Metalink support

[Matt Domsch]

On Sat, Mar 15, 2008 at 05:06:48PM +0100, Bram Neijt wrote:
> Oh, no, everything would be running on your own servers. You probably
> thought there would be an outside server involved because of the word
> "spider". I would just write some code, somebody with the right rights
> would have to run it.
> 
> With spider, I meant a script that would parse the HTML from the
> mirrormanager, so the mirrormanager wouldn't need a patch. Everything
> would of course run on your servers.
> 
> So the idea is:
> I write a script that uses a mirrorlist (which it gets from parsing
> the HTML, or in some other way) to generate .metalink files for all
> the .iso files on the mirror. The metalinks would just be placed on
> the mirror next to the .iso files.
> 
> Because the script would not have to do anything computational
> intensive (SHA1 sums have already been calculated) it could be run at
> any time you guys feel like it.
> 
> As I said above, I'd be happy to create the script, but somebody would
> have to say they would be willing to run it. I'm only asking: if I
> post the code, would anybody be able use it?

You don't want to spider this data all over again (the HTML pages are
really quite bare...) - you really want to get it from the MM database
directly.  I'd much prefer to see efforts aimed in that direction,
than for a throwaway implementation that doesn't use the MM database.

-Matt
-- 
Matt Domsch
Linux Technology Strategist, Dell Office of the CTO
linux.dell.com & www.dell.com/linux