update toml++ to v2.4.0

[marzer]

This PR updates toml++ to v2.4.0, as well as a few other minor bits:

• removed dependency on some of toml++'s internal macros
• updated benchmark script to show relative speedup/slowdown factor
• updated README to reflect TOML v1.0.0 status and added new benchmark data
• updated tl::optional
• made use of tl::optional over std::optional switched based on a configurable #define
• added .clang-format

RE std::optional: tl::optional is a great fallback but in general for this library's purposes you should probably prefer std::optional wherever it's available since it targets C++17 at a minimum and vendors are likely to have extensively tested their standard libraries on their specific platforms. Currently it uses tloptional for __APPLE__ platforms and std::optional everywhere else, per #14.

RE .clang-format: I generated it using the defaults (--style=llvm), only changing SortIncludes to false (header order shouldn't matter, but it often does, and letting a tool arbitrarily re-order headers can lead to very weird surprises).

I also removed 'unofficial' from the readme since I think this library is as close to offical a python toml++ wrapper as there's ever likely to be :)

[bobfang1992]

Thanks this looks great. I had a look at the changes I and they LGTM. I will merge it now.