SRS Ignored in GetCoverage

[warmerdam]

It seems that SRS is ignored in the GetCoverage, and the coverage is 
erroneously returned with the SRS of the map instead of the SRS of the layer
on return. 

For instance:

http://gdal.velocet.ca/cgi-bin/mswcs?REQUEST=GetCoverage&SERVICE=WCS&COVERAGE=TM&BBOX=448285,
4218488,463645,4233848&WIDTH=512&HEIGHT=512&FORMAT=GeoTIFF8&SRS=EPSG:32611

The layer is actually in EPSG 32611 (UTM 11 WGS84)  but the result file is just
WGS84 instead.

[sdlime]

You're correct, although it's CRS not SRS. I have got it handling RESPONSE_CRS,
which was easy since it's basically loading map->projection from a string. The
CRS is a bit more complicated I think...

[fwarmerdam]

I gather the CRS is the coordinate system of the BBOX.  My suggestion
for now is that we require the CRS and RESPONSE_CRS to match and if they
don't return an exception. 

As noted elsewhere, omitting REPSONSE_CRS is not causing the code to use
the CRS as the RESPONSE_CRS which is the action required by the spec. 

[sdlime]

Since CRS is required I know the fix. The only slight twist is when 
RESPONSE_CRS is present. In that case we have to project the BBOX before using 
it. Otherwise CRS and RESPONSE_CRS just modify the mapObj projection. I'll add 
the code today.

[sdlime]

Frank, have you had a chance to verify the fix (assuming I made it)?

Steve

[tomkralidis]

Any update on this one?

[tamas]

Replying to sdlime:

Since CRS is required I know the fix. The only slight twist is when 
RESPONSE_CRS is present. In that case we have to project the BBOX before using 
it. Otherwise CRS and RESPONSE_CRS just modify the mapObj projection. I'll add 
the code today.

Steve,

I've just run into the same problem that the BBOX is to be reprojected when both the CRS and the RESPONSE_CRS are passed. In addition RESX and the RESY should also be adjusted when the WIDTH and the HEIGHT are not specified in the URL. I attach a patch that solved my problem, however I'm not sure in that all of the consequences have been considered properly.

[sdlime]

Is this an issue in both 1.0 and the newer stuff Frank is working on? I will try to take a peek yet this week.

Steve

[tamas]

Replying to sdlime:

Is this an issue in both 1.0 and the newer stuff Frank is working on? I will try to take a peek yet this week.

Steve

I'm not sure about the 1.1 implication, but I've already posted a thread in the mapserver-dev list in this topic for comment.

​http://www.nabble.com/WCS-RESPONSE_CRS-and-BBOX-projection-td17328294.html

Tamas

[sdlime]

Status? My inclination is to close...

Steve

[dmorissette]

5.6 is over, moving to 6.0. Is this still an issue or not?