Commit Graph

4553 Commits

Author SHA1 Message Date
Denis.Zharkov 7645663d12 Deprecate (V)::a reference resolution to companion in FE 1.0
^KT-45315 Fixed
2021-06-25 18:41:14 +03:00
Anton Bannykh 1006eb0945 IR: fix offsets in constant folding lowering 2021-06-25 00:59:28 +03:00
Anton Bannykh 9d1a8b2d58 IR: fill type parameters in inline class static method delegation call 2021-06-25 00:59:27 +03:00
Anton Bannykh 61a08be6b4 JS IR: don't rename the local declarations
IR2JS handles clashing declarations anyway.
2021-06-25 00:59:25 +03:00
Anton Bannykh a56cc89010 JS IR: fill enum constructor call type parameters 2021-06-25 00:59:24 +03:00
Anton Bannykh eb0c13793b JS IR: introduce 'lower per module' mode
This mode is closer to how IC supposed to work - reusing work
from dependencies, not re-lowering them.
2021-06-25 00:59:23 +03:00
Anton Bannykh a4cb70af31 JS IR: gather statement origins in one place
Preparing to serialize lowered IR
2021-06-25 00:59:22 +03:00
Anton Bannykh 1b2f4ad071 dumpKotlinLike: add a flag to treat else as true
Deserialized IR always has `true` instead of `else` branch.
Dumping `else` as `true` helps comparing regular and serialized IR.
2021-06-25 00:59:20 +03:00
pyos 26e3237b8c JVM_IR: never rename public/protected fields
And if that causes a platform declaration clash, that's not a problem
that can be solved without breaking the ABI anyway.

 #KT-47412 Fixed
2021-06-24 17:41:06 +03:00
pyos 88320cbb05 JVM_IR: produce collection stubs in a stable order
This means not storing intermediate results in any HashSets.

 #KT-47411 Fixed
2021-06-24 14:47:49 +03:00
Dmitry Petrov 0104b1275f JVM_IR KT-47398 handle @EnhancedNullability String subject as in 1.0 2021-06-23 20:11:24 +03:00
pyos 537ce05bc9 JVM_IR: assume function reference adapters are tail-call
Meaning, they never need continuation objects. This shouldn't affect
correctness (if the assumption is valid, the continuation object
should always have been removed in the end), but the phantom
continuation sometimes left behind unused accessors (and in case of
inline function references, those accessors would refer to non-existent
functions) - see the modified test and Kotlin/kotlinx.coroutines#2769.
2021-06-23 11:18:25 +02:00
Dmitry Petrov c77884f067 Refactor SAM type handling, replace non-approximated arguments with * 2021-06-22 21:13:56 +03:00
Victor Petukhov 4aeabb6b0f Use upper bound aware type approximator for intersection types inside sam types in contravariant positions to build proper types in terms of subtyping 2021-06-22 21:13:55 +03:00
Georgy Bronnikov 40872dd4ca JVM_IR: fix parallel compilation
Non-static inline functions that return inline
classes need to be tracked when copying files for by-file lowering.

FIXED: KT-45680
2021-06-22 02:22:36 +03:00
Dmitry Petrov c19792e7c5 JVM_IR KT-47326 downcast field receiver on JvmField lowering 2021-06-21 19:17:54 +03:00
pyos cf660cf24a JVM_IR: expect unboxed return value from suspend default stubs
#KT-47206 Fixed
2021-06-21 18:53:31 +03:00
pyos 913ac289ed JVM_IR: extract boolean-valued DFS on IR elements into a helper 2021-06-21 16:19:24 +03:00
pyos ee4d9a89ea JVM_IR: read inline fun bytecode to detect unused property references
I.e. optimize `$$delegatedProperties` in the same way as the old
backend. This is more reliable than trying to match bytecode.
2021-06-21 16:19:23 +03:00
pyos bd6d96114b JVM_IR: generate shorter bytecode for reflected property references 2021-06-21 16:19:23 +03:00
Jinseong Jeon 4ec745abd7 FIR: relax simple type requirement in AbstractTypeMapper
The counterparts in old FE don't have such restriction, so it can
map a type that has a flexible type as a type argument to JVM type,
e.g., Stream<String..String?>
2021-06-21 13:49:40 +03:00
Anton Bannykh 48fe46303d JS IR: declare intrinsics for lowerings in stdlib
- simplifies lowered IR serialization
- removes some hacks in jsOperators.kt
- intrinsics can be directly invoked in stdlib
2021-06-18 15:26:52 +03:00
Roman Artemev 5bc884ce2b [KLIB] Fix fakeoverride builder
Set proper visibility in case of cross-module internal overrides.
TODO: KT-47192
2021-06-16 21:16:53 +03:00
Roman Artemev 0cb23a6e65 [KLIB] Fix serialization of loops with control statements in condition 2021-06-16 21:16:52 +03:00
Roman Artemev e5c6a5b80d [IR] Clean up Fake Override builder a bit 2021-06-16 21:16:51 +03:00
Roman Artemev 1831c0a14b [IR] Support disabling of rendering IrErrorType inner KotlinType 2021-06-16 21:16:50 +03:00
Roman Artemev 2eac442705 [IR] Support stable order in IR dump
That mode is need to test deserialization of klibs
2021-06-16 21:16:49 +03:00
Roman Artemev cb130181a2 [Psi2Ir] Fix overridden symbols of Delegated Properties 2021-06-16 21:16:47 +03:00
Roman Artemev 32bbc3b8ad [KLIB] Fix SimpleTypeKey in case of type abbreviation 2021-06-16 21:16:46 +03:00
Roman Artemev 2f772c7a7c [KLIB] Add IrStatementOrigin into IrConstructorCall message 2021-06-16 21:16:45 +03:00
Roman Artemev 16ea3d9ae9 [TEST] Add IrText klib tests 2021-06-16 21:16:44 +03:00
pyos aa47191de4 JVM: hide ExpressionCodegen in finally block generation during inlining 2021-06-16 12:13:32 +02:00
pyos 108bd01698 JVM: refine the stack spilling around inline calls
Not all suspend functions need it - only those with suspension points.
2021-06-16 12:13:32 +02:00
pyos b136acb185 JVM: move callSiteFile to InlineCallSiteInfo 2021-06-16 12:13:32 +02:00
pyos 6c1a5e1211 JVM: make inline function argument processing a bit shorter 2021-06-16 12:13:32 +02:00
pyos 392e4fba42 JVM: fix inlining of default lambdas of signature (Result) -> Any
They have no `invoke` bridge, and the overridden invoke expectes a boxed
`Result` as an argument.
2021-06-16 12:13:32 +02:00
pyos a0a14d9e25 JVM: remove descriptors from SourceCompilerForInline
Printing the call site source on errors isn't really useful anyway - the
old backend points to a specific PsiElement, and JVM_IR wraps the
exception in FunctionCodegen (and printing the source never worked with
JVM_IR anyway).
2021-06-16 12:13:32 +02:00
pyos 32ad747632 JVM: load default lambda method nodes immediately
The ones that are not needed are filtered out before DefaultLambda is
even constructed anyway, and this way we need fewer lateinit vars.
2021-06-16 12:13:32 +02:00
pyos 1e953eaf01 JVM: remove contextKind from SourceCompilerForInline
Guess what? It's only useful for the old backend's type mapper!
2021-06-16 12:13:32 +02:00
pyos 1109348b6d JVM: remove lookupLocation from SourceCompilerForInline 2021-06-16 12:13:32 +02:00
pyos b6e9f64e18 JVM: remove functionDescriptor from InlineCodegen 2021-06-16 12:13:32 +02:00
pyos d89e2d9f08 JVM: split inline intrinsics into "needed by JVM_IR" and "not"
or, alternatively, "suspend" and "not".
2021-06-16 12:13:32 +02:00
pyos eb4d831d27 JVM_IR: implement typeOf<T> as a codegen intrinsic
rather than a fake inline function.

Also, generate more correct instructions for typeOf. Not sure how that
even worked before - `aconst(Boolean)` isn't even valid.
2021-06-16 12:13:32 +02:00
pyos 7dbf08ae1c JVM: move loadCompiledInlineFunction out of InlineCodegen 2021-06-16 12:13:32 +02:00
pyos 8307367f90 JVM: don't thread callDefault through the inliner
It's only used by the old backend, and should be given to
PsiSourceCompilerForInline by PsiInlineCodegen.
2021-06-16 12:13:32 +02:00
Iaroslav Postovalov 0ba3ddcc29 Minor. Fix a typo, change visibility of internal object 2021-06-16 10:22:44 +03:00
Mads Ager ae03e2983d [JVM_IR] Keep track of catch variable gaps.
This ensures that catch variables are not visible in duplicated
finally blocks generated at an exit within the catch variable
scope.

^KT-46449 Fixed
2021-06-15 15:18:59 +02:00
Victor Petukhov 1b82227308 Don't do new captured type specific checks for old ones in the type checker
^KT-47143 Fixed
2021-06-15 10:07:20 +03:00
Mads Ager 6b6ba77869 [JVM_IR] Trim the expression test string for null assertions.
^KT-47166 Fixed.
2021-06-11 21:54:04 +03:00
Dmitry Petrov bf077b893c PSI2IR KT-47245 handle declaration statements in control structures 2021-06-11 16:42:31 +03:00