On 2/17/06, Jody Garnett <[EMAIL PROTECTED]> wrote:
> Simone Giannecchini wrote:
> > ---------- Forwarded message ----------
> > From: Simone Giannecchini <[EMAIL PROTECTED]>
> > Date: Feb 17, 2006 4:07 PM
> > Subject: Geoserver and crs
> > To: [EMAIL PROTECTED]
> >
> >
> > Hi list,
> > I have been working lately on a new version of the
> > wcs-wms_with_rasters and I ran into an annoying problem.
> >
> > In the getmap request, if I am right, the envelope is lon-lat
> > regardless the crs. This means that if I request a map with SRS
> > epsg:4326 we receive a bbox with lon first but we should internally
> > handle with lat first.
> >
> For WFS the request is always in the target coordinate system (which
> means client code has to convert from the advertised lon-lat bounds to
> the traget crs in order to make a request.
>
> What WMS can do depends a little bit on what version you are using, talk
> to richard and look at the different WMS strategies for the exact details.
> For example in WMS 1.3 your request can include the CRS and your
> bounding box in a GetMap request.

What I am talking about is basically wat follows.

If you take a look at the RendereUtilities class and you check the
worldToScreenTransform method you can see that it makes the assumption
that your crs is always Lon,Lat but if the underlying crs is lat,lon
it fails to compute the transform. This happens if you use EPSG:4326
for example.

When we parse the request parameters for a getmap in geoserver we
build a JTS Envelope object which is correctly lat, lan but when we
set the CRS using CRS.decode the axis order can be different as it
happens regularly with EPSG:4236 (see GetMapRequest file in geoserver
code). Hence when we build the worldToScreenTransform  inside the
Streaming renderer the transform is uncorrect as I sasid above. I am
investigating the part of the renderer where features are rendered in
order to understand how this problem has been overcome.

I fixed it creating a new worldToScreenTransform   method which takes
into account the crs.

I hope that someon can tell me something about it, in the meanwhile I .

Simone.






> > Despite of this in RenderUtilities the worldToScreenTransform assumes
> > that the crs is always lon,lat, is this an error or it is stupid me?
> >
> Um we use StreamingRenderer all the time with all sorts of projections.
> > I found other minor issues which I would like (in conjunction with
> > alessio) fix but this is an interesting/annoying one and I wanted to
> > hear someone else's opinion.
> >
>
>
>


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid3432&bid#0486&dat1642
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to