Anyone know how to work models?

[ducky]

Hi, I know this is a long shot, but if there is anyone out there that knows how to work models, a group of us would very much like some help in fixing a great map.

The map is FETCHBALL v1.0, a perfectly awesome map in every respect, except that there seems to be some problem with the goal models that causes objects (specifically, the ball) to bounce off what appear to be invisible walls.

If anybody could either take the map, fix the models, and UL the fixed version, or at least lay out what sort of steps one needs to take in order to perform the fix, we would be eternally grateful.

[Graydon]

I'll take a look into this.

[Deqlyn]

Make the goal a monster and it destroy the ball and respawn a new one at center court. I never played fetchball though.

[Pyro]

Well Deq, in Fetchball, the models that make up the goal have a sort of invisible wall on one side. So when the ball comes you might hit it and it will bounce off. If someone doesn't fix it another way to work around it is by adding something in front of that section so at least the players know for sure where to aim their ball.

[ducky]

yeah, that might work, but the problem is the invisible wall area is really quite big. Like, maybe 40% of the goal area. Plus it includes weird spots in mid air. I really have no idea what's causing it, besides I guess the models themselves.

[haravikk]

Are you sure it's part of the model itself? It's possible it's actually a bug since the model is like an enclosed box if I remember right? Myth's model collision code is quite primitive, and fairly confusing, certainly not something that'd likely to be fixed for 1.7!

Does this bug prevent the ball entering the model at all? Perhaps you could still detect when a ball /should/ score by adjusting the tests for it?

[Pyro]

It is like an invisible wall on one side of the model, nothing can get through on that section. It is consistant, perhaps just a mistake on whoever made the model for it.

[Graydon]

I've examined both the red and blue goal models in Loathing, Malice, and Meshwork. There are absolutely no triangles that exist in the area the bug is occurring. I messed around with ceiling heights as well for good measure, with no luck. The bug persists.

I have yet to try re-compiling the model files with apathy. New files may yield better results, but somehow I'm skeptical.

I came across some things during my exploration of the tags. It appears Vista had intentions of an 'animating' goal model. I think the idea would've been for the 5 red lights above each goal to light up and blink. If nobody has objections, regardless of whether this goal bug gets resolved or not, I'd like to implement this feature and put out a v1.1 update.

The other thing I noticed, is that the 'glass' panes, and the big banners on top of each goal, are only visible from one side. One would think you should be able to see them from both sides. Ducky: would you consider this worth fixing, or would it interfere in some way? I can't see a reason not to fix it.

Thoughts/Comments?

[ducky]

Damn!!! I'm at a loss. I really appreciate you looking into it, though. Hopefully maybe the recompile will do it otherwise.. damn... we always thought it was the models.

The lights sound cool! I never noticed that before.

The two-sided glass panes might be a bad idea, though, because a lot of times (like on defense) you're playing from behind the goal, and you don't want to see the wall you're looking for an angle on the play.

If the invisible wall doesn't have anything to do with the models, would that mean it has to be in the detection?

[haravikk]

I'd say it must be the collision-detection sometimes screwing up on enclosed models, perhaps if the top of the goals were reshaped somehow it would be less prone to this? Or possible if the model were altered such that it extended one world-unit further forward, without there actually being anything added there except maybe an overhang at the top of the glass-wall? The collision detection is a bit funny about the transition from one cell to another.

[Graydon]

I've gone ahead and modeled a work-around to the original problem. It solves it, however open's up a potential box of new 'bug'. However, if Vista had opted for this method in the original release, everyone would be used to this and not know any different.

What I've done is opened the goal right up. There's no longer any crossbar, and I modeled a 'back' to the goal.. so now it's just a recessed 'room' if you will. What this means, is that the ball can enter the scoring area from any vertical height... It's no longer limited to a crossbar. What this will result in is much more frequent shots from mid-map with like a 3-fetch blast making into the net and counting for a point. I'm not sure if there's a way we could synthetically create a crossbar and net above the scoring area with projectiles or lightning or something, to re-limit the actual goal window.

All that being said, it does fix the 'inviisble wall' issue, so any normal shots that _should_ go in, do go in.

Ducky: If you're interested in this concept at all, let me know and I'll post up some screenshots, and if I get it finished, a patch to a 'fetchball v1.1 beta'.

[Myrd]

For the record, there already has been a Fetchball v1.1, which apparently people didn't like that much and kept using 1.0.

So if a new release comes out of this, it should be v1.2 or something.

[Graydon]

K thanks Myrd. Does anyone know what 'bugs' v1.1 fixed? I'm curious.

[Pyro]

I have the following plugins but only have the readme for 1.0:

Fetchball_v1.0
Fetchball_v1.1
Fetchball 2.0d9b

[Myrd]

It seems the archive of Fetchball 1.1 on udogs ( http://hl.udogs.net/files/Gaming/%20Myt ... l_v1_1.zip ) does not come with a readme...