On Fri, 2004-10-22 at 16:42, Dalibor Topic wrote:
With no patches or all but #2, I have 12/144 regression test failures. With #2 thrown in the mix, I'm down to 4 errors and I can run my
classes. (no printf diffs were needed, since fixing arg passing means
printf doesn't get absurd arguments on a regular basis).
That sounds quite good. What are the remaining failures you get?
Preempt.java: thread 0 thread 1 thread 2
TestSerializable2.java: a = Serialized-0 magic 321 java.lang.NullPointerException a = Serialized1-10 magic 321 java.lang.NullPointerException a = Serialized2-10 magic 321 java.lang.NullPointerException
SoTimeout.java: Failure: Time out.
GCTest.fail: Internal error: caught an unexpected exception. Please check your CLASSPATH and your installation. java/lang/NullPointerException
I'm still getting the ocassional GCTest/WaitTest failure on x86-linux, so GCTest one may not be amd64 specific.
The other three don't happen on x86-linux for me, so they might be interesting. The TestSerializable2 one looks like it could be promising. You can use -vmdebug ELOOKUP to see where the NullPointerException is being created from and how it's propagated through the stack.
cheers, dalibor topic
_______________________________________________ kaffe mailing list [EMAIL PROTECTED] http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
