- Previous by thread: Re: bug in timezone???
- Next by thread: RE: bug in timezone???
Hi, guys!
Since I wrote that error message, let me see if I can explain it.
It's been at least 2 years since I wrote that so I may have forgotten
some stuff, but this is what I remember.
This is indeed a OS bug. It was discovered by Charles Labrec (you
still out there, Charles?) back in '97. Basically, the SyncCM
module does a whole lot of converting of dates to strings. There's
a Perl library that does this very efficiently, but it depends on
the results of the localtime(0) call to generate the value of
the epoch.
Unfortunately, there is (was?) a bug in the localtime call which
caused it to improperly handle daylight savings with certain
timezone settings when you get the time for the epoch, so that
causes this Perl library to fail. The root cause is the bug in
the localtime call. As a workaround, we chose to inform the user
instead of trying to fix the math (get the math wrong, hose thousands
of calendars...).
If you have a patch for this, please submit it!!
-Bharat
> -----Original Message-----
> From: Mike Kupfer [mailto:Mike.Kupfer@xxxxxxxxxxx]
> Sent: Tuesday, May 04, 1999 2:53 PM
> To: Dan Mick
> Cc: pilotmgr@xxxxxxxxxxxxxxxxxxxxxxx
> Subject: Re: bug in timezone???
>
>
> > Honestly, I keep asking: what about the message is unclear?
>
> Well, since you ask :-), there are a couple things that
> bother me about
> the message and the FAQ text.
>
> 1. There's no technical explanation of the bug, just a bald assertion
> that it's an OS bug and not a PilotManager bug. Okay, if that's
> really the case, why hasn't this been fixed yet in Solaris and
> whatever OS's provoke the message? Heck, why can't I even find a bug
> for this in the Sun bug database?
>
> 2. The recommended "fix" is to use a timezone like PST8PDT instead of
> US/Pacific. Why on Earth does this matter? On my Solaris 2.6 box
> /usr/share/lib/zoneinfo/PST8PDT and /usr/share/lib/zoneinfo/US/Pacific
> are hard links to each other, yet US/Pacific provoked a complaint when
> I first installed PilotManager, but PST8PDT did not.
>
> So the bottom line is, yes, the message and FAQ are very clear about
> what to do to make the message go away. On the other hand, the two
> issues I mentioned above lead me to wonder if things are as simple as
> the message and FAQ assert. So I'm not at all surprised to see that
> people continue to post queries about this problem.
>
> mike
> --
> Mike Kupfer kupfer@xxxxxxxxxxx
> NFS Engineering Speaking for
> myself, not for Sun.
> --------------------------------------------------------------
> ----------
> ***********************************************************
> * This is a public mailing list! *
> * Please do not publish Sun proprietary information here! *
> * - - - - - - - - - - - - - - *
> * http://www.moshpit.org/pilotmgr *
> ***********************************************************
>
|
|