There was a bug in LAME and a bug in my configure calls that meant that optimisations for LAME were not enabled. This is now turned on and should make the library significantly faster on macOS and Linux. Windows is unaffected.
1.5.1
Enable some optimisations, although it's a very minor speed bump and it appears that using the command line is still 15% faster. No idea why.
1.5.0
Switch the license to LGPLv3.
1.4.2
Support Python 3.11 and builds for Linux ARM64.
1.4.1
Add support for cross-compiling ARM64 binaries for macOS with Apple Silicon. Completely untested, but build appears to succeed and be of a reasonable size.
1.4.0
Add support for Python 3.10 and move to GitHub actions as Travis is basically unusable now.