MichaelSB

Is this a scenario (clang, libc++ on linux) crow should support? I could just "implement" the fix mentioned above. Or dig deeper and try to find out more about the...

I was busy with other things myself, so I had close to no time to dig deeper. For me this looks like: wvalue contains std::unique_ptrs. The code in question does...

I usually add crow as git submodule and add it with add_subdirectory(...) to my project. Like ``` cmake_minimum_required(VERSION 3.10) cmake_policy(VERSION 3.10) set(CMAKE_CXX_STANDARD 20) set(CMAKE_CXX_STANDARD_REQUIRED ON) set(CMAKE_CXX_EXTENSIONS OFF) project(example VERSION 1.0.0)...

I do basically the same. In my app I use a collection of std::unique_ptrs holding my allocated user_data. This way my instances are destroyed when the application shuts down without...

I agree on the shared_ptr. For me it would be worse. Also, when using it with some C-Api a shared_ptr would most likely be hard to implement.

I see... still I'm not convinced. shared_ptr gives the notion of something that is shared. In many cases nothing is shared. My userdata is unique, that's why I use unique_ptrs...

That's useful I think. What's the purpose of the string argument of the timeout handler? It's hardcoded to the string "timeout"?

Is there a reason for new char etc and not using a std::string or std::unique_ptr for the data (raii)? Also, I'm not fully convinced sending ~6MB of data in a...

Thanks. I'll see, if I can reproduce this issue in a minimal example. If so, I can try to fix it. You're using master or one of the releases?

Out of curiosity... which headers are missing? Just adding headers might not be enough. Most headers negotiate features, which probably are not implemented yet.