> At 05:54 AM 9/30/99 -0400, W Gerald Hicks wrote:
> >
> >> doing state machines with switch statements is a big mess.
> >
> >Still, you'll find a lot of them around. Do you have a favored
> >technique for coding complex state machines? (I'm a collector :)
>
>
> yes, state tables. Clean and easy to modify.
IMHO state tables are fine in theory. The problem is that the ``do
this'' bit sometimes needs to be split into two - one before the
state change and one after, and that same bit is frequently ``almost
the same'' as the ``do this'' bit for another transition.
Once you start coding it, you start to bring the common bits of code
into common routines, and eventually end up actually passing the
from/to states into those functions.
I found that redesigning the ppp(8) state machine eventually ended up
with lots of switch statements and a result that was nothing like I
had in mind when I started writing it ! It registers state
transition handlers to a certain extent, but there are too few
handlers and lots of ``if I'm in this state'' code.
--
Brian <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
<http://www.Awfulhak.org> <[EMAIL PROTECTED]>
Don't _EVER_ lose your sense of humour ! <[EMAIL PROTECTED]>
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
