Carl Trieloff escribió:
That's one for you folks who will actually have to apply the patches
to think about... It seems as though Manuel's Solaris changes are
doing very well and will cause the least upset, so maybe that one can
go first.
Well, not that sure now. The poller seems to work, but once I faked the
mutex initializer to go on even on failure, I found that the test_unit
crashed miserably with both SIGBUS or SIGSEGV, always trying to allocate
memory. Perhaps it's related with invalid mutexes, but I'm not still
sure. We will see once the mutex bug is ironed. On the other hand, the
broker starts and I was able to use a client on it, without problems,
apparently.
-Steve
Ack. will probably need to get a machine up to test that the merge does
not get broken
as we merge. or to start we could just make sure linux does not break,
and then let Manuel
check the Solaris build each time after a merge. I could chase a machine
down - but that
will take a few days. thoughts?
Are you doing x86 or SPARC?
Sparc, 64 bits builds.
Regards.
--
Manuel.
