I have disabled the offset parameter which was not having
much of the desired effect (dispersal of flow on convex areas).
There is still a bug in computation of 3D length of a flowline (2D length works) -
something is wrong in the code below but I must admit that I dont
quite understand the following line (280 in calc.c)
if (parm.lgout)
*l += parm.l3d ? hypot(length, oldz - p->z) : length;
return 1;
This is not a serious bug because 2D length should be sufficient for
most applications except for cases with steep terrain and low resolution.
Everything else seems to be running fine now (no G_calloc error).
Helena
Markus Neteler wrote:
helena,
I was taking a look at r.flow and discovered that I
get an unwanted error:r.flow help
r.flow Version 13 August 1995, update/fix October 1999FEHLER: G_calloc: out of memory
Looking into the code, I see:
io.c, line 127
default_offset = 0.0; /* fixed 20. May 2001 Helena */
default_offset_ans = (char *) G_calloc((int) log10( default_offset) + 2,
sizeof (char));
sprintf (default_offset_ans, "%f", default_offset);
offset_opt = (char *) G_calloc((int) log10( default_offset) + 4,
sizeof (char));AFAIK, log10(0.0) is always INF, which will then crash. Should it be log10(double)?
See you tomorrow
Markus