[PATCH v2 00/17] Improved text protocol, editor and keyboard examples

Jan Arne Petersen jpetersen at openismus.com
Thu Sep 13 07:09:05 PDT 2012

On 09/12/2012 11:38 PM, Kristian Høgsberg wrote:
> On Sun, Sep 09, 2012 at 11:08:29PM +0200, Jan Arne Petersen wrote:
>> From: Jan Arne Petersen <jpetersen at openismus.com>
>> Fixed some small bugs in still open patches, and added them here since they are
>> dependencies of the other changes.
>> Improve the editor example adding cursor, selection and pre-edit support.
>> Add a proper QWERTY keyboard layout example with upper- and lowercase
>> letters.
>> Enhance the protocol with a better set_surrounding_text request, support for
>> pre-edit strings, support for deleting text (delete_surrounding_text), support
>> for control keys and a request to reset text buffers.
>> Show the enhanced protocol in the editor and keyboard examples.
> The series looks good to me and I've commited it as is.  I did wonder
> about a couple of things in the text protocol though:
>  - Why is there a seat argument to text_model.deactivate?

The idea is that a text_model could maybe activated by different seats
at the same time, so that we can track regarding which seats the
text_model is still activated. But I am not really sure yet, if it would
not be better to have a text_model active for just one seat at a time. I
still need to figure that out.

>  - Should the text_model.activated event carry the surface and seat it
>    was activated for?  The protocol is asynchronous and we could in
>    theory have deactivated and activated another surface when the
>    client receives the activate event.

True that would make sense.

>  - Are text_model.enter/leave better names for
>    text_model.activated/deactivated?  It seems like these events are
>    similar to wl_keyboard.enter/leave, and I think it's better to be
>    consistent with that.

I agree we should name it enter and leave.

>  - The text_model.key event introduces keysym in the protocol and
>    assumes XKB keycodes.  I think that's fine, but I'm wondering if we
>    need to communicate the keycode namespace we're using there
>    somehow.

There is a patch with some additional documentation for the text
protocol, I tried to cover that there.

Best regards,
Jan Arne Petersen

More information about the wayland-devel mailing list