On Tuesday, 27 December 2011 at 19:31:58 UTC, Joshua Reusch wrote: > more general: is there any reason to not evaluate a pure > function with const arguments always at compile-time ? Halting problem?