← Back to team overview

multi-touch-dev team mailing list archive

Re: Peter Hutterer's thoughts on MT in X

 

On Wed, Oct 06, 2010 at 12:30:46PM -0700, Ping Cheng wrote:
> On Wed, Oct 6, 2010 at 6:51 AM, Chase Douglas
> <chase.douglas@xxxxxxxxxxxxx> wrote:
> >
> > If I remember correctly, actively grabbed events cannot be replayed.
> > Only passively grabbed events. Since passively grabbed events must start
> > within a particular window, this scenario will never occur.
> >
> > I suggest reading the XI protocol specifications and the X core protocol
> > specification, where the semantics and details of input grabbing are
> > discussed:
> 
> Chase, you are right. I have messed up the master/core and
> passive/slave in my explanation. I need to get the terms and semantics
> right before I can comment more (that would take some time).

master: virtual device representing either cursor or keyboard focus
slave: physical device (in most cases)
core: device that can send core events. term is essentially deprecated,
think of a master device as a core device. (slaves can never send core
events). best to not use the term anymore, see also "core event"
core event: core X protocol event sent down the wire.
passive grab: grab activated when a button/key combination is pressed on a
particular window. clients need to register a passive grab once, then will
activate repeatedly until ungrabbed. passive grabs terminate automatically
when the button/key is released
active grab: grab activated immediately on request by the client, terminates
when client ungrabs or disconnects.

That about sums it up, I think :)

Cheers,
  Peter

> What I wanted to make clear was:
> 
> The spec and implementation need to
> 
> 1.   provide an easy way for toolkits/client to support the most
> popular/used MT features, such as basic gestures;
> 2.   leave rooms for sophisticated MT toolkits/apps/clients to grow.
> 
> For item 1, a smooth transition from XI (single touch events) to
> XI2.1, and vice versa, has to be carefully considered, with the
> posibility to turn XI or/and XI2.1 report/events off. Active and
> passive grabbing has to be tracked and probably managed. I am not
> exactly sure at which level though. I would think it is in X server?
> 
> For item 2, I would think those sophisticated MT apps will occupy the
> whole desktop, i.e., there is only one MT app running and all MT
> events can only reach that app. So, there would be no competition or
> inconsistence there.
> 
> KISS is the rule that I always remind myself and others in my group to consider.
> 
> Ping



Follow ups

References