Commit Graph

7984 Commits

Author SHA1 Message Date
Mikhail Zarechenskiy b7d7d1eb01 Fix inline class type coercion in == with generic call
#KT-27393 Fixed
2018-10-12 12:50:49 +03:00
Mikhail Zarechenskiy 0201694f84 Propagate KotlinType into create method for coroutines
#KT-27526 Fixed
2018-10-12 12:50:47 +03:00
Mikhail Zarechenskiy a8abd8cceb Fix boxing/unboxing for generic functions on index expressions
Unsubstituted type should be used for coercion to box value of
 inline class type if it's needed. For the substituted value it's not
 known if it was a generic parameter or not.

 #KT-27502 Fixed
2018-10-12 11:58:56 +03:00
Anton Bannykh 44d56cb278 JS: fix val with backing field initialization in secondary constructors
This change reverts the AssignmentTranslator logic to a previous state
of "if we assign to a val, tranlate to backing field". Previously a
check whether or not we are inside of a constructor was added. The
check didn't detect secondary constructors, hence initializing of
val's with backing field started to work incorrectly.

The check itself was added in an attempt to prevent augmented assignment
operators to reference the backing field. The check seems to have been
wrong, because an augmented assignment could happen inside a construcotr.
A more correct fix was added later. It seems that it is safe now to
revert the logic back and rely on the frontend to only allow assignment
to a val property during initilization.
2018-10-10 17:25:55 +03:00
Anton Bannykh 7afb81a7df JVM_IR: fix ranges' test generator 2018-10-10 13:13:36 +03:00
Ilmir Usmanov 9af7316845 Add call checker to report error more granulary if possible
This, however, works only for calls of 'synchronized' only. Thus, it
does not support inline functions of any kind.
2018-10-09 22:55:51 +03:00
Dmitry Petrov b55b0c1ff1 Support local property delegation to inline class values
Also, add extra test on property delegation.
2018-10-03 11:09:48 +03:00
Dmitry Petrov 6fa436911a Support interface delegation to inline class values (KT-27318) 2018-10-03 11:09:47 +03:00
Dmitry Petrov 5480bf69e8 Support property delegation to inline class values (KT-27070) 2018-10-03 11:09:46 +03:00
Dmitry Petrov 70e60ea9bc Fix inline class coercion in string template with StringBuilder.append
Inline class values with array as a carrier type were not boxed.

 #KT-27113
2018-10-03 11:09:41 +03:00
Dmitry Petrov 8891733074 Generate const val initializers as IrConst expressions 2018-10-02 13:46:37 +03:00
Mikhael Bogdanov 4fb434c0de Move out jvm-specific part from ir-common
#KT-27005 Fixed
2018-10-02 11:12:21 +02:00
Dmitry Savvinov dd682bd37d Add test on data-flow
Idea is to intersect similar smartcasts on similar properties coming
from *different* instances (something like
'this.x == null && other.x == null' ).

It checks that we distinguish subjects of such smartcasts properly.

Currently behavior is undesired, see KT-27260
2018-10-02 11:05:06 +03:00
Alexander Udalov 009f18f1f4 Split AnalysisFlag values to AnalysisFlags and JvmAnalysisFlags
Declare AnalysisFlags in module 'frontend', and JvmAnalysisFlags in
module 'frontend.java', to avoid leaking Java-related logic to common
compiler code
2018-10-01 13:31:00 +02:00
Denis Zharkov 623c6803d6 Fix regression on smart casts in when on a sealed class
See the comment in PatternMatchingTypingVisitor

 #KT-27221 Fixed
2018-10-01 09:59:28 +03:00
Alexander Udalov 088cd4b5e3 Stop parsing argument in argfile when quote ends
There are two visible effects of this change:
1) If an empty argument is passed in quotes, it will be parsed as an
   empty string and handled by the compiler, which will report an error
   later. The specific error is not very important because it's an edge
   case anyway; at the moment, "source file or directory not found:" is
   reported which is no better than the "invalid flag:" error reported
   by javac in the similar case
2) It's no longer possible to split an argument into several parts and
   quote them separately, such as:

       "-langu"ag"e-"ver'sio'n 1.2

   No test added for this change in behavior since it's an even edgier
   case. Note that javac also prohibits this.

 #KT-27226 Fixed
2018-09-28 16:02:40 +02:00
Alexander Udalov 16dd6ebe61 Handle whitespaces in argfiles correctly
#KT-27218 Fixed
2018-09-28 16:02:40 +02:00
Dmitry Petrov 7b7cf39388 Fix inline classes coercion in 'if' expression
#KT-27132
2018-09-28 10:46:42 +03:00
Dmitry Petrov ab90b2b901 Fix nullability propagation in inline class type mapping
#KT-27096

See https://jetbrains.slack.com/archives/C06E082M6/p1537949572000100
2018-09-28 10:46:42 +03:00
Dmitry Petrov 0b23ddb947 Add test for KT-27140 2018-09-28 10:46:42 +03:00
Dmitry Petrov 0191e3d1cf Migrate to release coroutines 2018-09-27 21:59:31 +03:00
Dmitry Savvinov 99454aa78d Add test on smartcast from while-assigned var
Currently behavior is undesired, it will be fixed in the next commit
(see KT-27084)
2018-09-27 12:09:21 +03:00
Dmitry Petrov 7624dbbb20 Fix subject variable handling in specialized enum codegen
#KT-27161
2018-09-27 10:48:35 +03:00
Dmitry Petrov 0dd04c3424 Postpone companion object field visibility
Have to reconsider this issue because of interface companion objects.
2018-09-27 10:35:23 +03:00
Dmitry Petrov 792ff3c39e Generate private interface companion object as package-private synthetic 2018-09-27 10:35:23 +03:00
Dmitry Petrov 7949ac1080 Use common instance receiver generation logic for 'this' expression
Otherwise it would skip private companion object accessor generation.
2018-09-27 10:35:23 +03:00
Alexander Udalov 5713298108 Fix FQ name usage of experimental marker in UseExperimental
#KT-26366 Fixed
2018-09-26 18:36:01 +02:00
Mikhail Zarechenskiy 0d103e7f0c Allow using extensions with trivial-constraints in builder-inference
#KT-27079 Fixed
2018-09-26 11:52:23 +03:00
Mikhail Zarechenskiy 0da1b9b80f Revert "Migration change: temporarily specify type explicitly"
This reverts commit bc6e091004.

 This commit was needed to avoid bootstraping, now we can revert it
2018-09-26 11:51:57 +03:00
Mikhail Zarechenskiy 7c1d374ed5 Fix diagnostic tests about coroutine-inference 2018-09-26 11:51:55 +03:00
Dmitry Petrov ba32ed7404 Test: @JvmField in inline class companion 2018-09-25 11:15:29 +03:00
Dmitry Petrov bffe9e45e8 Add accessor tests: inline class <-> companion
KT-26858 was fixed somewhere in the middle.

 #KT26858
2018-09-25 11:15:29 +03:00
Dmitry Petrov 850d72f13f Fix accessor generation for inline class members
TODO 'generateMethodCallTo' is a bloody mess that deserves rewriting.
2018-09-25 11:15:29 +03:00
Dmitry Petrov edc8cf3ed0 Use proper type for context 'this'
Bug in 'generateThisOrOuterFromContext', it worked while instance of a
class was always an object.

 #KT-27078
2018-09-25 11:15:29 +03:00
Dmitry Petrov c1bb3479df More tests on casts on receiver of inline class type 2018-09-25 11:15:29 +03:00
Dmitry Petrov 88cc900dc7 Add test for private inline class method call from inline lambda 2018-09-25 11:15:29 +03:00
Dmitry Petrov 2e6d53a43f Fix extension receiver coercion for inline classes
#KT-27031
2018-09-25 11:15:29 +03:00
Dmitry Petrov 20d7210239 Fix outer class mapping for closures in inline class
#KT-27030
 #KT-27033
2018-09-25 11:15:29 +03:00
Dmitry Petrov 26602c8443 Minor: move tests 2018-09-25 11:15:29 +03:00
Dmitry Petrov f9a419f940 Fix companion method access from inline class
#KT-27025
2018-09-25 11:15:29 +03:00
Dmitry Petrov 1869ed09bc Fix companion field access from inline class
#KT-26858
2018-09-25 11:15:29 +03:00
Mikhail Zarechenskiy 6a860dd042 Fix blackbox codegen tests that use coroutine inference
Note that previously we implicitly ignored JS_IR backend when directive
 COMMON_COROUTINES_TEST was enabled, see kotlin/generators/util/coroutines.kt#L57
2018-09-24 12:10:56 +03:00
Anton Bannykh beaf6df8c5 [JS IR BE] support CharSequence and String methods 2018-09-21 18:16:03 +03:00
Alexander Udalov 3e90d367f2 Do not report experimental usage errors in import statements
#KT-25545 Fixed
2018-09-21 17:35:50 +03:00
Alexander Udalov 9d27771f86 Fix instance parameter type of inherited declaration in reflection
Both for callables obtained via reflection API (KClass.members etc) and
for callables obtained via ::-references, the instance parameter is now
the class which was used to construct the type at the left-hand side of
the reference, NOT the class where the callable is originally declared
as is known at compile-time. The reason is to reduce the difference in
behavior of KCallable.call vs FunctionN.invoke: the latter always
required the subclass instance for a fake override, and it's reasonable
that the former would require it as well.

Note that in Java reflection, behavior could differ in a similar case.
For a simple fake override, Class.getMethod would return the method
declared in the base class and that method will accept instances of the
base class in invoke. However, it's difficult to rely on this behavior
because if there's a bridge for a fake override in the derived class
(e.g. when overridden members have different signatures), the returned
Method object is accepting the derived class as the receiver. This just
confirms the fact that Java reflection operates on a different level of
abstraction, namely JVM methods in .class files, which is not applicable
to our use cases directly. Another reason not to replicate Java
reflection's behavior is the uncertainty as to which member is returned
in case there are several in the hierarchy for a given fake override:
see the "otherwise one of the methods is chosen arbitrarily" note in
javadoc on Class.getMethod.

 #KT-24170 Fixed
2018-09-21 17:32:43 +03:00
Alexander Udalov 3ca81b95c2 Use non-allowed actual defaults in backends to workaround compiler exception
It's difficult to fix KT-22818 until the IR comes along, so we're
providing a workaround where one can disable the
ACTUAL_FUNCTION_WITH_DEFAULT_ARGUMENTS error and provide default values
in the actual function, to avoid exception from the backend.

 #KT-22818
2018-09-21 17:30:05 +03:00
Dmitry Petrov ec8bdff5c7 Working infrastructure for external dependencies testing
(not really, but at least something)
2018-09-21 09:50:21 +03:00
Dmitry Petrov 0f898dc6dc Fix default interface generic extension fun call with inline class type
When we've determined that we are going to invoke an erased inline class
method ('static foo-impl(this, ...)' in inline class), keep track of it
so that the remaining type mapping logic proceeds correctly.

 #KT-26998 Fixed Target versions 1.3.20
2018-09-21 09:48:11 +03:00
Dmitry Petrov 8d2b1950e6 Fix Kotlin default interface methods calls for inline classes
When mapping callable method signature for erased inline class methods,
use original function descriptor instead of super declaration
(otherwise it would map to a default interface method with mismatching
signature).

When generating delegates to Kotlin default interface methods, keep
track of the original Kotlin types for delegating method arguments and
interface method arguments.

'original' for value parameters of fake overrides points to the
overridden function value parameters instead of the value parameter of
the unsubstituted function. This causes inconsistent type mapping for
inline classes implementing generic interfaces with default methods.

 #KT-25295 Fixed Target versions 1.3.20
 #KT-26931 Fixed Target versions 1.3.20
2018-09-21 09:48:11 +03:00
Dmitry Petrov d9593c7a34 Add tests for inline class variables captured in closure 2018-09-21 09:48:11 +03:00