On Mon, 27 Jul 1998, Herbert Xu wrote: > On Sun, Jul 26, 1998 at 03:28:54PM -0700, Tom wrote: > > > > On Mon, 27 Jul 1998, Herbert Xu wrote: > > > > > without creating the relevant xover header. I'm currently writing the code > > > to get xover info from articles when necessary as you may have more articles > > > than the remote server. > > > > Is that a good idea? It would seem very slow if you have to open every > > article to create NOV data on the fly. You should see how INN's nnrpd > > performs with a NOV database: very, very, very poorly, in fact basically > > unusable for production work. > > This is only going to happen if you don't have an xover record for an article > in cache which is highly unlikely except for expired xover records. Or flushed XOVER records. I can just imagine nntpcache grinding to a halt as it opens and parses headers of 2000 articles to generate some XOVER records. Also, if it is so rare, is it even worth doing? > Also, this is only going to happen once for each article until it expires > again. Or until they are flushed again. Of if the first reader in the group doesn't bother using XOVER, and just gets articles, then the second reader in the group using XOVER, forcing nntpcache to open and read a lot of files. > -- > Debian GNU/Linux 2.0 is out! ( http://www.debian.org/ ) > Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au> > Home Page: http://gondor.apana.org.au/~herbert/ > PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt > > Tom
