94de193993
FIR: introduce VALUE_PARAMETER_WITH_NO_TYPE_ANNOTATION check
...
Besides adding VALUE_PARAMETER_WITH_NO_TYPE_ANNOTATION,
we remove here NO_TYPE_FOR_TYPE_PARAMETER since it doesn't exist in FE1.0.
The name also doesn't make much sense.
From the usage it looks like it should have been
VALUE_PARAMETER_WITH_NO_TYPE_ANNOTATION instead.
2021-03-17 15:04:43 +03:00
315047b164
FIR checker: Make FirJvmExternalDeclarationChecker a
...
FirMemberDeclarationChecker.
Set isExternal in status for FirRegularClass. Invoke member declaration
checkers for FirRegularClass.
2021-03-17 10:45:26 +01:00
c3d2ce0c1f
FIR checker/IDE: Add checker and quickfix for WRONG_MODIFIER_TARGET.
...
There are only simple checks for `external` and `const` for now. The
rest of the checks (see ModifiersChecker in FE1.0) will be added later.
2021-03-17 10:45:26 +01:00
ebfc431733
FIR body resolve context: fix accessor scope handling
2021-03-17 12:13:23 +03:00
a90a5f6dd4
Check for backend when checking for suspend in fun interfaces
2021-03-16 22:37:44 +01:00
99a4779c6b
JVM IR: Don't produce line numbers in synthetic Enum members
2021-03-16 21:30:18 +01:00
bea1a8c422
JVM IR: Use Object.clone in Enum.values
2021-03-16 21:30:18 +01:00
e47715f52b
Minor. Add tests with returning inline class from SAM adapter
2021-03-16 20:57:05 +01:00
4c08d10cce
FIR checker: report expression of nullable type parameter as LHS of class literals
2021-03-16 21:56:10 +03:00
7fff4108b6
FIR checker: report KClass with incorrect type argument
2021-03-16 21:56:10 +03:00
bc2228d434
FIR checker: don't report errors on type parameter as LHS of class literals
...
If a type paramter is not reified or nullable, different errors will
be reported by FirGetClassCallChecker.
When determining whether type parameter wrapped in qualified access is a
standalone expression or not, we examine whether the checker context has
other qualified accesses in stack. Class literals (::class) is desugared
to FirGetClassCall, and thus not stacked as qualified access. Since
class literals are a special type of callable reference (a subtype of
qualified access), we should keep track of FirGetClassCall in a similar
way.
2021-03-16 21:56:09 +03:00
e29b40b07f
FIR checker: report TYPE_PARAMETER_AS_REIFIED
2021-03-16 21:56:09 +03:00
5ba5b63dee
FIR checker: report nullable expression as LHS of class literals
2021-03-16 21:56:09 +03:00
33c5b49632
FIR checker: report non-typed LHS of callable references
2021-03-16 21:56:08 +03:00
329be4f906
Add a test to for the situation where pathname of destination does not name a parent.
2021-03-16 18:44:12 +01:00
3a10ea4f88
JVM_IR: synchronize code generation for inline functions
2021-03-16 19:49:29 +03:00
decfaa3ba5
JVM_IR KT-44993 preserve inner expression type when fusing if-null
2021-03-16 15:30:45 +03:00
922419efb8
FIR CFG: route to exit of try main for throw in try main
...
to make the remaining part of try main marked as dead.
^KT-45475 Fixed
2021-03-16 15:11:56 +03:00
7e5b562b33
Reproduce KT-45475: no smartcast after conditional throw in try expression
2021-03-16 15:11:53 +03:00
cef859fced
Fix exponential pseudocode size on incorrect prefix increments and decrements
...
#KT-44153 Fixed
2021-03-16 12:25:23 +03:00
d74168fb8f
PSI2IR KT-44414 fix adapted reference to imported object member
2021-03-15 21:24:25 +03:00
e7129329d2
Revert "FIR: do not eagerly resolve class references in imported annotations"
...
This reverts commit 1a57d60f
2021-03-15 18:36:13 +03:00
fc36178f3a
Annotate kotlin.reflect.jvm.reflect with ExperimentalReflectionOnLambdas
...
This function was always experimental, as explained in its kdoc, but it
was introduced before opt-in requirement markers were supported. Thus,
breaking changes (such as in KT-42746) were always expected, and the
`@ExperimentalReflectionOnLambdas` annotation just makes it clearer.
#KT-45486 Fixed
2021-03-15 15:54:31 +01:00
43140db65e
FIR2IR: better filter out non-declared data class properties
2021-03-15 17:32:41 +03:00
1a57d60f68
FIR: do not eagerly resolve class references in imported annotations
2021-03-15 17:32:39 +03:00
e630e00e99
JVM_IR KT-44744 check accessibility of enum entry 'this'
2021-03-15 17:26:49 +03:00
8dce6f2ac9
FIR CFG: don't propagate deadness on the uncaught exception path
2021-03-15 15:54:22 +03:00
d27ecca0e9
FIR CFA: keep UncaughtExceptionPath separated
2021-03-15 15:54:21 +03:00
c641fa739b
FIR: provide 'replaceSource' only for FirQualifiedAccess inheritors
2021-03-15 15:33:38 +03:00
4d65b0ef5a
[FIR2IR] Unmute passing black box test
2021-03-15 15:12:11 +03:00
631b771251
[FIR] Report proper resolution diagnostics from arguments checks
2021-03-15 12:13:08 +03:00
6cc3fff48d
[FIR] Properly handle that type variable is contained in inv or contravar positions
...
#KT-45344 Fixed
2021-03-15 12:13:07 +03:00
3d1f4b8386
[FIR] Don't add @EnhancedVisibility to types with @Nullable annotation
2021-03-15 12:13:06 +03:00
64abb2501f
Minor. Add FIR_IDENTICAL directive
2021-03-15 06:19:22 +01:00
c9f41a2440
Minor. Ignore tests on WASM
2021-03-15 06:14:44 +01:00
149064803d
JVM_IR: Do not unbox Result argument inside java SAM adapters
...
#KT-45259
2021-03-15 06:14:42 +01:00
4099dfc7e0
JVM_IR: Do not unbox Result argument inside SAM adapters
...
#KT-45259 Fixed
2021-03-15 06:14:40 +01:00
95654bb9bc
Deprecate @JvmDefault
...
#KT-40392 Fixed
(cherry picked from commit c11f38688e
2021-03-13 08:24:17 +00:00
77c263fcc4
Mark kt45410.kt with TARGET_BACKEND: JVM
2021-03-13 02:30:24 +03:00
d3b42709fd
Revert "Value Classes: Forbid var properties with value class receivers"
...
This reverts commit f43899086a
2021-03-12 17:51:24 +01:00
6429ac17d8
Revert "Raise RESERVED_VAR_PROPERTY_OF_VALUE_CLASS to error"
...
This reverts commit bad197e075
2021-03-12 17:51:21 +01:00
050db6d454
[KLIB] Add box test for generic annotations
2021-03-12 19:03:04 +03:00
607a598f1a
[IR] Add irText test for generic annotations
2021-03-12 19:02:29 +03:00
ab8f64a3ee
[FIR] Mute failing tests
2021-03-12 18:47:34 +03:00
8588412a56
[JVM IR] Support break in do-while condition.
...
This breaks from the loop itself which is inconsistent with
what happens for breaks in while conditions.
Also, the frontend will report that code after the loop is
unreachable, which it isn't. :-\
However, those issues are covered in
https://youtrack.jetbrains.com/issue/KT-17728 , so for now
we follow the old backend to not "break" anyone. :)
Fixes KT-44412
2021-03-12 13:46:27 +01:00
d0d3b57366
Minor: mute test in FIR
2021-03-12 15:08:24 +03:00
91d2f32a57
Don't lose upper non-expected type constraints to include them to intersection type during finding the result type of the fixing type variable
2021-03-12 14:36:54 +03:00
b87c2a15b5
Don't stop constraints processing if all type variables have proper equality constraints
...
Those type variable may have other constraints after incorporation into which, the constraint error may appear
^KT-42042 Fixed
2021-03-12 14:36:54 +03:00
7f7bb70596
Don't fix a type variable into the intersection type if there is an explicit expected type
...
^KT-43303 Fixed
^KT-42396 Fixed
^KT-42472 Fixed
2021-03-12 14:36:53 +03:00
e06bacafad
JVM: fix inline cycle detection
...
1. use the correct descriptor in the old backend;
2. clear the temporary variables for arguments in the IR backend.
#KT-45292 Fixed
2021-03-12 11:49:42 +01:00
Tianyu Geng