* Fri Jan 04 2019 Todd R <toddrme2178@gmail.com>
- Update to Version 3.0.2
+ Fixed:
* On some systems, setup.py opened README.md with a non-unicode encoding. My fault for leaving the encoding flapping in the breeze. It's been fixed.
- Update to Version 3.0.1
+ Added:
* Travis testing for 3.7 and 3.8-dev. These needed OpenSSL, sudo and Xenial. 3.8-dev is allowed to fail.
+ Fixed:
* PyPI wasn't rendering markdown because I didn't tell it what format to use.
* Python 2 wasn't installing via pip because of a new utils package. It has been zapped.
+ Maintainers:
* TestPyPI version strings use .postN as the suffix instead of bN, and N counts from the latest tagged commit, which should be the last release
* Install from TestPyPI works via make install-testpypi
- Update to Version 3.0.0
+ Breaking:
* search(begin, end, strict) has been replaced with at(point), overlap(begin, end), and envelop(begin, end)
* extend(items) has been deleted, use update(items) instead
* Methods that take a strict=True/False argument now consistently default to strict=True
* Dropped support for Python 2.6, 3.2, and 3.3
* Add support for Python 3.5, 3.6, and 3.7
* Faster Interval overlap checking (@tuxzz, #56)
+ Updated README:
* new restructuring methods from 2.1.0
* example of from_tuples() added
* more info about chop(), split_overlaps(), merge_overlaps() and merge_equals().
+ Fixes:
* Node.from_tuples() will now raise an error if given an empty iterable. This should never happen, and it should error if it does.
* Interval.distance_to() gave an incorrect distance when passed the Interval's upper boundary
* Node.pop_greatest_child() sometimes forgot to rotate() when creating new child nodes. (@escalonn, #41, #42)
* IntervalTree.begin() and end() are O(1), not O(n). (@ProgVal, #40)
* intersection_update() and symmetric_difference() and symmetric_difference_update() didn't actually work. Now they do.
* collections.abc deprecation warning no longer happens
+ Maintainers:
* PyPi accepts Markdown! Woohoo!
* reorganize tests
* more tests added to improve code coverage (We're at 96%! Yay!)
* test for issue #4 had a broken import reference