1 | // String encode/decode helpers
|
---|
2 | 'use strict';
|
---|
3 |
|
---|
4 |
|
---|
5 | var utils = require('./common');
|
---|
6 |
|
---|
7 |
|
---|
8 | // Quick check if we can use fast array to bin string conversion
|
---|
9 | //
|
---|
10 | // - apply(Array) can fail on Android 2.2
|
---|
11 | // - apply(Uint8Array) can fail on iOS 5.1 Safari
|
---|
12 | //
|
---|
13 | var STR_APPLY_OK = true;
|
---|
14 | var STR_APPLY_UIA_OK = true;
|
---|
15 |
|
---|
16 | try { String.fromCharCode.apply(null, [ 0 ]); } catch (__) { STR_APPLY_OK = false; }
|
---|
17 | try { String.fromCharCode.apply(null, new Uint8Array(1)); } catch (__) { STR_APPLY_UIA_OK = false; }
|
---|
18 |
|
---|
19 |
|
---|
20 | // Table with utf8 lengths (calculated by first byte of sequence)
|
---|
21 | // Note, that 5 & 6-byte values and some 4-byte values can not be represented in JS,
|
---|
22 | // because max possible codepoint is 0x10ffff
|
---|
23 | var _utf8len = new utils.Buf8(256);
|
---|
24 | for (var q = 0; q < 256; q++) {
|
---|
25 | _utf8len[q] = (q >= 252 ? 6 : q >= 248 ? 5 : q >= 240 ? 4 : q >= 224 ? 3 : q >= 192 ? 2 : 1);
|
---|
26 | }
|
---|
27 | _utf8len[254] = _utf8len[254] = 1; // Invalid sequence start
|
---|
28 |
|
---|
29 |
|
---|
30 | // convert string to array (typed, when possible)
|
---|
31 | exports.string2buf = function (str) {
|
---|
32 | var buf, c, c2, m_pos, i, str_len = str.length, buf_len = 0;
|
---|
33 |
|
---|
34 | // count binary size
|
---|
35 | for (m_pos = 0; m_pos < str_len; m_pos++) {
|
---|
36 | c = str.charCodeAt(m_pos);
|
---|
37 | if ((c & 0xfc00) === 0xd800 && (m_pos + 1 < str_len)) {
|
---|
38 | c2 = str.charCodeAt(m_pos + 1);
|
---|
39 | if ((c2 & 0xfc00) === 0xdc00) {
|
---|
40 | c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00);
|
---|
41 | m_pos++;
|
---|
42 | }
|
---|
43 | }
|
---|
44 | buf_len += c < 0x80 ? 1 : c < 0x800 ? 2 : c < 0x10000 ? 3 : 4;
|
---|
45 | }
|
---|
46 |
|
---|
47 | // allocate buffer
|
---|
48 | buf = new utils.Buf8(buf_len);
|
---|
49 |
|
---|
50 | // convert
|
---|
51 | for (i = 0, m_pos = 0; i < buf_len; m_pos++) {
|
---|
52 | c = str.charCodeAt(m_pos);
|
---|
53 | if ((c & 0xfc00) === 0xd800 && (m_pos + 1 < str_len)) {
|
---|
54 | c2 = str.charCodeAt(m_pos + 1);
|
---|
55 | if ((c2 & 0xfc00) === 0xdc00) {
|
---|
56 | c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00);
|
---|
57 | m_pos++;
|
---|
58 | }
|
---|
59 | }
|
---|
60 | if (c < 0x80) {
|
---|
61 | /* one byte */
|
---|
62 | buf[i++] = c;
|
---|
63 | } else if (c < 0x800) {
|
---|
64 | /* two bytes */
|
---|
65 | buf[i++] = 0xC0 | (c >>> 6);
|
---|
66 | buf[i++] = 0x80 | (c & 0x3f);
|
---|
67 | } else if (c < 0x10000) {
|
---|
68 | /* three bytes */
|
---|
69 | buf[i++] = 0xE0 | (c >>> 12);
|
---|
70 | buf[i++] = 0x80 | (c >>> 6 & 0x3f);
|
---|
71 | buf[i++] = 0x80 | (c & 0x3f);
|
---|
72 | } else {
|
---|
73 | /* four bytes */
|
---|
74 | buf[i++] = 0xf0 | (c >>> 18);
|
---|
75 | buf[i++] = 0x80 | (c >>> 12 & 0x3f);
|
---|
76 | buf[i++] = 0x80 | (c >>> 6 & 0x3f);
|
---|
77 | buf[i++] = 0x80 | (c & 0x3f);
|
---|
78 | }
|
---|
79 | }
|
---|
80 |
|
---|
81 | return buf;
|
---|
82 | };
|
---|
83 |
|
---|
84 | // Helper (used in 2 places)
|
---|
85 | function buf2binstring(buf, len) {
|
---|
86 | // On Chrome, the arguments in a function call that are allowed is `65534`.
|
---|
87 | // If the length of the buffer is smaller than that, we can use this optimization,
|
---|
88 | // otherwise we will take a slower path.
|
---|
89 | if (len < 65534) {
|
---|
90 | if ((buf.subarray && STR_APPLY_UIA_OK) || (!buf.subarray && STR_APPLY_OK)) {
|
---|
91 | return String.fromCharCode.apply(null, utils.shrinkBuf(buf, len));
|
---|
92 | }
|
---|
93 | }
|
---|
94 |
|
---|
95 | var result = '';
|
---|
96 | for (var i = 0; i < len; i++) {
|
---|
97 | result += String.fromCharCode(buf[i]);
|
---|
98 | }
|
---|
99 | return result;
|
---|
100 | }
|
---|
101 |
|
---|
102 |
|
---|
103 | // Convert byte array to binary string
|
---|
104 | exports.buf2binstring = function (buf) {
|
---|
105 | return buf2binstring(buf, buf.length);
|
---|
106 | };
|
---|
107 |
|
---|
108 |
|
---|
109 | // Convert binary string (typed, when possible)
|
---|
110 | exports.binstring2buf = function (str) {
|
---|
111 | var buf = new utils.Buf8(str.length);
|
---|
112 | for (var i = 0, len = buf.length; i < len; i++) {
|
---|
113 | buf[i] = str.charCodeAt(i);
|
---|
114 | }
|
---|
115 | return buf;
|
---|
116 | };
|
---|
117 |
|
---|
118 |
|
---|
119 | // convert array to string
|
---|
120 | exports.buf2string = function (buf, max) {
|
---|
121 | var i, out, c, c_len;
|
---|
122 | var len = max || buf.length;
|
---|
123 |
|
---|
124 | // Reserve max possible length (2 words per char)
|
---|
125 | // NB: by unknown reasons, Array is significantly faster for
|
---|
126 | // String.fromCharCode.apply than Uint16Array.
|
---|
127 | var utf16buf = new Array(len * 2);
|
---|
128 |
|
---|
129 | for (out = 0, i = 0; i < len;) {
|
---|
130 | c = buf[i++];
|
---|
131 | // quick process ascii
|
---|
132 | if (c < 0x80) { utf16buf[out++] = c; continue; }
|
---|
133 |
|
---|
134 | c_len = _utf8len[c];
|
---|
135 | // skip 5 & 6 byte codes
|
---|
136 | if (c_len > 4) { utf16buf[out++] = 0xfffd; i += c_len - 1; continue; }
|
---|
137 |
|
---|
138 | // apply mask on first byte
|
---|
139 | c &= c_len === 2 ? 0x1f : c_len === 3 ? 0x0f : 0x07;
|
---|
140 | // join the rest
|
---|
141 | while (c_len > 1 && i < len) {
|
---|
142 | c = (c << 6) | (buf[i++] & 0x3f);
|
---|
143 | c_len--;
|
---|
144 | }
|
---|
145 |
|
---|
146 | // terminated by end of string?
|
---|
147 | if (c_len > 1) { utf16buf[out++] = 0xfffd; continue; }
|
---|
148 |
|
---|
149 | if (c < 0x10000) {
|
---|
150 | utf16buf[out++] = c;
|
---|
151 | } else {
|
---|
152 | c -= 0x10000;
|
---|
153 | utf16buf[out++] = 0xd800 | ((c >> 10) & 0x3ff);
|
---|
154 | utf16buf[out++] = 0xdc00 | (c & 0x3ff);
|
---|
155 | }
|
---|
156 | }
|
---|
157 |
|
---|
158 | return buf2binstring(utf16buf, out);
|
---|
159 | };
|
---|
160 |
|
---|
161 |
|
---|
162 | // Calculate max possible position in utf8 buffer,
|
---|
163 | // that will not break sequence. If that's not possible
|
---|
164 | // - (very small limits) return max size as is.
|
---|
165 | //
|
---|
166 | // buf[] - utf8 bytes array
|
---|
167 | // max - length limit (mandatory);
|
---|
168 | exports.utf8border = function (buf, max) {
|
---|
169 | var pos;
|
---|
170 |
|
---|
171 | max = max || buf.length;
|
---|
172 | if (max > buf.length) { max = buf.length; }
|
---|
173 |
|
---|
174 | // go back from last position, until start of sequence found
|
---|
175 | pos = max - 1;
|
---|
176 | while (pos >= 0 && (buf[pos] & 0xC0) === 0x80) { pos--; }
|
---|
177 |
|
---|
178 | // Very small and broken sequence,
|
---|
179 | // return max, because we should return something anyway.
|
---|
180 | if (pos < 0) { return max; }
|
---|
181 |
|
---|
182 | // If we came to start of buffer - that means buffer is too small,
|
---|
183 | // return max too.
|
---|
184 | if (pos === 0) { return max; }
|
---|
185 |
|
---|
186 | return (pos + _utf8len[buf[pos]] > max) ? pos : max;
|
---|
187 | };
|
---|