Commit Graph

57 Commits

Author SHA1 Message Date
Dmitry Savvinov 5cb949ad7f Fix language features in tests for gradual migration to 1.3 (part 2)
In 1.3, due to changes in language, testdata for some tests can be
different from 1.2

We want to simlultaneously test both versions, so instead of fixing
language version in such tests, we split them into two: one with fixed
1.2, another with fixed 1.3
2018-07-05 10:42:49 +03:00
Alexander Udalov 6e410cb182 Make TypeConstructor.isFinal return false for enums
The reason is that before dc02b2e3ab and 8a0dcca957,
TypeConstructor.isFinal for some class descriptors
(DeserializedClassDescriptor, LazyJavaClassDescriptor,
MutableClassDescriptor) were implemented as `isFinalClass` (which is
`modality == FINAL && kind != ENUM_CLASS`), and all others as
`modality == FINAL` or simply true/false. This led to differences in
behavior depending on the exact instance of the class descriptor.
Now that TypeConstructor.isFinal is always `modality == FINAL`, some
tests (PseudoValueTestGenerated) fail because the finality of some type
constructors changed and these tests render final vs non-final type
constructors differently.

In this commit, TypeConstructor.isFinal is now made to behave safer,
i.e. considering enum class type constructor to be non-final (as was the
case earlier for some ClassDescriptor instances). Some diagnostics might
disappear (e.g. FINAL_UPPER_BOUND) but it doesn't look like a big deal
2017-10-18 12:45:45 +02:00
Denis Zharkov edbf360852 Do not render trivial variables in DataFlow tests
The reason is that while there are optimizations for them
definition of their state is both rather simple and may be
confusing in testData
2017-10-02 10:04:16 +03:00
Denis Zharkov 477aeef90c Update test data for rendered CFA results after optimizations 2017-10-02 10:04:16 +03:00
Denis Zharkov 59f2ba98a6 Avoid creating SimpleType instances for default type 2017-07-20 10:53:09 +03:00
Denis Zharkov 02b40326cc Fix incorrect coroutines codegen behavior
If all the suspension calls in a suspend function were "hidden"
under the for-convention (iterator/next/hasNext) calls,
control-flow didn't find them, thus supposing that there is no
suspension points and there is no need to generate a coroutine state machine

The solution is to add relevant calls to CFG

 #KT-15824 Fixed
2017-01-25 13:54:01 +03:00
Mikhail Glukhikh b81268cca1 Control-flow analysis: do-while scope is ended after condition but before jump #KT-15334 Fixed 2017-01-10 17:57:43 +03:00
Nikolay Krasko 08d628537c Assign trailing and preceding whitespaces to lambda block until the last comment in parser 2016-10-27 14:26:10 +03:00
Mikhail Glukhikh a19d178867 Initialization analysis : more accurate handling of non-local variables while merging data #KT-14304 Fixed 2016-10-20 19:52:44 +03:00
Valentin Kipyatkov 8baa9ba1de No parameter names in type of lambda expression and anonymous function 2016-10-11 23:38:54 +03:00
Valentin Kipyatkov 41ee06ec96 Use parameter names in DescriptorRenderer 2016-10-11 23:38:49 +03:00
Mikhail Glukhikh f95dcad179 CFA minor: retain initialization state while analyzing variable declaration
(cherry picked from commit c89f4db)
2016-09-29 15:57:01 +03:00
Mikhail Glukhikh 436b0ec873 CFA merge for definition / initialization : more accurate handling of unknown variables #KT-13969 Fixed
(cherry picked from commit 4a96589)
2016-09-29 15:56:47 +03:00
Mikhail Glukhikh f8039249c6 CFA: additional jumps to catch / finally generated in the end of try / before exits from try #KT-5469 Fixed
Also #KT-13612 Fixed
(cherry picked from commit 7c188b3)
2016-08-31 19:28:09 +03:00
Mikhail Glukhikh 6a9d058db4 UNUSED_VALUE is now reported on expression only if this expression is unused in all (e.g. finally) branches #KT-9825 Fixed
(cherry picked from commit aac8e94)
2016-08-31 19:28:00 +03:00
Ilya Gorbunov 58a010464a Exception typealiases rendering in test output (CFG) 2016-08-17 12:40:50 +03:00
Mikhail Glukhikh 35446037bb UNINITIALIZED_ENUM_ENTRY compiler warning introduced #KT-2349 Fixed
(cherry picked from commit 76ac6d1)
2016-08-08 17:47:56 +03:00
Mikhail Glukhikh d24528f6bb Control flow analysis: nested finally blocks now appear in correct order in CFG #KT-4764 Fixed 2016-04-22 15:16:57 +03:00
Mikhail Glukhikh b975b7d26e More accurate CFG for local class constructors (now they are not considered unconditionally executed) #KT-10042 Fixed 2016-03-23 15:04:05 +03:00
Mikhail Glukhikh b805ce06c2 CFG exhaustive when else instruction for KT-8700 2015-12-26 10:46:01 +03:00
Mikhail Glukhikh ea3a65d0b8 Additional tests for KT-10243 2015-12-02 18:14:59 +03:00
Mikhail Glukhikh fe13f39de9 Use of uninitialized variables in lambdas / object literals / local functions is forbidden now #KT-4475 Fixed
Local declarations CFA: variable initialization information before them is now taken into account
2015-11-17 18:21:09 +03:00
Stanislav Erokhin ff00bde607 Removed dispatch receiver parameter for SyntheticFieldDescriptor 2015-10-27 15:36:38 +03:00
Mikhail Glukhikh d6988ad69e Get rid of FIELD_IDENTIFIER at syntax level, two errors dropped, a set of tests fixed / deleted #KT-9539 Fixed 2015-10-21 16:36:55 +03:00
Denis Zharkov 73799e2c3c Replace deprecated lambda syntax in testData
It's done with similar constructions where possible trying to preserve
intended behavior.
Some usages are removed because they test exactly the feature that
we are going to drop soon.
2015-09-25 08:29:25 +03:00
Mikhail Glukhikh 285d5d06c7 More precise value reassignment analysis. #KT-7418 Fixed. #KT-6902 Fixed.
Variable can now be "exactly initialized", "may be initialized" and "not initialized".
A set of relevant tests. Some fixed tests.
2015-09-02 12:07:18 +03:00
Alexey Sedunov 3d3b5591e1 Control-Flow: Improve debug text of ReadValueInstruction 2015-05-26 19:28:09 +03:00
Dmitry Jemerov 4bdf598bfe compiler testdata: s/trait/interface 2015-05-12 19:43:17 +02:00
Dmitry Jemerov f93eeeb563 fix tests by adding missing 'init' keywords to testdata 2015-04-29 14:07:05 +02:00
Denis Zharkov 7d963e8e07 Differ this/non-this instances for vars initialization in class
When deal with constructed object (not this) treat it like it's fully initialized.

Otherwise (this or access with no receiver) access instruction
should be handled as it was before.

 #KT-6788 Fixed
 #KT-4126 Fixed
2015-03-11 17:45:25 +03:00
Svetlana Isakova 7a5124215a Updated test data for Array<out T> vararg type 2014-12-12 23:08:28 +03:00
Alexey Sedunov a3a0edca58 Pseudocode: Drop CompilationErrorInstruction 2014-10-14 01:59:43 +04:00
Svetlana Isakova 976d5ab030 Pseudocode: don't write full label name when jump on it 2014-09-07 01:04:00 +04:00
Svetlana Isakova 61dc110cc7 Report 'break or continue outside a loop'
for break/continue outside a loop body (e.g. in loop condition)

 #KT-5724 Fixed
2014-09-07 01:03:57 +04:00
Svetlana Isakova 0b4f313b6d Jump on 'continue' to condition entry points for all loops 2014-09-05 18:31:53 +04:00
Svetlana Isakova 3d7c22f242 Added exact label names for labels 2014-09-05 18:31:53 +04:00
Alexey Sedunov 77bbf8f73b Pseudocode: Generate fake values for Unit/Nothing-typed expressions 2014-07-28 13:07:29 +04:00
Alexey Sedunov b0bf11c51e Pseudocode: Consume value of true condition in while-loop 2014-07-28 13:04:01 +04:00
Alexey Sedunov 487e7f2a91 Pseudocode: Replace {<: Any?} type predicate with * 2014-07-11 19:13:52 +04:00
Alexey Sedunov 4707f4a5e9 Pseudocode: Add debug information to magic instructions 2014-07-11 19:13:49 +04:00
Alexey Sedunov 0a8fa404bf Pseudocode: Generate additional returns for implicit return values (lambdas, functions with expression bodies) 2014-07-11 19:13:26 +04:00
Alexey Sedunov 1c75a5f642 Pseudocode: Generate instructions for superclass constructor calls. Consume value of delegate expression in the by-clause 2014-07-11 19:13:21 +04:00
Svetlana Isakova aa26db0538 Updated tests for control flow: call instruction stores call element
instead of callee expression
2014-07-08 14:57:38 +04:00
Alexey Sedunov f226d99d36 Pseudocode: Add valued instructions and unbound values to AbstractPseudoValueTest 2014-06-26 13:06:23 +04:00
Alexey Sedunov 5e2c2b4eca Pseudocode: Implement computation of expected type predicates 2014-06-23 19:32:25 +04:00
Svetlana Isakova 9480907514 Mark some expressions after processing their components
when it's semantically correct

Expressions: calls, 'for' loops, elvis operators
It's significant for reporting 'unreachable code' correctly
2014-06-21 12:26:32 +04:00
Alexey Sedunov 45bd3a520e Pseudocode: Generate pseudo-values for implicit receivers 2014-06-19 22:04:45 +04:00
Alexey Sedunov 168b32920b Pseudocode: Do not bind call value to callee expressions 2014-06-19 22:04:44 +04:00
Alexey Sedunov 4a5d2e6728 Add tests for element -> pseudo-value mapping 2014-06-03 18:17:07 +04:00
Alexey Sedunov d2c055e9da Pseudocode: Introduce pseudo-value analysis 2014-06-03 18:17:05 +04:00