Received mail timestamp is off by 7 hours

[Ian Smith]

On Wed, 2 Mar 2005 03:11:19 -0800, Ted Mittelstaedt wrote:

Loren wrote:

 > > little bit less reliable using local to UTC unless you are not affected
 > > by any daylight savings changes like Arizona in the US or, I'm
 > > sure, many
 > > other places around the world.

For a desktop or test machine, booting multiple OS at various times like
you said, I'd as likely run UTC CMOS too.  Enough to confuse already .. 

I've just some fun browsing again through bits of code from /usr/src
like sbin/adjkerntz/adjkerntz.c, sys/i386/isa/clock.c, sys/isa/rtc.h,
sys/i386/include/clock.h, just to scratch the surface.  Great stuff.

I'm yet to be convinced there's an even potential technical problem in
running local time RTC on a server that runs FreeBSD all the time.  I
won't labour personal preferences further, but don't feel vulnerable.

Yes, we do DST changes; I'd guess well under a millisecond twice a year.
Hey, you can lose your bootsector if the sky falls at the wrong instant! 

Ted writes:

 > There's no excuse for a mailserver to not be synced to a NTP source.

Absolutely, or indeed most anyserver.  Or even a multiboot workstation!

Ian out.