v3.7

• Fix issue where specially crafted inputs to encode() could
  take exceptionally long amount of time to process. [CVE-2024-3651]

Thanks to Guido Vranken for reporting the issue.

v3.6

• Fix regression to include tests in source distribution.

v3.5

• Update to Unicode 15.1.0
• String codec name is now "idna2008" as overriding the system codec
  "idna" was not working.
• Fix typing error for codec encoding
• "setup.py" has been restored for this release due to some downstream
  lack of adherence to PEP 517. Should be removed in a future release
  so please prepare accordingly.
• Removed reliance on a symlink for the "idna-data" tool to comport
  with PEP 517 and the Python Packaging User Guide for sdist archives.
• Added security reporting protocol for project

v3.4

• Update to Unicode 15.0.0
• Migrate to pyproject.toml for build information (PEP 621)
• Correct another instance where generic exception was raised instead of
  IDNAError for malformed input
• Source distribution uses zeroized file ownership for improved
  reproducibility

v3.3

• Update to Unicode 14.0.0
• Update to in-line type annotations
• Throw IDNAError exception correctly for some malformed input
• Advertise support for Python 3.10
• Improve testing regime on Github
• Fix Russian typo in documentation

v3.2

• Add type hints (Thanks, Seth Michael Larson!)
• Remove support for Python 3.4

v3.1

• Ensure license is included in package (Thanks, Julien Schueller)
• No longer mark wheel has universal (Thanks, Matthieu Darbois)
• Test on PowerPC using Travis CI

v3.0

• Python 2 is no longer supported (the 2.x branch supports Python 2,
  use "idna<3" in your requirements file if you need Python 2 support)
• Support for V2 UTS 46 test vectors.

v2.10

• Update to Unicode 13.0.0.
• Throws a more specific exception if "xn--" is provided as a label.
• This is expected to be the last version that supports Python 2.

v2.9

• Update to Unicode 12.1.0.
• Prohibit A-labels ending with a hyphen (Thanks, Julien Bernard!)
• Future-proofing: Test on Python 3.7 and 3.8, don't immediately
  fail should Python 4 come along.
• Made BSD 3-clause license clearer

v2.8

• Update to Unicode 11.0.0.
• Provide more specific exceptions for some malformed labels.

v2.7

• Update to Unicode 10.0.0.
• No longer accepts dot-prefixed domains (e.g. ".example") as valid.
  This is to be more conformant with the UTS 46 spec. Users should
  strip dot prefixes from domains before processing.

v2.6

• Allows generation of IDNA and UTS 46 table data for different
  versions of Unicode, by deriving properties directly from
  Unicode data.
• Ability to generate RFC 5892/IANA-style table data
• Diagnostic output of IDNA-related Unicode properties and
  derived calculations for a given codepoint
• Support for idna.version to report version
• Support for idna.idnadata.version and
  idna.uts46data.version to report Unicode version of
  underlying IDNA and UTS 46 data respectively.

v2.5

• Fix bug with Katakana middle dot context-rule (Thanks, Greg
  Shikhman.)

v2.4

• Restore IDNAError to be a subclass of UnicodeError, as some users of
  this library are only looking for the latter to catch invalid strings.

v2.3

• Fix bugs relating to deriving IDNAError from UnicodeError.
• More memory footprint improvements (Thanks, Alex Gaynor)

v2.2

• Made some changes to the UTS 46 data that should allow Jython to get around
  64kb Java class limits. (Thanks, John A. Booth and Marcin Płonka.)
• In Python 2.6, skip two tests that rely on data not present in that
  Python version's unicodedata module.
• Use relative imports to help downstream users.

v2.1

• Memory consumption optimizations. The library should consume significantly
  less memory through smarter data structures being used to represent
  relevant Unicode properties. Many thanks to Shivaram Lingamneni for this
  patch.
• Patches to make library work better with Python 2.6. The core library
  currently works however the unit testing does not. (Thanks, Robert
  Buchholz)
• Better affix all Unicode codepoint properties to a specific version.

v2.0

• Added support for Unicode IDNA Compatibility Processing (aka
  Unicode Technical Standard #46). Big thanks to Jon Ribbens
  who contributed this functionality.

v1.1

• Use IDNA properties from Unicode 6.3.0. Internet Architecture Board (IAB)
  issued statement recommending against the use of Unicode 7.0.0 until
  issues relating to U+08A1 codepoint are resolved. See http://goo.gl/Ed1n0K
• Identify some cases when label would be too longer to be a legal DNS name
  and raise an exception. (Thanks, Ed Lewis)