Skip to content

v4.3.1

Choose a tag to compare

@deemonic deemonic released this 03 Nov 10:32
· 37 commits to main since this release
c6dc5fc

Bug Fixes

  • Masks landing on the wrong characters (#56): the regex and phonetic drivers matched against normalized text but masked the original at the same positions. Whenever normalizing changed the length, every later mask was shifted. This affected:
    • German ä/ö/ü/ß and sch
    • Spanish ll
    • French œ/æ
    • Repeated whitespace in any language
Input Before After
a fuck this shit a ****k thi****it a **** this ****
Du blöde Scheiße, du Arschloch Du ******S******* du A******** Du ***** *******, du *********
llama mierda llama******a llama ******

Matched word text, positions and clean() now all refer to the right characters. No API changes.