ietf-nntp Response codes - debugging, authentication & unknown

Lee Kindness lkindness at csl.co.uk
Thu Aug 10 08:16:52 PDT 2000


Clive D.W. Feather writes:
 > Lee Kindness said:
 > >    If a client receives 
 > >    an unexpected response code from a command then it MUST use
 > >    the first digit of the response to determine the result. For
 > >    example 2xx would be a successful response, 4xx and 5xx
 > >    unsuccessful.
 > This forbids a client from saying "This server is broken, I'm
 > quitting.".

In which case it's more likely that the client is broken. This
requirement allows the NNTP spec to grow in the future - a new
response code could be added to an existing command without breaking
deployed clients. As it is now half wouldn't grok it - I don't think
this is desirable...

Another question is if NNTP should improve this way, or if all
additions must go via 'LIST EXTENSIONS'... But with the above wording
we're leaving the door open.

 > It also forbids it from saying "Oh, I know what a 407 response is: in this
 > context it's really a success".

In which case the 407 respose would be documented somewhere, and
explicitly checked for... Yes?

-- 
 Lee Kindness             lkindness at csl.co.uk
 Software Engineer        +44 (0)131 5575595
 Concept Systems Ltd.     http://www.csl.co.uk/



More information about the ietf-nntp mailing list