[Yum-devel] [Patch] Resolver Performance and Correctness

[Jeremy Katz]

On Wed, 2007-06-06 at 10:29 +0200, Florian Festi wrote:
> Jeremy Katz wrote:
> > On Tue, 2007-06-05 at 13:22 +0200, Florian Festi wrote:
> >>   * SqliteSack
> >>    * .prco uses empty tuple as default value
> >>      to destinct from empty list and avoid multiple db queries
> >>    * _search_cache
> > 
> > Didn't you previously have a patch to do this with None instead of an
> > empty tuple?  Also, what memory cost does the _search_cache end up
> > having?
>
> True. I moved to empty tuples as they are the closest distinguishable 
> replacement for an list. So the idea is that they cause less problems than 
> None as they also allow test for inclusion, iteration and so on. 

Ah, makes a lot of sense like this.

> I admit 
> that issuing an error when accessing the not yet loaded prco lists may be 
> wanted. But this is an different issue.
> 
> Single patch is attached. It also simplyfies the SQL statement which should 
> give another small speedup.

This one looks good to me.

Anyone not for it?

Jeremy