apple/swift#65253 (79e26970) made it so that the async and completion-handler variants of a member would both be matched by the same implementation, but the technique used composes poorly with typechecking work later in this series of commits. Reimplement it so that async variants are filtered out of `ObjCImplementationChecker::unmatchedRequirements` and then async candidates are compared against async alternatives later, during matching.
If function type of some declaration has a at least on type parameter
pack it could mean that number of parameters in "applied" type could
be different from that of "interface" type.
Avoid parsing the syntax tree up-front, and instead
only parse it when required, which happens when either:
1. ASTGen parsing is enabled (currently disabled
by default)
2. Round trip checking is enabled for a primary
file (enabled by default in a debug build,
except when dep scanning or doing an IDE
operation)
3. We need to evaluate a macro in that file
This change therefore means that we now no longer
need to parse the syntax tree for secondary files
by default unless we specifically need to evaluate
a macro in them (e.g if we need to lookup a member
on a decl with an attached macro). And the same
for primaries in release builds.
rdar://109283847
Sema now diagnoses @objcImpl implementations with:
• The wrong settability (i.e. a `let` used for a `readwrite` property)
• The wrong kind (i.e. a method used for a property)
Before this change it was possible to:
1. Call mutating methods on a consume result.
2. assign into a consume (e.x.: var x = ...; (consume x) = value.
From an implementation perspective, this involved just taking the logic I
already used for the CopyExpr and reusing it for ConsumeExpr with some small
tweaks.
rdar://109479440
Some notes:
1. I implemented this as a contextual keyword that can only apply directly to
lvalues. This ensures that we can still call functions called copy, define
variables named copy, etc. I added tests for both the c++ and swift-syntax based
parsers to validate this. So there shouldn't be any source breaks.
2. I did a little bit of type checker work to ensure that we do not treat
copy_expr's result as an lvalue. Otherwise, one could call mutating functions on
it or assign to it, which we do not want since the result of copy_value is
3. As expected, by creating a specific expr, I was able to have much greater
control of the SILGen codegen and thus eliminate extraneous copies and other
weirdness than if we used a function and had to go through SILGenApply.
rdar://101862423
Opened existentials should be erased to the dependent upper bound
if the dependent member can be reduced to a concrete type. This
allows the generic signature to support parameterized protocol types
and bound generic class types by producing a more specific constraint
instead of just a plain protocol or class.
previous checking for the expr being `self`
was checking for the decl being named `self`.
that meant you could do naughty things like:
```
let `self` = Self()
discard `self`
```
rdar://109376381
Make sure we look through a wrapping
`OptionalEvaluationExpr` and its injections when
looking for an assignment to mark a valid if/switch
source expression.
rdar://109305454
Subscripts today don't support any form of ownership specifier
for its parameters Since noncopyable types require such a
specifier, it's not helpful to emit an error suggesting to add
the specifier.
rdar://109233314
A bug was preventing you from writing `discard self`
in a consuming method of a generic noncopyable type.
The main cause was the `isPureMoveOnly` ignored
unbound generic types, claiming none of them are
noncopyable. Then I also needed to pass the contextual
type with the vars bound to the type checker.
rdar://108975216
You can overload a function based on its `async`-ness, and
resolution is carried out based on async-ness of calling
context.
But during protocol conformance checking, for an `async`
requirement we were accidentally choosing the non-`async
overload instead of the `async` one. The `async` one is
the choice people would expect, since the `async` requirement
is in essence the "context" that forwards to the underlying
witness. This intended behavior is also inferred from:
https://github.com/apple/swift/pull/40088
The problem boiled down to a bad check when categorizing the
witness matches prior to ranking them.
Resolves rdar://109135488 / https://github.com/apple/swift/issues/60318
* Parse `#<identifier>` attribute list as a `MacroExpansionDecl`
regardless of the position
* Diagnose whitespaces between `#` and the macro name.
* Correctly attach attributes to `MacroExpansionDecl`
* Fix `OrigDeclAttributes` to handle modifiers (use `getLocation()`
instead of `AtLoc`.)
Type checking is a TODO
rdar://107386648
Don't attempt to wrap `Any` into a single-element tuple to match
against a tuple with pack expansions, this conversion would be
handled by existential promotion if it's allowed, otherwise it
would produce an error.
Resolves: rdar://109160060
