- Fix TypeCheckExpr.cpp to be more careful when propagating sugar from an
argument to the result of the function. We don't want to propagate parens,
because they show up in diagnostics later.
- Restructure FailureDiagnosis::diagnoseFailure() to strictly process the tree
in depth first order. Before it would only do this if contextual typing was
unavailable, leading to unpredictable inconsistencies between diagnostics.
- Always perform diagnoseContextualConversionError early, as part of the thing
that calls the visitor, instead of in each visit method. This may change in
the future, but is a simplification for now.
- Make the operator processing code handle the "candidate is an exact match"
case by emitting a diagnostic indicating that the result type of the operator
must not match expectations, instead of emitting the silly things like
"binary operator '&' cannot be applied to two Int operands" which is obviously
false.
These changes lead to minor improvements across the testsuite, and should make the
diagnostics more predictable for more complex real-world ones, but I haven't gone
through the radars yet.
Major missing pieces:
- CallExpr isn't using the same logic that the operators are.
- When you have a near match (only one argument mismatches) we should specifically
complain about that argument, instead of spewing an entire argument list.
- The noescape function attr diagnostic is being emitted twice now.
Swift SVN r29733
- When diagnosing an error passing a noescape function to an escapeing function pointer,
return 'true' to avoid follow-on bogus error messages and notes being produced.
- Start classifying the closeness of overload matches, to allow future diagnostics to be
improved.
Note that the second regresses the testcase in test/Constraints/generics.swift because the
decomposeArgumentType function doesn't know to look through the archetype for the protocol
present on the operator, and thus thinks the function only takes one argument. Advice on
how to best detect this situation is appreciated.
Swift SVN r29731
Change the fix-it suggesting adding #available to be consistent with the fix-it
adding @available.
This changes "guard with version check" to "add if #available version check".
rdar://problem/21275857
Swift SVN r29648
Just like enums with integer raw values can get autoincrementing case values,
enums with string raw values get the name of the element. The name is /not/
prefixed with the enum type because the purpose is presumably to interoperate
with a string-based system, which may require either writing or printing the
raw value as a string.
If an enum's raw type is both integer literal convertible and string literal
convertible, the integer side wins. That is, elements without raw values
will get auto-incremented integer values, rather than string values, and will
produce an error if an auto-incremented value cannot be generated.
rdar://problem/15819953
Swift SVN r29542
The case where this comes up is when people name their app and framework
targets the same thing, or when they've renamed their test target module
in an attempt to avoid issues with NSClassFromString and differing
runtime names. We currently do various wrong things when this happens,
so just emit an error instead.
I left a hole for our overlays, which use '@exported import <the-current-module>'
to get at their Clang modules. The previous commit means this can be
replaced by -import-underlying-module, but that doesn't help our tests,
which use -enable-source-import for their overlays. Which we should stop doing.
rdar://problem/21254367
Swift SVN r29440
- Diagnose instead of crashing when doing this:
extension Array : MyProto where T : Equatable {}
- Fix crash when Self appeared in where clause
- Finally, improve the diagnostic for this:
extension Array<Foo> {}
Fixes <rdar://problem/21349794>.
Swift SVN r29379
This makes it clearer that expressions like "foo.myType.init()" are creating new objects, instead of invoking a weird-looking method. The last part of rdar://problem/21375845.
Swift SVN r29375
If 'x.init' appears as a member reference other than 'self.init' or 'super.init' within an initializer, treat it as a regular static member lookup for 'init' members. This allows a more explicit syntax for dynamic initializations; 'self.someMetatype()' looks too much like it's invoking a method. It also allows for partial applications of initializers using 'someMetatype.init' (though this needs some SILGen fixes, coming up next). While we're in the neighborhood, do some other correctness and QoI fixes:
- Only lookup initializers as members of metatypes, not instances, and add a fixit (instead of crashing) to insert '.dynamicType' if the initializer is found on an instance.
- Make it so that constructing a class-constrained archetype type correctly requires a 'required' or protocol initializer.
- Warn on unused initializer results. This seems to me like just the right thing to do, but is also a small guard against the fact that 'self.init' is now valid in a static method, but produces a newly-constructed value instead of delegating initialization (and evaluating to void).
Swift SVN r29344
Instead of forcing full application of '{super,self}.init' in the parser, and installing the RebindSelf semantic expr node early, make these constraints to Sema-time checks, and parse '<expr>.init' as a regular postfix production. This is a better separation of concerns, and also opens the door to supporting 'metatype.init()' in more general expression contexts (though that part still needs some follow-up sema work).
Swift SVN r29343
If P is a protocol, calling static methods or constructors
via values of type P.Protocol makes no sense, so let's prohibit
this.
Fixes <rdar://problem/21176676>.
Swift SVN r29338
Rename existentialConformsToSelf() to existentialTypeSupported(). This
predicate is the "protocol has no Self or associated type requirements"
check, which is a looser condition than self-conformance. This was being
tested to see if the user could refer to the protocol via an existential
type.
The new existentialConformsToSelf() now checks for protocol being @objc,
and for the absence of static methods. This is used as part of the
argument type matching logic in matchType() to determine if the
existential can be bound to a generic type parameter.
The latter condition is stricter, for two reasons:
1) We allow binding existentials to multiple type parameters all sharing
the same generic type parameter T, so we don't want the user to be
able to see any static methods on T.
2) There is an IRGen limitation whereby only existentials without witness
tables can be passed in this manner.
Using the above, the representsNonTrivialGenericParameter() function
has been renamed to canBindGenericParamToExistential(). It now allows
an existential type to be bound to a generic type parameter only under
the following circumstances:
A) If the generic type parameter has no conformances, the match is allowed.
B) If the generic type parameter has at least one conformance, then all
of the conformances on the generic type parameter must be
existentialConformsToSelf() (condition 1 above), and all conformances
on the existential must be @objc (condition 2 above).
Fixes <rdar://problem/18378390> and <rdar://problem/18683843>, and lays
the groundwork for fixing a few other related issues.
Swift SVN r29337
Special-casing these as MemberRefExprs created an asymmetry
where unbound archetype instance methods (<T : P> T.f) could
not be represented. Treating class and protocol methods
uniformly also eliminates a handful of special cases around
MemberRefExpr.
SILGen's RValue and call emission peepholes now have to know
about DeclRefExprs that point to protocol methods.
Finally, generalize the diagnostic for partially applied
mutating methods to any partially applied function with an
inout parameter, since this is not supported.
Fixes <rdar://problem/20564672>.
Swift SVN r29298
Now that generic subclasses of @objc classes are supported, dust off
Doug Gregor's fix for <rdar://problem/20385288>. It is now an error
to override an @objc declaration with something that cannot be
represented as @objc.
For example, the override of foo() here will not compile unless
it is explicitly marked @nonobjc:
func foo(i: Int) {}
...
override func foo(i: Int?) {}
Note that I updated IRGen to delete some logic for figuring out when
to emit @objc metadata. We can now rely on Sema to correctly set
isObjC(), instead of checking overrides ourselves. This was wrong
anyway, now that we can have @nonobjc overrides of @objc methods,
and vice versa.
Swift SVN r29263
If we end up trying to form a substitution where the replacement type
is an existential and there is a conformance to a non-@objc protocol
(i.e., a conformance where a witness table is required), complain in
Sema rather than crashing in IRGen. Fixes rdar://problem/21087341, but
the existential/generic interaction is still quite broken.
Swift SVN r29133
We currently complain about this type of thing:
class C<T> {
@objc func foo() -> [T]
}
However this is also not supported, but crashes in IRGen:
class C {
@objc func foo<T>() -> [T]
}
Also re-word a @nonobjc diagnostic and clean up some code for @objc and
@nonobjc.
Fixes <rdar://problem/19600602> and <rdar://problem/20886887>.
Swift SVN r29117
This permits, e.g., extending Array for all Hashable T's. However, it
does not permit extending Array for T == String. Part of
rdar://problem/21142043.
Swift SVN r29107
Now that we are using OptionSetType for option sets, all the support for
doing things the old way can die.
Note: the fix-it that used to apply to RawOptionSetType, it seemed to me,
should still apply to OptionSetType, so I switched it over instead of
removing it.
Swift SVN r29066
We're not sure when or if we want 'final' on protocol extension
members, so accept it but don't complain one way or another. We zap
this early on so that we don't end up printing it in generated
interfaces. Fixes rdar://problem/21112901.
Swift SVN r29040
Instead of the rather inscrutable "'foo()' is less available than protocol requires" be
more helpful: "protocol 'HasFoo' requires 'foo()' to be available on OS X 10.9 and newer".
Also add a note indicating where the conformance was introduced, as this may be in a
separate location from both the witness and requirement.
rdar://problem/20610411
Swift SVN r29028
Our implementation of partial_apply and currying is robust enough to handle these cases now. Mutating methods are still problematic since capturing would violate 'inout' semantics. (Maybe we could support 'mutating' partial applications as @noescape closures, some day.)
Swift SVN r28992
This isn't as straightforward as it should be, since EnumElementDecls aren't AbstractFunctionDecls, but luckily there's only one trivial curry level with a thin metatype parameter.
Swift SVN r28991
done, the rest of the infrastructure is all common and can be simplified. This
leaves us with a quite small and maintainable subsystem for diagnosing these
kinds of problems.
include/swift/AST/DiagnosticsSema.def | 28 ++-----
lib/Sema/CSDiag.cpp | 132 ++++++++++------------------------
2 files changed, 48 insertions(+), 112 deletions(-)
Swift SVN r28957
this is neutral w.r.t. diagnostics quality, but deletes a ton
of code:
include/swift/AST/DiagnosticsSema.def | 21 ++---------
lib/Sema/CSDiag.cpp | 64 ++--------------------------------
2 files changed, 9 insertions(+), 76 deletions(-)
Swift SVN r28956
that make vardecls and subscripts immutable. This makes the indirect cases
a lot more specific ("this is a get-only property" instead of "this is
immutable") and allows us to consolidate a bunch of code:
2 files changed, 45 insertions(+), 119 deletions(-)
Swift SVN r28954
This is more complex than it could be if ExtensionDecl and NominalTypeDecl
had a common ancestor in the Decl hierarchy, however this is not possible
right now because TypeDecl inherits from ValueDecl.
Fixes <rdar://problem/20981254>.
Swift SVN r28941
which tell you what the problem is, not just that you have one.
- Enhance diagnostics to be more specific about function calls producing
rvalues.
Swift SVN r28939
As far as I know, the compiler now only produces the useless "cannot
assign to the result of this expression" diagnostic in cases that don't
make sense to be an lvalue, like "4 = x".
Swift SVN r28935
When an optional requirement goes unfulfilled by a conformance, and
there is a non-@objc declaration with that name in the type/extension
declaring conformance, warn that it does not satisfy the optional
requirement. Sadly, this diagnostic is long because there are notes
for the two potential fixes: add @objc to try to conform, or move the
declaration elsewhere to silence the compiler.
Fixes rdar://problem/20219297.
Swift SVN r28908
