Repository navigation
Allow Chinese tags and document categories - #6
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d96b4ae09e
ℹ️ About Codex in GitHub
Your team has set up Codex to 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 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if (!window.categoryAclChoices) window.categoryAclChoices = {}; | ||
|
|
||
| if (!/^[가-힣a-zA-Z0-9\s_.-]+$/.test(cleanTag)) { | ||
| if (!/^[\p{Script=Han}가-힣a-zA-Z0-9\s_.-]+$/u.test(cleanTag)) { |
There was a problem hiding this comment.
When an application consumes the public @eoeoe2/wiki-shared/editor/autocomplete export, Chinese categories are still rejected client-side because the duplicate validator in packages/wiki-shared/src/editor/autocomplete.ts:1558 retains the Hangul/ASCII-only regex. Updating only this app-specific copy therefore leaves shared-package consumers unable to use the newly supported categories even though the server accepts them; keep the exported implementation in sync.
Useful? React with 👍 / 👎.
| * 입력은 배열 / JSON 배열 문자열 / 쉼표구분 문자열을 모두 허용한다. | ||
| */ | ||
| const TAG_VALID_RE = /^[가-힣a-zA-Z0-9 _.-]+$/; | ||
| const TAG_VALID_RE = /^[\p{Script=Han}가-힣a-zA-Z0-9 _.-]+$/u; |
There was a problem hiding this comment.
When a media tag contains newly accepted supplementary Han characters such as 𠀀, the existing trimmed.length > TAG_MAX_LENGTH check at line 41 counts each character as two UTF-16 code units. Consequently, a valid 26-character Extension-B tag is silently discarded as longer than the advertised 50-character limit, and the client performs the same incorrect check in src/client/common.ts; count Unicode code points consistently for the limit.
Useful? React with 👍 / 👎.
| // 컨테이너에 버블 UI + 자동완성(/api/media/search-tags)을 장착한다. | ||
| // 업로드 모달, 이미지 검색 모달, 이미지 문서 편집 모달에서 공통 사용한다. | ||
| const MEDIA_TAG_VALID_RE = /^[가-힣a-zA-Z0-9 _.-]+$/; | ||
| const MEDIA_TAG_VALID_RE = /^[\p{Script=Han}가-힣a-zA-Z0-9 _.-]+$/u; |
There was a problem hiding this comment.
For an accepted tag whose next character after the current query is supplementary Han, such as tag 中𠀀 with query 中, the autocomplete endpoint's upper bound ${q}\uffff in src/routes/media.ts:308-310 excludes it under SQLite's binary UTF-8 ordering because supplementary code points sort after U+FFFF. The newly allowed tag can therefore disappear from autocomplete while the user types its prefix; use a prefix query that spans the full Unicode range.
Useful? React with 👍 / 👎.
1 participant
Chinese media tags and document categories were rejected by Korean/ASCII-only validation. Accept Unicode Han characters in the frontend, API, batch operations, category ACL administration and MCP while retaining existing character and size restrictions. Update localized input guidance.
Validation: server/client typechecks, i18n tests and full production build pass. Direct sanitizer checks cover simplified/traditional and supplementary Han, mixed scripts, deduplication, array/JSON/comma input, limits and unsafe-character rejection.