How is long-running software supposed to deal with leap seconds that are decreed after the software was started, or even built?
OK, but then in what format should "system time" be stored? In NTP-synchronized "UTC time" that behaves erratically during leap-seconds? In "TAI", except computed from system UTC with maybe-wrong TAI-UTC adjustment?
-
-
And then, "system time" becomes even more "interesting" when various components (hwclock, kernel, libc, daemons, processes started at different times, etc.) each assume different versions of the leap second database, yet all pretend to talk in "unix time".
-
I have no idea about fixing time in the Unix world. Practicality aside, I'm imagining a simple system that only knows time as an integer: milliseconds after start. Complexities like leap second databases would belong outside of that, in the UI zone, a separate process.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
Read my blog!