René Meusel

Thanks for the detailed gist. Your code looks good to me. Personally, I also felt the pain of the inexplicable "bad_certificate" error when trying to use the ASIO wrapper in...

Re: ninja. I'm using it during development on Linux and macOS, and occasionally on Windows. Some minor things were fixed here and there since Botan 3.2 and, I guess, we...

I really like the idea! Perhaps an obvious question: did you consider modeling this after the builder abstractions I had cooked up in the drafts #4694 and #4318?

> I don't think the extra complexity of it makes sense here, where there is a single consumer of the builder output. Frankly, that's understandable. I'm not super happy about...

That's just my general two cents. If we start establishing builders throughout the API (which really makes a lot of sense in many places, IMHO), I would love to converge...

> Yeah what you have is how I initially wrote that code and of course was disappointed. Time for C++23 then? 😜 > I guess the type could implement a...

What actually bugs me quite a bit is the final instantiation that takes a number of positional arguments in `into_self_signed_cert()`: ```C++ Botan::CertificateParametersBuilder root_cert_params; auto cert = root_cert_params .add_common_name("Benchmark Root") .add_country("DE")...

> Pimpl with `shared_ptr` works great for immutable objects (eg `X509_Certificate`) but I would not consider it for anything that can be modified. Fully agree. That's the point I was...

Sounds good to me. We have a similar integration with CMake, where we install a Botan-config.cmake to be consumed by `find_package()`. Improving the integration with the python ecosystem would certainly...

The root cause is that we don't know the platform-default when registering this option. Currently `None` is passed as a placeholder which will be rectified at a later stage (in...