Treating special functions for `if`, `when`, `try`, `?:` as not accepting `Nothing` result type is incorrect.
Making so leads to cases with uninferred `Nothing` result type for inner calls and lost data flow info.
MissingDependencyClassChecker.collectDiagnostics now show only unique diagnostics
As per-file analyzer trace used in checker delegates to resolve session trace, diagnostics might be duplicated because of race condition:
1. If a non-checker thread performs analyze first, diagnostics for global elements will be stored in the resolve session trace only once.
2 If the checker threads comes to the analyze first, diagnostics will be stored in the local trace, and after that might be duplicated in the resolve session trace by other analyzers.
#KT-35578 Fixed
RequiresOptIn should be used now instead of Experimental, OptIn instead
of UseExperimental. See https://github.com/Kotlin/KEEP/pull/201.
This change adds the new declarations only to the stdlib, and supports
them in the compiler. Because of the way compiler loads annotations, we
need to bootstrap it first before deprecating the old annotations and
updating tests.
#KT-34647 Fixed
Call transformer previously ran completion of argument atoms
only for non-error candidates. This led to missing diagnostics,
i.e. from collection literal resolver.
Now arguments of calls resolved to error descriptor are completed,
with exception to not found provideDelegate calls.
provideDelegate's subatoms are not completed after failure, because
it is a part of delegate competion, which does not end with
unresolved provideDelegate.
Completing after provideDelegate failure removes constraint system
from resolved arguments, which breaks resolve for get/setValue.
^KT-33592 Fixed
Unresolved reference diagnositc was not reported on callable references
returned from if or when expression because of additional block wrapper.
^KT-30953 Fixed
With NO_LOCKS strategy we can easily end up in a situation when
constraint system for a generic call is built incorrectly,
producing flaky errors (or don't produce errors at all)
Now proper storage manager is injected for all cases except:
- IR
- Codegen
- Serialization plugin
- Fake local objects
Most likely, NO_LOCKS strategy for these cases is fine as at that point
the compiler works in one thread
#KT-34786 Fixed
FakeCallableDescriptorForObject instance, which is not ConstructorDescriptor,
is checked with call checkers when object's contents are being used.
Checker of missing supertypes should not be run against this fake descriptor.
Object's members belong to separate member scope, so their resolution doesn't
force containing class' supertypes evaluation. I.e. object's methods
may work fine even if some supertypes of containing class are missing.
Imports resolution is a rather time consuming task and doesn't have
`checkCanceled` on its computation path. So, if a thread is resolving
imports it will do it until the end of computation, which can lead to
unnecessary work and some IDE freezes.
#KT-34990 Fixed
Profiling has shown, that supertype hierarchy scan for all calls has considerable
performance cost. However, missing supertypes may be calculated only once per
descriptor which would help avoiding multiple supertype hierarchy scans for
resolved calls from the same class. New memoizer is injected into call completers
and checker contexts and then used for retrieving missing super classifiers.
#KT-19234 Fixed
Call checker and declaration checker are used in order to preserve backward compatibility.
Attempt to use classifier usage checker was not good enouth,
since not all errors found with it would actually be reported before.
For example types and constructor calls don't cause supertypes to resolve,
so missing supertypes would not lead to errors in case they are the only use of class name.
Updated tests failing due to missing Java dependencies in superclasses.
Substituting inferred type parameters with single substitutor leads
to incorrect behaviour in cases, when class' type parameters are used in constructor.
As a side effect of two-step substitution, intermediate descriptor is created,
which prevents incorrect substitution. To preserve this side effect, single
resulting substitutor was split into two substitutors: one for substituting fresh
variables and another for substituting inferred variables and known parameters.
^KT-32415 Fixed
Known type parameters appear after inheriting from class with type
parameters. Their substitution matters for inner class constructor,
because without substitution it's parameters will be type checked
against incorrect (original) parameter descriptor with unsubstituted
type parameters.
Skip creation of composite substitutor, if old substitutor is empty.
New substitutors return null in case they don't substitute a type,
but old type substitutors have explicit isEmpty method. Composite
substitutor with empty old substitutor leads to creation
of incorrect descriptor copies.
New behavior was introduced in ab506c1579 which breaks compilation on IDE module (see test example) for new inference.
Now we temporarily disable this optimisation. This is temporarily restriction which will be fixed once we'll have the same rules for lambdas completion
#KT-34889 Open
Motivation: missing visibility modifier is an error in visibility modifiers list, so we should highlight this list.
Including a name in the range is convenient for using alt+enter (you don't have to move cursor from name to fun/class/val keyword)
Also change NO_EXPLICIT_RETURN_TYPE_IN_API_MODE diagnostic range to 'declaration name' to match corresponding IDE inspection.
Fix stylistic problems and typos after review
Change CLI flag to -Xexplicit-api=strict|warning. 'Disable' state and 'mode' suffix are left out as implementation details.
Change intention title to 'make X public explicitly'
Do not report 'no explicit visibility' on property accessors
Set DECLARATION_SIGNATURE as a range for report
Rename internal diagnostic from _MIGRATION to _WARNING
Pavel Kirpichenkov