[boost] Re: value_initialized, deep const, and template instantiation

From: Jim Apple (japple_at_[hidden])
Date: 2004-02-01 12:46:34

Fernando Cacciola wrote:
> A deep-const wrapper with preserving-const will be such that:
 > [snip]
> value_initialized<> is a shallow-preserving-const wrapper:
 > [snip]
> Is this the behaviour you want?

Either one looks fine to me. I am concerned about the present hole, 
however, where const value_initialized<T> a has a user-defined 
conversion operator to T &.

[snip stuff I agree with]

> but can have the member functions, including conversion, without any
> worries about older compilers screaming ambiguity.

So, I was wrong. Compilers with EDG say:
--------------------------------------------------------------
struct A;
//typedef int A;

struct B {
   operator const A & ();
};

void g(const A &);
void g(A &);

int main() {
   g(B());
}
----------------------------------------------------------------
Fails with a claim of ambiguity, but succeeds when line 1 is commented 
out and line 2 is commented in.

So I have to think about that some more.

There's also the though of whether to provide template conversion 
functions - it allows any allowed conversions, but creates ambiguity in
----------------------------------------------------------------------
struct A;

struct bar {
   bar(A);
};

can_init<A> z;

bar x(z);
---------------------------------------------------------------------

Jim