svn commit: trunk/busybox/libbb

[Rob Landley]

> If you are still considering the user defined ioctl stuff I suggest you
> put you're efforts to something more productive.

So another vote for "not doing this at all".

> The shell script (or whatever) to translate a baudrate given by an end user
> to the correct ioctl values would be bigger than putting the table in
> busybox.

How do end users know that they can do 460800, but not 460900?  Nobody has yet 
answered this question.  End users are specifying variable baudrates for 
nonstandard speeds, yet they know what these nonstandard speeds are.  Where 
did they get them?  Are they looking in bits/termios.h directly, or is there 
some kind of documentation that lists "460800" which could also list x1004 as 
the way to get it?

> You would instead move some bytes out of busybox but add bloat to the
> system as a whole - probably forcing everybody caring about baudrates
> >115200 to have their own stty/speed_table implementation making merging
> and providing patches a hazzle.

Why?  You're assuming the need for a shell script here.  What is a use case 
where the shell script is actually required?

> /Johan

Rob
-- 
Never bet against the cheap plastic solution.