1 /*
2 +----------------------------------------------------------------------+
3 | Copyright (c) The PHP Group |
4 +----------------------------------------------------------------------+
5 | This source file is subject to version 3.01 of the PHP license, |
6 | that is bundled with this package in the file LICENSE, and is |
7 | available through the world-wide-web at the following url: |
8 | https://www.php.net/license/3_01.txt |
9 | If you did not receive a copy of the PHP license and are unable to |
10 | obtain it through the world-wide-web, please send a note to |
11 | license@php.net so we can mail you a copy immediately. |
12 +----------------------------------------------------------------------+
13 | Author: Bjørn Borud - Guardian Networks AS <borud@guardian.no> |
14 +----------------------------------------------------------------------+
15 */
16
17 #include "php.h"
18 #include <stdlib.h>
19 #include <errno.h>
20 #include <ctype.h>
21 #include "php_string.h"
22
23 /* Simple soundex algorithm as described by Knuth in TAOCP, vol 3 */
24 /* {{{ Calculate the soundex key of a string */
PHP_FUNCTION(soundex)25 PHP_FUNCTION(soundex)
26 {
27 char *str;
28 size_t i, _small, str_len, code, last;
29 char soundex[4 + 1];
30
31 static const char soundex_table[26] =
32 {0, /* A */
33 '1', /* B */
34 '2', /* C */
35 '3', /* D */
36 0, /* E */
37 '1', /* F */
38 '2', /* G */
39 0, /* H */
40 0, /* I */
41 '2', /* J */
42 '2', /* K */
43 '4', /* L */
44 '5', /* M */
45 '5', /* N */
46 0, /* O */
47 '1', /* P */
48 '2', /* Q */
49 '6', /* R */
50 '2', /* S */
51 '3', /* T */
52 0, /* U */
53 '1', /* V */
54 0, /* W */
55 '2', /* X */
56 0, /* Y */
57 '2'}; /* Z */
58
59 ZEND_PARSE_PARAMETERS_START(1, 1)
60 Z_PARAM_STRING(str, str_len)
61 ZEND_PARSE_PARAMETERS_END();
62
63 /* build soundex string */
64 last = -1;
65 for (i = 0, _small = 0; i < str_len && _small < 4; i++) {
66 /* convert chars to upper case and strip non-letter chars */
67 /* BUG: should also map here accented letters used in non */
68 /* English words or names (also found in English text!): */
69 /* esstsett, thorn, n-tilde, c-cedilla, s-caron, ... */
70 code = toupper((int)(unsigned char)str[i]);
71 if (code >= 'A' && code <= 'Z') {
72 if (_small == 0) {
73 /* remember first valid char */
74 soundex[_small++] = (char)code;
75 last = soundex_table[code - 'A'];
76 }
77 else {
78 /* ignore sequences of consonants with same soundex */
79 /* code in trail, and vowels unless they separate */
80 /* consonant letters */
81 code = soundex_table[code - 'A'];
82 if (code != last) {
83 if (code != 0) {
84 soundex[_small++] = (char)code;
85 }
86 last = code;
87 }
88 }
89 }
90 }
91 /* pad with '0' and terminate with 0 ;-) */
92 while (_small < 4) {
93 soundex[_small++] = '0';
94 }
95 soundex[_small] = '\0';
96
97 RETURN_STRINGL(soundex, _small);
98 }
99 /* }}} */
100
