On Thu, Jan 04, 2018 at 04:57:25PM -0700, Jonathan M Davis via Digitalmars-d wrote: > On Thursday, January 04, 2018 14:59:44 H. S. Teoh via Digitalmars-d wrote: [...] > > Are you sure that's non-CTFE? Isn't standard constant folding a > > part of CTFE? Anyway, the new section should clear up the > > confusion. > > As I understand it, CTFE and constant folding aren't really related at > all. Obviously, a value that was generated as part of CTFE could then > take part in constant folding if appropriate, because once the value > is generated, it really isn't any different from if that value were > written by hand. However, CTFE is never done as a compiler > optimization. It's only ever done because a value is required at > compile time. So, constant folding won't even trigger CTFE, just like > the compiler will never decide that it might be more efficient to call > a function at compile time and thus call it at compile time instead of > runtime. [...]
Ah, I see. That clears up some misunderstanding on my part. I'll update the article accordingly. Thanks! T -- Chance favours the prepared mind. -- Louis Pasteur