Repository navigation
0.1.2
Full Changelog: 0.1.1...0.1.2
中文
维护版:没有新增功能,.layer 数据格式没有变化。
修复
- 弹出面板(新建字段、筛选、排序、行高、颜色规则、列可见性、看板字段)改用自己的类名标记容器,不再依赖
:has()。表格每次重绘时,浏览器不必再重新匹配文档里所有弹窗容器。 - 输入法拦截不再读取已废弃的
KeyboardEvent.keyCode,改用isComposing和Process键值。中文输入法里按回车选词的行为与之前完全一致。
内部
- MCP server 用到的 Node 接口(
node:fs、node:path、process)集中到mcp/node-api.ts,并写明了确切签名。在没有加载 Node 类型声明的检查环境里,这些调用不会再被逐条判成 unsafe 值——社区目录的自动扫描曾因此报了 76 条。 - MCP server 一直在向客户端自报
0.1.0;现在跟随manifest.json,并且npm run check:version会把这一处一起校验,漏改就会失败。
升级:覆盖 main.js、manifest.json、styles.css(目录 layer-base)。数据不需要迁移。Release 另外附 layer-base-0.1.2.zip,里面就是这三个文件、不含外层目录。用 MCP server 的话,把它指向的 mcp-server.cjs 换成新版并重启客户端——客户端里显示的版本会从 0.1.0 变成 0.1.2。
English
A maintenance release: nothing new to use, and the .layer format is unchanged.
Fixed
- The anchored panels (new field, filter, sort, row height, colour rules, column visibility, board fields) mark their own container with a class instead of matching it with
:has(), so a table redraw no longer makes the browser re-match every modal container in the document. - The IME guard reads
isComposingand theProcesskey instead of the deprecatedKeyboardEvent.keyCode. Typing Chinese and pressing Enter to commit a candidate behaves exactly as it did.
Behind the scenes
- The Node surface the MCP server uses (
node:fs,node:path,process) now lives in one module with its signatures written out,mcp/node-api.ts. A checker that runs without Node's own declarations no longer reads every one of those calls as an unsafe value — the community directory's automated scan reported 76 of them. - The MCP server announced
0.1.0to every client. It now followsmanifest.json, andnpm run check:versionfails when the two drift.
Upgrading: replace main.js, manifest.json and styles.css in the layer-base folder. No data migration. The release also carries layer-base-0.1.2.zip, which holds exactly those three files with no wrapping folder. If you use the MCP server, replace the mcp-server.cjs you point it at and restart the client — the version it reports goes from 0.1.0 to 0.1.2.