Hello,

> > c/ my current setup makes it simpler to document the states that the
> >    system can be in - helps when being validated for FIPS and common
> >    criteria.   Simple is good.

I'd just point out that the two approaches under discussion don't differ
in the number of states the system may get into, AFAICS.

> this is something i can agree with.  althought i still think matt's
> design is superior, it has a head and a tail, while yours is just "run
> ahead and burn all the bridges behind us".

I don't like the "no way back" aspect either (I understand it may
be OK for Simon's needs, it is however not for mine and Lubomir's
at least).

Anyway, I'll post a fix for getttyent() API and updated init diff as
soon as I get a chance (and -- based on the feedback -- potentially
send-pr it later).

	Regards,
		-- Jachym Holecek