[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mapserver-dev
Subject:    [mapserver-dev] PJ pool
From:       "Paul Ramsey" <pramsey () cleverelephant ! ca>
Date:       2008-11-05 6:05:47
Message-ID: 30fe546d0811042205g4bec532aje0baab832a13e2d4 () mail ! gmail ! com
[Download RAW message or body]

So, I read over mappool.c while watching the election coverage and it
looks pretty convincing to me... I'd like to have one for projections.
 I feel silly just essentially cloning it, but it works, and a
slightly simpler version would serve the purpose.  Seems like the main
issues are thread safety, since these things will be global.

P.

On Tue, Nov 4, 2008 at 8:51 PM, Paul Ramsey <pramsey@cleverelephant.ca> wrote:
>
> I think the expedient answer is an in-memory cache of PJ structs,
> filled up on an "as needed basis".
>
> I think of it like this:
>
> You have a map file with 10 layers. Odds are, all/most of your data is
> going to be in the same SRS. We cache the first object in memory, and
> the rest use it. Instant 10x performance boost. Add in FCGI and the
> boost goes to infinity (or something). It's something that is
> self-contained in Mapserver and will work right away, on old versions
> of Proj.
>
> Paul
>
_______________________________________________
mapserver-dev mailing list
mapserver-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/mapserver-dev
[prev in list] [next in list] [prev in thread] [next in thread]