Re: [g-a-devel] [Kde-accessibility] [Accessibility] Re: [Accessibility-atspi] D-Bus AT-SPI - The way forward

[Willie Walker <William Walker Sun COM>]

Hi Rob:

Though I helped create the Java accessibility API proper, I'm not quite 
sure how the CORBA stuff is handled.  I've only worked on a few bugs in 
the java-access-bridge (see 
http://svn.gnome.org/svn/java-access-bridge/trunk), it's still a slight 
bit beyond me.

It's probably obvious, but the main code seems to live in these 
directories.  The one I play in the most is the 'impl' one:

http://svn.gnome.org/svn/java-access-bridge/trunk/bridge/org/GNOME/Accessibility/
http://svn.gnome.org/svn/java-access-bridge/trunk/impl/org/GNOME/Accessibility/
http://svn.gnome.org/svn/java-access-bridge/trunk/registry/org/GNOME/Accessibility/

As part of the build, I believe an idlgen pass is done that creates all 
the stub/skeleton stuff that lives under the idlgen directory.  I'm 
going to be digging a bit more in there this week, so I hope to learn more.

Will

Rob Taylor wrote:
> Michael Meeks wrote:
> > On Mon, 2007-12-10 at 21:57 +0000, Rob Taylor wrote:
> > > Supporting com.sun.java.accessibility shouldn't be hard, but we really
> > > need with some input from people who understand how accessibility is
> > > exposed by AWT/SWT/Swing..
> > 	Surely there should not be a per-toolkit wrt. simply bridging to a
> > different IPC mechanism. And indeed, the plus for Java of course would
> > be that it should be faster than using TCP sockets, more secure, and (of
> > course) will work on stock Linux systems (that disable IPv4/6 sockets).
>
> I think I could definitly do with a bit more of a background on how java
> toolkits expose themselves. Could you quickly go though what components
> do which bits and how corba currently ties in?
>
> > > > I'm a bit confused by the slowdown, though.  I thought that programs
> > > > that use UNIX sockets to connect to the ORBit2 server will continue to
> > > > do so even when TCP/IP is enabled.  My understanding was that enabling
> > > > TCP/IP with ORBit2 just made it possible for programs that want to use
> > > > TCP/IP to also be able to connect to the ORBit2 server (such as Java
> > > > programs).
> > > Well, the slowdown occurs when you disable local sockets, so no suprise
> > > there :)
> > 	I'd also expect a (small) slowdown just enabling IPv4 sockets, whether
> > they are used or not (and they're not preferred clearly), since in
> > itself that ~doubles the size of each object profile we marshal.
>
> Actually we did seem see that, but it was small enough a difference  (~
> 5%) that it could have been noise.
>
> Rob
> > 	HTH,
> >
> > 		Michael.