I have fixed a few things:
- r.thin writes NULL instead of 0 (src/raster/r.thin/cmd/io.c)
- r.digit writes NULL instead of zero
(fix applied in src/raster/r.in.poly/raster.c)
For the latter I am not sure how we treat 0 categories.
But the NULL fix was needed to avoid other problems.
- r.compress, r.null, r.clump, r.reclass: improved warning when
a reclassed map is found (tell user the name/mapset of the related
base map now)
- r.neighbors added test for reclassed maps since it obviously doesn't
work with those
On Mon, Nov 12, 2001 at 12:22:10AM +0100, Markus Neteler wrote:
Hi,
I have fixed a few things:
- r.thin writes NULL instead of 0 (src/raster/r.thin/cmd/io.c)
- r.digit writes NULL instead of zero
(fix applied in src/raster/r.in.poly/raster.c)
For the latter I am not sure how we treat 0 categories.
But the NULL fix was needed to avoid other problems.
I just realized that r.mfiler also outputs 0 instead
of NULL.
All r.thin, r.digit (r.in.poly), r.mfiler seem to use unusual
write routines (or not?). At least I am not very sure how
zero shall be treated. In general 0 should be supported.
Maybe my above changes switched that off for r.thin and r.digit?
r.mfiler is untouched so far.