Arnt Karlsen writes:
> 
> On Sat, 08 Jan 2005 14:47:20 +0800, Ivan wrote in message 
> <[EMAIL PROTECTED]>:
> 
> > G''day all.
> > 
> > I''ve written a client that drives FG using the native-fdm I/O
> > mechanism.
> > 
> > For the ''time'' variable, I've tried entering zero, and then entering
> > the value returned by (Win32's) GetTickCount() --> no difference.
> > 
> > However, interestingly, I noticed that FG starts off at midnight (in
> > its  internal world time) but the time advances at a phenomenal rate.
> > After a  couple of minutes I actually see the sky lighten up, followed
> > by the sun  rising in the east. Sunset follows about 2 min later.
> > 
> > What gives ??
> 
> ..a _guess_: the 32bit unix calendar ticks over sometime in 2038, 
> while the 32bit Wintendo calendar ticks over every 49? days, 
> I saw this given somewhere on the web as the reason Microsoft 
> used (they still do?) to recommend reboots about that often.

This is true a naive Win32 clock running of of timeGetTime() 
rolls over every 49 days but there are ways to prevent this
although I don't believe the FlightGear clock on Windows checks
for this.

However Ivan's problem is one of order of magnitude

SimGear / timing / timestamp.XXX is where this is spelled
out in code

Cheers

Norman


_______________________________________________
Flightgear-devel mailing list
[email protected]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d

Reply via email to