xorg-server again.. (synaptics seems broken)

Zbigniew Luszpinski zbiggy at o2.pl
Sat Mar 17 15:17:47 CET 2012


> xf86-input-synaptics seems to be broken with current xorg-server (it
> downgraded xorg-server and all worked.

How about doing some developer work to fix the problem and commit it to 
moonbase instead of workarounding it in zlocal? (check for new official 
release, check for developer release, check git, check synaptics ML). This 
is not so hard as it looks. This way more users will benefit.

The solution was on 2nd position I entitiled here: 'check for developer 
release'. I found it next to latest official release:
xf86-input-synaptics 1.5.0.tar.bz2
xf86-input-synaptics 1.5.99.901.tar.bz2

Works for me. This is RC release.

> Please review that, as synaptics is around in many notebooks, this
> might happen to a lot of users.

Not necessarily. I have netbook with touchpad. Its driven by evdev driver 
'out of the box'. (I have no xorg.conf so everything runs on defaults)
I have installed synaptics driver just because of your 'synaptics broken' 
mail otherwise I would continue using evdev. To be honest as long as 
touchpad works an ordinary user does not care.

Indeed synaptics driver gives a little better user experience by slowing 
down cursor movement so it does not move so fast and does not vibrate on 
screen when you keep finger on touchpad - you feel more human touch and 
control. And it is not only for synaptics as names says. It also supports 
for example my "ETPS/2 Elantech Touchpad" and ALPS. So they should change 
the name of driver from synaptics to touchpad.

> Btw. I think I wrote about testing
> things before updating them. Its okay if you don't have a notebook and
> can't test everything but please, ask others to test it then - waiting
> a few weeks longer to keep lunar in a stable state is far better than
> trying to update something just to be up to date without making sure it
> will work as expected.

We should have tool for autocollecting anonymous hardware info from users' 
boxes during lunar update to know what devices (and their drivers) should 
get special care and testing. Because there is at least one user of 
synaptics driver someone will test it before commit. I will keep it on my 
netbook to have an eye on it during future updates.

> At least.. for things like xorg-server, i had to downgrade three times
> now. That can't be the solution for a distribution.

This is because of lack of developers/testers - too small community. 
Florin commits a lot and thats great because thanks to him the distro is 
rolling. Of course 'shit happens' sometimes because he can not test 
everything but this is not big numbers. I usually thake some time in the 
evening/night to do lunar update on 3 machines to review whats new and 
have up to date PCs. Then adjust what failed. So in usually less than 10 
hours you have fix for things which sneaked out from Florin's testing.
I also make updates with unintended mistakes sometimes and fix them after 
bashing on ML.
Compile errors are not problem. Lunar will not update module which 
compilation failed so an user continues to use just older version he/she 
already have. When bug is reported it is fixed.



More information about the Lunar-dev mailing list