mmp
commented
7 years ago The intermediate value needs to be explicitly copied into a Float array; the ptex code always returns float values, but if Float is double, then we need to copy into a Float array to convert to doubles so that the cast is valid. Just pushed a fix to do just that.
Thanks for reporting this!
I used CMake to create a pbrt Visual Studio 2015 solution, with CMAKE_BUILD_TYPE "Release" and the solution build settings as Release - x64. From the Users Guide in the pbrt site, I uncommented "#define PBRT_FLOAT_AS_DOUBLE" in src/core/pbrt.h.
When I tried to build the solution, I got this error in the pbrt project, in the ptex.cpp file:
C2664: 'pbrt::RGBSpectrum pbrt::RGBSpectrum::FromRGB(const pbrt::Float [], pbrt::SpectrumType)': cannot convert argument 1 from 'float *' to 'const pbrt::Float[]'
The line is
return Spectrum::FromRGB(result);where result is a 'float' parameter in a function called "fromResult". Seems like defining PBRT_FLOAT_AS_DOUBLE makes 'Float' a double, and an implicit convesion from float to double* isn't allowed. So I made an explicit cast like below, and the build worked:
return Spectrum::FromRGB((const pbrt::Float*)result);Would this be considered a bug and a working bug fix? I've only built pbrt to try out a render, haven't seriously studied it yet. The ptex.cpp file looks relatively new, seems related to the March 5 texture cache feature you mention on your site. Don't see that file in my first pbrt build I made back in January.