Submitting patch for Unicode impl of strtok()
, since I don't have a
proper test setup, as yet.
-- Rolland
--- Rolland Santimano rollandsantimano@yahoo.com wrote:
Submitting patch for Unicode impl of
strtok()
, since I don't have a
proper test setup, as yet.-- Rolland
OK, attachment got stripped - please check inline patch.
Index: ext/standard/basic_functions.h
RCS file: /repository/php-src/ext/standard/basic_functions.h,v
retrieving revision 1.140
diff -u -r1.140 basic_functions.h
--- ext/standard/basic_functions.h 13 Aug 2005 02:23:29 -0000 1.140
+++ ext/standard/basic_functions.h 9 Sep 2005 18:13:25 -0000
@@ -153,9 +153,9 @@
HashTable *user_shutdown_function_names;
HashTable putenv_ht;
zval *strtok_zval;
-
char *strtok_string;
-
void *strtok_string; char *locale_string;
-
char *strtok_last;
-
void *strtok_last; char strtok_table[256]; ulong strtok_len; char str_ebuf[40];
Index: ext/standard/string.c
RCS file: /repository/php-src/ext/standard/string.c,v
retrieving revision 1.481
diff -u -r1.481 string.c
--- ext/standard/string.c 8 Sep 2005 14:07:40 -0000 1.481
+++ ext/standard/string.c 9 Sep 2005 18:13:54 -0000
@@ -1315,88 +1315,168 @@
Tokenize a string */
PHP_FUNCTION(strtok)
{
-
zval **args[2];
-
zval **tok, **str;
-
char *token;
-
char *token_end;
-
char *p;
-
char *pe;
-
void *tok, *str;
-
int32_t tok_len, str_len;
-
zend_uchar tok_type, str_type;
-
zval *zv;
-
char *token, *token_end, *p, *pe;
-
UChar *u_token, *u_p, *u_pe;
-
UChar32 ch, th;
-
int32_t start, end, i, j, rem_len;
-
int delim_found, token_present; int skipped = 0;
-
if (ZEND_NUM_ARGS() < 1 || ZEND_NUM_ARGS() > 2 ||
zend_get_parameters_array_ex(ZEND_NUM_ARGS(), args) == FAILURE) {
-
if (ZEND_NUM_ARGS() < 1 || ZEND_NUM_ARGS() > 2) { WRONG_PARAM_COUNT; }
-
if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "T|T",
-
&str, &str_len, &str_type,
-
&tok, &tok_len, &tok_type) == FAILURE) {
-
return;
-
}
-
switch (ZEND_NUM_ARGS()) { case 1:
-
tok = args[0];
-
tok = str;
-
tok_len = str_len;
-
tok_type = str_type; break; default: case 2:
-
str = args[0];
-
tok = args[1];
-
convert_to_string_ex(str);
-
zval_add_ref(str); if (BG(strtok_zval)) { zval_ptr_dtor(&BG(strtok_zval)); }
-
BG(strtok_zval) = *str;
-
BG(strtok_last) = BG(strtok_string) = Z_STRVAL_PP(str);
-
BG(strtok_len) = Z_STRLEN_PP(str);
-
MAKE_STD_ZVAL(zv);
-
if (str_type == IS_UNICODE) {
-
ZVAL_UNICODEL(zv, (UChar *)str, str_len, 1);
-
} else if (str_type == IS_BINARY) {
-
ZVAL_BINARYL(zv, (char *)str, str_len, 1);
-
} else {
-
ZVAL_STRINGL(zv, (char *)str, str_len, 1);
-
}
-
BG(strtok_zval) = zv;
-
if (str_type == IS_UNICODE) {
-
BG(strtok_last) = BG(strtok_string) = Z_USTRVAL_P(zv);
-
} else {
-
BG(strtok_last) = BG(strtok_string) = Z_STRVAL_P(zv);
-
}
-
BG(strtok_len) = str_len; break; }
-
p = BG(strtok_last); /* Where we start to search */
-
pe = BG(strtok_string) + BG(strtok_len);
-
if (!p || p >= pe) {
-
if (BG(strtok_zval) && tok_type != Z_TYPE_P(BG(strtok_zval))) {
-
php_error_docref(NULL TSRMLS_CC, E_WARNING, "Delimiter type must match string
type.");
RETURN_FALSE;
}
-
convert_to_string_ex(tok);
-
token = Z_STRVAL_PP(tok);
-
token_end = token + Z_STRLEN_PP(tok);
-
if (tok_type == IS_UNICODE) {
-
u_p = (UChar *)BG(strtok_last); /* Where we start to search */
-
u_pe = (UChar *)BG(strtok_string) + BG(strtok_len);
-
u_token = (UChar *)tok;
-
if (!u_p || u_p >= u_pe) {
-
RETURN_FALSE;
-
}
-
rem_len = u_pe - u_p;
-
while (token < token_end) {
-
STRTOK_TABLE(token++) = 1;
-
}
-
/* Skip leading delimiters */
-
while (STRTOK_TABLE(p)) {
-
if (++p >= pe) {
-
/* no other chars left */
-
/* Skip leading delimiters */
-
token_present = 0;
-
for (i = 0 ; (u_p + i) < u_pe ; ) {
-
delim_found = 0;
-
U16_NEXT(u_p, i, rem_len, ch);
-
for (j = 0 ; j < tok_len ; ) {
-
U16_NEXT(u_token, j, tok_len, th);
-
if ( ch == th ) {
-
delim_found = 1;
-
break;
-
}
-
}
-
if (delim_found == 0) {
-
U16_BACK_1(u_p, 0, i); /* U16_NEXT() post-incrs 'i' */
-
start = i;
-
token_present = 1;
-
break;
-
}
-
}
-
if (token_present == 0) { BG(strtok_last) = NULL;
-
RETVAL_FALSE;
-
goto restore;
-
RETURN_FALSE; }
-
skipped++;
-
}
-
/* We know at this place that *p is no delimiter, so skip it */
-
while (++p < pe) {
-
if (STRTOK_TABLE(p)) {
-
goto return_token;
-
/* Seek to next delimiter */
-
delim_found = 0;
-
for (i = start ; (u_p + i) < u_pe ; ) {
-
U16_NEXT(u_p, i, rem_len, ch);
-
for (j = 0 ; j < tok_len ; ) {
-
U16_NEXT(u_token, j, tok_len, th);
-
if ( ch == th ) {
-
delim_found = 1;
-
break;
-
}
-
}
-
if (delim_found) {
-
U16_BACK_1(u_p, 0, i); /* 'i' was beyond delimiter */
-
break;
-
}
-
}
-
end = i;
-
if (end - start) {
-
BG(strtok_last) = u_p + end;
-
RETURN_UNICODEL(u_p + start, end - start, 1);
-
} else {
-
BG(strtok_last) = NULL;
-
RETURN_FALSE; }
-
}
-
if (p - BG(strtok_last)) {
-return_token:
-
RETVAL_STRINGL(BG(strtok_last) + skipped, (p - BG(strtok_last)) - skipped, 1);
-
BG(strtok_last) = p + 1; } else {
-
RETVAL_FALSE;
-
BG(strtok_last) = NULL;
-
}
-
p = (char *)BG(strtok_last); /* Where we start to search */
-
pe = (char *)BG(strtok_string) + BG(strtok_len);
-
if (!p || p >= pe) {
-
RETURN_FALSE;
-
}
-
token = (char *)tok;
-
token_end = token + tok_len;
-
while (token < token_end) {
-
STRTOK_TABLE(token++) = 1;
-
}
-
/* Restore table -- usually faster then memset'ing the table on every invocation */
-
/* Skip leading delimiters */
-
while (STRTOK_TABLE(p)) {
-
if (++p >= pe) {
-
/* no other chars left */
-
BG(strtok_last) = NULL;
-
RETVAL_FALSE;
-
goto restore;
-
}
-
skipped++;
-
}
-
/* We know at this place that *p is no delimiter, so skip it */
-
while (++p < pe) {
-
if (STRTOK_TABLE(p)) {
-
goto return_token;
-
}
-
}
-
if (p - (char *)BG(strtok_last)) {
+return_token:
-
if (tok_type == IS_BINARY) {
-
RETVAL_BINARYL((char *)BG(strtok_last) + skipped, (p - (char
*)BG(strtok_last)) - skipped, 1);
-
} else {
-
RETVAL_STRINGL((char *)BG(strtok_last) + skipped, (p - (char
*)BG(strtok_last)) - skipped, 1);
-
}
-
BG(strtok_last) = p + 1;
-
} else {
-
RETVAL_FALSE;
-
BG(strtok_last) = NULL;
-
}
-
/* Restore table -- usually faster then memset'ing the table on every invocation
*/
restore:
-
token = Z_STRVAL_PP(tok);
-
while (token < token_end) {
-
STRTOK_TABLE(token++) = 0;
-
token = (char *)tok;
-
while (token < token_end) {
-
STRTOK_TABLE(token++) = 0;
-
} }
}
/* }}} */