[mythtv] LiveTV bug
bjm at lvcm.com
Sun Mar 2 11:52:45 EST 2003
Matt Zimmerman wrote:
> On Sat, Mar 01, 2003 at 09:28:41PM -0800, Bruce Markey wrote:
>>I'd hate to think that we'd need to stop and start every frontend and
>>backend on every machine each time one process dies.
> I'm not arguing that this isn't a bug (it is), but this is young code and
> fatal signals are considered hostile behaviour. Myth will choke on less at
> this point.
> Of course, if that message had come with a patch attached to it, I would not
> have blinked an eye.
Agreed. Before Isaac fixed the bug, the behavior was that
anytime the frontend was in live tv and died (SEGV, logout,
out of swapspace, system crash, etc.) the ringbuffer would
continue to record. I'd interpreted the original post as
'to reproduce the bug' send a kill signal rather than 'Doctor,
it hurts when I do this'.
I certainly didn't think you were arguing against fixing
bugs. I just wanted to add that the frontend will also need
to reconnect if the backend restarts.
More information about the mythtv-dev