Autogen Exclusions

[arno 1]

Haven't had the time to test it yet, but I have made a piece of scenery where I have all (ground)polygons in one Area and each have an own RefPoint with a tight v2 value. I haven't noticed any of these effects here. So I think it is interesting to have a look at that scenery again and see if I can produce the same results. I'll come back about it :).Arno

---

Member Netherlands 2000 Scenery Team[link:home.wanadoo.nl/arno.gerretsen]Arno's FlightSim World for scenery design hints, tips and other tricks...

[Guest GerrishGray]

Thanks DickAll very interesting. It all goes to show, more and more, that BGL is not really interpreted as a directly executable 'procedural' language and one cannot necessarily rely on the expected behaviour of, for example, jumping to a label.The SDK does describe how the end effect of drawing routines is to place objects into a list of active objects for display, and how this list is sorted for use by the painters algorithm. It also tells us some other similar information. But so many other details of how the BGL language is actually interpreted and implemented are left out or presented in a misleading way ...To be fair to MS, they are rather caught between the devil and the deep blue sea. They want to create a vaguely open architecture that allows third party add-ons because they know this is an important part of the success of FS, but at the same time they have to protect their intellectual property against piracy - not easy to strike a successful balance, I would imagine.Gives us an interesting hobby experimenting with it all, though, eh!To return to Autogen exclusion, what you say about the exclusion effect of some objects persisting even when the object is not displayed on screen shows that the alleged 'jump to the :fail label' is not really what happens at all. The engine still knows about the object that has been jumped over and still uses it for Autogen exclusion, it seems, even though it is removed from the display stack. It looks as though the code might be scanned separately for Autogen exclusion effects, and this could be consistent with the other rather strange behaviours such as the use of the dummy call.By the way, my own testing that revealed the 'not to spec' behaviour of multiple v2 checks did not expose the Autogen effect because I was using 3D objects with little Autogen exclusion effect anyway. The original objective of my test was to confirm (successfully, as it happens) that a large area with multiple objects having separate RefPoint's would not exclude Autogen over the whole area and the discovery of the incorrect v2 effect was just a side effect that I wasn't expecting!The never-ending story goes on and on ...CheersGerrish