[Xastir] WMSradar Transparency

[Chip G.]

On Sep 15, 2006, at 14:16, Tom Russo wrote:

> Are you running CVS xastir or one of the releases?

CVS on the latest update as of sometime this morning.

> What is your color depth?

"millions" which I think is 24bit.

> For several months running, the transparent pixel value returned by  
> the server has been 0x010101, not the 0x000000 that is explicitly  
> requested in the URL for the WMSRadar geo file.

I just checked my oldest running machine (xastir-wise) and it already  
had that modification to the geo file. This works, but the down side  
is that the rest of the world (large portions of Canada) come back as  
black.

> On unpredictable occasions, the pixel value comes back 0x000000  
> instead.  But these cases have been rare over the last few months.
>
> Setting TRANSPARENT 0x010101 in WMSRadar.geo (as it currently is in  
> CVS) has been working on 16 and 24 bit displays for several months  
> (I test it on both). If you're running one of those color depths  
> and are seeing opaque backgrounds, then either:
>   - There is some difference between your 16 or 24 bit display and  
> the two I test on.
>   - The WMS server is returning 0x000000 background pixels again  
> --- it probably won't do so for very long.

Or possibly the CVS update slipped in a 0x000000 as that is what this  
was compiled against this morning (-r RELEASE).

> I just tried WMSRadar on my home machine (a 24-bit color depth  
> display) and the stock .geo worked fine for it with transparent  
> background, eliminating the second case.

Yes, it's definitely 0x010101 right now, but that isn't what was in  
my geo file for some reason.

> I can think of only one thing in the map_geo code that might be  
> responsible for different behavior on your system and mine:  please  
> look at the file "magick-config.h" in whatever directory your  
> install of ImageMagick has placed its header files (/sw/include/ 
> magick, maybe?).  See what it says for #define QuantumDepth.  All  
> of my platforms have QuantumDepth 16, and there is special case  
> code in map_geo.c that takes care of QuantumDepth 8 --- I have been  
> unable to test that special case code on any platform, and it could  
> be broken.

It says 16. (Location is: /sw/include/magick)


-- Chip