Fixes point_search when unique is false

[danielpetri1]

This change implements point_search iteratively to ensure a stack overflow can not occur.

It also makes the method use the passed unique parameter, as the recursive call was previously defaulting unique to true:

itv = [(5...20), (15.6...20), (15.7...20), (15.7...20)]
t = IntervalTree::Tree.new(itv)
p t.search(15.7)     #=> [5...20, 15.6...20, 15.7...20]
p t.search(15.7, unique: false) #=> Should be [5...20, 15.6...20, 15.7...20, 15.7...20], not [5...20, 15.6...20, 15.7...20] again

Additionally, result.uniq is invoked at the end of the method for faster query speeds.

[ZimbiX]

Cheers; good thinking.

Given this includes the removal of #include?, from #13, let's rebase this once that PR is merged. In the meantime, would you mind adding a spec please? =)

[ZimbiX]

(I should have selected this, given the missing spec)

[danielpetri1]

Hi @ZimbiX,
Thanks for reviewing the code. I added 3 example to showcase where unique was failing in my view.

In addition to that I also changed the text of the case in which uniq defaults to true and added one more example for it. Before, it said something along the lines of "given intervals with duplicates, return the duplicates in the result" - I am assuming duplicates should be removed from the intervals in this PR.

Cheers!