Coleen,
I would suggest adding these sanity checks:
1063 } break;
1064
case JVM_CONSTANT_UnresolvedClassInError:
case JVM_CONSTANT_MethodHandleInError:
case JVM_CONSTANT_MethodTypeInError:
assert(0, "should have been eliminated by normalize_error_tag");
// fall through
1065 // Invalid is used as the tag for the second constant pool entry
1066 // occupied by JVM_CONSTANT_Double or JVM_CONSTANT_Long. It should
1067 // not be seen by itself.
1068 case JVM_CONSTANT_Invalid: // fall through
1069
1070 default:
1071 ShouldNotReachHere();
1072 break;
1073 }
Thanks
- Ioi
On 10/09/2013 07:04 PM, Coleen Phillimore wrote:
Thanks Dan!
Coleen
On 10/9/2013 6:39 PM, Daniel D. Daugherty wrote:
On 10/8/13 12:27 PM, Coleen Phillimore wrote:
Summary: Add missing cases.
I verified this with the test for
https://bugs.openjdk.java.net/browse/JDK-8022701 but don't want to
copy the test into with modifications the hotspot repository, still
working on refactoring the test for the JDK repository for after
this gets pushed.
open webrev at http://cr.openjdk.java.net/~coleenp/8025185/
src/share/vm/oops/constantPool.cpp
No comments.
Thumbs up.
Dan
bug link https://bugs.openjdk.java.net/browse/JDK-8025185
Thanks,
Coleen
