[GRASS-user] Fire Simulation bug

[Markus Neteler]

On Tue, Oct 20, 2009 at 10:16 PM, Adam Dershowitz
<adershowitz at exponent.com> wrote:
> I downloaded the 6.x firesimulation data set and demo script from:
> http://www.hpcc.nectec.or.th/grass/download/data6.php  (Yes, I am using the
> script on that page, not the one that is included in the compressed file)
> But, when I run it I get an error:
> It gets as far as this:
>  r.random my_spread n=10 raster_output=my_path
> Collecting Stats...
>  100%
> Writing raster map <my_path> ...
>  100%
> But then reports an error here:
> r.spreadpath -v x_input=my_spread.x y_input=my_spread.y \
>> output=my_path
> ERROR: option <output>: <my_path> exists.
> It is correct that my_path exists, because it was created in the prior step.
>  But the documentation for r.spreadpath says in part:
>      output=name Name of raster map layer to contain output. Also can be used as
>          the map layer of the input target points. If so used, the input target point
>          map will be overwritten by the output.
> It seems like the script is trying to use this as both the input and output,
> as it says should be allowed.  But it is instead giving an error.  I also
> tried to add --overwrite to r.spreadpath, but then I just get a completely
> yellow window for the output with no paths at all.
> Is there a bug in the script, or a bug in r.spreadpath?

I am not sure but in main.c I see

    /*  Set flag according to input */
    if (path_mapset != NULL) {
        if (head_start_pt == NULL)
            /*output layer exists and start pts are not given on cmd line */
            flag = 1;

        /* output layer exists and starting pts are given on cmd line */
        else
            flag = 2;
    }
    else
        flag = 3;               /* output layer does not previously exist */

which could interfere with the libgis magic to handle overwriting of files.
Perhaps r.spreadpath wasn't updated accordingly?

Markus