Curt,
You may be glad to know that today, after doing a SVN update of my
working copy, a subsequent release build succeeded and all the unit
tests completed successfully, including telnetappendertestcase, with no
hangs.
uname -srvmpio
Linux 2.6.23.14-64.fc7 #1 SMP Sun Jan 20 23:54:08 EST 2008 i686 i686
i386 GNU/Linux
thanks,
-Bernie Stumpf
Curt Arnold wrote at 10:02:
On Mar 26, 2008, at 5:44 AM, Moshe Matitya wrote:
On Friday, March 21, 2008 7:06 AM, Curt Arnold wrote:
There are much better things to do over the next 3 days than
to review a release candidate. I hope to prepare and post
one on Monday.
Is there a new ETA for this?
Moshe
At the moment, I'm in a quandary with ServerSocket. If you noticed,
I just suppressed a TelnetAppender test that was written in the last
week or so. The test is valid, it is just that the implementation of
ServerSocket behaves differently on different platforms and on some
platforms it will deadlock on close.
On Mac OS/X, a call to accept() returns an error if the listening
socket is closed on a different thread. On Linux (Ubuntu 6.06
specifically), a call to accept() does not return but continues to
wait does not return and the specified timeout seems to have no
affect. This results in the call to Thread::join on the
TelnetAppender::close to never return.
If blocking with timeout is not portable, then the accept logic
should be rewritten using non-blocking with polling (which hopefully
is portable). Unfortunately, I'm already fully committed today and
can not work on the issue.
