[PATCH 00/42] Multitouch support

Peter Hutterer peter.hutterer at who-t.net
Tue Dec 20 15:23:04 PST 2011

On Tue, Dec 20, 2011 at 11:53:41AM -0800, Chase Douglas wrote:
> On 12/14/2011 07:01 PM, Peter Hutterer wrote:
> > 
> > This is the set of patches to add multitouch support to the X server.
> > The lot can be found in the 'multitouch' branches of the following
> > repositories
> > 
> >     git://people.freedesktop.org/~whot/xserver
> >     git://people.freedesktop.org/~whot/inputproto
> >     git://people.freedesktop.org/~whot/xf86-input-evdev
> >     git://people.freedesktop.org/~whot/libXi
> > 
> > This patchset adds XI 2.2 support to the server.
> > For a full list of the additions see the protocol spec, but it includes
> > support for multitouch events, pointer emulation on selected touch events
> > and the semantics for grabbing and replaying touch sequences.
> > 
> > There are still a few rough edges here and there. For example Carlos
> > Garnacho (who is working on the GTK3 touch support) pointed out an active
> > grab issue with pointer-emulation today. Over the next days, I'll fix these
> > with follow-up patches.
> > This patchset should not affect regular pointer or keyboard processing
> > however.
> > 
> > Note that this patchset was essentially created from the diff against the
> > devel branch to get it down to a reasonable number of patches that people
> > can actually review. The original branch started from Daniel Stone's
> > initial work and piled up 188 patches on top of
> > 7528a6b88eb32098af4369a8bd9d70a808fa6f1c, my last syncpoint with master.
> > If you want to look at all these, they're on the multitouch-devel-history
> > branch.
> > The patches in this set are not the ones you could usually write out from
> > scratch since they handle a lot of corner-cases in the first revision.
> > Still, easier than reviewing the original patches.
> > 
> > Daniel and Chase both contributed to this patchset and I've tried to
> > correctly attribute (co-)authorship where applicable. My apologies for any
> > mistakes here, again the multitouch-devel-history branch is more accurate
> > here.
> > Daniel, Chase, I expect that my s-o-b on your patches should not be an
> > issue given that the source patches had your s-o-b on them?
> No issues here. I noticed a few things that I had written that weren't
> attributed, but they could have very well been written by you from
> scratch during the big rewrite :). It's hard to keep track of it all,
> and it's not worth much to me, so I'm happy to leave things as-is.
> If it matters for the certificate of origin legal matter, everything in
> this patch set that I have contributed but did not explicitly sign off
> on is implicitly:
> Signed-off-by: Chase Douglas <chase.douglas at canonical.com>
> I hope no one makes me go through and add this individually where
> required :).
> Unless I replied to a specific patch, all are:
> Reviewed-by: Chase Douglas <chase.douglas at canonical.com>
> To summarize, all are Reviewed-by (maybe with non-functional fixes)
> except 27 and 37.
> Can you summarize what is unimplemented (or needs fixing)? The following
> is what I remember:
> * Implementation of early reject/accept of passive touch grabs
> * What to do about overlapping touch selections when attaching a slave
>   device to a master device

I'm not sure what you mean here. Touch events are separately handled from
each device, so client C1 can have a selection on the SD and C2 on the MD.

The only thing we need to make sure is that when floating or attaching a
device all current touchpoints are destroyed (I _think_ this bit is missing,
at least I can't remember writing it)

> * Maybe a change to handling of explicit active grabs of touches

Active grab handling has at least one bug atm that I haven't tracked down
yet. I think once I've done that, active grab handling will be in a better
state than now and we can revise the protocol.

More information about the xorg-devel mailing list