On Thu, 2004-06-17 at 23:37, Ed K. wrote:
> Sorry for the delay.... But I was specifically recommending that you issue
> a HEAD command to a http mirror. This would be simpler then parsing the
> actual header.info
>
> {here starts a flame war}
> actually, I would encourage the use of only http mirrors, there is better
> support for caching when using proxy server. And, there is no support for
> the HEAD command for an ftp serfer. You software must support proxy server.
>
Flame war? which flame war? I agree with you. If I can save bandwith
using proxies (when the user has one) , then I'll do it.
And I just implemented the header.info parser.. Nothing much different
from the html parser I had before... I only had to change the regex to
\d:(.*)=(.*)
Now I only wish I knew what is the reason of this digit.... So far , I'm
simply ignoring it.
BTW , which directories do you think we should check? I'm thinking about
checking these directories:
updates/[$releasever,testing]
development
$releasever/SRPMS
$releasever/$arch/Fedora/RPMS
I'm not sure if checking the last two is a good idea , as they change
only when they are first created. Then they never change again... Maybe
I'll add a status file that will keep track of which mirrors have been
checked for the last two directories and have passed the check.
--
Pedro Macedo
