Commit Graph

31 Commits

Author SHA1 Message Date
Mikhael Bogdanov
4c59d161d4 Support basic reification in IR 2019-04-08 13:10:22 +02:00
Steven Schäfer
29b7da7c49 Implement support for KClass fields in annotation classes. 2019-03-26 11:54:16 +01:00
Ilmir Usmanov
39de3c2ab1 Fix test data on JVM_IR BE 2019-02-01 21:45:44 +03:00
Ilmir Usmanov
20e4efaeab Fix test data 2019-02-01 17:33:49 +03:00
Mads Ager
bb0cd92da0 JVM_IR: Fix codegeneration for missing branches.
This used to generate an ACONST_NULL instruction leading to incorrect
stack height for code such as:

```
if (condition) else 32
```
2019-01-28 15:41:48 +01:00
Mads Ager
690b8e0ac9 JVM_IR: Optimize null checks.
Introduce lowering to remove null checks for primitive type
expressions and replace them with true/false. Side-effects
are preserved.

Generate ifnull/ifnonnull instructions for null checks instead
of materializing a null literal for an equality check.
2019-01-19 09:43:43 +01:00
Dmitry Petrov
08abf5f2a2 Specialize unsigned range and progression iterator calls 2019-01-16 11:00:23 +03:00
Mads Ager
864b90f8c0 JVM_IR: Do not materialize negated boolean for branches.
Instead, flip the branch targets. This generates java byte code
such as:

   L2
    IFNE L3
    ALOAD 0
    INVOKEVIRTUAL A.getX ()F
    GOTO L4
   L3

instead of:

   L2
    IFNE L3
    ICONST_1
    GOTO L4
   L3
    ICONST_0
   L4
    IFEQ L5
    ALOAD 0
    INVOKEVIRTUAL A.getX ()F
    GOTO L6
   L5
2019-01-03 08:21:31 +01:00
Ilmir Usmanov
871134cff8 Update test data
#KT-28309
2018-12-28 14:18:39 +03:00
Mads Ager
3a11322506 Enable bytecode text tests for the JVM_IR backend. 2018-12-21 16:20:45 +01:00
Alexander Udalov
f5ff3d2fa9 Remove directives that have no effect from bytecode text tests
All bytecode text tests are run with stdlib in the classpath and only
for JVM backend, therefore directives WITH_RUNTIME, TARGET_BACKEND,
IGNORE_BACKEND are not needed
2018-12-20 12:53:24 +01:00
Dmitry Petrov
65881dda97 Fix tests after rebase 2018-09-05 12:20:57 +03:00
Mikhail Zarechenskiy
ec9d8e580e Leave boxing for compareTo/areEqual methods for inline classes
Inline classes can override methods and thus introduce side effects
2018-07-20 11:51:34 +03:00
Mikhail Zarechenskiy
01d9be65bc Perform bytecode optimisations for inline classes
#KT-23742 Fixed
2018-06-18 11:52:36 +03:00
Dmitry Petrov
39aa97eebf Use Intrinsics#compare when specializing compareTo for ints 2017-05-22 11:57:20 +03:00
Dmitry Petrov
e1b41eee15 Specialize Comparable#compareTo for boxed primitives
Support Comparable#compareTo for boxed primitive in redundant
boxing/unboxing analysis, along with CHECKCAST to java.lang.Comparable.

Note that we can do that for Float and Double, too, because
Float#compareTo(Float) and Double#compareTo(Double) are delegated to
Float#compare(float, float) and Double#compare(double, double),
respectively.

Fuse specialized comparison for integers with conditional jumps
if possible (both for Comparable#compareTo and Intrinsics#areEqual).

 #KT-11959 Fixed
2017-05-22 11:57:20 +03:00
Dmitry Petrov
1378b0cf05 Fix bytecode tests after new optimizations
- Turn some const conditions into non-const conditions
- Make sure inlined const values are used where required
(otherwise they are eliminated by POP backward propagation)
2017-05-16 17:28:43 +03:00
Dmitry Petrov
495fba43c0 Fuse primitive equality with safe call to avoid boxing
In code like 'a?.b == 42', we can immediately generate equality
comparison result when receiver is null (false for '==', true for '!='),
since the primitive value is definitely non-null.
Otherwise unnecessary boxing/unboxing is generated to handle possibly
null result of 'a?.b'.
2017-05-16 17:28:43 +03:00
Dmitry Petrov
c46164481a KT-15871 Unnecessary boxing for equality operator on inlined primitive values
Allow kotlin.jvm.internal.Intrinsics#areEqual for boxed values.
Rewrite to primitive equality.

NB we can't do that for Float and Double, because java.lang.Float#equals
and java.lang.Double#equals behave differently from primitive equality comparisons.
2017-03-13 09:04:31 +03:00
Dmitry Petrov
a087ea559f Eliminate redundant CHECKCAST instructions
CHECKCAST is redundant if the corresponding static type exactly matches the target type.
CHECKCAST instructions to-be-reified should not be eliminated.

KT-14811 Unnecessary checkcast generated in parameterized functions
KT-14963 unnecessary checkcast java/lang/Object
2017-03-13 09:04:31 +03:00
Dmitry Petrov
e0cea468fa KT-15862 Inline generic functions can unexpectedly box primitives
Previous version of the boxing/unboxing analysis treated merging boxed and non-boxed values as a hazard.
If such merged values are not used (e.g., early return + local variables reused in inlined calls),
corresponding boxing/unboxing operations still can be optimized out.

All information related to boxed value usage by instructions is moved to 'BoxedValueDescriptor'.
Introduce "tainted" (and "clean") boxed values, with the following rules:

  merge(B, B) = B, if unboxed types are compatible,
                T, otherwise

  merge(B, X) = T

  merge(T, X) = T

  where
    X is a non-boxed value,
    B is a "clean" boxed value,
    T is a "tainted" boxed value.

Postpone decision about value merge hazards until a "tainted" value is used.
2017-02-06 16:22:26 +03:00
Dmitry Petrov
3bbd8979e4 KT-9670: optimize Class <-> KClass wrapping/unwrapping as a special case of boxing/unboxing.
NB doesn't work for arrays of classes.
2016-03-04 10:14:21 +03:00
Denis Zharkov
914447b7eb Do not treat uninitialized value as a reason to retain boxing
See testData/simpleUnitializedMerge.kt
On exit from `if` we merge value in variable with slot 1 (x):
- from `if` body we get BoxedBasicValue
- from outer block we get UNITIALIZED_VALUE

So we just suppose `x` is unitialized after `if`
and there's no need to mark BoxedValue as unsafe to remove
because it's anyway can't be used after `if`

 #KT-6842 Fixed
2016-02-25 08:31:19 +03:00
Michael Bogdanov
b9c30c739e Test for obsolete KT-7224: Redundant result boxing on ?. call on a primitive
#KT-7224 Obsolete
2016-01-29 12:13:16 +03:00
Ilya Gorbunov
4d5ec9be3f Drop identityEquals from builtins, compiler and tests. 2016-01-22 05:54:38 +03:00
Ilya Gorbunov
5e6d16103e Remove deprecated ranges from tests and take into account that Byte and Short rangeTo now returns IntRange. 2015-11-11 03:53:31 +03:00
Dmitry Jemerov
1523d5bcbf removing static type assertions work in progress 2015-10-12 11:11:23 +02:00
Michael Bogdanov
841b199d49 Optimize comparison with zero and null 2015-04-07 17:28:01 +03:00
Denis Zharkov
654411a0b0 Refactored tests using Array constructor:
Some moved to tests with stdlib
Some changed to use arrayOfNulls
2014-12-11 16:04:03 +03:00
Denis Zharkov
da159c1e53 Simple version of redundant GOTO's removing 2014-09-26 12:37:33 +04:00
Denis Zharkov
0f5e29df9b tests added 2014-07-15 21:26:53 +04:00