d4e8a30dcf
[NI] Fix nullability for captured type in input types approximation
2017-08-07 18:00:37 +03:00
e5c87fdad7
[NI] Filter out duplicates types in commonSuperTypeForNotNullTypes
2017-08-07 18:00:36 +03:00
b336919db3
[NI] In NI mode, use return type from special call resolution for '!!'
2017-08-07 18:00:35 +03:00
397e04f382
[NI] Do not use return type from 'onlyResolvedCall'
2017-08-07 18:00:34 +03:00
5b2e66f0ca
[NI] Invoke checkType for arguments
...
checkType performs various actions required by BE
(e.g., record nullability assertions information for JVM BE).
2017-08-07 18:00:33 +03:00
dc453e6d8b
[NI] Update lambda argument expression type when return type is known
2017-08-07 18:00:32 +03:00
ba068c2d65
Use original descriptor for enum constructor symbol
2017-08-07 18:00:31 +03:00
def0816095
[NI] Write info about hidden descriptor to candidate
2017-08-07 18:00:30 +03:00
38d9123a57
[NI] Use weaker upper constraint for case {T & Any} <: S
2017-08-07 18:00:29 +03:00
18961cc5a9
[NI] Analyze lambda in independent context if it isn't call argument
2017-08-07 18:00:28 +03:00
8599763cb9
Do not perform type checks for function literals in visitor in NI mode
...
They are type-checked as arguments for the corresponding expressions.
2017-08-07 18:00:27 +03:00
fbad16567b
[NI] Fix number type adjustment for greatest lower bound calculation
...
Compute common supertype (regardless of whether the given set of
lower bounds contains number types, intersection types, or whatever).
If the result type S is a possibly nullable intersection X1 & ... & Xn,
N = {Xi | Xi is a primitive number type},
R = {Xi | Xi is not a primitive number type},
M = default primitive number type for {Nj},
then adjusted type T* = M & R1 & ... Rm with the same nullability as S.
NB: IntegerValueType(_) = Int & Byte & Short & Long
2017-08-07 18:00:26 +03:00
1ae4278cfd
Minor: fix formatting
2017-08-07 18:00:25 +03:00
bc2a8555a3
[NI] Resolve collection literals arguments as postponed ones
2017-08-07 18:00:24 +03:00
c0c94910e2
[NI] Fix StackOverflow in TypeApproximator
...
For cases like Enum<Captured(*)> where Captured(*) has supertype
Enum<Captured(*)> we get SO in TypeApproximator.
To prevent this argument depth was introduced.
2017-08-07 18:00:23 +03:00
9fa3bce73a
[NI] Fix special case of captured type handling in subtyping
...
Given the subtyping constraint
X <: CapturedType(in Y)
we should check only
X <: Y
if X contains type variables.
2017-08-07 18:00:22 +03:00
26cc08be65
Fix anonymous function literals handling in type checker
...
- [NI] Create type info directly if the expected type is functional.
- Properly handle suspended function expected type.
2017-08-07 18:00:21 +03:00
6facdcb7ea
Minor: lambda arguments formatting
2017-08-07 18:00:20 +03:00
badbf777df
[NI] Use data flow info from arguments in parenthesis to resolve lambda
...
Fixes 'javaObjectType' and 'javaPrimitiveType'
2017-08-07 18:00:19 +03:00
5eb56dca60
[NI] Check type compatibility before fixation.
...
If for type variable we have upper and lower bounds, then sometimes
our approximation before fixation give us incorrect result for type
variable and we should chose other bound as result.
Example: Int & Byte <: T <: Byte. If we run approximation for lower
bound we get Int as result and it isn't subtype of Byte.
2017-08-07 18:00:18 +03:00
d1263c5dc3
[NI] Coerce to unit for empty lambda block
...
Fixes 'kt3903'
2017-08-07 18:00:17 +03:00
cf55674109
[NI] Support fallback resolve to operator mod if needed
...
Fixes:
- assignmentOperations
- percentAsModOnBigIntegerWithoutRem
2017-08-07 18:00:16 +03:00
f578d07b00
[NI] Fix lambda return type when expected type is not functional type
...
Introduce a fresh type variable for lambda return type.
We can't set expected lambda return type to 'Any?', because we can't
infer the actual type from return expressions in lambda in that case.
2017-08-07 18:00:15 +03:00
f637ebe9ff
Minor: formatting for lambda arguments
2017-08-07 18:00:14 +03:00
b7c894a6d3
[NI] Capture types from smart-cast types.
...
Sometimes we have something like if (a is Foo<*>) a.bar()
where bar declared: fun <T> Foo<T>.bar().
For such case we should create receiver with possible types Capture(*).
2017-08-07 18:00:13 +03:00
b6bb171b67
[NI] Extension function wins member if it's infix call
...
Fixes 'infixFunctionOverBuiltinMember'
2017-08-07 18:00:12 +03:00
8670d2abba
[NI] Propagate information about safe call to receiver argument
2017-08-07 18:00:11 +03:00
72d14bfe0d
[NI] Record type of callable reference to trace
2017-08-07 18:00:10 +03:00
a303888e66
[NI] Fix expected return type for function literals
...
It should actually be Unit, as told in FunctionExpression comments
(was null, later defaulted to Any?).
2017-08-07 18:00:09 +03:00
0136be3500
[NI] Fix traversing delegated constructor call chain
...
Use original descriptor instead of possibly substituted one: even though
the substitution is always trivial here, delegated constructor call is
recorded for the original constructor descriptor. So the code that
traverses the delegated constructor call chain should use original
descriptor, too.
2017-08-07 18:00:08 +03:00
0634025229
[NI] Propagate known type substitutor to resolution part
...
This commit fixes 'functionNtoStringGeneric' and 'superConstructor'
2017-08-07 18:00:07 +03:00
c6fcbf6172
[NI] Check for unwrapped descriptor when generating accessible function
...
In new inference we get copy of sam constructor and therefore should check original descriptor in BE.
This fixes 'protectedSamConstructor' test
2017-08-07 18:00:06 +03:00
e12eb66572
[NI] Fix COERCION_TO_UNIT in NI.
2017-08-07 18:00:05 +03:00
de17604fe7
[NI] Bind all new resolved calls before update arguments.
2017-08-07 18:00:04 +03:00
42438bd363
[NI] Do not calculate lambda return type in Visitor if NI enabled.
2017-08-07 18:00:03 +03:00
85248676d2
[NI] Update lambda result arguments after resolution.
2017-08-07 18:00:02 +03:00
b9e9243e9d
[NI] Add hacks for property delegation resolution.
2017-08-07 18:00:00 +03:00
af138e7403
Minor. Extract object to inner class.
2017-08-07 17:59:59 +03:00
5bb61d2d02
[NI] Small improvements in callable reference resolution.
...
I'm not sure in this commit, but it fix some test -> let it be for now.
And yes, now I do not know correct way for callable reference resolution,
so for now it is just proposals.
2017-08-07 17:59:58 +03:00
7eabdeffb3
[NI] Fix variable to proper upper type even if direction is TO_SUBTYPE.
2017-08-07 17:59:57 +03:00
b3be21146c
[NI] Add more hacks.
2017-08-07 17:59:56 +03:00
0f0d834c23
[NI] Minor. Fix IOE.
2017-08-07 17:59:55 +03:00
55181541af
[NI] Change lambda analysis -- create arguments for return statements.
2017-08-07 17:59:54 +03:00
58f73bd82a
[NI] Use right DFI for arguments smart-casts.
2017-08-07 17:59:53 +03:00
30f7396803
[NI] Substitute lambda return type with current substitutor
...
We should do this because return type even it still not proper
can contain fixed type variables
2017-08-07 17:59:52 +03:00
89ac3bd5cf
[NI] Do not add useless constraints to constraint system
...
Skipped constraints:
- T <: Any, T >: Nothing
- T <: T, T <: T?, T >: T, T <: T!, T! <: T
2017-08-07 17:59:51 +03:00
3a25405088
[NI] Remove type variables for lambda arguments -- use existing instead.
2017-08-07 17:59:50 +03:00
ff6a28b64c
Add util method to BuildIns is kotlin.Function type.
2017-08-07 17:59:49 +03:00
acc6e48172
[NI] Added isSuspend for lambdas.
2017-08-07 17:59:48 +03:00
29e2a26ad6
[NI] Write fake ResolvedCall for callable reference.
2017-08-07 17:59:47 +03:00
Dmitry Petrov