Re: What TZ do you use to get right time w/ clockspeed?
Bennett Todd <[email protected]> Thu, 14 Nov 2002 14:33:11 -0500
| Newsgroups | gmane.comp.lib.libtai |
|---|---|
| Message-ID | <[email protected]> |
Thanks for the tips. I don't disagree that the system clock as set by clockspeed is right, and that the clock as set by ntp is wrong, in a fundamental sense of usability. However, it's not helpful when date(1) reports a time 22 seconds faster than (e.g.) date(1) run on an ntp-running system, or (e.g.) a GPS receiver, or whatever. There _is_ a notion of "what time it is" that's got real-world meaning to finer tolerance than 22 seconds:-). My problem turns out to be that as-shipped, Red Hat sets up for deliberately broken, POSIX-compatible handling of leap-seconds, where the system clock is supposed to slither around to track them. But it looks to me like simply setting the system-wide timezone to "right/US/Eastern" rather than "US/Eastern" fixes the problem. I don't know how widespread this problem is on systems, nor what fixes are required in different settings. Would this be a topic suitable for this list? And if we collect a range of correct answers for different platforms, perhaps that would be a document djb could put up alongside the clockspeed package; it caused me some embarrassment when I recommended clockspeed on a mailing list, and someone immediately came back and explained that it told the wrong time, and it took me a while to find the explanation. -Bennett
signature.asc
(application/pgp-signature, 189 B)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE90/p3HZWg9mCTffwRAlTCAKC+kSQ6GJdUEZL6PcLCrNhkreBdlQCg26zP P6zde/vsXEH4RLLOxsn8PFY= =T0Z/ -----END PGP SIGNATURE-----