From: Jay Ts <jay@jayts.cx>
Date: Mon, 14 Feb 2000 20:41:18 -0700 (MST)
A little while ago I reported an apparent y2k bug in the Linux kernel.
Alan Cox replied, saying that the problem of the year being reported
as 1910 couldn't possibly be due to the kernel, and that perhaps some
buggy application was setting the system date, causing it to propagate
outwards from there (i.e., bad values from time() and gettimeofday()).
I thought about it a few minutes ago realized that I might have used the
'nist' command (which querys the National Institute of Standards and
Technology's atomic clock over the Internet) to set the date on both of
those systems. And so I ran nist, and guess what it reported?
The root cause of this was (relatively) widely known. The issue is that
the NIST signal sends the date in two forms; one is a YYMMDD format, and
the other is in Julian days. The Julian day information is obviously
Y2K compliant, while the YYMMDD format isn't. Unfortuantely, it's a
pain to convert from Julian days to YYMMDD, so a lot of programs ---
including the sample programs distributed by NIST, opted to use the
YYMMDD, often with bad results on Y2K day.
- Ted
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:29 EST