v3.0.0

• ufoLib now requires the fs module, also know as pyfilesystem2. It will be installed automatically when installing ufoLib via pip.
  https://pypi.org/project/fs/
  /~https://github.com/PyFilesystem/pyfilesystem2
• Added support for reading and writing UFOZ, an alternative zip-compressed UFO file format.
• Moved test suite outside of the ufoLib package, in an external 'tests' folder.
• Removed deprecated function convertUFOFormatVersion1ToFormatVersion2.
• Changed license to MIT License.

NOTE: Except for bug fixes, v3.0 will be the last release for ufoLib as a standalone package. From now on, development will continue over at the /~https://github.com/fonttools/fonttools project.

v2.3.2

• Read features.fea as UTF-8 encoded text file.
• Fix issue when writing non-ASCII characters in a glyph's note text (googlefonts/fontmake#467).

v2.3.1

Fixed issue whereby python2 str (same as bytes) objects were written as property list's <data> elements instead of <string> elements.

Added use_builtin_types option to plistlib.load/dump to control whether <data> elements should be loaded as built-in bytes (this is the default when running ufoLib under python3), or whether they should be wrapped in a plistlib.Data object (default when running ufoLib under python2); when used with the dump function, this option controls whether bytes values should be written as <data> (python3) or as <string> (python2) elements.

#162
#163
#164

v2.3.0

• The lxml and singledispatch requirements are now optional. When not present, the built-in xml.etree.ElementTree library is used. To install ufoLib with lxml support you can do like this: pip install ufoLib[lxml].
• Added new ufoLib.plistlib implementation that works the same way on both python 2 and python 3, and matches the python3's plistlib API (#160).
• Added new ufoLib.etree module that exports the same ElementTree interface using either lxml or xml.etree (built-in) as backends, depending on the availability of former (#161).
• glifLib: use double quotes in XML declaration like previous ufoLib versions (#159).
• Set up test coverage and upload coverage report to codecov.io.

v2.2.4

Fixed issue when parsing GLIF xml from a Unicode string containing an explicit XML encoding declaration. The latter is ignored and UTF-8 encoding is always assumed (#157).

v2.2.3

• Fixed issue introduced in v2.2 whereby the XML declaration was not written at the top of GLIF files and in the string returned from writeGlyphToString function (#157).

v2.2.2

• Minor fixes to validation
• Make validation on the default for read and write (prior it was off for write)

v2.2.1

• Fixes issue whereby the required lxml dependency was not correctly specified in the setup.py's install_requires.

v2.2.0

• Added options to turn validation off on read and write. This can be globally, or per bit of data. For UFOWriter, validation is on by default. For UFOReader, validation is off by default.
• Switched to lxml from ElementTree. lxml is now a required dependency.
• Optimizations to glyphLib (@adrientetar)
• When converting to UFO3, ufoLib will now remove the older @mmk group prefixes

v2.1.1

• Clarify license, copyright and trademark strings for test fonts. They
  now explicitly point to the license of the package (#100).
• Ignore renameGlyphSet and deleteGlyphSet when formatVersion < 3 (#91).
• Make fontLibValidator errors more verbose.