#yeah its a copy paste from engine code

1 messages · Page 1 of 1 (latest)

wary obsidian
#

Better to make a thread I think.
Let's start by the beginning: the poly you get from the nav link, when you debug draw it, does it make sense? does it look correct?

floral warren
#

The poly from the navlink has 2 vertices which seem to mostly correspond with the endpoints of the link

#

It doesnt seem to account for the Z axis difference, the connected mesh is offset by 30 units upwards, but the link location only has a Z of 1, might be correct though

#

Its still all the same tile and the step height is more than 30 units

wary obsidian
#

Ummm I find very strange you end up with a poly of 2 nav links. If you are getting a poly ref form the nav, it should have 3 or more vertices. The link itself doesn't exists in the nav data aside "poly ref A points to poly ref B"

#

About the Z axis, not sure tbh

floral warren
#

I think the navlink is a unique poly, which is why there is the dtpolytype enum, to identify if it is an offmesh polygon or a normal one

floral warren
#

So I tried walking through the polys from the path corridor of a path that walks the same navlink, it seems the data is distinctly different between the one I found this way vs the other one, even though the poly points are very similar

wary obsidian
#

umm interesting. I'm a bit overloaded now but super curious about trying what you are trying.
Any chance you can post here the code you use to finde the poly etc:?

floral warren
#

So ive figured out this is not a default navmesh issue. After some digging it turns out this happens due to world partition and/or streaming navmesh. I'm not quite sure what exactly is going wrong

#

The code I've written for finding these polys works as expected if its run at the correct time when everything is loaded