Opened 18 years ago
Closed 14 years ago
#1922 closed defect (worksforme)
[MapServer-msautotest]GetCoverage with GEOTIFF_16 format is offset
Reported by: | Owned by: | warmerdam | |
---|---|---|---|
Priority: | high | Milestone: | |
Component: | msautotest | Version: | 4.10 |
Severity: | normal | Keywords: | wcs |
Cc: | warmerdam |
Description (last modified by )
I run the wcs_simple.map msautotest and for the GetCoverage request with a GEOTIFF_16 format I get an image that has a certain offset on Windows. I'm referring to wcs_16.tif image. I suspect that it is related to MS1916 bug since my Windows build is maybe not as recent as the Linux one. Tested on MS beta 3. Query string: QUERY_STRING="map=wcs_simple.map&SERVICE=WCS&VERSION=1.0.0& REQUEST=GetCoverage&WIDTH=120&HEIGHT=90&FORMAT=GEOTIFF_16& BBOX=0,0,400,300&COVERAGE=grey&CRS=EPSG:32611" diff output from gdalinfo of expected and actual result: 22,23c22,23 < Origin = (-0.000000,300.000000) < Pixel Size = (3.33333333,-3.33333333) --- > Origin = (-1.680672,301.685393) > Pixel Size = (3.36134454,-3.37078652) 27,30c27,30 < Upper Left ( -0.000, 300.000) (121d29'19.48"W, 0d 0'9.74"N) < Lower Left ( -0.0000000, 0.0000000) (121d29'19.48"W, 0d 0'0.01"N) < Upper Right ( 400.000, 300.000) (121d29'6.58"W, 0d 0'9.74"N) < Lower Right ( 400.000, 0.000) (121d29'6.58"W, 0d 0'0.01"N) --- > Upper Left ( -1.681, 301.685) (121d29'19.53"W, 0d 0'9.80"N) > Lower Left ( -1.6806723, -1.6853933) (121d29'19.53"W, 0d 0'0.05"S) > Upper Right ( 401.681, 301.685) (121d29'6.52"W, 0d 0'9.80"N) > Lower Right ( 401.681, -1.685) (121d29'6.52"W, 0d 0'0.05"S)
Change History (4)
comment:2 by , 17 years ago
Cc: | added |
---|
Replying to nsavard@mapgears.com:
I run the wcs_simple.map msautotest and for the GetCoverage request with a GEOTIFF_16 format I get an image that has a certain offset on Windows. I'm referring to wcs_16.tif image. I suspect that it is related to MS1916 bug since my Windows build is maybe not as recent as the Linux one. Tested on MS beta 3. Query string: QUERY_STRING="map=wcs_simple.map&SERVICE=WCS&VERSION=1.0.0& REQUEST=GetCoverage&WIDTH=120&HEIGHT=90&FORMAT=GEOTIFF_16& BBOX=0,0,400,300&COVERAGE=grey&CRS=EPSG:32611" diff output from gdalinfo of expected and actual result: 22,23c22,23 < Origin = (-0.000000,300.000000) < Pixel Size = (3.33333333,-3.33333333) --- > Origin = (-1.680672,301.685393) > Pixel Size = (3.36134454,-3.37078652) 27,30c27,30 < Upper Left ( -0.000, 300.000) (121d29'19.48"W, 0d 0'9.74"N) < Lower Left ( -0.0000000, 0.0000000) (121d29'19.48"W, 0d 0'0.01"N) < Upper Right ( 400.000, 300.000) (121d29'6.58"W, 0d 0'9.74"N) < Lower Right ( 400.000, 0.000) (121d29'6.58"W, 0d 0'0.01"N) --- > Upper Left ( -1.681, 301.685) (121d29'19.53"W, 0d 0'9.80"N) > Lower Left ( -1.6806723, -1.6853933) (121d29'19.53"W, 0d 0'0.05"S) > Upper Right ( 401.681, 301.685) (121d29'6.52"W, 0d 0'9.80"N) > Lower Right ( 401.681, -1.685) (121d29'6.52"W, 0d 0'0.05"S)
I ran the msautotest against MS4.99, the differences between the expected and actual results are now:
22c22 < Origin = (-0.000000000000001,299.999999999999940) ---
Origin = (0.000000000000000,300.000000000000000)
27,28c27,28 < Upper Left ( -0.000, 300.000) (121d29'19.48"W, 0d 0'9.74"N) < Lower Left ( -0.0000000, -0.0000000) (121d29'19.48"W, 0d 0'0.00"S) ---
Upper Left ( 0.000, 300.000) (121d29'19.48"W, 0d 0'9.74"N) Lower Left ( 0.0000000, 0.0000000) (121d29'19.48"W, 0d 0'0.01"N)
30c30 < Lower Right ( 400.000, -0.000) (121d29'6.58"W, 0d 0'0.00"S) ---
Lower Right ( 400.000, 0.000) (121d29'6.58"W, 0d 0'0.01"N)
What should I do with the msautotest to take care of these results on Windows?
comment:3 by , 14 years ago
Description: | modified (diff) |
---|---|
Keywords: | wcs added |
Status: | new → assigned |
Investigating.
comment:4 by , 14 years ago
Resolution: | → worksforme |
---|---|
Status: | assigned → closed |
I was unable to reproduce the indicated problem on windows, but I did find that the geotiff_16 tests are quite sensitive to the version of GDAL used due to modest variations in how GDAL produces geotiff through versions. I have switched the test to AAIGRID format for three tests in trunk (r10065) which seems to help.