⚠ Archived content — this site is no longer maintained.   Current WebKit documentation is at docs.webkit.org.

Timeline



Nov 15, 2007:

11:47 PM Changeset in webkit [27846] by mrowe@apple.com
  • 2 edits in trunk/JavaScriptCore

2007-11-15 Mark Rowe <​mrowe@apple.com>

Gtk build fix.

  • kjs/Parser.cpp:
11:44 PM Changeset in webkit [27845] by mrowe@apple.com
  • 4 edits in trunk/JavaScriptCore

Mac build and header search path sanity fix.

Reviewed by Sam Weinig and Tim Hatcher.

Move base setting for HEADER_SEARCH_PATHS into Base.xcconfig, and extend
it in JavaScriptCore.xcconfig. This removes the need to override it on a
per-target basis inside the .xcodeproj file.

  • Configurations/Base.xcconfig:
  • Configurations/JavaScriptCore.xcconfig:
  • JavaScriptCore.xcodeproj/project.pbxproj:
11:08 PM Changeset in webkit [27844] by mrowe@apple.com
  • 2 edits in trunk/JavaScriptCore

2007-11-15 Mark Rowe <​mrowe@apple.com>

Qt build fix.

  • kjs/Parser.h:
11:00 PM Changeset in webkit [27843] by adele@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Oliver.

Fix for <rdar://problem/5566652> CrashTracer: [USER] 3 crashes in Safari at HTMLSelectElement::menuListDefaultEventHandler (reproducible on mactc30.com)

Test: fast/forms/menulist-no-renderer-onmousedown.html

  • html/HTMLSelectElement.cpp: (WebCore::HTMLSelectElement::defaultEventHandler): Nil check the renderer here. None of the default behavior makes sense if there's no renderer. (WebCore::HTMLSelectElement::listBoxDefaultEventHandler): Moved the nil check to the main defaultEventHandler.

LayoutTests:

Reviewed by Oliver.

Test for <rdar://problem/5566652> CrashTracer: [USER] 3 crashes in Safari at HTMLSelectElement::menuListDefaultEventHandler (reproducible on mactc30.com)

  • fast/forms/menulist-no-renderer-onmousedown-expected.txt: Added.
  • fast/forms/menulist-no-renderer-onmousedown.html: Added.
10:54 PM Changeset in webkit [27842] by ggaren@apple.com
  • 16 edits in trunk/JavaScriptCore

Reviewed by Eric Seidel.

Another round of grammar / parsing cleanup.


  1. Created distinct parser calls for parsing function bodies vs programs. This will help later with optimizing global variable access.


  1. Turned Parser into a singleton. Cleaned up Lexer's singleton interface.


  1. Modified Lexer to free a little more memory when done lexing. (Added FIXMEs for similar issues that I didn't fix.)


  1. Changed Lexer::makeIdentifier and Lexer::makeUString to start respecting the arguments passed to them. (No behavior change, but this problem could have caused serious problems for an unsuspecting user of these functions.)


  1. Removed KJS_DEBUG_MEM because it was bit-rotted.


  1. Removed Parser::prettyPrint because the same work was simpler to do at the call site.


  1. Some renames:


"Parser::accept" => "Parser::didFinishParsing"
"Parser::sid" => "Parser::m_sourceID"
"Lexer::doneParsing" => "Lexer::clear"
"sid" => "sourceId"
"lineno" => "lineNo"


  • JavaScriptCore.exp:
  • kjs/Parser.cpp: (KJS::Parser::Parser): (KJS::Parser::parseProgram): (KJS::Parser::parseFunctionBody): (KJS::Parser::parse): (KJS::Parser::didFinishParsing): (KJS::parser):
  • kjs/Parser.h: (KJS::Parser::sourceId):
  • kjs/function.cpp: (KJS::GlobalFuncImp::callAsFunction):
  • kjs/function_object.cpp: (FunctionObjectImp::construct):
  • kjs/grammar.y:
  • kjs/interpreter.cpp: (KJS::Interpreter::checkSyntax): (KJS::Interpreter::evaluate):
  • kjs/interpreter.h:
  • kjs/lexer.cpp: (kjsyylex): (KJS::lexer): (KJS::Lexer::Lexer): (KJS::Lexer::~Lexer): (KJS::Lexer::scanRegExp): (KJS::Lexer::doneParsing): (KJS::Lexer::makeIdentifier): (KJS::Lexer::makeUString):
  • kjs/lexer.h: (KJS::Lexer::pattern): (KJS::Lexer::flags): (KJS::Lexer::sawError):
  • kjs/nodes.cpp: (KJS::Node::Node): (KJS::FunctionBodyNode::FunctionBodyNode):
  • kjs/nodes.h:
  • kjs/testkjs.cpp: (prettyPrintScript): (kjsmain):
  • kjs/ustring.cpp:
  • kjs/ustring.h:
10:25 PM Changeset in webkit [27841] by Antti Koivisto
  • 2 edits in trunk/WebCore

Reviewed by Adele.

Fix <rdar://problem/5601598>
The movie tag should do initialization on need and not during startup.


Initialize Movie MIME type hash on demand

  • platform/MIMETypeRegistry.cpp: (WebCore::initialiseSupportedMovieMIMETypes): (WebCore::initialiseMIMETypeRegistry): (WebCore::MIMETypeRegistry::isSupportedMovieMIMEType): (WebCore::MIMETypeRegistry::getSupportedMovieMIMETypes):
10:18 PM Changeset in webkit [27840] by mrowe@apple.com
  • 2 edits in trunk/WebCore

2007-11-15 Mark Rowe <​mrowe@apple.com>

Build fix for Qt on Windows.

  • xml/XPathStep.cpp: (WebCore::XPath::Step::nodesInAxis):
10:15 PM Changeset in webkit [27839] by timothy@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Mark Rowe.

Bug 14761: Web Inspector leaks JS objects and DOM nodes
​http://bugs.webkit.org/show_bug.cgi?id=14761

  • page/InspectorController.cpp: (WebCore::InspectorController::~InspectorController): Change the tear down order so that inspectorDestroyed() is called before InspectorController private data is set to Zero. The inspectorDestroyed() call ends up closing the WebView and triggering InspectorController::close() which unprotects m_scriptObject. This didn't happen when the private data was cleared before calling inspectorDestroyed().
10:10 PM Changeset in webkit [27838] by mrowe@apple.com
  • 2 edits in trunk/WebCore

2007-11-15 Mark Rowe <​mrowe@apple.com>

Gtk build fix.

  • WebCore.pro:
10:04 PM Changeset in webkit [27837] by ap@webkit.org
  • 4 edits
    2 adds in trunk

Reviewed by Darin.

​http://bugs.webkit.org/show_bug.cgi?id=15989
XPath queries with predicates incorrectly retains the current node across unions

Test: fast/xpath/union-context-node.xhtml

  • xml/XPathPath.cpp: (WebCore::XPath::LocationPath::evaluate): Restore context after evaluation.
  • xml/XPathStep.cpp: (WebCore::XPath::Step::evaluate): Do not backup context, as we can easily re-create it.
10:02 PM Changeset in webkit [27836] by ap@webkit.org
  • 1 edit
    2 adds in trunk/LayoutTests

Reviewed by Darin.

​http://bugs.webkit.org/show_bug.cgi?id=15448
Import XPath test suite from amachang.art-code.org

Test by Hitoshi Amano <​seijro@gmail.com>.

  • fast/xpath/xpath-functional-test-expected.txt: Added.
  • fast/xpath/xpath-functional-test.html: Added.
10:00 PM Changeset in webkit [27835] by ap@webkit.org
  • 3 edits
    2 adds in trunk

Reviewed by Darin.

​http://bugs.webkit.org/show_bug.cgi?id=15988
REGRESSION: XPath preceding-axis query misses nested elements

Test: fast/xpath/preceding-axis.xhtml

  • xml/XPathStep.cpp: (WebCore::XPath::Step::nodesInAxis): Hopefully correct this time.
5:21 PM Changeset in webkit [27834] by timothy@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Mark Rowe and Sam.

This corrects a couple of issues in the Web Inspector where selected search results
would not have the right text color when the window is inactive or the results are focused.

  • page/inspector/inspector.css:
5:04 PM Changeset in webkit [27833] by timothy@apple.com
  • 3 edits in trunk/WebCore

Reviewed by Sam.

Bug 16007: REGRESSION: Disclosure triangle click zone for resource categories is off by 10 pixels
​http://bugs.webkit.org/show_bug.cgi?id=16007

Makes the click width an option on TreeElement and make ResourceCategoryTreeElement
a true subclass of TreeElement. The click width for ResourceCategoryTreeElement is
20 pixels, and the default is 10 pixels.

  • page/inspector/ResourceCategory.js:
  • page/inspector/treeoutline.js:
4:42 PM Changeset in webkit [27832] by andersca@apple.com
  • 5 edits in trunk/WebCore

Reviewed by Brady.

Pass a VoidCallback to Database::transaction and Database::changeVersion.
This is not yet hooked up to the database machinery.


  • WebCore.vcproj/WebCore.vcproj:
  • bindings/js/JSDatabaseCustom.cpp: (WebCore::JSDatabase::changeVersion): (WebCore::JSDatabase::transaction):
  • storage/Database.cpp: (WebCore::Database::changeVersion): (WebCore::Database::transaction):
  • storage/Database.h:
4:40 PM Changeset in webkit [27831] by oliver@apple.com
  • 3 edits in trunk/JavaScriptCore

<rdar://problem/5601548> REGRESSION: All SourceElements and their children leak after a syntax error

Reviewed by Darin.

Add a stub node to maintain the Vector of SourceElements until assignment.

4:17 PM Changeset in webkit [27830] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-15 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Abstract most of RMATCH into MatchStack functions.

SunSpider claims this, combined with the last 2 patches was a 1% speedup, 10% for dna-regexp.

  • pcre/pcre_exec.cpp: (MatchStack::canUseStackBufferForNextFrame): (MatchStack::allocateNextFrame): (MatchStack::pushNewFrame): (MatchStack::frameIsStackAllocated): (MatchStack::popCurrentFrame): (MatchStack::unrollAnyHeapAllocatedFrames): (match):
4:16 PM Changeset in webkit [27829] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-15 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Remove RETURN_ERROR, add MatchStack

  • pcre/pcre_exec.cpp: (MatchStack::MatchStack): (MatchStack::unrollAnyHeapAllocatedFrames): (matchError): (match):
4:14 PM Changeset in webkit [27828] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-15 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Clean up match function to match WebKit style

  • JavaScriptCore.xcodeproj/project.pbxproj:
  • pcre/pcre_exec.cpp: (match):
4:07 PM Changeset in webkit [27827] by timothy@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Adam.

Bug 16005: Hovering in the breadcrumbs causes jumpy behavior
​http://bugs.webkit.org/show_bug.cgi?id=16005

Hovering over cumbs no longer exposes new crumbs. Clicking on a collapsed
crumb will expose as many hidden crumbs as possible to the user. Also crumbs
that have ID attributes will compact to the ID over the tag name.

  • page/inspector/DocumentPanel.js:
3:53 PM Changeset in webkit [27826] by andersca@apple.com
  • 5 edits
    2 adds
    1 delete in trunk/WebCore

Reviewed by Sam.

Make the VoidCallback DOM interface behave more like our other DOM interfaces.


  • bindings/js/JSCustomVoidCallback.h: Added.
  • bindings/js/JSCustomVoidCallback.cpp: Added. (WebCore::JSCustomVoidCallback::JSCustomVoidCallback): (WebCore::JSCustomVoidCallback::handleEvent): VoidCallback implementation that wraps a JS object.


(WebCore::toVoidCallback):
New method that creates a VoidCallback implementation given a JS object.


  • bindings/scripts/CodeGeneratorJS.pm: Specify that VoidCallback can fail conversion and add a custom conversion function.


  • html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::movieCuePointReached): Call handleEvent on the callback.


(WebCore::HTMLMediaElement::removeCuePoint):
Comment out parts of this method, it has already been removed in the HTML5 spec and it relies
on overriding VoidCallback equality.


  • html/VoidCallback.cpp: Removed.


  • html/VoidCallback.h: (WebCore::VoidCallback::VoidCallback): (WebCore::VoidCallback::~VoidCallback): Make this an abstract class with a pure virtual handleEvent method.
3:42 PM Changeset in webkit [27825] by Adam Roben
  • 2 edits in trunk/WebCore

Fix <rdar://5485108> Type-select while a <select> menu is open doesn't work

There were two things broken here:

  • WebView now requires a WM_CHAR message to be sent before it will create a keyboard event with a charCode, but we were only sending a WM_KEYDOWN event.
  • The popup menu was not updating its focused index when the <select> element's selectedIndex changed.

Reviewed by Adele.

  • platform/win/PopupMenuWin.cpp: (WebCore::PopupMenu::updateFromElement): Update the focused index from the client's selected index. (WebCore::PopupWndProc): When we receive a WM_CHAR message for a printable character, manufacture a WM_KEYDOWN message for it and post both that and the current WM_CHAR message to the WebView.
3:23 PM Changeset in webkit [27824] by beidson@apple.com
  • 2 edits in trunk/WebCore

Build fix

  • storage/DatabaseTracker.cpp: (WebCore::DatabaseTracker::fullPathForDatabase): (WebCore::DatabaseTracker::databaseNamesForOrigin):
2:54 PM Changeset in webkit [27823] by justin.garcia@apple.com
  • 1 edit in trunk/LayoutTests/platform/mac/fast/events/5056619-expected.txt

Forgot this new expected result in my previous check-in.

2:29 PM Changeset in webkit [27822] by beidson@apple.com
  • 1 edit in trunk/WebCore/ChangeLog

Reviewed by John and Sam

Stubbing out everything required for a WebKit API for databases

Biggest change in WebCore is that we passed around SecurityOriginData as the representation
of a SecurityOrigin, instead of a String or a SecurityOrigin itself (which is an opaque object
that has a different purpose besides just hanging on to the data)

  • WebCore.base.exp:
  • WebCore.xcodeproj/project.pbxproj:
  • platform/SecurityOrigin.cpp:
  • platform/SecurityOrigin.h:
  • platform/SecurityOriginData.h: Added. Simple object to encapsulate the 3 pieces of data in the SecurityOrigin tuple (WebCore::SecurityOriginData::SecurityOriginData): (WebCore::SecurityOriginData::protocol): (WebCore::SecurityOriginData::host): (WebCore::SecurityOriginData::port): (WebCore::SecurityOriginData::toString): Simple concatenation of the 3 components, colon separated
  • storage/Database.cpp: (WebCore::Database::Database): Use SecurityOriginData instead
  • storage/DatabaseTracker.cpp: (WebCore::DatabaseTracker::fullPathForDatabase): (WebCore::DatabaseTracker::databaseNamesForOrigin): (WebCore::DatabaseTracker::deleteDatabasesWithOrigin): (WebCore::DatabaseTracker::deleteDatabase):
  • storage/DatabaseTracker.h:
  • storage/DatabaseTrackerClient.h: Added. For dispatching notifications up to WebKit (WebCore::DatabaseTrackerClient::~DatabaseTrackerClient):2007-11-15 Jon Honeycutt <​jhoneycutt@apple.com>
2:27 PM Changeset in webkit [27821] by beidson@apple.com
  • 12 edits
    5 copies
    3 moves
    3 adds in trunk

WebCore:

Reviewed by John and Sam

Stubbing out everything required for a WebKit API for databases

Biggest change in WebCore is that we passed around SecurityOriginData as the representation
of a SecurityOrigin, instead of a String or a SecurityOrigin itself (which is an opaque object
that has a different purpose besides just hanging on to the data)

  • WebCore.base.exp:
  • WebCore.xcodeproj/project.pbxproj:
  • platform/SecurityOrigin.cpp:
  • platform/SecurityOrigin.h:
  • platform/SecurityOriginData.h: Added. Simple object to encapsulate the 3 pieces of data in the SecurityOrigin tuple (WebCore::SecurityOriginData::SecurityOriginData): (WebCore::SecurityOriginData::protocol): (WebCore::SecurityOriginData::host): (WebCore::SecurityOriginData::port): (WebCore::SecurityOriginData::toString): Simple concatenation of the 3 components, colon separated
  • storage/Database.cpp: (WebCore::Database::Database): Use SecurityOriginData instead
  • storage/DatabaseTracker.cpp: (WebCore::DatabaseTracker::fullPathForDatabase): (WebCore::DatabaseTracker::databaseNamesForOrigin): (WebCore::DatabaseTracker::deleteDatabasesWithOrigin): (WebCore::DatabaseTracker::deleteDatabase):
  • storage/DatabaseTracker.h:
  • storage/DatabaseTrackerClient.h: Added. For dispatching notifications up to WebKit (WebCore::DatabaseTrackerClient::~DatabaseTrackerClient):

WebKit:

Reviewed by John

Stubbing out everything required for a WebKit API for databases

  • WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

Reviewed by John

Stubbing out everything required for a WebKit API for databases

These interfaces seem to provide everything we need for UI and management at the browser level

  • Misc/WebDatabaseManager.h: Removed.
  • Misc/WebDatabaseManager.mm: Removed.
  • Misc/WebDatabaseManagerPrivate.h: Removed.
  • Misc/WebSecurityOrigin.mm: Added. Object that acts as a container for the "SecurityOrigin tuple" (protocol, domain, and port) (-[WebSecurityOriginPrivate initWithProtocol:domain:port:]): (-[WebSecurityOriginPrivate dealloc]): (-[WebSecurityOrigin initWithProtocol:domain:]): (-[WebSecurityOrigin initWithProtocol:domain:port:]): (-[WebSecurityOrigin protocol]): (-[WebSecurityOrigin domain]): (-[WebSecurityOrigin port]): (-[WebSecurityOrigin dealloc]): (-[WebSecurityOrigin _initWithWebCoreSecurityOriginData:WebCore::]):
  • Misc/WebSecurityOriginInternal.h: Added.
  • Misc/WebSecurityOriginPrivate.h: Added.
  • Storage/WebDatabaseManager.mm: Added. (+[WebDatabaseManager sharedWebDatabaseManager]): (-[WebDatabaseManager origins]): Get a list of all origins currently tracked (-[WebDatabaseManager detailsForOrigin:]): Get the current usage and current quota for the given origin (-[WebDatabaseManager databasesWithOrigin:]): Get all databases for a certain origin (-[WebDatabaseManager detailsForDatabase:withOrigin:]): Get all details about a specific database (-[WebDatabaseManager setQuota:forOrigin:]): Change origin-wide quota (-[WebDatabaseManager deleteAllDatabases]): (-[WebDatabaseManager deleteAllDatabasesWithOrigin:]): (-[WebDatabaseManager deleteDatabase:withOrigin:]):
  • Storage/WebDatabaseManagerPrivate.h: Added.
  • Storage/WebDatabaseManagerInternal.h: Added. (WebKitInitializeDatabasesIfNecessary): One-time initialization of database-related things
  • Storage/WebDatabaseTrackerClient.h: Added. Stubbed out client for notifications
  • Storage/WebDatabaseTrackerClient.mm: Added. (WebDatabaseTrackerClient::sharedWebDatabaseTrackerClient): (WebDatabaseTrackerClient::WebDatabaseTrackerClient): (WebDatabaseTrackerClient::~WebDatabaseTrackerClient): (WebDatabaseTrackerClient::dispatchDidModifyOrigin): (WebDatabaseTrackerClient::dispatchDidModifyDatabase):
  • WebView/WebView.mm: (-[WebView _commonInitializationWithFrameName:groupName:]): Do one-time initialization of Database-related things here
2:23 PM Changeset in webkit [27820] by jhoneycutt@apple.com
  • 2 edits in trunk/WebCore

2007-11-15 Jon Honeycutt <​jhoneycutt@apple.com>

Reviewed by Anders.

NP_ASFILEONLY streams should not buffer data in m_deliveryData.

  • plugins/win/PluginStreamWin.cpp: (WebCore::PluginStreamWin::didReceiveData):
2:02 PM Changeset in webkit [27819] by andersca@apple.com
  • 1 edit in trunk/LayoutTests/fast/dom/Window/window-onFocus.html

Revert accidental change. andersca--

1:52 PM Changeset in webkit [27818] by justin.garcia@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Adele Peterson.

<rdar://problem/5497643> Crash at Node::isDescendantOf when switching out of Edit HTML Source mode


A textarea that contained the selection was removed but the selection wasn't cleared,
and we'd crash in code that assumed a valid, in-document selection.

  • editing/SelectionController.cpp: (WebCore::removingNodeRemovesPosition): Clear the selection if the node being removed is the shadowAncestorNode of the node that contains the position, not just if the node being removed contains that shadowAncestorNode.

LayoutTests:

Reviewed by Adele Peterson.

<rdar://problem/5497643> Crash at Node::isDescendantOf when switching out of Edit HTML Source mode

  • editing/selection/5497643-expected.txt: Added.
  • editing/selection/5497643.html: Added.
1:35 PM Changeset in webkit [27817] by Adam Roben
  • 2 edits in trunk/WebKitTools

Make run-safari actually work on Windows

  • Scripts/run-safari: Fixed the order of arguments to cp, and added a chdir call.
1:22 PM Changeset in webkit [27816] by Nikolas Zimmermann
  • 2 edits in trunk/WebCore

Reviewed by Maciej.
Fix logic error: calculateGlyphBoundaries takes RTL text into account itself, no need to adjust offsets before.

12:27 PM Changeset in webkit [27815] by andersca@apple.com
  • 4 edits
    2 adds in trunk

WebCore:

Reviewed by Brady.

<rdar://problem/5562470>
openDatabase does not work when the version string is empty

If the string is empty, pass a real empty string to sqlite3_bind_text16.


  • platform/sql/SQLiteStatement.cpp: (WebCore::SQLiteStatement::bindText):

LayoutTests:

Reviewed by Brady.

<rdar://problem/5562470>
openDatabase does not work when the version string is empty

  • storage/open-database-empty-version-expected.txt: Added.
  • storage/open-database-empty-version.html: Added.
12:00 PM Changeset in webkit [27814] by sfalken@apple.com
  • 1 copy in tags/Safari-525.1b

New tag.

11:49 AM Changeset in webkit [27813] by andersca@apple.com
  • 6 edits in trunk/WebCore

Reviewed by Brady.

Get rid of SQLiteStatement::bindText and rename bindText16 to bindText.
Remove the copy parameter from bindText and bindBlob and always copy passed in data instead.

  • loader/icon/IconDatabase.cpp: (WebCore::IconDatabase::setIconIDForPageURLInSQLDatabase): (WebCore::IconDatabase::removePageURLFromSQLDatabase): (WebCore::IconDatabase::getIconIDForIconURLFromSQLDatabase): (WebCore::IconDatabase::addIconURLToSQLDatabase): (WebCore::IconDatabase::getImageDataForIconURLFromSQLDatabase): (WebCore::IconDatabase::writeIconSnapshotToSQLDatabase):
  • platform/sql/SQLiteStatement.cpp: (WebCore::SQLiteStatement::prepare):
  • platform/sql/SQLiteStatement.h:
  • storage/Database.cpp: (WebCore::setTextValueInDatabase):
  • storage/DatabaseTracker.cpp: (WebCore::DatabaseTracker::fullPathForDatabase): (WebCore::DatabaseTracker::databaseNamesForOrigin): (WebCore::DatabaseTracker::addDatabase):
10:39 AM Changeset in webkit [27812] by sfalken@apple.com
  • 2 edits in trunk/JavaScriptCore

Windows build fix.

12:01 AM Changeset in webkit [27811] by beidson@apple.com
  • 2 edits in trunk/WebCore

WebCore:

Rubberstamped by Sam

  • platform/SecurityOrigin.h: Lets make this an unsigned short and *not* rule out half the possible ports?

Nov 14, 2007:

9:59 PM Changeset in webkit [27810] by ap@webkit.org
  • 7 edits in trunk/JavaScriptCore

Reviewed by Darin.

​http://bugs.webkit.org/show_bug.cgi?id=15982
Improve JSString UTF-8 decoding

  • API/JSStringRef.cpp: (JSStringCreateWithUTF8CString): Use strict decoding, return 0 on error.
  • wtf/unicode/UTF8.cpp: (WTF::Unicode::convertUTF16ToUTF8): (WTF::Unicode::convertUTF8ToUTF16):
  • wtf/unicode/UTF8.h: Made these function names start with a lower case letter.
  • kjs/ustring.cpp: (KJS::UString::UTF8String): Updated for the above renaming.
  • bindings/c/c_utility.cpp: (KJS::Bindings::convertUTF8ToUTF16WithLatin1Fallback): Renamed to highlight the difference from convertUTF8ToUTF16 in wtf/unicode. (KJS::Bindings::convertNPStringToUTF16): Updated for the above renaming. (KJS::Bindings::identifierFromNPIdentifier): Ditto.
  • bindings/c/c_utility.h: Made convertUTF8ToUTF16WithLatin1Fallback() a file static.
9:49 PM Changeset in webkit [27809] by ap@webkit.org
  • 4 edits
    4 adds in trunk

Reviewed by Justin Garcia.

​http://bugs.webkit.org/show_bug.cgi?id=15781
REGRESSION: Ligatures fail to form when typing in Devanagari (because WebKit can't handle
a marked range that covers half of a composed character sequence)

Test: platform/mac/editing/input/devanagari-ligature.html

This fix is somewhat of a hack, as it asks editing commands to work with invalid selections.
However, this is not entirely new for them, as Roman accents are typed in a similar manner.

In the future, we probably want to make commands work with ranges (or Positions explicitly).

  • editing/Editor.cpp: (WebCore::Editor::selectComposition): Force selection to composition range.
  • editing/InsertTextCommand.cpp: (WebCore::InsertTextCommand::input): Force ending selection to inserted text range.
9:17 PM Changeset in webkit [27808] by weinig@apple.com
  • 2 edits in trunk/JavaScriptCore

Rubber-stamped by Anders.

Fix the Xcode project file after it was messed up in r27402.

7:52 PM Applications using WebKit edited by info@erikjohansson.com
Added Coda to "Web Development Applications" (diff)
6:40 PM Changeset in webkit [27807] by eric@webkit.org
  • 2 edits in trunk/JavaScriptCore

2007-11-14 Eric Seidel <​eric@webkit.org>

Reviewed by Oliver.

More PCRE style cleanup.

  • pcre/pcre_compile.cpp: (compile_regex):
6:02 PM Changeset in webkit [27806] by Adam Roben
  • 2 edits in trunk/JavaScriptCore

Clean up the bison conflict checking script

Reviewed by Geoff.

  • DerivedSources.make:
6:00 PM Changeset in webkit [27805] by eric@webkit.org
  • 4 edits in trunk/JavaScriptCore

2007-11-14 Eric Seidel <​eric@webkit.org>

Reviewed by Geoff.

Another round of PCRE cleanups: inlines

SunSpider claims that this, combined with my previous PCRE cleanup were a 0.7% speedup, go figure.

  • pcre/pcre_compile.cpp: (jsRegExpCompile):
  • pcre/pcre_exec.cpp: (match): (jsRegExpExecute):
  • pcre/pcre_internal.h: (PUT): (GET): (PUT2): (GET2): (isNewline):
5:55 PM Changeset in webkit [27804] by beidson@apple.com
  • 1 edit
    1 add in trunk/WebKitSite

For future use

  • images/DroseraPic.png: Added.
5:34 PM Changeset in webkit [27803] by Adam Roben
  • 3 edits in trunk/WebKitTools

Updates to Safari launching now that 3.0.4 is released

Reviewed by Sam.

  • FindSafari/FindSafari.cpp: (getWebViewCLSID): Use version-independent ProgID.
  • Scripts/run-safari: Use run-webkit-nightly.cmd.
5:17 PM Changeset in webkit [27802] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2007-11-14 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

Give PCRE a (small) bath.
Fix some formating and break things off into separate functions
​http://bugs.webkit.org/show_bug.cgi?id=15993

  • pcre/pcre_compile.cpp: (calculateCompiledPatternLengthAndFlags): (printCompiledRegExp): (returnError): (jsRegExpCompile):
  • pcre/pcre_internal.h: (compile_data::compile_data):
5:04 PM Changeset in webkit [27801] by Antti Koivisto
  • 2 edits in trunk/LayoutTests

Try to make this test less timing dependent.

  • http/tests/media/video-play-stall.html:
3:48 PM Changeset in webkit [27800] by Beth Dakin
  • 7 edits
    5 adds in trunk

WebCore:

Reviewed by Hyatt.

Fix for <rdar://problem/5540855> REGRESSION: Combination of client-
side image map and <a> tag is not working properly (15522)

  • html/HTMLMapElement.cpp: (WebCore::HTMLMapElement::checkDTD): Allow <map> to have both block and inline children rather than just block children. This matches Firefox.

LayoutTests:

Reviewed by Hyatt.

Tests for <rdar://problem/5540855> REGRESSION: Combination of
client-side image map and <a> tag is not working properly (15522)

New test:

  • fast/images/image-map-anchor-children.html: Added.
  • fast/images/resources/boston.gif: Added.
  • platform/mac/fast/images/image-map-anchor-children- expected.checksum: Added.
  • platform/mac/fast/images/image-map-anchor-children-expected.png: Added.
  • platform/mac/fast/images/image-map-anchor-children-expected.txt: Added.


This test was broken and is fixed now! I had to edit the test
itself because it was written in the html that the two <map>
examples were failing, but now they succeed.

  • fast/invalid/residual-style.html:
  • platform/mac/fast/invalid/residual-style-expected.checksum:
  • platform/mac/fast/invalid/residual-style-expected.png:
  • platform/mac/fast/invalid/residual-style-expected.txt:
2:59 PM Changeset in webkit [27799] by ggaren@apple.com
  • 3 edits in trunk/JavaScriptCore

Reviewed by Eric Seidel.


Cleaned up the JavaScript grammar a bit.


  1. Changed BlockNode to always hold a child vector (which may be empty), eliminating a few NULL-check branches in the common execution case.


  1. Changed the Block production to correctly report its starting and ending line numbers to the debugger. (It used to report its ending line as its starting line.) Also, removed duplicate line-reporting code inside the BlockNode constructor.


  1. Moved curly braces up from FunctionBody production into parent productions. (I had to move the line number reporting code, too, since it depends on the location of the curly braces.) This matches the ECMA spec more closely, and makes some future changes I plan easier.


  1. Fixed statementList* convenience functions to deal appropriately with empty Vectors.

SunSpider reports a small and statistically insignificant speedup.

  • kjs/grammar.y:
  • kjs/nodes.cpp: (KJS::statementListPushFIFO): (KJS::statementListGetDeclarations): (KJS::statementListInitializeDeclarationStack): (KJS::statementListInitializeVariableAccessStack): (KJS::BlockNode::BlockNode): (KJS::BlockNode::optimizeVariableAccess): (KJS::BlockNode::getDeclarations): (KJS::BlockNode::execute): (KJS::FunctionBodyNode::initializeDeclarationStacks): (KJS::FunctionBodyNode::optimizeVariableAccess):
2:57 PM Changeset in webkit [27798] by Adam Roben
  • 2 edits in trunk/WebKit/win

Change Interfaces to be a "Static Library" project

Being a Utility project meant that Interfaces was rebuilding every
time (Utility projects are supposed to handle their own dependencies).
Interfaces isn't really a static library, but it means that VS will
handle dependencies for us (and not complain about missing manifests
like it did when the project was an Application).

  • WebKit.vcproj/Interfaces.vcproj:
2:56 PM Changeset in webkit [27797] by Adam Roben
  • 2 edits in trunk/WebKit/win

Shut up FixMIDLHeaders.pl

  • WebKit.vcproj/FixMIDLHeaders.pl:
2:33 PM Changeset in webkit [27796] by justin.garcia@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Alexey Proskuryakov.

<rdar://problem/5546763> CrashTracer: [USER] 362 crashes at WebCore::DeleteSelectionCommand::mergeParagraphs

  • editing/DeleteSelectionCommand.cpp: (WebCore::DeleteSelectionCommand::handleGeneralDelete): Removed an irrelevant FIXME. (WebCore::DeleteSelectionCommand::mergeParagraphs): If the block that contained the end of the selection hasn't been removed but has been emptied by deletion, we would to try and fail to create a VisiblePosition inside that block, which could lead to a crash. If that happens, there's no content in the block to move, so just remove the block and return. Preserve m_needPlaceholder during the call to moveParagraphs, since it may change it and since it does its own placeholder insertion when necessary. (WebCore::DeleteSelectionCommand::doApply): No need to check m_needPlaceholder before calling mergeParagraphs, because it handles preserving m_needPlaceholder when it calls moveParagraphs.

LayoutTests:

Reviewed by Alexey Proskuryakov.


<rdar://problem/5546763> CrashTracer: [USER] 362 crashes at WebCore::DeleteSelectionCommand::mergeParagraphs

  • editing/deleting/5546763-expected.txt: Added.
  • editing/deleting/5546763.html: Added.
2:30 PM Changeset in webkit [27795] by andersca@apple.com
  • 4 edits in trunk

WebKitTools:

Reviewed by Adam.

<rdar://problem/5309081>
In DRT, "plugin.logDestroy = true" not working on Windows.


  • DumpRenderTree/win/TestNetscapePlugin/main.c: (NPP_Destroy):

LayoutTests:

Reviewed by Adam.

<rdar://problem/5309081>
In DRT, "plugin.logDestroy = true" not working on Windows.

Remove now succeeding test from the skipped list.


  • platform/win/Skipped:
2:09 PM Changeset in webkit [27794] by timothy@apple.com
  • 2 edits in trunk/WebKitLibraries

Remove the WebKit dependancy from WebKitSystemInterface.

2:08 PM Changeset in webkit [27793] by Adam Roben
  • 1 edit in trunk/WebKit/win/ChangeLog

Fix ChangeLog order

2:07 PM Changeset in webkit [27792] by Adam Roben
  • 2 edits in trunk/WebKit/win

Change Interfaces to use a Utility configuration

It had previously been marked as an "Application (.exe)", which is
wrong but worked with most versions of Visual Studio.

Rubberstamped by Steve.

  • WebKit.vcproj/Interfaces.vcproj:
1:44 PM Changeset in webkit [27791] by Antti Koivisto
  • 3 edits in trunk/LayoutTests

Try to make this test less timing dependent.

  • media/video-currentTime-set-expected.txt:
  • media/video-currentTime-set.html:
12:36 PM Changeset in webkit [27790] by eric@webkit.org
  • 4 edits in trunk

2007-11-14 Eric Seidel <​eric@webkit.org>

Reviewed by Sam.

  • sunspider: add --shark-cache for L2 Cache Miss profiling
11:19 AM Changeset in webkit [27789] by timothy@apple.com
  • 3 edits in trunk/WebCore

Reviewed by Adam.

Bug 14380: Long DOM ancestry breadcrumb lists get cut off
​http://bugs.webkit.org/show_bug.cgi?id=14380

The breadcumbs will now be compacted and collapsed if there isn't enough room
to show everything. The collapsing algorithm always affects the crumbs that
are farthest away from the selected or hovered crumb first.

  • page/inspector/DocumentPanel.js:
  • page/inspector/inspector.css:
11:10 AM Changeset in webkit [27788] by andersca@apple.com
  • 6 edits in trunk

WebKit/win:

Reviewed by Adam.

Get the new focus window from the wParam instead of calling GetFocus().
Also, send blur even if there is no focused frame.


  • WebView.cpp: (WebViewWndProc):

WebKitTools:

Reviewed by Adam.

<rdar://problem/5141186>
window.layoutTestController.setWindowIsKey is not implemented in DRT.


Implement setWindowIsKey.


  • DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::setWindowIsKey):

LayoutTests:

Reviewed by Adam.

  • platform/win/Skipped: Remove fast/dom/Window/window-onFocus.html since it succeeds now.
10:53 AM Changeset in webkit [27787] by Adam Roben
  • 2 edits in trunk/WebKit/win

Another build fix for systems without Cygwin in their PATH

  • WebKit.vcproj/Interfaces.vcproj: Add Cygwin to the PATH in the Post-Build Event.
10:36 AM Changeset in webkit [27786] by andersca@apple.com
  • 4 edits in trunk/WebCore

Use the correct include path.


  • platform/Cursor.h:
  • plugins/win/PluginPackageWin.h:
  • plugins/win/PluginStreamWin.h:
10:11 AM Applications using WebKit edited by tristan@apple.com
(diff)
10:10 AM Applications using WebKit edited by tristan@apple.com
(diff)
8:33 AM BuildingGtk edited by zecke2@selfish.org
(diff)
8:32 AM BuildingGtk edited by zecke2@selfish.org
(diff)
Note: See TracTimeline for information about the timeline view.