Commit Graph

9267 Commits

Author SHA1 Message Date
pyos 82899e6243 JVM_IR: reuse MethodNodes for inline functions in same module
This fixes the weird cases when a class gets overwritten by an imperfect
copy, reduces the number of classes in the output if an inline function
contains an inline call that causes it to have regenerated anonymous
objects, and makes inlining of same module functions a bit faster in
general. On the other hand, this may increase memory footprint a bit
because classes cannot be flushed to the output jar, as the inliner
would not be able to locate classes for anonymous objects if they have
already been unloaded from memory.
2020-04-08 07:04:16 +02:00
Mikhail Zarechenskiy 86fd4da567 [UAST] Fix mapping types containing type aliases
#KT-27935 Fixed
2020-04-08 01:25:14 +03:00
Mikhail Zarechenskiy 73e1ddc505 Allow Result as a return type if one enabled inline classes explicitly
#KT-38042 Fixed
2020-04-08 01:25:13 +03:00
Kevin Bierhoff 7448761dfd only stub default constructor when compiling against .java source files 2020-04-07 17:48:39 +02:00
Mads Ager 03fb49bb38 [JVM_IR] Improve debugging behavior of inline functions
Specifically, this commit improves the stepping behavior of the IR
backend around functions with defaults.

  - Improved line numbers in the default handler itself for better
    stepping when inlined.

  - Improved source information on default arguments

  - Improved test coverage of stepping behavior in old and IR backends.

Improves the stepping behaviour around inline methods with default
arguments. In particular, we now accurately step through the
evaluation of default arguments, but do _not_ spuriously show the exit
from the $default handler.
2020-04-07 16:52:45 +02:00
pyos 516692008f JVM: add tests fixed by the last two commits
One was broken on JVM, the other on JVM_IR.
2020-04-07 15:42:41 +02:00
pyos c650c9570f JVM: use conditional suspension point markers 2020-04-07 15:42:41 +02:00
Dmitry Petrov dae8872aba KT-37779 Treat named value arguments in vararg as arguments with '*' 2020-04-07 15:40:43 +03:00
Mikhail Glukhikh 565340667c Fix black box test data generation (IGNORE_BACKEND_FIR) 2020-04-06 18:25:04 +03:00
Vyacheslav Gerasimov ca0f59dbc4 Use linux endings in generated tests 2020-04-06 14:06:15 +03:00
Dmitry Petrov d5ace43614 KT-37986 Force boxing of inline class returned from function reference
KT-37998 Provide KotlinType for safe call
2020-04-04 01:32:45 +03:00
pyos 5ed845d0b4 JVM_IR: reuse same bodies for suspend funs and $$forInline versions 2020-04-03 19:51:45 +02:00
Ilya Gorbunov 8194652793 Allow placing Suppress on a type parameter
#KT-12448
2020-04-03 19:36:37 +03:00
Dmitry Petrov a422a12407 Minor: update testData
In BridgeTest::testFakeOverrideMisleadingImplementation,
order of generated bridges has changed slightly
(but the overall set of generated bridges remains the same).
2020-04-03 18:31:39 +03:00
Alexander Udalov bb1a12e28e JVM IR: use correct owner for callable references in optimized multifile classes
The owner should always be the facade class, because the part class is
package-private and will be inaccessible from other package.

Note that in the old backend, function references already do have the
facade as the owner correctly, but property references don't, this is
reported as KT-37972.
2020-04-03 15:37:56 +02:00
Igor Yakovlev a71fd0e6d9 Support for showing rendered doc comments in editor
Fixed #KT-37361
2020-04-03 12:38:51 +03:00
pyos a75408b1ed JVM_IR: use non-nullable types as SAM super types
The wrappers for conversions to nullable and non-nullable SAM types are
the same, so differentiating those has no effect other than class name
collisions.
2020-04-02 15:40:17 +02:00
Dmitry Petrov 24b8495e00 KT-30419 Box inline classes in return types of covariant overrides
Use boxed version of an inline class in return type position for
covariant and generic-specialized overrides.

Also fixes KT-35234 and KT-31585.
2020-04-01 22:59:14 +03:00
Mathias Quintero 34a64d9171 Making var arg kParameters default to empty array no argument given
fixes KT-29969
2020-04-01 16:00:16 +02:00
Alexander Udalov 3269a7e693 Change behavior of equals/hashCode on adapted function references
Function references are now equal if they refer to the same function,
and if the parameter/return type adaptation, which happens when a
reference is used where some function type is expected, is exactly the
same. This includes the number of expected positional parameters (which
can be affected by defaults/varargs), whether the coercion of vararg
parameter to Array type happened, and whether the coercion of return
type to Unit happened.

 #KT-37543 Fixed
2020-04-01 14:18:49 +02:00
pyos e98bdc6f8e JVM: preserve call site markers when inlining lambdas
and default functions into their own stubs.

Fixes #KT-35006
2020-03-31 16:06:57 +02:00
pyos 9d21800d8f JVM: fix SMAP range extension logic
If `mapLineNumber` was called in non-monotonic order, e.g. N then N+2
then N+1, the first two calls created a range that spans [N; N+2] but
the third call did not reuse it.
2020-03-31 16:06:57 +02:00
Mikhail Zarechenskiy 38a719cb22 [NI] Fix trace manipulations for builder inference and ::-expressions
For a class literal Type::class we are resolving Type as a constructor,
 getting all diagnostics (about missing arguments, for example) and then
 just not committing this trace with errors

 #KT-37626 Fixed
2020-03-31 15:50:58 +03:00
Pavel Kirpichenkov a416fde814 [NI] Move abstract class instantiation check to call checkers
This way the check works for callable reference arguments.
Also candidate applicability during resolution does not change compared to the old inference.
^KT-37530 Fixed
2020-03-31 12:04:01 +03:00
Dmitry Petrov cec64a2ec7 KT-37861 'this' is uninitialized in constructor default parameters 2020-03-30 17:23:15 +03:00
Igor Chevdar d808ef10b2 Added some tests on local classes in inline bodies 2020-03-28 15:26:19 +03:00
Igor Chevdar 90abf1fda0 [JS_IR] Don't capture the bound receiver of a CR 2020-03-28 15:26:19 +03:00
Dmitry Petrov 042424d599 KT-27524 Don't box (some) inline classes in suspend fun return
If an inline class is mapped to a reference type (or an array), it's Ok
to treat JVM view on a suspend function as returning a value of
corresponding inline class (although in reality it returns 'Any?'
because of COROUTINE_SUSPENDED).
2020-03-27 18:29:36 +03:00
Mads Ager 5570a5fe74 [JVM_IR] Use iinc for incrementing Int variables.
Fix line number generation for assignments where the right-hand
side of the assignment is not on the same line.

Fix line number generation for intrinsics functions where the
function is not on the same line as the last argument.

Be careful to not break stepping behavior with the iinc
optimizations.
2020-03-27 18:17:53 +03:00
Steven Schäfer 58685be4e2 IR: Don't use IrStringConcatenation for ordinary toString calls
We can only use IrStrinConcatentation to represent calls to Any?.toString
and toString calls on primitive types. Otherwise, x.toString() and "$x"
are observably different when x is a non-null type with null value
(e.g., an @NotNull value coming from Java).
2020-03-27 17:31:48 +03:00
Mikhail Glukhikh 7bfd354a77 Don't report UNUSED_PARAMETER in main from object #KT-37718 Fixed 2020-03-27 16:46:59 +03:00
Roman Artemev 6e01ec8dd3 [IR] Fix translation of synthetic generic java properties
- Compute substituted accessor descriptor to avoid unbound type parameters
2020-03-27 10:52:33 +03:00
Pavel Kirpichenkov e39b69839b [NI] Make constraint check for type variables with complex dependency
Additional check for trivial constraints is needed to make lambda
analysis before outer variable fixation to Nothing(?)

^KT-37627 Fixed
2020-03-26 17:06:37 +03:00
Mark Punzalan 4234fa79c0 [JVM IR] Use append(Char) for 1-length string literals in string
templates and plus concatenations.

This is slightly more efficient and mirrors the behavior of the non-IR
backend for templates (but not for plus concatenations).

#KT-36638 Fixed
2020-03-26 12:39:39 +03:00
Mikhail Glukhikh 547bb64b18 Fix BB test auto-generation (ignore backend FIR) 2020-03-26 10:31:14 +03:00
Dmitriy Novozhilov f18cef0a12 [FIR] Update ignore rules of generated backed tests for primitive 2020-03-25 18:07:28 +03:00
Dmitriy Novozhilov 6c9dd8bb22 [FIR] Reorganize modules of raw fir builders 2020-03-25 18:07:27 +03:00
pyos be37e7135a Add a test for SMAPs with interleaved files 2020-03-25 14:03:32 +01:00
pyos 4558d48481 JVM: add a language feature to omit *E between SMAP strata
Fixes #KT-37704
2020-03-25 10:33:59 +01:00
Jinseong Jeon acbe3126b1 FIR: consider all functions in scope when computing dispatch receiver parameter. 2020-03-25 09:20:07 +03:00
Pavel Kirpichenkov adf3351a8b [NI] Don't skip constraints with projections during variable fixation
Those constraints are only restored for variable fixation stage because of the
influence on pefrormance.

^KT-37650 Fixed
^KT-37380 Fixed
2020-03-24 16:13:49 +03:00
Pavel Kirpichenkov dc18c62dbc [NI] Filter out type variable from its dependencies
Don't take into account complex variable dependency on itself when determining fixation status.
^KT-37621 Fixed
2020-03-24 12:49:18 +03:00
Dmitry Petrov afacb4b4b2 KT-37604 Use proper type when generating constructor call 2020-03-24 11:31:47 +03:00
Mikhail Zarechenskiy ec4d9d2f1f [NI] Prefer between two complex variables one with proper lower bounds
This reverts commit d4d35bb766.
2020-03-24 11:13:55 +03:00
Mikhail Glukhikh a4c7619c89 [FIR2IR] Introduce & use FirBuiltInsPackageFragment
Without this commit, JVM name mapping logic in BE does not work for FIR,
because FIR cannot use old BuiltInsPackageFragmentImpl descriptor.
In this commit we add our own implementation thus fixing
a pack of FIR black box tests.
2020-03-24 10:37:53 +03:00
Mikhail Zarechenskiy 191fb02bf6 [NI] Consider intersection type with number type as Nothing
Currently, only for "in": In<in Int & A> == In<in Nothing> == In<*>

 #KT-37302 Fixed
2020-03-23 16:39:21 +03:00
Dmitry Petrov b0e916769d Add test for KT-37570 2020-03-23 12:27:58 +03:00
Georgy Bronnikov 4742057b51 IR: generate lambdas in place in Psi2Ir
even when arguments are to be rearranged.
Lambdas have no side effects, and storing them in temporary variables
prevents processing in the backend
(such as inserting continuation parameter in AddContinuationLowering).
2020-03-23 11:09:30 +03:00
Mikhail Zarechenskiy d4d35bb766 Revert "[NI] Prefer between two complex variables one with proper lower bounds"
This reverts commit ced8a92629.
2020-03-22 17:16:22 +03:00
Mikhail Zarechenskiy ced8a92629 [NI] Prefer between two complex variables one with proper lower bounds
#KT-37554 Fixed
2020-03-20 15:44:53 +03:00