Follow-up to https://github.com/apple/swift/pull/65048
`getDesugaredType` unwraps sugar types that appear in sequence,
to remove sugar from nested positions we need to get a canonical type.
Thanks to @slavapestov for pointing it out.
Generic type aliases, unless desugared, could bring unrelated type variables
into the scope i.e. `TypeAlias<$T, $U>.Context` is actually `_Context<$U>`.
These variables could be inferrable only after the the body the closure is
solved. To avoid that, let's adjust `TypeVariableRefFinder` to desugar types
before collecting referenced type variables.
Resolves: rdar://107835060
Previously, when creating availability attributes for synthesized declarations
we would identify some set of reference declarations that the synthesized
declaration should be as-available-as. The availability attributes of the
reference declarations would then be merged together to create the attributes
for the synthesized declaration. The problem with this approach is that the
reference declarations themselves may implicitly inherit availability from their
enclosing scopes, so the resulting merged attributes could be incomplete. The
fix is to walk though the enclosing scopes of the reference declarations,
merging the availability attributes found at each level.
Additionally, the merging algorithm that produces inferred availability
attributes failed to deal with conflicts between platform specific and platform
agnostic availability. Now the merging algorithm notices when platform agnostic
availability should take precedence and avoids generating conflicting platform
specific attributes.
Resolves rdar://106575142
Current logic attempts simplification of the base type only if it's
a type variable or a dependent member type. That's valid for correct
code but not for invalid code e.g. `(($T) -> Void).Element` is not
going to be simplified even if `$T` is bound, which causes crashes
in diagnostic mode because `matchTypes` is going to re-introduce
constraint with partially resolved dependent member types and by
doing so make it stale forever which trips constraint system state
verification logic.
Resolves: rdar://105149979
Introduce SingleValueStmtExpr, which allows the
embedding of a statement in an expression context.
This then allows us to parse and type-check `if`
and `switch` statements as expressions, gated
behind the `IfSwitchExpression` experimental
feature for now. In the future,
SingleValueStmtExpr could also be used for e.g
`do` expressions.
For now, only single expression branches are
supported for producing a value from an
`if`/`switch` expression, and each branch is
type-checked independently. A multi-statement
branch may only appear if it ends with a `throw`,
and it may not `break`, `continue`, or `return`.
The placement of `if`/`switch` expressions is also
currently limited by a syntactic use diagnostic.
Currently they're only allowed in bindings,
assignments, throws, and returns. But this could
be lifted in the future if desired.
It's ok to drop the global-actor qualifier `@G` from a function's type if:
- the cast is happening in a context isolated to global-actor `G`
- the function value will not be `@Sendable`
- the function value is not `async`
It's primarily safe to drop the attribute because we're already in the
same isolation domain. So it's OK to simply drop the global-actor
if we prevent the value from later leaving that isolation domain.
This means we no longer need to warn about code like this:
```
@MainActor func doIt(_ x: [Int], _ f: @MainActor (Int) -> ()) {
x.forEach(f)
// warning: converting function value of type '@MainActor (Int) -> ()' to '(Int) throws -> Void' loses global actor 'MainActor'
}
```
NOTE: this implementation is a bit gross in that the constraint solver
might emit false warnings about casts it introduced that are actually
safe. This is mainly because closure isolation is only fully determined
after constraint solving. See the FIXME's for more details.
resolves rdar://94462333
In some situations `getContextualType` for a contextual type
locator is going to return then empty type. This happens because
e.g. optional-some patterns and patterns with incorrect type don't
have a contextual type for initialization expression but use
a conversion with contextual locator nevertheless to indicate
the purpose. This doesn't affect non-ambiguity diagnostics
because mismatches carry both `from` and `to` types.
Resolves: rdar://problem/103739206
All of the type variables referenced by a type had to be
handled by `inferVariables` otherwise it would to possible
to bring non-representative variable into scope which in
turn later would get simplied into a variable that doesn't
exist in the active scope and solver would crash.
Resolves: rdar://83418797
SE-0110 allows tuple slat between function types. The solver needs to
account for that when trying to infer parameter types from context
while resolving a closure in order to propagate types and speed up
type-checking.
Resolves: https://github.com/apple/swift/issues/62390
Inject `buildBlock` call into `return`, `buildFinalResult`, and
`build{Optional, Either}` to take advantage of contextual information
and maintain compatibility with current inference behavior.
Invalid syntax could introduce type variables into sequence which
have to be brought into scope otherwise solver is going to crash.
Resolves: rdar://100753270
If a (partial) solution has a type variable it could only be unbound
if `FreeTypeVariableBinding` is set to `Allow`, in all other cases
solution would either have a fully resolved type or a hole.
`applySolution` shouldn't second guess `finalize()` and just apply
a solution as given. This is very important for multi-statement closures
because their elements are solved in isolation and opaque value types
inferred for their result could contain not-yet-resolved type variables
from outer context in their substitution maps (which it totally legal
under bi-directional inference).
Update `applyPropertyWrapperToParameter` to set types to projected
and wrapped values and allow `TypeVariableRefFinder` to skip decls
with implicit property wrappers that are not yet resolved.
Reference to `makeIterator` in for-in loop context needs to be
treated as a reference to a witness of `Sequence#makeIterator`
requirement, otherwise it could lead to problems with retroactive
conformances.
Using `computeWrappedValueType` is incorrect because
that return a type of the wrapped variable and not
the *wrapper* variable (one that starts with `_`).
Resolves: https://github.com/apple/swift/issues/61017
If the failure is not reflected in constraint system it would let
the solver to form a _valid_ solution as if the constraint between
the type variable and the unresolved dependent member type never
existed.
Resolves: https://github.com/apple/swift/issues/60649
`ExprRewriter::coerceToType` should canonicalize contextual type before
attempting to use it for value abstraction, because sugared type could
have typealias references that hide their underlying opaque result types.
Resolves: rdar://98577451
To support referencing projected and/or wrapped var
in a closure solver needs to lookup a type of their
originator and based on the wrapper type compute and
assign types to projection and/or wrapper.
Resolves: https://github.com/apple/swift/issues/59295
Resolves: rdar://94506352
To make this test work, fix an issue in `ConstraintSystem::salvage` where a
threshold breach during solving went unnoticed due to exiting on ambiguity
before reaching the `isTooComplex` check. Address this by moving the
`isTooComplex` check to before we start processing solutions, and stick another
one in `findBestSolution` for short-circuiting while we're here.
Property wrappers trigger initializer synthesis. Synthesized
initializers should not be re-typechecked when encountered e.g.
while re-solving closure with a different contextual type.
Resolves: https://github.com/apple/swift/issues/59294
Resolves: rdar://94506352
`RelabelArguments` cannot possibly diagnose this issue because there
are no argument lists in this case. Let's report contextual mismatch
instead.
Resolves: https://github.com/apple/swift/issues/59058
lit.py currently allows any substring of `target_triple` to be used as a
feature in REQUIRES/UNSUPPORTED/XFAIL. This results in various forms of
the OS spread across the tests and is also somewhat confusing since they
aren't actually listed in the available features.
Modify all OS-related features to use the `OS=` version that Swift adds
instead. We can later remove `config.target_triple` so that these don't
the non-OS versions don't work in the first place.
