[GRASS-dev] Re: WinGrass testing results 6.3.0rc2

[Moritz Lennert]

Michael Barton wrote:
> This helps a LOT. I think it also explains why NVIZ doesn't work (although I
> don't understand why it DOES work on some systems). NVIZ is launched from a
> script.  > IIRC, there at least some scripts get *.bat equivalents. Is/was this
> the case for the script named "nviz"?

Yes and no. The *.bat equivalents of the shell scripts are acutally just 
a call to the GRASS_SH to launch the shell script, where GRASS_SH will 
be msys' sh.exe for those who installed it and configured the 
grass63.bat file.

Nviz' case is different as the nviz.bat file directly calls 
%GISBASE%\etc\nviz2.2\nviz.exe and so doesn't need a shell. So, you do 
not need to have msys installed to launch nviz.


> Everyone here has installed mysys as
> suggested in the README. But it is nice to know that it is NOT needed. The
> problem then is that you can't run any of the scripts and some of these are
> important.

Obviously they are, although most are "just" a combination of different 
grass commands with some awk/sed/etc magic thrown in as a bonus. So many 
of the scripts can be reproduced by hand if really needed.

The idea for the future is to replace all the shell scripts with 
equivalents in a more portable language such as python. As python will 
be an obligatory dependency for anyone who wants to use the gui, it 
sounds like the best choice for replacement.

> 
> If I understand you correctly, if we modify the grass63.bat to make sure
> that it correctly indicates the location of mysys (we are just installing it
> in the default location), scripts (including nviz) *should* run.

As mentioned above, nviz should run in any case. If it doesn't then 
there is a problem and we need to identify it.

I'll try to find time tomorrow to squat a colleague's computer which has 
nothing (neither grass not tcltk) installed, yet, in order to test some 
more.

Benjamin also had problems with nviz 
(http://grass.itc.it/pipermail/grass5/2007-October/033804.html), but 
this was in what I would call a "non-standard" environment as he tried 
to work from within msys.

> 
> Another question for the general group here. I'm about to try to convince a
> classroom lab manager to put this on the computers in the lab. Is there any
> problem to installing mysys in a tightly managed lab environment? Are we
> likely to run into any permission problems trying to run GRASS in such an
> environment (i.e., where students do not have write permission outside their
> own home directories)?

No idea.

Moritz