431 for TAKETHIS

Russ Allbery rra at stanford.edu
Fri Sep 17 12:07:12 UTC 1999


Fabien Tassin <fta at oleane.net> writes:

> It's not the goal of TAKETHIS to deal with retries. If you really want a
> try-again-later the best way to do it is to use CHECK or IHAVE methods.

This is largely due to the problem that "try again later" was designed to
handle.  If you have multiple peers offering you articles at the same time
with CHECK, you may get an article offered from both peers at once.  If
you just accept from one peer and refuse from the other, and the first
peer never sends you the article, you may never get the article at all.
So you send "go ahead" to one peer and send a deferral to the other.

If you've just received the entire article with TAKETHIS, there's no
reason to do that; just accept and store the article that you just got and
reject/refuse additional copies.

Deferrals were designed for the specific situation of multiple feeds
offering the same articles at near the same time; they're not really a
general deferral mechanism.

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


More information about the inn-bugs mailing list