Prevent accidental wallmounted_to_dir poisoning by mods

[TheEt1234]

• Goal of the PR

I'll copy paste the commit's description (but slightly edited to correct things):

Prior to this commit, if you used a function like core.wallmounted_to_dir, and modified its output, it would modify all of the output in the future

So, something like this:

local v = core.wallmounted_to_dir(0)
v.y = v.y + 1
core.debug(v:to_string()) -- result: (0, 2, 0)
local v2 = core.wallmounted_to_dir(0)
core.debug(v2:to_string()) -- result: (0, 2, 0)

This would result in some bugs that are really hard to catch if you don't know this undocumented behavior

---

Now that i'm writing this i realize i made an error in the commit description, line 9 of the commit description was supposed to have result: (0, 2, 0) not result: (0, 1, 0), and in line 10 i have put core.get_wallmounted_to_dir(0) when i should've put core.wallmounted_to_dir(0), i'm never making commit descriptions again

I don't know how to change the commit message description, besides just making a new fork and PR, is this a problem?

---

• How does the PR work?
  Uses vector.copy on outputs of core.wallmounted_to_dir, core.fourdir_to_dir, core.facedir_to_dir so that can't happen
• Does it resolve any reported issue?
  Not in the issue tracker, but in discord (i reported it):
  https://discord.com/channels/369122544273588224/369123175583186964/1341121598895423591
  The content of the message:

i think i just experienced possibly the weirdest engine bug ever?
`core.wallmounted_to_dir(1)` for some reason wasn't `vector.new(0,-1,0)` but was `vector.new(0,1,0)` instead, 

it starts to happen completely randomly

(i was completely clueless about it back then)

• Does this relate to a goal in the roadmap?
  No

To do

This PR is Ready for Review.

How to test

local v = core.wallmounted_to_dir(0)
v.y = v.y + 1
local v2 = core.wallmounted_to_dir(0)
assert(v2.y == 1, "test failed") 

[sfan5]

I don't know how to change the commit message description, besides just making a new fork and PR, is this a problem?

It's possible to edit commits in-place, but we can also just strip the description upon merging.
IMO there is no need for the commit description to be that verbose anyway since interested people can just look up the PR a commit came from.