Use the accessor names as written in Objective-C.

[varungandhi-apple]

Not 100% sure if my approach taken here is correct. It is unclear to me what the semantics of getParsedAccessor are supposed to be in the presence of user-written accessor names in Objective-C (as opposed to those written in Swift). Should they be passed through? Or not passed through because they are considered implicit? I've assumed that the existing behavior of getParsedAccessor is correct, so we use getAccessor instead.

Fixes rdar://problem/55519276.

Related: commit 0c5d52d (PR #26461) where getParsedAccessor was introduced.

[varungandhi-apple]

@swift-ci please test

[swift-ci]

Build failed
Swift Test Linux Platform
Git Sha - 3cfb05da280afbcfb7626cf6d3c3a381554f4b53

[varungandhi-apple]

@swift-ci please test

[swift-ci]

Build failed
Swift Test OS X Platform
Git Sha - 3cfb05da280afbcfb7626cf6d3c3a381554f4b53

[swift-ci]

Build failed
Swift Test Linux Platform
Git Sha - 3cfb05da280afbcfb7626cf6d3c3a381554f4b53

[rjmccall]

IIRC, Slava was aiming to eliminate uses of the raw getAccessor method. I think getParsedAccessor is meant to be solely about what was written in source, so it's reasonable to say that we shouldn't be using it in this semantic method, but it's not clear to me that there's currently a better method, since we (apparently) can't assert as a precondition that this is an @objc property. Slava would have thoughts.

[slavapestov]

I think the right fix is to not mark imported accessors that correspond to objc methods as implicit. Only bona fide synthesized accessors should be implicit (eg, those for stored properties of structs, or a modify accessor synthesized when a property witnesses a mutable protocol requirement, etc).

And John is right, we should not be introducing new usages of getAccessor(), and existing usages should mostly be removed (except for a few places such as the AST verifier where we don't want to trigger synthesis).

[varungandhi-apple]

@swift-ci please test Linux platform

[varungandhi-apple]

@swift-ci please test

[varungandhi-apple]

Note: I've added a new commit as I think @edymtt cherry-picked the first two commits elsewhere, so I didn't want to mess with history by squashing the third commit on top of the second one, just in case it causes issues.

[swift-ci]

Build failed
Swift Test Linux Platform
Git Sha - 3b9dbd5

[swift-ci]

Build failed
Swift Test OS X Platform
Git Sha - 3b9dbd5

[varungandhi-apple]

Note for anyone else reading: the imported accessor is explicit in the existing code. Which is correct behavior. Hence, I didn't make any changes there.