xref: /PHP-5.4/ext/intl/grapheme/grapheme_util.c (revision 9ad97cd4)
1 /*
2    +----------------------------------------------------------------------+
3    | PHP Version 5                                                        |
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: Ed Batutis <ed@batutis.com>                                  |
14    +----------------------------------------------------------------------+
15  */
16 
17 /* {{{ includes */
18 #ifdef HAVE_CONFIG_H
19 #include "config.h"
20 #endif
21 
22 #include <php.h>
23 #include "grapheme.h"
24 #include "grapheme_util.h"
25 #include "intl_common.h"
26 
27 #include <unicode/utypes.h>
28 #include <unicode/ucol.h>
29 #include <unicode/ustring.h>
30 #include <unicode/ubrk.h>
31 #include <unicode/usearch.h>
32 
33 #include "ext/standard/php_string.h"
34 
ZEND_EXTERN_MODULE_GLOBALS(intl)35 ZEND_EXTERN_MODULE_GLOBALS( intl )
36 
37 /* }}} */
38 
39 /* {{{ grapheme_close_global_iterator - clean up */
40 void
41 grapheme_close_global_iterator( TSRMLS_D )
42 {
43 	UBreakIterator *global_break_iterator = INTL_G( grapheme_iterator );
44 
45 	if ( NULL != global_break_iterator ) {
46 		ubrk_close(global_break_iterator);
47 	}
48 }
49 /* }}} */
50 
51 /* {{{ grapheme_substr_ascii f='from' - starting point, l='length' */
grapheme_substr_ascii(char * str,int str_len,int f,int l,int argc,char ** sub_str,int * sub_str_len)52 void grapheme_substr_ascii(char *str, int str_len, int f, int l, int argc, char **sub_str, int *sub_str_len)
53 {
54     *sub_str = NULL;
55 
56     if (argc > 2) {
57         if ((l < 0 && -l > str_len)) {
58             return;
59         } else if (l > str_len) {
60             l = str_len;
61         }
62     } else {
63         l = str_len;
64     }
65 
66     if (f > str_len || (f < 0 && -f > str_len)) {
67         return;
68     }
69 
70     if (l < 0 && (l + str_len - f) < 0) {
71         return;
72     }
73 
74     /* if "from" position is negative, count start position from the end
75      * of the string
76      */
77     if (f < 0) {
78         f = str_len + f;
79         if (f < 0) {
80             f = 0;
81         }
82     }
83 
84 
85     /* if "length" position is negative, set it to the length
86      * needed to stop that many chars from the end of the string
87      */
88     if (l < 0) {
89         l = (str_len - f) + l;
90         if (l < 0) {
91             l = 0;
92         }
93     }
94 
95     if (f >= str_len) {
96         return;
97     }
98 
99     if ((f + l) > str_len) {
100         l = str_len - f;
101     }
102 
103     *sub_str = str + f;
104     *sub_str_len = l;
105 
106     return;
107 }
108 /* }}} */
109 
110 #define STRPOS_CHECK_STATUS(status, error) 							\
111 	if ( U_FAILURE( (status) ) ) { 									\
112 		intl_error_set_code( NULL, (status) TSRMLS_CC ); 			\
113 		intl_error_set_custom_msg( NULL, (error), 0 TSRMLS_CC ); 	\
114 		if (uhaystack) { 											\
115 			efree( uhaystack ); 									\
116 		} 															\
117 		if (uneedle) { 												\
118 			efree( uneedle ); 										\
119 		} 															\
120 		if(bi) { 													\
121 			ubrk_close (bi); 										\
122 		} 															\
123 		if(src) {													\
124 			usearch_close(src);										\
125 		}															\
126 		return -1; 													\
127 	}
128 
129 
130 /* {{{ grapheme_strpos_utf16 - strrpos using utf16*/
grapheme_strpos_utf16(unsigned char * haystack,int32_t haystack_len,unsigned char * needle,int32_t needle_len,int32_t offset,int32_t * puchar_pos,int f_ignore_case,int last TSRMLS_DC)131 int grapheme_strpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int32_t *puchar_pos, int f_ignore_case, int last TSRMLS_DC)
132 {
133 	UChar *uhaystack = NULL, *uneedle = NULL;
134 	int32_t uhaystack_len = 0, uneedle_len = 0, char_pos, ret_pos, offset_pos = 0;
135 	unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
136 	UBreakIterator* bi = NULL;
137 	UErrorCode status;
138 	UStringSearch* src = NULL;
139 	UCollator *coll;
140 
141 	if(puchar_pos) {
142 		*puchar_pos = -1;
143 	}
144 	/* convert the strings to UTF-16. */
145 
146 	status = U_ZERO_ERROR;
147 	intl_convert_utf8_to_utf16(&uhaystack, &uhaystack_len, (char *) haystack, haystack_len, &status );
148 	STRPOS_CHECK_STATUS(status, "Error converting input string to UTF-16");
149 
150 	status = U_ZERO_ERROR;
151 	intl_convert_utf8_to_utf16(&uneedle, &uneedle_len, (char *) needle, needle_len, &status );
152 	STRPOS_CHECK_STATUS(status, "Error converting input string to UTF-16");
153 
154 	/* get a pointer to the haystack taking into account the offset */
155 	status = U_ZERO_ERROR;
156 	bi = grapheme_get_break_iterator(u_break_iterator_buffer, &status TSRMLS_CC );
157 	STRPOS_CHECK_STATUS(status, "Failed to get iterator");
158 	status = U_ZERO_ERROR;
159 	ubrk_setText(bi, uhaystack, uhaystack_len, &status);
160 	STRPOS_CHECK_STATUS(status, "Failed to set up iterator");
161 
162 	status = U_ZERO_ERROR;
163 	src = usearch_open(uneedle, uneedle_len, uhaystack, uhaystack_len, "", bi, &status);
164 	STRPOS_CHECK_STATUS(status, "Error creating search object");
165 
166 	if(f_ignore_case) {
167 		coll = usearch_getCollator(src);
168 		status = U_ZERO_ERROR;
169 		ucol_setAttribute(coll, UCOL_STRENGTH, UCOL_SECONDARY, &status);
170 		STRPOS_CHECK_STATUS(status, "Error setting collation strength");
171 		usearch_reset(src);
172 	}
173 
174 	if(offset != 0) {
175 		offset_pos = grapheme_get_haystack_offset(bi, offset);
176 		if(offset_pos == -1) {
177 			status = U_ILLEGAL_ARGUMENT_ERROR;
178 			STRPOS_CHECK_STATUS(status, "Invalid search offset");
179 		}
180 		status = U_ZERO_ERROR;
181 		usearch_setOffset(src, offset_pos, &status);
182 		STRPOS_CHECK_STATUS(status, "Invalid search offset");
183 	}
184 
185 
186 	if(last) {
187 		char_pos = usearch_last(src, &status);
188 		if(char_pos < offset_pos) {
189 			/* last one is beyound our start offset */
190 			char_pos = USEARCH_DONE;
191 		}
192 	} else {
193 		char_pos = usearch_next(src, &status);
194 	}
195 	STRPOS_CHECK_STATUS(status, "Error looking up string");
196 	if(char_pos != USEARCH_DONE && ubrk_isBoundary(bi, char_pos)) {
197 		ret_pos = grapheme_count_graphemes(bi, uhaystack,char_pos);
198 		if(puchar_pos) {
199 			*puchar_pos = char_pos;
200 		}
201 	} else {
202 		ret_pos = -1;
203 	}
204 
205 	if (uhaystack) {
206 		efree( uhaystack );
207 	}
208 	if (uneedle) {
209 		efree( uneedle );
210 	}
211 	ubrk_close (bi);
212 	usearch_close (src);
213 
214 	return ret_pos;
215 }
216 
217 /* }}} */
218 
219 /* {{{ grapheme_ascii_check: ASCII check */
grapheme_ascii_check(const unsigned char * day,int32_t len)220 int grapheme_ascii_check(const unsigned char *day, int32_t len)
221 {
222 	int ret_len = len;
223 	while ( len-- ) {
224 	if ( *day++ > 0x7f )
225 		return -1;
226 	}
227 
228 	return ret_len;
229 }
230 
231 /* }}} */
232 
233 /* {{{ grapheme_split_string: find and optionally return grapheme boundaries */
grapheme_split_string(const UChar * text,int32_t text_length,int boundary_array[],int boundary_array_len TSRMLS_DC)234 int grapheme_split_string(const UChar *text, int32_t text_length, int boundary_array[], int boundary_array_len TSRMLS_DC )
235 {
236 	unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
237 	UErrorCode		status = U_ZERO_ERROR;
238 	int ret_len, pos;
239 	UBreakIterator* bi;
240 
241 	bi = grapheme_get_break_iterator((void*)u_break_iterator_buffer, &status TSRMLS_CC );
242 
243 	if( U_FAILURE(status) ) {
244 		return -1;
245 	}
246 
247 	ubrk_setText(bi, text, text_length,	&status);
248 
249 	pos = 0;
250 
251 	for ( ret_len = 0; pos != UBRK_DONE; ) {
252 
253 		pos = ubrk_next(bi);
254 
255 		if ( pos != UBRK_DONE ) {
256 
257 			if ( NULL != boundary_array && ret_len < boundary_array_len ) {
258 				boundary_array[ret_len] = pos;
259 			}
260 
261 			ret_len++;
262 		}
263 	}
264 
265 	ubrk_close(bi);
266 
267 	return ret_len;
268 }
269 /* }}} */
270 
271 /* {{{ grapheme_count_graphemes */
grapheme_count_graphemes(UBreakIterator * bi,UChar * string,int32_t string_len)272 int32_t grapheme_count_graphemes(UBreakIterator *bi, UChar *string, int32_t string_len)
273 {
274 	int ret_len = 0;
275 	int pos = 0;
276 	UErrorCode		status = U_ZERO_ERROR;
277 
278 	ubrk_setText(bi, string, string_len, &status);
279 
280 	do {
281 
282 		pos = ubrk_next(bi);
283 
284 		if ( UBRK_DONE != pos ) {
285 			ret_len++;
286 		}
287 
288 	} while ( UBRK_DONE != pos );
289 
290 	return ret_len;
291 }
292 /* }}} */
293 
294 
295 /* {{{ 	grapheme_get_haystack_offset - bump the haystack pointer based on the grapheme count offset */
grapheme_get_haystack_offset(UBreakIterator * bi,int32_t offset)296 int grapheme_get_haystack_offset(UBreakIterator* bi, int32_t offset)
297 {
298 	int32_t pos;
299 	int32_t (*iter_op)(UBreakIterator* bi);
300 	int iter_incr;
301 
302 	if ( 0 == offset ) {
303 		return 0;
304 	}
305 
306 	if ( offset < 0 ) {
307 		iter_op = ubrk_previous;
308 		ubrk_last(bi); /* one past the end */
309 		iter_incr = 1;
310 	}
311 	else {
312 		iter_op = ubrk_next;
313 		iter_incr = -1;
314 	}
315 
316 	pos = 0;
317 
318 	while ( pos != UBRK_DONE && offset != 0 ) {
319 
320 		pos = iter_op(bi);
321 
322 		if ( UBRK_DONE != pos ) {
323 			offset += iter_incr;
324 		}
325 	}
326 
327 	if ( offset != 0 ) {
328 		return -1;
329 	}
330 
331 	return pos;
332 }
333 /* }}} */
334 
335 /* {{{ grapheme_strrpos_ascii: borrowed from the php ext/standard/string.c */
336  int32_t
grapheme_strrpos_ascii(unsigned char * haystack,int32_t haystack_len,unsigned char * needle,int32_t needle_len,int32_t offset)337 grapheme_strrpos_ascii(unsigned char *haystack, int32_t haystack_len, unsigned char *needle, int32_t needle_len, int32_t offset)
338 {
339 	unsigned char *p, *e;
340 
341 	if (offset >= 0) {
342 		p = haystack + offset;
343 		e = haystack + haystack_len - needle_len;
344 	} else {
345 		p = haystack;
346 		if (needle_len > -offset) {
347 			e = haystack + haystack_len - needle_len;
348 		} else {
349 			e = haystack + haystack_len + offset;
350 		}
351 	}
352 
353 	if (needle_len == 1) {
354 		/* Single character search can shortcut memcmps */
355 		while (e >= p) {
356 			if (*e == *needle) {
357 				return (e - p + (offset > 0 ? offset : 0));
358 			}
359 			e--;
360 		}
361 		return -1;
362 	}
363 
364 	while (e >= p) {
365 		if (memcmp(e, needle, needle_len) == 0) {
366 			return (e - p + (offset > 0 ? offset : 0));
367 		}
368 		e--;
369 	}
370 
371 	return -1;
372 }
373 
374 /* }}} */
375 
376 /* {{{ grapheme_get_break_iterator: get a clone of the global character break iterator */
grapheme_get_break_iterator(void * stack_buffer,UErrorCode * status TSRMLS_DC)377 UBreakIterator* grapheme_get_break_iterator(void *stack_buffer, UErrorCode *status TSRMLS_DC )
378 {
379 	int32_t buffer_size;
380 
381 	UBreakIterator *global_break_iterator = INTL_G( grapheme_iterator );
382 
383 	if ( NULL == global_break_iterator ) {
384 
385 		global_break_iterator = ubrk_open(UBRK_CHARACTER,
386 											NULL,	/* icu default locale - locale has no effect on this iterator */
387 											NULL,	/* text not set in global iterator */
388 											0,		/* text length = 0 */
389 											status);
390 
391 		INTL_G(grapheme_iterator) = global_break_iterator;
392 	}
393 
394 	buffer_size = U_BRK_SAFECLONE_BUFFERSIZE;
395 
396 	return ubrk_safeClone(global_break_iterator, stack_buffer, &buffer_size, status);
397 }
398 /* }}} */
399 
400 /*
401  * Local variables:
402  * tab-width: 4
403  * c-basic-offset: 4
404  * End:
405  * vim600: fdm=marker
406  * vim: noet sw=4 ts=4
407  */
408 
409