[NNTP] Snapshot 4
Andrew - Supernews
andrew at supernews.net
Wed Dec 1 02:38:06 PST 2004
>>>>> "Russ" == Russ Allbery <rra at stanford.edu> writes:
>> 3) The command isn't used very frequently (LIST ACTIVE is the only
>> existing LIST subcommand that gets heavy usage)
Russ> This is where I don't think that it does. If clients do what
Russ> we're going to tell them to do, CAPABILITIES will be used at
Russ> least once a session, perhaps more. That may still not qualify
Russ> as very frequently, but I think it's still as frequent as LIST
Russ> ACTIVE or perhaps more.
Once per session isn't what I call "very frequently". LIST ACTIVE is
often used thousands of times per session or more (LIST ACTIVE
group.name being a common way to check for new arrivals, being faster
than GROUP). Anything used only a couple of times per session is down
in the noise in comparison to LIST ACTIVE and the common top-level
commands (ARTICLE, OVER/XOVER, GROUP etc.)
Russ> Certainly my intention is to hard-code the output of the
Russ> command except for SASL mechanisms and the needed switches to
Russ> add or remove commands that aren't available for a particular
Russ> session and make the command reasonably fast. Note that
Russ> currently in INN, we weren't able to share code for LIST
Russ> EXTENSIONS with the other LIST commands (most of which do all
Russ> share code).
INN's implementation of the other LIST subcommands is possibly a bit
more rigid than mine is.
--
Andrew, Supernews
http://www.supernews.com
More information about the ietf-nntp
mailing list