As pointed out to me after class, initializing the arrays in
DepthFirstPaths.java probably needs time proportional to V (unless we
know that their memory is zeroed out already, which we probably
don't).  So, I should not have said that DepthFirstPaths requires O(E)
time.  It is also O(V+E) time, just like CC.


In ../book/CC.java (from http://algs4.cs.princeton.edu/code/algs4.jar)
we found a bug.  Where it says this:
        size[v]++;
it should say this:
        size[count]++;
But this bug is fixed in their other online versions of CC.java:
   http://algs4.cs.princeton.edu/41undirected/CC.java.html
   http://algs4.cs.princeton.edu/41undirected/CC.java
And the textbook completely omits the size method and its array.