[gdal-dev] gdalwarp trunk and FWtools giving different results

[Joaquim Luis]

Hi,

I am very puzzled with this difference between the outputs of gdalwarp 
form a couple of days old Trunk version versus FWtools 2.2.0

In both cases the command is the following
gdalwarp -s_srs "+proj=merc +R=6371008.7714" -t_srs EPSG:4326 alg.tiff 
alg_geo.tiff

but the trunk version has

Upper Left  (  -8.0862815,  37.3460021) (  8d 5'10.61"W, 37d20'45.61"N)
Lower Left  (  -8.0862815,  37.0644230) (  8d 5'10.61"W, 37d 3'51.92"N)
Upper Right (  -7.7343076,  37.3460021) (  7d44'3.51"W, 37d20'45.61"N)
Lower Right (  -7.7343076,  37.0644230) (  7d44'3.51"W, 37d 3'51.92"N)
Center      (  -7.9102945,  37.2052125) (  7d54'37.06"W, 37d12'18.77"N)

whilst  FW

Upper Left  (  -8.0862815,  37.1605907) (  8d 5'10.61"W, 37d 9'38.13"N)
Lower Left  (  -8.0862815,  36.8792107) (  8d 5'10.61"W, 36d52'45.16"N)
Upper Right (  -7.7339340,  37.1605907) (  7d44'2.16"W, 37d 9'38.13"N)
Lower Right (  -7.7339340,  36.8792107) (  7d44'2.16"W, 36d52'45.16"N)
Center      (  -7.9101077,  37.0199007) (  7d54'36.39"W, 37d 1'11.64"N)

Note the differences in latitude.
The FW solution, as far as I can tell from plotting a coastline, is the 
correct solution.

What can explain those differences.
NOTE, I do have GDAL_DATA pointing to the SVN data directory

Joaquim Luis