Fix and small optimization

[Ward727a]

So as I said here: Redot-Experimental#11 (comment) there was a bug when we get 2 functions on the same line in trait file and script file. This is now fixed!

Also I did a small optimization on the get_function_names_recursively function inside gdscript_editor.cpp this should allow us to gain ~10% (from 9.8µs to 8.8µs).

In fact: The fix and the optimization are linked.

Before when Redot was getting the function name inside the script, it was checking for the function name inside the script file (.gd) and added them to a HashMap<int, String> (int = starting line of the function, string = function name), then it was checking for the function name inside the traits files (.gdt), but then if 1 function was on the same line than the script file, the function inside the traits files would "override" it.

Now they are "pre-concat" then added to a vector.

Some explanation:

First, a simple example:

# MyTrait.gdt
trait_name my_trait extends Node

func MyTraitFunction():
    pass
    
# MyScript.gd
extends Node
uses my_trait
func _process():
    print("hello world!")

func _ready():
    pass

Here we have 2 functions on the same line (3): MyTraitFunction and _process, Redot would first get the function from the script so MyScript.gd here, giving this result:

Function[3] = "_process"
Function[6] = "_ready"

Then Redot would get the function from the trait, here MyTrait.gdt, but as the function MyTraitFunction is on the same line, here is the result:

Function[3] = "MyTraitFunction"
Function[6] = "_ready"

We can see that we losed "_process".

In this fix, as we use vector, we directly format the string as awaited for next function, with the new version, we would get something like that:

Function = ["_process:3", "_ready:6", "MyTraitFunction:3"]

[decryptedchaos]

So, i'm probably gonna have to just pull the diff here and do fix this PR on my side because i had to create a new branch for the trait feature called feature/traits-fix

No big deal i can fix it, but it does mess up the credit on the log so i'll just reference this PR

[Ward727a]

Okk, no problem at all, don't worry :)

[decryptedchaos]

Closing this, We need to re-open a new PR with the correct branch this branch was very broken and i had to create a new branch to fix it