From ebbc897b69f906d88a99768a2fff7661e2894670 Mon Sep 17 00:00:00 2001 From: Nathan Whitaker <17734409+nathanwhit@users.noreply.github.com> Date: Fri, 15 Mar 2024 17:24:13 -0700 Subject: [PATCH] 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. --- ext/node/polyfills/string_decoder.ts | 605 ++++++++++-------- tests/node_compat/config.jsonc | 1 + tests/node_compat/test.ts | 1 + .../test/parallel/test-string-decoder.js | 292 +++++++++ tools/node_compat/TODO.md | 1 - 5 files changed, 631 insertions(+), 269 deletions(-) create mode 100644 tests/node_compat/test/parallel/test-string-decoder.js diff --git a/ext/node/polyfills/string_decoder.ts b/ext/node/polyfills/string_decoder.ts index 507a994bbc..ef83b6fc92 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"; -import { - normalizeEncoding as castEncoding, - notImplemented, -} from "ext:deno_node/_utils.ts"; +// Logic and comments translated pretty much one-to-one from node's impl +// (https://github.com/nodejs/node/blob/ba06c5c509956dc413f91b755c1c93798bb700d4/src/string_decoder.cc) -enum NotImplemented { - "ascii", - "latin1", - "utf16le", -} +import { Buffer, constants } from "node:buffer"; +import { normalizeEncoding as castEncoding } from "ext:deno_node/_utils.ts"; +import { + 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; +function normalizeBuffer(buf: Buffer) { + if (!ArrayBufferIsView(buf)) { + throw new ERR_INVALID_ARG_TYPE( + "buf", + ["Buffer", "TypedArray", "DataView"], + buf, + ); + } + if (isBufferType(buf)) { + return buf; + } else { + return Buffer.from( + buf.buffer, + ); + } } -/* - * 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, +function bufferToString( 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; + 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 (--j < i || nb === -2) return 0; - nb = utf8CheckByte(buf[j]); - if (nb >= 0) { - if (nb > 0) self.lastNeed = nb - 2; - return nb; - } - 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; - } - return 0; + return buf.toString(encoding as Any, start, end); } -/* - * 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, - buf: Buffer, -): string | undefined { - if ((buf[0] & 0xc0) !== 0x80) { - self.lastNeed = 0; - return "\ufffd"; - } - 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"; +// 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; + 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); + } + + if (prepend.length === 0) { + return rest; + } else { + return prepend + rest; + } } else { - i = 0; - } - if (i < buf.length) { - return r - ? r + this.text(normalizedBuffer, i) - : this.text(normalizedBuffer, i); - } - return r || ""; -} - -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]; - } else { - this.lastChar[0] = buf[buf.length - 2]; - this.lastChar[1] = buf[buf.length - 1]; - } - 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); - } - return r; -} - -function simpleWrite( - this: StringDecoderBase, - buf: Buffer | string, -): string { - if (typeof buf === "string") { - return buf; - } - return buf.toString(this.encoding); -} - -function simpleEnd(this: GenericDecoder, buf?: Buffer): string { - return buf && buf.length ? this.write(buf) : ""; -} - -class StringDecoderBase { - public lastChar: Buffer; - public lastNeed = 0; - public lastTotal = 0; - constructor(public encoding: string, nb: number) { - this.lastChar = Buffer.allocUnsafe(nb); + return bufferToString(buf, this.encoding, bufIdx, bufEnd); } } -class Base64Decoder extends StringDecoderBase { - public end = base64End; - public fillLast = utf8FillLastIncomplete; - public text = base64Text; - public write = utf8Write; +function flush(this: StringDecoder) { + const enc = this.enc; - constructor(encoding?: string) { - super(normalizeEncoding(encoding), 3); + 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; } + + if (this[kBufferedBytes] === 0) { + return ""; + } + + const ret = bufferToString( + this.lastChar, + this.encoding, + 0, + this[kBufferedBytes], + ); + + this[kBufferedBytes] = 0; + this[kMissingBytes] = 0; + + return ret; } -class GenericDecoder extends StringDecoderBase { - public end = simpleEnd; - public fillLast = undefined; - public text = utf8Text; - public write = simpleWrite; - - constructor(encoding?: string) { - super(normalizeEncoding(encoding), 4); - } +enum Encoding { + Utf8, + Base64, + Utf16, + Ascii, + Latin1, + Hex, } -class Utf8Decoder extends StringDecoderBase { - public end = utf8End; - public fillLast = utf8FillLastComplete; - public text = utf8Text; - public write = utf8Write; +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; + + decode: (buf: Buffer) => string; + + [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, 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 }; diff --git a/tests/node_compat/config.jsonc b/tests/node_compat/config.jsonc index 4d89b1a89a..bf254faf6d 100644 --- a/tests/node_compat/config.jsonc +++ b/tests/node_compat/config.jsonc @@ -614,6 +614,7 @@ "test-stream3-cork-uncork.js", "test-stream3-pause-then-read.js", "test-streams-highwatermark.js", + "test-string-decoder.js", "test-timers-api-refs.js", "test-timers-args.js", "test-timers-clear-null-does-not-throw-error.js", diff --git a/tests/node_compat/test.ts b/tests/node_compat/test.ts index bafb14db2c..04a85f1135 100644 --- a/tests/node_compat/test.ts +++ b/tests/node_compat/test.ts @@ -82,6 +82,7 @@ async function runTest(t: Deno.TestContext, path: string): Promise { "-A", "--quiet", //"--unsafely-ignore-certificate-errors", + "--unstable-unsafe-proto", "--unstable-bare-node-builtins", "--v8-flags=" + v8Flags.join(), "runner.ts", diff --git a/tests/node_compat/test/parallel/test-string-decoder.js b/tests/node_compat/test/parallel/test-string-decoder.js new file mode 100644 index 0000000000..84ac71aac8 --- /dev/null +++ b/tests/node_compat/test/parallel/test-string-decoder.js @@ -0,0 +1,292 @@ +// deno-fmt-ignore-file +// deno-lint-ignore-file + +// Copyright Joyent and Node contributors. All rights reserved. MIT license. +// Taken from Node 18.12.1 +// This file is automatically generated by `tools/node_compat/setup.ts`. Do not modify this file manually. + +// Copyright Joyent, Inc. and other Node contributors. +// +// Permission is hereby granted, free of charge, to any person obtaining a +// copy of this software and associated documentation files (the +// "Software"), to deal in the Software without restriction, including +// without limitation the rights to use, copy, modify, merge, publish, +// distribute, sublicense, and/or sell copies of the Software, and to permit +// persons to whom the Software is furnished to do so, subject to the +// following conditions: +// +// The above copyright notice and this permission notice shall be included +// in all copies or substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS +// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN +// NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR +// OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE +// USE OR OTHER DEALINGS IN THE SOFTWARE. + +'use strict'; +const common = require('../common'); +const assert = require('assert'); +const inspect = require('util').inspect; +const StringDecoder = require('string_decoder').StringDecoder; + +// Test default encoding +let decoder = new StringDecoder(); +assert.strictEqual(decoder.encoding, 'utf8'); + +// Should work without 'new' keyword +const decoder2 = {}; +StringDecoder.call(decoder2); +assert.strictEqual(decoder2.encoding, 'utf8'); + +// UTF-8 +test('utf-8', Buffer.from('$', 'utf-8'), '$'); +test('utf-8', Buffer.from('¢', 'utf-8'), '¢'); +test('utf-8', Buffer.from('€', 'utf-8'), '€'); +test('utf-8', Buffer.from('𤭢', 'utf-8'), '𤭢'); +// A mixed ascii and non-ascii string +// Test stolen from deps/v8/test/cctest/test-strings.cc +// U+02E4 -> CB A4 +// U+0064 -> 64 +// U+12E4 -> E1 8B A4 +// U+0030 -> 30 +// U+3045 -> E3 81 85 +test( + 'utf-8', + Buffer.from([0xCB, 0xA4, 0x64, 0xE1, 0x8B, 0xA4, 0x30, 0xE3, 0x81, 0x85]), + '\u02e4\u0064\u12e4\u0030\u3045' +); + +// Some invalid input, known to have caused trouble with chunking +// in https://github.com/nodejs/node/pull/7310#issuecomment-226445923 +// 00: |00000000 ASCII +// 41: |01000001 ASCII +// B8: 10|111000 continuation +// CC: 110|01100 two-byte head +// E2: 1110|0010 three-byte head +// F0: 11110|000 four-byte head +// F1: 11110|001'another four-byte head +// FB: 111110|11 "five-byte head", not UTF-8 +test('utf-8', Buffer.from('C9B5A941', 'hex'), '\u0275\ufffdA'); +test('utf-8', Buffer.from('E2', 'hex'), '\ufffd'); +test('utf-8', Buffer.from('E241', 'hex'), '\ufffdA'); +test('utf-8', Buffer.from('CCCCB8', 'hex'), '\ufffd\u0338'); +test('utf-8', Buffer.from('F0B841', 'hex'), '\ufffdA'); +test('utf-8', Buffer.from('F1CCB8', 'hex'), '\ufffd\u0338'); +test('utf-8', Buffer.from('F0FB00', 'hex'), '\ufffd\ufffd\0'); +test('utf-8', Buffer.from('CCE2B8B8', 'hex'), '\ufffd\u2e38'); +test('utf-8', Buffer.from('E2B8CCB8', 'hex'), '\ufffd\u0338'); +test('utf-8', Buffer.from('E2FBCC01', 'hex'), '\ufffd\ufffd\ufffd\u0001'); +test('utf-8', Buffer.from('CCB8CDB9', 'hex'), '\u0338\u0379'); +// CESU-8 of U+1D40D + +// V8 has changed their invalid UTF-8 handling, see +// https://chromium-review.googlesource.com/c/v8/v8/+/671020 for more info. +test('utf-8', Buffer.from('EDA0B5EDB08D', 'hex'), + '\ufffd\ufffd\ufffd\ufffd\ufffd\ufffd'); + +// UCS-2 +test('ucs2', Buffer.from('ababc', 'ucs2'), 'ababc'); + +// UTF-16LE +test('utf16le', Buffer.from('3DD84DDC', 'hex'), '\ud83d\udc4d'); // thumbs up + +// Additional UTF-8 tests +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('E1', 'hex')), ''); + +// A quick test for lastChar, lastNeed & lastTotal which are undocumented. +assert(decoder.lastChar.equals(new Uint8Array([0xe1, 0, 0, 0]))); +assert.strictEqual(decoder.lastNeed, 2); +assert.strictEqual(decoder.lastTotal, 3); + +assert.strictEqual(decoder.end(), '\ufffd'); + +// ArrayBufferView tests +const arrayBufferViewStr = 'String for ArrayBufferView tests\n'; +const inputBuffer = Buffer.from(arrayBufferViewStr.repeat(8), 'utf8'); +for (const expectView of common.getArrayBufferViews(inputBuffer)) { + assert.strictEqual( + decoder.write(expectView), + inputBuffer.toString('utf8') + ); + assert.strictEqual(decoder.end(), ''); +} + +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('E18B', 'hex')), ''); +assert.strictEqual(decoder.end(), '\ufffd'); + +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('\ufffd')), '\ufffd'); +assert.strictEqual(decoder.end(), ''); + +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('\ufffd\ufffd\ufffd')), + '\ufffd\ufffd\ufffd'); +assert.strictEqual(decoder.end(), ''); + +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('EFBFBDE2', 'hex')), '\ufffd'); +assert.strictEqual(decoder.end(), '\ufffd'); + +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('F1', 'hex')), ''); +assert.strictEqual(decoder.write(Buffer.from('41F2', 'hex')), '\ufffdA'); +assert.strictEqual(decoder.end(), '\ufffd'); + +// Additional utf8Text test +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.text(Buffer.from([0x41]), 2), ''); + +// Additional UTF-16LE surrogate pair tests +decoder = new StringDecoder('utf16le'); +assert.strictEqual(decoder.write(Buffer.from('3DD8', 'hex')), ''); +assert.strictEqual(decoder.write(Buffer.from('4D', 'hex')), ''); +assert.strictEqual(decoder.write(Buffer.from('DC', 'hex')), '\ud83d\udc4d'); +assert.strictEqual(decoder.end(), ''); + +decoder = new StringDecoder('utf16le'); +assert.strictEqual(decoder.write(Buffer.from('3DD8', 'hex')), ''); +assert.strictEqual(decoder.end(), '\ud83d'); + +decoder = new StringDecoder('utf16le'); +assert.strictEqual(decoder.write(Buffer.from('3DD8', 'hex')), ''); +assert.strictEqual(decoder.write(Buffer.from('4D', 'hex')), ''); +assert.strictEqual(decoder.end(), '\ud83d'); + +decoder = new StringDecoder('utf16le'); +assert.strictEqual(decoder.write(Buffer.from('3DD84D', 'hex')), '\ud83d'); +assert.strictEqual(decoder.end(), ''); + +// Regression test for https://github.com/nodejs/node/issues/22358 +// (unaligned UTF-16 access). +decoder = new StringDecoder('utf16le'); +assert.strictEqual(decoder.write(Buffer.alloc(1)), ''); +assert.strictEqual(decoder.write(Buffer.alloc(20)), '\0'.repeat(10)); +assert.strictEqual(decoder.write(Buffer.alloc(48)), '\0'.repeat(24)); +assert.strictEqual(decoder.end(), ''); + +// Regression tests for https://github.com/nodejs/node/issues/22626 +// (not enough replacement chars when having seen more than one byte of an +// incomplete multibyte characters). +decoder = new StringDecoder('utf8'); +assert.strictEqual(decoder.write(Buffer.from('f69b', 'hex')), ''); +assert.strictEqual(decoder.write(Buffer.from('d1', 'hex')), '\ufffd\ufffd'); +assert.strictEqual(decoder.end(), '\ufffd'); +assert.strictEqual(decoder.write(Buffer.from('f4', 'hex')), ''); +assert.strictEqual(decoder.write(Buffer.from('bde5', 'hex')), '\ufffd\ufffd'); +assert.strictEqual(decoder.end(), '\ufffd'); + +assert.throws( + () => new StringDecoder(1), + { + code: 'ERR_UNKNOWN_ENCODING', + name: 'TypeError', + message: 'Unknown encoding: 1' + } +); + +assert.throws( + () => new StringDecoder('test'), + { + code: 'ERR_UNKNOWN_ENCODING', + name: 'TypeError', + message: 'Unknown encoding: test' + } +); + +assert.throws( + () => new StringDecoder('utf8').write(null), + { + code: 'ERR_INVALID_ARG_TYPE', + name: 'TypeError', + message: 'The "buf" argument must be an instance of Buffer, TypedArray,' + + ' or DataView. Received null' + } +); + +if (common.enoughTestMem) { + assert.throws( + () => new StringDecoder().write(Buffer.alloc((process.arch === 'ia32' ? 0x18ffffe8 : 0x1fffffe8) + 1).fill('a')), + { + code: 'ERR_STRING_TOO_LONG', + } + ); +} + +assert.throws( + () => new StringDecoder('utf8').__proto__.write(Buffer.from('abc')), // eslint-disable-line no-proto + { + code: 'ERR_INVALID_THIS', + } +); + +// Test verifies that StringDecoder will correctly decode the given input +// buffer with the given encoding to the expected output. It will attempt all +// possible ways to write() the input buffer, see writeSequences(). The +// singleSequence allows for easy debugging of a specific sequence which is +// useful in case of test failures. +function test(encoding, input, expected, singleSequence) { + let sequences; + if (!singleSequence) { + sequences = writeSequences(input.length); + } else { + sequences = [singleSequence]; + } + const hexNumberRE = /.{2}/g; + sequences.forEach((sequence) => { + const decoder = new StringDecoder(encoding); + let output = ''; + sequence.forEach((write) => { + output += decoder.write(input.slice(write[0], write[1])); + }); + output += decoder.end(); + if (output !== expected) { + const message = + `Expected "${unicodeEscape(expected)}", ` + + `but got "${unicodeEscape(output)}"\n` + + `input: ${input.toString('hex').match(hexNumberRE)}\n` + + `Write sequence: ${JSON.stringify(sequence)}\n` + + `Full Decoder State: ${inspect(decoder)}`; + assert.fail(message); + } + }); +} + +// unicodeEscape prints the str contents as unicode escape codes. +function unicodeEscape(str) { + let r = ''; + for (let i = 0; i < str.length; i++) { + r += `\\u${str.charCodeAt(i).toString(16)}`; + } + return r; +} + +// writeSequences returns an array of arrays that describes all possible ways a +// buffer of the given length could be split up and passed to sequential write +// calls. +// +// e.G. writeSequences(3) will return: [ +// [ [ 0, 3 ] ], +// [ [ 0, 2 ], [ 2, 3 ] ], +// [ [ 0, 1 ], [ 1, 3 ] ], +// [ [ 0, 1 ], [ 1, 2 ], [ 2, 3 ] ] +// ] +function writeSequences(length, start, sequence) { + if (start === undefined) { + start = 0; + sequence = []; + } else if (start === length) { + return [sequence]; + } + let sequences = []; + for (let end = length; end > start; end--) { + const subSequence = sequence.concat([[start, end]]); + const subSequences = writeSequences(length, end, subSequence, sequences); + sequences = sequences.concat(subSequences); + } + return sequences; +} diff --git a/tools/node_compat/TODO.md b/tools/node_compat/TODO.md index 88cadfc1c3..3d8306988e 100644 --- a/tools/node_compat/TODO.md +++ b/tools/node_compat/TODO.md @@ -2184,7 +2184,6 @@ NOTE: This file should not be manually edited. Please edit `tests/node_compat/co - [parallel/test-stream3-pipeline-async-iterator.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-stream3-pipeline-async-iterator.js) - [parallel/test-string-decoder-end.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-string-decoder-end.js) - [parallel/test-string-decoder-fuzz.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-string-decoder-fuzz.js) -- [parallel/test-string-decoder.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-string-decoder.js) - [parallel/test-stringbytes-external.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-stringbytes-external.js) - [parallel/test-structuredClone-global.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-structuredClone-global.js) - [parallel/test-sync-fileread.js](https://github.com/nodejs/node/tree/v18.12.1/test/parallel/test-sync-fileread.js)