[GRASS-stats] Re: spgrass6 and winGRASS 6.4.0 RC5

[Roger Bivand]

On Thu, 3 Sep 2009, alexandre villers wrote:

> Dear members,
>
> Roger Bivand advised me to post problems on the R / GRASS interface on this list.
> I was using spgrass6 and R 2.9.2 with win GRASS 6.3.0 without much problems.
> I downloaded winGRASS 6.4 and tried to run again old script and the initGRASS doesn't work.
>
> loc<-initGRASS("C:/GRASS", home=tempdir(), override=T) #intialise GRASS dans un dossier de travail dans tempdir
> Input is not proper UTF-8, indicate encoding !
> Bytes: 0xE9 0x66 0x69 0x6E
> Error : 1: Input is not proper UTF-8, indicate encoding !
> Bytes: 0xE9 0x66 0x69 0x6E
>
> Erreur dans parseGRASS(cmd) : g.region not parsed
>
> Roger advised me this could come from file path problem so I also tried, instead of home=tempdir() home="C:/temp" and
>
> loc<-initGRASS("C:/GRASS", home="C:/temp/", override=T) #intialise GRASS dans un dossier de travail dans tempdir
> Input is not proper UTF-8, indicate encoding !
> Bytes: 0xE9 0x66 0x69 0x6E
> Error : 1: Input is not proper UTF-8, indicate encoding !
> Bytes: 0xE9 0x66 0x69 0x6E
>
> Erreur dans parseGRASS(cmd) : g.region not parsed

Stay with your previous working version of GRASS.

Next time, please include the output of traceback() run immediately after 
the error, and the output of sessionInfo().

My guess is that the error is occurring in the last, unimportant, line of 
initGRASS(), and that the location has been created successfully. What 
fails is attempting to read the XML produced by g.region(.exe) 
--interface-description, which appears to include invalid characters. I 
think that the error messages are coming from xmlTreeParse(), but don't 
know without traceback() output. This will likely affect other GRASS 
modules too.

As I wrote on R-sig-geo, I have tried to reproduce the problem with 
winGRASS 6.4.0 RC5 and R 2.9.2 without success - everything worked for me. 
If you can redirect the output of g.region(.exe) --interface-description 
to a file (with >) at the shell prompt, zip the file, and attach it to me 
off-list, I'll take a look.

The problem still looks locale-related. If GRASS was internationalised 
between RC4 and RC5 with reference to the contents of what 
--interface-description returns, and the contents are malformed, that 
would explain it. If this is the case (with apologies) make sure that your 
computer knows that it is not in a locale using non ACSII signs, or find 
the right locale for GRASS to match the one R is expecting.

Roger

>
> Any idea ?
>
> Alex
>
>
> Alexandre Villers
> PhD. Student
> Team Biodiversity
> CEBC CNRS
> 79360 Beauvoir sur Niort
>
>
> __________ Information from ESET Mail Security, version of virus signature database 4392 (20090903) __________
>
> The message was checked by ESET Mail Security.
> http://www.eset.com
>
>
>

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no