Shawn Walker wrote:
> Brock Pytlik wrote:
>> Reasons for sticking with existing approach:
>> Smaller indexes, at least so far. (40M vs 240M on my local system,
>> 272M vs 4.2G on ipkg as reported by du)
>
> Disk space is really the least of my concerns.
In general I agree, but differences of this magnitude seemed reasonable
to call attention to.
>
>> Constant depot memory usage for all searches. Using pylucene makes
>> the depot grow when searches are done for things like p* (up to 710
>> size, 650M rss).
>
> I'm quite concerned about this memory usage compared to our current
> usage. Our current usage is much lower for a build 101 depot because
> of your change to use a generator-based approach.
>
Yeah, I'm concerned, and honestly a bit disappointed by this as well.
It's possible some of this will get better when I rewhack the pylucene
method of storing things to be more sensible given it's way of storing
things and also so that it can return the entire action instead of the
current results.
> The rest of what you describe is actually great to hear.
>
Brock
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
