[mapserver-dev] Single-pass query RFC updated...

[Steve Lime]

I had thought about that. I think that will break more apps since most folks only use msLayerGetShape() in
the context of queries. Which is worse? I do like the more explicit name of msLayerResultSetGetShape(). 

Note that for many drivers the  msLayerResultSetGetShape() would just wrap msLayerGetShape() since
the result set indexes reference the main dataset. Shapefiles would work like this. Any driver that didn't/couldn't
maintain a result set handle could do it this way and would work like do now.

Steve

>>> On 6/5/2009 at 3:12 PM, in message <4A297C21.7050009 at mapgears.com>, Daniel
Morissette <dmorissette at mapgears.com> wrote:
> Steve Lime wrote:
>> Hi all: I've updated RFC 52 
> (http://mapserver.org/development/rfc/ms-rfc-52.html) to reflect work done 
> at the code
>> sprint and subsequent testing, thoughts about query files etc... I'd like to 
> move on this relatively soon before it gets to hard to merge things. Please 
> take a look and pass along comments. Paul R. and Jim K. can comment 
> specifically on the performance improvements they've seen with PostGIS and 
> Oracle respectively.
>> 
> 
> I like this overall. Except that the change in the behavior of 
> msLayerGetShape() could break some MapScript applications. Could we not 
> leave msLayerGetShape() unchanged and instead add a new 
> msResultSetGetShape() that acts on the current result set?
> 
> About the OGR driver, my first reaction is that I don't expect any 
> significant problem (famous last words...)
> 
> Daniel