On Mon, Mar 11, 2002 at 02:50:13PM +0100, Richard Atterer wrote: > > But just now a relatively elegant solution occurred to me which allows > me to use the fallbacks *and* download 10 files at a time. I'll change > the next version of jigdo-lite to work like that. It'll also solve the > problem of *modified* (rather than disappeared) files on servers. >
Cool! Sometimes all you need is someone like me to ask a dumb question.. :^) > > > 2) allow the entry/selection of multiple mirrors, so that I can > > > distribute my load aross multiple mirrors. > > I see no way to do this with the current jigdo-lite. Eventually, I > hope to finish the jigdo GUI app which will do efficient HTTP > pipelining and maybe also concurrent downloads from 2 servers. Hmm, I see what you mean; using multiple servers *concurrently* would be a major change in jigdo-lite. I was just thinking of rotating through various mirrors on each call to wget, so as to download fewer total megabtyes from each server. I thought I would take the file list output by jigdo-file, use sed to replace the mirrors with "USMIRROR/" or "NONusMIRROR/", and then, within fetch(), read a url and replace those strings with a mirror uri selected from a list (probably using sed again, in reverse). Now I'm not sure whether it's worth it though; most mirrors manage their own bandwidth to what they consider acceptable limits on their own. Cheers, -(-- John -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
