[gentoo-dev] Retries for emerge and lo issue.

[Chris Davies]

Hi,

I'm just installing Gentoo at this moment (it takes a fair amount of time 
on a modem :), and I've had a few ideas about what I'd like to add to the 
emerge script if possible. Do you accept patches from anyone, or does a 
perspective developer have to go through some debian-esque procedure?

Anyway, my idea was essentially to do with emerge scheduling and retries 
when doing 'emerge system', and just retries in general when doing any 
other emerge. I thought since you had three mirrors (at least one of which 
appears somewhat unreliable) then emerge could check each one in order of 
increasing ping time to establish which site to get a package from. I also 
thought that if it got error responses from all three sites it might queue 
the item for later download if it wasn't needed for a dependency 
immediately, perhaps this could be controlled by a MAXRETRIES variable in 
global make configuration file. Do you think this would be worth 
implementing, or do you think it would make the system to complex?

Also I noticed that the loopback interface is not started by default, if 
the user doesn't start it then they'll notice about the time that Perl 
finishes downloading. Is there any reason that this interface isn't started 
by the init process?

Anyway, thanks for the time spent on this nifty distribution.
C.Davies
(c.davies@×××××××.org)