input transformations

[Deron Johnson]

Keith Packard wrote:

>On Mon, 2007-02-05 at 20:04 -0500, David Reveman wrote:
>
>  
>
>>From the talks keithp has given on the subject I've understood that
>>implementing this properly is very hard, if not impossible. Is that not
>>the case anymore?
>>    
>>
>
>The hard part was making it asynchronous and routing it through the X
>protocol itself. I did a small fraction of that implementation before
>abandoning it. Creating a synchronous back-channel would work, although
>I have concerns about performance in that case.
>  
>

In the plug-in picker case which I mentioned in my previous mail, the LG 
picker would
use a synchronous back-channel (via sockets) which Keith mentions. I 
think I know a
way of making this fast. This will involve routing all events through 
the external picker
in a "pre-picking" step and then later in the event pipeline if the 
pre-picked information
gets out of date to use the back-channel to go back to the external 
picker for the
up-to-date information. Since the pre-picked information gets 
out-of-date infrequently
this approach will have the same good performance behavior of today's LG 
pipeline.
But, "of course, "the proof is in the pudding" and I haven't yet gotten 
the time to prove
this with a prototype. But if we put in place the plug-in picker 
interface I mentioned
in my previous mail we will be able to experiment with external pickers 
in the future
while satisfying the needs of today.