Reimar Döffinger

I have no idea how it is supposed to work, not even how the allocator this pointer connects to the storage. I guess the part that should be aligned is...

Sorry, I missed the information at the start that this is only an issue when allocated on stack. The issue should in this case trigger on older Linux 32-bit systems...

There are 3 main stumbling blocks: - it seems there is no API designed for how to set filter options in a generic way. While it's possible to just add...

> Yeah, one might want to use the auto-detection capability of the library My understanding is that the library does not have this capability, only the binary. But if there...

I had a bit of a look. The good news: it's easy to add support for extra parameters via the zstd_params struct. The bad news: the library functions to set...

I think it's hard-coded because much of the C compression API uses "char" so making it generic would cause lots of other issues. Also this isn't just bzip2, isn't it?...

Note: I am not speaking with any particular authority and certainly not for the boost project. Maybe I misunderstand, but I think it returns exactly the right thing: The bzip2...

The polymorphism is to allow replacing the allocator, e.g. to allocate from a fixed memory pool on some embedded device. I.e. HOW the memory is allocated, not which TYPE. And...

But that is exactly the purpose of the non_blocking_adapter? When a nonblocking sink returns 0 it should mean to try again later, and that's what it does. In case if...

It's a bit hard to say without knowing why the sink device would become "broken" and how it is designed to report that. It could throw an exception when it...