r3.mapcalc error

[mmetz]

the error occurs with the new temporal gis tutorial, tested on different machines:

r3.mapcalc --o expr="t5ym = (t1ym[0,0,-2] + t1ym[0,0,-1] + t1ym[0,0,0] + t1ym[0,0,1] + t1ym[0,0,2])/5.0"

ERROR: ERROR: Rast3d_get_double_region: error ERROR: Rast3d_get_double_region: Rast3d_get_double_region:in Rast3d_get_tile_ptr
error in Rast3d_get_tile_ptr
 error in Rast3d_get_tile_ptr
WARNING: Unable to flush index
WARNING: Unable to create 3D raster map <t5ym>

Any ideas?

[huhabla]

Unfortunately i can not reproduce this error and i did not found any memory leak with valgrind on my 64Bit system.

Can you please try r49789. I don't know if i have fixed the error, but at least a more detailed error message will appear.

[glynn]

Replying to mmetz:

Any ideas?

Either the map is corrupted, or there's a bug in the raster3d library. Requesting a cell which is outside of the map should just produce a null value. FWIW, I don't think that r49789 will affect this, as Rast3d_nearest_neighbor() already performs bounds checks before calling Rast3d_get_value_region().

[huhabla]

Replying to glynn:

Replying to mmetz:

Any ideas?

Either the map is corrupted, or there's a bug in the raster3d library. Requesting a cell which is outside of the map should just produce a null value. FWIW, I don't think that r49789 will affect this, as Rast3d_nearest_neighbor() already performs bounds checks before calling Rast3d_get_value_region().

This is true indeed. I have removed the redundant spatial boundary check from Rast3d_nearest_neighbor, as IMHO this should be done in Rast3d_get_value_regions.

Looks like a bug in Rast3d_coord2tile_index() or Rast3d_get_tile_ptr() ....

[martinl]

No activity for a long time. Closing. Feel free to reopen if needed.