/src/php-src/ext/standard/metaphone.c

Source (jump to first uncovered line)
/*
   +----------------------------------------------------------------------+
   | Copyright (c) The PHP Group                                          |
   +----------------------------------------------------------------------+
   | This source file is subject to version 3.01 of the PHP license,      |
   | that is bundled with this package in the file LICENSE, and is        |
   | available through the world-wide-web at the following url:           |
   | http://www.php.net/license/3_01.txt                                  |
   | If you did not receive a copy of the PHP license and are unable to   |
   | obtain it through the world-wide-web, please send a note to          |
   | license@php.net so we can mail you a copy immediately.               |
   +----------------------------------------------------------------------+
   | Author: Thies C. Arntzen <thies@thieso.net>                          |
   +----------------------------------------------------------------------+
*/

/*
  Based on CPANs "Text-Metaphone-1.96" by Michael G Schwern <schwern@pobox.com>
*/

#include "php.h"

static int metaphone(unsigned char *word, size_t word_len, zend_long max_phonemes, zend_string **phoned_word, int traditional);

/* {{{ Break english phrases down into their phonemes */
PHP_FUNCTION(metaphone)
{
  zend_string *str;
  zend_string *result = NULL;
  zend_long phones = 0;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(phones)
  ZEND_PARSE_PARAMETERS_END();

  if (metaphone((unsigned char *)ZSTR_VAL(str), ZSTR_LEN(str), phones, &result, 1) == 0) {
    RETVAL_STR(result);
  } else {
    if (result) {
      zend_string_free(result);
    }
    RETURN_FALSE;
  }
}
/* }}} */

/*
   this is now the original code by Michael G Schwern:
   i've changed it just a slightly bit (use emalloc,
   get rid of includes etc)
  - thies - 13.09.1999
*/

/*-----------------------------  */
/* this used to be "metaphone.h" */
/*-----------------------------  */

/* Special encodings */
#define  SH   'X'
#define  TH   '0'

/*-----------------------------  */
/* end of "metaphone.h"          */
/*-----------------------------  */

/*----------------------------- */
/* this used to be "metachar.h" */
/*----------------------------- */

/* Metachar.h ... little bits about characters for metaphone */
/*-- Character encoding array & accessing macros --*/
/* Stolen directly out of the book... */
static const char _codes[26] =
{
  1, 16, 4, 16, 9, 2, 4, 16, 9, 2, 0, 2, 2, 2, 1, 4, 0, 2, 4, 4, 1, 0, 0, 0, 8, 0
/*  a  b c  d e f g  h i j k l m n o p q r s t u v w x y z */
};


#define ENCODE(c) (isalpha(c) ? _codes[((toupper(c)) - 'A')] : 0)

#define isvowel(c)  (ENCODE(c) & 1)    /* AEIOU */

/* These letters are passed through unchanged */
#define NOCHANGE(c) (ENCODE(c) & 2)   /* FJMNR */

/* These form diphthongs when preceding H */
#define AFFECTH(c)  (ENCODE(c) & 4)    /* CGPST */

/* These make C and G soft */
#define MAKESOFT(c) (ENCODE(c) & 8)    /* EIY */

/* These prevent GH from becoming F */
#define NOGHTOF(c)  (ENCODE(c) & 16)  /* BDH */

/*----------------------------- */
/* end of "metachar.h"          */
/*----------------------------- */

/* I suppose I could have been using a character pointer instead of
 * accesssing the array directly... */

/* Look at the next letter in the word */
#define Next_Letter (toupper(word[w_idx+1]))
/* Look at the current letter in the word */
#define Curr_Letter (toupper(word[w_idx]))
/* Go N letters back. */
#define Look_Back_Letter(n) (w_idx >= n ? toupper(word[w_idx-n]) : '\0')
/* Previous letter.  I dunno, should this return null on failure? */
#define Prev_Letter (Look_Back_Letter(1))
/* Look two letters down.  It makes sure you don't walk off the string. */
#define After_Next_Letter (Next_Letter != '\0' ? toupper(word[w_idx+2]) \
                           : '\0')
#define Look_Ahead_Letter(n) (toupper(Lookahead((char *) word+w_idx, n)))


/* Allows us to safely look ahead an arbitrary # of letters */
/* I probably could have just used strlen... */
static char Lookahead(char *word, int how_far)
{
  char letter_ahead = '\0'; /* null by default */
  int idx;
  for (idx = 0; word[idx] != '\0' && idx < how_far; idx++);
  /* Edge forward in the string... */

  letter_ahead = word[idx]; /* idx will be either == to how_far or
                 * at the end of the string
                 */
  return letter_ahead;
}


/* phonize one letter
 * We don't know the buffers size in advance. On way to solve this is to just
 * re-allocate the buffer size. We're using an extra of 2 characters (this
 * could be one though; or more too). */
#define Phonize(c)  { \
            if (p_idx >= max_buffer_len) { \
              *phoned_word = zend_string_extend(*phoned_word, 2 * sizeof(char) + max_buffer_len, 0); \
              max_buffer_len += 2; \
            } \
            ZSTR_VAL(*phoned_word)[p_idx++] = c; \
            ZSTR_LEN(*phoned_word) = p_idx; \
          }
/* Slap a null character on the end of the phoned word */
#define End_Phoned_Word { \
              if (p_idx == max_buffer_len) { \
                *phoned_word = zend_string_extend(*phoned_word, 1 * sizeof(char) + max_buffer_len, 0); \
                max_buffer_len += 1; \
              } \
              ZSTR_VAL(*phoned_word)[p_idx] = '\0'; \
              ZSTR_LEN(*phoned_word) = p_idx; \
            }
/* How long is the phoned word? */
#define Phone_Len (p_idx)

/* Note is a letter is a 'break' in the word */
#define Isbreak(c)  (!isalpha(c))

/* {{{ metaphone */
static int metaphone(unsigned char *word, size_t word_len, zend_long max_phonemes, zend_string **phoned_word, int traditional)
{
  int w_idx = 0;        /* point in the phonization we're at. */
  size_t p_idx = 0;       /* end of the phoned phrase */
  size_t max_buffer_len = 0;    /* maximum length of the destination buffer */

/*-- Parameter checks --*/
  /* Negative phoneme length is meaningless */

  if (max_phonemes < 0)
    return -1;

  /* Empty/null string is meaningless */
  /* Overly paranoid */
  /* assert(word != NULL && word[0] != '\0'); */

  if (word == NULL)
    return -1;

/*-- Allocate memory for our phoned_phrase --*/
  if (max_phonemes == 0) { /* Assume largest possible */
    max_buffer_len = word_len;
    *phoned_word = zend_string_alloc(sizeof(char) * word_len + 1, 0);
  } else {
    max_buffer_len = max_phonemes;
    *phoned_word = zend_string_alloc(sizeof(char) * max_phonemes + 1, 0);
  }


/*-- The first phoneme has to be processed specially. --*/
  /* Find our first letter */
  for (; !isalpha(Curr_Letter); w_idx++) {
    /* On the off chance we were given nothing but crap... */
    if (Curr_Letter == '\0') {
      End_Phoned_Word
        return SUCCESS; /* For testing */
    }
  }

  switch (Curr_Letter) {
    /* AE becomes E */
  case 'A':
    if (Next_Letter == 'E') {
      Phonize('E');
      w_idx += 2;
    }
    /* Remember, preserve vowels at the beginning */
    else {
      Phonize('A');
      w_idx++;
    }
    break;
    /* [GKP]N becomes N */
  case 'G':
  case 'K':
  case 'P':
    if (Next_Letter == 'N') {
      Phonize('N');
      w_idx += 2;
    }
    break;
    /* WH becomes W,
       WR becomes R
       W if followed by a vowel */
  case 'W':
    if (Next_Letter == 'R') {
      Phonize(Next_Letter);
      w_idx += 2;
    } else if (Next_Letter == 'H' || isvowel(Next_Letter)) {
      Phonize('W');
      w_idx += 2;
    }
    /* else ignore */
    break;
    /* X becomes S */
  case 'X':
    Phonize('S');
    w_idx++;
    break;
    /* Vowels are kept */
    /* We did A already
       case 'A':
       case 'a':
     */
  case 'E':
  case 'I':
  case 'O':
  case 'U':
    Phonize(Curr_Letter);
    w_idx++;
    break;
  default:
    /* do nothing */
    break;
  }



  /* On to the metaphoning */
  for (; Curr_Letter != '\0' &&
     (max_phonemes == 0 || Phone_Len < (size_t)max_phonemes);
     w_idx++) {
    /* How many letters to skip because an eariler encoding handled
     * multiple letters */
    unsigned short int skip_letter = 0;


    /* THOUGHT:  It would be nice if, rather than having things like...
     * well, SCI.  For SCI you encode the S, then have to remember
     * to skip the C.  So the phonome SCI invades both S and C.  It would
     * be better, IMHO, to skip the C from the S part of the encoding.
     * Hell, I'm trying it.
     */

    /* Ignore non-alphas */
    if (!isalpha(Curr_Letter))
      continue;

    /* Drop duplicates, except CC */
    if (Curr_Letter == Prev_Letter &&
      Curr_Letter != 'C')
      continue;

    switch (Curr_Letter) {
      /* B -> B unless in MB */
    case 'B':
      if (Prev_Letter != 'M')
        Phonize('B');
      break;
      /* 'sh' if -CIA- or -CH, but not SCH, except SCHW.
       * (SCHW is handled in S)
       *  S if -CI-, -CE- or -CY-
       *  dropped if -SCI-, SCE-, -SCY- (handed in S)
       *  else K
       */
    case 'C':
      if (MAKESOFT(Next_Letter)) { /* C[IEY] */
        if (After_Next_Letter == 'A' &&
          Next_Letter == 'I') { /* CIA */
          Phonize(SH);
        }
        /* SC[IEY] */
        else if (Prev_Letter == 'S') {
          /* Dropped */
        } else {
          Phonize('S');
        }
      } else if (Next_Letter == 'H') {
        if ((!traditional) && (After_Next_Letter == 'R' || Prev_Letter == 'S')) { /* Christ, School */
          Phonize('K');
        } else {
          Phonize(SH);
        }
        skip_letter++;
      } else {
        Phonize('K');
      }
      break;
      /* J if in -DGE-, -DGI- or -DGY-
       * else T
       */
    case 'D':
      if (Next_Letter == 'G' &&
        MAKESOFT(After_Next_Letter)) {
        Phonize('J');
        skip_letter++;
      } else
        Phonize('T');
      break;
      /* F if in -GH and not B--GH, D--GH, -H--GH, -H---GH
       * else dropped if -GNED, -GN,
       * else dropped if -DGE-, -DGI- or -DGY- (handled in D)
       * else J if in -GE-, -GI, -GY and not GG
       * else K
       */
    case 'G':
      if (Next_Letter == 'H') {
        if (!(NOGHTOF(Look_Back_Letter(3)) ||
            Look_Back_Letter(4) == 'H')) {
          Phonize('F');
          skip_letter++;
        } else {
          /* silent */
        }
      } else if (Next_Letter == 'N') {
        if (Isbreak(After_Next_Letter) ||
          (After_Next_Letter == 'E' &&
           Look_Ahead_Letter(3) == 'D')) {
          /* dropped */
        } else
          Phonize('K');
      } else if (MAKESOFT(Next_Letter) &&
             Prev_Letter != 'G') {
        Phonize('J');
      } else {
        Phonize('K');
      }
      break;
      /* H if before a vowel and not after C,G,P,S,T */
    case 'H':
      if (isvowel(Next_Letter) &&
        !AFFECTH(Prev_Letter))
        Phonize('H');
      break;
      /* dropped if after C
       * else K
       */
    case 'K':
      if (Prev_Letter != 'C')
        Phonize('K');
      break;
      /* F if before H
       * else P
       */
    case 'P':
      if (Next_Letter == 'H') {
        Phonize('F');
      } else {
        Phonize('P');
      }
      break;
      /* K
       */
    case 'Q':
      Phonize('K');
      break;
      /* 'sh' in -SH-, -SIO- or -SIA- or -SCHW-
       * else S
       */
    case 'S':
      if (Next_Letter == 'I' &&
        (After_Next_Letter == 'O' ||
         After_Next_Letter == 'A')) {
        Phonize(SH);
      } else if (Next_Letter == 'H') {
        Phonize(SH);
        skip_letter++;
      } else if ((!traditional) && (Next_Letter == 'C' && Look_Ahead_Letter(2) == 'H' && Look_Ahead_Letter(3) == 'W')) {
        Phonize(SH);
        skip_letter += 2;
      } else {
        Phonize('S');
      }
      break;
      /* 'sh' in -TIA- or -TIO-
       * else 'th' before H
       * else T
       */
    case 'T':
      if (Next_Letter == 'I' &&
        (After_Next_Letter == 'O' ||
         After_Next_Letter == 'A')) {
        Phonize(SH);
      } else if (Next_Letter == 'H') {
        Phonize(TH);
        skip_letter++;
      } else if (!(Next_Letter == 'C' && After_Next_Letter == 'H')) {
        Phonize('T');
      }
      break;
      /* F */
    case 'V':
      Phonize('F');
      break;
      /* W before a vowel, else dropped */
    case 'W':
      if (isvowel(Next_Letter))
        Phonize('W');
      break;
      /* KS */
    case 'X':
      Phonize('K');
      Phonize('S');
      break;
      /* Y if followed by a vowel */
    case 'Y':
      if (isvowel(Next_Letter))
        Phonize('Y');
      break;
      /* S */
    case 'Z':
      Phonize('S');
      break;
      /* No transformation */
    case 'F':
    case 'J':
    case 'L':
    case 'M':
    case 'N':
    case 'R':
      Phonize(Curr_Letter);
      break;
    default:
      /* nothing */
      break;
    }           /* END SWITCH */

    w_idx += skip_letter;
  }              /* END FOR */

  End_Phoned_Word;

  return 0;
}                /* END metaphone */
/* }}} */

Coverage Report

Created: 2022-10-06 21:30

Line	Count	Source (jump to first uncovered line)
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		\| http://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: Thies C. Arntzen <thies@thieso.net> \|
14		+----------------------------------------------------------------------+
15		*/
16
17		/*
18		Based on CPANs "Text-Metaphone-1.96" by Michael G Schwern <schwern@pobox.com>
19		*/
20
21		#include "php.h"
22
23		static int metaphone(unsigned char word, size_t word_len, zend_long max_phonemes, zend_string *phoned_word, int traditional);
24
25		/* {{{ Break english phrases down into their phonemes */
26		PHP_FUNCTION(metaphone)
27	0	{
28	0	zend_string *str;
29	0	zend_string *result = NULL;
30	0	zend_long phones = 0;
31
32	0	ZEND_PARSE_PARAMETERS_START(1, 2)
33	0	Z_PARAM_STR(str)
34	0	Z_PARAM_OPTIONAL
35	0	Z_PARAM_LONG(phones)
36	0	ZEND_PARSE_PARAMETERS_END();
37
38	0	if (metaphone((unsigned char *)ZSTR_VAL(str), ZSTR_LEN(str), phones, &result, 1) == 0) {
39	0	RETVAL_STR(result);
40	0	} else {
41	0	if (result) {
42	0	zend_string_free(result);
43	0	}
44	0	RETURN_FALSE;
45	0	}
46	0	}
47		/* }}} */
48
49		/*
50		this is now the original code by Michael G Schwern:
51		i've changed it just a slightly bit (use emalloc,
52		get rid of includes etc)
53		- thies - 13.09.1999
54		*/
55
56		/----------------------------- /
57		/* this used to be "metaphone.h" */
58		/----------------------------- /
59
60		/* Special encodings */
61		#define SH 'X'
62		#define TH '0'
63
64		/----------------------------- /
65		/* end of "metaphone.h" */
66		/----------------------------- /
67
68		/----------------------------- /
69		/* this used to be "metachar.h" */
70		/----------------------------- /
71
72		/* Metachar.h ... little bits about characters for metaphone */
73		/-- Character encoding array & accessing macros --/
74		/* Stolen directly out of the book... */
75		static const char _codes[26] =
76		{
77		1, 16, 4, 16, 9, 2, 4, 16, 9, 2, 0, 2, 2, 2, 1, 4, 0, 2, 4, 4, 1, 0, 0, 0, 8, 0
78		/* a b c d e f g h i j k l m n o p q r s t u v w x y z */
79		};
80
81
82	0	#define ENCODE(c) (isalpha(c) ? _codes[((toupper(c)) - 'A')] : 0)
83
84	0	#define isvowel(c) (ENCODE(c) & 1) /* AEIOU */
85
86		/* These letters are passed through unchanged */
87		#define NOCHANGE(c) (ENCODE(c) & 2) /* FJMNR */
88
89		/* These form diphthongs when preceding H */
90	0	#define AFFECTH(c) (ENCODE(c) & 4) /* CGPST */
91
92		/* These make C and G soft */
93	0	#define MAKESOFT(c) (ENCODE(c) & 8) /* EIY */
94
95		/* These prevent GH from becoming F */
96	0	#define NOGHTOF(c) (ENCODE(c) & 16) /* BDH */
97
98		/----------------------------- /
99		/* end of "metachar.h" */
100		/----------------------------- /
101
102		/* I suppose I could have been using a character pointer instead of
103		* accesssing the array directly... */
104
105		/* Look at the next letter in the word */
106	0	#define Next_Letter (toupper(word[w_idx+1]))
107		/* Look at the current letter in the word */
108	0	#define Curr_Letter (toupper(word[w_idx]))
109		/* Go N letters back. */
110	0	#define Look_Back_Letter(n) (w_idx >= n ? toupper(word[w_idx-n]) : '\0')
111		/* Previous letter. I dunno, should this return null on failure? */
112	0	#define Prev_Letter (Look_Back_Letter(1))
113		/* Look two letters down. It makes sure you don't walk off the string. */
114	0	#define After_Next_Letter (Next_Letter != '\0' ? toupper(word[w_idx+2]) \
115	0	: '\0')
116	0	#define Look_Ahead_Letter(n) (toupper(Lookahead((char *) word+w_idx, n)))
117
118
119		/* Allows us to safely look ahead an arbitrary # of letters */
120		/* I probably could have just used strlen... */
121		static char Lookahead(char *word, int how_far)
122	0	{
123	0	char letter_ahead = '\0'; /* null by default */
124	0	int idx;
125	0	for (idx = 0; word[idx] != '\0' && idx < how_far; idx++);
126		/* Edge forward in the string... */
127
128	0	letter_ahead = word[idx]; /* idx will be either == to how_far or
129		* at the end of the string
130		*/
131	0	return letter_ahead;
132	0	}
133
134
135		/* phonize one letter
136		* We don't know the buffers size in advance. On way to solve this is to just
137		* re-allocate the buffer size. We're using an extra of 2 characters (this
138		* could be one though; or more too). */
139	0	#define Phonize(c) { \
140	0	if (p_idx >= max_buffer_len) { \
141	0	phoned_word = zend_string_extend(phoned_word, 2 * sizeof(char) + max_buffer_len, 0); \
142	0	max_buffer_len += 2; \
143	0	} \
144	0	ZSTR_VAL(*phoned_word)[p_idx++] = c; \
145	0	ZSTR_LEN(*phoned_word) = p_idx; \
146	0	}
147		/* Slap a null character on the end of the phoned word */
148	0	#define End_Phoned_Word { \
149	0	if (p_idx == max_buffer_len) { \
150	0	phoned_word = zend_string_extend(phoned_word, 1 * sizeof(char) + max_buffer_len, 0); \
151	0	max_buffer_len += 1; \
152	0	} \
153	0	ZSTR_VAL(*phoned_word)[p_idx] = '\0'; \
154	0	ZSTR_LEN(*phoned_word) = p_idx; \
155	0	}
156		/* How long is the phoned word? */
157	0	#define Phone_Len (p_idx)
158
159		/* Note is a letter is a 'break' in the word */
160	0	#define Isbreak(c) (!isalpha(c))
161
162		/* {{{ metaphone */
163		static int metaphone(unsigned char word, size_t word_len, zend_long max_phonemes, zend_string *phoned_word, int traditional)
164	0	{
165	0	int w_idx = 0; /* point in the phonization we're at. */
166	0	size_t p_idx = 0; /* end of the phoned phrase */
167	0	size_t max_buffer_len = 0; /* maximum length of the destination buffer */
168
169		/-- Parameter checks --/
170		/* Negative phoneme length is meaningless */
171
172	0	if (max_phonemes < 0)
173	0	return -1;
174
175		/* Empty/null string is meaningless */
176		/* Overly paranoid */
177		/* assert(word != NULL && word[0] != '\0'); */
178
179	0	if (word == NULL)
180	0	return -1;
181
182		/-- Allocate memory for our phoned_phrase --/
183	0	if (max_phonemes == 0) { /* Assume largest possible */
184	0	max_buffer_len = word_len;
185	0	phoned_word = zend_string_alloc(sizeof(char) word_len + 1, 0);
186	0	} else {
187	0	max_buffer_len = max_phonemes;
188	0	phoned_word = zend_string_alloc(sizeof(char) max_phonemes + 1, 0);
189	0	}
190
191
192		/-- The first phoneme has to be processed specially. --/
193		/* Find our first letter */
194	0	for (; !isalpha(Curr_Letter); w_idx++) {
195		/* On the off chance we were given nothing but crap... */
196	0	if (Curr_Letter == '\0') {
197	0	End_Phoned_Word
198	0	return SUCCESS; /* For testing */
199	0	}
200	0	}
201
202	0	switch (Curr_Letter) {
203		/* AE becomes E */
204	0	case 'A':
205	0	if (Next_Letter == 'E') {
206	0	Phonize('E');
207	0	w_idx += 2;
208	0	}
209		/* Remember, preserve vowels at the beginning */
210	0	else {
211	0	Phonize('A');
212	0	w_idx++;
213	0	}
214	0	break;
215		/* [GKP]N becomes N */
216	0	case 'G':
217	0	case 'K':
218	0	case 'P':
219	0	if (Next_Letter == 'N') {
220	0	Phonize('N');
221	0	w_idx += 2;
222	0	}
223	0	break;
224		/* WH becomes W,
225		WR becomes R
226		W if followed by a vowel */
227	0	case 'W':
228	0	if (Next_Letter == 'R') {
229	0	Phonize(Next_Letter);
230	0	w_idx += 2;
231	0	} else if (Next_Letter == 'H' \|\| isvowel(Next_Letter)) {
232	0	Phonize('W');
233	0	w_idx += 2;
234	0	}
235		/* else ignore */
236	0	break;
237		/* X becomes S */
238	0	case 'X':
239	0	Phonize('S');
240	0	w_idx++;
241	0	break;
242		/* Vowels are kept */
243		/* We did A already
244		case 'A':
245		case 'a':
246		*/
247	0	case 'E':
248	0	case 'I':
249	0	case 'O':
250	0	case 'U':
251	0	Phonize(Curr_Letter);
252	0	w_idx++;
253	0	break;
254	0	default:
255		/* do nothing */
256	0	break;
257	0	}
258
259
260
261		/* On to the metaphoning */
262	0	for (; Curr_Letter != '\0' &&
263	0	(max_phonemes == 0 \|\| Phone_Len < (size_t)max_phonemes);
264	0	w_idx++) {
265		/* How many letters to skip because an eariler encoding handled
266		* multiple letters */
267	0	unsigned short int skip_letter = 0;
268
269
270		/* THOUGHT: It would be nice if, rather than having things like...
271		* well, SCI. For SCI you encode the S, then have to remember
272		* to skip the C. So the phonome SCI invades both S and C. It would
273		* be better, IMHO, to skip the C from the S part of the encoding.
274		* Hell, I'm trying it.
275		*/
276
277		/* Ignore non-alphas */
278	0	if (!isalpha(Curr_Letter))
279	0	continue;
280
281		/* Drop duplicates, except CC */
282	0	if (Curr_Letter == Prev_Letter &&
283	0	Curr_Letter != 'C')
284	0	continue;
285
286	0	switch (Curr_Letter) {
287		/* B -> B unless in MB */
288	0	case 'B':
289	0	if (Prev_Letter != 'M')
290	0	Phonize('B');
291	0	break;
292		/* 'sh' if -CIA- or -CH, but not SCH, except SCHW.
293		* (SCHW is handled in S)
294		* S if -CI-, -CE- or -CY-
295		* dropped if -SCI-, SCE-, -SCY- (handed in S)
296		* else K
297		*/
298	0	case 'C':
299	0	if (MAKESOFT(Next_Letter)) { /* C[IEY] */
300	0	if (After_Next_Letter == 'A' &&
301	0	Next_Letter == 'I') { /* CIA */
302	0	Phonize(SH);
303	0	}
304		/* SC[IEY] */
305	0	else if (Prev_Letter == 'S') {
306		/* Dropped */
307	0	} else {
308	0	Phonize('S');
309	0	}
310	0	} else if (Next_Letter == 'H') {
311	0	if ((!traditional) && (After_Next_Letter == 'R' \|\| Prev_Letter == 'S')) { /* Christ, School */
312	0	Phonize('K');
313	0	} else {
314	0	Phonize(SH);
315	0	}
316	0	skip_letter++;
317	0	} else {
318	0	Phonize('K');
319	0	}
320	0	break;
321		/* J if in -DGE-, -DGI- or -DGY-
322		* else T
323		*/
324	0	case 'D':
325	0	if (Next_Letter == 'G' &&
326	0	MAKESOFT(After_Next_Letter)) {
327	0	Phonize('J');
328	0	skip_letter++;
329	0	} else
330	0	Phonize('T');
331	0	break;
332		/* F if in -GH and not B--GH, D--GH, -H--GH, -H---GH
333		* else dropped if -GNED, -GN,
334		* else dropped if -DGE-, -DGI- or -DGY- (handled in D)
335		* else J if in -GE-, -GI, -GY and not GG
336		* else K
337		*/
338	0	case 'G':
339	0	if (Next_Letter == 'H') {
340	0	if (!(NOGHTOF(Look_Back_Letter(3)) \|\|
341	0	Look_Back_Letter(4) == 'H')) {
342	0	Phonize('F');
343	0	skip_letter++;
344	0	} else {
345		/* silent */
346	0	}
347	0	} else if (Next_Letter == 'N') {
348	0	if (Isbreak(After_Next_Letter) \|\|
349	0	(After_Next_Letter == 'E' &&
350	0	Look_Ahead_Letter(3) == 'D')) {
351		/* dropped */
352	0	} else
353	0	Phonize('K');
354	0	} else if (MAKESOFT(Next_Letter) &&
355	0	Prev_Letter != 'G') {
356	0	Phonize('J');
357	0	} else {
358	0	Phonize('K');
359	0	}
360	0	break;
361		/* H if before a vowel and not after C,G,P,S,T */
362	0	case 'H':
363	0	if (isvowel(Next_Letter) &&
364	0	!AFFECTH(Prev_Letter))
365	0	Phonize('H');
366	0	break;
367		/* dropped if after C
368		* else K
369		*/
370	0	case 'K':
371	0	if (Prev_Letter != 'C')
372	0	Phonize('K');
373	0	break;
374		/* F if before H
375		* else P
376		*/
377	0	case 'P':
378	0	if (Next_Letter == 'H') {
379	0	Phonize('F');
380	0	} else {
381	0	Phonize('P');
382	0	}
383	0	break;
384		/* K
385		*/
386	0	case 'Q':
387	0	Phonize('K');
388	0	break;
389		/* 'sh' in -SH-, -SIO- or -SIA- or -SCHW-
390		* else S
391		*/
392	0	case 'S':
393	0	if (Next_Letter == 'I' &&
394	0	(After_Next_Letter == 'O' \|\|
395	0	After_Next_Letter == 'A')) {
396	0	Phonize(SH);
397	0	} else if (Next_Letter == 'H') {
398	0	Phonize(SH);
399	0	skip_letter++;
400	0	} else if ((!traditional) && (Next_Letter == 'C' && Look_Ahead_Letter(2) == 'H' && Look_Ahead_Letter(3) == 'W')) {
401	0	Phonize(SH);
402	0	skip_letter += 2;
403	0	} else {
404	0	Phonize('S');
405	0	}
406	0	break;
407		/* 'sh' in -TIA- or -TIO-
408		* else 'th' before H
409		* else T
410		*/
411	0	case 'T':
412	0	if (Next_Letter == 'I' &&
413	0	(After_Next_Letter == 'O' \|\|
414	0	After_Next_Letter == 'A')) {
415	0	Phonize(SH);
416	0	} else if (Next_Letter == 'H') {
417	0	Phonize(TH);
418	0	skip_letter++;
419	0	} else if (!(Next_Letter == 'C' && After_Next_Letter == 'H')) {
420	0	Phonize('T');
421	0	}
422	0	break;
423		/* F */
424	0	case 'V':
425	0	Phonize('F');
426	0	break;
427		/* W before a vowel, else dropped */
428	0	case 'W':
429	0	if (isvowel(Next_Letter))
430	0	Phonize('W');
431	0	break;
432		/* KS */
433	0	case 'X':
434	0	Phonize('K');
435	0	Phonize('S');
436	0	break;
437		/* Y if followed by a vowel */
438	0	case 'Y':
439	0	if (isvowel(Next_Letter))
440	0	Phonize('Y');
441	0	break;
442		/* S */
443	0	case 'Z':
444	0	Phonize('S');
445	0	break;
446		/* No transformation */
447	0	case 'F':
448	0	case 'J':
449	0	case 'L':
450	0	case 'M':
451	0	case 'N':
452	0	case 'R':
453	0	Phonize(Curr_Letter);
454	0	break;
455	0	default:
456		/* nothing */
457	0	break;
458	0	} /* END SWITCH */
459
460	0	w_idx += skip_letter;
461	0	} /* END FOR */
462
463	0	End_Phoned_Word;
464
465	0	return 0;
466	0	} /* END metaphone */
467		/* }}} */