From: Brian D H. <bdh...@c4...> - 2001-12-28 02:48:24
|
Chuck, What I'm seeing are interlaces of what should be a seperate packet following positions from KC7O-9. My thought is that it is somehow crashing the packet parser. I need to go back through the source and see where we discard overly long packets. I'm guessing that the IGATE taking the packets off RF and inserting them on the Inet stream has MIC-E conversion turned on. You can see that it's a MIC-E station, but it's coming across as full posit packets. THX/BDH On Thu, 2001-12-27 at 20:11, Chuck Byam wrote: > On Thursday 27 December 2001 07:48 pm, Brian D Heaton wrote: > : Attached is a clip of some interesting packets coming across the stream > : right now. Is it possible the "/" at the end of the packets is causing > : a problem??? > : > : 73/N5VFF > : > : ------------- LOG CLIP---------------------- > : KC7O-9>ST1UYS,WIDE3-2:`.7qqzv>/ > : KC7O-9>APRS,WIDE3:@280034z3415.91N/11827.85W>090/059/Mic-E/M1/Enroute>/ > > I don't know if there are "killer packets" coming across, but I too see this > on 'first' as well. It seems that it runs well for a day or so then will > start dying for no apparent reason with the pattern noted. I too have been > trying to capture something, anything, that would help to explain this, but > have yet been able to explain or capture anything (packets or debugging info) > that has been helpful. > > On a side note; I'll post my current development code that 'first' has been > running for the past couple weeks. It includes a number of changes and > several new classes (exception, mutex, assertion, and a redone cpqueue > class). I've also made a change in configure to include Dale's -pthread flag > suggestion. I haven't seen any improvement in performance but as Dale noted, > those "in the know" suggest this as opposed to -D_REENTRANT -l/-L. > > I'd really like to redo history and make this a "real" class. I've been able > to demonstrate that many of the segfaults have occurred in or around calls to > it's functions. The new mutex class has all but eliminated spin locks by > making mutex's local within methods, so when the method exits a mutex is > automatically removed if it hasn't already when the method loses scope (kinda > cool). > > Chuck > > _______________________________________________ > Aprsd-users mailing list > Apr...@li... > https://lists.sourceforge.net/lists/listinfo/aprsd-users > |