openjdk6 - FreeBSD 10.0-RELEASE - dies with "exited on signal 6"

[Jung-uk Kim]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2014-01-30 07:53:41 -0500, Kenneth Culver wrote:
> Disabling background compilation fixed my issues.
> 
> On Wednesday, January 29, 2014, Sevan / Venture37
> <venture37 at gmail.com> wrote:
> 
>> On 29 Jan 2014, at 01:13, Huang Wen Hui <huanghwh at gmail.com
>> <javascript:;>> wrote:
>> 
>>> I got the same problem on FreeBSD 10, disable Background
>>> Compilation
>> seems to fixed this problem.
>> 
>> Patched & rebuild openjdk6, I was able to complete the opennms
>> built process so I thought I'd restart it to double check, build
>> process froze with java process in a state of uwait. Aborted
>> after 4 hours & restarted once more.

We are well aware of this problem and trying hard to fix it ATM.
However, debugging process is very slow because I cannot reproduce the
problem at all.  ale@ is providing some core dumps and I am trying to
figure out what's happening.  I know some objects are getting
free()'ed earlier than expected, which is ultimately causing SIGSEGV
or SIGBUS *later*.  So, I need more information.  Please send the
following information *privately*.

grep -B 3 -C 8 CPU: /var/run/dmesg.boot
cd /usr/ports/java/openjdk6 && make -V CC -V CFLAGS -V CXX -V CXXFLAGS

In the mean time, it seems disabling background compilation (-Xbatch)
significantly reduces chance of crashes as you already found out.
Obviously, disabling JIT compilation (-Xint) should always work.

Thanks,

Jung-uk Kim
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (FreeBSD)

iQEcBAEBAgAGBQJS6pM6AAoJEHyflib82/FGV3cIAJUr0Ra1+Dfeuj/cOmKsWfF4
ETn5eZWoN/Fpw+wUbLXyy47WyssL3yEYykXNBhaAKUEoTlFy8wQ6+ovJw83bSaSk
dvxg7ZGfAh/Q8MD8pxRSFnkn+SNVOmu4hVp3RyYFMWnz1ml5MRUF3yt7Da3XIwum
a2Yxn78K1Q1Rq1bbwSS9caRK+UbVH1SzSHTUUc3M+hoLbP08vhJXhENQNEd1YRxc
AxHwM3SO6rX5Ff6jv6bZyylZd6l4GJ8KXxCitlnhWNkyS2vh8iA62scVq8KUiNdt
XNV/P0nKfXPOh+sSPTxRaWoFrFJHzYimaihANWq6X2lO5yz8ee70VWt8akDCWEY=
=GMnm
-----END PGP SIGNATURE-----