I beleave there are serious bugs in the new object handling. Example,
trying to compile testsuite/java.net/SocketSendReceiveTest.java:
> ~/japhar/bin/javac SocketSendReceiveTest.java
fatal signal handler called, signal = 11
java frame: sun/tools/javac/SourceClass.noteUsedBy (pc = 7)
java frame: sun/tools/java/Environment.resolve (pc = 129)
java frame: sun/tools/tree/IdentifierExpression.toResolvedType (pc = 50)
java frame: sun/tools/tree/IdentifierExpression.toType (pc = 7)
java frame: sun/tools/tree/DeclarationStatement.checkBlockStatement (pc = 16)
java frame: sun/tools/tree/CompoundStatement.check (pc = 48)
java frame: sun/tools/tree/TryStatement.check (pc = 43)
java frame: sun/tools/tree/Statement.checkBlockStatement (pc = 9)
java frame: sun/tools/tree/CompoundStatement.check (pc = 48)
java frame: sun/tools/tree/Statement.checkMethod (pc = 32)
java frame: sun/tools/javac/SourceField.check (pc = 744)
java frame: sun/tools/javac/SourceClass.checkFields (pc = 694)
java frame: sun/tools/javac/SourceClass.checkInternal (pc = 231)
java frame: sun/tools/javac/SourceClass.check (pc = 73)
java frame: sun/tools/javac/Main.compile (pc = 1043)
java frame: sun/tools/javac/Main.main (pc = 53)
SourceClass.noteUsedBy() opcode 7 is aload_0.
Another test program manged to crash in native frame
java/net/PlainDatagramSocketImpl.datagramSocketCreate() as
object->data was NULL.
BTW: My thread initialization and termination fixes seems to be
working. I'll commit them today.
--
##> Petter Reinholdtsen <## | [EMAIL PROTECTED]
