On Thu, Mar 28, 2024 at 8:18 PM Robert Ramey via Boost < boost@lists.boost.org> wrote:
Hi Robert,
I’m sorry you are feeling that Boost.Serialization might be EOL. I actually do use it and have found its stability to be a great selling point over the years. I confess that I’m not a retro-compiler person so I
On 3/28/24 2:23 PM, Brook Milligan via Boost wrote: personally don’t use any old compat code, but particularly in this case I expect many others might. Of course, it is your decision I suppose, but it would be a great loss were the library to be deprecated.
Thanks for your steady work on it.
You welcome.
I don't think it's so much the serialization library. I'm guessing that a lot libraries have this concerns. But many older libraries are not maintained so no one raises the issue. Or maybe they are so well written that they never need maintenance or evolution. Or maybe no one uses them any more. Without some real data, there's no way to tell.
My take is that it's time for serialization 2.0 built on c++26 reflection. This is the perfect domain to test the reflection facilities being proposed for c++26. There's a fork of clang that supports it so it's technically possible. Jeff