Re: [Flightgear-devel] New route manager?

[James T. <zak...@ma...>]

On 9 Oct 2009, at 01:06, Curtis Olson wrote:

> What the original route manager did was constantly compute the  
> heading to the next waypoint and dump that value into /autopilot/ 
> settings/true-heading-deg
> Then the autopilot could be activated in true-heading-follow mode  
> and you would fly right to the destination.

Yep, understood - but I'm not sure it's a feature you should be  
abusing in that fashion :)

The route manager *does* still compute the heading; it's dumping  
values into random other subsystems I have concerns about.

> I just did a cvs update and I see a number of fixes, but I'm still  
> confused about how to make the autopilot follow what the new route  
> manager is doing?
>
> I opened up the gps dialog box and checked "NAV Slave", but when I  
> put the autopilot into NAV1 CDI Course follow mode, it flew a  
> heading of 40 degrees where the proper heading would be closer to 274.
>
> Is there still a missing connection in the code, or is the missing  
> connection in my head? :-)

There are still several missing connections, depending on which  
aircraft you're using. I have more updates pending, but I think I need  
to resolve this mess the 'proper' way (so that aircraft / panels /  
instruments that want to be realistic don't have to do so many hacks  
in Nasal), but also provide a fallback that uses the current behaviour  
that you want.

Just for one example, as I understand it /autopilot/settings/true- 
heading-deg is not 'standard' property - you're hoping that the given  
autopilot (I assume including the generic one) respects this property.  
For the route-manager to make these kinds of assumptions about a  
particular aircraft's autopilot configuration is inherently dangerous...

Give me a few days to think over the best solution that doesn't  
compromise the realistic setup, but keeps the unrealistic-but-quick- 
and-easy approach working :)

James