On 12/20/2010 08:28 AM, Peter Hutterer wrote:
> On Sun, Dec 19, 2010 at 05:04:51PM +0100, Alberto Milone wrote:
>> @@ -1754,7 +1754,8 @@ ComputeDeltas(SynapticsPrivate *priv, const struct
>> SynapticsHwState *hw,
>> if (inside_area && moving_state && !priv->palm &&
>> !priv->vert_scroll_edge_on && !priv->horiz_scroll_edge_on &&
>> !priv->vert_scroll_twofinger_on && !priv->horiz_scroll_twofinger_on &&
>> - !priv->circ_scroll_on && priv->prevFingers == hw->numFingers) {
>> + !priv->circ_scroll_on && priv->prevFingers == hw->numFingers &&
>> + hw->millis > HIST(0).millis) {
>
> shouldn't we reset the history in this case too? it is invalid after all
> (not sure if it has an effect though)
Protecting against timer overruns is nothing but good sense. But it
should happen seldom enough that anything but crashing the machine is
acceptable. It might be better to handle this in all instances where
history is queried, i.e. only handling a strictly positive dtime.
BTW I notice we somehow messed up the return value of ComputeDeltas
(delay), probably some cleanup missing.
Cheers,
Simon
_______________________________________________
[email protected]: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: http://lists.x.org/mailman/listinfo/xorg-devel
