Small example of livelock regression in garbage collector for GCJ 3.3 under Win32

Mohan Embar gnustuff@thisiscool.com
Tue May 20 17:17:00 GMT 2003


Wait a minute,

>Thread 6 (thread 2644.0x1358):
>#0  0x7ffe0304 in ?? ()
>#1  0x77f7561d in _libwsock32_a_iname ()
>#2  0x00413bd5 in gnu::gcj::runtime::FinalizerThread::finalizerReady()

This seems eerily symptomatic of this:

http://gcc.gnu.org/ml/java/2003-05/msg00187.html

When certain exceptions are thrown, the stack seems to get corrupted.
During my horror with GCJ 3.4, I ended up in _libwsock32_a_iname ()
after the world came to an end. 

Unfortunately, I haven't had time to delve into this.

Ranjit, does your intuition tell you that this could have
anything to do with this?:

http://gcc.gnu.org/ml/java/2003-03/msg00048.html

Can you explain the --enable-sjlj-exceptions issue to me more?
Should we be building with this? It isn't in my build scripts,
which I originally got from Ranjit.

-- Mohan
http://www.thisiscool.com/
http://www.animalsong.org/






More information about the Java mailing list