[GRASSLIST:7362] Re: [bug #3112] grass has been running for three days

[Hamish]

> >> grass has been running for three days to create watershed related
> >> maps. 
> >> At this moment, it has been in this stage for a long time (almost
> >> one day):
> >> ~~~~~~~~~~
> >> Running: /usr/lib/grass/etc/water/ram el="el_5" t=100 se="str_5" 
> >> ha="hbas20_50"
> >> SECTION 1a (of 5): Initiating Memory.
> >> SECTION 1b (of 5): Determining Offmap Flow. Percent Complete:
> >> SECTION 2: A * Search.                      Percent complete:
> >> SECTION 3: Accumulating Surface Flow.       Percent complete:
> >> SECTION 4: Watershed determination.
> >> SECTION 5: Closing Maps.
> >> ~~~~~~~~~
> >>
> >> I am not sure if  grass will be able to complete it's job ?
> >> or is it a good idea to cancel the job and start again with 
> >> smaller region? 
> >> what do you suggest ?
...
> > prcessing very large maps is quite slow with r.whatershed (I guess
> > you are using this command!).
..
> as far as I see r.terraflow does not produce half basin.
> I need a module that creates half basin other than r.watershed.


Looks like the while loop in raster/r.watershed/ram/close_maps2.c
line 38 is never satisfied.

Maybe same endless call to close_array_seg() is what causes bug #3112?

Things often report 98% when actually done as G_percent() is often 
called in the wrong place or doesn't get called after exit. (e.g. watch
r.terrflow run)  So it is likely it finished & then got stuck on the
cleanup step.


-- how big is the test region?
-- r.info:  CELL map, DCELL map?
    (idea- try a CELL version [r.mapcalc 'newmap=int(oldmap)'])


that's about as deep as I can look into it right now.


Hamish