[mythtv] usability bug in mythfrontend

Keith Irwin kirwin at ncsu.edu
Tue Jun 28 23:44:21 UTC 2005


On Tuesday, June 28, 2005, at 07:32  PM, Micah F. Galizia wrote:

>> 1) I would imagine that most people with a real remote
>> (i.e. with a channel up and down button) would have done this?
>
> Actually, if you use the Linux input layer, the channel up/down
> scancodes are 402 and 403.  XInput, allowing scancodes no greater than
> 255, is incapable of receiving those keys.  Probably works fine for 
> LIRC
> though.

I believe that those are actually the key codes, not scan codes.  
There's generally a two part translation: scan codes -> key codes -> 
characters.  So far as I can tell there is no program for X or tty 
which actually handles key codes above 255.  As I couldn't get lirc 
working, I wound up hacking the gpio ir-kbd drivers (which produce key 
codes rather than scan codes) to produce key codes in the 0-255 range.  
This was the only way I could get my remote working.  It's not a mythtv 
problem, obviously, but it is a big pain.  It may also be something 
that we could adjust mythtv to fix since X applications do not have to 
use xkb's key mappings.  I looked at fixing xkb to handle it, but the 
assumption that a key code is just one byte seemed to be embedded all 
over the place.  There were one byte types, 255's and 0xff's in a 
number of places throughout the code.  The whole thing was really 
sloppy.

If anyone has a better idea on how to fix this without hacking the 
drivers, please let me know.  I'll probably look at a mythtv specific 
solution at some point in the future when I have more time.

Keith Irwin



More information about the mythtv-dev mailing list