i hunt the text bugs that quietly break software everywhere: unicode, ime, encoding. hardest to catch, hardest to reproduce: cjk and japanese. most of the proof below traces back there.
| merged into OSS | open, in review | failures catalogued |
|---|---|---|
| 113 | 145 | 97 |
Read live from the GitHub API on 2026-07-04. Nothing on this line is hand-typed.
corpus · fixtures · benchmark · report · proof · hub · dev.to · X · i18n-swarm · glovrex



