Stabilize core::convert::identity

[Centril]

r? @SimonSapin

fixes #53500

This is waiting for FCP to complete but in the interim it would be good to review.

[SimonSapin]

r+, pending FCP completion

[clarfonthey]

I feel like this particularly would be better as flatten with an explicit specialisation for Option, rather than filter_map.

[Centril]

I disagree.

First, I think that .filter_map(identity) is clearer wrt. intent (and also semantics). I can clearly tell that filtering is happening because it says so in the name. I know that .filter_map takes f: T -> Option<U> and with identity I fix T = Option<U> and thus I get back all the Somes. The .flatten() method doesn't tell me any of this, in particular you cannot tell without knowing that Option is being operated on what the semantics are. Thus the reasoning footprint is greater with .flatten().

Second, even tho .flatten() is more general than join :: Monad m => m (m a) -> m a I think the primary use of the method should be for monadic join and using it for other purposes will break some people's minds (e.g. mine).

Third, I think that we shouldn't hold up stabilization on this issue; we can resolve it in a subsequent PR or issue if you feel strongly.

[Centril]

FCP has completed, and therefore:

@bors r=SimonSapin p=1

[bors]

📌 Commit e75dab7 has been approved by SimonSapin

[bors]

⌛ Testing commit e75dab7 with merge 1a3a3df...

[bors]

☀️ Test successful - checks-travis, status-appveyor
Approved by: SimonSapin
Pushing 1a3a3df to master...