Type on the right-hand side of the element conversion/pattern match
should be allowed to have holes to be able to diagnose failures with
structurally incompatible types.
Resolves: rdar://problem/60832876
If the type used in for-each loop doesn't conform to `Sequence`
let's not try to diagnose anything about its element since such
diagnostics would include unresolved types and actual problem
(missing conformance) would already be diagnosed.
Make TypeChecker::typeCheckPattern() return the computed type, rather
than returning an "error" flag. More importantly, make it functional,
so that it doesn't set the type it computes on the pattern.
Use UnresolvedType as a placeholder for types that need to be
inferred, rather than a null type. This allows us to produce
structural types involving types that need to be inferred.
Note that with this change as-is, we get a number of duplicated
diagnostics, because typeCheckPattern() will be called multiple times
for the same pattern and will emit some diagnostics each time. This
will be addressed in a subsequent commit.
Rather than having the type checker look for the specific witness to
next() when type checking the for-each loop, which had the effect of
devirtualizing next() even when it shouldn't be, leave the formation
of the next() reference to SILGen. There, form it as a witness
reference, so that the SIL optimizer can choose whether to
devirtualization (or not).
Introduce a new kind of constraint, the "value witness" constraint,
which captures a reference to a witness for a specific protocol
conformance. It otherwise acts like a more restricted form of a "value
member" constraint, where the specific member is known (as a
ValueDecl*) in advance.
The constraint is effectively dependent on the protocol
conformance itself; if that conformance fails, mark the type variables
in the resolved member type as "holes", so that the conformance
failure does not cascade.
Note that the resolved overload for this constraint always refers to
the requirement, rather than the witness, so we will end up recording
witness-method references in the AST rather than concrete references,
and leave it up to the optimizers to perform devirtualization. This is
demonstrated by the SIL changes needed in tests, and is part of the
wider resilience issue with conformances described by
rdar://problem/22708391.
* [TypeChecker] Enclosing stubs protocol note within editor mode
* [test] Removing note from test where there is no -diagnostics-editor-mode flag
* Formatting modified code
* [tests] Fixing tests under validation-tests
Patch up all the places that are making a syntactic judgement about the
isInvalid() bit in a ValueDecl. They may continue to use that query,
but most guard themselves on whether the interface type has been set.
This is an amalgam of simplifications to the way VarDecls are checked
and assigned interface types.
First, remove TypeCheckPattern's ability to assign the interface and
contextual types for a given var decl. Instead, replace it with the
notion of a "naming pattern". This is the pattern that semantically
binds a given VarDecl into scope, and whose type will be used to compute
the interface type. Note that not all VarDecls have a naming pattern
because they may not be canonical.
Second, remove VarDecl's separate contextual type member, and force the
contextual type to be computed the way it always was: by mapping the
interface type into the parent decl context.
Third, introduce a catch-all diagnostic to properly handle the change in
the way that circularity checking occurs. This is also motivated by
TypeCheckPattern not being principled about which parts of the AST it
chooses to invalidate, especially the parent pattern and naming patterns
for a given VarDecl. Once VarDecls are invalidated along with their
parent patterns, a large amount of this diagnostic churn can disappear.
Unfortunately, if this isn't here, we will fail to catch a number of
obviously circular cases and fail to emit a diagnostic.
Under non-editor mode, the fixit for inserting protocol stubs is associated with a note
pointing to the missing protocol member declaration which could stay in a separate file from
the conforming type, leading to the behavior of rdar://51534405. This change checks if
the fixit is in a separate file and issues another note to carry the fixit if so.
rdar://51534405
When we determine that an optional value needs to be unwrapped to make
an expression type check, use notes to provide several different
Fix-It options (with descriptions) rather than always pushing users
toward '!'. Specifically, the errors + Fix-Its now looks like this:
error: value of optional type 'X?' must be unwrapped to a value of
type 'X'
f(x)
^
note: coalesce using '??' to provide a default when the optional
value contains 'nil'
f(x)
^
?? <#default value#>
note: force-unwrap using '!' to abort execution if the optional
value contains 'nil'
f(x)
^
!
Fixes rdar://problem/42081852.
Rather than relying on the NameAliasType we get by default for references
to non-generic typealiases, use BoundNameAliasType consistently to handle
references to typealiases that are formed by the type checker.
Early on in the development of conditional conformances, we put in a bunch
of assertions for code that uses `conformsToProtocol()` but wasn't handling
conditional conformances. Now, `conformsToProtocol()` handles conditional
conformances by default, and these assertions are tripping up code
that would work.
Remove the assertions and add test cases that used to trip up the
assertions (but now work).
From the Swift documentation:
"If you define an optional variable without providing a default value,
the variable is automatically set to nil for you."
This fixes several issues:
- By default parent types of alias types are not printed which results in
- Erroneous fixits, for example when casting to 'Notification.Name' from a string, which ends up adding erroneous cast
as "Name(rawValue: ...)"
- Hard to understand types in code-completion results and diagnostics
- When printing with 'fully-qualified' option typealias types are printed erroneously like this "<PARENT>.Type.<TYPEALIAS>"
The change make typealias printing same as nominal types and addresses the above.
Avoid calling lookupMemberType() with an existential base
altogether.
With the previous resolveTypeInContext() patch, a compiler_crasher
regressed and hit this, because the behavior of lookupMemberType()
changed to return nullptr in fewer cases.
as a failure to convert the individual operand, since the operator
is likely conceptually generic in some way and the choice of any
specific overload is probably arbitrary.
Since we now fall back to a better-informed diagnostics point, take
advantage of this to generate a specialized diagnostic when trying to
compare values of function type with ===.
Fixes rdar://25666129.
This reverts commit 073f427942,
i.e. it reapplies 35ba809fd0 with a
test fix to expect an extra note in one place.
as a failure to convert the individual operand, since the operator
is likely conceptually generic in some way and the choice of any
specific overload is probably arbitrary.
Since we now fall back to a better-informed diagnostics point, take
advantage of this to generate a specialized diagnostic when trying to
compare values of function type with ===.
Fixes rdar://25666129.
call expression onto a callee when it was a binary expression. Doing this
requires improving the diagnostics for when the contextual result type is
incompatible with all candidates, but that is general goodness all around.
This fixes:
<rdar://problem/22333090> QoI: Propagate contextual information in a call to operands
and improves a number of diagnostics where the problem is that an operator
is used in a context that expects a type that it cannot produce.
Swift SVN r31891
we process contextual constraints when producing diagnostic. Formerly,
we would aggressively drop contextual type information on the floor under
the idea that it would reduce constraints on the system and make it more
likely to be solvable. However, this also has the downside of introducing
ambiguity into the system, and some expr nodes (notably closures) cannot
usually be solved without that contextual information.
In the new model, expr diagnostics are expected to handle the fact that
contextual information may be present, and bail out without diagnosing an
error if that is the case. This gets us more information into closures,
allowing more specific return type information, e.g. in the case in
test/expr/closure/closures.swift.
This approach also produces more correct diagnostics in a bunch of other
cases as well, e.g.:
- var c = [:] // expected-error {{type '[_ : _]' does not conform to protocol 'DictionaryLiteralConvertible'}}
+ var c = [:] // expected-error {{expression type '[_ : _]' is ambiguous without more context}}
and the examples in test/stmt/foreach.swift, test/expr/cast/as_coerce.swift,
test/expr/cast/array_iteration.swift, etc.
That said, this another two steps forward, one back thing. Because we
don't handle propagating sametype constraints from results of calls to their
arguments, we regress a couple of (admittedly weird) cases. This is now
tracked by:
<rdar://problem/22333090> QoI: Propagate contextual information in a call to operands
There is also the one-off narrow case tracked by:
<rdar://problem/22333281> QoI: improve diagnostic when contextual type of closure disagrees with arguments
Swift SVN r31319
Take expression depth and preorder traversal index into account when
deciding which unresolved overload to complain about, rather than giving
up if there are two exprs with the same number of overloads. Don't
consider solutions with fixes when emitting ambiguous-system
diagnostics.
Swift SVN r30931
down to call argument lists that have more than one operand (heavily leveraging
"computeTupleShuffle"). This resolves a great number of QoI radars, including
things like:
<rdar://problem/19981782> QoI: poor diagnostic for call to memcmp with UInt length parameter
where we used to produce:
error: cannot invoke 'memcmp' with an argument list of type '([UInt8], [UInt8], UInt)'
return memcmp(left, right, UInt(left.count)) == 0
^
note: expected an argument list of type '(UnsafePointer<Void>, UnsafePointer<Void>, Int)'
but now we produce:
error: cannot convert value of type 'UInt' to expected argument type 'Int'
return memcmp(left, right, UInt(left.count)) == 0
^~~~~~~~~~~~~~~~
which is more "to the point"
Swift SVN r30930
machinery, instead of in multiple places in CSSolver and CSDiags. This leads
to more predictable behavior (e.g. by removing the UnboundGenericParameter
failure kind) and eliminates a class of "'_' is not convertible to 'FooType'"
diagnostics.
Swift SVN r30923
other constraints intentionally ripped off, tell the recursive solution that
we can tolerate an ambiguous result. The point of this walk is not to
produce a concrete type for the subexpression, it is to expose any structural
errors within that subsystem that don't depend on the contextual constraints.
Swift SVN r30917
- Produce more specific diagnostics relating to different kinds of invalid
- add a testcase, nfc
- Reimplement FailureDiagnosis::diagnoseGeneralMemberFailure in terms of
Not including r30787 means that we still generate bogus diagnostics like:
[1, 2, 3].doesntExist(0) // expected-error {{type 'Int2048' does not conform to protocol 'IntegerLiteralConvertible'}}
But it is an existing and separable problem from the issues addressed here.
Swift SVN r30819
r30787 causes our tests to time out; the other commits depend on r30787.
Revert "revert part of my previous patch."
Revert "Produce more specific diagnostics relating to different kinds of invalid"
Revert "add a testcase, nfc"
Revert "- Reimplement FailureDiagnosis::diagnoseGeneralMemberFailure in terms of"
Revert "Fix places in the constraint solver where it would give up once a single "
Swift SVN r30805
performMemberLookup, eliminating a ton of duplicated logic, but keeping the
same general behavior.
- Now that r30787 landed, we can have diagnoseGeneralMemberFailure inform
clients when a member lookup fails due to referencing a candidate decl of
ErrorType (i.e, it is already invalid somehow). When this happens, there is
no reason to diagnose a problem, because the original issue has been diagnosed
and anything we produce now is just garbage.
The second point cleans up a bunch of bogus diagnostics in the testsuite, which are
*actually* due to upstream error that are already diagnosed.
Swift SVN r30789
"unavoidable failure" path, along with Failure::DoesNotHaveNonMutatingMember and
just doing some basic disambiguation in CSDiags.
This provides some benefits:
- Allows us to plug in much more specific diagnostics for the existing "only has
mutating members" diagnostic, including producing notes for why the base expr
isn't mutable (see e.g. test/Sema/immutability.swift diffs).
- Corrects issues where we'd drop full decl name info for selector references.
- Wordsmiths diagnostics to not complain about "values of type Foo.Type" instead
complaining about "type Foo"
- Where before we would diagnose all failures with "has no member named", we now
distinguish between when there is no member, and when you can't use it. When you
can't use it, you get a vauge "cannot use it" diagnostic, but...
- This provides an infrastructure for diagnosing other kinds of problems (e.g.
trying to use a private member or a static member from an instance).
- Improves a number of cases where failed type member constraints would produce uglier
diagnostics than a different constraint failure would.
- Resolves a number of rdars, e.g. (and probably others):
<rdar://problem/20294245> QoI: Error message mentions value rather than key for subscript
Swift SVN r30715
get the same wording, fixing <rdar://problem/21964599> Different diagnostics for the same issue
While I'm in the area, remove some dead code.
Swift SVN r30713
