| Leo Repp | 58b9f11 | 2021-11-22 11:57:47 +0100 | [diff] [blame^] | 1 | "use strict"; |
| 2 | var Buffer = require("safer-buffer").Buffer; |
| 3 | |
| 4 | // Export Node.js internal encodings. |
| 5 | |
| 6 | module.exports = { |
| 7 | // Encodings |
| 8 | utf8: { type: "_internal", bomAware: true}, |
| 9 | cesu8: { type: "_internal", bomAware: true}, |
| 10 | unicode11utf8: "utf8", |
| 11 | |
| 12 | ucs2: { type: "_internal", bomAware: true}, |
| 13 | utf16le: "ucs2", |
| 14 | |
| 15 | binary: { type: "_internal" }, |
| 16 | base64: { type: "_internal" }, |
| 17 | hex: { type: "_internal" }, |
| 18 | |
| 19 | // Codec. |
| 20 | _internal: InternalCodec, |
| 21 | }; |
| 22 | |
| 23 | //------------------------------------------------------------------------------ |
| 24 | |
| 25 | function InternalCodec(codecOptions, iconv) { |
| 26 | this.enc = codecOptions.encodingName; |
| 27 | this.bomAware = codecOptions.bomAware; |
| 28 | |
| 29 | if (this.enc === "base64") |
| 30 | this.encoder = InternalEncoderBase64; |
| 31 | else if (this.enc === "cesu8") { |
| 32 | this.enc = "utf8"; // Use utf8 for decoding. |
| 33 | this.encoder = InternalEncoderCesu8; |
| 34 | |
| 35 | // Add decoder for versions of Node not supporting CESU-8 |
| 36 | if (Buffer.from('eda0bdedb2a9', 'hex').toString() !== '💩') { |
| 37 | this.decoder = InternalDecoderCesu8; |
| 38 | this.defaultCharUnicode = iconv.defaultCharUnicode; |
| 39 | } |
| 40 | } |
| 41 | } |
| 42 | |
| 43 | InternalCodec.prototype.encoder = InternalEncoder; |
| 44 | InternalCodec.prototype.decoder = InternalDecoder; |
| 45 | |
| 46 | //------------------------------------------------------------------------------ |
| 47 | |
| 48 | // We use node.js internal decoder. Its signature is the same as ours. |
| 49 | var StringDecoder = require('string_decoder').StringDecoder; |
| 50 | |
| 51 | if (!StringDecoder.prototype.end) // Node v0.8 doesn't have this method. |
| 52 | StringDecoder.prototype.end = function() {}; |
| 53 | |
| 54 | |
| 55 | function InternalDecoder(options, codec) { |
| 56 | StringDecoder.call(this, codec.enc); |
| 57 | } |
| 58 | |
| 59 | InternalDecoder.prototype = StringDecoder.prototype; |
| 60 | |
| 61 | |
| 62 | //------------------------------------------------------------------------------ |
| 63 | // Encoder is mostly trivial |
| 64 | |
| 65 | function InternalEncoder(options, codec) { |
| 66 | this.enc = codec.enc; |
| 67 | } |
| 68 | |
| 69 | InternalEncoder.prototype.write = function(str) { |
| 70 | return Buffer.from(str, this.enc); |
| 71 | } |
| 72 | |
| 73 | InternalEncoder.prototype.end = function() { |
| 74 | } |
| 75 | |
| 76 | |
| 77 | //------------------------------------------------------------------------------ |
| 78 | // Except base64 encoder, which must keep its state. |
| 79 | |
| 80 | function InternalEncoderBase64(options, codec) { |
| 81 | this.prevStr = ''; |
| 82 | } |
| 83 | |
| 84 | InternalEncoderBase64.prototype.write = function(str) { |
| 85 | str = this.prevStr + str; |
| 86 | var completeQuads = str.length - (str.length % 4); |
| 87 | this.prevStr = str.slice(completeQuads); |
| 88 | str = str.slice(0, completeQuads); |
| 89 | |
| 90 | return Buffer.from(str, "base64"); |
| 91 | } |
| 92 | |
| 93 | InternalEncoderBase64.prototype.end = function() { |
| 94 | return Buffer.from(this.prevStr, "base64"); |
| 95 | } |
| 96 | |
| 97 | |
| 98 | //------------------------------------------------------------------------------ |
| 99 | // CESU-8 encoder is also special. |
| 100 | |
| 101 | function InternalEncoderCesu8(options, codec) { |
| 102 | } |
| 103 | |
| 104 | InternalEncoderCesu8.prototype.write = function(str) { |
| 105 | var buf = Buffer.alloc(str.length * 3), bufIdx = 0; |
| 106 | for (var i = 0; i < str.length; i++) { |
| 107 | var charCode = str.charCodeAt(i); |
| 108 | // Naive implementation, but it works because CESU-8 is especially easy |
| 109 | // to convert from UTF-16 (which all JS strings are encoded in). |
| 110 | if (charCode < 0x80) |
| 111 | buf[bufIdx++] = charCode; |
| 112 | else if (charCode < 0x800) { |
| 113 | buf[bufIdx++] = 0xC0 + (charCode >>> 6); |
| 114 | buf[bufIdx++] = 0x80 + (charCode & 0x3f); |
| 115 | } |
| 116 | else { // charCode will always be < 0x10000 in javascript. |
| 117 | buf[bufIdx++] = 0xE0 + (charCode >>> 12); |
| 118 | buf[bufIdx++] = 0x80 + ((charCode >>> 6) & 0x3f); |
| 119 | buf[bufIdx++] = 0x80 + (charCode & 0x3f); |
| 120 | } |
| 121 | } |
| 122 | return buf.slice(0, bufIdx); |
| 123 | } |
| 124 | |
| 125 | InternalEncoderCesu8.prototype.end = function() { |
| 126 | } |
| 127 | |
| 128 | //------------------------------------------------------------------------------ |
| 129 | // CESU-8 decoder is not implemented in Node v4.0+ |
| 130 | |
| 131 | function InternalDecoderCesu8(options, codec) { |
| 132 | this.acc = 0; |
| 133 | this.contBytes = 0; |
| 134 | this.accBytes = 0; |
| 135 | this.defaultCharUnicode = codec.defaultCharUnicode; |
| 136 | } |
| 137 | |
| 138 | InternalDecoderCesu8.prototype.write = function(buf) { |
| 139 | var acc = this.acc, contBytes = this.contBytes, accBytes = this.accBytes, |
| 140 | res = ''; |
| 141 | for (var i = 0; i < buf.length; i++) { |
| 142 | var curByte = buf[i]; |
| 143 | if ((curByte & 0xC0) !== 0x80) { // Leading byte |
| 144 | if (contBytes > 0) { // Previous code is invalid |
| 145 | res += this.defaultCharUnicode; |
| 146 | contBytes = 0; |
| 147 | } |
| 148 | |
| 149 | if (curByte < 0x80) { // Single-byte code |
| 150 | res += String.fromCharCode(curByte); |
| 151 | } else if (curByte < 0xE0) { // Two-byte code |
| 152 | acc = curByte & 0x1F; |
| 153 | contBytes = 1; accBytes = 1; |
| 154 | } else if (curByte < 0xF0) { // Three-byte code |
| 155 | acc = curByte & 0x0F; |
| 156 | contBytes = 2; accBytes = 1; |
| 157 | } else { // Four or more are not supported for CESU-8. |
| 158 | res += this.defaultCharUnicode; |
| 159 | } |
| 160 | } else { // Continuation byte |
| 161 | if (contBytes > 0) { // We're waiting for it. |
| 162 | acc = (acc << 6) | (curByte & 0x3f); |
| 163 | contBytes--; accBytes++; |
| 164 | if (contBytes === 0) { |
| 165 | // Check for overlong encoding, but support Modified UTF-8 (encoding NULL as C0 80) |
| 166 | if (accBytes === 2 && acc < 0x80 && acc > 0) |
| 167 | res += this.defaultCharUnicode; |
| 168 | else if (accBytes === 3 && acc < 0x800) |
| 169 | res += this.defaultCharUnicode; |
| 170 | else |
| 171 | // Actually add character. |
| 172 | res += String.fromCharCode(acc); |
| 173 | } |
| 174 | } else { // Unexpected continuation byte |
| 175 | res += this.defaultCharUnicode; |
| 176 | } |
| 177 | } |
| 178 | } |
| 179 | this.acc = acc; this.contBytes = contBytes; this.accBytes = accBytes; |
| 180 | return res; |
| 181 | } |
| 182 | |
| 183 | InternalDecoderCesu8.prototype.end = function() { |
| 184 | var res = 0; |
| 185 | if (this.contBytes > 0) |
| 186 | res += this.defaultCharUnicode; |
| 187 | return res; |
| 188 | } |