Re: Protocol questions - INPUT REQUIRED.

[Gerry Duprey <gerry@xxxxxxxxxxx>]

D'oh!

Everywhere I said CR, I really mean LF (i.e. chr(10)).

Sorry about the mental slip-up :-)

Gerry

Gerry Duprey wrote:
> Howdy,
>
>
>>Actually... Not really - as I believe a couple of the
implementations use
>>the { and } bracses to delimit the header and body.
>>
>>That being said - the protocol document is unclear in this regards,
so let's
>>open it to the floor - thoughts?
>
>
> Hmmm.  my take on this is that anything textual should be able to
follow the
> = in a name=value pair.  The use of the { and } characters right now
in
> terms of header/body separation is that they are the first character
after a
> CR and before another CR.
>
> I think even small devices can handle a parse check that says "if
a { is
> received and it's the first character of a line, it's special.  If
not, it's
> just another character".
>
> I tend agree with the original premise that what follows the = must be
> textual (only chars between 32 and 126 ASCII) and the value ends with
a CR.
>
> I think putting other restriction on the characters in the value
portion
> could be problematic (not so with the names -- those are expected to
be more
> limited and follow a more restrictive set of rules).
>
> Well, that's my opinion :-)
>
> Gerry

--
Gerry Duprey
Ann Arbor, MI 48103
http://www.cdp1802.org