[PATCH wayland-protocols 1/2] Introduce wp_relative_pointer interface

[Jonas Ådahl]

On Mon, Nov 23, 2015 at 03:24:20PM -0600, Derek Foreman wrote:
> On 17/11/15 04:09 AM, Jonas Ådahl wrote:
> > The wp_relative_pointer interface is an extension to the wl_pointer
> > interface created from wl_seat. It has the same focus, but adds the
> > functionality of sending relative pointer motions unhindered by
> > constraints such as monitor edges or other barriers. It also contains
> > unaccelerated pointer motion information.
> > 
> > Signed-off-by: Jonas Ådahl <jadahl at gmail.com>
> > ---
> > 
> > There has alreay been RBs on this protocol, but I'd like to get them
> > explicitly re-RB:ed considering that the protocol has changed some, even
> > though the changes have not been major ones.
> > 
> > 
> > Changes since the last version sent as weston patches:
> > 
> > Interfaces and protocol renamed according to wayland-protocols naming
> > conventions.
> > 
> > Accelerated and unaccelerated motion deltas have the data type
> > wl_fixed_t now again. 64 bit fixed point deltas was deemed unnecessary,
> > and wl_double_fixed_t was dropped.
> > 
> > 64 bit timestamp argument name suffix changed from msb/lsb to hi/lo.
> > 
> > Various other minor changes.
> > 
> > 
> > Jonas
> > 
> > 
> >  Makefile.am                                        |   1 +
> >  unstable/relative-pointer/README                   |   4 +
> >  .../relative-pointer-unstable-v1.xml               | 122 +++++++++++++++++++++
> >  3 files changed, 127 insertions(+)
> >  create mode 100644 unstable/relative-pointer/README
> >  create mode 100644 unstable/relative-pointer/relative-pointer-unstable-v1.xml
> > 
> > diff --git a/Makefile.am b/Makefile.am
> > index f1bac16..2db6bd5 100644
> > --- a/Makefile.am
> > +++ b/Makefile.am
> > @@ -5,6 +5,7 @@ nobase_dist_pkgdata_DATA =							\
> >  	unstable/text-input/text-input-unstable-v1.xml				\
> >  	unstable/input-method/input-method-unstable-v1.xml			\
> >  	unstable/xdg-shell/xdg-shell-unstable-v5.xml				\
> > +	unstable/relative-pointer/relative-pointer-unstable-v1.xml		\
> >  	$(NULL)
> >  
> >  pkgconfigdir = $(libdir)/pkgconfig
> > diff --git a/unstable/relative-pointer/README b/unstable/relative-pointer/README
> > new file mode 100644
> > index 0000000..64c42a1
> > --- /dev/null
> > +++ b/unstable/relative-pointer/README
> > @@ -0,0 +1,4 @@
> > +Relative pointer protocol
> > +
> > +Maintainers:
> > +Jonas Ådahl <jadahl at gmail.com>
> > diff --git a/unstable/relative-pointer/relative-pointer-unstable-v1.xml b/unstable/relative-pointer/relative-pointer-unstable-v1.xml
> > new file mode 100644
> > index 0000000..0411d1f
> > --- /dev/null
> > +++ b/unstable/relative-pointer/relative-pointer-unstable-v1.xml
> > @@ -0,0 +1,122 @@
> > +<?xml version="1.0" encoding="UTF-8"?>
> > +<protocol name="relative_pointer_unstable_v1">
> > +
> > +  <copyright>
> > +    Copyright © 2014      Jonas Ådahl
> > +    Copyright © 2015      Red Hat Inc.
> > +
> > +    Permission is hereby granted, free of charge, to any person obtaining a
> > +    copy of this software and associated documentation files (the "Software"),
> > +    to deal in the Software without restriction, including without limitation
> > +    the rights to use, copy, modify, merge, publish, distribute, sublicense,
> > +    and/or sell copies of the Software, and to permit persons to whom the
> > +    Software is furnished to do so, subject to the following conditions:
> > +
> > +    The above copyright notice and this permission notice (including the next
> > +    paragraph) shall be included in all copies or substantial portions of the
> > +    Software.
> > +
> > +    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> > +    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> > +    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
> > +    THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> > +    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> > +    FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
> > +    DEALINGS IN THE SOFTWARE.
> > +  </copyright>
> > +
> > +  <interface name="zwp_relative_pointer_manager_v1" version="1">
> > +    <description summary="get relative pointer objects">
> > +      A global interface used for getting the relative pointer object for a
> > +      given seat.
> 
> for a given pointer?  (I don't see any place here where we're passing a
> seat...)
> 
> > +
> > +      Warning! The protocol described in this file is experimental and backward
> > +      incompatible changes may be made. Backward compatible changes may be added
> > +      together with the corresponding interface version bump.  Backward
> > +      incompatible changes are done by bumping the version number in the
> > +      protocol and interface names and resetting the interface version.  Once
> > +      the protocol is to be declared stable, the 'z' prefix and the version
> > +      number in the protocol and interface names are removed and the interface
> > +      version number is reset.
> > +    </description>
> > +
> > +    <request name="destroy" type="destructor">
> > +      <description summary="destroy the relative pointer manager object">
> > +	Used by the client to notify the server that it will no longer use this
> > +	relative pointer manager object.
> > +      </description>
> > +    </request>
> > +
> > +    <request name="get_relative_pointer">
> > +      <description summary="get a relative pointer object">
> > +	Create a relative pointer interface given a wl_pointer object. See the
> > +	wp_relative_pointer interface for more details.
> > +      </description>
> > +
> > +      <arg name="id" type="new_id" interface="zwp_relative_pointer_v1"/>
> > +      <arg name="pointer" type="object" interface="wl_pointer"/>
> > +    </request>
> > +  </interface>
> > +
> > +  <interface name="zwp_relative_pointer_v1" version="1">
> > +    <description summary="relative pointer object">
> > +      A wp_relative_pointer object is an extension to the wl_pointer interface
> > +      used for emitting relative pointer events. It shares the same focus as
> > +      wl_pointer objects of the same seat and will only emit events when it has
> > +      focus.
> > +    </description>
> > +
> > +    <request name="destroy" type="destructor">
> > +      <description summary="release the relative pointer object"/>
> > +    </request>
> > +
> > +    <event name="relative_motion">
> > +      <description summary="relative pointer motion">
> > +	Relative x/y pointer motion from the pointer of the seat associated with
> > +	this object.
> > +
> > +	A relative motion is in the same dimension as regular wl_pointer motion
> > +	events, except they do not represent an absolute position. For example,
> > +	moving a pointer from (x, y) to (x', y') would have the equivalent
> > +	relative motion (x' - x, y' - y). If a pointer motion caused the
> > +	absolute pointer position to be clipped by for example the edge of the
> > +	monitor, the relative motion is unaffected by the clipping and will
> > +	represent the unclipped motion.
> > +
> > +	This event also contains non-accelerated motion deltas. The
> > +	non-accelerated delta is, when applicable, the regular pointer motion
> > +	delta as it was before having applied motion acceleration and other
> > +	transformations such as normalization.
> > +
> > +	Note that the non-accelerated delta does not represent 'raw' events as
> > +	they were read from some device. Pointer motion acceleration is device-
> > +	and configuration-specific and non-accelerated deltas and accelerated
> > +	deltas may have the same value on some devices.
> > +
> > +	Relative motions are not coupled to wl_pointer.motion events, and can be
> > +	sent in combination with such events, but also independently. There may
> > +	also be scenarious where wl_pointer.motion is sent, but there is no
> > +	relative motion. The order of an absolute and relative motion event
> > +	originating from the same physical motion is not guaranteed.
> 
> It still makes me a little grumpy that we're potentially going to post a
> pile of absolute events (potentially 1000+ per second with a ridiculous
> gamer mouse) that a client ignores, but still get sent and trigger a
> callback.  I think it would be nice to have a way to disable the
> absolute motion events, possibly as a side effect of creating a relative
> pointer.
> 
> > +
> > +	If the client needs button events or focus state, it can receive them
> > +	from a wl_pointer object of the same seat that the wp_relative_pointer
> > +	object is associated with.
> 
> button presses are still strictly ordered though, right?  That is, I
> might see
> 
> rel
> abs
> button
> 
> or:
> abs
> rel
> button
> 
> but I'll never see
> abs
> button
> rel

We'd see this if there was an abs that didn't come with a rel, then a
button event, then a rel that didn't come with an abs.

> 
> for a situation where the button occurred after the rel motion?

Right. If the button happened before the rel motion, the rel event will
come before the button.

> 
> I liked the discussion about framing rel pointer events inside
> wl_pointer.frame... I think that's something that can be bolted on later
> without breaking any backwards compatibility?

Yes, I think it will be good to add that later when we have made a
wayland release with that change included.

> 
> I guess a stream where wl_pointer didn't generate anything and
> wp_relative_pointer did could potentially look weird to an application
> not expecting it.
> 
> I hope there won't be any problems stemming from button presses coming
> with 32 bit timestamps while rel motion comes with 64 bit.  These could
> appear to be non-monotonic to a client?

I suppose that could be problematic, but we can't fixe that without
extending wl_pointer.button really.

> 
> > +      </description>
> > +
> > +      <arg name="utime_hi" type="uint"
> > +	   summary="high 32 bits of a 64 bit timestamp with microsecond granularity"/>
> > +      <arg name="utime_lo" type="uint"
> > +	   summary="low 32 bits of a 64 bit timestamp with microsecond granularity"/>
> > +      <arg name="dx" type="fixed"
> > +	   summary="the x component of the motion vector"/>
> > +      <arg name="dy" type="fixed"
> > +	   summary="the y component of the motion vector"/>
> > +      <arg name="dx_unaccel" type="fixed"
> > +	   summary="the x component of the unaccelerated motion vector"/>
> > +      <arg name="dy_unaccel" type="fixed"
> > +	   summary="the y component of the unaccelerated motion vector"/>
> 
> Utterly worthless bike shed painting:
> Personally, I'd like to see the unaccelerated deltas posted as "dx" and
> "dy" and the accelerated ones as "dx_accel" or "dy_accel", since if
> you're using the rel pointer interface there's a good chance you're
> writing a first person game where accelerated deltas are terrible.
> 
> Qualifying all the names would be fine too, and have nothing simply
> called "dx" and "dy", then developers are more likely to think for 5
> seconds instead of just grabbing dx and dy.

The reason for having dx/dy suffix free is because they are simply the
relative variant of the regular x/y. If that means accelerated or not
accelerated is an implementation detail, so if we leave dx/dy to just
whatever x/y is, we don't have to define what "_accel" actually means.

> 
> Anyway, low priority whining aside, this looks really good to me, and
> closes a functional hole we've had for far too long.
> 
> Reviewed-By: Derek Foreman <derekf at osg.samsung.com>


Thanks for your reviews!


Jonas

> 
> 
> > +    </event>
> > +  </interface>
> > +
> > +</protocol>
> > 
>