[expr.unary.op]/3: the value of a pointer should be "pointer to", not "address of"

[languagelawyer]

http://eel.is/c++draft/expr.unary.op#3.sentence-4:

Otherwise, if the type of the expression is T, the result has type “pointer to T” and is a prvalue that is the address of the designated object ([intro.memory]) or a pointer to the designated function.

1. «The result of a prvalue is the value that the expression stores into its context.» ([basic.lval]). The lvalue can be of type T, but denote an object of type U, so the type of the result of applying the address-of operator to it will be pointer to U, not pointer to T (and I'm unsure if we even can talk about the type of the result. The result is "pointer to object", the type of the prvalue is pointer to T, but I don't think this makes sense to say that the type of the result is pointer to U). pointer to T is the type of the prvalue, not its result.
2. the prvalue has the value "pointer to the designated object or function", not "the address of the designated object or a pointer to the designated function."

The note following the sentence above (http://eel.is/c++draft/expr.unary.op#3.note-1) also needs a fix:

In particular, the address of an object of type “cv T” is “pointer to cv T”

I don't think that addresses have types.

[jensmaurer]

I don't follow item 1, but I think item 2 and the note could use some embellishment to properly fit into the pointer value framework of [basic.compound].

Suggestion: "Otherwise, if the type of the expression is T, the result is a prvalue of type "pointer to T" whose value is a pointer to the designated object [intro.memory] or function. [ Note: In particular, taking the address of a variable of type "cv T" yields a pointer of type "pointer to cv T". ]"

[jensmaurer]

Editorial meeting: Point 1 is wrong; for the rest, my suggestion is good.