On Mon, Feb 06, 2012 at 02:59:03PM +0100, Nicolas Cellier wrote:
> 2012/2/6 David T. Lewis <lewis at mail.msen.com>:
> >
> > This still leaves us with the problem of converting from local time to UTC
> > in the image, which is problematic during a daylight savings time transition
> > (two values of the UTC seconds clock map to the same local seconds clock
> > value, which one is correct?). It is better to use UTC because translating
> > from UTC to local seconds is unambiguous.
> >
> 
> No, if the primitive also fill the localOffset, then you'll get twice
> the same wallClockTime, but with a different localOffset.
> In this case, conversion to UTC is trivial.

Apologies, you are right of course.

> > I think it is best to leave the complexity of translating to julian
> > day, seconds, etc on the image side rather that put it in the VM, unless
> > there is a measurable performance advantage to putting it in the VM.
> >
> 
> I don't know. The julianDayNumber sounds like assembler to me ;)
> 	
> 	p := (monthIndex - 14) quo: 12.
> 	q := year + 4800 + p.
> 	r := monthIndex - 2 - (12 * p).
> 	s := (year + 4900 + p) quo: 100.
> 
> 	julianDayNumber :=
>  		( (1461 * q) quo: 4 ) +
> 			( (367 * r) quo: 12 ) -
>  				( (3 * s) quo: 4 ) +
>  					( day - 32075 ).
> 
> Which date would overflow? Not that soon I guess.
> 
> Also, I'm not sure that duplicating the time zone and DST handling in
> the image is a good idea.
> What advantage over OS would this provide?

If you do it in the image you can get it done right now, and if you
make a mistake you can fix it. If you do it in the VM, you will wait
for everyone involved to come to agreement, implement the code, and
distribute supported VMs.

As far as time zone and DST handling, you cannot realistically do
this in the VM except for the specific case of reporting the current
time, which involves nothing more than reporting the offset from UTC
in the current time zone at the current time. If the platform is
not able to provide this support, then you just report an offset
of zero.

Dave