[Experiment] One-way closure parameters

[DougGregor]

Introduce an experimental mode (behind the flag
experimental-one-way-closure-params) that places one-way
constraints between closure parameter types and references to those
parameters within the body of the closure. The intent here is to
break up constraint systems further, potentially improving type
checking performance and making way for larger closure bodies to be
supported.

This is a source-breaking change when the body of a single-expression
closure is used to determine the parameter types. One obvious example
is when there is no contextual type, e.g.,

let _ = { $0 + 1 }

this type-checks today because 1 becomes Int, which matches the
+ overload with the type (Int, Int) -> Int, determining the
parameter type Int for the closure. Such code would not type-check
with one-way constraints.

[DougGregor]

@swift-ci please test source compatibility

[xedin]

@DougGregor This failure is actually want is preventing my PR for BindParam as well because currently it's possible to infer inout from @lvalue used in the body e.g. _ = { $0 = 1 }. I managed to make that one work but there are more complicated examples of the behavior is SwiftPM and in the test suite.

[DougGregor]

@DougGregor This failure is actually want is preventing my PR for BindParam as well because currently it's possible to infer inout from @lvalue used in the body e.g. _ = { $0 = 1 }. I managed to make that one work but there are more complicated examples of the behavior is SwiftPM and in the test suite.

Ah, tricky, tricky. Maybe we can (temporarily?) patch things up in SwiftDriver and SwiftPM to at least get a complete set of results, but it might be that one-way constraints cannot be used for single-expression closures. Or maybe for closures involving $ variables.

[xedin]

Yeah, if I remember correctly there were some compatibility suite failures with that as well.

[DougGregor]

Here's just the experimental flag, which includes the semantics we're going to need elsewhere. The "one way by default" experiment is over at #32222.

[DougGregor]

@swift-ci please smoke test

[DougGregor]

@swift-ci please smoke test

[DougGregor]

@swift-ci smoke test Linux