[GRASS5] Re: d.rgb: different display in gis.m and d.m

[Glynn Clements]

Michael Barton wrote:

> I *THINK* you've hit a bug I reported in d.rgb and d.his about a month back.
> I'm pretty sure that they treat "0" as NULL. This is probably a holdover
> from the old 'pre-NULL' days of GRASS. Does your map use "0"s? If so, try
> changing to another value and see what happens.
> 
> I don't know why it's showing up in gis.m and not in d.m, but probably has
> to do with the x11 vs. PNG/PPM display driver architecture. I'm adding Glynn
> to the copy list because he may have some insight into what is going on.

The PNG driver uses the fallback implementation for RGB rasters, which
converts the R/G/B values to internal colour numbers then uses the
RASTER_INT operation. This does the wrong thing if the internal colour
number is zero (and black on an RGB device probably will be zero) and
overlay mode is being used.

I'll change it to provide its own RGB raster operation.

-- 
Glynn Clements <glynn at gclements.plus.com>