Crossover functions not working as intended for integer sequences?

[prconlin]

The following tests were conducted in Visual Studio Code v1.78.0 with the Julia language support extension v1.47.2, using Julia v1.9.0 and Evolutionary.jl v0.11.1

I ran a few basic tests on the crossover functions using a sequential parent and a null parent (see below). The results were not as expected.

a = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
b = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]

Evolutionary.DC(a,b) produces the following two children:

c = [0, 2, 3, 4, 0, 0, 0, 0, 0, 10]
d = [0, 2, 0, 4, 0, 6, 7, 8, 9, 0]

It's easy to see how c results from randomly assembling the elements of a and b. However, d is supposed to be the inverse of c, but this does not appear to be the case. Unless I misunderstand the concept, the two children produced from this operation should be:

c  = [0, 2, 3, 4, 0, 0, 0, 0, 0, 10]
d' = [1, 0, 0, 0, 5, 6, 7, 8, 9, 0]

Some of the other crossover functions appear to have more severe issues:
Evolutionary.CX(a,b) fails with: BoundsError: attempt to access 10-element Vector{Bool} at index [0]
Evolutionary.OX1(a,b) fails catastrophically and hangs the terminal, forcing me to kill it.
Evolutionary.OX2(a,b) fails with the same error as .CX

I can confirm that Evolutionary.PMX(a,b) works exactly as expected. Evolutionary.SSX(a,b) seems to work, although I'm admittedly unsure what the SSX function is supposed to produce.

I appreciate your development of this module and hope you can help sort out these issues.