Fix DisjointSets constructor error

[c-p-murphy]

Address error introduced by #567 in situations like the following:

DisjointSets(x % 2 == 0 ? x+1//x : x*im for x = 1:10)

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (master@2680058). Click here to learn what that means.
The diff coverage is n/a.

@@            Coverage Diff            @@
##             master     #568   +/-   ##
=========================================
  Coverage          ?   87.95%           
=========================================
  Files             ?       32           
  Lines             ?     2109           
  Branches          ?        0           
=========================================
  Hits              ?     1855           
  Misses            ?      254           
  Partials          ?        0

Impacted Files	Coverage Δ
src/disjoint_set.jl	94.33% <ø> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2680058...7809b81. Read the comment docs.

[c-p-murphy]

Just realized this actually still needs work, so don't worry about reviewing or merging it yet.

[oxinabox]

Definately add that test case . I should have checked coverage before merging

[c-p-murphy]

Everything seems to be working as expected now, but I'm not entirely confident that this is the best approach to solving the problem. Both this and the Set constructor rely on these lines in arrays.jl which seems a bit counterintuitive. If it doesn't seem like a problem to you though, I'm happy to have this merged and then see about opening a new PR down the road if a more elegant solution becomes apparent.

Either way, thanks for taking the time to review this.

[oxinabox]

Looks good to me.
Can you bump the patch version so I can tag a release?