[gpm]alternate wmode for GPM synaptics

[Peter Berg Larsen]

On Mon, 22 Jul 2002, Charl P. Botha wrote:

> Things that I still didn't like even with your new gpm:
> 1. Corner tapping still doesn't work for me.  I expect a change to the edge
>    limit values would fix this.

The codes to detect edges are nearly identical (as they both have their
origin from the same source:). So try modifing gpm-syn.conf.


> 3. My pad has left and right buttons, and then a button in the middle for up
>    and down.  GPM still maps this as left, right, middle and fourth, which I
>    don't find very useful.

Is this a type of "4 way botton" or is it something a I dont know? (do you
have a link with description?) All the bottons I have knowledge of are all
configurable except #fingertaps (I will send a patch soon). (though some
only through gpm's config)


> 2. The mouse would sporadically move across the screen during a tap; this is
>    VERY irritating.  I have no clue what causes this.

> 4. Somehow, the two and three-finger tap detection with the xfree driver
>    (and port) never fails me, whereas on the new GPM this has happened
>    often.  This must be a fault on my part, as both my port and the new GPM
>    make use of the hardware multi-finger detection.  Could it be some of the
>    extra software logic that you use?     

Yes, there are two extra layer of software logic that can be turn off (I
dont remember the variables). The reason for this is multiple: 

The touchpad (I have at least) is very bad to detect multi fingers when
placed vertical to each other (not horisontal), but when I uses the
touchpad I have a 30deg angle in the wrist which means that a multifinger
tap often fails, and becomes on large finger.

Secondly, (which migh explain 2) when doing a multitap, one finger tends
to touch the touchpad before the others, followed by the others. The
touchpad (I have) the sends info about the first finger, when the second
gets closer the x,y values changes toward to other finger and the
fingerwidth enlarges until it becomes one finger. Therefor it is offen
that when using multifingers the cursor moves a little, which is
compensated for (in software) when actually detecting multifingers.

Thirdly, I dont like edge motion, so when I reach the edge in a tap hold, 
I place onether finger on the pad, which without the above compensation
flings the cursor far away.  

It might be that some of this is now done in hardware in your touchpad,
which would lead to the observed behaviour.


> In anycase, I think that the work you have done is really great!  However,
> for now I'm going to stick to my very simple patch as It Works For Me(tm).
> I'm still going to keep an eye on your progress...

> I expect that these things can be generally alleviated by mucking around in
> the config. 

Except for those mention in the previos mail, I belive you can have
exactly the same behaviour by "mucking around in the config". If not
please do tell.


Peter
--
E-Mail:       pebl@math.ku.dk 
Real name:    Peter Berg Larsen 
Where:        Department of Computer Science, Copenhagen Uni., Denmark