| Leo Repp | 58b9f11 | 2021-11-22 11:57:47 +0100 | [diff] [blame^] | 1 | "use strict"; |
| 2 | var Buffer = require("safer-buffer").Buffer; |
| 3 | |
| 4 | // UTF-7 codec, according to https://tools.ietf.org/html/rfc2152 |
| 5 | // See also below a UTF-7-IMAP codec, according to http://tools.ietf.org/html/rfc3501#section-5.1.3 |
| 6 | |
| 7 | exports.utf7 = Utf7Codec; |
| 8 | exports.unicode11utf7 = 'utf7'; // Alias UNICODE-1-1-UTF-7 |
| 9 | function Utf7Codec(codecOptions, iconv) { |
| 10 | this.iconv = iconv; |
| 11 | }; |
| 12 | |
| 13 | Utf7Codec.prototype.encoder = Utf7Encoder; |
| 14 | Utf7Codec.prototype.decoder = Utf7Decoder; |
| 15 | Utf7Codec.prototype.bomAware = true; |
| 16 | |
| 17 | |
| 18 | // -- Encoding |
| 19 | |
| 20 | var nonDirectChars = /[^A-Za-z0-9'\(\),-\.\/:\? \n\r\t]+/g; |
| 21 | |
| 22 | function Utf7Encoder(options, codec) { |
| 23 | this.iconv = codec.iconv; |
| 24 | } |
| 25 | |
| 26 | Utf7Encoder.prototype.write = function(str) { |
| 27 | // Naive implementation. |
| 28 | // Non-direct chars are encoded as "+<base64>-"; single "+" char is encoded as "+-". |
| 29 | return Buffer.from(str.replace(nonDirectChars, function(chunk) { |
| 30 | return "+" + (chunk === '+' ? '' : |
| 31 | this.iconv.encode(chunk, 'utf16-be').toString('base64').replace(/=+$/, '')) |
| 32 | + "-"; |
| 33 | }.bind(this))); |
| 34 | } |
| 35 | |
| 36 | Utf7Encoder.prototype.end = function() { |
| 37 | } |
| 38 | |
| 39 | |
| 40 | // -- Decoding |
| 41 | |
| 42 | function Utf7Decoder(options, codec) { |
| 43 | this.iconv = codec.iconv; |
| 44 | this.inBase64 = false; |
| 45 | this.base64Accum = ''; |
| 46 | } |
| 47 | |
| 48 | var base64Regex = /[A-Za-z0-9\/+]/; |
| 49 | var base64Chars = []; |
| 50 | for (var i = 0; i < 256; i++) |
| 51 | base64Chars[i] = base64Regex.test(String.fromCharCode(i)); |
| 52 | |
| 53 | var plusChar = '+'.charCodeAt(0), |
| 54 | minusChar = '-'.charCodeAt(0), |
| 55 | andChar = '&'.charCodeAt(0); |
| 56 | |
| 57 | Utf7Decoder.prototype.write = function(buf) { |
| 58 | var res = "", lastI = 0, |
| 59 | inBase64 = this.inBase64, |
| 60 | base64Accum = this.base64Accum; |
| 61 | |
| 62 | // The decoder is more involved as we must handle chunks in stream. |
| 63 | |
| 64 | for (var i = 0; i < buf.length; i++) { |
| 65 | if (!inBase64) { // We're in direct mode. |
| 66 | // Write direct chars until '+' |
| 67 | if (buf[i] == plusChar) { |
| 68 | res += this.iconv.decode(buf.slice(lastI, i), "ascii"); // Write direct chars. |
| 69 | lastI = i+1; |
| 70 | inBase64 = true; |
| 71 | } |
| 72 | } else { // We decode base64. |
| 73 | if (!base64Chars[buf[i]]) { // Base64 ended. |
| 74 | if (i == lastI && buf[i] == minusChar) {// "+-" -> "+" |
| 75 | res += "+"; |
| 76 | } else { |
| 77 | var b64str = base64Accum + buf.slice(lastI, i).toString(); |
| 78 | res += this.iconv.decode(Buffer.from(b64str, 'base64'), "utf16-be"); |
| 79 | } |
| 80 | |
| 81 | if (buf[i] != minusChar) // Minus is absorbed after base64. |
| 82 | i--; |
| 83 | |
| 84 | lastI = i+1; |
| 85 | inBase64 = false; |
| 86 | base64Accum = ''; |
| 87 | } |
| 88 | } |
| 89 | } |
| 90 | |
| 91 | if (!inBase64) { |
| 92 | res += this.iconv.decode(buf.slice(lastI), "ascii"); // Write direct chars. |
| 93 | } else { |
| 94 | var b64str = base64Accum + buf.slice(lastI).toString(); |
| 95 | |
| 96 | var canBeDecoded = b64str.length - (b64str.length % 8); // Minimal chunk: 2 quads -> 2x3 bytes -> 3 chars. |
| 97 | base64Accum = b64str.slice(canBeDecoded); // The rest will be decoded in future. |
| 98 | b64str = b64str.slice(0, canBeDecoded); |
| 99 | |
| 100 | res += this.iconv.decode(Buffer.from(b64str, 'base64'), "utf16-be"); |
| 101 | } |
| 102 | |
| 103 | this.inBase64 = inBase64; |
| 104 | this.base64Accum = base64Accum; |
| 105 | |
| 106 | return res; |
| 107 | } |
| 108 | |
| 109 | Utf7Decoder.prototype.end = function() { |
| 110 | var res = ""; |
| 111 | if (this.inBase64 && this.base64Accum.length > 0) |
| 112 | res = this.iconv.decode(Buffer.from(this.base64Accum, 'base64'), "utf16-be"); |
| 113 | |
| 114 | this.inBase64 = false; |
| 115 | this.base64Accum = ''; |
| 116 | return res; |
| 117 | } |
| 118 | |
| 119 | |
| 120 | // UTF-7-IMAP codec. |
| 121 | // RFC3501 Sec. 5.1.3 Modified UTF-7 (http://tools.ietf.org/html/rfc3501#section-5.1.3) |
| 122 | // Differences: |
| 123 | // * Base64 part is started by "&" instead of "+" |
| 124 | // * Direct characters are 0x20-0x7E, except "&" (0x26) |
| 125 | // * In Base64, "," is used instead of "/" |
| 126 | // * Base64 must not be used to represent direct characters. |
| 127 | // * No implicit shift back from Base64 (should always end with '-') |
| 128 | // * String must end in non-shifted position. |
| 129 | // * "-&" while in base64 is not allowed. |
| 130 | |
| 131 | |
| 132 | exports.utf7imap = Utf7IMAPCodec; |
| 133 | function Utf7IMAPCodec(codecOptions, iconv) { |
| 134 | this.iconv = iconv; |
| 135 | }; |
| 136 | |
| 137 | Utf7IMAPCodec.prototype.encoder = Utf7IMAPEncoder; |
| 138 | Utf7IMAPCodec.prototype.decoder = Utf7IMAPDecoder; |
| 139 | Utf7IMAPCodec.prototype.bomAware = true; |
| 140 | |
| 141 | |
| 142 | // -- Encoding |
| 143 | |
| 144 | function Utf7IMAPEncoder(options, codec) { |
| 145 | this.iconv = codec.iconv; |
| 146 | this.inBase64 = false; |
| 147 | this.base64Accum = Buffer.alloc(6); |
| 148 | this.base64AccumIdx = 0; |
| 149 | } |
| 150 | |
| 151 | Utf7IMAPEncoder.prototype.write = function(str) { |
| 152 | var inBase64 = this.inBase64, |
| 153 | base64Accum = this.base64Accum, |
| 154 | base64AccumIdx = this.base64AccumIdx, |
| 155 | buf = Buffer.alloc(str.length*5 + 10), bufIdx = 0; |
| 156 | |
| 157 | for (var i = 0; i < str.length; i++) { |
| 158 | var uChar = str.charCodeAt(i); |
| 159 | if (0x20 <= uChar && uChar <= 0x7E) { // Direct character or '&'. |
| 160 | if (inBase64) { |
| 161 | if (base64AccumIdx > 0) { |
| 162 | bufIdx += buf.write(base64Accum.slice(0, base64AccumIdx).toString('base64').replace(/\//g, ',').replace(/=+$/, ''), bufIdx); |
| 163 | base64AccumIdx = 0; |
| 164 | } |
| 165 | |
| 166 | buf[bufIdx++] = minusChar; // Write '-', then go to direct mode. |
| 167 | inBase64 = false; |
| 168 | } |
| 169 | |
| 170 | if (!inBase64) { |
| 171 | buf[bufIdx++] = uChar; // Write direct character |
| 172 | |
| 173 | if (uChar === andChar) // Ampersand -> '&-' |
| 174 | buf[bufIdx++] = minusChar; |
| 175 | } |
| 176 | |
| 177 | } else { // Non-direct character |
| 178 | if (!inBase64) { |
| 179 | buf[bufIdx++] = andChar; // Write '&', then go to base64 mode. |
| 180 | inBase64 = true; |
| 181 | } |
| 182 | if (inBase64) { |
| 183 | base64Accum[base64AccumIdx++] = uChar >> 8; |
| 184 | base64Accum[base64AccumIdx++] = uChar & 0xFF; |
| 185 | |
| 186 | if (base64AccumIdx == base64Accum.length) { |
| 187 | bufIdx += buf.write(base64Accum.toString('base64').replace(/\//g, ','), bufIdx); |
| 188 | base64AccumIdx = 0; |
| 189 | } |
| 190 | } |
| 191 | } |
| 192 | } |
| 193 | |
| 194 | this.inBase64 = inBase64; |
| 195 | this.base64AccumIdx = base64AccumIdx; |
| 196 | |
| 197 | return buf.slice(0, bufIdx); |
| 198 | } |
| 199 | |
| 200 | Utf7IMAPEncoder.prototype.end = function() { |
| 201 | var buf = Buffer.alloc(10), bufIdx = 0; |
| 202 | if (this.inBase64) { |
| 203 | if (this.base64AccumIdx > 0) { |
| 204 | bufIdx += buf.write(this.base64Accum.slice(0, this.base64AccumIdx).toString('base64').replace(/\//g, ',').replace(/=+$/, ''), bufIdx); |
| 205 | this.base64AccumIdx = 0; |
| 206 | } |
| 207 | |
| 208 | buf[bufIdx++] = minusChar; // Write '-', then go to direct mode. |
| 209 | this.inBase64 = false; |
| 210 | } |
| 211 | |
| 212 | return buf.slice(0, bufIdx); |
| 213 | } |
| 214 | |
| 215 | |
| 216 | // -- Decoding |
| 217 | |
| 218 | function Utf7IMAPDecoder(options, codec) { |
| 219 | this.iconv = codec.iconv; |
| 220 | this.inBase64 = false; |
| 221 | this.base64Accum = ''; |
| 222 | } |
| 223 | |
| 224 | var base64IMAPChars = base64Chars.slice(); |
| 225 | base64IMAPChars[','.charCodeAt(0)] = true; |
| 226 | |
| 227 | Utf7IMAPDecoder.prototype.write = function(buf) { |
| 228 | var res = "", lastI = 0, |
| 229 | inBase64 = this.inBase64, |
| 230 | base64Accum = this.base64Accum; |
| 231 | |
| 232 | // The decoder is more involved as we must handle chunks in stream. |
| 233 | // It is forgiving, closer to standard UTF-7 (for example, '-' is optional at the end). |
| 234 | |
| 235 | for (var i = 0; i < buf.length; i++) { |
| 236 | if (!inBase64) { // We're in direct mode. |
| 237 | // Write direct chars until '&' |
| 238 | if (buf[i] == andChar) { |
| 239 | res += this.iconv.decode(buf.slice(lastI, i), "ascii"); // Write direct chars. |
| 240 | lastI = i+1; |
| 241 | inBase64 = true; |
| 242 | } |
| 243 | } else { // We decode base64. |
| 244 | if (!base64IMAPChars[buf[i]]) { // Base64 ended. |
| 245 | if (i == lastI && buf[i] == minusChar) { // "&-" -> "&" |
| 246 | res += "&"; |
| 247 | } else { |
| 248 | var b64str = base64Accum + buf.slice(lastI, i).toString().replace(/,/g, '/'); |
| 249 | res += this.iconv.decode(Buffer.from(b64str, 'base64'), "utf16-be"); |
| 250 | } |
| 251 | |
| 252 | if (buf[i] != minusChar) // Minus may be absorbed after base64. |
| 253 | i--; |
| 254 | |
| 255 | lastI = i+1; |
| 256 | inBase64 = false; |
| 257 | base64Accum = ''; |
| 258 | } |
| 259 | } |
| 260 | } |
| 261 | |
| 262 | if (!inBase64) { |
| 263 | res += this.iconv.decode(buf.slice(lastI), "ascii"); // Write direct chars. |
| 264 | } else { |
| 265 | var b64str = base64Accum + buf.slice(lastI).toString().replace(/,/g, '/'); |
| 266 | |
| 267 | var canBeDecoded = b64str.length - (b64str.length % 8); // Minimal chunk: 2 quads -> 2x3 bytes -> 3 chars. |
| 268 | base64Accum = b64str.slice(canBeDecoded); // The rest will be decoded in future. |
| 269 | b64str = b64str.slice(0, canBeDecoded); |
| 270 | |
| 271 | res += this.iconv.decode(Buffer.from(b64str, 'base64'), "utf16-be"); |
| 272 | } |
| 273 | |
| 274 | this.inBase64 = inBase64; |
| 275 | this.base64Accum = base64Accum; |
| 276 | |
| 277 | return res; |
| 278 | } |
| 279 | |
| 280 | Utf7IMAPDecoder.prototype.end = function() { |
| 281 | var res = ""; |
| 282 | if (this.inBase64 && this.base64Accum.length > 0) |
| 283 | res = this.iconv.decode(Buffer.from(this.base64Accum, 'base64'), "utf16-be"); |
| 284 | |
| 285 | this.inBase64 = false; |
| 286 | this.base64Accum = ''; |
| 287 | return res; |
| 288 | } |
| 289 | |
| 290 | |