This makes it more consistent and fixes some
overlooked corner cases. Also it was decided
on the last equality applicability DM
(KT-62646) that we'd like
`is`/`!is`/`as`/`as?` to work similarly
to `===`/`!==`.
Also note that it now gives a clearer
explaination of why some corner cases work
the way they do. For example,
`FirPsiDiagnosticTestGenerated.testLambdaInLhsOfTypeOperatorCall`
yields `UNCHECKED_CAST` instead of
`CAST_NEVER_SUCCEEDS`, because
`toTypeInfo()` replaces all type arguments
with star projections, even when the argument
is not a type parameter. This is because
it has been desided to work this way in
KT-57779.
In
`FirPsiOldFrontendDiagnosticsTestGenerated..NeverSucceeds#testNoGenericsRelated`
the diagnostic is introduced, because
`t2 as FC1` and `FC1` is a final class with
no `T5` supertype.
`UNCHECKED_CAST` in
`FirPsiOldFrontendDiagnosticsTestGenerated.testSmartCast`
disappeared, because previously we didn't
take smartcasts into account.
Note that
`FirPsiOldFrontendDiagnosticsTestGenerated.testMappedSubtypes`
is a false positive. It appears because `isSubtypeOf()` doesn't
take into account platform types in supertypes of the given types
(doesn't map them).
They are not `FirRegularClassSymbol`, but are
final. `FirClassSymbol` is enough, because
`toSymbol()` returns `FirClassifierSymbol`, and:
- type parameters are always expected to have
subtypes
- typealiases have been expanded in the
previous line
It was decided to forbid such comparisons,
as we know how `===` works. Also, added some more
test cases, just for comparison.
Reusing the proper `canHaveSubtypes()`
from `TypeUtils` prevents a breaking change
in:
- `comparingTripleWithPair.kt`
- `comparisonOfGenericInterfaceWithGenericClass.kt`
But it does lead to warnings
(instead of errors) in
`incompatibleEnumEntryClasses.kt`, which is an
unrelated mistake that will be fixed in the next
commit.
The refactoring in `canHaveSubtypes()` is purely
cosmetic - otherwise reading these conditions is hard
(and they don't fit my screen vertically).
^KT-62646
^KT-65541
^KT-57779
This fixes some cases where we infer some type variable inside one
of the branches to Nothing instead of the expected type because Nothing
appeared in some other branch.
Specifically, we add an equality instead of a subtype constraint during
completion of calls to synthetic functions for if/when, try and !!.
We don't do it when the call contains a (possibly nested) elvis or is
inside the RHS of an assignment.
Otherwise, we would prevent some smart-casts.
#KT-65882 Fixed
This fixes some type argument mismatch errors caused by a captured type
being approximated and then captured again.
Some places need to be adapted to work with captured types that
previously only worked with approximated types.
#KT-62959 Fixed
When both sides of an equality expression are null or Nothing?, only the
right side is receiving implications within the data-flow. Make sure the
left side has implications add as well. This is important for boolean
conditions when implications from both sides need to be combined.
^KT-63535 Fixed
Terminating a CFG node because the result is Nothing should be reserved
for explicit Nothing type definitions, and not apply when smartcasting.
This allows boolean expressions to propagate implications correctly even
when the RHS is impossible or will never be executed.
^KT-47931 Fixed
This uses the same approach as
INFERRED_TYPE_VARIABLE_INTO_EMPTY_INTERSECTION where we use a visitor
to find a call to a symbol that contains the type variable in question.
#KT-56140 Fixed
Entering a `finally` block can happen from many different places:
through an exception, a jump, or normal exit from the `try` block. When
in the `finally` block, all DFA flows must be merged to have correct
smart casting. However, after the `finally` block, if exiting normally
or because of a jump, the combined flow from within the `finally` block
should not be used, but rather an alternate flow which combines the
correct flows from before the `finally` block.
```
try {
str as String // Potential cast exception
} finally {
str.length // Shouldn`t be resolved
}
str.length // Should be resolved
```
When building DFA flows, track the start of possible alternate flows,
and continue building them until they end. Both of these situations are
now marked on CFGNodes via interfaces.
When building the default DFA flow, and the source node is the end node
of alternate flows, attempt to use the alternate flow with the same edge
label instead of the default flow of the source node.
#KT-56888 Fixed
Previously, when a candidate was found with an applicability that is
better than the current best applicability, all previous candidates were
thrown away. Now we keep them, unless the new applicability is
successful. If no successful candidates are found, we fully resolve all
the unsuccessful ones and select the ones with the least bad
applicability. This improves diagnostics for unresolved calls.
#KT-57844 Fixed
The change is needed for the parallel resolution (^KT-55750), so we can resolve the declaration
under a lock that is specific to this declaration.
Previously, if LL FIR was resolving some FirClass, LL FIR resolved all its children too, and it had no control over what parts of the FIR tree were modified.
The same applied to the designation path, sometimes the classes on the designation path
might be unexpectedly (and without lock) modified.
This commit introduces LLFirResolveTarget, which specifies which exact declarations should be resolved during the lazy resolution of the declaration.
All elements outside the declarations specified for resolve in LLFirResolveTarget, should not be modified.
The logic of lazy transformers is the following:
- Go to target declaration collecting all scopes from the file and containing classes
- Resolve only declarations that are specified by the LLFirResolveTarget, performing the resolve under a separate lock for each declaration
^KT-56543
^KT-57619 Fixed
The compiler should only report diagnostics for
comparisons over builtins and identity-less types,
other incompatibilities should be reported
via inspections.
It's ok that in `equalityChecksOnIntegerTypes`
instead of `EQUALITY_NOT_APPLICABLE_WARNING` we get
`EQUALITY_NOT_APPLICABLE`, because
`ProperEqualityChecksInBuilderInferenceCalls`
is already active by default.
This change also replaces the notion of a representative superclass
with the least upper bound.
This makes complex types like
intersection/flexible transparent to
RULES1-based compatibility checks.
One way to look at it is to think
that this is an automatic way of handling
type parameters: automatic picking of
"interesting" bounds, and checking them against one another.
Note that `TypeIntersector.intersectTypes`
for `Int` and `T` where `T` is a type parameter
may return both `{Int & T}` or `null`
depending on `T`-s bounds. At the same time,
for type parameters `T` and `K` it will
always return `{T & K}`.
`ConeTypeIntersector.intersectTypes`, on the
other hand, will always return `{Int & T}`
irrespectively of the bounds. Meaning, the two
intersectors differ in corner cases.
`lowerBoundIfFlexible` call in `isLiterallyTypeParameter` is backed by
the `equalityOfFlexibleTypeParameters` test.
^KT-35134 #fixed-in-k2
^KT-22499 #fixed-in-k2
^KT-46383 #fixed-in-k2
The only case when behavior is change is described at
computeNonTrivialTypeArgumentForScopeSubstitutor
The idea is to avoid depending on the presence of @UnsafeVariance
and instead approximate captured types in covariant argument positions
before building substitution scopes
It's correct because for Captured(*) <: Supertype,
Out<Captured(*)> <: Out<Supertype> and when we've got @UnsafeVariance
value parameters at Out, it's ok to allow passing Supertype there.
^KT-57602 Fixed
^KT-54894 Fixed
Some of the changed tests may duplicate other existing diagnostics,
but that should not be reason not to report them at all.
There might be another job to be done to avoid diagnostic duplications
While it is theoretically useful to know that `{ while(true) {} }`
returns Nothing, CFG node deadness is not precise enough to do that: if
the entire lambda is dead, it's no longer possible to find out whether
the loop is terminating. Besides, `while (true)` and `if (true)` are
pretty much the only constructs like that anyway.
Note that this commit does not affect resolution for lambdas that end in
a Nothing-returning expression, e.g. `throw`.
If a certain type statement is true on loop entry and all continue
paths, then it is also true on exit if the condition did not reassign
the variable.
^KT-7676 tag fixed-in-k2
It's also not a backwards jump in do-while, unless it's in the loop's
condition, which is a stupid "feature" IMO. As you can probably tell
from the comments added in this commit.
This makes the `returns() implies` checker slightly cleaner, and also
fixes the case that I've missed where in RHS of `x ?:` type of `x` was
not set to `Nothing?`.
If the right-hand side is evaluated at all, then in its flow those
statements were already approved. Re-approving them erases the effect of
reassignments.
^KT-28369 tag fixed-in-k2
In theory, forking persistent flows should be cheap because of object
reuse, so the proposal here is to start from scratch and prove
redundancy of forks on a case-by-case basis. Something something better
safe than sorry.
^KT-28333 tag fixed-in-k2
^KT-28489 tag fixed-in-k2
Nikolay Lunyak