Again a belated report, most of the work involved a pretty evasive bug present on the actual live support in feng (the fenice one forwardported more or less), strangely looks like we are missing a lock somewhere and from time to time _something_ segfaults. Since this bug is hard to reproduce we are wasting lots of time and effort basically trying every debugging facility that comes to mind, as result I updated the configure to support mudflap and duma, so far the bug manages to escape the nets.

Alessandro is still overhauling libnemesi and fixing issues, hopefully once he is done we'll prepare an automated regression/stress test for feng.

Dario is busy with some of his doctorate duty and has almost exposed the seek functionality to the RTSP layer while Alessandro prepared a simple wrapper function to make seeking using libnemesi simple, soon he'll add some safety checks in order to make easier handle seek failures.