[gdal-dev] ECW JPEG2000 SDK v5.0 now available

[Even Rouault]

> 
> Even,
> 
> I see the ecw_29 test case your introduced in r26176 (for #4822) now fails.
>  The current results seem to supersample in a different way that GDAL
> normally would - an off by a half issue that think is a reasonable
> interpretation of how to supersample.  Should the test be modified with a
> different case for ECW 5?  Or perhaps the test is just too fragile to be
> worthwhile?

I've fixed ecw_29 and ecw_41.

-- 
Geospatial professional services
http://even.rouault.free.fr/services.html