andrei Fri Sep 22 17:47:10 2006 UTC
Modified files:
/php-src unicode-progress.txt
/php-src/ext/standard string.c
Log:
Unicode support in count_chars(). Note that due to the size of the
Unicode character set only mode=1 is supported.
http://cvs.php.net/viewvc.cgi/php-src/unicode-progress.txt?r1=1.47&r2=1.48&diff_format=u
Index: php-src/unicode-progress.txt
diff -u php-src/unicode-progress.txt:1.47 php-src/unicode-progress.txt:1.48
--- php-src/unicode-progress.txt:1.47 Wed Sep 20 20:30:19 2006
+++ php-src/unicode-progress.txt Fri Sep 22 17:47:09 2006
@@ -16,10 +16,6 @@
string.c
--------
- count_chars()
- Params API. Do we really want to go through the whole Unicode table?
- May need to use hashtable instead of array.
-
hebrev(), hebrevc()
Figure out if this is something we can use ICU for, internally.
Check with Zeev.
@@ -122,7 +118,6 @@
max()
range()
shuffle()
- strrchr()
end(), prev(), next(), reset(), current(), key()
@@ -150,6 +145,7 @@
bin2hex()
chr()
chunk_split()
+ count_chars()
dirname()
explode()
implode()
http://cvs.php.net/viewvc.cgi/php-src/ext/standard/string.c?r1=1.588&r2=1.589&diff_format=u
Index: php-src/ext/standard/string.c
diff -u php-src/ext/standard/string.c:1.588 php-src/ext/standard/string.c:1.589
--- php-src/ext/standard/string.c:1.588 Wed Sep 20 20:30:19 2006
+++ php-src/ext/standard/string.c Fri Sep 22 17:47:09 2006
@@ -18,7 +18,7 @@
+----------------------------------------------------------------------+
*/
-/* $Id: string.c,v 1.588 2006/09/20 20:30:19 andrei Exp $ */
+/* $Id: string.c,v 1.589 2006/09/22 17:47:09 andrei Exp $ */
/* Synced with php 3.0 revision 1.193 1999-06-16 [ssb] */
@@ -6295,79 +6295,115 @@
}
/* }}} */
-/* {{{ proto mixed count_chars(string input [, int mode])
+/* {{{ proto mixed count_chars(string input [, int mode]) U
Returns info about what characters are used in input */
PHP_FUNCTION(count_chars)
{
- zval **input, **mode;
+ zstr input;
+ int input_len;
+ zend_uchar type;
+ long mode = 0;
int chars[256];
- int ac=ZEND_NUM_ARGS();
- int mymode=0;
+ HashTable uchars;
+ UChar32 cp;
+ int *uchar_cnt_ptr, uchar_cnt;
unsigned char *buf;
- int len, inx;
+ int inx;
char retstr[256];
int retlen=0;
- if (ac < 1 || ac > 2 || zend_get_parameters_ex(ac, &input, &mode) ==
FAILURE) {
- WRONG_PARAM_COUNT;
+ if (zend_parse_parameters(ZEND_NUM_ARGS(), "t|l", &input, &input_len,
+ &type, &mode) ==
FAILURE) {
+ return;
}
- convert_to_string_ex(input);
-
- if (ac == 2) {
- convert_to_long_ex(mode);
- mymode = Z_LVAL_PP(mode);
-
- if (mymode < 0 || mymode > 4) {
+ if (ZEND_NUM_ARGS() > 1) {
+ if (mode < 0 || mode > 4) {
php_error_docref(NULL TSRMLS_CC, E_WARNING, "Unknown
mode.");
RETURN_FALSE;
}
+ if (UG(unicode) && mode != 1) {
+ php_error_docref(NULL TSRMLS_DC, E_WARNING, "Only
mode=1 is supported with Unicode strings");
+ }
}
- len = Z_STRLEN_PP(input);
- buf = (unsigned char *) Z_STRVAL_PP(input);
- memset((void*) chars, 0, sizeof(chars));
+ if (type == IS_UNICODE) {
+ UChar buf[3];
+ int buf_len;
- while (len > 0) {
- chars[*buf]++;
- buf++;
- len--;
- }
+ zend_hash_init(&uchars, 0, NULL, NULL, 0);
- if (mymode < 3) {
- array_init(return_value);
- }
+ inx = 0;
+ while (inx < input_len) {
+ U16_NEXT_UNSAFE(input.u, inx, cp);
+ if (zend_hash_index_find(&uchars, cp,
(void**)&uchar_cnt_ptr) == SUCCESS) {
+ (*uchar_cnt_ptr)++;
+ } else {
+ uchar_cnt = 1;
+ zend_hash_index_update(&uchars, cp, &uchar_cnt,
sizeof(int), NULL);
+ }
+ }
- for (inx = 0; inx < 256; inx++) {
- switch (mymode) {
- case 0:
- add_index_long(return_value, inx, chars[inx]);
- break;
- case 1:
- if (chars[inx] != 0) {
- add_index_long(return_value, inx,
chars[inx]);
- }
- break;
- case 2:
- if (chars[inx] == 0) {
- add_index_long(return_value, inx,
chars[inx]);
- }
- break;
- case 3:
- if (chars[inx] != 0) {
- retstr[retlen++] = inx;
- }
- break;
- case 4:
- if (chars[inx] == 0) {
- retstr[retlen++] = inx;
- }
- break;
+ if (mode < 3) {
+ array_init(return_value);
}
- }
- if (mymode >= 3 && mymode <= 4) {
- RETURN_STRINGL(retstr, retlen, 1);
+ for (zend_hash_internal_pointer_reset(&uchars);
+ zend_hash_get_current_data(&uchars,
(void**)&uchar_cnt_ptr) == SUCCESS;
+ zend_hash_move_forward(&uchars)) {
+
+ zend_hash_get_current_key(&uchars, NULL, (ulong*)&cp,
0);
+
+ buf_len = zend_codepoint_to_uchar(cp, buf);
+ buf[buf_len] = 0;
+ add_u_assoc_long_ex(return_value, IS_UNICODE,
ZSTR(buf), buf_len+1, *uchar_cnt_ptr);
+ }
+
+ zend_hash_destroy(&uchars);
+ } else {
+ buf = (unsigned char *) input.s;
+ memset((void*) chars, 0, sizeof(chars));
+
+ while (input_len > 0) {
+ chars[*buf]++;
+ buf++;
+ input_len--;
+ }
+
+ if (mode < 3) {
+ array_init(return_value);
+ }
+
+ for (inx = 0; inx < 256; inx++) {
+ switch (mode) {
+ case 0:
+ add_index_long(return_value, inx,
chars[inx]);
+ break;
+ case 1:
+ if (chars[inx] != 0) {
+ add_index_long(return_value,
inx, chars[inx]);
+ }
+ break;
+ case 2:
+ if (chars[inx] == 0) {
+ add_index_long(return_value,
inx, chars[inx]);
+ }
+ break;
+ case 3:
+ if (chars[inx] != 0) {
+ retstr[retlen++] = inx;
+ }
+ break;
+ case 4:
+ if (chars[inx] == 0) {
+ retstr[retlen++] = inx;
+ }
+ break;
+ }
+ }
+ if (mode >= 3 && mode <= 4) {
+ RETURN_STRINGL(retstr, retlen, 1);
+ }
}
}
/* }}} */
--
PHP CVS Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
