[GRASSLIST:3972] SV: Some questions about 5.3 and r.cva - got it to compile

[Lars Forseth]

Hi!

Thanks for your answers. Late last evening I got it to compile. Had to edit "make_point3.o"; the file Paul Kelly pointet out the compile error came from. Commenting out the reference to G_malloc()in that file and line, it then compiled! I ran a quick test just running it on a .dem file and nothing else, generating a random raster cva raster. It worked, it seems!

Regarding including r.cva in GRASS I can just say that it would be great. Problem is; as long as it is'nt under any particular licence (GPL, BSD) I guess it must be considered as essentially Mark's private work and Intelectual Property. So Mark's permission must be gotten to make this happen. I'l mail him also on this! If he could be persuaded to publish the module under the GPL it could be included(?) and others could improve on it. 

Regards!
larsf


-----Opprinnelig melding--------------------------------------
Fra: Benjamin Ducke [mailto:benducke at compuserve.de]
Sendt: 19. juli 2004 18:33
Til: Lars Forseth; grasslist at baylor.edu
Emne: Re: [GRASSLIST:3953] Some questions about 5.3 and r.cva


Hi,

I believe there are some instructions included in the r.cva
source package that tell you to adjust the binary paths
in your GRASS source tree (INSTALL.txt).
If you don't the compiled modules are likely to end
up in your source tree instead of the GRASS 5 binary
installation tree (usually /usr/local/grass53 or similar) and
you won't be able to execute them.

However, if you get it to compile and install it still won't
be very useful because it is essentially GRASS 4.X code and
does not handle floating point raster, GRASS 5 sites format
or NULL values.
I have done a conversion of r.cva to GRASS 5.0/3/7
that works with floating point und NULL values and also
uses the new sites format.
I have tried to email Mark a while ago to ask him if I was
allowed to distribute this updated version but got no
reply so far.

I am a bit reluctant to just publish his code without his
expressed affirmation.
I will try to contact him again about it and hopefully
get a reply (Mark, do you read this -- please answer!).

Cheers,

Benjamin

On Mon, 19 Jul 2004 09:48:59 +0200
"Lars Forseth" <lars.forseth at ntfk.no> wrote:

> Hi!
> 
> I'm wondering if anyone has tried to compile Mark Lake's r.cva module on 5.3. Is that possible?
> 
> Reason behind the question is that I yesterday tried to compile 5.0.3 on a pc (AMD processor, Suse 8.2 pro) in my home. After some trouble with diverse libraries (GD, tclkt and BLAS/Lapack) missing I got it to compile. But BLAS/LAPACK was not found it reported at the end of ./configure (but installed from SUSE 8.2 DVD, are they broken/too old?)
> 
> Then I tried to compile r.cva, following the instructions. However, issuing gmake5 made nothing happen, error being unrecognized command. 'Locate gmake5' found it, so issuing it with full path made it execute, but it then processed the whole sourcetree (a lot of modules flashed by! All?). I'l try another machine this evening, and certainly be back with other questions then!
> 
> 
> regards
> larsf
> 
> 
> -------------------------------------------
> Lars Forseth
> Archaeologist Nord-Trøndelag County Council
> Tlf: 74111270 Mob: 97528143
> -------------------------------------------
>