Re: [Flightgear-devel] Re: Next FlightGear Release - upcoming.

[Durk T. <d.t...@xs...>]

On Friday 10 March 2006 14:38, Melchior FRANZ wrote:
> * Melchior FRANZ -- Thursday 09 March 2006 19:43:
> > (D) "-1000 ft" crash. Don't remember the exact message. Doesn't
> >     seem related to any "special" subsystem.
>
>  "Attempting to schedule tiles for bogus lon and lat  = (-1000,0)
>   This is a FATAL error.  Exiting!"
>
> triggered by FGTileMgr::schedule_needed(). Which process sets
> lon=-1000 and lat=0 and requests a tile for that?
>
> m.
>
I'm able to trigger this error message quite consistently by enabling traffic 
manager/AImodels. If the first few AIModels that are required to be loaded 
are non-existent, FlightGear crashes with this error message (always 
reporting the same coordinates, as mentioned above).

Sofare, I've only been able to make an educated guess that generating an error 
condition in the main thread before or at the time the tile loader is running 
triggers this message and a subsequent abort. I'm still in the dark what's 
actually going on (possibly triggered by setting errno????).

FWIW, once scenery loading is complete, the FlightGear main thread/ AIModels 
subsystem happily ignores the loading of non-existent AIModels, as desiged. 
Fixing this, or at least resolving the interaction between any error 
condition set by AIModels in the main thread and an abort in the tile loader 
thread is currently at my todo list. But as I mentioned above, I don't really 
have a clue right now.

Cheers,
Durk