[GRASS-dev] r.cross 0 output inconsistency

[Michel Wortmann]

Hi devs,
I found a strange r.cross behaviour, the output includes a random 0 
category without any label information. Here an example in the NC sample 
location:

g.version -rg
 > version=7.0.0
 > date=2015
 > revision=00000
 > build_date=2015-01-24
 > build_platform=x86_64-unknown-linux-gnu
 > libgis_revision=64007
 > libgis_date="2015-01-08 20:08:36 +0100 (Thu, 08 Jan 2015) "

# set region so that all maps have non-NULL values
g.region n=224470 s=219020 w=633040 e=640160

r.cross basins,landuse,soils out=test__

r.mapcalc exp='test__0=if(test__==0)'
r.stats test__0 -c
 > 0 335445
 > 1 189

These 0 areas change when the region is changed or a mask is applied. 
The only correlation with the input data I could find was the fact that 
one of them always seems to have the category 1 in these areas. That 
seems like an inconsistent behaviour and annoying if you expect non-zero 
output or rely on the label information of the output.

Best regards,
Michel