On 4/7/2015 6:30 PM, Andrei Alexandrescu wrote:
Classical interpretation via tree walking does not use a bytecode. -- Andrei

And this is proven to work by CTFE. (CTFE walks the AST.)

Note that walking the AST, instead of generating bytecode, is not at all why CTFE is slow. CTFE is slow because it allocates memory for every step in the computation.

Reply via email to