rtems-980923 is available

[eric at skatter.USask.Ca (Eric Norum)]

joel at oarcorp.com wrote:
> <off the list>
>
> Now that I look at it, I am not sure your patch is correct.
>
> If we take the true case of the if, then you should return 0. If I
> am reading the patch correctly, the routine always returns 1.
>
> --joel
>

Yeah.
I made the change too quickly.

I don't really care what gets returned.  My drivers don't need the  
value anyhow. This seems to be causing a lot of friction, so you may  
as well leave it the way Aleksey has it in the 0923 snapshot.

I find that the variable name nToSend implies something that's not  
strictly true -- there may in fact be lots more to send (after the  
ring buffer wraps around to the beginning again), or there may be  
less to send (since the device write routine has already had a crack  
at the characters) -- but I really don't see any point in further  
arguments about this.

---
Eric Norum                                 eric at skatter.usask.ca
Saskatchewan Accelerator Laboratory        Phone: (306) 966-6308
University of Saskatchewan                 FAX:   (306) 966-6058
Saskatoon, Canada.