Re: FIX FOR HAMM: timezone problem
>>>>> "DS" == Dale Scheetz <dwarf@polaris.net> writes:
DS> We are working on a report of failure in US/central WRT Daylight
DS> Savings Time, right?
DS> There is one variable we haven't nailed down yet. The hardware
DS> clock can be set either to local time or GMT (UTC). As I
DS> remember, the failure only happens when the clock is set to one
DS> of these two. (Memory says Local Time is the broken one)
Not quite correct. I remember this problem, as I was hit by it :)
The problem was with the SystemV CST6CDT timezone. It has an incorrect
changeover rule. At the time of the switchover, I had my timezone set
to the SystemV/CST6CDT value, and after the switchover it was still
thinking it was set to CST. ntpdate synched the clock correctly with
UTC, but the displayed time was 1 hour off, since it was displaying
CST, not CDT. Switching the timezone to US/Central caused the date to
be displayed correctly.
I just checked, and now CST6CDT correctly gives CDT, so the problem is
an incorrect transition rule in the SystemV/CCST6CDT timezone file.
The hardware clock issue is a red herring. I only read my clock at
boot time, and my machine had not been booted for several days. I do
write the time back immediately after setting the kernel clock with
ntpdate, though.
-Larry
--
Larry Daffner | Linux: Unleash the workstation in your PC!
vizzie@airmail.net / http://web2.airmail.net/vizzie/
Ray's Rule of Precision:
Measure with a micrometer. Mark with chalk. Cut with an axe.
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: