'r.line' problem

[Jim Westervelt]

Jinn-Guey Lay writes (jinn at uhunix.uhcc.Hawaii.Edu):

# In my test, the result from 'r.line' is not quite right.
# The program does convert all the streams but the result
# looks like spagehtti. It always connects the last point of
# an arc to the first point of the next arc with a straight
# line.

Yes, r.line and r.thin are still in the alpha directory because they are
not quite right.   The problem with r.line noted above is the only problem
that I have seen - as far as I know it has not been fixed.   Does look like
a simple fix by the symptoms.   We'll all be most interested in a fix.  Many
of us would be more interested in visiting Hawaii to do the fix.  Your
choice.

Let me note that the sister program r.thin needs fixing too.  It thins
features down into single lines.  The raster results of r.thin should 
feed directly into r.line (which converts raster lines to vector lines).
Often however r.line complains that the lines are not thinned and that
the map should be fed through r.thin (which is where it just came from).
Interestingly, feeding the result of r.thin back through r.thin results in
a DIFFERENT output.   The rules for thinning alters an already thinned map.

Any takers on fixing this problem will be blessed.