Re: [webkit-dev] Need help in understanding how webkit handles clicks event

[Alex Russell]

The PDF handling stuff is related to a mimetype filter. The web server  
sends a particular mime type along with each document, and if it's not  
something that the browser knows how to deal with directly, it'll  
interrogate all of its plugins to see if they want to handle it. When  
plugins are registered they provide a list of types they'll handle. See:



http://developer.apple.com/DOCUMENTATION/InternetWeb/Conceptual/WebKit_PluginProgTopic/Concepts/AboutPlugins.html

And:

http://trac.webkit.org/browser/trunk/WebCore/plugins

Regards

On Jan 28, 2009, at 5:25 PM, Meryl Silverburgh wrote:


Hi,

Can you please how webkit handles a mouse click from a user?
For example, when user clicks an anchor, it will load the web page.
But if the links points to  a pdf document, it will load the document
using pdf reader.

Can you please tell me where in the code handles that?


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

[webkit-dev] Need help in understanding how webkit handles clicks event

[Meryl Silverburgh]

Hi,

Can you please how webkit handles a mouse click from a user?
For example, when user clicks an anchor, it will load the web page.
But if the links points to  a pdf document, it will load the document
using pdf reader.

Can you please tell me where in the code handles that?

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

Re: [webkit-dev] Need help in running WebKit in debugger mode

[Darin Adler]

On Jan 28, 2009, at 10:57 AM, ying lcs wrote:


bootstrap_register(): failed 1100 (0x44c) 'Permission denied', port =
0x7d03, name = 'com.apple.Safari.ServiceProvider'


That's a harmless warning you get any time you run two copies of  
Safari at the same time, when both try to add the Safari items to the  
Service menu. There’s no way for Safari to prevent this from being  
written to the console.


You can avoid it by quitting Safari before using the WebKit launcher  
to launch a new Safari with the built WebKit, or you can just ignore  
the warning.


-- Darin

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

[webkit-dev] Need help in running Webkit in debugger mode

[ying lcs]

HI,

I am trying to run Webkit in debug under Mac, but I get the following error:

$ WebKitTools/Scripts/run-safari --debug
Starting Safari with DYLD_FRAMEWORK_PATH set to point to built WebKit
in /Volumes/test/WebKit/WebKitBuild/Debug.
2009-01-28 10:51:04.314 Safari[36717:80f] *** CFMessagePort:
bootstrap_register(): failed 1100 (0x44c) 'Permission denied', port =
0x7d03, name = 'com.apple.Safari.ServiceProvider'
See /usr/include/servers/bootstrap_defs.h for the error codes.


>From the '/usr/include/servers/bootstrap_defs.h'
#define BOOTSTRAP_NOT_PRIVILEGED1100

I expected it runs in a debugger so that I can step thru the code. But
it just run as an application.  Can you please tell me what am I
missing?

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

[webkit-dev] Fwd: Getting more buildbots green

[Marc-Antoine Ruel]

Idea
One of the advantages of keeping the tree green is to enable the
implementation of a "try server". We've been using that for a while and even
with its caveat, it's particularly useful when multiple platforms are
supported by a code base.

You can see it in action there:
http://build.chromium.org/buildbot/try-server/waterfall

How we use it
When someone uploads a patch to our code review
toolwith gcl.py, it also implicitly
calls trychange.py to send it to the try
server. The try server has 2 communication methods, a svn server to hold the
.diff files (which enable access control and logging) or direct HTTP
connection (a bit faster).

Try jobs can also be manually called. In webkit's case, it'd look like this:
trychange.py --url https://bugs.webkit.org/attachment.cgi?id=26067
--patchlevel
1

It then sends back an email for each platform with the equivalent waterfall
"strip" for the build. I attached an example at the end of this email.


Sources
The try server 
master.cfg
and
its related scripts
.

trychange.py, gcl.py and
gclient.py
.

gclient needs to be used to do a functional try server checkout. (sorry)
The commands are:
  gclient config
http://src.chromium.org/svn/trunk/tools/buildbot/config/master.tryserver/
  gclient sync

Our code review tool
sources.
Not directly related, just FYI.


Applied on webkit
In the webkit case, we could probably extract something much simpler. It
doesn't need gclient, gcl or the code review tool. Only something similar to
trychange.py and the try server scripts.

I attached a sample try result email.

M-A


 CUT HERE 
http://build.chromium.org/buildbot/try-server/

TRY FAILED

http://build.chromium.org/buildbot/try-server/builders/win/builds/1106

win
Build 
1106
'maruel: disconnect' try job svnkill
stdioupdate
scripts
stdiotaskkill
stdioupdate
patch
stdio
compile
stdioStart
Crash Handler
stdioipc_tests
stdioinstaller_unittests
stdiounit_tests
13 disabled
stdiodebugger_test
stdioui_tests
crashed or hung
stdiotest_shell_tests
1 disabled
stdio
base_unittests
1 disabled
stdionet_unittests
9 disabled
stdiomedia_unittests
stdio
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev

Re: [webkit-dev] What's up with the Chromium bots?

[Darin Fisher]

On Wed, Jan 28, 2009 at 7:49 AM, Darin Adler  wrote:

> On Jan 26, 2009, at 12:11 PM, Darin Fisher wrote:
>
>  If we do remove the bots from the front page, I hope everyone understands
>> that we intend to re-enable them ASAP.
>>
>
> My take on this: Green builds are great on the buildbot page page. Red
> builds are not. If it's straightforward to change the Chromium bots to be
> green then we should do it; I don't care how little they are doing, but I
> want to make sure that they are expected to pass, not expected to fail. I
> wasn't pushing to remove them.
>
>-- Darin
>
>
It looks like the buildbots got removed.  I agree that it only makes sense
to put them back up once they have a shot at being green.  We'll work on
getting to that point soon.

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

[webkit-dev] Restrictions

[Nacho Abejaro]

Hello, I have a test application that works as a kind of demo browser, where
I change the source of all images from any HTML page using a custom
Javascript that I introduce in it through WebKit. My Javascript just does
that:

newSource = "file:///c:/test.jpg";
var images = document.getElementsByTagName("img")
for (var i=0; ihttp://anyserver/anyfile.jpg";), it works!

So, my question is: is there now any kind of restrictions on Webkit about
directly changing the image sources with Javascript and pointing them to
local files?

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

Re: [webkit-dev] Get the HTTP "request" and "response" when loading a URL

[Gustavo Noronha Silva]

Hello,

On Tue, 2009-01-27 at 09:36 +0100, ferrety ferrety wrote:
> When using "./Programs/GtkLauncher" to load a web page, let say
> "http://www.cnn.com";,
> I'd like it to output both HTTP "request" it sends to the server and
> "response" from that server,
> both printed on the shell.

That is mostly not supported API-wise as of yet, on the GTK+ port. You
can take a look at this bug report to know where things stand:

https://bugs.webkit.org/show_bug.cgi?id=18608

Take a look also at the frame loaders rework:

https://bugs.webkit.org/show_bug.cgi?id=17066

See you,

-- 
Gustavo Noronha Silva 
GNOME contributor: http://www.gnome.org/

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

Re: [webkit-dev] What's up with the Chromium bots?

[Darin Adler]

On Jan 26, 2009, at 12:11 PM, Darin Fisher wrote:

If we do remove the bots from the front page, I hope everyone  
understands that we intend to re-enable them ASAP.


My take on this: Green builds are great on the buildbot page page. Red  
builds are not. If it's straightforward to change the Chromium bots to  
be green then we should do it; I don't care how little they are doing,  
but I want to make sure that they are expected to pass, not expected  
to fail. I wasn't pushing to remove them.


-- Darin

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

[webkit-dev] Simulate clicks on the rendering Window : WebKit GTK

[ferrety ferrety]

Hi List,

I'm looking for a simple solution to simulate clicks on the
"./Programs/GtkLauncher" rendering window
from left-top corner to the right-bottom corner, pixel after pixel.

Is this doable with WebKit GTK?

N.b: is someone could help on my previous question too about
displaying HTTP headers when
loading about, It'll be very helpfull for me.

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