jsql-neo 6.4.0-beta.3 → 6.4.0-beta.4

package/README.md

@@ -3,7 +3,7 @@
 > **One engine to rule them all** — a Rust-powered embedded database that speaks your language:
 > MySQL, PostgreSQL, MongoDB, Redis, SQL, TypeScript, **and the browser**.
 
-> **v6.4.0-beta.3** — official release build · [github.com/vexify-org/JSQL-neo](https://github.com/vexify-org/JSQL-neo) · [jsql.vexify.top](https://jsql.vexify.top)
+> **v6.4.0-beta.4** — official release build · [github.com/vexify-org/JSQL-neo](https://github.com/vexify-org/JSQL-neo) · [jsql.vexify.top](https://jsql.vexify.top)
 
 ![Engines](https://img.shields.io/badge/engines-Native%20%7C%20WASM%20%7C%20Pure%20JS-7ee787)
 ![MySQL](https://img.shields.io/badge/protocol-MySQL%20compatible-1f6feb)

package/index.js

@@ -1,5 +1,5 @@
 /**
- * JSQL-NEO v6.4.0-beta.3 — Rust-Powered Embedded Database (WASM + HTTP)
+ * JSQL-NEO v6.4.0-beta.4 — Rust-Powered Embedded Database (WASM + HTTP)
  *
  * @example
  * const jsql = require('jsql-neo');

package/lib/sql.js

@@ -136,8 +136,22 @@ function tokenize(sql) {
   let i = 0;
   const n = sql.length;
 
+  // 获取位置 i 处的完整 Unicode 字符（处理代理对，如 emoji）
+  // 返回 { char, length }，char 是完整字符，length 是 UTF-16 码元数
+  function charAt(pos) {
+    const c = sql[pos];
+    const code = sql.charCodeAt(pos);
+    if (code >= 0xD800 && code <= 0xDBFF && pos + 1 < n) {
+      const next = sql.charCodeAt(pos + 1);
+      if (next >= 0xDC00 && next <= 0xDFFF) {
+        return { char: sql.slice(pos, pos + 2), length: 2 };
+      }
+    }
+    return { char: c, length: 1 };
+  }
+
   while (i < n) {
-    const c = sql[i];
+    const { char: c, length: cLen } = charAt(i);
 
     if (c === ' ' || c === '\t' || c === '\n' || c === '\r') { i++; continue; }
 
@@ -188,6 +202,37 @@ function tokenize(sql) {
       continue;
     }
 
+    // 16 进制字面量：0x1F / 0XFF / X'1F' / x'FF'
+    // 标准 SQL：0x 前缀或 X'' 包裹，解析成十进制整数
+    if ((c === '0' && (sql[i + 1] === 'x' || sql[i + 1] === 'X')) ||
+        ((c === 'X' || c === 'x') && sql[i + 1] === "'")) {
+      if (c === '0') {
+        // 0x 前缀形式：0x1F, 0XFF
+        let j = i + 2;
+        while (j < n && /[0-9A-Fa-f]/.test(sql[j])) j++;
+        if (j === i + 2) throw new Error(`Invalid hexadecimal literal at position ${i}`);
+        const hexStr = sql.slice(i + 2, j);
+        const numVal = parseInt(hexStr, 16);
+        tokens.push(new SQLToken('number', numVal, i));
+        i = j;
+        continue;
+      } else {
+        // X'...' 形式：X'1F', x'FF'
+        let j = i + 2;
+        let hexStr = '';
+        while (j < n && sql[j] !== "'") {
+          if (!/[0-9A-Fa-f]/.test(sql[j])) throw new Error(`Invalid hexadecimal literal at position ${j}`);
+          hexStr += sql[j];
+          j++;
+        }
+        if (j >= n) throw new Error(`Unterminated hexadecimal literal at position ${i}`);
+        const numVal = parseInt(hexStr, 16);
+        tokens.push(new SQLToken('number', numVal, i));
+        i = j + 1;
+        continue;
+      }
+    }
+
     if (/[0-9]/.test(c) || (c === '.' && /[0-9]/.test(sql[i + 1] || ''))) {
       let j = i;
       let isFloat = false;
@@ -219,9 +264,16 @@ function tokenize(sql) {
       continue;
     }
 
-    if (/[a-zA-Z_$]/.test(c)) {
-      let j = i;
-      while (j < n && /[a-zA-Z0-9_$]/.test(sql[j])) j++;
+    // Unicode 标识符：标准 SQL 允许非 ASCII 字母做列名/表名（中文、emoji 等）
+    // \p{L} = Unicode 字母，\p{So} = Unicode 其他符号（含 emoji），\p{M} = 组合标记
+    // 注意：不包含 \p{Sm}（数学符号如 + = < >）和 \p{P}（标点如 ( ) , ;）
+    if (/[\p{L}\p{So}\p{Sc}\p{M}_$]/u.test(c)) {
+      let j = i + cLen;
+      while (j < n) {
+        const { char: jc, length: jLen } = charAt(j);
+        if (!/[\p{L}\p{N}\p{So}\p{Sc}\p{M}_$]/u.test(jc)) break;
+        j += jLen;
+      }
       const word = sql.slice(i, j);
       const upper = word.toUpperCase();
       const tok = new SQLToken(KEYWORDS.has(upper) ? 'keyword' : 'ident', word, i);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "jsql-neo",
-  "version": "6.4.0-beta.3",
+  "version": "6.4.0-beta.4",
   "description": "JSQL-NEO — Rust-powered embedded database with WASM, REST API, B-Tree indexes, WAL, crash recovery",
   "main": "index.js",
   "types": "index.d.ts",