ben-craig

Seems reasonable. I vaguely recall running into some cases where there were spaces between `operator` and the op type. Put together a PR and we can look at it!

I would argue that all times are unreliable. CPU or user time can still be unreliable just based off of what the rest of the system is doing. If you...

I discuss something along these lines here: https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p2268r0.html#constexpr_to_consteval I think the idea can work, but there's a need for someone to test it, and to describe the behavior differences between...

First, it's fine for an implementation to implement the integer charconv code in headers if they so choose. The floating point charconv code is the crazy expensive stuff. Second, "freestanding"...

I agree with your reading @frederick-vs-ja . It is not intended to exclude the C `` headers. I'll file an LWG issue on it.

https://cplusplus.github.io/LWG/issue4049

Hi alecazam. I authored the template set reporting. The std::unordered_map entries are for instantiating the class, and not any of the methods. You can see how std::function::function also shows up,...

`new` and `delete` during constant evaluation works fine in freestanding. Those are only a problem if they leak into runtime. There's discussion of keeping constexpr new and delete working in...

If std::allocator isn't an option, then what do you use in hosted environments during constant evaluation? Another potential alternative for you is to use the new[] operator to allocate union...