Please report new issues athttps://github.com/DOCGroup
This bug documents a change in behavior from previous versions of TAO: the ORB used to flush its message queues on shutdown(), the problem is that shutdown is too late, and flushing can take a *very* long time, possibly even dead-lock the ORB because of a flow-controlled peer. IMHO if the application needs some sort of flushing then they should provide it using a no-op operation, or _is_a() or anything similar.
Accepted.