Kay Ramme will take care of the re-opened issue.

-Stephan

On 12/14/09 11:05, Huaidong Qiu wrote:
Yes, I those are two different UNO remote call session, Could you
please talk about your thought about fix this?

On Mon, Dec 14, 2009 at 5:32 PM, Stephan Bergmann
<stephan.bergm...@sun.com> wrote:
Thank you for the stack, now I see what the problem is.  There is no
recursive call back across the remote bridge (which should have worked).
 Instead, osl_waitCondition (called from cppu_threadpool::JobQueue::enter)
is broken in that it calls PeekMessage, which continues to dispatch Windows
events, which in turn causes the logically blocked UNO thread to make a
second, unrelated remote call, which is not expected to happen per the
design of UNO and consequently leads to disaster.

The PeekMessage was apparently added when fixing
<http://qa.openoffice.org/issues/show_bug.cgi?id=92756>.  I re-opened that
issue now.

-Stephan

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@framework.openoffice.org
For additional commands, e-mail: dev-h...@framework.openoffice.org

Reply via email to