Re: finger, again

At 9:53 AM 3/6/95, Jared_Rhine@hmc.edu wrote:
>It might make sense to restrict the outgoing request to only a single line,
>as I believe the finger protocol works (at least, I've never been able to
>get a server to parse more than one line).  This would reduce the
>possibility of a port 25 (SMTP) spoof.

The current (and future) draft says: "Clients should not
decode CR and LF characters in a URL."

>Speaking of end-of-lines, does your
>finger proposal specify how the <request> is terminated?

No, because...

>Does the finger
>protocol spec?

Yes, it does: a CRLF. Thus, the client reads the URL and turns it into a
request. In the case of a mindful client, this means decoding, checking for
security, checking for sanity, slapping on a CRLF, and sending it on its
way.

--Paul Hoffman
--Proper Publishing

Received on Monday, 6 March 1995 12:05:28 UTC