[Xastir] Serial KISS interface - long delays

[J. Lance Cotton]

Curt Mills, WE7U wrote:

> Another possibility is to edit the code yourself to make a custom
> build.  interface.c, line 3481 is the place:
> 
> 
> //fprintf(stderr,"Char pacing ");
> usleep(25000); // character pacing, 25ms per char.  20ms doesn't work for PicoPacket.
> 
> 
> Crank it down to some lower value and test it out.  Make sure
> whatever value you choose that you don't go too fast and overrun
> your TNC's buffer.  If your TNC can keep up at max speed then you
> can comment out the usleep() call.

I actually found that and first changed the usleep value to 2500 (10 times 
less) and it didn't seem to have much effect. I commented out the line and it 
still didn't seem to have much of an effect. I didn't measure it, but it may 
have been 0.5 to 1 seconds less. I will measure it tonight and see

I supposed I could just up my serial port speed to 57600 and test that also. 
Then we can find out where the real delays are, rather than the actual 
transmission. Or maybe I need to look for more excuses to get a 'scope.

> We also have 100ms waits in output_my_data() at lines 5397 and 5499,
> and a 1.5 second wait, same function, at 5511.  None of those apply
> to serial KISS TNC's though, but they do apply to other types of
> TNC's, including AX25.  AX25 devices should probably be excluded
> from the usleep() calls as well.
> 

-- 
J. Lance Cotton, KJ5O
http://map.findu.com/kj5o-14
joe at lightningflash.net