[PATCH] evdev: pass the valuators to the X server for ButtonPress/Release events.

[Oliver McFadden]

On Fri, 2010-02-26 at 01:53 +0100, ext Peter Hutterer wrote:
> On Thu, Feb 25, 2010 at 11:45:59AM +0100, Julien Cristau wrote:
> > On Thu, Feb 25, 2010 at 12:35:00 +0200, Oliver McFadden wrote:
> > 
> > > On Thu, 2010-02-25 at 11:25 +0100, ext Julien Cristau wrote:
> > > > On Thu, Feb 25, 2010 at 10:21:07 +0200, Oliver McFadden wrote:
> > > > 
> > > > > Previously this was not implemented because it requires the X server has
> > > > > the xf86PostButtonEventP() function which was not released at the time.
> > > > > 
> > > > Can you either keep the code working with older X servers with an ifdef,
> > > > or update configure.ac to require a new enough xorg-server.pc?
> > > 
> > > Yeah, this is probably a good idea. I'll dig through the X server log
> > > later today and check which version is required.
> > > 
> > > Also, which is preferred; an #if check on XORG_VERSION_CURRENT or a
> > > configure.ac check?
> > > 
> > I'd say either use xorg-server >= $version in configure.ac's
> > PKG_CHECK_MODULES call (if you want to require a new server), or add an
> > AC_CHECK_DECL(xf86PostButtonEventP) and a #ifdef (if you want to keep
> > the driver working with older server).
> 
> I'd be simpler to have an ABI_XINPUT_VERSION check for a version that
> already includes this call. Afaict, this is ABI_XINPUT_VERSION >= 8.
> This looses out on the commits in between (8da0ff2d5108666..895f40792a14d8)
> but I think that's a negligible issue.

Ok, I can do the patch that way. AC_CHECK_* was giving me some trouble
anyway. I will wait for your comments on the previous email first,
though.