fix(ext/node): Reimplement StringDecoder to match node's behavior (#22933)

Fixes #22158.

Basically reimplements the whole `StringDecoder` with a much more direct
translation (read like one-to-one) of node's current logic. The old
implementation was closer to node's super old impl and it was too hard
to keep the code structure while matching the behavior of their new
logic.

This adds support for UTF-16LE, ascii, and latin1. 

This also enables the node_compat test, which now passes without
modification.

1 files changed, 331 insertions, 262 deletions

diff --git a/ext/node/polyfills/string_decoder.ts b/ext/node/polyfills/string_decoder.ts
index 507a994bb..ef83b6fc9 100644
--- a/ext/node/polyfills/string_decoder.ts
+++ b/ext/node/polyfills/string_decoder.ts
@@ -23,23 +23,38 @@
 // TODO(petamoriken): enable prefer-primordials for node polyfills
 // deno-lint-ignore-file prefer-primordials
 
-import { Buffer } from "node:buffer";
+// Logic and comments translated pretty much one-to-one from node's impl
+// (https://github.com/nodejs/node/blob/ba06c5c509956dc413f91b755c1c93798bb700d4/src/string_decoder.cc)
+
+import { Buffer, constants } from "node:buffer";
+import { normalizeEncoding as castEncoding } from "ext:deno_node/_utils.ts";
 import {
-  normalizeEncoding as castEncoding,
-  notImplemented,
-} from "ext:deno_node/_utils.ts";
-
-enum NotImplemented {
-  "ascii",
-  "latin1",
-  "utf16le",
-}
+  ERR_INVALID_ARG_TYPE,
+  ERR_INVALID_THIS,
+  ERR_UNKNOWN_ENCODING,
+  NodeError,
+} from "ext:deno_node/internal/errors.ts";
+
+import { primordials } from "ext:core/mod.js";
+const {
+  ArrayBufferIsView,
+  ObjectDefineProperties,
+} = primordials;
+
+const { MAX_STRING_LENGTH } = constants;
+
+// to cast from string to `BufferEncoding`, which doesn't seem nameable from here
+// deno-lint-ignore no-explicit-any
+type Any = any;
 
 function normalizeEncoding(enc?: string): string {
   const encoding = castEncoding(enc ?? null);
-  if (encoding && encoding in NotImplemented) notImplemented(encoding);
-  if (!encoding && typeof enc === "string" && enc.toLowerCase() !== "raw") {
-    throw new Error(`Unknown encoding: ${enc}`);
+  if (!encoding) {
+    if (typeof enc !== "string" || enc.toLowerCase() !== "raw") {
+      throw new ERR_UNKNOWN_ENCODING(
+        enc as Any,
+      );
+    }
   }
   return String(encoding);
 }
@@ -49,295 +64,349 @@ function normalizeEncoding(enc?: string): string {
  */
 
 function isBufferType(buf: Buffer) {
-  return buf instanceof ArrayBuffer && buf.BYTES_PER_ELEMENT;
+  return buf instanceof Buffer && buf.BYTES_PER_ELEMENT;
 }
 
-/*
- * Checks the type of a UTF-8 byte, whether it's ASCII, a leading byte, or a
- * continuation byte. If an invalid byte is detected, -2 is returned.
- */
-function utf8CheckByte(byte: number): number {
-  if (byte <= 0x7f) return 0;
-  else if (byte >> 5 === 0x06) return 2;
-  else if (byte >> 4 === 0x0e) return 3;
-  else if (byte >> 3 === 0x1e) return 4;
-  return byte >> 6 === 0x02 ? -1 : -2;
-}
-
-/*
- * Checks at most 3 bytes at the end of a Buffer in order to detect an
- * incomplete multi-byte UTF-8 character. The total number of bytes (2, 3, or 4)
- * needed to complete the UTF-8 character (if applicable) are returned.
- */
-function utf8CheckIncomplete(
-  self: StringDecoderBase,
-  buf: Buffer,
-  i: number,
-): number {
-  let j = buf.length - 1;
-  if (j < i) return 0;
-  let nb = utf8CheckByte(buf[j]);
-  if (nb >= 0) {
-    if (nb > 0) self.lastNeed = nb - 1;
-    return nb;
-  }
-  if (--j < i || nb === -2) return 0;
-  nb = utf8CheckByte(buf[j]);
-  if (nb >= 0) {
-    if (nb > 0) self.lastNeed = nb - 2;
-    return nb;
+function normalizeBuffer(buf: Buffer) {
+  if (!ArrayBufferIsView(buf)) {
+    throw new ERR_INVALID_ARG_TYPE(
+      "buf",
+      ["Buffer", "TypedArray", "DataView"],
+      buf,
+    );
   }
-  if (--j < i || nb === -2) return 0;
-  nb = utf8CheckByte(buf[j]);
-  if (nb >= 0) {
-    if (nb > 0) {
-      if (nb === 2) nb = 0;
-      else self.lastNeed = nb - 3;
-    }
-    return nb;
+  if (isBufferType(buf)) {
+    return buf;
+  } else {
+    return Buffer.from(
+      buf.buffer,
+    );
   }
-  return 0;
 }
 
-/*
- * Validates as many continuation bytes for a multi-byte UTF-8 character as
- * needed or are available. If we see a non-continuation byte where we expect
- * one, we "replace" the validated continuation bytes we've seen so far with
- * a single UTF-8 replacement character ('\ufffd'), to match v8's UTF-8 decoding
- * behavior. The continuation byte check is included three times in the case
- * where all of the continuation bytes for a character exist in the same buffer.
- * It is also done this way as a slight performance increase instead of using a
- * loop.
- */
-function utf8CheckExtraBytes(
-  self: StringDecoderBase,
+function bufferToString(
   buf: Buffer,
-): string | undefined {
-  if ((buf[0] & 0xc0) !== 0x80) {
-    self.lastNeed = 0;
-    return "\ufffd";
+  encoding?: string,
+  start?: number,
+  end?: number,
+): string {
+  const len = (end ?? buf.length) - (start ?? 0);
+  if (len > MAX_STRING_LENGTH) {
+    throw new NodeError("ERR_STRING_TOO_LONG", "string exceeds maximum length");
   }
-  if (self.lastNeed > 1 && buf.length > 1) {
-    if ((buf[1] & 0xc0) !== 0x80) {
-      self.lastNeed = 1;
-      return "\ufffd";
-    }
-    if (self.lastNeed > 2 && buf.length > 2) {
-      if ((buf[2] & 0xc0) !== 0x80) {
-        self.lastNeed = 2;
-        return "\ufffd";
+  return buf.toString(encoding as Any, start, end);
+}
+
+// the heart of the logic, decodes a buffer, storing
+// incomplete characters in a buffer if applicable
+function decode(this: StringDecoder, buf: Buffer) {
+  const enc = this.enc;
+
+  let bufIdx = 0;
+  let bufEnd = buf.length;
+
+  let prepend = "";
+  let rest = "";
+
+  if (
+    enc === Encoding.Utf8 || enc === Encoding.Utf16 || enc === Encoding.Base64
+  ) {
+    // check if we need to finish an incomplete char from the last chunk
+    // written. If we do, we copy the bytes into our `lastChar` buffer
+    // and prepend the completed char to the result of decoding the rest of the buffer
+    if (this[kMissingBytes] > 0) {
+      if (enc === Encoding.Utf8) {
+        // Edge case for incomplete character at a chunk boundary
+        // (see https://github.com/nodejs/node/blob/73025c4dec042e344eeea7912ed39f7b7c4a3991/src/string_decoder.cc#L74)
+        for (
+          let i = 0;
+          i < buf.length - bufIdx && i < this[kMissingBytes];
+          i++
+        ) {
+          if ((buf[i] & 0xC0) !== 0x80) {
+            // We expected a continuation byte, but got something else.
+            // Stop trying to decode the incomplete char, and assume
+            // the byte we got starts a new char.
+            this[kMissingBytes] = 0;
+            buf.copy(this.lastChar, this[kBufferedBytes], bufIdx, bufIdx + i);
+            this[kBufferedBytes] += i;
+            bufIdx += i;
+            break;
+          }
+        }
+      }
+
+      const bytesToCopy = Math.min(buf.length - bufIdx, this[kMissingBytes]);
+      buf.copy(
+        this.lastChar,
+        this[kBufferedBytes],
+        bufIdx,
+        bufIdx + bytesToCopy,
+      );
+
+      bufIdx += bytesToCopy;
+
+      this[kBufferedBytes] += bytesToCopy;
+      this[kMissingBytes] -= bytesToCopy;
+
+      if (this[kMissingBytes] === 0) {
+        // we have all the bytes, complete the char
+        prepend = bufferToString(
+          this.lastChar,
+          this.encoding,
+          0,
+          this[kBufferedBytes],
+        );
+        // reset the char buffer
+        this[kBufferedBytes] = 0;
       }
     }
-  }
-}
 
-/*
- * Attempts to complete a multi-byte UTF-8 character using bytes from a Buffer.
- */
-function utf8FillLastComplete(
-  this: StringDecoderBase,
-  buf: Buffer,
-): string | undefined {
-  const p = this.lastTotal - this.lastNeed;
-  const r = utf8CheckExtraBytes(this, buf);
-  if (r !== undefined) return r;
-  if (this.lastNeed <= buf.length) {
-    buf.copy(this.lastChar, p, 0, this.lastNeed);
-    return this.lastChar.toString(this.encoding, 0, this.lastTotal);
-  }
-  buf.copy(this.lastChar, p, 0, buf.length);
-  this.lastNeed -= buf.length;
-}
+    if (buf.length - bufIdx === 0) {
+      // we advanced the bufIdx, so we may have completed the
+      // incomplete char
+      rest = prepend.length > 0 ? prepend : "";
+      prepend = "";
+    } else {
+      // no characters left to finish
 
-/*
- * Attempts to complete a partial non-UTF-8 character using bytes from a Buffer
- */
-function utf8FillLastIncomplete(
-  this: StringDecoderBase,
-  buf: Buffer,
-): string | undefined {
-  if (this.lastNeed <= buf.length) {
-    buf.copy(this.lastChar, this.lastTotal - this.lastNeed, 0, this.lastNeed);
-    return this.lastChar.toString(this.encoding, 0, this.lastTotal);
-  }
-  buf.copy(this.lastChar, this.lastTotal - this.lastNeed, 0, buf.length);
-  this.lastNeed -= buf.length;
-}
+      // check if the end of the buffer has an incomplete
+      // character, if so we write it into our `lastChar` buffer and
+      // truncate buf
+      if (enc === Encoding.Utf8 && (buf[buf.length - 1] & 0x80)) {
+        for (let i = buf.length - 1;; i--) {
+          this[kBufferedBytes] += 1;
+          if ((buf[i] & 0xC0) === 0x80) {
+            // Doesn't start a character (i.e. it's a trailing byte)
+            if (this[kBufferedBytes] >= 4 || i === 0) {
+              // invalid utf8, we'll just pass it to the underlying decoder
+              this[kBufferedBytes] = 0;
+              break;
+            }
+          } else {
+            // First byte of a UTF-8 char, check
+            // to see how long it should be
+            if ((buf[i] & 0xE0) === 0xC0) {
+              this[kMissingBytes] = 2;
+            } else if ((buf[i] & 0xF0) === 0xE0) {
+              this[kMissingBytes] = 3;
+            } else if ((buf[i] & 0xF8) === 0xF0) {
+              this[kMissingBytes] = 4;
+            } else {
+              // invalid
+              this[kBufferedBytes] = 0;
+              break;
+            }
 
-/*
- * Returns all complete UTF-8 characters in a Buffer. If the Buffer ended on a
- * partial character, the character's bytes are buffered until the required
- * number of bytes are available.
- */
-function utf8Text(this: StringDecoderBase, buf: Buffer, i: number): string {
-  const total = utf8CheckIncomplete(this, buf, i);
-  if (!this.lastNeed) return buf.toString("utf8", i);
-  this.lastTotal = total;
-  const end = buf.length - (total - this.lastNeed);
-  buf.copy(this.lastChar, 0, end);
-  return buf.toString("utf8", i, end);
-}
+            if (this[kBufferedBytes] >= this[kMissingBytes]) {
+              // We have enough trailing bytes to complete
+              // the char
+              this[kMissingBytes] = 0;
+              this[kBufferedBytes] = 0;
+            }
 
-/*
- * For UTF-8, a replacement character is added when ending on a partial
- * character.
- */
-function utf8End(this: Utf8Decoder, buf?: Buffer): string {
-  const r = buf && buf.length ? this.write(buf) : "";
-  if (this.lastNeed) return r + "\ufffd";
-  return r;
-}
+            this[kMissingBytes] -= this[kBufferedBytes];
+            break;
+          }
+        }
+      } else if (enc === Encoding.Utf16) {
+        if ((buf.length - bufIdx) % 2 === 1) {
+          // Have half of a code unit
+          this[kBufferedBytes] = 1;
+          this[kMissingBytes] = 1;
+        } else if ((buf[buf.length - 1] & 0xFC) === 0xD8) {
+          // 2 bytes out of a 4 byte UTF-16 char
+          this[kBufferedBytes] = 2;
+          this[kMissingBytes] = 2;
+        }
+      } else if (enc === Encoding.Base64) {
+        this[kBufferedBytes] = (buf.length - bufIdx) % 3;
+        if (this[kBufferedBytes] > 0) {
+          this[kMissingBytes] = 3 - this[kBufferedBytes];
+        }
+      }
 
-function utf8Write(
-  this: Utf8Decoder | Base64Decoder,
-  buf: Buffer | string,
-): string {
-  if (typeof buf === "string") {
-    return buf;
-  }
-  if (buf.length === 0) return "";
-  let r;
-  let i;
-  // Because `TypedArray` is recognized as `ArrayBuffer` but in the reality, there are some fundamental difference. We would need to cast it properly
-  const normalizedBuffer: Buffer = isBufferType(buf) ? buf : Buffer.from(buf);
-  if (this.lastNeed) {
-    r = this.fillLast(normalizedBuffer);
-    if (r === undefined) return "";
-    i = this.lastNeed;
-    this.lastNeed = 0;
-  } else {
-    i = 0;
-  }
-  if (i < buf.length) {
-    return r
-      ? r + this.text(normalizedBuffer, i)
-      : this.text(normalizedBuffer, i);
-  }
-  return r || "";
-}
+      if (this[kBufferedBytes] > 0) {
+        // Copy the bytes that make up the incomplete char
+        // from the end of the buffer into our `lastChar` buffer
+        buf.copy(
+          this.lastChar,
+          0,
+          buf.length - this[kBufferedBytes],
+        );
+        bufEnd -= this[kBufferedBytes];
+      }
+
+      rest = bufferToString(buf, this.encoding, bufIdx, bufEnd);
+    }
 
-function base64Text(this: StringDecoderBase, buf: Buffer, i: number): string {
-  const n = (buf.length - i) % 3;
-  if (n === 0) return buf.toString("base64", i);
-  this.lastNeed = 3 - n;
-  this.lastTotal = 3;
-  if (n === 1) {
-    this.lastChar[0] = buf[buf.length - 1];
+    if (prepend.length === 0) {
+      return rest;
+    } else {
+      return prepend + rest;
+    }
   } else {
-    this.lastChar[0] = buf[buf.length - 2];
-    this.lastChar[1] = buf[buf.length - 1];
+    return bufferToString(buf, this.encoding, bufIdx, bufEnd);
   }
-  return buf.toString("base64", i, buf.length - n);
 }
 
-function base64End(this: Base64Decoder, buf?: Buffer): string {
-  const r = buf && buf.length ? this.write(buf) : "";
-  if (this.lastNeed) {
-    return r + this.lastChar.toString("base64", 0, 3 - this.lastNeed);
+function flush(this: StringDecoder) {
+  const enc = this.enc;
+
+  if (enc === Encoding.Utf16 && this[kBufferedBytes] % 2 === 1) {
+    // ignore trailing byte if it isn't a complete code unit (2 bytes)
+    this[kBufferedBytes] -= 1;
+    this[kMissingBytes] -= 1;
   }
-  return r;
-}
 
-function simpleWrite(
-  this: StringDecoderBase,
-  buf: Buffer | string,
-): string {
-  if (typeof buf === "string") {
-    return buf;
+  if (this[kBufferedBytes] === 0) {
+    return "";
   }
-  return buf.toString(this.encoding);
-}
 
-function simpleEnd(this: GenericDecoder, buf?: Buffer): string {
-  return buf && buf.length ? this.write(buf) : "";
-}
+  const ret = bufferToString(
+    this.lastChar,
+    this.encoding,
+    0,
+    this[kBufferedBytes],
+  );
 
-class StringDecoderBase {
-  public lastChar: Buffer;
-  public lastNeed = 0;
-  public lastTotal = 0;
-  constructor(public encoding: string, nb: number) {
-    this.lastChar = Buffer.allocUnsafe(nb);
-  }
-}
+  this[kBufferedBytes] = 0;
+  this[kMissingBytes] = 0;
 
-class Base64Decoder extends StringDecoderBase {
-  public end = base64End;
-  public fillLast = utf8FillLastIncomplete;
-  public text = base64Text;
-  public write = utf8Write;
+  return ret;
+}
 
-  constructor(encoding?: string) {
-    super(normalizeEncoding(encoding), 3);
-  }
+enum Encoding {
+  Utf8,
+  Base64,
+  Utf16,
+  Ascii,
+  Latin1,
+  Hex,
 }
 
-class GenericDecoder extends StringDecoderBase {
-  public end = simpleEnd;
-  public fillLast = undefined;
-  public text = utf8Text;
-  public write = simpleWrite;
+const kBufferedBytes = Symbol("bufferedBytes");
+const kMissingBytes = Symbol("missingBytes");
 
-  constructor(encoding?: string) {
-    super(normalizeEncoding(encoding), 4);
-  }
-}
+type StringDecoder = {
+  encoding: string;
+  end: (buf: Buffer) => string;
+  write: (buf: Buffer) => string;
+  lastChar: Buffer;
+  lastNeed: number;
+  lastTotal: number;
+  text: (buf: Buffer, idx: number) => string;
+  enc: Encoding;
 
-class Utf8Decoder extends StringDecoderBase {
-  public end = utf8End;
-  public fillLast = utf8FillLastComplete;
-  public text = utf8Text;
-  public write = utf8Write;
+  decode: (buf: Buffer) => string;
 
-  constructor(encoding?: string) {
-    super(normalizeEncoding(encoding), 4);
-  }
-}
+  [kBufferedBytes]: number;
+  [kMissingBytes]: number;
+
+  flush: () => string;
+};
 
 /*
  * StringDecoder provides an interface for efficiently splitting a series of
  * buffers into a series of JS strings without breaking apart multi-byte
  * characters.
  */
-export class StringDecoder {
-  public encoding: string;
-  public end: (buf?: Buffer) => string;
-  public fillLast: ((buf: Buffer) => string | undefined) | undefined;
-  public lastChar: Buffer;
-  public lastNeed: number;
-  public lastTotal: number;
-  public text: (buf: Buffer, n: number) => string;
-  public write: (buf: Buffer) => string;
-
-  constructor(encoding?: string) {
-    const normalizedEncoding = normalizeEncoding(encoding);
-    let decoder: Utf8Decoder | Base64Decoder | GenericDecoder;
-    switch (normalizedEncoding) {
-      case "utf8":
-        decoder = new Utf8Decoder(encoding);
-        break;
-      case "base64":
-        decoder = new Base64Decoder(encoding);
-        break;
-      default:
-        decoder = new GenericDecoder(encoding);
-    }
-    this.encoding = decoder.encoding;
-    this.end = decoder.end;
-    this.fillLast = decoder.fillLast;
-    this.lastChar = decoder.lastChar;
-    this.lastNeed = decoder.lastNeed;
-    this.lastTotal = decoder.lastTotal;
-    this.text = decoder.text;
-    this.write = decoder.write;
+export function StringDecoder(this: Partial<StringDecoder>, encoding?: string) {
+  const normalizedEncoding = normalizeEncoding(encoding);
+  let enc: Encoding = Encoding.Utf8;
+  let bufLen = 0;
+  switch (normalizedEncoding) {
+    case "utf8":
+      enc = Encoding.Utf8;
+      bufLen = 4;
+      break;
+    case "base64":
+      enc = Encoding.Base64;
+      bufLen = 3;
+      break;
+    case "utf16le":
+      enc = Encoding.Utf16;
+      bufLen = 4;
+      break;
+    case "hex":
+      enc = Encoding.Hex;
+      bufLen = 0;
+      break;
+    case "latin1":
+      enc = Encoding.Latin1;
+      bufLen = 0;
+      break;
+    case "ascii":
+      enc = Encoding.Ascii;
+      bufLen = 0;
+      break;
   }
+  this.encoding = normalizedEncoding;
+  this.lastChar = Buffer.allocUnsafe(bufLen);
+  this.enc = enc;
+  this[kBufferedBytes] = 0;
+  this[kMissingBytes] = 0;
+  this.flush = flush;
+  this.decode = decode;
 }
-// Allow calling StringDecoder() without new
-const PStringDecoder = new Proxy(StringDecoder, {
-  apply(_target, thisArg, args) {
-    // @ts-ignore tedious to replicate types ...
-    return Object.assign(thisArg, new StringDecoder(...args));
+
+/**
+ * Returns a decoded string, omitting any incomplete multi-bytes
+ * characters at the end of the Buffer, or TypedArray, or DataView
+ */
+StringDecoder.prototype.write = function write(buf: Buffer): string {
+  if (typeof buf === "string") {
+    return buf;
+  }
+  const normalizedBuf = normalizeBuffer(buf);
+  if (this[kBufferedBytes] === undefined) {
+    throw new ERR_INVALID_THIS("StringDecoder");
+  }
+  return this.decode(normalizedBuf);
+};
+
+/**
+ * Returns any remaining input stored in the internal buffer as a string.
+ * After end() is called, the stringDecoder object can be reused for new
+ * input.
+ */
+StringDecoder.prototype.end = function end(buf: Buffer): string {
+  let ret = "";
+  if (buf !== undefined) {
+    ret = this.write(buf);
+  }
+  if (this[kBufferedBytes] > 0) {
+    ret += this.flush();
+  }
+  return ret;
+};
+
+// Below is undocumented but accessible stuff from node's old impl
+// (node's tests assert on these, so we need to support them)
+StringDecoder.prototype.text = function text(
+  buf: Buffer,
+  offset: number,
+): string {
+  this[kBufferedBytes] = 0;
+  this[kMissingBytes] = 0;
+  return this.write(buf.subarray(offset));
+};
+
+ObjectDefineProperties(StringDecoder.prototype, {
+  lastNeed: {
+    configurable: true,
+    enumerable: true,
+    get(this: StringDecoder): number {
+      return this[kMissingBytes];
+    },
+  },
+  lastTotal: {
+    configurable: true,
+    enumerable: true,
+    get(this: StringDecoder): number {
+      return this[kBufferedBytes] + this[kMissingBytes];
+    },
   },
 });
 
-export default { StringDecoder: PStringDecoder };
+export default { StringDecoder };