On Fri, Dec 9, 2016 at 3:52 AM, Rene Gollent <anevilyak@xxxxxxxxx> wrote:
On Dec 8, 2016 4:45 PM, "Dario Casalinuovo" <b.vitruvio@xxxxxxxxx> wrote:
In the case of the crash, the node was itself timesource and since the
media_kit has that confused pseudo-locking strategy....it happens from time
to time that the timesource is being updated (PubblishTime()) while the
ControlLoop() is calling RealTimeFor().
That implies that the locking is still broken, and you've simply changed
the timing during which the race condition can occur without actually
fixing it.