I'm not sure [class.dtor] p18 can be turned into a note. [class.dtor] p4 doesn't seem to exclude destructors (see also cplusplus/CWG#435).

[class.dtor#4] simply has no relevance here, no?
[class.dtor#18] (normatively) reads as follows:

Once a destructor is invoked for an object, the object's lifetime ends; the behavior is undefined if the destructor is invoked for an object whose lifetime has ended ([basic.life]).

The part before the semicolon is covered by [basic.life#1.4], while the part following it - by [basic.life#6.2] or [basic.life#7.2] with the exception of class-type objects going through a construction/destruction process which is denoted there to be covered separately (the remainder of the reasoning being that destructors consitute non-static member functions). Not having the destructors excluded from the list of what is allowed to call in [class.cdtor#4] constitutes a stand-alone issue: the one CWG2757 addresses (again, the intention of prohibiting the destructor calls is clearly there: the bottom third of the example demonstrates exactly that).

Edit: Nevermind all that: I've got the point. Your objection is fair, @frederick-vs-ja: the part following the semicolon indeed does introduce a new constraint. The problem is that it doesn't state anything about calling a destructor for an object whose lifetime hasn't yet begun, so the referred issue is still useful. And, given a more generic prohibition will be adopted, [class.dtor#18] would again constitute a redundancy in full.