CTFE isn't really an interpreter in the traditional sense. I believe it
just uses the parsed source in its IR form, which is different from
normal byte code interpreters (and also slower). I'm guessing here
though, so I could be wrong.
Whether I'm right or wrong though, I don't see much use for this. As you
said:
- CTFE is limited.
- You'd need nearly the whole D compiler, as well as the standard
library and runtime.
- It would be slower than other scripting languages (unless given lots
of attention).
- D wasn't made to be a scripting language.
The most important thing is simply that CTFE is a function evaluator.
For scripting, that's pretty much useless. You need an entire virtual
machine that you can call into, one that has state and can bind into
your main application.
I think this would be a lot of work for little gain.
