[Yum-devel] ranked multiterm searching

[seth vidal]

On Tue, 2007-06-05 at 11:38 -0400, James Antill wrote:
> On Tue, 2007-06-05 at 17:12 +0200, Florian Festi wrote:
> > Hi!
> > 
> > >  Well one of the reasons google can get away with this is that they only
> > > display the first X matches. Doing that in yum search, by default, would
> > > probably up the usability too.
> > >  Maybe yum search --all-matches, for the everything?
> > 
> > If you do that you need a parameter where to start with the results to be 
> > able to see the "next pages" and even an parameter for the "page 
> > size"/number of results.
> 
>  Well, I'm not sure you'd want to go that far (at least not before yum
> search is near instant). I was thinking more like two cases:

what version of yum are you using? With 3.2.0 search isn't quite instant
but it is very, very fast. 2-4 seconds when I'm using 'time'. Its faster
when you run it as a user b/c it doesn't touch the repos.

> 1. yum search foo - output "small" number of results, and if there is
> more than that a message of "X of Y more results, use --all-matches to
> see them all".


we're making what searchGenerator() has to hand back very, very, very
ornate by doing this. Think of all the returned data that the caller
will need to take into account?

-sv