[PATCH weston 1/6] xdg-shell: define the present/present_from_event() requests

Daniel Stone daniel at fooishbar.org
Wed Sep 30 12:00:25 PDT 2015


On Wednesday, 30 September 2015, Derek Foreman <derekf at osg.samsung.com>
wrote:

> On 30/09/15 12:01 PM, Bill Spitzak wrote:
> > My main concern is that programmers are going to search for the word
> > "raise" to figure out how to raise windows. They are not going to search
> > for "present" or "indicate_activity" or any other bogus word. I suspect
> > that if you don't name this "raise" then a function called "raise" will
> > be added as an extension very soon, and probably break all your careful
> > plans of allowing the compositor to do something other than raise.
>
> Well, X calls similar functionality "UrgencyHint"...  That's actually
> more descriptive than imperative.  The idea is that we let the
> compositor know the surface needs attention, not that we tell it what to
> do.
>
> Programmers shouldn't be trying to figure out how to raise a window
> anyway, they should be searching for words like notify.  But if we put
> the word "raise" prominently in the documentation, I guess google should
> sort it out for them.  ;)
>
> In any event, I dislike "present" less than I dislike "raise", so I'm
> happy enough with the patch as presented.  You do raise a good point
> though.
>
> I'm done.  Really.  Others can continue to fight about this if they so
> choose, I've had my say.


 +1 from me, with optional bike shed to request_attention or note_urgency,
or anything of their ilk. I don't like 'present' overlapping with the
meaning of showing buffers on screen. Urgency / requesting attention have
prior art in X11/ICCCM/EWMH. Anyone designing protocols is unlikely to be
unaware of this, will have it pointed out to them in review, and can grep
the docs for 'raise' anyway.

Pending resolution of zero-as-invalid-serial, subsequent flattening of
requests, documentation that the extension relies on this behaviour:
Reviewed-by: Daniel Stone <daniels at collabora.com>

(On the assumption that any serial passed will be from xdg-shell or
similar; input serials would require a wl_seat to be passed, but if you're
reacting to input, then you by definition have the user's attention.)

Cheers,
Daniel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/wayland-devel/attachments/20150930/5f77d95c/attachment.html>


More information about the wayland-devel mailing list