That is, if there's a problem with a witness, and the witness comes
from a different extension from the conformance (or the original type,
when the conformance is on an extension), put the main diagnostic on
the conformance, with a note on the witness. This involves some
shuffling and rephrasing of existing diagnostics too.
There's a few reasons for this change:
- More context. It may not be obvious why a declaration in file
A.swift needs to be marked 'public' if you can't see the conformance
in B.swift.
- Better locations for imported declarations. If you're checking a
conformance in a source file but the witness came from an imported
module, it's better to put the diagnostic on the part you have
control over. (This is especially true in Xcode, which can't display
diagnostics on imported declarations in the source editor.)
- Plays better with batch mode. Without this change, you can have
diagnostics being reported in file A.swift that are tied to a
conformance declared in file B.swift. Of course the contents of
A.swift also affect the diagnostic, but compiling A.swift on its
own wouldn't produce the diagnostic, and so putting it there is
problematic.
The change does in some cases make for a worse user experience,
though; if you just want to apply the changes and move on, the main
diagnostic isn't in the "right place". It's the note that has the info
and possible fix-it. It's also a slightly more complicated
implementation.
The diagnostic change is harmless. The diagnostic that is no longer being
emitted was intended for multi-file conformance checking only, and we still
emit a diagnostic for the overall conformance failure with missing witnesses.
Typealiases in protocol extensions can be used to satisfy
associated type requirements. However, when they don’t meet all
of the requirements placed on the associated type, fall back to
the normal inference path rather than failing outright.
Fixes SR-6609 / rdar://problem/36038033.
Introduce an ugly hack where a typealias in a protocol extension that
has the name `_Default_Foo` can be found by associated type inference for
an associated type named `Foo`, respecting the constrains of the protocol
extension in which that typealias resides.
When a protocol closes off an associated type by tying it to a concrete
type (e.g., via a same-type constraint), allow associated type inference
to use that information to infer the associated type rather than
requiring the user to restate the obvious.
Fixes SR-6640.
During associated type inference, also look for default associated type
witnesses from overridden associated types, so that one need not
redeclare default associated types at every level in a protocol hierarchy.
Within the compiler, we use the term "layout constraint" for any
constraint that affects the layout of a type parameter that has that
constraint. However, the only user-visible constraint is "AnyObject",
and calling that a layout constraint is confusing. Drop the term
"layout" from diagnostics.
Fixes rdar://problem/35295372.
As part of type witness inference, allow us to infer a generic parameter
as the type witness for an associated type, when the associated type has
the same name as the generic parameter.
We likely want the more-general rule that generic parameters are always
visible as members of their enclosing type, but I'll tackle that separately.
This works around an order dependency that affected the source
compatibility suite.
If unqualified name lookup finds an associated type, but resolution to
the type witness fails, produce a diagnostic rather than silently
failing. Fixes the crash in SR-5825, SR-5881, and SR-5905.
It's conceivable that we could resolve this with suitably global
associated type inference... but that's far off and it's best not to
crash until then.
The base mutability of storage is part of the signature, so be sure
to compute that during validation. Also, serialize it as part of
the storage declaration, and fix some places that synthesize
declarations to set it correctly.
We had two slightly different codepaths to diagnose ': class'
in an inheritance clause where it is not supported.
For generic parameters, we would fix the 'class' to 'AnyObject',
but for associated types we didn't do this. Perform the fix in
all cases where it makes sense and remove one of the two
diagnostics.
Fixes a problem related to presence of InOutType in function parameters
which diagnostics related to generic parameter requirements didn't handle
correctly, and improves diagnostics for unsatisfied generic requirements
in operator applications, which we didn't attempt to diagnose at all.
Resolves: rdar://problem/33477726
* Update usage checking to account for __shared parameters as immutable
* Allow pattern type checking to resolve Shared parameters to the appropriate parameter specifiers
* Add the __shared protocol requirement restriction
Swift 3 allowed a requirement to be satisfied by an unavailable
witness, which doesn't make sense. We've been warning about it in
Swift 3 for a while; make it an error in Swift 4.
Whenever we form a potential archetype that is unresolved (because it
names a member wasn't known at the time the potential archetype was
formed), create a corresponding delayed requirement to resolve the
potential archetype. This ensures that all potential archetypes get a
chance to be resolve, fixing the
nested type should have matched associated type
assertion in rdar://problem/31401161 (and others).
(...is constrained to be a subtype of another)
Previously the compiler would just mark the entry in the inheritance
clause invalid and move on without emitting any errors; in certain
circumstances in no-asserts builds this could actually lead to
everything working "correctly" if all conforming types happened to
pick the same concrete type for both associated types. In Swift 4 this
can actually be enforced with a same-type requirement, which will
guarantee that the two associated types are the same even in generic
contexts.
This fix avoids assertions and crashes, but the diagnostic is still
incorrect, and in the simple case of the inheritance clause it's
redundant. Doing something better and possibly even downgrading it to
a warning in Swift 3 mode is tracked by rdar://problem/32409449.
Initial patch by Slava, fixed up by me.
Recursive concrete and superclass constraints are detected
per-equivalence-class; record them that way.
Use that information to drop recursive concrete and superclass
constraints from the resulting signature, which frees the canonical
generic signature builder from having to worry about such recursive
constraints. This eliminates the invalid-code crashes introduced in
the prior commit that disabled finalization for the canonical GSBs, as
well as fixing one other random crash-on-invalid.
Looks like subscript validation wasn't checking or setting
ValueDecl::isBeingValidated(). This caused a stack overflow
with associated type inference. The way to trigger this
during normal use is to invoke the fixit for adding
missing protocol requirements -- they're written in terms
of the protocol's associated type, which the user might
not have defined yet.
Fixes <rdar://problem/26680599>.
When a requirement mentions a concrete type, that type might utter
other types (e.g., Set<T>) that infer requirements (here, T:
Hashable). Perform requirement inference for such types.
Part of rdar://problem/31520386.
ArchetypeBuilder::finalize() is needed to tie up any loose ends before
requesting a generic signature or generic environment. Make sure it
gets called consistently.
Introduce an algorithm to canonicalize and minimize same-type
constraints. The algorithm itself computes the equivalence classes
that would exist if all explicitly-provided same-type constraints are
ignored, and then forms a minimal, canonical set of explicit same-type
constraints to reform the actual equivalence class known to the type
checker. This should eliminate a number of problems we've seen with
inconsistently-chosen same-type constraints affecting
canonicalization.
Now that NameAliasTypes desugar to interface types, it is possible
to have a protocol requirement type contain a NameAliasType which
contains an associated type:
protocol P {
associatedtype Element
typealias Elements = [Element]
func process(elements: Elements)
}
In Swift 3, the typealias would be desugared at name lookup time
in this case, but this is no longer the case, as a result associated
type inference stopped working in this example.
Fixes <https://bugs.swift.org/browse/SR-3641>.
When enumerating requirements, always use the archetype anchors to
express requirements. Unlike "representatives", which are simply there
to maintain the union-find data structure used to track equivalence
classes of potential archetypes, archetype anchors are the
ABI-stable canonical types within a fully-formed generic signature.
The test case churn comes from two places. First, while
representatives are *often* the same as the archetype anchors, they
aren't *always* the same. Where they differ, we'll see a change in
both the printed generic signature and, therefore, it's
mangling.
Additionally, requirement inference now takes much greater
care to make sure that the first types in the requirement follow
archetype anchor ordering, so actual conformance requirements occur in
the requirement list at the archetype anchor---not at the first type
that is equivalent to the anchor---which permits the simplification in
IRGen's emission of polymorphic arguments.
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.
- In functions called from resolveType(), consistently
use a Type() return value to indicate 'unsatisfied
dependency', and ErrorType to indicate failure.
- Plumb the unsatisfiedDependency callback through the
resolution of the arguments of BoundGenericTypes, and
also pass down the options.
- Before doing a conformance check on the argument of a
BoundGenericType, kick off a TypeCheckSuperclass request
if the type in question is a class. This ensures we don't
recurse through NominalTypeDecl::prepareConformanceTable(),
which wants to see a class with a valid superclass.
- The ResolveTypeOfDecl request was assuming that
the request was satisfied after calling validateDecl().
This is not the case when the ITC is invoked from a
recursive call to validateDecl(), hack this up by returning
*true* from isResolveTypeDeclSatisfied(); otherwise we
assert in satisfy(), and we can't make forward progress
in this case anyway.
- Fix a bug in cycle breaking; it seems if we don't invoke
the cycle break callback on all pending requests, we end
up looping forever in an outer call to satisfy().
- Remove unused TR_GlobalTypeAlias option.
We no longer need a separate "pass" that creates an archetype builder
that inherits context archetypes, because we no longer ever inherit
context archetypes.
This function did three things:
- In debug builds, record an association between the newly-created
context archetypes and the current DeclContext.
- Set the accessibility of the GenericTypeParamDecls as appropriate.
- Re-check the types written in the GenericParamList.
The last step was not needed, because we no longer serialize
GenericParamLists, or care if the RequirementRepr contains valid
types at all. The other two have been moved elsewhere.
First, ensure all ParamDecls that are synthesized from scratch are given
both a contextual type and an interface type.
For ParamDecls written in source, add a new recordParamType() method to
GenericTypeResolver. This calls setType() or setInterfaceType() as
appropriate.
Interestingly enough a handful of diagnostics in the test suite have
improved. I'm not sure why, but I'll take it.
The ParamDecl::createUnboundSelf() method is now only used in the parser,
and no longer sets the type of the self parameter to the unbound generic
type. This was wrong anyway, since the type was always being overwritten.
This allows us to remove DeclContext::getSelfTypeOfContext().
Also, ensure that FuncDecl::getBodyResultTypeLoc() always has an interface
type for synthesized declarations, eliminating a mapTypeOutOfContext()
call when computing the function interface type in configureInterfaceType().
Finally, clean up the logic for resolving the DynamicSelfType. We now
get the interface or contextual type of 'Self' via the resolver, instead
of always getting the contextual type and patching it up inside
configureInterfaceType().
The previous patches regressed a test where we used to diagnose
(poorly) a circular associated type, like so:
associatedtype e: e
With the error "inheritance from non-protocol, non-class type 'e'".
This error went away, because we end up not setting the interface
type of the associated type early enough. Instead, we return an
ErrorType from resolveTypeInContext() and diagnose nothing.
With this patch, emit a diagnostic at the point where the ErrorType
first appears.
Also, remove the isRecursive() bit from AssociatedTypeDecl, and
remove isBeingTypeChecked() which duplicates a bit with the same
name in Decl.
This fixes a regression from the previous patch which got
rid of PrintOptions::StripDynamicSelf.
When printing protocol declarations with a BaseType set in
PrintOptions, we can end up with a DynamicSelfType wrapping
a non-class type, if the protocol requirement returned
Self.
Note that this changes the diagnostic for missing protocol
requirements slightly; we used to sometimes refer to 'Self'
even if the conforming type is not a class, which is not
accepted by the type checker anyway. I believe the new
diagnostics are more correct.
PR #5857 started rejecting generic requirements that adding
constraints directly to 'Self', which means the requirements would be
unsatisfiable by some models. At the time that commit was merged, we
had thought the compiler crashed on all instances of this problem.
It turns out that, with assertions disabled, these protocols would be
accepted and could be used, so downgrade the error to a 'deprecated'
warning in Swift 3 compatibility mode.
We had two egregious errors in our handling of superclass constraints:
* We were recording superclass constraints on non-representative
potential archetypes, which meant they would get ignored.
* When two equivalence classes got merged, we wouldn't add the
superclass constraint to the representative.
Drive-by fix noticed while addressing rdar://problem/29075927.
