c8d68acc9e
* chore: prefer es-toolkit helpers and lint the mechanical cases AGENTS.md now names the es-toolkit helpers to reach for instead of hand-written equivalents, and the exceptions: a single clear native call or a measured hot path. The new open-pencil/prefer-es-toolkit rule rejects filter(Boolean) and Set round trips on arrays, the two cases that need no type information, and the existing 45 sites use compact and uniq. tools/ci/policy runs before dependencies are installed, so the rule is off there. * refactor: deduplicate diagnostic categories with uniq * fix: keep es-toolkit out of serialized Playwright callbacks The codemod rewrote a filter(Boolean) inside a page.evaluate callback, which Playwright runs in the page where the compact import does not exist. The spec filters there again, and the rule now skips callbacks passed to evaluate, $eval, $$eval, evaluateHandle, addInitScript and waitForFunction, and filter calls on iterators from values, keys, entries and matchAll, which compact cannot take. * test: write the prefer-es-toolkit cases like the other rule tests Short standalone snippets, as in the base64 and JSON rule tests, instead of a declaration prefix on every case and inline object types.
16 lines
312 B
JSON
16 lines
312 B
JSON
{
|
|
"name": "@open-pencil/tauri-menu-tools",
|
|
"private": true,
|
|
"type": "module",
|
|
"scripts": {
|
|
"test": "bun test tests"
|
|
},
|
|
"dependencies": {
|
|
"es-toolkit": "^1.52.0"
|
|
},
|
|
"devDependencies": {
|
|
"@open-pencil/package-artifacts-tools": "workspace:*",
|
|
"@open-pencil/vue": "workspace:*"
|
|
}
|
|
}
|