Re: [webkit-dev] Fwd: [GTK]Scroll deficiency

2009-06-02 Thread Xiong
Hi,
I am happy to hear that you solve the issue.

I am also sorry to say i can not search the bug
report, and never submit the patch with my sloth.

-Xiong

2009/6/2 Srinivas Rao M Hamse msrini...@gmail.com

 Hi Xiong,

 My platform is WebKiton gtk+directfb. I am noticing that the issue is still
 present in the current Webkit nightly R-44111 build. I saw your sugested
 patch :

   In function ChromeClient::scroll()

 //   gdk_window_move_region(window, moveRegion, delta.width(),
 delta.height());
 //   gdk_region_offset(moveRegion, delta.width(), delta.height());
   gdk_window_move_region(window, moveRegion, moveRect.width,
 moveRect.height);
   gdk_region_offset(moveRegion, moveRect.width, moveRect.height);

 After applying it, I am not finding the scroll deficiency issue.
 I wish to know if this fix is working, why is it not landed yet. What is
 the plan for this bug.
 What is the bug number for Gtk+Directfb platform.

 Thanks and regards,
 Srinivas Rao.M http://msrinirao.blogspot.com





 On Wed, Apr 22, 2009 at 8:04 AM, Xiong browse...@gmail.com wrote:



 -- Forwarded message --
 From: Xiong browse...@gmail.com
 Date: Wed, Apr 22, 2009 at 9:42 AM
 Subject: Re: [webkit-dev] [GTK]Scroll deficiency
 To: Frank Thomsen f...@trifork.com
 Cc: Brent Fulgham bfulg...@gmail.com


 Hi Frank

 I am working on gtk+directfb and Brent working on Window+cairo, and
 you ?

 As you see, Brent and i have two separate bugs, because the
 workarounds he described
 do not help me.

 For now, i have resolved the problem in my platform.

 The reason why the screen display confusion is that GDK do not refresh
 the
 overlapping region when scrolling.

 So, change the codes as following:

 In function ChromeClient::scroll()

 //   gdk_window_move_region(window, moveRegion, delta.width(),
 delta.height());
 //   gdk_region_offset(moveRegion, delta.width(), delta.height());
   gdk_window_move_region(window, moveRegion, moveRect.width,
 moveRect.height);
   gdk_region_offset(moveRegion, moveRect.width, moveRect.height);

 After do that, the problem was gone, at least in my
 platform(Gtk+directfb).

 Good luck.

 -xiong

 2009/4/17 Frank Thomsen f...@trifork.com

 Hi guys

 Yesterday I wrote that I had solved the problem - at least on my machine.
 Well, I was wrong. Apparently the redraw-problem only exists in
 release-mode.
 Could you please tell me what you have done to circumvent this problem?
 You have talked about both commenting out code and doing something with the
 ... background..? Whatever it is I cannot find it in the mailing list, so I
 would really appreciate if you could enlighten me. And do you know if this
 has been reported as a bug? Can't seem to find that either...

 Thanks in advance!
 Frank

 -Original Message-
 From: webkit-dev-boun...@lists.webkit.org [mailto:
 webkit-dev-boun...@lists.webkit.org] On Behalf Of Brent Fulgham
 Sent: 31. marts 2009 01:04
 To: xiong
 Cc: webkit-dev@lists.webkit.org
 Subject: Re: [webkit-dev] [GTK]Scroll deficiency

 Hi Xiong,

 Thanks for trying these suggestions.

 On Sun, Mar 29, 2009 at 8:24 PM, xiong browse...@gmail.com wrote:
  Brent,
 
  In my platform, all the hack as you metioned are valueless.
 
  Bad news to me.
 
  Now, i only have a clue on cario for the scroll problem.

 I think we may have two separate (but perhaps related) bugs, because
 the workarounds I described do not help your problem.  Furthermore,
 the change you suggested (comment out the if (contentChanged) test
 in ChromeClient::repaint ()) does not resolve my problem, while it
 does fix yours.

 Thanks,

 -Brent
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev




 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev




 --
 Srinivas Rao M  Hamse


___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] Newbie question: any sample codes to use C bridge

2009-05-26 Thread Xiong
Hi,
Maybe you need touch the directory of JavaScriptCore/API/ and the file of
JavaScriptCore/API/test/testapi.c.

Thanks.

-Xiong

On Tue, May 26, 2009 at 5:26 PM, Harry Zhang zhangha...@gmail.com wrote:

 Dear all, can anyone kindly share some code snippets on how to create and
 attach an JS object, e.g. window.foo to window object via C bridge?
 (btw, my dev environment is Ubuntu, i want to enable this in webkitgtk+)

 Many thanks in advance!

 Regards,
 Harry

 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] JSObjectCallAsFunctionCallback question

2009-05-17 Thread Xiong
Hi, Su

Yes, this is the key point of the problem.

In my program, many Object use the same staticFunc entry[Program
Architecture].

So, in the getValue_cb callback function, we need do extra two things:

1.Get the Object name, here is myObject; [WE CAN NOW]
2.Get the propertyName of the method, here is getValues; [WE CAN NOT NOW]

After both get the Object and method's name, we can identify another
callback function.

Therefore, implement a way to get the propertyName of the method via
'function object' is needed.

Thanks.

-Xiong

On Fri, May 15, 2009 at 3:28 PM, Zhe Su james...@gmail.com wrote:

 AFAIK, there is no way to achieve what you want by using staticFunc
 structure. Of course, if you use different C functions for each staticFunc
 entry, you actually don't need the name.

 Regards
 James Su


 On Fri, May 15, 2009 at 3:24 PM, Xiong browse...@gmail.com wrote:

 Thanks Su.

 I think i know you mean.

 To get the method name [getValue] via attach a private data to function
 object which create by JSObjectMake().

 Yes, it's a better way, but not for me.

 In my program, we have many Object and it's method and now all in
 staticFunc structure.

 Thanks again.

 -Xiong


 On Fri, May 15, 2009 at 11:57 AM, Zhe Su james...@gmail.com wrote:

 If you create a JSObject with JSObjectMake() function, you can have a
 private data associated to the newly created JSObject. You can store
 anything you want in the private data. So following approach might be
 feasible for you:

 1. define a special JSClass for your function, which has finalize and
 callAsFunction set. In finalize() method, you need destroy the private data
 associated to the JSObject, if necessary. callAsFunction points to your
 getValue_cb.
 2. create the getValue function object by calling JSObjectMake() with the
 function name getValue as private data.
 3. attach the function object to myObject with name getValue. It might
 be done in myObject class's initialize() method.
 4. in getValue_cb function you can retrieve the private data (which
 contains the name) by calling JSObjectGetPrivate(function).

 The only drawback of this approach: you can't declare the function
 statically in staticFunctions.

 Regards
 James Su

 On Fri, May 15, 2009 at 10:53 AM, Xiong browse...@gmail.com wrote:

 Thanks Darin.

 For the first problem, as you said, all the functions are objects, and
 also have **properties**.

 So,can we add a member for the function object to store the method
 name[here is getValues] in theoretical ?

 And this also like the JSObjectRef object variable in the callback
 prototype,we can get the class name like this:

 toJS(object)-className().data()

 Thanks.

 -Xiong


 On Fri, May 15, 2009 at 10:29 AM, Darin Adler da...@apple.com wrote:

 On May 14, 2009, at 7:03 PM, Xiong wrote:

  1what is the purpose of the JSObjectRef function variable? In my
 case, we never use.


 It’s fine to not use this if you don’t need to. In JavaScript,
 functions are objects. Like other objects they can have properties. This
 argument is passed to your C implementation so that it can get access to
 properties of the function object if it has some reason to. In many cases
 there is no need to do this.

  2In this function, i want to get the propertyName of the method,
 in this case, we need getValues, can be implemented ?


 No, there's no way to do that, for a good architectural reason.

 A JavaScript function is a first class object. Using JSObjectMake will
 create an object with a prototype that has functions stored in its
 properties, but there’s nothing to prevent the JavaScript program from
 taking that function and storing it somewhere else, in a property with
 another name. When the function is called, it’s just an object, and 
 there’s
 no way to go back in time and find out where that object came from.

 Consider the similar situation of a data property. When you get the
 length of an array, it's a number 1, not a number 1 that knows it came 
 from
 a property named length.

-- Darin



 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev





___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] JSObjectCallAsFunctionCallback question

2009-05-15 Thread Xiong
Thanks Su.

I think i know you mean.

To get the method name [getValue] via attach a private data to function
object which create by JSObjectMake().

Yes, it's a better way, but not for me.

In my program, we have many Object and it's method and now all in staticFunc
structure.

Thanks again.

-Xiong

On Fri, May 15, 2009 at 11:57 AM, Zhe Su james...@gmail.com wrote:

 If you create a JSObject with JSObjectMake() function, you can have a
 private data associated to the newly created JSObject. You can store
 anything you want in the private data. So following approach might be
 feasible for you:

 1. define a special JSClass for your function, which has finalize and
 callAsFunction set. In finalize() method, you need destroy the private data
 associated to the JSObject, if necessary. callAsFunction points to your
 getValue_cb.
 2. create the getValue function object by calling JSObjectMake() with the
 function name getValue as private data.
 3. attach the function object to myObject with name getValue. It might be
 done in myObject class's initialize() method.
 4. in getValue_cb function you can retrieve the private data (which
 contains the name) by calling JSObjectGetPrivate(function).

 The only drawback of this approach: you can't declare the function
 statically in staticFunctions.

 Regards
 James Su

 On Fri, May 15, 2009 at 10:53 AM, Xiong browse...@gmail.com wrote:

 Thanks Darin.

 For the first problem, as you said, all the functions are objects, and
 also have **properties**.

 So,can we add a member for the function object to store the method
 name[here is getValues] in theoretical ?

 And this also like the JSObjectRef object variable in the callback
 prototype,we can get the class name like this:

 toJS(object)-className().data()

 Thanks.

 -Xiong


 On Fri, May 15, 2009 at 10:29 AM, Darin Adler da...@apple.com wrote:

 On May 14, 2009, at 7:03 PM, Xiong wrote:

  1what is the purpose of the JSObjectRef function variable? In my
 case, we never use.


 It’s fine to not use this if you don’t need to. In JavaScript, functions
 are objects. Like other objects they can have properties. This argument is
 passed to your C implementation so that it can get access to properties of
 the function object if it has some reason to. In many cases there is no need
 to do this.

  2In this function, i want to get the propertyName of the method, in
 this case, we need getValues, can be implemented ?


 No, there's no way to do that, for a good architectural reason.

 A JavaScript function is a first class object. Using JSObjectMake will
 create an object with a prototype that has functions stored in its
 properties, but there’s nothing to prevent the JavaScript program from
 taking that function and storing it somewhere else, in a property with
 another name. When the function is called, it’s just an object, and there’s
 no way to go back in time and find out where that object came from.

 Consider the similar situation of a data property. When you get the
 length of an array, it's a number 1, not a number 1 that knows it came from
 a property named length.

-- Darin



 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev



___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


[webkit-dev] JSObjectCallAsFunctionCallback question

2009-05-14 Thread Xiong
Hello all,

I have implement a JS object named myObject with the JavaScriptCore(JSC)
API, and it has a method named getValues(),  called as following:
myObject.getValues();

And it define like this:
JSStaticFunction staticFunc[] = {
{getValues, getValue_cb, kJSPropertyAttributeReadOnly},
{NULL, NULL, NULL, NULL}
};

The prototype of getValue_cb callback function like this:
JSValueRef getValue_cb(JSContextRef context, JSObjectRef function,
JSObjectRef object,
size_t argumentCount, const JSValueRef arguments[],
JSValueRef *exception)

I have two problems about the getValue_cb callback function:
1what is the purpose of the JSObjectRef function variable? In my case,
we never use.
2In this function, i want to get the propertyName of the method, in this
case, we need getValues, can be implemented ?

Appreciate any clues~

Thanks.

-Xiong
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] JSObjectCallAsFunctionCallback question

2009-05-14 Thread Xiong
Thanks Darin.

For the first problem, as you said, all the functions are objects, and also
have **properties**.

So,can we add a member for the function object to store the method
name[here is getValues] in theoretical ?

And this also like the JSObjectRef object variable in the callback
prototype,we can get the class name like this:

toJS(object)-className().data()

Thanks.

-Xiong

On Fri, May 15, 2009 at 10:29 AM, Darin Adler da...@apple.com wrote:

 On May 14, 2009, at 7:03 PM, Xiong wrote:

  1what is the purpose of the JSObjectRef function variable? In my
 case, we never use.


 It’s fine to not use this if you don’t need to. In JavaScript, functions
 are objects. Like other objects they can have properties. This argument is
 passed to your C implementation so that it can get access to properties of
 the function object if it has some reason to. In many cases there is no need
 to do this.

  2In this function, i want to get the propertyName of the method, in
 this case, we need getValues, can be implemented ?


 No, there's no way to do that, for a good architectural reason.

 A JavaScript function is a first class object. Using JSObjectMake will
 create an object with a prototype that has functions stored in its
 properties, but there’s nothing to prevent the JavaScript program from
 taking that function and storing it somewhere else, in a property with
 another name. When the function is called, it’s just an object, and there’s
 no way to go back in time and find out where that object came from.

 Consider the similar situation of a data property. When you get the length
 of an array, it's a number 1, not a number 1 that knows it came from a
 property named length.

-- Darin


___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] JSObjectCallAsFunctionCallback question

2009-05-14 Thread Xiong
Darin, thanks for you help.

I think we must implement the extra member for function object in my
program architecture.

Maybe more work.

Thanks again.

-Xiong

On Fri, May 15, 2009 at 11:32 AM, Darin Adler da...@apple.com wrote:

 On May 14, 2009, at 7:53 PM, Xiong wrote:

  So,can we add a member for the function object to store the method
 name[here is getValues] in theoretical ?


 Yes, in theory you could do that. Although that’s not the typical pattern.

  And this also like the JSObjectRef object variable in the callback
 prototype,we can get the class name like this:

 toJS(object)-className().data()


 No, you can’t do that. Those three functions are part of the internals and
 not part of the JavaScriptCore API.

 Instead, if you want to know if an object has a particular class you can
 use JSValueIsObjectOfClass.

-- Darin


___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] How to detect the parameters in object tag?

2009-05-12 Thread Xiong
Hi,

For the first question, you can implement a NPAPI plugin.

You can get all the parameters in the plugin-side functions
API(NPPluginFuncs-newp).

Thanks.

-Xiong

On Wed, May 13, 2009 at 8:56 AM, naixuan guan guannaix...@gmail.com wrote:

 Hi, everyone

 Recently, I met a HTML page which has a object tag like this:

 OBJECT id=player style=LEFT: 0px; WIDTH:100%; TOP: 0px; HEIGHT:100%
 codebase= classid=
 param name=URL value=?
 PARAM NAME=ShowUI VALUE=1
 PARAM NAME=volume VALUE=90
 PARAM NAME=autoPlay VALUE=2
 PARAM NAME=mute VALUE=0
 PARAM NAME=DropEnable VALUE=0
 PARAM NAME=EnableFullScreen VALUE=0
 PARAM NAME=ShowVideo VALUE=1
 param name=loop value=2/
 /OBJECT
 script language=JavaScript
 document.player.URL = urlpath+file;
 /Script

 I have two questions
 1.how to detect the parameters in object tag?
 2.specially, how to detect the URL param which is changed by Javascript?

 Thank you!

 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


[webkit-dev] Background color issue

2009-04-27 Thread Xiong
Dear all,

I am working webkit over directfb backend on arm platform - using
GtkLauncher.

The problem I face is that the background color of loading image is always
WHITE, i am really want to be BLACK.

BTW:
I use the gtk_widget_modify_bg() to change color of GTKWINDOW and
WebKitWebView,
but it looks like valueless.

Also, i connect to expose-event signal for WebKitWebView and setting the
background color in following code:
[
cairo_set_source_rgb(cr, 0, 0, 0);
cairo_paint(cr);
]
This way was failed also.

The Gtk+ API does not change the background color, so may be the webkit
kernel do something i think.

Any ideas where to look to fix this issue ?

Appreciate any clues~

-Xiong
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


[webkit-dev] Fwd: [GTK]Scroll deficien cy

2009-04-21 Thread Xiong
-- Forwarded message --
From: Xiong browse...@gmail.com
Date: Wed, Apr 22, 2009 at 9:42 AM
Subject: Re: [webkit-dev] [GTK]Scroll deficiency
To: Frank Thomsen f...@trifork.com
Cc: Brent Fulgham bfulg...@gmail.com


Hi Frank

I am working on gtk+directfb and Brent working on Window+cairo, and you
?

As you see, Brent and i have two separate bugs, because the workarounds
he described
do not help me.

For now, i have resolved the problem in my platform.

The reason why the screen display confusion is that GDK do not refresh
the
overlapping region when scrolling.

So, change the codes as following:

In function ChromeClient::scroll()

//   gdk_window_move_region(window, moveRegion, delta.width(),
delta.height());
//   gdk_region_offset(moveRegion, delta.width(), delta.height());
  gdk_window_move_region(window, moveRegion, moveRect.width,
moveRect.height);
  gdk_region_offset(moveRegion, moveRect.width, moveRect.height);

After do that, the problem was gone, at least in my
platform(Gtk+directfb).

Good luck.

-xiong

2009/4/17 Frank Thomsen f...@trifork.com

Hi guys

 Yesterday I wrote that I had solved the problem - at least on my machine.
 Well, I was wrong. Apparently the redraw-problem only exists in
 release-mode.
 Could you please tell me what you have done to circumvent this problem? You
 have talked about both commenting out code and doing something with the ...
 background..? Whatever it is I cannot find it in the mailing list, so I
 would really appreciate if you could enlighten me. And do you know if this
 has been reported as a bug? Can't seem to find that either...

 Thanks in advance!
 Frank

 -Original Message-
 From: webkit-dev-boun...@lists.webkit.org [mailto:
 webkit-dev-boun...@lists.webkit.org] On Behalf Of Brent Fulgham
 Sent: 31. marts 2009 01:04
 To: xiong
 Cc: webkit-dev@lists.webkit.org
 Subject: Re: [webkit-dev] [GTK]Scroll deficiency

 Hi Xiong,

 Thanks for trying these suggestions.

 On Sun, Mar 29, 2009 at 8:24 PM, xiong browse...@gmail.com wrote:
  Brent,
 
  In my platform, all the hack as you metioned are valueless.
 
  Bad news to me.
 
  Now, i only have a clue on cario for the scroll problem.

 I think we may have two separate (but perhaps related) bugs, because
 the workarounds I described do not help your problem.  Furthermore,
 the change you suggested (comment out the if (contentChanged) test
 in ChromeClient::repaint ()) does not resolve my problem, while it
 does fix yours.

 Thanks,

 -Brent
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev

___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [GTK]Scroll deficiency

2009-03-29 Thread xiong
Brent,

I just start to resolve the problem.

The debug build and background color hack you mentioned, i also
need time to try.

Please wait a reply.

I hope the member in the mail-list to join in the work, when see the subject
if you have the same symptom.

Thanks.

-Xiong

On Fri, Mar 27, 2009 at 1:54 PM, Brent Fulgham bfulg...@gmail.com wrote:

 Xiong,
 I'm sorry you are encountering the problem, but I must admit I am glad
 someone else is seeing this issue as now we can pool our efforts to resolve
 it.

 Since you are using directfb stuff, and I am on top of Windows, it seems
 the culprit is somewhere in the Cairo logic.  Perhaps the dirty regions are
 not being computed properly (I have encountered a few places where the
 coordinate systems in Cairo and Windows don't agree).

 However, the kludge of setting a background color (does this work for you
 as well?) does not fit into the 'bad coordinate' theory, as I don't see how
 the presence of a background color in the rendering instructions would
 affect this.

 The fact that it does not happen for me in Debug builds makes me wonder if
 maybe there is some kind of improper initialization or buffer overrun that
 is involved.

 -Brent


 On Mar 26, 2009, at 10:48 PM, xiong wrote:

 Webkit r40777 Release builds for me.

 On Fri, Mar 27, 2009 at 1:25 PM, Brent Fulgham bfulg...@gmail.com wrote:

 I wonder if your issue is related to one I documented on Windows (using
 Cairo backend) a few months ago.  Currently this only happens for me in
 Release builds, but it's still quite annoying.

 https://bugs.webkit.org/show_bug.cgi?id=22891

 I attached a few images in my bug for comparison.

 The easiest way to reproduce is to go to the http://planet.webkit.org and
 try to scroll up/down the page using the scroll bar.  After a short while
 you will find areas of the page not rendered.

 Thanks,

 -Brent


 On Mar 26, 2009, at 9:54 PM, browserwk wrote:

  Dear all:

 I am able to run webkit over directfb backend on arm platform - using
 GtkLauncher .

 The problem I face is that when I open
 http://www.google.com/intl/en/privacy.html using this browser and scroll
 the page; it results the visual area is not updated when scrolling, and
 the screen display confusion.

 Any ideas where to look to fix this issue ?

 Appreciate any clues~

 FYI:
 [Following address are the screenshots]

 Render OK when not scroll:

 http://picasaweb.google.com/lh/photo/ybhIcahTEgsfBmMzXM13-g?feat=directlink

 Render error when scroll:

 http://picasaweb.google.com/lh/photo/0B8fi1ZTZ7hWH0S86a5RJA?feat=directlink

 http://picasaweb.google.com/lh/photo/_ScHyJBOmJ8-xbRB_aZ2NQ?feat=directlink

 http://picasaweb.google.com/lh/photo/Ael-gWdHEVnEtu_qbMFOZw?feat=directlink

 BTW:
 I am using following version of packages and running DirectFB without
 any acceleration on my embeded target.
 WebKit-r40777
 gtk+-2.12.12
 cairo-1.4.14
 DirectFB-1.2.6

 Thanks.

 -xiong
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev





___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [GTK]Scroll deficiency

2009-03-29 Thread xiong
Brent,

In my platform, all the hack as you metioned are valueless.

Bad news to me.

Now, i only have a clue on cario for the scroll problem.

Thanks

-xiong

On Mon, Mar 30, 2009 at 10:02 AM, xiong browse...@gmail.com wrote:

 Brent,

 I just start to resolve the problem.

 The debug build and background color hack you mentioned, i also
 need time to try.

 Please wait a reply.

 I hope the member in the mail-list to join in the work, when see the
 subject if you have the same symptom.

 Thanks.

 -Xiong


 On Fri, Mar 27, 2009 at 1:54 PM, Brent Fulgham bfulg...@gmail.com wrote:

 Xiong,
 I'm sorry you are encountering the problem, but I must admit I am glad
 someone else is seeing this issue as now we can pool our efforts to resolve
 it.

 Since you are using directfb stuff, and I am on top of Windows, it seems
 the culprit is somewhere in the Cairo logic.  Perhaps the dirty regions are
 not being computed properly (I have encountered a few places where the
 coordinate systems in Cairo and Windows don't agree).

 However, the kludge of setting a background color (does this work for you
 as well?) does not fit into the 'bad coordinate' theory, as I don't see how
 the presence of a background color in the rendering instructions would
 affect this.

 The fact that it does not happen for me in Debug builds makes me wonder if
 maybe there is some kind of improper initialization or buffer overrun that
 is involved.

 -Brent


 On Mar 26, 2009, at 10:48 PM, xiong wrote:

 Webkit r40777 Release builds for me.

 On Fri, Mar 27, 2009 at 1:25 PM, Brent Fulgham bfulg...@gmail.comwrote:

 I wonder if your issue is related to one I documented on Windows (using
 Cairo backend) a few months ago.  Currently this only happens for me in
 Release builds, but it's still quite annoying.

 https://bugs.webkit.org/show_bug.cgi?id=22891

 I attached a few images in my bug for comparison.

 The easiest way to reproduce is to go to the http://planet.webkit.organd 
 try to scroll up/down the page using the scroll bar.  After a short
 while you will find areas of the page not rendered.

 Thanks,

 -Brent


 On Mar 26, 2009, at 9:54 PM, browserwk wrote:

  Dear all:

 I am able to run webkit over directfb backend on arm platform - using
 GtkLauncher .

 The problem I face is that when I open
 http://www.google.com/intl/en/privacy.html using this browser and
 scroll
 the page; it results the visual area is not updated when scrolling, and
 the screen display confusion.

 Any ideas where to look to fix this issue ?

 Appreciate any clues~

 FYI:
 [Following address are the screenshots]

 Render OK when not scroll:

 http://picasaweb.google.com/lh/photo/ybhIcahTEgsfBmMzXM13-g?feat=directlink

 Render error when scroll:

 http://picasaweb.google.com/lh/photo/0B8fi1ZTZ7hWH0S86a5RJA?feat=directlink

 http://picasaweb.google.com/lh/photo/_ScHyJBOmJ8-xbRB_aZ2NQ?feat=directlink

 http://picasaweb.google.com/lh/photo/Ael-gWdHEVnEtu_qbMFOZw?feat=directlink

 BTW:
 I am using following version of packages and running DirectFB without
 any acceleration on my embeded target.
 WebKit-r40777
 gtk+-2.12.12
 cairo-1.4.14
 DirectFB-1.2.6

 Thanks.

 -xiong
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev






___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [GTK]Scroll deficiency

2009-03-26 Thread xiong
Webkit r40777 Release builds for me.

On Fri, Mar 27, 2009 at 1:25 PM, Brent Fulgham bfulg...@gmail.com wrote:

 I wonder if your issue is related to one I documented on Windows (using
 Cairo backend) a few months ago.  Currently this only happens for me in
 Release builds, but it's still quite annoying.

 https://bugs.webkit.org/show_bug.cgi?id=22891

 I attached a few images in my bug for comparison.

 The easiest way to reproduce is to go to the http://planet.webkit.org and
 try to scroll up/down the page using the scroll bar.  After a short while
 you will find areas of the page not rendered.

 Thanks,

 -Brent


 On Mar 26, 2009, at 9:54 PM, browserwk wrote:

  Dear all:

 I am able to run webkit over directfb backend on arm platform - using
 GtkLauncher .

 The problem I face is that when I open
 http://www.google.com/intl/en/privacy.html using this browser and scroll
 the page; it results the visual area is not updated when scrolling, and
 the screen display confusion.

 Any ideas where to look to fix this issue ?

 Appreciate any clues~

 FYI:
 [Following address are the screenshots]

 Render OK when not scroll:

 http://picasaweb.google.com/lh/photo/ybhIcahTEgsfBmMzXM13-g?feat=directlink

 Render error when scroll:

 http://picasaweb.google.com/lh/photo/0B8fi1ZTZ7hWH0S86a5RJA?feat=directlink

 http://picasaweb.google.com/lh/photo/_ScHyJBOmJ8-xbRB_aZ2NQ?feat=directlink

 http://picasaweb.google.com/lh/photo/Ael-gWdHEVnEtu_qbMFOZw?feat=directlink

 BTW:
 I am using following version of packages and running DirectFB without
 any acceleration on my embeded target.
 WebKit-r40777
 gtk+-2.12.12
 cairo-1.4.14
 DirectFB-1.2.6

 Thanks.

 -xiong
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev



___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev