December 2006 - webkit-dev - lists.webkit.org

Changes to GraphicsContext
by David Hyatt 02 Jan '07

02 Jan '07

I changed how the graphics context works today to avoid updating CoreGraphics's fill and stroke colors in every drawGlyphs call. Basically I changed GraphicsContextCG to be more like GraphicsContextQt, which was already behaving the way I wanted CG to behave. Now when you set a fill or stroke color on a GraphicsContext, it is updated immediately in the underlying native context (in both CG and Qt). What was formerly known as the Pen has now been broken down into stroke style, color and thickness. Thickness was changed to a float so that arc drawing could take a float-based thickness when stroking. drawArc was reworked and renamed to strokeArc and its false concept of filling was removed (it was not really using filling). Text in HTML is now being drawn using the fill color and not the pen (stroke) color. This allows for SVG to fill and stroke separately and for text drawing to do either. A new textDrawingMode API has been added to reflect filling, stroking and clipping of text. The API allows for bits and lets you set any combination of those three bits. SVG is now using this API to distinguish between filling and stroking of text. For those of you working with native CG and Qt code directly (as is sometimes the case in RenderTheme*** classes), you will need to be especially careful now. Fill and stroke colors (for speed) are held as Colors in GraphicsContext's cross-platform state, and so if you set the CG fill color directly (and don't go through the cross- platform API), you will now be out of sync. To avoid any issues you should either use the cross-platform API everywhere you can, or simply bracket your uses of native CG code with save and restore calls. Anyone working on ports other than CG and Qt (I patched those two) will need to replace your uses of Pen with the new stroke APIs. In addition you will need to make sure your font rendering code uses the fill color now and not the pen (stroke) color. If you plan on supporting SVG, you will also need to support stroking of text using the stroke color. dave (hyatt(a)apple.com)

3 2

feature request-webkit should send update events on minimization
by Gurmit Teotia 30 Dec '06

30 Dec '06

Hi, We've developed an application using WebKit and Flash. As soon as we're minimizing the main window, flash application was stopping the execution. It was obvious as WebKit was stopping update events upon minimization. WebDashboardBehaviorAlwaysSendActiveNullEventsToPlugIns is declared but is not used anywhere, looks redundant code. Later I achieved that by removing the "WebNetscapePluginEmbeddedView" instance from default notification center for notifications-NSWindowDidMiniaturizeNotification and NSWindowDidResignKeyNotification. To me it does not matter if flash play at same speed while window does not have focus. I don't know till when this solution will work so I'd like to suggest that you must give some option in public API so that WebKit does not stop update events upon minimization. I don't know if this is the right place to suggest feature request. If it's not then please suggest me the address. Thanks, Gurmit

2 3

CanvasRenderingContext2D.cpp
by Mike Reed 22 Dec '06

22 Dec '06

We're all excited to make <canvas> work on our port. However, it looks like the portable part of the implementation does not call through to the platform version of GraphicsContext, but it still has #if PLATFORM(CG) around all the good stuff. Is that a gut feel for when this will be made more "portable"? (to quote one of the comments from that file) // FIXME: Do this through platform-independent GraphicsContext API. thanks, mike

3 3

Reducing Bugs
by Shawn Smith 22 Dec '06

22 Dec '06

I am attempting to reduce the vox compose page into a collection of issues. I have run into a problem I would like some help with. For example I have the following type of issue in the Java Script Console: Undefined Value: line 328 Which gives me this: return event.stop();},eventBeforeUnload:function(event){if(this.changed || this.editor.changed){event.returnValue = Template.process("UnsavedChanges");return event.returnValue;} Now, most likely in the scope of this event handler "this" doesn't support a particular method or function. The challenge is figuring out which one and what the context of the problem is. I tried using Drosera but it becomes unresponsive with the size of the java script file on my PPC G4. Any other good ideas on how to identify more quickly the problem? Thanks Shawn -- Shawn Smith shawnsmithj(a)gmail.com

3 2

JavascriptCore questions
by Edward Rowe 21 Dec '06

21 Dec '06

Hi Webkit folks - I'm Ed Rowe, and I work on the Apollo project at Apollo. I've got some questions about JavascriptCore that I'm hoping to get answered. The background is that I am trying to write a full-featured subclass of JSObject that forwards to an object in a different script engine (Flash's ActionScript in this case) and vice versa. Doing this correctly requires me to understand a number of details, some of which I admit seem somewhat pedantic. I have read all of the code for JSValue, JSObject, most of the binding code, etc. in order to figure things out on my own to the degree that it is possible. Now for the questions: Interpreters, ExecStates, Binding --------------------------------- - ConvertValueToNPVariant() and convertValueToObjcValue(), which are used in the plugin interface binding code, both contain special logic that basically says "if wrapping up a window object, store the RootObject associated with the window object in the wrapper rather than the current RootObject". Can you tell me the purpose of this logic? I have theories, but nothing definitive. - The JavaScriptObject struct contains pointers to RootObjects. I may be missing something, but it appears to me that a JavaScriptObject struct can outlive the RootObjects that it points at, which could lead to the use of freed memory. Can this happen? - What are the rules around exceptions in the ExecState retrieved from Interpreter::globalExec()? It appears that most of the uses of it assume that there is no exception already in the ExecState when it is retrieved and clear any exception and before returning ensure that it contains no exception. Is this correct? - When a plugin is unloaded, what happens if there are existing RuntimeObjectImp instances that reference objects provided by the plugin? Is there logic to prevent a plugin from being unloaded if there are extant objects that it has provided? I haven't been able to find any code that ensures this. - There are a variety of places where exec->dynamicInterpreter()->globalObject() is called. For example, this happens in FunctionProtoFunc::callAsFunction() to handle the apply case when the this arg is NULL. In such cases, is it correct to use the global obj from the current ExecState's Interpreter as opposed to the global object at the root of the scope chain? ECMA 262 section 15.3.4.4 seems somewhat ambiguous on this subject, but my reading leads me to think that it would be more correct to pull from the scope chain. JSObject and JSValue -------------------- - Can you confirm that NULL is never an immediate JSValue and can be used to mean "no JSValue"? - Is it legal for JSObject::hasProperty(), JSObject::canPut(), or JSObject::deleteProperty() to set an exception in the passed ExecState? I cannot find any current code doing this, but I'm not sure if it is a case I need to handle or not. - Is it legal for JSObject::get() to ever return NULL? I cannot find any current code doing this, but I'm not sure if it is a case I need to handle or not. - Is it legal to set an exception in the passed ExecState in PropertySlot's GetValueFunc? I can't find any precedent in the code for a propgetter (or get itself for that matter) to throw. - Is it legal for JSObject::call() to ever return NULL? I cannot find any current code doing this, but I'm not sure if it is a case I need to handle or not. - Is the result of JSValue::type() expected to be immutable for a given JSValue instance? That is, can a particular instance return different types at different times? - Are the results of JSObject::implementsConstruct(), JSObject::implementsCall(), and JSObject::implementsHasInstance() expected to be immutable for a given JSObject instance? That is, can a particular instance return different values at different times? - The default implementation of JSObject::defaultValue() and JSObject::toPrimitive() can return an Error object if an exception occurs. This would seem to violate ECMA 262 sections 8.6.2.6 and 9.1 which state that defaultvalue() and toPrimitive() return non-Object types. Looking at some of the callers of toPrimitive(), it appears that some are not prepared to handle getting an object back. I'll file this as a bug if you guys like. - Neither JSObject::propertyIsEnumerable() nor JSObject::getPropertyAttributes() are virtual, making it impossible for a subclass to override the default behavior. Any objection to making propertyIsEnumerable() virtual, and making getPropertyAttributes() private? Thanks in advance for the help! Ed

2 3

Improving ability to filter the tags produced by editable webkit?
by Dan Wood 21 Dec '06

21 Dec '06

Hi folks, I've been working with the webkit editing APIs for well over a year now, and I've long had some frustrations with the quality of the HTML markup that is produced when one edits a block of text with webkit. I've mentioned this over the months to various webkats and webkittens, but I haven't really gotten anywhere with it, so I thought I would post my thoughts here and see if I can get some discussion going, and maybe we all can draw up a spec on how to improve WebKit. I have a number of concerns about the HTML markup that I'd lke to address: * Being able to control/prevent insertion of apple-only and/or webkit- only tags and styles * Being able to control/prevent certain kinds of tags and style tags from being inserted, to keep markup simpler and perhaps prevent certain adjustments like changing fonts/colors * Having control over how "physical" attributes (like boldface) get marked up (e.g. <b> or <strong> or <div style="font-weight:bold;">) * Better normalization of tags so you never get two identical, adjacent style spans; they would be coalesced into one. * Semi-intelligent mapping of "physical" attributes to predefined styles classes * Better use of CSS short-hand, e.g. use the "font:" property instead of font-family and font-size * Be able to specify how plain text is dealt with when it's pasted in; is it blocked within <pre> tags, separated by <br/> tags, or each line enclosed in <div> tags. As a launching point, I wanted to point out some existing APIs in NSAttributedString, namely -[NSAttributedString dataFromRange:documentAttributes:error:]. You can pass in NSHTMLTextDocumentType to convert an attributed string to HTML, and you can then pass in NSExcludedElementsDocumentAttribute, described as "An NSArray object containing NSString objects, representing HTML elements not to be used in generated HTML." These attributes are documented on this page <http://developer.apple.com/releasenotes/ Cocoa/AppKit.html>. WebKit actually uses this technique in -[WebHTMLView _documentFragmentFromPasteboard:...], for what it's worth. I think that ideally how this would best work would be a method or some methods in the editing delegate, where you could specify the kind of markup that is allowed and specify other options. This would, of course, affect the actual editing behavior while editing happens, e.g. if you were to exclude the special apple/webkit tags that provide fine control over the text (such as allowing multiple spaces), then the text would reflect that while editing. Comments? -- Dan Wood Karelia Software — Sandvox for the Mac http://www.karelia.com/ Liberty means responsibility. That is why most men dread it. — George Bernard Shaw

2 1

Impossible to compile Webkit on Windows
by Robert Błaut 21 Dec '06

21 Dec '06

Hello Probably after changeset: > 2006-12-16 Adele Peterson <adele(a)apple.com> > RS by Adam. > Removed RenderLineEdit, TextField, WebCoreTextField classes, and > use of isNonWidgetTextField, it's impossible to compile Webkit on Windows :( Errors: RenderLineEdit.cpp c1xx : fatal error C1083: Cannot open source file: '..\..\rendering\RenderLineEdit.cpp': No such file or directory TemporaryLinkStubs.cpp ..\..\platform\win\TemporaryLinkStubs.cpp(80) : fatal error C1083: Cannot open include file: 'TextField.h': No such file or directory FontWin.cpp ..\..\platform\win\FontWin.cpp(51) : error C2039: 'pen' : is not a member of 'WebCore::GraphicsContext' d:\cygwin\home\quiris\Webkit\WebCore\platform\graphics\GraphicsContext.h(81) : see declaration of 'WebCore::GraphicsContext' ..\..\platform\win\FontWin.cpp(51) : error C2228: left of '.color' must have class/struct/union GraphicsContextCairo.cpp ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(131) : error C2228: left of '.style' must have class/struct/union type is ''unknown-type'' ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(131) : error C2653: 'Pen' : is not a class or namespace name ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(131) : error C2065: 'NoPen' : undeclared identifier ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(131) : error C3861: 'pen': identifier not found [...] ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(344) : error C3861: 'pen': identifier not found ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(345) : error C2228: left of '.width' must have class/struct/union type is ''unknown-type'' ..\..\platform\graphics\cairo\GraphicsContextCairo.cpp(345) : error C3861: 'pen': identifier not found GraphicsContext.cpp ..\..\platform\graphics\GraphicsContext.cpp(344) : error C2761: 'void WebCore::GraphicsContext::setPlatformStrokeColor(const WebCore::Color &)' : member function redeclaration not allowed ..\..\platform\graphics\GraphicsContext.cpp(345) : error C2447: '{' : missing function header (old-style formal list?) -- Regards Robert Błaut

2 1

webkit download
by nukern 20 Dec '06

20 Dec '06

Hi I am unable to download webkit source from svn://anonsvn.opensource.apple.com both through and tortise svn version. The directions given here http://webkit.org/building/checkout.html does not work Kindly help me download the source regards nukern

3 3

Greetings from the Apollo team
by Chris Brichford 18 Dec '06

18 Dec '06

Greetings, My name is Chris Brichford. I am an engineer working on Adobe's Apollo project. Apollo is a cross OS/cross device runtime for creating desktop rich internet applications. You can learn about Apollo by visiting http://www.adobe.com/go/apollo <http://www.adobe.com/go/apollo> As you may have heard we are using WebKit to implement HTML rendering capability in Apollo. I have been and will continue working on integrating WebKit with Apollo. I'd also like to introduce Ed Rowe ( erowe at adobe dot com ) who is the leading the Apollo effort and Brent Rosenquist ( brosenqu at adobe dot com ) who is working with me on integrating WebKit with Apollo. The three of us have been lurking on IRC and the email lists for a little while now and we will be becoming more active in the coming new year. Specifically we are hoping to get top of tree WebKit merged into Apollo and then start the process of contributing our changes back to the community. Initially our changes to the core code will be limited to minor bug fixes and small changes needed to enable our port of WebKit. We look forward to getting know and working with you all in the coming months. Chris Brichford Adobe Systems Inc.

2 1

Re: [webkit-dev] webkit download
by nukern 18 Dec '06

18 Dec '06

Hi Even with cygwin client I cannot connect Get the following message svn: Can't connect to host 'anonsvn.opensource.apple.com':Connection timed out I cant ping 'anonsvn.opensource.apple.com' either. Does this node exist? I have been trying since 4 days now, with same result. Thanks nukern On 12/18/06, Eric Seidel <eric(a)eseidel.com> wrote: > The instructions posted on webkit.org should work for the CYGWIN > client. I'm not sure about Tortoise. > > http://trac.webkit.org/projects/webkit/wiki/BuildingOnWindows > > Has additional instructions for working in a windows environment. > > You might also try asking in #webkit as you're bound to find someone > there who has built WebKIt on windows. > http://webkit.org/contact.html > > -eric > > On Dec 18, 2006, at 1:16 AM, nukern wrote: > > > Hi > > > > I am unable to download webkit source from svn:// > > anonsvn.opensource.apple.com > > both through and tortise svn version. > > > > The directions given here > > http://webkit.org/building/checkout.html > > > > does not work > > > > Kindly help me download the source > > > > regards > > nukern > > _______________________________________________ > > webkit-dev mailing list > > webkit-dev(a)lists.webkit.org > > http://lists.webkit.org/mailman/listinfo/webkit-dev > >

2 1