Currently inference logic only checked direct equivalence class members
associated with a "work-in-progress" type variable, but each member can
have local equivalences as well that need to be accounted for.
Resolves: rdar://75978086
Add logic to `matchTypes` to insert an implicit conversion to/from
CGFloat type when two nominal types are matched and one of them is
of `CGFloat` type.
Currently `getFunctionArgApplyInfo` expects a locator with `ApplyArgToParam`
element to identify location of the argument. `InOutExpr` could only be used
in argument positions but it doesn't have the same locator format as non-inout
arguments, so `getFunctionArgApplyInfo` needs to do some digging in the AST
to retrieve that information.
Resolves: rdar://75146811
While it is very convenient to default the ExtInfo state when creating
new function types, it also make the intent unclear to those looking to
extend ExtInfo state. For example, did a given call site intend to have
the default ExtInfo state or does it just happen to work? This matters a
lot because function types are regularly unpacked and rebuilt and it's
really easy to accidentally drop ExtInfo state.
By changing the ExtInfo state to an optional, we can track when it is
actually needed.
The existing overloading rules strongly prefer async functions within
async contexts, and synchronous functions in synchronous contexts.
However, when there are other differences in the
signature, particularly parameters of function type that differ in
async vs. synchronous, the overloading rule would force the use of the
synchronous function even in cases where the synchronous function
would be better. An example:
func f(_: (Int) -> Int) { }
func f(_: (Int) async -> Int) async { }
func g(_ x: Int) -> Int { -x }
func h() async {
f(g) // currently selects async f, want to select synchronous f
}
Effect the semantics change by splitting the "sync/async mismatch"
score in the constraint system into an "async in sync mismatch" score
that is mostly disqualifying (because the call will always fail) and a
less-important score for "sync used in an async context", which also
includes conversion from a synchronous function to an asynchronous
one. This way, only synchronous functions are still considered within
a synchronous context, but we get more natural overloading behavior
within an asynchronous context. The end result is intended to be
equivalent to what one would get with reasync:
func f(_: (Int) async -> Int) async { ... }
Addresses rdar://74289867.
Currently bindings where inferred on every `bindTypeVariable` call,
but that's wasteful because not all binds are always correct. To
avoid unnecessary inference traffic let's wait until re-activated
constraints are simplified and notify binding inference about new
fixed type only if all of them are successful.
when it has property wrapper parameters.
The property wrapper type will be replaced with either the wrapped-value
or projected-value type, depending on the argument label/parameter name,
and CSApply will build a thunk to construct the property wrapper and call
the function.
`PotentialBindings` lost most of its responsibilities,
and are no longer comparable. Their main purpose now
is binding and metadata tracking (introduction/retraction).
New `BindingSet` type is something that represents a set
of bindings at the current step of the solver.
Instead of requiring result type of the member to conform to declaring protocol,
as originally proposed by SE-0299, let's instead require `Self` to be bound to
some concrete type in extension that declares a static member.
This would make sure that:
1. Members are only visible on a particular type;
2. There is no ambiguity regarding what base of the member chain is going to be.
Produce a tailored diagnostic when it has been established that
base type of a static member reference on protocol metatype doesn't
conform to a required protocol.
Since it's now possible to refer to static members declared on a protocol
metatype if result type conforms to the protocol we need to adjust failure
detection to identify that conformance failure means and invalid reference
in certain situations.
Rename `openUnboundGenericTypes` to `convertInferableTypes`. In addition to unbound generics, this method also converts placeholder types to fresh type variables.
The isGetterMutating() and isSetterMutating() checks already return
false if the storage is a static property or defined on a reference
type, so the isStatic() and hasReferenceSemantics() checks were not
needed.
SILGen verification crash when no other async functions are called in a function body not marked as async when a for-await-in syntax is used. This corrects the diagnostics to ensure that the type check effects sees the proper async call that the for-await-in syntax infers.
for-await-in syntax was missing a diagnistic hint for inserting a try when the protocol conformance of the sequence shows a potential of throwing. (also there was a superfluous check for the nominal type in that type check (which was removed).
for-await-in syntax and for-try-await-in syntax did not infer async or throwing to closure constraints, both were added to properly identify the asyncy-ness/throwy-ness of the iteration.
Currently the pattern is to collect the type variables and then unique
them. Instead of asking clients to do uniquing, let's just accept a set
as an argument.
Create a new namespace - `swift::constraints::inference` and associate
`PotentialBinding` with it. This way it would be possible for constraint
graph to operate on `PotentialBinding(s)` in the future.
Unresolved member lookup is allowed to perform implicit optional
unwrap of a base type to find members. Previously if there were
any members directly on `Optional`, lookup would stop there. But
since SR-13815 it became possible for solver to attempt members
found on unwrapped type even if there are viable ones on
`Optional` as well.
New score kind has been introduced to guard against possible ambiguities
with new scheme - `SK_UnresolvedMemberViaOptional`. It's used very
time member found via base type unwrap is attempted. Unfortunately,
doing so can lead to behavior changes in existing code because it's
possible that base was wrapped into optional implicitly based on
context e.g. unresolved member passed in as an argument to a parameter
of optional type.
To fix situations like that, `SK_UnresolvedMemberViaOptional` should
only be increased if there is a mix of members to attempt - both directly
on `Optional` and on unwrapped type, in all other cases score should stay
the same because there could be no ambiguity.
Resolves: rdar://73027153
