[PATCH 2/2] Keep virtual core pointer touch class around if new slave doesn't have one

[Peter Hutterer]

On Mon, Feb 13, 2012 at 04:00:48PM -0800, Chase Douglas wrote:
> The VCP may have active touch grabs. The touch records must be kept so
> these touch grabs may be accepted/rejected in the future. This means the
> touch class list will not represent the touch class of the attached
> slave device if it does not have a touch class, but we already were
> breaking that assumption by keeping a separate touches array for the
> VCP.
> 
> Signed-off-by: Chase Douglas <chase.douglas at canonical.com>
> ---
> I haven't seen this cause a crash in the wild, and I haven't stress tested it,
> but I don't see how accept/reject could work after we switch to a non-touch
> slave device without this change.
 
yeah, unfortunately that explanation makes sense. merged, thank you.

Cheers,
  Peter

>  Xi/exevents.c |    9 +++------
>  1 files changed, 3 insertions(+), 6 deletions(-)
> 
> diff --git a/Xi/exevents.c b/Xi/exevents.c
> index 51f991c..069a52c 100644
> --- a/Xi/exevents.c
> +++ b/Xi/exevents.c
> @@ -699,13 +699,10 @@ DeepCopyPointerClasses(DeviceIntPtr from, DeviceIntPtr to)
>          to->touch->touches = to_touches_array_tmp;
>          to->touch->num_touches = to_num_touches_tmp;
>          to->touch->sourceid = from->id;
> -    } else if (to->touch)
> -    {
> -        ClassesPtr classes;
> -        classes = to->unused_classes;
> -        classes->touch = to->touch;
> -        to->touch      = NULL;
>      }
> +    /* Don't remove touch class if from->touch is non-existent. The to device
> +     * may have an active touch grab, so we need to keep the touch class record
> +     * around. */
>  }
>  
>  /**
> -- 
> 1.7.9
>