fix: 修复行缘、切分还原与语言解析并同步文档

保留逐字 margin／jiya.bias 字体后备方案；补齐 demo 重排、量测边界说明及共享规则 1.0.1 反例向量。

验证：完整构建通过，Node 376 项、WKWebView 64 项及原生 demo 9 项通过；Firefox 仍暂缓。

rules/README.md

@@ -40,6 +40,15 @@ Unicode profile 另行記錄。更新 Unicode 範圍時，即使 JSON 結構不
 字符覆寫固定為追加語義：不得經 override 移除或重分類內置字符。若未來需要破壞此
 規則，屬 major 變更，須先作產品政策決策，不能由單一 adapter 私自實作。
 
+## 1.0.1 語言契約澄清
+
+`zh-explicit-script` 及 `zh-traditional-region` 指 [BCP 47 的 script／region 位置](https://www.rfc-editor.org/rfc/rfc5646.html#section-2.1)，
+不能從 extension／private-use 內容搜尋同名字串。Web 曾把 `zh-Hans-x-hant` 誤判成
+繁體，現修正為簡體；空語言標籤歸 `other`。新增七個反例向量，既有 55 個向量結果、
+字符集合、預設政策、schema 與 adapter 義務均未改，故依上述 patch 規則升為 1.0.1。
+HTML 如何繼承 `lang` 是 Web adapter 契約，不寫入共享規則。未來 adapter 仍須通過
+全部 62 個向量；本次未驗收其他 adapter。
+
 ## Adapter 邊界
 
 字符集合一律以大寫十六進位 Unicode code point 的閉區間／單點表示；陣列須按起點

rules/golden-vectors.json

@@ -1,7 +1,7 @@
 {
   "$schema": "./golden-vectors.schema.json",
   "schemaVersion": 1,
-  "rulesetVersion": "1.0.0",
+  "rulesetVersion": "1.0.1",
   "vectors": [
     {
       "id": "locale-ja",
@@ -63,6 +63,48 @@
       "input": { "lang": "en" },
       "expected": { "langClass": "other" }
     },
+    {
+      "id": "locale-hans-private-hant",
+      "category": "language-classification",
+      "input": { "lang": "zh-Hans-x-hant" },
+      "expected": { "langClass": "zh-Hans" }
+    },
+    {
+      "id": "locale-tw-private-hans",
+      "category": "language-classification",
+      "input": { "lang": "zh-TW-x-hans" },
+      "expected": { "langClass": "zh-Hant" }
+    },
+    {
+      "id": "locale-latn-private-tw",
+      "category": "language-classification",
+      "input": { "lang": "zh-Latn-x-tw" },
+      "expected": { "langClass": "zh-Hans" }
+    },
+    {
+      "id": "locale-extension-hant",
+      "category": "language-classification",
+      "input": { "lang": "zh-CN-u-co-hant" },
+      "expected": { "langClass": "zh-Hans" }
+    },
+    {
+      "id": "locale-extlang-hans",
+      "category": "language-classification",
+      "input": { "lang": "zh-cmn-Hans-TW" },
+      "expected": { "langClass": "zh-Hans" }
+    },
+    {
+      "id": "locale-hant-cn",
+      "category": "language-classification",
+      "input": { "lang": "zh-Hant-CN" },
+      "expected": { "langClass": "zh-Hant" }
+    },
+    {
+      "id": "locale-empty",
+      "category": "language-classification",
+      "input": { "lang": "" },
+      "expected": { "langClass": "other" }
+    },
     {
       "id": "punct-stop",
       "category": "punctuation-classification",

rules/juzhen-rules.json

@@ -1,7 +1,7 @@
 {
   "$schema": "./juzhen-rules.schema.json",
   "schemaVersion": 1,
-  "rulesetVersion": "1.0.0",
+  "rulesetVersion": "1.0.1",
   "unicodeProfile": {
     "version": "17.0.0",
     "coverage": "CJK Unified Ideographs through Extension J, compatibility ideographs, kana, bopomofo, radicals, and enclosed CJK characters"