It's been suggested on time-nuts that (presumably in the general case) clock_combine is a no-op and that sys_offset is determined by a single source -- speer I assume. Either by design or because of some error.
Is this correct (incorrect)? If so has it always been correct (incorrect)? Certainly the documentation, the code and looking at a debug log suggest to me that clock_combine is doing what one would expect. I do see that "prefer" overrides clock_combine so the very common PPS system would use a single clock offset. -- Paul _______________________________________________ questions mailing list [email protected] http://lists.ntp.org/listinfo/questions
