ietf-nntp HDR

[Andrew Gierth]

>>>>> "Russ" == Russ Allbery <rra at stanford.edu> writes:

 > Andrew Gierth <andrew at erlenstar.demon.co.uk> writes:
 >> So do we (using Lines as an example)

 >> a) require that "HDR Lines" return only actual Lines headers from the
 >> article, which implies that it could return 503 if the server only
 >> has actual linecounts available

 >> b) define "HDR Lines" as returning either the actual linecount or the
 >> Lines: header at the server's whim

 >> c) define "HDR Lines" as referring to the actual Lines header, and provide
 >> an alternative way of specifying the "lines" and "bytes" overview
 >> fields (e.g. "HDR :Lines" - ':' is not valid in a header name, or
 >> alternatively require "HDR Lines:" to get the real header)

 >> I would go for something like (c), because existing practice isn't an
 >> issue for HDR, and the "XHDR Lines" behaviour is both a) very useful
 >> and b) a nasty wart.

 Russ> There is a fourth option which we should consider, namely

 Russ>   d) define HDR Lines and HDR Bytes to be special exceptions
 Russ>   that return the overview information rather than article
 Russ>   headers, thereby making the actual Lines: and Bytes: headers
 Russ>   in the articles, if any, not retrievable by HDR.

what about servers that don't have an overview? (I don't think diablo
does, as I recall it stores the actual headers of the article in its
index). These would have to work around the issue by "fixing" the
Lines: header so that it showed the correct value.

-- 
Andrew.