mol_dump_lib 0.0.869 → 0.0.871

package/node.mjs

@@ -3088,12 +3088,41 @@ var $;
 "use strict";
 var $;
 (function ($) {
-    const TextEncoder = globalThis.TextEncoder ?? $node.util.TextEncoder;
-    const encoder = new TextEncoder();
-    function $mol_charset_encode(value) {
-        return encoder.encode(value);
+    let buf = new Uint8Array(2 ** 12);
+    function $mol_charset_encode(str) {
+        const capacity = str.length * 3;
+        if (buf.byteLength < capacity)
+            buf = new Uint8Array(capacity);
+        return buf.slice(0, $mol_charset_encode_to(str, buf));
     }
     $.$mol_charset_encode = $mol_charset_encode;
+    function $mol_charset_encode_to(str, buf, from = 0) {
+        let pos = from;
+        for (let i = 0; i < str.length; i++) {
+            let code = str.charCodeAt(i);
+            if (code < 0x80) {
+                buf[pos++] = code;
+            }
+            else if (code < 0x800) {
+                buf[pos++] = 0xc0 | (code >> 6);
+                buf[pos++] = 0x80 | (code & 0x3f);
+            }
+            else if (code < 0xd800 || code >= 0xe000) {
+                buf[pos++] = 0xe0 | (code >> 12);
+                buf[pos++] = 0x80 | ((code >> 6) & 0x3f);
+                buf[pos++] = 0x80 | (code & 0x3f);
+            }
+            else {
+                const point = ((code - 0xd800) << 10) + str.charCodeAt(++i) + 0x2400;
+                buf[pos++] = 0xf0 | (point >> 18);
+                buf[pos++] = 0x80 | ((point >> 12) & 0x3f);
+                buf[pos++] = 0x80 | ((point >> 6) & 0x3f);
+                buf[pos++] = 0x80 | (point & 0x3f);
+            }
+        }
+        return pos - from;
+    }
+    $.$mol_charset_encode_to = $mol_charset_encode_to;
 })($ || ($ = {}));
 
 ;

package/node.test.js

@@ -3079,12 +3079,41 @@ var $;
 "use strict";
 var $;
 (function ($) {
-    const TextEncoder = globalThis.TextEncoder ?? $node.util.TextEncoder;
-    const encoder = new TextEncoder();
-    function $mol_charset_encode(value) {
-        return encoder.encode(value);
+    let buf = new Uint8Array(2 ** 12);
+    function $mol_charset_encode(str) {
+        const capacity = str.length * 3;
+        if (buf.byteLength < capacity)
+            buf = new Uint8Array(capacity);
+        return buf.slice(0, $mol_charset_encode_to(str, buf));
     }
     $.$mol_charset_encode = $mol_charset_encode;
+    function $mol_charset_encode_to(str, buf, from = 0) {
+        let pos = from;
+        for (let i = 0; i < str.length; i++) {
+            let code = str.charCodeAt(i);
+            if (code < 0x80) {
+                buf[pos++] = code;
+            }
+            else if (code < 0x800) {
+                buf[pos++] = 0xc0 | (code >> 6);
+                buf[pos++] = 0x80 | (code & 0x3f);
+            }
+            else if (code < 0xd800 || code >= 0xe000) {
+                buf[pos++] = 0xe0 | (code >> 12);
+                buf[pos++] = 0x80 | ((code >> 6) & 0x3f);
+                buf[pos++] = 0x80 | (code & 0x3f);
+            }
+            else {
+                const point = ((code - 0xd800) << 10) + str.charCodeAt(++i) + 0x2400;
+                buf[pos++] = 0xf0 | (point >> 18);
+                buf[pos++] = 0x80 | ((point >> 12) & 0x3f);
+                buf[pos++] = 0x80 | ((point >> 6) & 0x3f);
+                buf[pos++] = 0x80 | (point & 0x3f);
+            }
+        }
+        return pos - from;
+    }
+    $.$mol_charset_encode_to = $mol_charset_encode_to;
 })($ || ($ = {}));
 
 ;
@@ -10004,10 +10033,23 @@ var $;
 var $;
 (function ($) {
     $mol_test({
-        'encode utf8 string'() {
-            const str = 'Hello, ΧΨΩЫ';
-            const encoded = new Uint8Array([72, 101, 108, 108, 111, 44, 32, 206, 167, 206, 168, 206, 169, 208, 171]);
-            $mol_assert_like($mol_charset_encode(str), encoded);
+        'encode empty'() {
+            $mol_assert_equal($mol_charset_encode(''), new Uint8Array([]));
+        },
+        'encode 1 octet'() {
+            $mol_assert_equal($mol_charset_encode('F'), new Uint8Array([0x46]));
+        },
+        'encode 2 octet'() {
+            $mol_assert_equal($mol_charset_encode('Б'), new Uint8Array([0xd0, 0x91]));
+        },
+        'encode 3 octet'() {
+            $mol_assert_equal($mol_charset_encode('ह'), new Uint8Array([0xe0, 0xa4, 0xb9]));
+        },
+        'encode 4 octet'() {
+            $mol_assert_equal($mol_charset_encode('𐍈'), new Uint8Array([0xf0, 0x90, 0x8d, 0x88]));
+        },
+        'encode surrogate pair'() {
+            $mol_assert_equal($mol_charset_encode('😀'), new Uint8Array([0xf0, 0x9f, 0x98, 0x80]));
         },
     });
 })($ || ($ = {}));