[gdal-dev] Amplitude virtual bands for complex datasets ?

[Even Rouault]

Le mardi 05 juillet 2016 09:34:31, Bugbuster a écrit :
> Dear Even,
> 
> Sorry for my late answer, I was onto something different yesterday.
> 
> I agree with you, and I did change all function declarations in order to
> "export" them in the DLL.
> Thus, all functions in both pixfunplugin.c and pixelfunctions.c are now
> declared as "C" functions with the CPL_DLL export tag.
> I checked with Dependency Walker that all functions are indeed exported
> correctly.
> What is really weird is that *all I/F variables* in RealPixFunc() function
> (for example) seem to be correct *but the 1sr one papoSources*.

For those who followed this thread, the issue has been reported and fixed in
https://trac.osgeo.org/gdal/ticket/6599

It turned out not being a Windows specific issue, but a general one that could 
affect any VRT derived band made of a single SimpleSource, when using dataset 
level raster I/O such as with gdal_translate.

Even

-- 
Spatialys - Geospatial professional services
http://www.spatialys.com