ietf-nntp Minor corrections to 11.4

[Russ Allbery]

Charles Lindsey <chl at clw.cs.man.ac.uk> writes:
> Russ Allbery <rra at stanford.edu> writes:

>>       Clients SHOULD use the eight-digit date form with four digits for
>>       the year.  The six-digit form is provided solely for backwards
>>       compatibility.

>> I'd even be comfortable with making that a MUST, but existing practice
>> definitely supports a SHOULD.

*grumble*  Since writing this, I checked INN 1.7.2, and it doesn't support
four-digit dates.  So definitely not a MUST.  I still think a SHOULD is in
order, but the client's going to have to be prepared to fall back on a two
digit date, at least for a while longer.  :/

> Various early versions of INN (inlcuding one server I sometimes connect
> to) interpret the 2 digit form incorrectly since y2k, as does the model
> NNTP implementation. So when people fix those bugs, they might as well
> switch to 4 digits regardless and save the hassle of that "nearest the
> current year" hack (which is a bit of a pig to implement).

Does the reference implementation support four-digit years?

200 nntp.stanford.edu InterNetNews NNRP server INN 1.7.2 08-Dec-1997 ready
newgroups 20000701 000000
501 Usage: NEWGROUPS yymmdd hhmmss ["GMT"] [<distributions>]

It's going to take a *long* time for INN 1.7 to disappear; it was
incredibly stable.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>