@swf2es/runtime 0.1.4 → 0.1.5

package/dist/avm2/utf8.js

@@ -0,0 +1,168 @@
+// UTF-8 as avmplus reads and writes it: an invalid byte reads as the
+// character of its value, a lone surrogate writes as U+FFFD. ByteArray,
+// AMF and the shell's natives share it.
+//
+// Translated from avmplus' core/UnicodeUtils.cpp, this file is subject to
+// the Mozilla Public License, v. 2.0: http://mozilla.org/MPL/2.0/.
+/** The host's UTF-8 encoder, which writes a lone surrogate as U+FFFD, as utf8() does. */
+const encoder = new TextEncoder();
+/**
+ * Below this length, a string encodes faster here than through
+ * TextEncoder, whose call costs more than the loop saves (about 90
+ * characters in V8).
+ */
+const ENCODER_LENGTH = 96;
+/**
+ * As UnicodeUtils::Utf16ToUtf8: a surrogate pair as four bytes, a lone
+ * surrogate as U+FFFD. avmshell's writeUTFBytes loses characters around a
+ * lone one instead; this keeps them, and the UTF-8 valid.
+ */
+export function utf8(s) {
+    if (s.length >= ENCODER_LENGTH) {
+        return encoder.encode(s);
+    }
+    // ASCII, byte for byte.
+    let ascii = true;
+    for (let i = 0; i < s.length && ascii; i++) {
+        ascii = s.charCodeAt(i) < 0x80;
+    }
+    if (ascii) {
+        const out = new Uint8Array(s.length);
+        for (let i = 0; i < s.length; i++) {
+            out[i] = s.charCodeAt(i);
+        }
+        return out;
+    }
+    const out = new Uint8Array(s.length * 3);
+    let n = 0;
+    for (let i = 0; i < s.length; i++) {
+        let ch = s.charCodeAt(i);
+        if (ch < 0x80) {
+            out[n++] = ch;
+            continue;
+        }
+        if (ch < 0x800) {
+            out[n++] = 0xc0 | ((ch >> 6) & 0x1f);
+            out[n++] = 0x80 | (ch & 0x3f);
+            continue;
+        }
+        if (ch >= 0xd800 && ch <= 0xdbff) {
+            const ch2 = i + 1 < s.length ? s.charCodeAt(i + 1) : 0;
+            if (ch2 >= 0xdc00 && ch2 <= 0xdfff) {
+                i++;
+                const ucs4 = ((ch - 0xd800) << 10) + (ch2 - 0xdc00) + 0x10000;
+                out[n++] = 0xf0 | ((ucs4 >> 18) & 0x07);
+                out[n++] = 0x80 | ((ucs4 >> 12) & 0x3f);
+                out[n++] = 0x80 | ((ucs4 >> 6) & 0x3f);
+                out[n++] = 0x80 | (ucs4 & 0x3f);
+                continue;
+            }
+            ch = 0xfffd;
+        }
+        else if (ch >= 0xdc00 && ch <= 0xdfff) {
+            ch = 0xfffd;
+        }
+        out[n++] = 0xe0 | ((ch >> 12) & 0x0f);
+        out[n++] = 0x80 | ((ch >> 6) & 0x3f);
+        out[n++] = 0x80 | (ch & 0x3f);
+    }
+    return out.subarray(0, n);
+}
+/**
+ * As UnicodeUtils::Utf8ToUtf16, not strict: a byte that does not start a
+ * valid sequence is the character of its value, and a four-byte sequence
+ * may start with 0xF8 to 0xFF, as Flash Player 9 and 10 read them.
+ */
+/** The length from which fromUtf8 first looks for ASCII. */
+const ASCII_LENGTH = 16;
+export function fromUtf8(bytes) {
+    const n = bytes.length;
+    // ASCII, as most text is: its bytes are its characters. Below
+    // ASCII_LENGTH bytes, a character at a time, until one is not.
+    if (n < ASCII_LENGTH) {
+        let s = "";
+        let k = 0;
+        for (; k < n && bytes[k] < 0x80; k++) {
+            s += String.fromCharCode(bytes[k]);
+        }
+        if (k === n) {
+            return s;
+        }
+    }
+    else {
+        let ascii = 0;
+        while (ascii < n && bytes[ascii] < 0x80) {
+            ascii++;
+        }
+        if (ascii === n) {
+            let s = "";
+            for (let k = 0; k < n; k += 8192) {
+                s += String.fromCharCode.apply(null, bytes.subarray(k, k + 8192));
+            }
+            return s;
+        }
+    }
+    const out = [];
+    let i = 0;
+    while (i < n) {
+        const c = bytes[i];
+        const left = n - i;
+        switch (c >> 4) {
+            case 12:
+            case 13: {
+                if (left >= 2 && (bytes[i + 1] & 0xc0) === 0x80) {
+                    const ch = ((c << 6) & 0x7c0) | (bytes[i + 1] & 0x3f);
+                    if (ch >= 0x80) {
+                        out.push(ch);
+                        i += 2;
+                        continue;
+                    }
+                }
+                break;
+            }
+            case 14: {
+                if (left >= 3 && (bytes[i + 1] & 0xc0) === 0x80 && (bytes[i + 2] & 0xc0) === 0x80) {
+                    const ch = ((c << 12) & 0xf000) | ((bytes[i + 1] << 6) & 0xfc0) | (bytes[i + 2] & 0x3f);
+                    if (ch >= 0x800) {
+                        out.push(ch);
+                        i += 3;
+                        continue;
+                    }
+                }
+                break;
+            }
+            case 15: {
+                if (left >= 4 &&
+                    (bytes[i + 1] & 0xc0) === 0x80 &&
+                    (bytes[i + 2] & 0xc0) === 0x80 &&
+                    (bytes[i + 3] & 0xc0) === 0x80) {
+                    const ch = ((c << 18) & 0x1c0000) |
+                        ((bytes[i + 1] << 12) & 0x3f000) |
+                        ((bytes[i + 2] << 6) & 0xfc0) |
+                        (bytes[i + 3] & 0x3f);
+                    if (ch >= 0x10000) {
+                        out.push((((ch - 0x10000) >> 10) & 0x3ff) + 0xd800, ((ch - 0x10000) & 0x3ff) + 0xdc00);
+                        i += 4;
+                        continue;
+                    }
+                }
+                break;
+            }
+            default:
+                break;
+        }
+        // Not a sequence: the byte itself.
+        out.push(c);
+        i++;
+    }
+    return fromCodes(out);
+}
+/** A string of UTF-16 code units, in chunks, as String.fromCharCode takes only so many arguments. */
+export function fromCodes(units) {
+    let s = "";
+    for (let k = 0; k < units.length; k += 8192) {
+        s += String.fromCharCode(...units.slice(k, k + 8192));
+    }
+    return s;
+}
+//# sourceMappingURL=utf8.js.map

package/dist/avm2/utf8.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"utf8.js","sourceRoot":"","sources":["../../src/avm2/utf8.ts"],"names":[],"mappings":"AAAA,qEAAqE;AACrE,wEAAwE;AACxE,wCAAwC;AACxC,EAAE;AACF,0EAA0E;AAC1E,mEAAmE;AAMnE,yFAAyF;AACzF,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAC;AAElC;;;;GAIG;AACH,MAAM,cAAc,GAAG,EAAE,CAAC;AAE1B;;;;GAIG;AACH,MAAM,UAAU,IAAI,CAAC,CAAS;IAC5B,IAAI,CAAC,CAAC,MAAM,IAAI,cAAc,EAAE,CAAC;QAC/B,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;IAC3B,CAAC;IAED,wBAAwB;IACxB,IAAI,KAAK,GAAG,IAAI,CAAC;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC;IACjC,CAAC;IAED,IAAI,KAAK,EAAE,CAAC;QACV,MAAM,GAAG,GAAG,IAAI,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QACrC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAClC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAC3B,CAAC;QAED,OAAO,GAAG,CAAC;IACb,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,UAAU,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACzC,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClC,IAAI,EAAE,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QACzB,IAAI,EAAE,GAAG,IAAI,EAAE,CAAC;YACd,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;YACd,SAAS;QACX,CAAC;QAED,IAAI,EAAE,GAAG,KAAK,EAAE,CAAC;YACf,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;YACrC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,EAAE,GAAG,IAAI,CAAC,CAAC;YAC9B,SAAS;QACX,CAAC;QAED,IAAI,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,EAAE,CAAC;YACjC,MAAM,GAAG,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACvD,IAAI,GAAG,IAAI,MAAM,IAAI,GAAG,IAAI,MAAM,EAAE,CAAC;gBACnC,CAAC,EAAE,CAAC;gBACJ,MAAM,IAAI,GAAG,CAAC,CAAC,EAAE,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,GAAG,GAAG,MAAM,CAAC,GAAG,OAAO,CAAC;gBAC9D,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC;gBACxC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC;gBACxC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;gBACvC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC,CAAC;gBAChC,SAAS;YACX,CAAC;YAED,EAAE,GAAG,MAAM,CAAC;QACd,CAAC;aAAM,IAAI,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,EAAE,CAAC;YACxC,EAAE,GAAG,MAAM,CAAC;QACd,CAAC;QAED,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC;QACtC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;QACrC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,EAAE,GAAG,IAAI,CAAC,CAAC;IAChC,CAAC;IAED,OAAO,GAAG,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;AAC5B,CAAC;AAED;;;;GAIG;AACH,4DAA4D;AAC5D,MAAM,YAAY,GAAG,EAAE,CAAC;AAExB,MAAM,UAAU,QAAQ,CAAC,KAAiB;IACxC,MAAM,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;IACvB,8DAA8D;IAC9D,+DAA+D;IAC/D,IAAI,CAAC,GAAG,YAAY,EAAE,CAAC;QACrB,IAAI,CAAC,GAAG,EAAE,CAAC;QACX,IAAI,CAAC,GAAG,CAAC,CAAC;QACV,OAAO,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,GAAG,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC;YACrC,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACrC,CAAC;QAED,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;YACZ,OAAO,CAAC,CAAC;QACX,CAAC;IACH,CAAC;SAAM,CAAC;QACN,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,OAAO,KAAK,GAAG,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,GAAG,IAAI,EAAE,CAAC;YACxC,KAAK,EAAE,CAAC;QACV,CAAC;QAED,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;YAChB,IAAI,CAAC,GAAG,EAAE,CAAC;YACX,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC;gBACjC,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,CAAwB,CAAC,CAAC;YAC3F,CAAC;YAED,OAAO,CAAC,CAAC;QACX,CAAC;IACH,CAAC;IAED,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QACb,MAAM,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACnB,MAAM,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QACnB,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;YACf,KAAK,EAAE,CAAC;YACR,KAAK,EAAE,EAAE,CAAC;gBACR,IAAI,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;oBAChD,MAAM,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;oBACtD,IAAI,EAAE,IAAI,IAAI,EAAE,CAAC;wBACf,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;wBACb,CAAC,IAAI,CAAC,CAAC;wBACP,SAAS;oBACX,CAAC;gBACH,CAAC;gBACD,MAAM;YACR,CAAC;YACD,KAAK,EAAE,EAAE,CAAC;gBACR,IAAI,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;oBAClF,MAAM,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;oBACxF,IAAI,EAAE,IAAI,KAAK,EAAE,CAAC;wBAChB,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;wBACb,CAAC,IAAI,CAAC,CAAC;wBACP,SAAS;oBACX,CAAC;gBACH,CAAC;gBACD,MAAM;YACR,CAAC;YACD,KAAK,EAAE,EAAE,CAAC;gBACR,IACE,IAAI,IAAI,CAAC;oBACT,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,IAAI;oBAC9B,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,IAAI;oBAC9B,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,IAAI,EAC9B,CAAC;oBACD,MAAM,EAAE,GACN,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,GAAG,QAAQ,CAAC;wBACtB,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,GAAG,OAAO,CAAC;wBAChC,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC;wBAC7B,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;oBACxB,IAAI,EAAE,IAAI,OAAO,EAAE,CAAC;wBAClB,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC,GAAG,KAAK,CAAC,GAAG,MAAM,EAAE,CAAC,CAAC,EAAE,GAAG,OAAO,CAAC,GAAG,KAAK,CAAC,GAAG,MAAM,CAAC,CAAC;wBACvF,CAAC,IAAI,CAAC,CAAC;wBACP,SAAS;oBACX,CAAC;gBACH,CAAC;gBACD,MAAM;YACR,CAAC;YACD;gBACE,MAAM;QACV,CAAC;QAED,mCAAmC;QACnC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACZ,CAAC,EAAE,CAAC;IACN,CAAC;IAED,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC;AACxB,CAAC;AAED,qGAAqG;AACrG,MAAM,UAAU,SAAS,CAAC,KAAe;IACvC,IAAI,CAAC,GAAG,EAAE,CAAC;IACX,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC;QAC5C,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,OAAO,CAAC,CAAC;AACX,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@swf2es/runtime",
-  "version": "0.1.4",
+  "version": "0.1.5",
   "description": "AVM2 and AVM1 language runtime used by generated code",
   "keywords": [
     "actionscript",
@@ -41,6 +41,6 @@
     "LICENSE.third-party"
   ],
   "dependencies": {
-    "@swf2es/format": "0.1.4"
+    "@swf2es/format": "0.1.5"
   }
 }

package/src/avm1/errors.ts

@@ -23,7 +23,7 @@ export class Halt extends Error {
  * own check (adl; Ruffle's continue_if_base_clip_exists). One instance,
  * thrown as a signal and never shown.
  */
-export class BaseRemoved extends Error {
+class BaseRemoved extends Error {
   override name = "Avm1BaseRemoved";
 }
 

package/src/avm1/numbers.ts

@@ -52,16 +52,13 @@ export function numberToString(n: number): string {
     return "NaN";
   }
 
-  if (n === Number.POSITIVE_INFINITY) {
-    return "Infinity";
-  }
-
-  if (n === Number.NEGATIVE_INFINITY) {
-    return "-Infinity";
-  }
-
-  if (n === 0) {
-    return "0";
+  switch (n) {
+    case Number.POSITIVE_INFINITY:
+      return "Infinity";
+    case Number.NEGATIVE_INFINITY:
+      return "-Infinity";
+    case 0:
+      return "0";
   }
 
   if (n >= -2147483648 && n <= 2147483647 && Number.isInteger(n)) {

package/src/avm2/amf.ts

@@ -15,10 +15,11 @@
 import type { AsObject, Value } from "./descriptors.js";
 import { FloatValue } from "./float.js";
 import { NS_PackageInternal, NS_Public, namespace, publicQname, qname } from "./names.js";
-import { type Bytes, bytesOf, fromUtf8, utf8 } from "./natives/bytearray.js";
+import { type Bytes, bytesOf } from "./natives/bytearray.js";
 import { xmlToXMLString } from "./natives/xml/strings.js";
 import type { Runtime } from "./runtime.js";
 import type { Traits } from "./traits.js";
+import { fromUtf8, utf8 } from "./utf8.js";
 
 const kUndefined = 0;
 const kNull = 1;

package/src/avm2/names.ts

@@ -11,9 +11,6 @@ import { NO_CACHE, type PropertyCache } from "./property-cache.js";
 /** Namespace kinds, numbered as the compiler's NS_* constants. */
 export const NS_Public = 0;
 export const NS_PackageInternal = 1;
-export const NS_Protected = 2;
-export const NS_Explicit = 3;
-export const NS_StaticProtected = 4;
 export const NS_Private = 5;
 
 export class Namespace {
@@ -87,7 +84,6 @@ export const CONSTANT_Multiname = 0x09;
 export const CONSTANT_MultinameA = 0x0e;
 export const CONSTANT_MultinameL = 0x1b;
 export const CONSTANT_MultinameLA = 0x1c;
-export const CONSTANT_TypeName = 0x1d;
 
 /**
  * A name as a property lookup uses it: its namespaces, each with the API

package/src/avm2/natives/bytearray.ts

@@ -2,12 +2,12 @@
 // DataInput/DataOutput implement them. The buffer keeps avmplus' capacity:
 // it grows as avmplus' Grower grows it, zeroing what a reallocation adds,
 // so bytes a shorter length left behind come back when the length grows
-// within the capacity, as they do in avmplus. UTF-8 is avmplus' own: an
-// invalid byte reads as the character of its value.
+// within the capacity, as they do in avmplus. UTF-8 is avmplus' own
+// (utf8.ts).
 //
-// Translated from avmplus' core/ByteArrayGlue.cpp, core/DataIO.cpp and
-// core/UnicodeUtils.cpp, this file is subject to the Mozilla Public
-// License, v. 2.0: http://mozilla.org/MPL/2.0/.
+// Translated from avmplus' core/ByteArrayGlue.cpp and core/DataIO.cpp,
+// this file is subject to the Mozilla Public License, v. 2.0:
+// http://mozilla.org/MPL/2.0/.
 import {
   CompressedDataError,
   CompressedDataTooLargeError,
@@ -23,6 +23,7 @@ import type { AsObject, Method, Value } from "../descriptors.js";
 import type { IndexHook } from "../hooks.js";
 import type { Runtime } from "../runtime.js";
 import type { Traits } from "../traits.js";
+import { fromCodes, fromUtf8, utf8 } from "../utf8.js";
 import { type Natives, registerNativeClass } from "./define.js";
 
 const kGrowthIncr = 4096;
@@ -653,194 +654,6 @@ export function knownSharedByteArray(rt: Runtime, id: number): AsObject | undefi
   return sharedObjectsOf(rt).get(id)?.deref();
 }
 
-// TextEncoder is the web's and node's alike, though not ECMAScript's,
-// whose library alone the runtime is typed against.
-declare const TextEncoder: new () => { encode(s: string): Uint8Array };
-
-/** The host's UTF-8 encoder, which writes a lone surrogate as U+FFFD, as utf8() does. */
-const encoder = new TextEncoder();
-
-/**
- * Below this length, a string encodes faster here than through
- * TextEncoder, whose call costs more than the loop saves (about 90
- * characters in V8).
- */
-const ENCODER_LENGTH = 96;
-
-/**
- * As UnicodeUtils::Utf16ToUtf8: a surrogate pair as four bytes, a lone
- * surrogate as U+FFFD. avmshell's writeUTFBytes loses characters around a
- * lone one instead; this keeps them, and the UTF-8 valid.
- */
-export function utf8(s: string): Uint8Array {
-  if (s.length >= ENCODER_LENGTH) {
-    return encoder.encode(s);
-  }
-
-  // ASCII, byte for byte.
-  let ascii = true;
-  for (let i = 0; i < s.length && ascii; i++) {
-    ascii = s.charCodeAt(i) < 0x80;
-  }
-
-  if (ascii) {
-    const out = new Uint8Array(s.length);
-    for (let i = 0; i < s.length; i++) {
-      out[i] = s.charCodeAt(i);
-    }
-
-    return out;
-  }
-
-  const out = new Uint8Array(s.length * 3);
-  let n = 0;
-  for (let i = 0; i < s.length; i++) {
-    let ch = s.charCodeAt(i);
-    if (ch < 0x80) {
-      out[n++] = ch;
-      continue;
-    }
-
-    if (ch < 0x800) {
-      out[n++] = 0xc0 | ((ch >> 6) & 0x1f);
-      out[n++] = 0x80 | (ch & 0x3f);
-      continue;
-    }
-
-    if (ch >= 0xd800 && ch <= 0xdbff) {
-      const ch2 = i + 1 < s.length ? s.charCodeAt(i + 1) : 0;
-      if (ch2 >= 0xdc00 && ch2 <= 0xdfff) {
-        i++;
-        const ucs4 = ((ch - 0xd800) << 10) + (ch2 - 0xdc00) + 0x10000;
-        out[n++] = 0xf0 | ((ucs4 >> 18) & 0x07);
-        out[n++] = 0x80 | ((ucs4 >> 12) & 0x3f);
-        out[n++] = 0x80 | ((ucs4 >> 6) & 0x3f);
-        out[n++] = 0x80 | (ucs4 & 0x3f);
-        continue;
-      }
-
-      ch = 0xfffd;
-    } else if (ch >= 0xdc00 && ch <= 0xdfff) {
-      ch = 0xfffd;
-    }
-
-    out[n++] = 0xe0 | ((ch >> 12) & 0x0f);
-    out[n++] = 0x80 | ((ch >> 6) & 0x3f);
-    out[n++] = 0x80 | (ch & 0x3f);
-  }
-
-  return out.subarray(0, n);
-}
-
-/**
- * As UnicodeUtils::Utf8ToUtf16, not strict: a byte that does not start a
- * valid sequence is the character of its value, and a four-byte sequence
- * may start with 0xF8 to 0xFF, as Flash Player 9 and 10 read them.
- */
-/** The length from which fromUtf8 first looks for ASCII. */
-const ASCII_LENGTH = 16;
-
-export function fromUtf8(bytes: Uint8Array): string {
-  const n = bytes.length;
-  // ASCII, as most text is: its bytes are its characters. Below
-  // ASCII_LENGTH bytes, a character at a time, until one is not.
-  if (n < ASCII_LENGTH) {
-    let s = "";
-    let k = 0;
-    for (; k < n && bytes[k] < 0x80; k++) {
-      s += String.fromCharCode(bytes[k]);
-    }
-
-    if (k === n) {
-      return s;
-    }
-  } else {
-    let ascii = 0;
-    while (ascii < n && bytes[ascii] < 0x80) {
-      ascii++;
-    }
-
-    if (ascii === n) {
-      let s = "";
-      for (let k = 0; k < n; k += 8192) {
-        s += String.fromCharCode.apply(null, bytes.subarray(k, k + 8192) as unknown as number[]);
-      }
-
-      return s;
-    }
-  }
-
-  const out: number[] = [];
-  let i = 0;
-  while (i < n) {
-    const c = bytes[i];
-    const left = n - i;
-    switch (c >> 4) {
-      case 12:
-      case 13: {
-        if (left >= 2 && (bytes[i + 1] & 0xc0) === 0x80) {
-          const ch = ((c << 6) & 0x7c0) | (bytes[i + 1] & 0x3f);
-          if (ch >= 0x80) {
-            out.push(ch);
-            i += 2;
-            continue;
-          }
-        }
-        break;
-      }
-      case 14: {
-        if (left >= 3 && (bytes[i + 1] & 0xc0) === 0x80 && (bytes[i + 2] & 0xc0) === 0x80) {
-          const ch = ((c << 12) & 0xf000) | ((bytes[i + 1] << 6) & 0xfc0) | (bytes[i + 2] & 0x3f);
-          if (ch >= 0x800) {
-            out.push(ch);
-            i += 3;
-            continue;
-          }
-        }
-        break;
-      }
-      case 15: {
-        if (
-          left >= 4 &&
-          (bytes[i + 1] & 0xc0) === 0x80 &&
-          (bytes[i + 2] & 0xc0) === 0x80 &&
-          (bytes[i + 3] & 0xc0) === 0x80
-        ) {
-          const ch =
-            ((c << 18) & 0x1c0000) |
-            ((bytes[i + 1] << 12) & 0x3f000) |
-            ((bytes[i + 2] << 6) & 0xfc0) |
-            (bytes[i + 3] & 0x3f);
-          if (ch >= 0x10000) {
-            out.push((((ch - 0x10000) >> 10) & 0x3ff) + 0xd800, ((ch - 0x10000) & 0x3ff) + 0xdc00);
-            i += 4;
-            continue;
-          }
-        }
-        break;
-      }
-      default:
-        break;
-    }
-
-    // Not a sequence: the byte itself.
-    out.push(c);
-    i++;
-  }
-
-  return fromCodes(out);
-}
-
-/** A string of UTF-16 code units, in chunks, as String.fromCharCode takes only so many arguments. */
-function fromCodes(units: number[]): string {
-  let s = "";
-  for (let k = 0; k < units.length; k += 8192) {
-    s += String.fromCharCode(...units.slice(k, k + 8192));
-  }
-
-  return s;
-}
-
 /** The bytes up to the first NUL, as avmplus reads a C string. */
 function toNul(bytes: Uint8Array): Uint8Array {
   const nul = bytes.indexOf(0);

package/src/avm2/natives/regexp.ts

@@ -212,30 +212,28 @@ function variableLookbehind(source: string): boolean {
   /** The length of the item at i, which it steps past. */
   const item = (): number => {
     const c = source[i++];
-    if (c === "(") {
-      return group();
-    }
+    switch (c) {
+      case "(":
+        return group();
+      case "^":
+      case "$":
+        return 0;
+      case "[":
+        if (source[i] === "^") {
+          i++;
+        }
 
-    if (c === "^" || c === "$") {
-      return 0;
-    }
+        // A ] first is a literal.
+        if (source[i] === "]") {
+          i++;
+        }
 
-    if (c === "[") {
-      if (source[i] === "^") {
-        i++;
-      }
+        while (i < source.length && source[i] !== "]") {
+          i += source[i] === "\\" ? 2 : 1;
+        }
 
-      // A ] first is a literal.
-      if (source[i] === "]") {
         i++;
-      }
-
-      while (i < source.length && source[i] !== "]") {
-        i += source[i] === "\\" ? 2 : 1;
-      }
-
-      i++;
-      return 1;
+        return 1;
     }
 
     if (c !== "\\") {
@@ -324,62 +322,66 @@ function nestedTooDeep(source: string): boolean {
   for (let i = 0; i < source.length; i++) {
     const c = source[i];
     let item = 2;
-    if (c === "(") {
-      const capture = source[i + 1] !== "?" || source.startsWith("(?P<", i);
-      outer.push(held);
-      held += previous + (capture ? 5 : 3);
-      previous = 0;
-      if (held > WORKSPACE) {
-        return true;
-      }
-
-      // Past the group's kind: ?: ?= ?! ?<= ?<! ?P<name> or flags.
-      if (source[i + 1] === "?") {
-        const end = source.slice(i + 2).search(/[:=!>)]/);
-        i = end < 0 ? source.length : i + 2 + end;
-      }
-
-      continue;
-    }
-
-    if (c === ")") {
-      held = outer.pop() ?? held;
-      // The group is the next one's previous item, but what it holds is gone.
-      previous = 0;
-      continue;
-    }
+    switch (c) {
+      case "(": {
+        const capture = source[i + 1] !== "?" || source.startsWith("(?P<", i);
+        outer.push(held);
+        held += previous + (capture ? 5 : 3);
+        previous = 0;
+        if (held > WORKSPACE) {
+          return true;
+        }
 
-    if (c === "|") {
-      previous = 0;
-      continue;
-    }
+        // Past the group's kind: ?: ?= ?! ?<= ?<! ?P<name> or flags.
+        if (source[i + 1] === "?") {
+          const end = source.slice(i + 2).search(/[:=!>)]/);
+          i = end < 0 ? source.length : i + 2 + end;
+        }
 
-    if (c === "\\") {
-      i++;
-    } else if (c === "[") {
-      // A class is its opcode and a 32-byte bitmap.
-      item = 33;
-      i++;
-      if (source[i] === "^") {
-        i++;
+        continue;
       }
-
-      // A ] first is a literal.
-      if (source[i] === "]") {
+      case ")":
+        held = outer.pop() ?? held;
+        // The group is the next one's previous item, but what it holds is gone.
+        previous = 0;
+        continue;
+      case "|":
+        previous = 0;
+        continue;
+      case "\\":
         i++;
-      }
+        break;
+      case "[":
+        // A class is its opcode and a 32-byte bitmap.
+        item = 33;
+        i++;
+        if (source[i] === "^") {
+          i++;
+        }
 
-      while (i < source.length && source[i] !== "]") {
-        if (source[i] === "\\") {
+        // A ] first is a literal.
+        if (source[i] === "]") {
           i++;
         }
 
-        i++;
-      }
-    } else if (c === "." || c === "^" || c === "$") {
-      item = 1;
-    } else if ("*+?{".includes(c)) {
-      continue;
+        while (i < source.length && source[i] !== "]") {
+          if (source[i] === "\\") {
+            i++;
+          }
+
+          i++;
+        }
+        break;
+      case ".":
+      case "^":
+      case "$":
+        item = 1;
+        break;
+      case "*":
+      case "+":
+      case "?":
+      case "{":
+        continue;
     }
 
     previous = item;

package/src/avm2/natives/shell-workers.ts

@@ -11,7 +11,7 @@
 
 import type { Value } from "../descriptors.js";
 import type { ShellWorkerInit, ShellWorkersMemory } from "../options.js";
-import { fromUtf8, utf8 } from "./bytearray.js";
+import { fromUtf8, utf8 } from "../utf8.js";
 import { Sync, type SyncMemory } from "./sync.js";
 
 /** avmplus' Isolate::State. */

package/src/avm2/natives/shell.ts

@@ -9,15 +9,14 @@ import type { ClassHook } from "../hooks.js";
 import type { ShellThread } from "../options.js";
 import type { Runtime } from "../runtime.js";
 import { setStaticVar } from "../traits.js";
+import { fromUtf8, utf8 } from "../utf8.js";
 import {
   byteArrayCapacity,
   bytesOf,
-  fromUtf8,
   GLOBAL_MEMORY_MIN_SIZE,
   knownSharedByteArray,
   setDomainMemory,
   sharedByteArray,
-  utf8,
 } from "./bytearray.js";
 import { syncObject, syncOf, syncRef } from "./concurrent.js";
 import { elements, type Natives, registerNativeClass } from "./define.js";