No worry. Once a bug is identified, it can be fixed. :)
Anyway, I've found the work around i.e. by increasing the threshold
using -Dsun.reflect.inflationThreshold=100000000.
On 11 Aug 2009, at 22:28, Rémi Forax wrote:
Le 11/08/2009 21:15, Wei Yin Teo a écrit :
It doesn't work. :(
java.dyn.NoAccessException: receiver must be in caller class:
InstanceTest/27271057.<init>()void, from test.AnonymousClassLoading
at sun.dyn.MemberName.newNoAccessException(MemberName.java:421)
at java.dyn.MethodHandles.unreflectImpl(MethodHandles.java:496)
at java.dyn.MethodHandles
$Lookup.unreflectConstructor(MethodHandles.java:408)
at
test.AnonymousClassLoading.testLoading(AnonymousClassLoading.java:50)
This is another bug :(
Not your lucky day :)
This one is funny because this is a bug totally in the mind of the
programmer.
There is a mismatch, the reflection API says "constructor dot invoke"
which for the VM means "call <init>()" but it means create a new
instance.
So here, the code tries to just called the <init> method
and failed because you are not in caller class.
To summarize, unreflectConstructor() is currently broken :(
Rémi
_______________________________________________
mlvm-dev mailing list
mlvm-dev@openjdk.java.net
http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev
_______________________________________________
mlvm-dev mailing list
mlvm-dev@openjdk.java.net
http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev
