Input thread on EVoC

[Daniel Stone]

Hi,

On Thu, Jun 03, 2010 at 03:51:40PM -0300, Fernando Carrijo wrote:
> What do you guys think about making use of EVoC to bring back to life the
> concept of the input thread? I had an offline conversation with Tiago and
> Peter about this, and it seems that it is a truly desirable feature.
> 
> The original implementation, due to Tiago's efforts on GSoC, was lost after
> annarchy's apocalypse; but he somehow managed to keep intact a considerable
> part of the code. So, the question which subsumes the former is: given that
> the revitalization of the input thread doesn't seem to demand much work, do
> you guys have any pending expectations in this regard that might amend the
> original idea in such a way that it could sound attractive both from the X
> Server's and from EVoC's points of view?
> 
> I would really appreciate any comments.

If you're really (really, really) keen, you could try to thread not only
input event generation, but event delivery. :) You could get away with
it fairly easily at first[0] by just having a single mutex which is
taken both by the input thread, and by any operation which modifies the
window stack, input shape, etc, in any way.

Cheers,
Daniel

[0]: At least for the hardware cursor; software loses out due to the
     need to do rendering.