Change scanning.scanResolution recommended setting to character for agglutinative languages#2308
Conversation
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 8f867a37e8
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| "path": "scanning.scanResolution", | ||
| "value": "character" |
There was a problem hiding this comment.
Restore word search resolution for agglutinative presets
This replacement removes the previous translation.searchResolution = "word" override for Finnish (and the same pattern appears in hu, id, mn, and tr), so these languages now fall back to the global default translation.searchResolution = "letter" (ext/data/schemas/options-schema.json). In letter mode, Translator._getNextSubstring truncates one character at a time (ext/js/language/translator.js), which substantially increases per-scan lookup attempts and introduces many partial-prefix matches compared with the prior word-boundary behavior, especially on long agglutinative words; keeping an explicit word search resolution avoids this regression.
Useful? React with 👍 / 👎.
Kuuuube
left a comment
There was a problem hiding this comment.
Im fine merging this. Might need some followup for translation.searchResolution. Not entirely sure but doesn't seem like it will cause any issues as is.

Source is this Wikipedia category.
Not sure what translation.searchResolution does; removed it completely for agglutinatives.