* First, we were treating the return value of HTTPResponse.read() as a
  string when it is really a byte string.

* Second, the urlparse library was renamed to urllib.parse in Python 3,
  so I made the appropriate substitution.

Change-Id: I0b23716e990eca8ddf77742f378ab54eda3d3717
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/2307509Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
Commit-Queue: Dan McArdle <dmcardle@chromium.org>