3 Dec
2006
3 Dec
'06
4:23 a.m.
On 12/2/06, Steven Watanabe <steven@providere-consulting.com> wrote:
Well, even whether it works is debatable.
template<class X> void do_stuff(X*, X*); template<class X, class Y, class Z, class W> void f(vec4<X, Y, Z, W>&, vec4<X, Y, Z, W>&) { do_stuff(&arg1, &arg2); }
Very good point. Did you have an alternate solution? It could possibly just be documented that it didn't work, or the address-of operator could static assert when the types are references. I don't have a strong opinion on this, in fact I hadn't even thought of it until Michael Marcin brought it up. --Michael Fawcett