Concerning the interest of the subject, since WL is an interpreted language, the compilation bootstrapping (wikipedia Compiler bootstraping) is not necessarily an applicable paradigm (although it will eventually happen in the near future). Nevertheless, considering the interpretation as a compilation substitute, I guess that this is somehow related... isn't it?
It is my understanding that the WL runtime library is being revamped mostly with C++ (or code of the same family), I guess with the purpose of having the fastest implementation possible. Nevertheless, I don't think it would be completely out of the question to have a runtime written on WL and compiled with the WL (compiled) compilation library... Most likely not as fast, as the compiler automatic optimization is eventually not capable of compensating for some astute human brain speed optimization strategies, but it could probably facilitate the port of the runtime to other platforms? (well, I guess that the programming language of the runtime is already LLVM compilation "compatible", facilitating the port to other platforms...). Would love to have a micro^2 kernel (one micro for the subset, another micro for the device aspect) running on a ESP32, both running binary and the interpreter (similar to micropython); but since there's still no LLVM translation to Xtensa... (in a few years...)
