Opened 19 years ago
Closed 17 years ago
#1372 closed defect (fixed)
if RESX not equal to RESY only first band contains data
Reported by: | Owned by: | warmerdam | |
---|---|---|---|
Priority: | high | Milestone: | 4.10.3 release |
Component: | WCS Server | Version: | 4.5 |
Severity: | normal | Keywords: | |
Cc: |
Description (last modified by )
If resx not equal to resy then first band contains data, all other bands are blank, the map file is as below and a sample request is http://huby/cgi-bin/admin?REQUEST=GetCoverage&SERVICE=WCS&Version=1.0.0&Coverage=Ozone__Potential_vorticity&time=2002-09-21T12:00:00&FORMAT=GTIFF_FLOAT&RESX=3.75&RESY=2.5&CRS=EPSG:4326 LAYER NAME 'Ozone__Potential_vorticity' STATUS OFF TYPE RASTER DUMP TRUE TILEINDEX 'Ozone__Potential_vorticity_idx' PROJECTION "init=epsg:4326" END METADATA wcs_label 'Met Office Ozone data' ows_extent '-1.875 -91.25 358.125 91.25' wcs_resolution '3.75 -2.5' ows_srs 'EPSG:4326' wcs_formats 'GTIFF_FLOAT' wcs_nativeformat 'NetCDF' wcs_bandcount '27' wcs_rangeset_axes 'bands' wcs_rangeset_label 'Pressure (hPa) Levels' wcs_rangeset_name 'bands' wcs_rangeset_description '0.09999999403953552 0.19999998807907104 0.30000001192092896 0.5 0.699999988079071 1.0 2.0 3.0 5.0 7.0 10.0 20.0 30.0 50.0 70.0 100.0 150.0 200.0 250.0 300.0 400.0 500.0 600.0 700.0 850.0 975.0 1000.0 ' wcs_timeposition '2002-09-21T12:00:00,2002-09-22T12:00:00,2002-09-23T12:00:00,2002-09-24T12:00:00,2002-09-25T12:00:00,2002-09-26T12:00:00,2002-09-27T12:00:00,2002-09-28T12:00:00,2002-09-29T12:00:00,2002-10-01T12:00:00,2002-10-02T12:00:00,2002-10-03T12:00:00' wcs_timeitem 'time' END END Many thanks for your help, quite happy to take on some work :-)
Attachments (1)
Change History (11)
by , 19 years ago
Attachment: | mapresample.c added |
---|
comment:1 by , 19 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Attached proposed update, please let me know if it is ok and if it will be included in CVS (I only have read access to CVS). Many thanks for the help, not sure of the procedure for adding fixes, let me know if I haven't done it correctly!
comment:2 by , 19 years ago
Resolution: | fixed |
---|---|
Status: | closed → reopened |
Reopening until the patch is applied. Cc'ing Frank as well since that is his code. Thanks for doing the legwork on this one Norman... Steve
comment:3 by , 19 years ago
Cc: | added |
---|
comment:4 by , 19 years ago
Owner: | changed from | to
---|---|
Status: | reopened → new |
Norman, Somehow I snoozed through this. I looked at your patch, and it looks generally OK, but I think it is necessary to move the band loop inside the per-pixel loop to avoid redoing the transformation calculations for each band to be processed.
comment:5 by , 19 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
I have applied a variation on the suggested patch in CVS-HEAD and it seems to work. I have also created an msautotest test to excercise this issue. It is the gdal/nonsquare_multiraw.map test. I have also backported the fix into the 4.6.x branch.
comment:6 by , 19 years ago
Resolution: | fixed |
---|---|
Status: | closed → reopened |
I checked out MapServer completely and recompiled, and the bug is still there with the change Frank has put in to mapresample.c. Reapplying the patch I attached previously fixes the problem! So I have reopened the bug. thanks for the help.
comment:7 by , 19 years ago
Cc: | added |
---|
comment:8 by , 17 years ago
Description: | modified (diff) |
---|---|
Milestone: | → 4.10.3 release |
Priority: | high → highest |
This needs to be review promptly, hopefully for a 4.10.3 release.
comment:9 by , 17 years ago
Priority: | highest → high |
---|
Norman,
Would it be possible to get details to reproduce this problem? I wonder if the issue might be related to the one reported and fixed in #2180?
Rather than patch this blindly, I'd really like to reproduce it at my end before I change anything.
comment:10 by , 17 years ago
Resolution: | → fixed |
---|---|
Status: | reopened → closed |
Tested with version 4.99 on a single data file, and with a tileindex, returning data for all bands.
Fixed in trunk.
Note:
See TracTickets
for help on using tickets.
updated version of mapresample.c with iteration over available bands