[mapserver-dev] Couple of ideas for 6.2...

Stephen Woodbridge woodbri at swoodbridge.com
Wed Feb 15 13:17:43 EST 2012


+1 on all of these. I would like to see added to the various drivers 
like PostGIS. A long time ago I hacked the postgis code to support scale 
so I could select an appropriate simplified geometry column based on the 
scale like:

DATA "the_geom (select gid, case when !scale!>10000000 then geom10m
                                  when !scale!>1000000 then geom1m
                                  else the_geom end as the_geom
                   from mytable) as foo using SRID=4326 using unique gid"

I hear you are not implementing all the driver specific hacks, but your 
part is the first step and may be others all add the hacks.

-Steve W

On 2/15/2012 1:05 PM, Lime, Steve D (DNR) wrote:
> Hi all: I’ve got a couple of ideas for minor enhancements I’d like to
> add into 6.2 and am looking for feedback. I don’t think
> they quite rise to RFC status:
> Adding runtime constants to common expression handling:
> Basically it would be nice to expose a few runtime constants to
> MapServer logical expressions. Things like computed
> extent, scale, image size, and cellsizer. Syntax would be something like
> !name! following after the !bbox!
> hack in PostGIS. While I’m only interested in MapServer logical
> expressions it’s reasonable that certain drivers could
> support them as well within their DATA/FILTER statements.
> I’m proposing !bbox! (as define in PostGIS code), !minx! (etc…),
> !scale!, !width!, !height!, and !cellsize!.
> One use for this would be triggering scaling on parameters passed to
> style GEOMTRANSFORMs (e.g a buffer).
> Adding classgroup as a CGI parameter:
> I’ve recently discovered the usefulness of class GROUP’s in setting up a
> mapfile to render nicely with and without an
> image background. It’s very handy to manage 1 layer definition with
> alternative display configs. What sucks is triggering
> the use of a group. Maybe I’m missing something but I had to do
> URL-based config on each layer that supported
> the different class groups. It would be helpful to define a classgroup
> once and then for all layers with classes that use
> it we set the layer CLASSGROUP. If a layer doesn’t use it then no change
> is made.
> Thoughts?
> Steve
> _______________________________________________
> mapserver-dev mailing list
> mapserver-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev

More information about the mapserver-dev mailing list