Re: Wayland

[Piñeiro <apinheiro igalia com>]

Hi Jason,

On 05/07/2012 04:58 AM, Jason White wrote:
> Are there people with backgrounds in accessibility development working with
> the Wayland project?

As far as I know, no. At least  no Wayland developer provided any
feedback on gnome-accessibility, gnu-accessibility or any other of the
accessibility-related mailing lists I'm subscribed to.

Anyway, we have Wayland in consideration. Some historic context:

We had a brief thread about this on 2010. It started on
gnome-accessibility-list:
https://mail.gnome.org/archives/gnome-accessibility-list/2010-November/msg00054.html

Although it also touched gnome-shell mailing lists.
https://mail.gnome.org/archives/gnome-shell-list/2010-November/msg00040.html

Due this conversation, we include Wayland as one of the things to talk
about in the first ATK/AT-SPI2 hackfest, and we created a bug to track it:
https://bugzilla.gnome.org/show_bug.cgi?id=649556

But the conclusions of the Hackfest didn't provide too much light:
https://bugzilla.gnome.org/show_bug.cgi?id=649556#c1

So we know basic things, like that we want X/Wayland totally transparent
to ATK, and that probably AT-SPI2 will need to take that into account,
but we lack the details. Why we didn't do anything else here? Well,
basically because we are already overwhelmed with other stuff.

>
> In particular, the accessibility infrastructure will need to be fully
> supported on systems that run Wayland without an X server. 

Unfourtunately, that means that current GNOME accessibility
infrastructure will not work on Wayland right now.

> The latest Wayland
> update that I've read mentions keyboard handling as one of the aspects of
> Wayland development that is still in flux, hence there may be opportunities
> for those who understand the requirements of AT-SPI 2, Orca etc. in this area to
> make their requirements known before the relevant APIs are set.

One of the current problems big problems is related to key events. In a
short, as XEvIE is dead, we can't track key events directly from X
server, so we are doing a key snooping on the application, and
forwarding it via DBUS. This is not really performant, can have some
collateral effects, and is the only remaining thing that makes X not
really transparent to ATK. Since we started to talk about that, I have
asking about how this could be done on Wayland. As you mention that key
handling is still in flux, I guess that it is a good moment to ask.


> http://lwn.net/Articles/491509/

Thanks for the link, I will read it shortly.

>
> I know that it will remain possible to run X as a Wayland client, but, given
> the current thinking of X and Wayland developers, the norm will be for desktop
> environments to use Wayland directly without an X server.
>
Yes, you are right. It seems that Wayland is the future (not sure if
middle or long term yet), so it is good to really think about it.

Thanks

BR

-- 
Alejandro Piñeiro Iglesias