As described in the manual pages, the current r.drain program
can not accept more than one pair of x,y coordinates as the
starting drain locations.
There is another bug - the current r.drain does not take
a map layer as the source of starting locations.
Now these two bugs have been fixed, temporally call r.drain.rsc
at our Remote Sensing Center. If enough responses about this,
I'll either put the fixed code into our ANONYMOUS FTP site or
sent it to the CERL.
BTW, I think that a combination of r.cost, r.drain, etc., doesn't
always provide a one-route path. I want to code a 'r.path' program
based on r.cost, but would like to hear enough positive feedback
about its usage and the need first.
Jianping Xu jianp@ocean.rutgers.edu
Rutgers University