[Xcb] What needs to be done for keyboard support?

[Jeremy A. Kolb]

On Thu, 7 Jun 2007, Stefan O'Rear wrote:

> On Thu, Jun 07, 2007 at 10:41:28AM +0200, Vincent Torri wrote:
> > On Wed, 6 Jun 2007, Barton C Massey wrote:
> > 
> > > In message <20070606222247.GA3110 at localhost.localdomain> you wrote:
> > 
> > >> Having already looked at the xcb@ archives, I see the presense of a
> > >> util/keysym library, however it is not in Debian sid - is this just
> > >> immaturity, or is that project dead?
> > >
> > > A little of both, I think.  No one has worked on it in a
> > > long time as far as I can recall, and it is pretty primitive
> > > in the functionality that it offers.
> > 
> > Jamey and me have worked on it. I use it for ecore and it works well.
> 
> Great, I'll have to take a look at it.
> 
> > >> What needs to be done to implement a usable keyboard
> > >> support system for XCB?  I am willing to help.
> > >
> > > I think the current library would be "usable", but what
> > > really wants to be done is for someone to open the Xkb can
> > > of worms.  Several of us have peeked into the can, and then
> > > slammed the lid shut. :-)
> > >
> > > Probably the best plan would be for somebody to propose a
> > > reasonable architecture for XCB keyboard input, and then we
> > > can talk about how to implement it.  Want to look at this
> > > for us?
> > 
> > I've already tried to write the xml description of xkb, but that extension 
> > is very difficult to port ( maybe even impossible currently).
> > 
> > I think that once the python parser is finished, it will be easier to 
> > write that protocol.
> 
> Uh... What does rewriting the XML parser have to do with XKB support?
> 
> Stefan

Hopefully the new parser will be easier to play with.  The big barrier 
with XSLT is that few of us know it and readability sucks.

Jeremy