From: cwilson@Xenon.Stanford.EDU (Christopher S. Wilson) Subject: Re: Wrong timestamp on DOS [FAQ editor: please note] Date: Sun, 14 Feb 1993 11:19:43 GMT
In article <Feb.13.15.18.59.1993.4525@geneva.rutgers.edu> hedrick@geneva.rutgers.edu (Charles Hedrick) writes:
[...]
>cwilson@Xenon.Stanford.EDU (Christopher S. Wilson) writes:
[I explain how Linux interprets my mounted DOS filesystem
timestamps incorrectly as Universal Time, not local time and ask
for an explanation.]
[Charles Hedrick explains that the kernel doesn't include all the
code for handling timezones that applications get in a standard
library because it would a lot to put in the kernel.]
[He goes on to explain how to use the -d and -t options of date can
be used to give the kernel timezone info and the problems that
arise from that.]
Thanks for the explanation and possible solutions. Unfortunately, I
don't have -d or -t options on my date command. I'm using the current
version of SLS from tsx11.mit.edu and I found that it gave me two
different date commands, /bin/date and /usr/bin/date. /usr/bin/date
doesn't recognize either. /bin/date doesn't recognize -t at all and
it thinks -d is telling it there is a date string following. Both of
these binaries are timestamped September. [Are you reading this
Peter? Can I assume that one of these binaries is in there by
mistake?]
Is there a more recent version of date I should be using?
I agree that having the CMOS clock on Universal time is a much better
idea if it weren't for DOS. I'm just a bit reluctant to have DOS
running with the CMOS clock set that way. You never know when some
badly behaved program is going to look at the CMOS clock for itself
instead of going through DOS. But I think that's what I'll end up
doing.
Again, thanks, I really do appreciate all the info.