[unord.hash] Redundant wording LWG 3512

[JohelEGP]

[unord.hash]p5.4 ends with "shall not throw an exception unless hash<Key> is a program-defined specialization that depends on at least one program-defined type."

"that depends on at least one program-defined type" is redundant given [namespace.std]p2: "Unless explicitly prohibited, a program may add a template specialization for any standard library class template to namespace std provided that (a) the added declaration depends on at least one program-defined type ..."

[jwakely]

I agree those words are redundant, but is that wording even true? If the hash function is hash<optional<T>> then it's not a program-defined specialization, but it can throw if hash<T> can throw.

Rather than fixing the redundancy as an editorial issue, we need to fix the falsehood as an LWG issue.

[JohelEGP]

You're right. Fortunately, the redundant part comes after the falsehood part.

[Dani-Hub]

A corresponding LWG issue exists now: LWG 3512