"Peder Holt" <peder.holt@gmail.com> wrote
To be precise, m*N/2
To deduce the type of an expression int* a[10], sizeof(foo<n>(a)), n=1..N the compiler needs to generate the following types (names may vary) encode_array<T[10],1>
encode_array<T[10],2> encode_pointer<T*,1>
encode_array<T[10],3> encode_pointer<T*,2> encode_impl<int,1>
encode_array<T[10],4> encode_pointer<T*,3> encode_impl<int,2> encode_dummy ... -> N.
that is, N*m/2 types needs to be generated by the compiler. Please correct me if I'm wrong.
In my implementation there is no encode_array<T[10], 1>, encode_array<T[10],2>, etc. There is: template<class V, class T, int Size> struct encode_type_impl<V, const T[Size]> Which is the same type throughout BOOST_TYPEOF invocation, and therefore gets instantiated only once. So I insist that the order of my algorithm is O(N), which results from the need to instantiate "foo<n>(expr)" for every n from 0 to N-1. If you came up with an O(M) rather than O(N) algorithm, where M is the real size of the encoding sequence, this is really great for simple types. I'll do my best to understand your solution and adapt it to my implementation (if possible). For now, it really looks more like a miracle to me...
I think i am closing in on the ultimate typeof implementation.
The "ultimate" implementation will come from the compiler vendors :) Regards, Arkadiy