slope issues

physics

• Objects that aren't players are really bad at moving up and down slopes. (P_ZMovement runs the check and SlopeLaunch at the start of the function, while P_PlayerZMovement runs it later on. This may be why.)
• Ceiling (floor in reverse gravity) bouncing for non-player objects doesn't account for slopes. Needs fixed to.
• Fire trails on slopes don't work properly. Don't spawn going uphill, spawn but don't hurt going downhill.
• P_ButteredSlope doesn't get the object's actual gravity. Fix it. (Make a function to get an object's gravity if we don't already have one??)
• (lllloooowwww priority, feature request) Slide momentum up a wall from a steep enough slope so that quarter-pipes can at least slightly make physical sense. It would work like a slope launch and not actually stick the player to the wall.

rendering

• FOF planes don't layer properly when slopes are involved. Probably a simple check that needs tweaked.
• FOF planes also don't light properly. Similar issue, I believe.
• Sprites occasionally appear through walls that are near slopes. I bet I just misspelled something in the clipping code or some shit.
• Slope planes render with wrong height values when visportals are visible on-screen. (Compute and store relative viewx/y/z/angle when creating the visplane?)
• (low priority) Sloped FOFs with midtextures that have holes in them are completely broken. Texture alignment does not skew, but FOF holes do. Going to be a difficult fix.

ogl rendering (need to work with chee on)

• Midtextures + slopes == bugs.
• Sloped light blocks aren't properly rendered and do not affect sprites properly.

other

• Add the new height-related functions and slope structs and eeeeverything else to Lua metatables so that modders can access them.

Physics issue 3 fixed in this commit on my slope-fixes branch: 3058648f

mentioned in merge request !50 (merged)

I added the slope struct to Lua on another branch, though it's still pretty buggy/incomplete: b0cbc8ab

Do the first two rendering issues occur in slope-fixes? I have no idea if the sloped FOF issues I fixed the other day are related or not, but it might be worth checking

Disregard the previous comment, they're not related as far as I'm aware

Assignee removed

@MonsterIestyn Confirm rendering issue 3 is fixed in slope-fixes?

Well, from the tests I've done I believe I've fixed all cases of sprites displaying through walls adjacent to slopes that I know of... that is:

• walls of slopes above the player's viewz, but at least partially below the player's floor (see c67683eb)
• "closed door" walls involving slopes, which includes thok barriers IIRC (see 6623a914 and a82c19ad)

I think rendering issue 2 should also be fixed as of 6bda1a57

Fix for rendering issue 4: 3c5a8b80

Not part of slope-fixes because there's other problems with portals that I intend to sort out soon anyway

mentioned in commit ac3de70c

mentioned in commit bb9098b5

Thanks to toaster, physics issue 4 should be sorted now: fa002e58

No check off, though - not merged in yet.

On the other hand, you do seem to have merged in a fix to some of that midtexture weirdness in OGL. Does that get checked off?

Diagnosed what is probably a big chunk of physics issue 1.

https://gfycat.com/WhisperedVelvetyArmedcrab (go into frame step-by-step to see things the way i'm looking at it)

Whenever the Crawla's frame is updating (marked by the tread pixels changing), it's stepping INTO the slope. It then doesn't get pushed upwards until the next frame. Investigating further down as we speak.

Nah the issue is in P_ZMovement

I actually moved the whole "if (mo->standingslope)" thing like it is in P_PlayerZMovement, and there was no meaningful improvement in the crawla's motion up the slope.

EDIT: Committed so we're all on the same page.

I don't think the real issue is with P_ZMovement, it's probably more to do with the fact crawlas don't use momx/momy to move horizontally, they use A_Chase (which calls P_Move) to move by simply "teleporting" to their position plus their mobjinfo speed value in a particular direction.

The problem probably lies with P_Move not taking slopes into consideration like P_XYMovement does, if it did it wouldn't have that "stepping" problem going up or down. That is, if I'm correct.

Physics issue #1 (closed) got fixed for 2.1.16.

Just curious, is there an example of Rendering issue #5 to check out?

Physics issue 5 done in internal.

@RedEnchilada or @chi.miru, either of you able to confirm if either of the two ogl rendering issues have been fixed?

marked the checklist item Add the new height-related functions and slope structs and eeeeverything else to Lua metatables so that modders can access them. as completed

Lua support for slope now exists thanks to !307 (merged) being merged in

mentioned in merge request !339 (merged)