Changes from 0.4.3
- Character encoding conversion is now spontaneous, no more buffering via temporary file.
- Rewritten RTF filter. It's now tested to work with real RTF document.
- Process characters as Unicode internally, so that characters not present in TIS-620 are not lost in output.
- Fix potential buffer overflow vulnerability in Mule mode.
- Updated word break dictionary.
- Significant source clean-ups.
- Switch to XZ tarball compression.