Hallo Denys!

> Why do you think O_NONBLOCK is wrong?
>
> agetty uses it (always).
> mgetty uses it by default.

Do not compare with mgetty, mgetty uses a completely different approach
then other getty implementation and has special code to monitor and
control the modem lines (not provided by Busybox getty).


>> INITSTR was never designed to be used for modem initialization tasks.
> Says who?

It was a suggestion from my experience and knowledge of over 20 year of
Unix serial line communication and programming on an uncounted number of
platforms. The pure old standard getty lacked that INITSTR completely,
but there are other getty implementations with INITSTR functionality and
none of them worked properly if used for modem initialization purposes
except if designed specially for that case (like mgetty).
 

> I mean, I can't just take your work for that.
> Show me that most/all gettys which have this option
> explicitly say that this data are not meant to be used
> to initialize modem.

I do neither have access to the documentation of all getty
implementations I used nor like to spend time searching for such
documentation. It was a suggestion, consider to use or drop it. Your
decision.

IMHO. The current Busybox getty is modeled close to standard getty
functionality and as this is error prune on none local lines if INITSTR
is used for modem initialization. Either drop that (modem init) feature
or getty needs rework to function proper on none local lines (which
currently fails). That was the intention of my note (from context of
thread).

--
Harald
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox

Reply via email to