Hovercraft Pathing Oddities

[REVENGE]

When testing a map with some vertical cliffs, I discovered that hovercraft are able to path up some seemingly vertical sections of terrain.





However, on a different map with similar vertical sections (Higher Grounds v2), I was unable to reproduce the issue.

Any thoughts?

[Kloot]

Slopes need to be a minimum of two heightmap squares wide before the engine can reliably detect them.

[jK]

Slopes need to be a minimum of two heightmap squares wide before the engine can reliably detect them.

I thought I fixed that ages ago

[knorke]

On this map hovers can cross the supersteep cliff between land and water:
http://zero-k.info/Maps/Detail/25394
replay: http://zero-k.info/Battles/Detail/148156 (another is linked in map comments, did not watch it)
That is with 91.0 though.

[hoijui]

i propose this image as the very first thing on the engine dev wiki

the essence

[REVENGE]

Nice.

Thanks guys, shall I file a Mantis report?

[jK]

@slopemap
https://github.com/spring/spring/blob/d ... p.cpp#L491
-> see I fixed that ages ago. before it just computed the slope for just one square. since a while now it computes all slopes and saves the average one to the slopemap (I also tried a max filter but that broke pathing on many maps for huge tanks)

When slopes aren't detected correctly atm it is not the fault of the slopemap, instead it is either fault of the pathfinder or incorrect updated heightmap rectangles.