[GRASS-dev] Re: Problem when loading GRASS 6 raster with large integer values into R via r.out.bin

Hamish hamish_b at yahoo.com
Wed Sep 24 05:51:08 EDT 2008


Roger Bivand wrote:
> Yes in principle, but do we know whether users are still on 6.2, or have 
> migrated to 6.3 and there is no further need to condition on feature 
> creep? Remember that the R/GRASS interface isn't GRASS, it's an interface. 
> I can look at reading the GRASS_VERBOSE shell variable and using it to set 
> verbosity if present. The -q versus --q change killed the tests/examples 
> when it happened, so I'd prefer conservatism (do nothing) to using GRASS 
> facilities that vary between GRASS versions.

FWIW, --quiet and --verbose are silently ignored in GRASS 6.2.3 so
newer scripts could be shared. Prior to that it will throw an error
complaining about an unknown flag.

the GRASS_VERBOSE env var will safely be ignored if the version of GRASS
doesn't know to look for it. Probably what you should use, should you
choose to worry about verbosity levels.

-q vs --q? All -q flags in GRASS 6.x should be retained exactly for those
backwards compatibility reasons. If one has been removed then it's a bug.
Any command line script written for GRASS 6.0.0 should still work with
any GRASS 6.99.x. We've taken a lot of disciplined effort to ensure that.

One huge exception might be the change of r.out.gdal.sh to r.out.gdal (C
version), WRT observing overall map vs. current region bounds, but I am
not sure of that, probably you know better than me if that is actually
the case and causes pain.


Hamish



      



More information about the grass-dev mailing list