Does the kd_tree need SIZE and MAX_LEVEL?

[wucke13]

1. Is there a way to express the one through the other?
2. Is there a significant performance hit when SIZE is rather an attribute of Tree?
3. Same question re when MAX_LEVEL is an attribute of Tree?
4. If the const generics stay, can we use something like static_assertions to ensure that they are not set wrongly (e.g. SIZE=1024, MAX_LEVEL=9)?

[sevenautumns]

Is there a way to express the one through the other?

Not as long as we can not use const generics to calculate the size of arrays at compile time

Is there a significant performance hit when SIZE is rather an attribute of Tree?

We are not in no_std anymore, should we make it an attribute because we cant allocate arrays from attributes

If the const generics stay, can we use something like static_assertions to ensure that they are not set wrongly (e.g. SIZE=1024, MAX_LEVEL=9)?

Yes

[sevenautumns]

@wucke13 I am fairly certain, we could get around SIZE as a const generic but not around MAX_LEVEL.
Doing this would enable us to test and benchmark the search function with different tree sizes up to an maximum of 2^(MAX_LEVEL) nodes

[wucke13]

That sounds good!

[sevenautumns]

With this only MAX_LEVEL remains. I do not think there is any way of verifying that MAX_LEVEL is correct at compile time.

[sevenautumns]

For a given MAX_LEVEL the maximal length of the given Node-Array is 2^(MAX_LEVEL) - 1