Actually they are not that related, you may solve a problem with each of
them but this is different.
What makes CTFE awesome is that you use same function for runtime and
compile-time, without a single change.
The answer to your question is just in this very definition i believe.
It is not easy to decide which approach to implement in a library
because both have drawbacks.
And one fills the holes of the other.
Can anybody give informed advice as to which one is preferable in
practice? Or should a library provide both?
I can't guarantee the "informed" part but,
Between your two implementations it is obvious which one, after you answer
the question:
What you want it to be, runtime? compile-time? both?
