api-expected.txt.
        * platform/mac/fast/events/objc-event-api.html: Copied from fast/events/objc-event-api.html.
        * platform/mac/fast/events/objc-keyboard-event-creation-expected.txt: Copied from fast/events/objc-keyboard-event-creation-expected.txt.
        * platform/mac/fast/events/objc-keyboard-event-creation.html: Copied from fast/events/objc-keyboard-event-creation.html.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by John.

        Rename WebKitShrinksStandaloneImagesToFitKey to WebKitShrinksStandaloneImagesToFit.

        This is safe to do because the preference is off by default and Safari 3, which is the only client that turns it on,
        is using the setter and not messing around with NSUserDefaults.

        * WebView/WebPreferenceKeysPrivate.h:
        * WebView/WebPreferences.m:
        (+[WebPreferences initialize]):
        (-[WebPreferences shrinksStandaloneImagesToFit]):
        (-[WebPreferences setShrinksStandaloneImagesToFit:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Windows build fix

        Removed some unreachable code (ironically, the code was some
        ASSERT_NOT_REACHED()s).

        * pcre/pcre_compile.cpp:
        (compile_branch):
        * pcre/pcre_exec.cpp:
        (match):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28201 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam.

        Don't install event handlers or set the zoom cursor if auto-shrink has been disabled.

        * loader/ImageDocument.cpp:
        (WebCore::ImageDocument::createDocumentStructure):
        (WebCore::ImageDocument::imageChanged):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28200 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Removing some fixed tests

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28199 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Build fix that works...?

        * WebChromeClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - Build fix.  Added additional includes for VS Express to the Release
        project.

        * Drosera/win/Drosera.vcproj/Drosera.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28197 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Better build fix for Gtk

        * WebCoreSupport/ChromeClientGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Build fix

        * WebChromeClient.cpp:
        (WebChromeClient::requestQuotaIncreaseForNewDatabase):
        (WebChromeClient::requestQuotaIncreaseForDatabaseOperation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28195 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKitTools:

        Reviewed by Adam.

        <rdar://problem/5132005>
        setUserStyleSheetEnabled is not fully implemented in Windows DRT.

        * DumpRenderTree/win/DumpRenderTree.vcproj:
        Add shlwapi.lib

        * DumpRenderTree/win/LayoutTestControllerWin.cpp:
        (LayoutTestController::setUserStyleSheetEnabled):
        Implement this.

        (appendComponentToPath):
        New method which wraps the Win32 API PathAppend.

        (followShortcuts):
        New method which checks if a file points to a shortcut and
        follows the shortcut.

        (resolveCygwinPath):
        New method that takes a cygwin unix-style path and returns the Win32 path.

        (cfStringRefToWString):

        (LayoutTestController::setUserStyleSheetLocation):
        Implement this.

LayoutTests:

        Reviewed by Adam.

        <rdar://problem/5132005>
        setUserStyleSheetEnabled is not fully implemented in Windows DRT.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28194 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff.

        Fix for <rdar://problem/5620151> Disable JavaScript for SVG as
        image for <img> and CSS images

        * platform/graphics/svg/SVGImage.cpp:
        (WebCore::SVGImage::dataChanged): When we create the empty clients,
        explicitly disable JavaScript in the settings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28193 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Anders

        Support for <rdar://problem/5556381> and <rdar://problem/5556379>

        The Database feature in the engine needs to support delegate calls for policy decisions and
        also provide notifications when origins and databases change their state.

        This patch also polishes off the missing features of the management API

        * page/Chrome.cpp: Implement the two UIDelegate methods
        (WebCore::Chrome::requestQuotaIncreaseForNewDatabase): Ask for more space to create a new database if it won't fit
        (WebCore::Chrome::requestQuotaIncreaseForDatabaseOperation): Ask for more space to complete an in-progress operation
        * page/Chrome.h:
        * page/ChromeClient.h:
        * platform/graphics/svg/SVGImageEmptyClients.h:
        (WebCore::SVGEmptyChromeClient::requestQuotaIncreaseForNewDatabase):
        (WebCore::SVGEmptyChromeClient::requestQuotaIncreaseForDatabaseOperation):

        * storage/Database.cpp:
        (WebCore::Database::openDatabase): Check to see if this database can be opened - the quota must be high
          enough and if it isn't, the UIDelegate should have a change to bump the quota.
          Also update the UI-relevant details (display name and estimated size) upon successful opening of the database

        * storage/DatabaseDetails.h: Remove the version parameter as it is a programatic detail of a
          site database and is not important to API clients
        (WebCore::DatabaseDetails::DatabaseDetails):
        (WebCore::DatabaseDetails::isValid):
        (WebCore::DatabaseDetails::name):

        * storage/DatabaseTracker.cpp:
        (WebCore::DatabaseTracker::openTrackerDatabase): Tweaked the schema here - there was a horrible bug with the old schema that would
          prevent you from having two databases of the same name from two different origins.  Also simplify the origin-management schema
        (WebCore::DatabaseTracker::canEstablishDatabase): Added.  Does some estimated size vs quota checks, and asks the UI delegate for
          more space if necessary
        (WebCore::DatabaseTracker::hasEntryForOrigin):
        (WebCore::DatabaseTracker::establishEntryForOrigin): Establishes a tracker entry for the given origin with the current default quota
          Also notifies the client of the new origin
        (WebCore::DatabaseTracker::setDatabaseDetails): Update the display name and estimated size for the given database
        (WebCore::DatabaseTracker::fullPathForDatabase): Tweak to add the ability to get the path without creating it - for management purposes
        (WebCore::DatabaseTracker::populateOrigins): Populate origins from the Origins table instead of the Databases table
        (WebCore::DatabaseTracker::origins):
        (WebCore::DatabaseTracker::detailsForNameAndOrigin): For API management
        (WebCore::DatabaseTracker::usageForDatabase):
        (WebCore::DatabaseTracker::usageForOrigin):
        (WebCore::DatabaseTracker::quotaForOrigin):
        (WebCore::DatabaseTracker::setQuota): Notify the client
        (WebCore::DatabaseTracker::addDatabase): Notify the client
        (WebCore::DatabaseTracker::deleteAllDatabases):
        (WebCore::DatabaseTracker::deleteDatabasesWithOrigin):
        (WebCore::DatabaseTracker::deleteDatabase):
        (WebCore::DatabaseTracker::deleteDatabaseFile):
        * storage/DatabaseTracker.h:

WebKit/gtk:

        Keep it building with new client method

        * WebCoreSupport/ChromeClientGtk.cpp:
        (WebKit::ChromeClient::requestQuotaIncreaseForNewDatabase):
        (WebKit::ChromeClient::requestQuotaIncreaseForDatabaseOperation):
        * WebCoreSupport/ChromeClientGtk.h:

WebKit/mac:

        Reviewed by Anders

        Support for <rdar://problem/5556381> and <rdar://problem/5556379>

        Hook up UI Delegate calls for the database engine feature and other small tweaks

        * Storage/WebDatabaseManager.mm:
        (-[WebDatabaseManager detailsForDatabase:withOrigin:]): Renamed databaseName parameter to databaseIdentifier for clarity
        (-[WebDatabaseManager deleteDatabase:withOrigin:]): Renamed databaseName parameter to databaseIdentifier for clarity
        * Storage/WebDatabaseManagerPrivate.h:

        * Storage/WebDatabaseTrackerClient.h:
        * Storage/WebDatabaseTrackerClient.mm:
        (WebDatabaseTrackerClient::dispatchDidModifyDatabase): Renamed databaseName parameter to databaseIdentifier for clarity

        * WebCoreSupport/WebChromeClient.h:
        * WebCoreSupport/WebChromeClient.mm:
        (WebChromeClient::requestQuotaIncreaseForNewDatabase): Call through to the UI Delegate
        (WebChromeClient::requestQuotaIncreaseForDatabaseOperation): Ditto

        * WebView/WebUIDelegatePrivate.h: Added the two UI Delegate methods

        * WebView/WebView.mm:
        (CallDelegateReturningUnsignedLongLong):
        (CallUIDelegateReturningUnsignedLongLong):
        * WebView/WebViewInternal.h:

WebKit/qt:

        Keep it building with new client method

        * WebCoreSupport/ChromeClientQt.cpp:
        (WebCore::ChromeClientQt::requestQuotaIncreaseForNewDatabase):
        (WebCore::ChromeClientQt::requestQuotaIncreaseForDatabaseOperation):
        * WebCoreSupport/ChromeClientQt.h:

WebKit/win:

        Keep it building with new client method

        * WebChromeClient.cpp:
        (ChromeClient::requestQuotaIncreaseForNewDatabase):
        (ChromeClient::requestQuotaIncreaseForDatabaseOperation):
        * WebChromeClient.h:

WebKit/wx:

        Keep it building with new client method

        * WebKitSupport/ChromeClientWx.cpp:
        (WebCore::ChromeClient::requestQuotaIncreaseForNewDatabase):
        (WebCore::ChromeClient::requestQuotaIncreaseForDatabaseOperation):
        * WebKitSupport/ChromeClientWx.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28192 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        Fixed <rdar://5133828> fast/frames/iframe-window-focus.html output is lowercase

        * DumpRenderTree/win/EventSender.cpp:
        (keyDownCallback):
        uppercase letters were being sent as lowercase letters without the shift key down.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28191 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix typo

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28190 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - Windows build fix. VS express needs some love.

        * Drosera/win/Drosera.vcproj/Drosera.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28189 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Beth Dakin and Darin Adler.

        - manual test for <rdar://problem/5346394> Contextmenu event doesn't fire on body element inside frame

        * manual-tests/context-click-unfocused-frame.html: Added.

WebKit/win:

        Reviewed by Beth Dakin and Darin Adler.

        - fix <rdar://problem/5346394> Contextmenu event doesn't fire on body element inside frame

        * WebView.cpp:
        (WebView::handleContextMenuEvent): Send the context menu event to the
        frame containing the node that was hit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28188 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam.

        - Added recursive runloop guards.

        * DefaultDelegates/WebScriptDebugServer.m:
        (-[WebScriptDebugServer suspendProcessIfPaused]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28187 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-29  Eric Seidel  <eric@webkit.org>

        Reviewed by Mark Rowe.

        Fix for --guard crash of fast/js/regexp-charclass-crash introduced by r28151.

        * pcre/pcre_compile.cpp:
        (is_anchored):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28186 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Gtk build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28185 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-29  Eric Seidel  <eric@webkit.org>

        Reviewed by Mark Rowe.

        Attempt to fix non-mac builds after PCRE cleanup.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
        * JavaScriptCoreSources.bkl:
        * pcre/pcre.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28184 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Sam.

        Give consistent naming to the RegExp options/compile flags

        * pcre/pcre_compile.cpp:
        (compile_branch):
        (is_anchored):
        (find_firstassertedchar):
        (printCompiledRegExp):
        (jsRegExpCompile):
        * pcre/pcre_exec.cpp:
        (jsRegExpExecute):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28183 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Sam.

        Give consistent naming to the RegExp options/compile flags

        * pcre/pcre_compile.cpp:
        (compile_branch):
        (is_anchored):
        (find_firstassertedchar):
        (printCompiledRegExp):
        (jsRegExpCompile):
        * pcre/pcre_exec.cpp:
        (jsRegExpExecute):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Sam.

        Pull first_byte and req_byte optimizations out into separate static funtions, SunSpider reported this as a win.

        * pcre/pcre_exec.cpp:
        (tryFirstByteOptimization):
        (tryRequiredByteOptimization):
        (jsRegExpExecute):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28181 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        give PCRE_MULTILINE a better name: OptionMatchAcrossMultipleLines

        * pcre/pcre_compile.cpp:
        (compile_branch):
        (is_anchored):
        (printCompiledRegExp):
        (jsRegExpCompile):
        * pcre/pcre_exec.cpp:
        (jsRegExpExecute):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28180 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Oliver.

        Deprecate jsRegExpExecute's offset-vector fallback code

        * pcre/pcre_exec.cpp:
        (jsRegExpExecute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28179 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-28  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Centralize code for subjectPtr adjustments using inlines, only ever check for a single trailing surrogate (as UTF16
 only allows one), possibly fix PCRE bugs involving char classes and garbled UTF16 strings.

        * pcre/pcre_exec.cpp:
        (match):
        (jsRegExpExecute):
        * pcre/pcre_internal.h:
        (getPreviousChar):
        (movePtrToPreviousChar):
        (movePtrToNextChar):
        (movePtrToStartOfCurrentChar):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28178 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Make cur_is_word and prev_is_word locals, and change OP_ANY to OP_ANY_CHAR for clarity

        * pcre/pcre_compile.cpp:
        (find_fixedlength):
        (compile_branch):
        (canApplyFirstCharOptimization):
        * pcre/pcre_exec.cpp:
        (match):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28177 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Mitz & Maciej.

        Change _NC operators to use _IGNORING_CASE for clarity

        * pcre/pcre_compile.cpp:
        (find_fixedlength):
        (compile_branch):
        (find_firstassertedchar):
        * pcre/pcre_exec.cpp:
        (match):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28176 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Mitz.

        Remove branch from return

        * pcre/pcre_compile.cpp:
        (compile_branch):
        * pcre/pcre_exec.cpp:
        (match):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28175 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Add repeatInformationFromInstructionOffset inline

        * pcre/pcre_exec.cpp:
        (repeatInformationFromInstructionOffset):
        (match):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28174 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Remove no longer used error code JSRegExpErrorMatchLimit

        * kjs/regexp.cpp:
        (KJS::RegExp::match):
        * pcre/pcre.h:
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28173 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-28  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        change getChar* functions to return result and push 'c' into local scopes for clarity

        * pcre/pcre_compile.cpp:
        (compile_branch):
        (calculateCompiledPatternLengthAndFlags):
        * pcre/pcre_exec.cpp:
        (match):
        * pcre/pcre_internal.h:
        (getChar):
        (getCharAndAdvance):
        (getCharAndLength):
        (getCharAndAdvanceIfSurrogate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28172 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Sam.

        Make i locally scoped for better code clarity

        * pcre/pcre_exec.cpp:
        (match):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-28  Eric Seidel  <eric@webkit.org>

        Reviewed by Sam.

        Comment cleanup

        * pcre/pcre_exec.cpp:
        (match):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28170 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Give subjectPtr and instructionPtr sane names, reduce size of MatchFrame for a 0.2% speedup.

        * pcre/pcre_compile.cpp:
        (compile_branch):
        (calculateCompiledPatternLengthAndFlags):
        * pcre/pcre_exec.cpp:
        (match_ref):
        (MatchStack::pushNewFrame):
        (getUTF8CharAndIncrementLength):
        (match):
        * pcre/pcre_internal.h:
        (getChar):
        (getCharAndAdvance):
        (getCharAndLength):
        (getCharAndAdvanceIfSurrogate):
        * pcre/pcre_xclass.cpp:
        (getUTF8CharAndAdvancePointer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28169 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-11-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Sam.

        Small speedup (0.7%) by simplifying canUseStackBufferForNextFrame() check

        * pcre/pcre_exec.cpp:
        (MatchStack::MatchStack):
        (MatchStack::popCurrentFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28168 268f45cc-cd09-0410-ab3c-d52691b4dbfc