Re: Proposal for new ISO C 9x time API
Ever watched xntpd try to handle a leap second? Wobble, wobble, wobble, wobble, wobble.
NTP on an "Olson-right" system, with the addition of a single time2posix() call, sails smoothly through a leap second by ignoring the leap warning indicator and all its convoluted logic.
Bradley White writes:
NTP on an "Olson-right" system, with the addition of a single time2posix() call, sails smoothly through a leap second
Yup. The same is true of sntpclock, my libtai-based SNTP client, with no code changes. See http://pobox.com/~djb/clockspeed.html.
by ignoring the leap warning indicator and all its convoluted logic.
If you want to use NTP _during_ a leap second then you have to pay attention to the indicator. sntpclock does this, though I haven't had a chance to test the code. Of course, there's a good chance that the server has been screwed up by xntpd, so it's better to run on autopilot for a while. ---Dan 1000 recipients, 28.8 modem, 10 seconds. http://pobox.com/~djb/qmail/mini.html
participants (2)
-
Bradley White -
D. J. Bernstein