Physics2D.LinecastNonAlloc vS Physics2D.Linecast

According to the API LinecastNonAlloc can avoid overhead cost of allocating memory for the
RaycastHit2D array. But I still have to declare and allocate my own RaycastHit2D (used as argument Physics2D.LinecastNonAlloc ) to array when I use this function so I don't see / understand the benefit ?

 

I think the benefit is for when you already know your RaycastHit2D, so in that case instead of reallocating it again you just use the non allocating function, I don't think it has any other use.

 

The benefit is that it is not allocating data which is feeding the garbage collector. Hint: it's always better.

 

I think what you are misisng is, the array should be allocated outside of the function call, and kept around for the life of the object that is doing the searching, so it can 'recycle' the array

 

Well it just populates it up until array length, instead of doing allocations with every call. I wish all physics tests including sweeps had this option since it's not fun to see your title slow down due to core loop allocations.

 

very much agreed, honestly I was shocked when I first started using unity that they didnt do this, and very happy when I saw the new 2D calls have it as an option!

 

And @Unity ... how about all the other calls, guys? don't stop now.

 

Something else I'd like to see, that is included in almost every other game engine is an ignore parameter for these functions as well, so you can ignore one specific collider, without having to have the whole list marshaled over to c# and checked.

 

Are any of the above two requests listed in the wish list thingie? If so what are they called?
I'll add my vote if they are since it sounds like those are two very worthy optimizers.

 

OK great.

I have one class that are handling grounding. Can I have raycasthits2d[] as public / private variable in this class and initiate the size in the awake() function? (tried it but shows null when I use it in other class methods in this class later on)