Marc Boucek

AFAIK these utf8 conversion problems are all fixed already. We added some extensive tests to ensure files with unicode characters can be loaded. The folder `ni-media/audiostream/test/test_files/user_files` contains all sorts of...

Yes I agree, this is a hidden requirement and totally not obvious for clients. I would treat this issue separately from the C++ 17 std::filesystem support though. I would suggest...

Sorry for the late reply. Are you still facing this issue? We have a CI pipeline setup for MSVC2019. I just checked the logs and found the last successful build...

Hey Neil Thanks for the feedback! We did some initial benchmarking to verify the performance of boost::iostreams. We also benchmarked the pcm iterator performance ( against raw loops ) while...

We do generally support reading alac encoded mp4 files on windows. We have several alac files that get tested on each commit in the windows CI: ie: https://github.com/NativeInstruments/ni-media/blob/master/audiostream/test/test_files/reference_files/sin440.1.44100.44100.s16le.alac.m4a Could you...

> We might be able to work around this by patching boost, but it would be preferable to fix it in ni-media in a way that no redefinition of BOOST_IOSTREAMS_BASIC_STREAMBUF...

> Yes, the layout of the class itself doesn't change, but some of its associated RTTI data does, specifically the "RTTI Base Class Array". Which makes sense, because you are...

> But also, I don't think this would help, because unless I'm mistaken the problem not only exists with ifstream.cpp, but with any translation unit that includes an ni-media header....

I don't think there is an ODR violation in ni-media. By redefining BOOST_IOSTREAMS_BASIC_STREAMBUF we inject our own streambuf class. While ugly I believe this is perfectly legal as the class...

You are right. Seeing no ODR violation in ni-media itself, I jumped to conclusions too quickly thinking that it must be ni-media built with /LTCG that is leaking that (supposedly)...