In the general case, this is done by reverse engineering the "best"
places for requirements to go from the requirement signature.
Conformance/superclass requirements like Self: Foo and Self.T: Bar defer
to the inheritance clause if they appear there, or are attached to the
protocol where clause or T (respectively) if not. A conformance
requirement like Self.T.U: Baz will go on T (if T is declared in the
protocol being printed).
Same-type requirements always go in where clauses, and specifically a
where clause of an associated type that is mentioned in them, so
something simple like Self.T.U == Int goes on the T associated type
definition, and similarly Self.T.U == Self.V will go on V (it's kinda
nonsense, but also more directly connected to V). There's a left-bias
for cases without an "obvious" choice, meaning something more
complicated like Self.T.U == Foo<Self.V> will end up on T.
Requirements that don't fit elsewhere will go on the
protocol (e.g. Self.AssocTypeFromSuperProtocol == Int).
Use this in ProtocolDecl::requiresClassSlow(), and hope its presence
discourages more potentially-infinitely-recursive walks of the
inherited-protocols lists.
Previously some decls (TypeAliasDecl and ExtensionDecl) had bits
explicitly marking whether they've been validated, while other decls
just deduced this from hasInterfaceType. The doing the latter doesn't
work when the interface type can be computed before doing full
validation (such as protocols and associatedtypes, which have trivial
interface types), and so an explicit bit is adopted for all decls.
The list of directly inherited protocols of a ProtocolDecl is already
encoded in the requirement signature, as conformance constraints where
the subject is Self. Gather the list from there rather than separately
computing/storing the list of "inherited protocols".
The root cause is that NormalProtocolConformance::forEachValueWitness()
needs to skip protocol members that are not requirements.
Otherwise we end up passing such a non-requirement member down to
NormalProtocolConformance::getWitness() and hit an assert when we
cannot find it.
It looks like this code path was only ever hit from SourceKit.
The fix moves TypeChecker::isRequirement() to a method on ValueDecl,
and calls it in the right places.
Fixes <https://bugs.swift.org/browse/SR-3815>.
This is in preparation for generic subscripts, which will also
expose methods like getGenericSignature(), and so on.
ExtensionDecl, GenericTypeDecl and AbstractFunctionDecl all share
code. Instead of copy and pasting it yet again into SubscriptDecl,
factor it out into a common base class.
There are more yaks to shave here, but this is a step in the right
direction.
This function was returning an ArrayRef pointing into a data structure
that is easily mutated via code walking over that ArrayRef, which
could cause spooky side effects, particularly during
deserialization. Perform a defensive copy to eliminate such side
effects.
Storing this separately is unnecessary since we already
serialize the enum element's interface type. Also, this
eliminates one of the few remaining cases where we serialize
archetypes during AST serialization.
This is a generic signature that stores exactly the requirements that a
protocol decl introduces, not letting them be implied by the Self :
Protocol requirement, nor storing any requirements introduced by the
protocols requirements.
Specifically, suppose we have
protocol Foo {}
protocol Bar {}
protocol Baz {
associatedtype X : Foo
}
protocol Qux: Baz {
associatedtype X : Bar
}
The normal generic signature and (canonical) protocol requirement
signature of `Baz` will be, respectively
<Self where Self : Baz>
<Self where Self : Baz, Self.X : Foo>
And for `Qux`, they will be:
<Self where Self : Qux>
<Self where Self : Qux, Self : Baz, Self.X : Bar>
Note that the `Self.X : Foo` requirement is not listed.
For the moment, this is unused except for `-debug-generic-signatures`.
This is a temporary stop-gap for getting round-trip parsing
off the ground. The real fix, not re-injecting declarations
into an if-config's declaration context, is a deep dive and
is still ongoing.
Parsing declaration list (e.g. member list of nominal decl) is very
different from comma separated list, because it's elements are separated with
new-line or semi-colon. There's no good reason to consolidate them.
Also, declaration list in 'extension' or inside of decl '#if' didn't
emit diagnostics for consecutive declarations on a line.
class C {
#if true
var value: Int = 42 func foo() {}
#endif
}
extension C {
func bar() {} subscript(i: Int) -> Int {
return 24
}
}
This change consolidates declaration list parsing for
members of nominal decls, extensions, and inside of '#if'.
In addition, removed unnecessary property 'TrailingSemiLoc' from decl.
This commit introduces new kind of requirements: layout requirements.
This kind of requirements allows to expose that a type should satisfy certain layout properties, e.g. it should be a trivial type, have a given size and alignment, etc.
This reverts the contents of #5778 and replaces it with a far simpler
implementation of condition resolution along with canImport. When
combined with the optimizations in #6279 we get the best of both worlds
with a performance win and a simpler implementation.
- Most immediately, we now have `withoutActuallyEscaping` as a supported way to temporarily reference a nonescaping closure as if it were escapable, and we plan to break the ABI for escaping and nonescaping closures so that the old `unsafeBitCast` workaround no longer works.
- `unsafeBitCast` is also commonly used to kludge pointers into different types, but we have more semantically meaningful APIs for type punning now. Guide users towards those APIs.
- Suggest more specific and type-safe operations, like `bitPattern:` initializers or `unsafeDowncast`, for the situations where `unsafeBitCast` is being used to avoid dynamic type checks or reinterpret numerical bits.
* Pack the bits for IfConfigDecls into Decl
* Don't open symbols into a module when evaluating canImport statements
The module loaders now have API to check whether a given module can be
imported without importing the referenced module. This provides a
significant speed boost to condition resolution and no longer
introduces symbols from the referenced module into the current context
without the user explicitly requesting it.
The definition of ‘canImport’ does not necessarily mean that a full
import without error is possible, merely that the path to the import is
visible to the compiler and the module is loadable in some form or
another.
Note that this means this check is insufficient to guarantee that you
are on one platform or another. For those kinds of checks, use
‘os(OSNAME)’.
Protocol methods returning optionals, metatypes and functions
returning 'Self' are now handled correctly in Sema when
accessed with a base of existential type, eg:
protocol Clonable {
func maybeClone() -> Self?
func cloneMetatype() -> Self.Type
func getClonerFunc() -> () -> Self
}
let c: Clonable = ...
let _: Clonable = c.maybeClone()
let _: Clonable.Type = c.cloneMetatype()
Previously, we were unable to model this properly, for
several reasons:
- When opening the function type, we replaced the return
value in openedFullType _unconditionally_ with the
existential type. This was broken if the return type
was not the 'Self' parameter, but rather an optional,
metatype or function type involving the 'Self' parameter.
- It was also broken because we lost information; if the
return type originally contained an existential, we had
no way to draw the distinction. The 'hasArchetypeSelf()'
method was a hint that the original type contained a
type parameter, but it was inaccurate in some subtle cases.
- Since we performed this replacement on both the
'openedFullType' and 'openedType', CSApply had to "undo"
it to replace the existential type with the opened
existential archetype. This is because while the formal
type of the access returns the existential type, in
reality it returns the opened type, and CSApply has to
insert the correct ErasureExpr.
This was also done unconditionally, but even if it were
done more carefully, it would do the wrong thing because
of the 'loss of information' above.
- There was something fishy going on when we were dealing
with a constructor that was declared on the Optional type
itself, as seen in the fixed type_checker_crasher.
- TypeBase::eraseOpenedExistential() would transform a
MetatypeType of an opened existential into a MetatypeType
of an existential, rather than an ExistentialMetatypeType
of the existential type.
The new logic has the following improvements:
- When opening a function type, we replace the 'Self' type
parameter with the existential type in 'openedType', but
*not* 'openedFullType'. This simplifies CSApply, since it
doesn't have to "undo" this transformation when figuring
out what coercions to perform.
- We now only use replaceCovariantResultType() when working
with Self-returning methods on *classes*, which have more
severe restrictions than protocols. For protocols, we walk
the type using Type::transform() and handle all the cases
properly.
- Not really related to the above, but there was a whole
pile of dead code here concerning associated type references.
Note that SILGen still needs support for function conversions
involving an existential erasure; in the above Clonable example,
Sema now models this properly but SILGen still crashes:
let _: () -> Clonable = c.getClonerFunc()
Fixes <rdar://problem/28048391>, progress on <rdar://problem/21391055>.
Fixes SR-2757.
Variables in capture lists are treated as 'let' constants, which can
result in misleading, incorrect diagnostics. Mark them as such in order
to produce better diagnostics, by adding an extra parameter to the
VarDecl initializer.
Alternatively, these variables could be marked as implicit, but that
results in other diagnostic problems: capture list variables that are
never used produce warnings, but these warnings aren't normally emitted for
implicit variables. Other assertions in the compiler also misfire when
these variables are treated as implicit.
Another alternative would be to walk up the AST and determine whether
the `VarDecl`, but there doesn't appear to be a way to do so.
Previously, validateDecl() would check if the declaration had an
interface type and use that as an indication not to proceed.
However for functions we can only set an interface type after
checking the generic signature, so a recursive call to validateDecl()
on a function would "steal" the outer call and complete validation.
For generic types, this meant we could have a declaration with a
valid interface type but no generic signature.
Both cases were problematic, so narrow workarounds were put in
place with additional new flags. This made the code harder to
reason about.
This patch consolidates the flags and establishes new invariants:
- If validateDecl() returns and the declaration has no interface
type and the isBeingValidated() flag is not set, it means one
of the parent contexts is being validated by an outer recursive
call.
- If validateDecl() returns and the declaration has the
isBeingValidated() flag set, it may or may not have an interface
type. In this case, the declaration itself is being validated
by an outer recursive call.
- If validateDecl() returns and the declaration has an interface
type and the isBeingValidated() flag is not set, it means the
declaration and all of its parent contexts are fully validated
and ready for use.
In general, we still want name lookup to find things that have an
interface type but are not in a valid generic context, so for this
reason nominal types and associated types get an interface type as
early as possible.
Most other code only wants to see fully formed decls, so a new
hasValidSignature() method returns true iff the interface type is
set and the isBeingValidated() flag is not set.
For example, while resolving a type, we can resolve an unqualified
reference to a nominal type without a valid signature. However, when
applying generic parameters, the hasValidSignature() flag is used
to ensure we error out instead of crashing if the generic signature
has not yet been formed.
- The DeclContext versions of these methods have equivalents
on the DeclContext class; use them instead.
- The GenericEnvironment versions of these methods are now
static methods on the GenericEnvironment class. Note that
these are not made redundant by the instance methods on
GenericEnvironment, since the static methods can also be
called with a null GenericEnvironment, in which case they
just assert that the type is fully concrete.
- Remove some unnecessary #includes of ArchetypeBuilder.h
and GenericEnvironment.h. Now changes to these files
result in a lot less recompilation.
- TypeAliasDecl::getAliasType() is gone. Now, getDeclaredInterfaceType()
always returns the NameAliasType.
- NameAliasTypes now always desugar to the underlying type as an
interface type.
- The NameAliasType of a generic type alias no longer desugars to an
UnboundGenericType; call TypeAliasDecl::getUnboundGenericType() if you
want that.
- The "lazy mapTypeOutOfContext()" hack for deserialized TypeAliasDecls
is gone.
- The process of constructing a synthesized TypeAliasDecl is much simpler
now; instead of calling computeType(), setInterfaceType() and then
setting the recursive properties in the right order, just call
setUnderlyingType(), passing it either an interface type or a
contextual type.
In particular, many places weren't setting the recursive properties,
such as the ClangImporter and deserialization. This meant that queries
such as hasArchetype() or hasTypeParameter() would return incorrect
results on NameAliasTypes, which caused various subtle problems.
- Finally, add some more tests for generic typealiases, most of which
fail because they're still pretty broken.
