[GRASS-windows] Compiling and testing i.class

[Daniel Rodríguez Pérez]

Hello everybody.

I am new to this mail-list and not too experienced with WinGrass, so my 
apologies in advance for any "newby thing" in what follows.

I am using Grass as the software of a remote sensing course and I am 
trying to use the supervised classification module i.class.

I've seen in the winGrass Wiki that this module is still to be fixed for 
Windows. It does not compile "cleanly" due to some mingw signal trapping 
limitations (¿win32 limitations?).

 From the source code, I have guessed that the problem is due to 
SIGALARM. It is trapped to avoid infinite time iterations in one loop 
(instead of checking the time elapsed in every step, which could be very 
time consuming ¿?). So I commented out the signal trap and the program 
compiled neatly and runs from the command line (although no GUI is 
launched when called from the menu ¿?).

However, I have tried to work out some classification example to test 
it, and here comes my problem: I haven't found any step-by-step example 
to follow. A friend of mine tried following the "Grass Book", but it is 
not too explicit about this issue. When passing the parameters, i.class 
always complains about some empty on nonexistent subgroup (¿may be there 
is a problem defining groups and subgroups?).

I would like to hear any ideas.

PS: Can anybody also tell me about how to contribute the code changes to 
the winGrass developers (assuming i.class eventually works).

Thanks in advance.

-- 
Daniel Rodriguez-Perez, <daniel at dfmf.uned.es>

Depto. Fisica Matematica y de Fluidos
Facultad de Ciencias, UNED
Tel. +34 913987127