-
Bug
-
Resolution: Duplicate
-
P3
-
9
-
x86
-
os_x
FULL PRODUCT VERSION :
java version "1.9.0-ea"
Java(TM) SE Runtime Environment (build 1.9.0-ea-b74)
Java HotSpot(TM) 64-Bit Server VM (build 1.9.0-ea-b74, mixed mode)
ADDITIONAL OS VERSION INFORMATION :
Darwin x86_64
A DESCRIPTION OF THE PROBLEM :
The included program has different behaviour when compiled with the jdk9 javac than with earlier versions.
Specifically, javac9 resolves the invocation inside f() as g(Object...), and javac8 resolves g(M[]).
This is probably related to https://bugs.openjdk.java.net/browse/JDK-8039214.
REGRESSION. Last worked in version 8u60
ADDITIONAL REGRESSION INFORMATION:
java version "1.8.0_60-ea"
Java(TM) SE Runtime Environment (build 1.8.0_60-ea-b13)
Java HotSpot(TM) 64-Bit Server VM (build 25.60-b12, mixed mode)
STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
$JAVA_8_HOME/bin/javac Test.java
$JAVA_8_HOME/bin/java Test
one
$JAVA_9_HOME/bin/javac Test.java
$JAVA_9_HOME/bin/java Test
two
EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
I expected the same behaviour from both compilers.
ACTUAL -
The program behaved differently when compiled with javac9.
REPRODUCIBILITY :
This bug can be reproduced always.
---------- BEGIN SOURCE ----------
public class Test {
interface M<T> {}
public static void main(String[] args) {
f((M<Object>) null);
}
static <T> void f(M<? super T>... xs) {
g(xs);
}
public static <E> void g(M<? super E>... xs) {
System.err.println("one");
}
public static <E> void g(E... xs) {
System.err.println("two");
}
}
---------- END SOURCE ----------
CUSTOMER SUBMITTED WORKAROUND :
Explicit type parameters:
static <T> void f(M<? super T>... xs) {
Test.<M<? super T>>g(xs);
}
java version "1.9.0-ea"
Java(TM) SE Runtime Environment (build 1.9.0-ea-b74)
Java HotSpot(TM) 64-Bit Server VM (build 1.9.0-ea-b74, mixed mode)
ADDITIONAL OS VERSION INFORMATION :
Darwin x86_64
A DESCRIPTION OF THE PROBLEM :
The included program has different behaviour when compiled with the jdk9 javac than with earlier versions.
Specifically, javac9 resolves the invocation inside f() as g(Object...), and javac8 resolves g(M[]).
This is probably related to https://bugs.openjdk.java.net/browse/JDK-8039214.
REGRESSION. Last worked in version 8u60
ADDITIONAL REGRESSION INFORMATION:
java version "1.8.0_60-ea"
Java(TM) SE Runtime Environment (build 1.8.0_60-ea-b13)
Java HotSpot(TM) 64-Bit Server VM (build 25.60-b12, mixed mode)
STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
$JAVA_8_HOME/bin/javac Test.java
$JAVA_8_HOME/bin/java Test
one
$JAVA_9_HOME/bin/javac Test.java
$JAVA_9_HOME/bin/java Test
two
EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
I expected the same behaviour from both compilers.
ACTUAL -
The program behaved differently when compiled with javac9.
REPRODUCIBILITY :
This bug can be reproduced always.
---------- BEGIN SOURCE ----------
public class Test {
interface M<T> {}
public static void main(String[] args) {
f((M<Object>) null);
}
static <T> void f(M<? super T>... xs) {
g(xs);
}
public static <E> void g(M<? super E>... xs) {
System.err.println("one");
}
public static <E> void g(E... xs) {
System.err.println("two");
}
}
---------- END SOURCE ----------
CUSTOMER SUBMITTED WORKAROUND :
Explicit type parameters:
static <T> void f(M<? super T>... xs) {
Test.<M<? super T>>g(xs);
}
- duplicates
-
JDK-8159680 Inference failure with unchecked subtyping and arrays
- Closed
- relates to
-
JDK-8033718 Inference ignores capture variable as upper bound
- Closed
-
JDK-8039214 Inference should not map capture variables to their upper bounds
- Closed