When trying to figure out errors from an import failure, the nullability
completeness warnings would clutter the output making it difficult to
identify the errors. Sprinkle the declaarations with
`_Null_unspecified` to maintain the current nullability semantics and
silence the warnings. NFC.
ground work for the syntactic bridging peephole.
- Pass source and dest formal types to the bridging routines in addition
to the dest lowered type. The dest lowered type is still necessary
in order to handle non-standard abstraction patterns for the dest type.
- Change bridging abstraction patterns to store bridged formal types
instead of the formal type.
- Improve how SIL type lowering deals with import-as-member patterns.
- Fix some AST bugs where inadequate information was being stored in
various expressions.
- Introduce the idea of a converting SGFContext and use it to regularize
the existing id-as-Any conversion peephole.
- Improve various places in SILGen to emit directly into contexts.
Bitcast the AnyObject result to AnyObject?, then call our new helper function, so that we can handle nils without choking. Fixes rdar://problem/27874026.
This is support for SE-0069: Mutability and Foundation Value Types.
In cases where someone has overridden a method that takes, e.g.
'NSURL', the override will no longer be valid, because the system
class will now use the value type 'URL' instead. If an override's
full name matches exactly, the compiler will offer fix-its for any
uses of reference types where value types are now preferred.
(This must be a direct match; subclasses, including the mutable
variants of many Foundation types, will need to be updated by hand.)
One wrinkle here is the use of generics. In Swift 2, Objective-C
generics weren't imported at all, so it's unlikely that the overriding
method will have the correct generic arguments. Simple migration
might insert the "bound" type, but it can't know what specific type
might be more appropriate. Therefore, the logic to add the fix-it
ignores generic arguments, assuming the parent's type is correct.
rdar://problem/26183575
This lets us eliminate the _getObjectiveCType() value witness, which
was working around the lack of proper type witness metadata in witness
tables. Boilerplate -= 1.
Use the swift_bridge Clang attribute and paired _ObjectiveCBridgeable
conformance to provide bridging for a an Objective-C class to a Swift
value type. All of these code paths have been exercised already by the
generalizations for String/Array/Dictionary/Set, so there are no
compiler changes required here, but this verifies that we haven't
missed any ad hoc rules for those bridged types along the critical
path.
Checkpoint that we've implemented the bulk of rdar://problem/24659951.
Eliminate the String/Array/Dictionary/Set special cases, using
_ObjectiveCBridgeable conformances instead.
Note: Clang update required for the test to success.
