multi-touch-dev team mailing list archive

[Duncan McGreggor <duncan.mcgreggor@xxxxxxxxxxxxx>]

On 08/17/2010 11:48 PM, Rafi Rubin wrote:
> Not present on 2.6.35-{14,15,16} for x86_64 on my xt.
> 
> $ cat /sys/class/input/event*/dev*/cap*/abs                     ~  1:14AM 
> holly
> 0
> 0
> 0
> 0
> 0
> 0
> 1000003
> 73010000000003
> 30000000003
> 0
> 0
> 0
> $ uname -a                                                      ~  1:15AM 
> holly
> Linux holly 2.6.35-15-generic #21-Ubuntu SMP Wed Aug 11 17:06:48 UTC 2010 
> x86_64 GNU/Linux
> 
> 
> I will start a 32b image download and will have to figure out a way to 
> test multiple kernels.  I guess I can blow away swap or something for 
> another install.
> 
> Nothing terribly obvious in the delta of the source for -13 -> -15 
> (assuming I grabbed the correct git tree).
> 
> Anyway the functions in questions are input_bits_to_string 
> (input.c:814ish) and input_print_bitmap (1218) or maybe 
> input_seq_print_bitmap.
> 
> Anyway should be easy enough to figure out, and we probably should do to 
> increase consistency.  However, it might also make sense to adjust the 
> script to account for buggy kernels which might not get fixed before 
> people start testing and asking questions.
> 
> Rafi

Interesting... so it looks like the difference here is "0000000", 7
zeros. Well, that and an extra space (in my output). But I don't know
what's going on under the hood to either account for that or to adjust
the script for it.

What actually concerns me more, though, is if the content of
device/capabilities/abs is used in grail or gevdev. Could my MT
experience be getting seriously borked as a result of this?

Thanks,

d