UNPKG

6.31 kBJavaScriptView Raw
1"use strict";
2var Buffer = require("safer-buffer").Buffer;
3
4// Export Node.js internal encodings.
5
6module.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
25function 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
43InternalCodec.prototype.encoder = InternalEncoder;
44InternalCodec.prototype.decoder = InternalDecoder;
45
46//------------------------------------------------------------------------------
47
48// We use node.js internal decoder. Its signature is the same as ours.
49var StringDecoder = require('string_decoder').StringDecoder;
50
51if (!StringDecoder.prototype.end) // Node v0.8 doesn't have this method.
52 StringDecoder.prototype.end = function() {};
53
54
55function InternalDecoder(options, codec) {
56 this.decoder = new StringDecoder(codec.enc);
57}
58
59InternalDecoder.prototype.write = function(buf) {
60 if (!Buffer.isBuffer(buf)) {
61 buf = Buffer.from(buf);
62 }
63
64 return this.decoder.write(buf);
65}
66
67InternalDecoder.prototype.end = function() {
68 return this.decoder.end();
69}
70
71
72//------------------------------------------------------------------------------
73// Encoder is mostly trivial
74
75function InternalEncoder(options, codec) {
76 this.enc = codec.enc;
77}
78
79InternalEncoder.prototype.write = function(str) {
80 return Buffer.from(str, this.enc);
81}
82
83InternalEncoder.prototype.end = function() {
84}
85
86
87//------------------------------------------------------------------------------
88// Except base64 encoder, which must keep its state.
89
90function InternalEncoderBase64(options, codec) {
91 this.prevStr = '';
92}
93
94InternalEncoderBase64.prototype.write = function(str) {
95 str = this.prevStr + str;
96 var completeQuads = str.length - (str.length % 4);
97 this.prevStr = str.slice(completeQuads);
98 str = str.slice(0, completeQuads);
99
100 return Buffer.from(str, "base64");
101}
102
103InternalEncoderBase64.prototype.end = function() {
104 return Buffer.from(this.prevStr, "base64");
105}
106
107
108//------------------------------------------------------------------------------
109// CESU-8 encoder is also special.
110
111function InternalEncoderCesu8(options, codec) {
112}
113
114InternalEncoderCesu8.prototype.write = function(str) {
115 var buf = Buffer.alloc(str.length * 3), bufIdx = 0;
116 for (var i = 0; i < str.length; i++) {
117 var charCode = str.charCodeAt(i);
118 // Naive implementation, but it works because CESU-8 is especially easy
119 // to convert from UTF-16 (which all JS strings are encoded in).
120 if (charCode < 0x80)
121 buf[bufIdx++] = charCode;
122 else if (charCode < 0x800) {
123 buf[bufIdx++] = 0xC0 + (charCode >>> 6);
124 buf[bufIdx++] = 0x80 + (charCode & 0x3f);
125 }
126 else { // charCode will always be < 0x10000 in javascript.
127 buf[bufIdx++] = 0xE0 + (charCode >>> 12);
128 buf[bufIdx++] = 0x80 + ((charCode >>> 6) & 0x3f);
129 buf[bufIdx++] = 0x80 + (charCode & 0x3f);
130 }
131 }
132 return buf.slice(0, bufIdx);
133}
134
135InternalEncoderCesu8.prototype.end = function() {
136}
137
138//------------------------------------------------------------------------------
139// CESU-8 decoder is not implemented in Node v4.0+
140
141function InternalDecoderCesu8(options, codec) {
142 this.acc = 0;
143 this.contBytes = 0;
144 this.accBytes = 0;
145 this.defaultCharUnicode = codec.defaultCharUnicode;
146}
147
148InternalDecoderCesu8.prototype.write = function(buf) {
149 var acc = this.acc, contBytes = this.contBytes, accBytes = this.accBytes,
150 res = '';
151 for (var i = 0; i < buf.length; i++) {
152 var curByte = buf[i];
153 if ((curByte & 0xC0) !== 0x80) { // Leading byte
154 if (contBytes > 0) { // Previous code is invalid
155 res += this.defaultCharUnicode;
156 contBytes = 0;
157 }
158
159 if (curByte < 0x80) { // Single-byte code
160 res += String.fromCharCode(curByte);
161 } else if (curByte < 0xE0) { // Two-byte code
162 acc = curByte & 0x1F;
163 contBytes = 1; accBytes = 1;
164 } else if (curByte < 0xF0) { // Three-byte code
165 acc = curByte & 0x0F;
166 contBytes = 2; accBytes = 1;
167 } else { // Four or more are not supported for CESU-8.
168 res += this.defaultCharUnicode;
169 }
170 } else { // Continuation byte
171 if (contBytes > 0) { // We're waiting for it.
172 acc = (acc << 6) | (curByte & 0x3f);
173 contBytes--; accBytes++;
174 if (contBytes === 0) {
175 // Check for overlong encoding, but support Modified UTF-8 (encoding NULL as C0 80)
176 if (accBytes === 2 && acc < 0x80 && acc > 0)
177 res += this.defaultCharUnicode;
178 else if (accBytes === 3 && acc < 0x800)
179 res += this.defaultCharUnicode;
180 else
181 // Actually add character.
182 res += String.fromCharCode(acc);
183 }
184 } else { // Unexpected continuation byte
185 res += this.defaultCharUnicode;
186 }
187 }
188 }
189 this.acc = acc; this.contBytes = contBytes; this.accBytes = accBytes;
190 return res;
191}
192
193InternalDecoderCesu8.prototype.end = function() {
194 var res = 0;
195 if (this.contBytes > 0)
196 res += this.defaultCharUnicode;
197 return res;
198}