What steps will reproduce the problem?

1. Postion any point on a map passing WKT parameter:

http://wms.latlon.org/?format=image/png&layers=osm&width=800&height=600&bbox=-83,-41.8,67,68.8&wkt=POINT(-8.55%2042.85)

The above point is positioned in Ireland (as a green mark). But in fact, the point coordinates are from Spain (-8.55 W, 42.85 N).

2. Change any of the latitude limits of the bounding box (2nd & 4th values of bbox), but don't change the POINT coordinates:

http://wms.latlon.org/?format=image/png&layers=osm&width=800&height=600&bbox=-83,11.8,67,68.8&wkt=POINT(-8.55%2042.85)

Now the point is drawn at the sea, outside of Ireland

3. So, the bounding box is affecting wkt positioning, I guess. Changing bbox should only have affected map boundaries and stretching (for a given width/height), but not the point position relative to background map.

What is the expected output? A well positioned point over a background map (based on its WSG84 geographical coordinates)

What do you see instead? A bad positioned point

What version of the product are you using?
tWMS project demo server at http://wms.latlon.org

On what operating system? http://wms.latlon.org operating system (not mine)

Please provide any additional information below.

My own installation (twms source 0.01q on Ubuntu 10.04, web.py standalone installation) behaves in the same wrong way

• same result if I specify EPSG:4326 as the mapped bounding box coordinates, by adding this to the URLs: &srs=EPSG:4326

• same result using latest code downloaded with mercurial: hg clone https://twms.googlecode.com/hg/ twms