On 02/26/2015 02:32 AM, patrick keshishian wrote:
On 2/25/15, Ulf Brosziewski<ulf.brosziew...@t-online.de> wrote:
On 02/25/2015 11:53 PM, patrick keshishian wrote:
Hi,
On 2/25/15, joshua stein<j...@openbsd.org> wrote:
On Tue, 24 Feb 2015 at 12:32:10 -0800, patrick keshishian wrote:
I'm noticing slight annoyance with recent update from 20141121
snapshot to 20150217.
My touchpad, while two-finger scrolling (up/down) sometimes ignores
the scrolls. I have to lift my fingers and retry the gesture to
initiate
the scroll.
There was a change to pms(4) (r1.57) but it should only affect
elantech touchpads, so it's probably the r1.11 change to the
synaptics xorg driver which affects all of them:
http://cvsweb.openbsd.org/cgi-bin/cvsweb/xenocara/driver/xf86-input-synaptics/src/wsconscomm.c
Can you try recompiling that driver with that last revision backed
out and see if it fixes the problem?
Thanks for the reply and hint.
Took me a while longer as I made the mistake of rebuilding
the entire xenocara.
Indeed this revision seems to be the cause of the issue; at
least I've not been able to run into the issue in the last 10
minutes or so, whilst trying my best.
Thanks to Henrik Friedrichsen for also confirming the issue
and the fix.
Cheers,
--patrick
Unfortunately that "fix" might reintroduce other problems, at least
with other touchpads, and maybe even with yours. For example, if you
start a scrolling operation by making two touches at the same time,
I would expect that it begins with a "jump". Would you mind check that?
I can't make any tests myself, there is no synaptics touchpad around
here.
Hi,
I'm not sure what you mean exactly by "making two touches
at the same time".
The way I initiate a scroll is by placing two fingers (typically my
index and middle fingers) on the touchpad and moving them
up/down or left/right.
It has never begun with any "jump".
The change which prompted this thread, was absolutely annoying.
e.g., While reading a PDF document. I'd have to remove my fingers
and touch the touchpad again, very frequently, in order to continue
scrolling. Where as before (and now with the revision rollback) I can
completely concentrate on reading the document.
Best,
--patrick
Hi Patrick,
thanks for the reply. What I meant was this: There can be a (possibly
small) time interval between the contacts, or this interval can be
(nearly) zero. In the first case, the "finger count" of the driver
should change from 0 to 1, then from 1 to 2. If the hardware is not
extremely accurate it may change directly from 0 to 2 in the second
case. If the hardware *is* very accurate, it may be difficult to produce
these "double contacts". The test I had in mind could look like this: You
move the cursor over a scrollable document, lift the moving finger, then
make a "double contact" on a position that has a noticeable (vertical)
distance from the previous contact point. If the document moves/"jumps"
when the double contact is made, there's something wrong. This can
happen when the driver doesn't "count" properly.
Regards,
Ulf
