Re: problem with sntpclock
Alex Stevens <[email protected]> Thu, 11 Jul 2002 14:35:13 -0500
| Newsgroups | gmane.comp.lib.libtai |
|---|---|
| Message-ID | <[email protected]> |
On Wed, Jul 10, 2002 at 06:27:37PM +0000, Thorsten Glaser wrote: > begin electrogrammati illius D. J. Bernstein > > >Thorsten Glaser writes: > >> Without the rdate patch AND the leapsecs file, the Unix clock is set to UTC, > > > >No. The time_t used by xntpd and other leap-second-ignorant applications > >is not UTC. It jumps backwards. It cannot be reliably converted to UTC, > > Uh, sorry, yes, I forgot that. Is "pseudo-UTC" or "NTP" better? xntpd notwithstanding, NTP is reset by UTC. Thus, although NTP has no history of leapseconds since its timescale is reset when UTC leaps, NTP time is not 'pseudo-UTC' at all. In view of the different ways time is handled in the UNIX world, NTP's strategy of letting the OS/apps determine leapseconds is not completely insane. -- __________________ [email protected] ------------------ "Yeah, I'm drunk all right, and you're crazy. Tomorrow I'll be sober, but you'll still be crazy for the rest of your life." -W.C. Fields