ietf-nntp DEBUG command (9xx)

[Lee Kindness]

Stan O. Barber writes:
 > [ x9x ]
 > The question is still open (unless I missed it) on what a client
 > should do should a server present these results and that server is
 > understood to be in "normal circumstances" mode by the
 > client. Should the client ignore these responses or treat them as
 > protocol errors? I would suggest the latter, but again, I am open
 > to other possibilities as long as we can get some consensus.

This is really opening up a can of worms. Does a x9x respose replace a 
normal response? Is it in addition to a normal reponse?

I think it is best to indicate that x9x responses are LOCAL EXTENSIONS 
and not toil on their usage in the specification. If a server is using 
x9x debug responses then they can knock up a client to parse them
(hell, telnet will do).

x9x responses are not in common usage and were never fully specified
by RFC-977.

Perhaps if debugging responses are thought to useful, and desirable,
we could define a DEBUG extension. If a client wants debug information 
then they would do a LIST EXTENSIONS and check for DEBUG in the
response. A subsequent DEBUG ON command would enable the debugging
responses (and DEBUG OFF disable them). The debugging responses would
be of the form 9xx and would be in addition to normal resposes - thus
if 9xx is read then you must read another line looking for the real
response (and another if its 9xx):

 [C] LIST EXTENSIONS
 [S] 202 Extensions:
 [S]  DEBUG
 [S] .
 [C] DEBUG ON
 [S] 200 Debug responses enabled
 [C] LIST
 [S] 900 0.2 seconds to retrieve newsgroup list
 [S] 901 130 moderated groups, 9873 unmoderated.
 [S] 215 list of newsgroups follow:
 [S] alt.fan.nntp-authors
 [S] and.so.on
 [S] .