On Mon, Feb 20, 2012 at 10:17 PM, Jerry Lundström <[email protected]> wrote: > I think it might be MutexFactory::i() that bangs, even if you dont see > it in the back trace since it could be optimized away.
The auto_ptr() is destroyed during atexit(). It may e.g. be that it does not set it to NULL. Thus the invalid pointer. The save way is not to have any dependencies between two different auto_ptr() objects during their destruction phase. If there is a need for a dependency, then it should be solved in a different way. E.g. a parent singleton object which takes care of the destruction order or, as in this case, move the MutexFactory object into the SoftHSMInternal since it is the only one who uses it. // Rickard _______________________________________________ Opendnssec-user mailing list [email protected] https://lists.opendnssec.org/mailman/listinfo/opendnssec-user
