[GRASS5] GRASS 5.0.0pre3 selected binaries available

Andreas Lange Andreas.Lange at Rhein-Main.de
Tue Jan 22 11:43:43 EST 2002


Markus Neteler wrote:
> 
> Dear GRASS community,
> 
> selected binaries available for GRASS 5.0.0pre3
> are available now on the GRASS web sites.
> At time from
> http://grass.itc.it/index2.html
> 
> on the mirrors tomorrow.
> 
> Available:
>  - Linux (i86)
>  - SUN/Solaris
>  - Windows/Cygwin (needs Xserver)
> 
> Thanks to Andreas Lange (Linux, Windows/Cygwin) and Palle Haastrup
> (SUN/Solaris)!
> 
> Promised:
>  - Irix (maybe next week)
> 
> Desired:
>  - Windows/Cygwin with generic driver (maybe just an update package?)
>  - MacOS X
> 

Hi again,

i have to add some comments on the binaries i prepared.

Linux (Red Hat 7.0)
The Linux bin was prepared according to the instructions (first without
postgresql, so that nviz does not depend on pgsql, then with postgresql
and "make pre-compile; gmake5 src.garden/grass.postgresql", so that the
modules are created. If we want to support this feature in future i want
to suggest that we automate this somehow. Its just too much work.

I fear that we will get some problems with libgdal.1.1.so, as this
depends on:
libstdc++-libc6.2-2.so.3 (and exactly this version).

When i tested my binary on SuSE 7.0, r.in.gdal did not run, but echoed
that the library "libgdal.1.1.so" was not found. It was there, but the
library for this libstdc++ version is missing on this SuSE release. 

That is IMHO the reason why we have no real choice of using C++ code,
this would mean that we have to standardize on gcc/g++ and distribute a
binary version for each different combination of linux system and
library version.

Cygwin:
The cygwin version comes without r.in.gdal, because the latest CVS
version did not compile (ld produces stackdump). Had no time to work on
this. nviz did compile, but it does not run (even does not echo any
error message). So it is of no relevance if it is compiled with
pgsql-support.  


I think that we should find some solution for automating all this. I
believe that we should distribute GRASS mainly as binary. We need a tool
to make some sort of "nightly builds" from the CVS tree, maybe once a
week from the CVS snapshot. This would help to speed up the release
cycle without putting another burden on the GRASS developers. 

cu,

Andreas

-- 
Andreas Lange, 65187 Wiesbaden, Germany, Tel. +49 611 807850
url: http://mitglied.tripod.de/AndreasLange
mail: Andreas.Lange_at_Rhein-Main.de - A.C.Lange_at_GMX.net



More information about the grass-dev mailing list