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

Source
/*
   +----------------------------------------------------------------------+
   | Copyright © The PHP Group and Contributors.                          |
   +----------------------------------------------------------------------+
   | This source file is subject to the Modified BSD License that is      |
   | bundled with this package in the file LICENSE, and is available      |
   | through the World Wide Web at <https://www.php.net/license/>.        |
   |                                                                      |
   | SPDX-License-Identifier: BSD-3-Clause                                |
   +----------------------------------------------------------------------+
   | Authors: Rasmus Lerdorf <rasmus@php.net>                             |
   |          Stig Sæther Bakken <ssb@php.net>                          |
   |          Zeev Suraski <zeev@php.net>                                 |
   +----------------------------------------------------------------------+
 */

#include <stdio.h>
#include "php.h"
#include "php_string.h"
#include "php_variables.h"
#include <locale.h>
#ifdef HAVE_NL_LANGINFO
# include <langinfo.h>
#endif
#ifdef HAVE_LANGINFO_H
# include <langinfo.h>
#endif

#include "scanf.h"
#include "zend_API.h"
#include "zend_execute.h"
#include "basic_functions.h"
#include "zend_smart_str.h"
#include <Zend/zend_exceptions.h>
#ifdef ZTS
#include "TSRM.h"
#endif

/* For str_getcsv() support */
#include "ext/standard/file.h"
/* For php_next_utf8_char() */
#include "ext/standard/html.h"
#include "ext/random/php_random.h"

#ifdef __SSE2__
#include "Zend/zend_bitset.h"
#endif

#include "zend_simd.h"

/* localeconv mutex */
#ifdef ZTS
static MUTEX_T locale_mutex = NULL;
#endif

/* {{{ php_hex2bin */
static zend_string *php_hex2bin(const unsigned char *old, const size_t oldlen)
{
  size_t target_length = oldlen >> 1;
  zend_string *str = zend_string_alloc(target_length, 0);
  unsigned char *ret = (unsigned char *)ZSTR_VAL(str);
  size_t i, j;

  for (i = j = 0; i < target_length; i++) {
    unsigned char c = old[j++];
    unsigned char l = c & ~0x20;
    int is_letter = ((unsigned int) ((l - 'A') ^ (l - 'F' - 1))) >> (8 * sizeof(unsigned int) - 1);
    unsigned char d;

    /* basically (c >= '0' && c <= '9') || (l >= 'A' && l <= 'F') */
    if (EXPECTED((((c ^ '0') - 10) >> (8 * sizeof(unsigned int) - 1)) | is_letter)) {
      d = (l - 0x10 - 0x27 * is_letter) << 4;
    } else {
      zend_string_efree(str);
      return NULL;
    }
    c = old[j++];
    l = c & ~0x20;
    is_letter = ((unsigned int) ((l - 'A') ^ (l - 'F' - 1))) >> (8 * sizeof(unsigned int) - 1);
    if (EXPECTED((((c ^ '0') - 10) >> (8 * sizeof(unsigned int) - 1)) | is_letter)) {
      d |= l - 0x10 - 0x27 * is_letter;
    } else {
      zend_string_efree(str);
      return NULL;
    }
    ret[i] = d;
  }
  ret[i] = '\0';

  return str;
}
/* }}} */

#ifdef ZTS
/* read the decimal point through nl_langinfo() (thread-safe), instead of taking the lock. */
PHPAPI char localeconv_decimal_point(void)
{
#if defined(HAVE_NL_LANGINFO) && (defined(__GLIBC__) || defined(__MUSL__))
  return *nl_langinfo(RADIXCHAR);
#else
  struct lconv lc;
  localeconv_r(&lc);
  return *lc.decimal_point;
#endif
}
#endif

/* {{{ localeconv_r
 * glibc's localeconv is not reentrant, so lets make it so ... sorta */
PHPAPI struct lconv *localeconv_r(struct lconv *out)
{

#ifdef ZTS
  tsrm_mutex_lock( locale_mutex );
#endif

  /* localeconv doesn't return an error condition */
  *out = *localeconv();

#ifdef ZTS
  tsrm_mutex_unlock( locale_mutex );
#endif

  return out;
}
/* }}} */

#ifdef ZTS
/* {{{ PHP_MINIT_FUNCTION */
PHP_MINIT_FUNCTION(localeconv)
{
  locale_mutex = tsrm_mutex_alloc();
  return SUCCESS;
}
/* }}} */

/* {{{ PHP_MSHUTDOWN_FUNCTION */
PHP_MSHUTDOWN_FUNCTION(localeconv)
{
  tsrm_mutex_free( locale_mutex );
  locale_mutex = NULL;
  return SUCCESS;
}
/* }}} */
#endif

/* {{{ Converts the binary representation of data to hex */
PHP_FUNCTION(bin2hex)
{
  zend_string *result;
  zend_string *data;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(data)
  ZEND_PARSE_PARAMETERS_END();

  result = zend_bin2hex_str((unsigned char *) ZSTR_VAL(data), ZSTR_LEN(data));

  RETURN_STR(result);
}
/* }}} */

/* {{{ Converts the hex representation of data to binary */
PHP_FUNCTION(hex2bin)
{
  zend_string *result, *data;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(data)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(data) % 2 != 0) {
    php_error_docref(NULL, E_WARNING, "Hexadecimal input string must have an even length");
    RETURN_FALSE;
  }

  result = php_hex2bin((unsigned char *)ZSTR_VAL(data), ZSTR_LEN(data));

  if (!result) {
    php_error_docref(NULL, E_WARNING, "Input string must be hexadecimal string");
    RETURN_FALSE;
  }

  RETVAL_STR(result);
}
/* }}} */

static void php_spn_common_handler(INTERNAL_FUNCTION_PARAMETERS, bool is_strspn) /* {{{ */
{
  zend_string *s11, *s22;
  zend_long start = 0, len = 0;
  bool len_is_null = 1;

  ZEND_PARSE_PARAMETERS_START(2, 4)
    Z_PARAM_STR(s11)
    Z_PARAM_STR(s22)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(start)
    Z_PARAM_LONG_OR_NULL(len, len_is_null)
  ZEND_PARSE_PARAMETERS_END();

  size_t remain_len = ZSTR_LEN(s11);
  if (start < 0) {
    start += remain_len;
    if (start < 0) {
      start = 0;
    }
  } else if ((size_t) start > remain_len) {
    start = remain_len;
  }

  remain_len -= start;
  if (!len_is_null) {
    if (len < 0) {
      len += remain_len;
      if (len < 0) {
        len = 0;
      }
    } else if ((size_t) len > remain_len) {
      len = remain_len;
    }
  } else {
    len = remain_len;
  }

  if (len == 0) {
    RETURN_LONG(0);
  }

  if (is_strspn) {
    RETURN_LONG(php_strspn(ZSTR_VAL(s11) + start /*str1_start*/,
            ZSTR_VAL(s22) /*str2_start*/,
            ZSTR_VAL(s11) + start + len /*str1_end*/,
            ZSTR_VAL(s22) + ZSTR_LEN(s22) /*str2_end*/));
  } else {
    RETURN_LONG(php_strcspn(ZSTR_VAL(s11) + start /*str1_start*/,
            ZSTR_VAL(s22) /*str2_start*/,
            ZSTR_VAL(s11) + start + len /*str1_end*/,
            ZSTR_VAL(s22) + ZSTR_LEN(s22) /*str2_end*/));
  }
}
/* }}} */

/* {{{ Finds length of initial segment consisting entirely of characters found in mask. If start or/and length is provided works like strspn(substr($s,$start,$len),$good_chars) */
PHP_FUNCTION(strspn)
{
  php_spn_common_handler(INTERNAL_FUNCTION_PARAM_PASSTHRU, /* is_strspn */ true);
}
/* }}} */

/* {{{ Finds length of initial segment consisting entirely of characters not found in mask. If start or/and length is provide works like strcspn(substr($s,$start,$len),$bad_chars) */
PHP_FUNCTION(strcspn)
{
  php_spn_common_handler(INTERNAL_FUNCTION_PARAM_PASSTHRU, /* is_strspn */ false);
}
/* }}} */

#ifdef HAVE_NL_LANGINFO
/* {{{ Query language and locale information */
PHP_FUNCTION(nl_langinfo)
{
  zend_long item;
  char *value;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_LONG(item)
  ZEND_PARSE_PARAMETERS_END();

  switch(item) { /* {{{ */
#ifdef ABDAY_1
    case ABDAY_1:
    case ABDAY_2:
    case ABDAY_3:
    case ABDAY_4:
    case ABDAY_5:
    case ABDAY_6:
    case ABDAY_7:
#endif
#ifdef DAY_1
    case DAY_1:
    case DAY_2:
    case DAY_3:
    case DAY_4:
    case DAY_5:
    case DAY_6:
    case DAY_7:
#endif
#ifdef ABMON_1
    case ABMON_1:
    case ABMON_2:
    case ABMON_3:
    case ABMON_4:
    case ABMON_5:
    case ABMON_6:
    case ABMON_7:
    case ABMON_8:
    case ABMON_9:
    case ABMON_10:
    case ABMON_11:
    case ABMON_12:
#endif
#ifdef MON_1
    case MON_1:
    case MON_2:
    case MON_3:
    case MON_4:
    case MON_5:
    case MON_6:
    case MON_7:
    case MON_8:
    case MON_9:
    case MON_10:
    case MON_11:
    case MON_12:
#endif
#ifdef AM_STR
    case AM_STR:
#endif
#ifdef PM_STR
    case PM_STR:
#endif
#ifdef D_T_FMT
    case D_T_FMT:
#endif
#ifdef D_FMT
    case D_FMT:
#endif
#ifdef T_FMT
    case T_FMT:
#endif
#ifdef T_FMT_AMPM
    case T_FMT_AMPM:
#endif
#ifdef ERA
    case ERA:
#endif
#ifdef ERA_YEAR
    case ERA_YEAR:
#endif
#ifdef ERA_D_T_FMT
    case ERA_D_T_FMT:
#endif
#ifdef ERA_D_FMT
    case ERA_D_FMT:
#endif
#ifdef ERA_T_FMT
    case ERA_T_FMT:
#endif
#ifdef ALT_DIGITS
    case ALT_DIGITS:
#endif
#ifdef INT_CURR_SYMBOL
    case INT_CURR_SYMBOL:
#endif
#ifdef CURRENCY_SYMBOL
    case CURRENCY_SYMBOL:
#endif
#ifdef CRNCYSTR
    case CRNCYSTR:
#endif
#ifdef MON_DECIMAL_POINT
    case MON_DECIMAL_POINT:
#endif
#ifdef MON_THOUSANDS_SEP
    case MON_THOUSANDS_SEP:
#endif
#ifdef MON_GROUPING
    case MON_GROUPING:
#endif
#ifdef POSITIVE_SIGN
    case POSITIVE_SIGN:
#endif
#ifdef NEGATIVE_SIGN
    case NEGATIVE_SIGN:
#endif
#ifdef INT_FRAC_DIGITS
    case INT_FRAC_DIGITS:
#endif
#ifdef FRAC_DIGITS
    case FRAC_DIGITS:
#endif
#ifdef P_CS_PRECEDES
    case P_CS_PRECEDES:
#endif
#ifdef P_SEP_BY_SPACE
    case P_SEP_BY_SPACE:
#endif
#ifdef N_CS_PRECEDES
    case N_CS_PRECEDES:
#endif
#ifdef N_SEP_BY_SPACE
    case N_SEP_BY_SPACE:
#endif
#ifdef P_SIGN_POSN
    case P_SIGN_POSN:
#endif
#ifdef N_SIGN_POSN
    case N_SIGN_POSN:
#endif
#ifdef DECIMAL_POINT
    case DECIMAL_POINT:
#elif defined(RADIXCHAR)
    case RADIXCHAR:
#endif
#ifdef THOUSANDS_SEP
    case THOUSANDS_SEP:
#elif defined(THOUSEP)
    case THOUSEP:
#endif
#ifdef GROUPING
    case GROUPING:
#endif
#ifdef YESEXPR
    case YESEXPR:
#endif
#ifdef NOEXPR
    case NOEXPR:
#endif
#ifdef YESSTR
    case YESSTR:
#endif
#ifdef NOSTR
    case NOSTR:
#endif
#ifdef CODESET
    case CODESET:
#endif
      break;
    default:
      php_error_docref(NULL, E_WARNING, "Item '" ZEND_LONG_FMT "' is not valid", item);
      RETURN_FALSE;
  }
  /* }}} */

  value = nl_langinfo(item);
  if (value == NULL) {
    RETURN_FALSE;
  } else {
    RETURN_STRING(value);
  }
}
#endif
/* }}} */

/* {{{ Compares two strings using the current locale */
PHP_FUNCTION(strcoll)
{
  zend_string *s1, *s2;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(s1)
    Z_PARAM_STR(s2)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_LONG(strcoll((const char *) ZSTR_VAL(s1),
                      (const char *) ZSTR_VAL(s2)));
}
/* }}} */

/* {{{ php_charmask
 * Fills a 256-byte bytemask with input. You can specify a range like 'a..z',
 * it needs to be incrementing.
 * Returns: FAILURE/SUCCESS whether the input was correct (i.e. no range errors)
 */
static inline zend_result php_charmask(const unsigned char *input, size_t len, char *mask)
{
  const unsigned char *end;
  unsigned char c;
  zend_result result = SUCCESS;

  memset(mask, 0, 256);
  for (end = input+len; input < end; input++) {
    c=*input;
    if ((input+3 < end) && input[1] == '.' && input[2] == '.'
        && input[3] >= c) {
      memset(mask+c, 1, input[3] - c + 1);
      input+=3;
    } else if ((input+1 < end) && input[0] == '.' && input[1] == '.') {
      /* Error, try to be as helpful as possible:
         (a range ending/starting with '.' won't be captured here) */
      if (end-len >= input) { /* there was no 'left' char */
        php_error_docref(NULL, E_WARNING, "Invalid '..'-range, no character to the left of '..'");
        result = FAILURE;
        continue;
      }
      if (input+2 >= end) { /* there is no 'right' char */
        php_error_docref(NULL, E_WARNING, "Invalid '..'-range, no character to the right of '..'");
        result = FAILURE;
        continue;
      }
      if (input[-1] > input[2]) { /* wrong order */
        php_error_docref(NULL, E_WARNING, "Invalid '..'-range, '..'-range needs to be incrementing");
        result = FAILURE;
        continue;
      }
      /* FIXME: better error (a..b..c is the only left possibility?) */
      php_error_docref(NULL, E_WARNING, "Invalid '..'-range");
      result = FAILURE;
      continue;
    } else {
      mask[c]=1;
    }
  }
  return result;
}
/* }}} */

static zend_always_inline bool php_is_whitespace(unsigned char c)
{
  return c <= ' ' && (c == ' ' || c == '\f' || c == '\n' || c == '\r' || c == '\t' || c == '\v' || c == '\0');
}

/* {{{ php_trim_int()
 * mode 1 : trim left
 * mode 2 : trim right
 * mode 3 : trim left and right
 * what indicates which chars are to be trimmed. NULL->default (' \f\t\n\r\v\0')
 */
static zend_always_inline zend_string *php_trim_int(zend_string *str, const char *what, size_t what_len, int mode)
{
  const char *start = ZSTR_VAL(str);
  const char *end = start + ZSTR_LEN(str);
  char mask[256];

  if (what) {
    if (what_len == 1) {
      char p = *what;
      if (mode & 1) {
        while (start != end) {
          if (*start == p) {
            start++;
          } else {
            break;
          }
        }
      }
      if (mode & 2) {
        while (start != end) {
          if (*(end-1) == p) {
            end--;
          } else {
            break;
          }
        }
      }
    } else {
      php_charmask((const unsigned char *) what, what_len, mask);

      if (mode & 1) {
        while (start != end) {
          if (mask[(unsigned char)*start]) {
            start++;
          } else {
            break;
          }
        }
      }
      if (mode & 2) {
        while (start != end) {
          if (mask[(unsigned char)*(end-1)]) {
            end--;
          } else {
            break;
          }
        }
      }
    }
  } else {
    if (mode & 1) {
      while (start != end) {
        if (php_is_whitespace((unsigned char)*start)) {
          start++;
        } else {
          break;
        }
      }
    }
    if (mode & 2) {
      while (start != end) {
        if (php_is_whitespace((unsigned char)*(end-1))) {
          end--;
        } else {
          break;
        }
      }
    }
  }

  if (ZSTR_LEN(str) == end - start) {
    return zend_string_copy(str);
  } else if (end - start == 0) {
    return ZSTR_EMPTY_ALLOC();
  } else {
    return zend_string_init(start, end - start, 0);
  }
}
/* }}} */

/* {{{ php_trim_int()
 * mode 1 : trim left
 * mode 2 : trim right
 * mode 3 : trim left and right
 * what indicates which chars are to be trimmed. NULL->default (' \f\t\n\r\v\0')
 */
PHPAPI zend_string *php_trim(zend_string *str, const char *what, size_t what_len, int mode)
{
  return php_trim_int(str, what, what_len, mode);
}
/* }}} */

/* {{{ php_do_trim
 * Base for trim(), rtrim() and ltrim() functions.
 */
static zend_always_inline void php_do_trim(INTERNAL_FUNCTION_PARAMETERS, int mode)
{
  zend_string *str;
  zend_string *what = NULL;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_STR(what)
  ZEND_PARSE_PARAMETERS_END();

  ZVAL_STR(return_value, php_trim_int(str, (what ? ZSTR_VAL(what) : NULL), (what ? ZSTR_LEN(what) : 0), mode));
}
/* }}} */

/* {{{ Strips whitespace from the beginning and end of a string */
PHP_FUNCTION(trim)
{
  php_do_trim(INTERNAL_FUNCTION_PARAM_PASSTHRU, 3);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(trim, 1)
{
  zval str_tmp;
  zend_string *str;

  Z_FLF_PARAM_STR(1, str, str_tmp);

  ZVAL_STR(return_value, php_trim_int(str, /* what */ NULL, /* what_len */ 0, /* mode */ 3));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

ZEND_FRAMELESS_FUNCTION(trim, 2)
{
  zval str_tmp, what_tmp;
  zend_string *str, *what;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_STR(2, what, what_tmp);

  ZVAL_STR(return_value, php_trim_int(str, ZSTR_VAL(what), ZSTR_LEN(what), /* mode */ 3));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
  Z_FLF_PARAM_FREE_STR(2, what_tmp);
}

/* {{{ Removes trailing whitespace */
PHP_FUNCTION(rtrim)
{
  php_do_trim(INTERNAL_FUNCTION_PARAM_PASSTHRU, 2);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(rtrim, 1)
{
  zval str_tmp;
  zend_string *str;

  Z_FLF_PARAM_STR(1, str, str_tmp);

  ZVAL_STR(return_value, php_trim_int(str, /* what */ NULL, /* what_len */ 0, /* mode */ 2));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

ZEND_FRAMELESS_FUNCTION(rtrim, 2)
{
  zval str_tmp, what_tmp;
  zend_string *str, *what;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_STR(2, what, what_tmp);

  ZVAL_STR(return_value, php_trim_int(str, ZSTR_VAL(what), ZSTR_LEN(what), /* mode */ 2));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
  Z_FLF_PARAM_FREE_STR(2, what_tmp);
}

/* {{{ Strips whitespace from the beginning of a string */
PHP_FUNCTION(ltrim)
{
  php_do_trim(INTERNAL_FUNCTION_PARAM_PASSTHRU, 1);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(ltrim, 1)
{
  zval str_tmp;
  zend_string *str;

  Z_FLF_PARAM_STR(1, str, str_tmp);

  ZVAL_STR(return_value, php_trim_int(str, /* what */ NULL, /* what_len */ 0, /* mode */ 1));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

ZEND_FRAMELESS_FUNCTION(ltrim, 2)
{
  zval str_tmp, what_tmp;
  zend_string *str, *what;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_STR(2, what, what_tmp);

  ZVAL_STR(return_value, php_trim_int(str, ZSTR_VAL(what), ZSTR_LEN(what), /* mode */ 1));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
  Z_FLF_PARAM_FREE_STR(2, what_tmp);
}

/* {{{ Wraps buffer to selected number of characters using string break char */
PHP_FUNCTION(wordwrap)
{
  zend_string *text;
  char *breakchar = "\n";
  size_t newtextlen, chk, breakchar_len = 1;
  size_t alloced;
  zend_long current = 0, laststart = 0, lastspace = 0;
  zend_long linelength = 75;
  bool docut = 0;
  zend_string *newtext;

  ZEND_PARSE_PARAMETERS_START(1, 4)
    Z_PARAM_STR(text)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(linelength)
    Z_PARAM_STRING(breakchar, breakchar_len)
    Z_PARAM_BOOL(docut)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(text) == 0) {
    RETURN_EMPTY_STRING();
  }

  if (breakchar_len == 0) {
    zend_argument_must_not_be_empty_error(3);
    RETURN_THROWS();
  }

  if (linelength == 0 && docut) {
    zend_argument_value_error(4, "cannot be true when argument #2 ($width) is 0");
    RETURN_THROWS();
  }

  /* Special case for a single-character break as it needs no
     additional storage space */
  if (breakchar_len == 1 && !docut) {
    newtext = zend_string_finit(ZSTR_VAL(text), ZSTR_LEN(text), 0, /* fallible */ true);
    if (UNEXPECTED(!newtext)) {
      RETURN_OOM();
    }

    laststart = lastspace = 0;
    for (current = 0; current < (zend_long)ZSTR_LEN(text); current++) {
      if (ZSTR_VAL(text)[current] == breakchar[0]) {
        laststart = lastspace = current + 1;
      } else if (ZSTR_VAL(text)[current] == ' ') {
        if (current - laststart >= linelength) {
          ZSTR_VAL(newtext)[current] = breakchar[0];
          laststart = current + 1;
        }
        lastspace = current;
      } else if (current - laststart >= linelength && laststart != lastspace) {
        ZSTR_VAL(newtext)[lastspace] = breakchar[0];
        laststart = lastspace + 1;
      }
    }

    RETURN_NEW_STR(newtext);
  } else {
    /* Multiple character line break or forced cut */
    if (linelength > 0) {
      chk = (size_t)(ZSTR_LEN(text)/linelength + 1);
      newtext = zend_string_safe_falloc(chk, breakchar_len, ZSTR_LEN(text), 0, /* fallible */ true);
      if (UNEXPECTED(!newtext)) {
        RETURN_OOM();
      }
      alloced = ZSTR_LEN(text) + chk * breakchar_len + 1;
    } else {
      chk = ZSTR_LEN(text);
      alloced = ZSTR_LEN(text) * (breakchar_len + 1) + 1;
      newtext = zend_string_safe_falloc(ZSTR_LEN(text), breakchar_len + 1, 0, 0, /* fallible */ true);
      if (UNEXPECTED(!newtext)) {
        RETURN_OOM();
      }
    }

    /* now keep track of the actual new text length */
    newtextlen = 0;

    laststart = lastspace = 0;
    for (current = 0; current < (zend_long)ZSTR_LEN(text); current++) {
      if (chk == 0) {
        alloced += (size_t) (((ZSTR_LEN(text) - current + 1)/linelength + 1) * breakchar_len) + 1;
        newtext = zend_string_fextend(newtext, alloced, 0, /* fallible */ true);
        if (UNEXPECTED(!newtext)) {
          RETURN_OOM();
        }
        chk = (size_t) ((ZSTR_LEN(text) - current)/linelength) + 1;
      }
      /* when we hit an existing break, copy to new buffer, and
       * fix up laststart and lastspace */
      if (ZSTR_VAL(text)[current] == breakchar[0]
        && current + breakchar_len <= ZSTR_LEN(text)
        && !strncmp(ZSTR_VAL(text) + current, breakchar, breakchar_len)) {
        memcpy(ZSTR_VAL(newtext) + newtextlen, ZSTR_VAL(text) + laststart, current - laststart + breakchar_len);
        newtextlen += current - laststart + breakchar_len;
        current += breakchar_len - 1;
        laststart = lastspace = current + 1;
        chk--;
      }
      /* if it is a space, check if it is at the line boundary,
       * copy and insert a break, or just keep track of it */
      else if (ZSTR_VAL(text)[current] == ' ') {
        if (current - laststart >= linelength) {
          memcpy(ZSTR_VAL(newtext) + newtextlen, ZSTR_VAL(text) + laststart, current - laststart);
          newtextlen += current - laststart;
          memcpy(ZSTR_VAL(newtext) + newtextlen, breakchar, breakchar_len);
          newtextlen += breakchar_len;
          laststart = current + 1;
          chk--;
        }
        lastspace = current;
      }
      /* if we are cutting, and we've accumulated enough
       * characters, and we haven't see a space for this line,
       * copy and insert a break. */
      else if (current - laststart >= linelength
          && docut && laststart >= lastspace) {
        memcpy(ZSTR_VAL(newtext) + newtextlen, ZSTR_VAL(text) + laststart, current - laststart);
        newtextlen += current - laststart;
        memcpy(ZSTR_VAL(newtext) + newtextlen, breakchar, breakchar_len);
        newtextlen += breakchar_len;
        laststart = lastspace = current;
        chk--;
      }
      /* if the current word puts us over the linelength, copy
       * back up until the last space, insert a break, and move
       * up the laststart */
      else if (current - laststart >= linelength
          && laststart < lastspace) {
        memcpy(ZSTR_VAL(newtext) + newtextlen, ZSTR_VAL(text) + laststart, lastspace - laststart);
        newtextlen += lastspace - laststart;
        memcpy(ZSTR_VAL(newtext) + newtextlen, breakchar, breakchar_len);
        newtextlen += breakchar_len;
        laststart = lastspace = lastspace + 1;
        chk--;
      }
    }

    /* copy over any stragglers */
    if (laststart != current) {
      memcpy(ZSTR_VAL(newtext) + newtextlen, ZSTR_VAL(text) + laststart, current - laststart);
      newtextlen += current - laststart;
    }

    ZSTR_VAL(newtext)[newtextlen] = '\0';
    /* free unused memory */
    newtext = zend_string_ftruncate(newtext, newtextlen, 0, /* fallible */ true);
    if (UNEXPECTED(!newtext)) {
      RETURN_OOM();
    }

    RETURN_NEW_STR(newtext);
  }
}
/* }}} */

/* {{{ php_explode */
PHPAPI void php_explode(const zend_string *delim, zend_string *str, HashTable *parts, zend_long limit)
{
  const char *p1 = ZSTR_VAL(str);
  const char *endp = ZSTR_VAL(str) + ZSTR_LEN(str);
  const char *p2 = php_memnstr(ZSTR_VAL(str), ZSTR_VAL(delim), ZSTR_LEN(delim), endp);
  zval  tmp;

  if (p2 == NULL) {
    ZVAL_STR_COPY(&tmp, str);
    zend_hash_next_index_insert_new(parts, &tmp);
  } else {
    zend_hash_real_init_packed(parts);
    ZEND_HASH_FILL_PACKED(parts) {
      do {
        /* Each piece costs a zval plus a zend_string, so stop once over the limit. */
        if (ZEND_HAS_OOM()) {
          ZEND_HASH_FILL_FINISH();
          return;
        }
        ZEND_HASH_FILL_GROW();
        ZEND_HASH_FILL_SET_STR(zend_string_init_fast(p1, p2 - p1));
        ZEND_HASH_FILL_NEXT();
        p1 = p2 + ZSTR_LEN(delim);
        p2 = php_memnstr(p1, ZSTR_VAL(delim), ZSTR_LEN(delim), endp);
      } while (p2 != NULL && --limit > 1);

      if (p1 <= endp) {
        ZEND_HASH_FILL_GROW();
        ZEND_HASH_FILL_SET_STR(zend_string_init_fast(p1, endp - p1));
        ZEND_HASH_FILL_NEXT();
      }
    } ZEND_HASH_FILL_END();
  }
}
/* }}} */

/* {{{ php_explode_negative_limit */
static void php_explode_negative_limit(const zend_string *delim, zend_string *str, HashTable *parts, zend_long limit)
{
#define EXPLODE_ALLOC_STEP 64
  const char *p1 = ZSTR_VAL(str);
  const char *endp = ZSTR_VAL(str) + ZSTR_LEN(str);
  const char *p2 = php_memnstr(ZSTR_VAL(str), ZSTR_VAL(delim), ZSTR_LEN(delim), endp);
  zval  tmp;

  if (p2 == NULL) {
    /*
    do nothing since limit <= -1, thus if only one chunk - 1 + (limit) <= 0
    by doing nothing we return empty array
    */
  } else {
    size_t allocated = EXPLODE_ALLOC_STEP, found = 0;
    zend_long i, to_return;
    const char **positions = emalloc(allocated * sizeof(char *));

    positions[found++] = p1;
    do {
      /* One pointer per delimiter, then a zval and a zend_string per piece. */
      if (ZEND_HAS_OOM()) {
        efree((void *)positions);
        return;
      }
      if (found >= allocated) {
        allocated = found + EXPLODE_ALLOC_STEP;/* make sure we have enough memory */
        positions = erealloc(ZEND_VOIDP(positions), allocated*sizeof(char *));
      }
      positions[found++] = p1 = p2 + ZSTR_LEN(delim);
      p2 = php_memnstr(p1, ZSTR_VAL(delim), ZSTR_LEN(delim), endp);
    } while (p2 != NULL);

    to_return = limit + found;
    /* limit is at least -1 therefore no need of bounds checking : i will be always less than found */
    for (i = 0; i < to_return; i++) { /* this checks also for to_return > 0 */
      if (ZEND_HAS_OOM()) {
        break;
      }
      ZVAL_STRINGL_FAST(&tmp, positions[i], (positions[i+1] - ZSTR_LEN(delim)) - positions[i]);
      zend_hash_next_index_insert_new(parts, &tmp);
    }
    efree((void *)positions);
  }
#undef EXPLODE_ALLOC_STEP
}
/* }}} */

/* {{{ Splits a string on string separator and return array of components. If limit is positive only limit number of components is returned. If limit is negative all components except the last abs(limit) are returned. */
PHP_FUNCTION(explode)
{
  zend_string *str, *delim;
  zend_long limit = ZEND_LONG_MAX; /* No limit */
  zval tmp;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(delim)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(limit)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(delim) == 0) {
    zend_argument_value_error(1, "must not be empty, use str_split() to split a string into characters");
    RETURN_THROWS();
  }

  array_init(return_value);

  if (ZSTR_LEN(str) == 0) {
    if (limit >= 0) {
      ZVAL_EMPTY_STRING(&tmp);
      zend_hash_index_add_new(Z_ARRVAL_P(return_value), 0, &tmp);
    }
    return;
  }

  if (limit > 1) {
    php_explode(delim, str, Z_ARRVAL_P(return_value), limit);
  } else if (limit < 0) {
    php_explode_negative_limit(delim, str, Z_ARRVAL_P(return_value), limit);
  } else {
    ZVAL_STR_COPY(&tmp, str);
    zend_hash_index_add_new(Z_ARRVAL_P(return_value), 0, &tmp);
  }
}
/* }}} */

/* {{{ php_implode */
PHPAPI void php_implode(const zend_string *glue, HashTable *pieces, zval *return_value)
{
  zval         *tmp;
  uint32_t      numelems;
  zend_string  *str;
  char         *cptr;
  size_t        len = 0;
  struct {
    zend_string *str;
    zend_long    lval;
  } *strings, *ptr;
  ALLOCA_FLAG(use_heap)

  numelems = zend_hash_num_elements(pieces);

  if (numelems == 0) {
    RETURN_EMPTY_STRING();
  } else if (numelems == 1) {
    /* loop to search the first not undefined element... */
    ZEND_HASH_FOREACH_VAL(pieces, tmp) {
      RETURN_STR(zval_get_string(tmp));
    } ZEND_HASH_FOREACH_END();
  }

  ptr = strings = do_alloca((sizeof(*strings)) * numelems, use_heap);

  uint32_t flags = ZSTR_GET_COPYABLE_CONCAT_PROPERTIES(glue);

  /* Converting an element may call __toString(), which can destroy pieces. */
  GC_TRY_ADDREF(pieces);

  ZEND_HASH_FOREACH_VAL(pieces, tmp) {
    if (EXPECTED(Z_TYPE_P(tmp) == IS_STRING)) {
      ptr->str = Z_STR_P(tmp);
      len += ZSTR_LEN(ptr->str);
      ptr->lval = 0;
      flags &= ZSTR_GET_COPYABLE_CONCAT_PROPERTIES(ptr->str);
      ptr++;
    } else if (UNEXPECTED(Z_TYPE_P(tmp) == IS_LONG)) {
      zend_long val = Z_LVAL_P(tmp);

      ptr->str = NULL;
      ptr->lval = val;
      ptr++;
      if (val <= 0) {
        len++;
      }
      while (val) {
        val /= 10;
        len++;
      }
    } else {
      ptr->str = zval_get_string_func(tmp);
      len += ZSTR_LEN(ptr->str);
      ptr->lval = 1;
      flags &= ZSTR_GET_COPYABLE_CONCAT_PROPERTIES(ptr->str);
      ptr++;
    }
  } ZEND_HASH_FOREACH_END();

  /* numelems cannot be 0, we checked above */
  /* The glue is repeated once per element, so the result can dwarf the input. */
  str = zend_string_safe_falloc(numelems - 1, ZSTR_LEN(glue), len, 0, /* fallible */ true);
  if (UNEXPECTED(!str)) {
    while (ptr > strings) {
      ptr--;
      if (ptr->str && ptr->lval) {
        zend_string_release_ex(ptr->str, 0);
      }
    }
    free_alloca(strings, use_heap);
    RETURN_OOM();
  }
  GC_ADD_FLAGS(str, flags);
  cptr = ZSTR_VAL(str) + ZSTR_LEN(str);
  *cptr = 0;

  while (1) {
    ptr--;
    if (EXPECTED(ptr->str)) {
      cptr -= ZSTR_LEN(ptr->str);
      memcpy(cptr, ZSTR_VAL(ptr->str), ZSTR_LEN(ptr->str));
      if (ptr->lval) {
        zend_string_release_ex(ptr->str, 0);
      }
    } else {
      char *oldPtr = cptr;
      char oldVal = *cptr;
      cptr = zend_print_long_to_buf(cptr, ptr->lval);
      *oldPtr = oldVal;
    }

    if (ptr == strings) {
      break;
    }

    cptr -= ZSTR_LEN(glue);
    if (ZSTR_LEN(glue) == 1) {
      *cptr = ZSTR_VAL(glue)[0];
    } else {
      memcpy(cptr, ZSTR_VAL(glue), ZSTR_LEN(glue));
    }
  }

  free_alloca(strings, use_heap);
  GC_TRY_DTOR_NO_REF(pieces);
  RETURN_NEW_STR(str);
}
/* }}} */

/* {{{ Joins array elements placing glue string between items and return one string */
PHP_FUNCTION(implode)
{
  zend_string *arg1_str = NULL;
  HashTable *arg1_array = NULL;
  zend_array *pieces = NULL;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_ARRAY_HT_OR_STR(arg1_array, arg1_str)
    Z_PARAM_OPTIONAL
    Z_PARAM_ARRAY_HT_OR_NULL(pieces)
  ZEND_PARSE_PARAMETERS_END();

  if (pieces == NULL) {
    if (arg1_array == NULL) {
      zend_type_error(
        "%s(): If argument #1 ($separator) is of type string, "
        "argument #2 ($array) must be of type array, null given",
        get_active_function_name()
      );
      RETURN_THROWS();
    }

    arg1_str = ZSTR_EMPTY_ALLOC();
    pieces = arg1_array;
  } else {
    if (arg1_str == NULL) {
      zend_argument_type_error(1, "must be of type string, array given");
      RETURN_THROWS();
    }
  }

  php_implode(arg1_str, pieces, return_value);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(implode, 1)
{
  zval *pieces;

  /* Manual parsing for more accurate error message. */
  if (!zend_parse_arg_array(arg1, &pieces, /* null_check */ false, /* or_object */ false)) { \
    zend_type_error(
      "%s(): If argument #1 ($separator) is of type string, "
      "argument #2 ($array) must be of type array, null given",
      get_active_function_name()
    );
    goto flf_clean; \
  }

  zend_string *str = ZSTR_EMPTY_ALLOC();

  php_implode(str, Z_ARR_P(pieces), return_value);

flf_clean:;
}

ZEND_FRAMELESS_FUNCTION(implode, 2)
{
  zval str_tmp;
  zend_string *str;
  zval *pieces;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_ARRAY_OR_NULL(2, pieces);

  if (!pieces) {
    zend_type_error(
      "%s(): If argument #1 ($separator) is of type string, "
      "argument #2 ($array) must be of type array, null given",
      get_active_function_name()
    );
    goto flf_clean;
  }

  php_implode(str, Z_ARR_P(pieces), return_value);

flf_clean:;
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

#define STRTOK_TABLE(p) BG(strtok_table)[(unsigned char) *p]

/* {{{ Tokenize a string */
PHP_FUNCTION(strtok)
{
  zend_string *str, *tok = NULL;
  char *token;
  char *token_end;
  char *p;
  char *pe;
  size_t skipped = 0;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_STR_OR_NULL(tok)
  ZEND_PARSE_PARAMETERS_END();

  if (!tok) {
    tok = str;
  } else {
    if (BG(strtok_string)) {
      zend_string_release(BG(strtok_string));
    }
    BG(strtok_string) = zend_string_copy(str);
    BG(strtok_last) = ZSTR_VAL(str);
    BG(strtok_len) = ZSTR_LEN(str);
  }

  if (!BG(strtok_string)) {
    /* String to tokenize not set. */
    php_error_docref(NULL, E_WARNING, "Both arguments must be provided when starting tokenization");
    RETURN_FALSE;
  }

  p = BG(strtok_last); /* Where we start to search */
  pe = ZSTR_VAL(BG(strtok_string)) + BG(strtok_len);
  if (p >= pe) {
    /* Reached the end of the string. */
    RETURN_FALSE;
  }

  token = ZSTR_VAL(tok);
  token_end = token + ZSTR_LEN(tok);

  while (token < token_end) {
    STRTOK_TABLE(token++) = 1;
  }

  /* Skip leading delimiters */
  while (STRTOK_TABLE(p)) {
    if (++p >= pe) {
      /* no other chars left */
      goto 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;
    }
  }

  if (p - BG(strtok_last)) {
return_token:
    RETVAL_STRINGL(BG(strtok_last) + skipped, (p - BG(strtok_last)) - skipped);
    BG(strtok_last) = p + 1;
  } else {
return_false:
    RETVAL_FALSE;
    zend_string_release(BG(strtok_string));
    BG(strtok_string) = NULL;
  }

  /* Restore table -- usually faster then memset'ing the table on every invocation */
  token = ZSTR_VAL(tok);
  while (token < token_end) {
    STRTOK_TABLE(token++) = 0;
  }
}
/* }}} */

/* {{{ Makes a string uppercase */
PHP_FUNCTION(strtoupper)
{
  zend_string *arg;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(arg)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_STR(zend_string_toupper(arg));
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(strtoupper, 1)
{
  zval str_tmp;
  zend_string *str;

  Z_FLF_PARAM_STR(1, str, str_tmp);

  RETVAL_STR(zend_string_toupper(str));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

/* {{{ Makes a string lowercase */
PHP_FUNCTION(strtolower)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_STR(zend_string_tolower(str));
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(strtolower, 1)
{
  zval str_tmp;
  zend_string *str;

  Z_FLF_PARAM_STR(1, str, str_tmp);

  RETVAL_STR(zend_string_tolower(str));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

PHP_FUNCTION(str_increment)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(str) == 0) {
    zend_argument_must_not_be_empty_error(1);
    RETURN_THROWS();
  }
  if (!zend_string_only_has_ascii_alphanumeric(str)) {
    zend_argument_value_error(1, "must be composed only of alphanumeric ASCII characters");
    RETURN_THROWS();
  }

  zend_string *incremented = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), /* persistent */ false);
  size_t position = ZSTR_LEN(str)-1;
  bool carry = false;

  do {
    char c = ZSTR_VAL(incremented)[position];
    /* We know c is in ['a', 'z'], ['A', 'Z'], or ['0', '9'] range from zend_string_only_has_ascii_alphanumeric() */
    if (EXPECTED( c != 'z' && c != 'Z' && c != '9' )) {
      carry = false;
      ZSTR_VAL(incremented)[position]++;
    } else { /* if 'z', 'Z', or '9' */
      carry = true;
      if (c == '9') {
        ZSTR_VAL(incremented)[position] = '0';
      } else {
        ZSTR_VAL(incremented)[position] -= 25;
      }
    }
  } while (carry && position-- > 0);

  if (UNEXPECTED(carry)) {
    zend_string *tmp = zend_string_alloc(ZSTR_LEN(incremented)+1, 0);
    memcpy(ZSTR_VAL(tmp) + 1, ZSTR_VAL(incremented), ZSTR_LEN(incremented));
    ZSTR_VAL(tmp)[ZSTR_LEN(incremented)+1] = '\0';
    switch (ZSTR_VAL(incremented)[0]) {
      case '0':
        ZSTR_VAL(tmp)[0] = '1';
        break;
      default:
        ZSTR_VAL(tmp)[0] = ZSTR_VAL(incremented)[0];
        break;
    }
    zend_string_efree(incremented);
    RETURN_NEW_STR(tmp);
  }
  RETURN_NEW_STR(incremented);
}


PHP_FUNCTION(str_decrement)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(str) == 0) {
    zend_argument_must_not_be_empty_error(1);
    RETURN_THROWS();
  }
  if (!zend_string_only_has_ascii_alphanumeric(str)) {
    zend_argument_value_error(1, "must be composed only of alphanumeric ASCII characters");
    RETURN_THROWS();
  }
  if (ZSTR_LEN(str) >= 1 && ZSTR_VAL(str)[0] == '0') {
    zend_argument_value_error(1, "\"%s\" is out of decrement range", ZSTR_VAL(str));
    RETURN_THROWS();
  }

  zend_string *decremented = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), /* persistent */ false);
  size_t position = ZSTR_LEN(str)-1;
  bool carry = false;

  do {
    char c = ZSTR_VAL(decremented)[position];
    /* We know c is in ['a', 'z'], ['A', 'Z'], or ['0', '9'] range from zend_string_only_has_ascii_alphanumeric() */
    if (EXPECTED( c != 'a' && c != 'A' && c != '0' )) {
      carry = false;
      ZSTR_VAL(decremented)[position]--;
    } else { /* if 'a', 'A', or '0' */
      carry = true;
      if (c == '0') {
        ZSTR_VAL(decremented)[position] = '9';
      } else {
        ZSTR_VAL(decremented)[position] += 25;
      }
    }
  } while (carry && position-- > 0);

  if (UNEXPECTED(carry || (ZSTR_VAL(decremented)[0] == '0' && ZSTR_LEN(decremented) > 1))) {
    if (ZSTR_LEN(decremented) == 1) {
      zend_string_efree(decremented);
      zend_argument_value_error(1, "\"%s\" is out of decrement range", ZSTR_VAL(str));
      RETURN_THROWS();
    }
    zend_string *tmp = zend_string_alloc(ZSTR_LEN(decremented) - 1, 0);
    memcpy(ZSTR_VAL(tmp), ZSTR_VAL(decremented) + 1, ZSTR_LEN(decremented) - 1);
    ZSTR_VAL(tmp)[ZSTR_LEN(decremented) - 1] = '\0';
    zend_string_efree(decremented);
    RETURN_NEW_STR(tmp);
  }
  RETURN_NEW_STR(decremented);
}

#if defined(PHP_WIN32)
static bool _is_basename_start(const char *start, const char *pos)
{
  if (pos - start >= 1
      && *(pos-1) != '/'
      && *(pos-1) != '\\') {
    if (pos - start == 1) {
      return true;
    } else if (*(pos-2) == '/' || *(pos-2) == '\\') {
      return true;
    } else if (*(pos-2) == ':'
      && _is_basename_start(start, pos - 2)) {
      return true;
    }
  }
  return false;
}
#endif

/* {{{ php_basename */
PHPAPI zend_string *php_basename(const char *s, size_t len, const char *suffix, size_t suffix_len)
{
  const char *basename_start;
  const char *basename_end;

  if (CG(ascii_compatible_locale)) {
    basename_end = s + len - 1;

    /* Strip trailing slashes */
    while (basename_end >= s
#ifdef PHP_WIN32
      && (*basename_end == '/'
        || *basename_end == '\\'
        || (*basename_end == ':'
          && _is_basename_start(s, basename_end)))) {
#else
      && *basename_end == '/') {
#endif
      basename_end--;
    }
    if (basename_end < s) {
      return ZSTR_EMPTY_ALLOC();
    }

    /* Extract filename */
    basename_start = basename_end;
    basename_end++;
    while (basename_start > s
#ifdef PHP_WIN32
      && *(basename_start-1) != '/'
      && *(basename_start-1) != '\\') {

      if (*(basename_start-1) == ':' &&
        _is_basename_start(s, basename_start - 1)) {
        break;
      }
#else
      && *(basename_start-1) != '/') {
#endif
      basename_start--;
    }
  } else {
    /* State 0 is directly after a directory separator (or at the start of the string).
     * State 1 is everything else. */
    int state = 0;

    basename_start = s;
    basename_end = s;
    while (len > 0) {
      int inc_len = (*s == '\0' ? 1 : php_mblen(s, len));

      switch (inc_len) {
        case 0:
          goto quit_loop;
        case 1:
#ifdef PHP_WIN32
          if (*s == '/' || *s == '\\') {
#else
          if (*s == '/') {
#endif
            if (state == 1) {
              state = 0;
              basename_end = s;
            }
#ifdef PHP_WIN32
          /* Catch relative paths in c:file.txt style. They're not to confuse
             with the NTFS streams. This part ensures also, that no drive
             letter traversing happens. */
          } else if ((*s == ':' && (s - basename_start == 1))) {
            if (state == 0) {
              basename_start = s;
              state = 1;
            } else {
              basename_end = s;
              state = 0;
            }
#endif
          } else {
            if (state == 0) {
              basename_start = s;
              state = 1;
            }
          }
          break;
        default:
          if (inc_len < 0) {
            /* If character is invalid, treat it like other non-significant characters. */
            inc_len = 1;
            php_mb_reset();
          }
          if (state == 0) {
            basename_start = s;
            state = 1;
          }
          break;
      }
      s += inc_len;
      len -= inc_len;
    }

quit_loop:
    if (state == 1) {
      basename_end = s;
    }
  }

  if (suffix != NULL && suffix_len < (size_t)(basename_end - basename_start) &&
      memcmp(basename_end - suffix_len, suffix, suffix_len) == 0) {
    basename_end -= suffix_len;
  }

  return zend_string_init(basename_start, basename_end - basename_start, 0);
}
/* }}} */

/* {{{ Returns the filename component of the path */
PHP_FUNCTION(basename)
{
  char *string, *suffix = NULL;
  size_t   string_len, suffix_len = 0;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STRING(string, string_len)
    Z_PARAM_OPTIONAL
    Z_PARAM_STRING(suffix, suffix_len)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_STR(php_basename(string, string_len, suffix, suffix_len));
}
/* }}} */

/* {{{ php_dirname
   Returns directory name component of path */
PHPAPI size_t php_dirname(char *path, size_t len)
{
  return zend_dirname(path, len);
}
/* }}} */

static zend_always_inline void _zend_dirname(zval *return_value, zend_string *str, zend_long levels)
{
  zend_string *ret;

  ret = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), 0);

  if (levels == 1) {
    /* Default case */
#ifdef PHP_WIN32
    ZSTR_LEN(ret) = php_win32_ioutil_dirname(ZSTR_VAL(ret), ZSTR_LEN(str));
#else
    ZSTR_LEN(ret) = zend_dirname(ZSTR_VAL(ret), ZSTR_LEN(str));
#endif
  } else if (levels < 1) {
    zend_argument_value_error(2, "must be greater than or equal to 1");
    zend_string_efree(ret);
    RETURN_THROWS();
  } else {
    /* Some levels up */
    size_t str_len;
    do {
#ifdef PHP_WIN32
      ZSTR_LEN(ret) = php_win32_ioutil_dirname(ZSTR_VAL(ret), str_len = ZSTR_LEN(ret));
#else
      ZSTR_LEN(ret) = zend_dirname(ZSTR_VAL(ret), str_len = ZSTR_LEN(ret));
#endif
    } while (ZSTR_LEN(ret) < str_len && --levels);
  }

  RETURN_NEW_STR(ret);
}

/* {{{ Returns the directory name component of the path */
PHP_FUNCTION(dirname)
{
  zend_string *str;
  zend_long levels = 1;

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

  _zend_dirname(return_value, str, levels);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(dirname, 1)
{
  zval str_tmp;
  zend_string *str;

  Z_FLF_PARAM_STR(1, str, str_tmp);

  _zend_dirname(return_value, str, 1);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

ZEND_FRAMELESS_FUNCTION(dirname, 2)
{
  zval str_tmp;
  zend_string *str;
  zend_long levels;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_LONG(2, levels);

  _zend_dirname(return_value, str, levels);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

/* {{{ Returns information about a certain string */
PHP_FUNCTION(pathinfo)
{
  char *path;
  size_t path_len;
  zend_long opt = PHP_PATHINFO_ALL;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STRING(path, path_len)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(opt)
  ZEND_PARSE_PARAMETERS_END();

  if (opt < PHP_PATHINFO_DIRNAME || opt > PHP_PATHINFO_ALL) {
    zend_argument_value_error(2, "must be one of the PATHINFO_* constants");
    RETURN_THROWS();
  }

  if (opt < PHP_PATHINFO_ALL && (opt & (opt - 1))) {
    zend_argument_value_error(2, "must be only one of the PATHINFO_* constants");
    RETURN_THROWS();
  }

  if (opt == PHP_PATHINFO_DIRNAME) {
    char *dirname = estrndup(path, path_len);
    php_dirname(dirname, path_len);
    RETVAL_STRING_FAST(dirname);
    efree(dirname);
    return;
  }

  zend_string *basename = php_basename(path, path_len, NULL, 0);
  if (opt == PHP_PATHINFO_BASENAME) {
    RETURN_STR(basename);
  }

  const char *p = zend_memrchr(ZSTR_VAL(basename), '.', ZSTR_LEN(basename));
  size_t extension_len = p ? ZSTR_LEN(basename) - (p - ZSTR_VAL(basename)) - 1 : 0;
  if (opt == PHP_PATHINFO_EXTENSION) {
    if (p) {
      RETVAL_STRINGL(p + 1, extension_len);
    } else {
      RETVAL_EMPTY_STRING();
    }
    zend_string_release_ex(basename, 0);
    return;
  }

  zend_string *filename = p
    ? zend_string_init(ZSTR_VAL(basename), p - ZSTR_VAL(basename), 0)
    : zend_string_copy(basename);
  if (opt == PHP_PATHINFO_FILENAME) {
    zend_string_release_ex(basename, 0);
    RETURN_STR(filename);
  }

  ZEND_ASSERT(opt == PHP_PATHINFO_ALL);
  array_init(return_value);

  char *dirname = estrndup(path, path_len);
  php_dirname(dirname, path_len);
  if (*dirname) {
    add_assoc_string(return_value, "dirname", dirname);
  }
  efree(dirname);

  add_assoc_str(return_value, "basename", basename);
  if (p) {
    add_assoc_stringl(return_value, "extension", p + 1, extension_len);
  }
  add_assoc_str(return_value, "filename", filename);
}
/* }}} */

/* {{{ php_stristr
   case insensitive strstr */
PHPAPI char *php_stristr(const char *s, const char *t, size_t s_len, size_t t_len)
{
  return (char*)php_memnistr(s, t, t_len, s + s_len);
}
/* }}} */

static size_t php_strspn_strcspn_common(const char *haystack, const char *characters, const char *haystack_end, const char *characters_end, bool must_match)
{
  /* Fast path for short strings.
   * The table lookup cannot be faster in this case because we not only have to compare, but also build the table.
   * We only compare in this case.
   * Empirically tested that the table lookup approach is only beneficial if characters is longer than 1 character. */
  if (characters_end - characters == 1) {
    /* Avoid memchr() overhead for short spans and an immediate match. */
    if (!must_match && haystack_end - haystack >= 16 && *haystack != *characters) {
      const char *ptr = memchr(haystack, (unsigned char) *characters, haystack_end - haystack);
      return ptr ? (size_t) (ptr - haystack) : (size_t) (haystack_end - haystack);
    }
    const char *ptr = haystack;
    while (ptr < haystack_end && (*ptr == *characters) == must_match) {
      ptr++;
    }
    return ptr - haystack;
  }

  /* Every character in characters will set a boolean in this lookup table.
   * We'll use the lookup table as a fast lookup for the characters in characters while looping over haystack. */
  bool table[256];
  /* Use multiple small memsets to inline the memset with intrinsics, trick learned from glibc. */
  memset(table, 0, 64);
  memset(table + 64, 0, 64);
  memset(table + 128, 0, 64);
  memset(table + 192, 0, 64);

  while (characters < characters_end) {
    table[(unsigned char) *characters] = true;
    characters++;
  }

  const char *ptr = haystack;
  while (ptr < haystack_end && table[(unsigned char) *ptr] == must_match) {
    ptr++;
  }

  return ptr - haystack;
}

/* {{{ php_strspn */
PHPAPI size_t php_strspn(const char *haystack, const char *characters, const char *haystack_end, const char *characters_end)
{
  return php_strspn_strcspn_common(haystack, characters, haystack_end, characters_end, true);
}
/* }}} */

/* {{{ php_strcspn */
PHPAPI size_t php_strcspn(const char *haystack, const char *characters, const char *haystack_end, const char *characters_end)
{
  return php_strspn_strcspn_common(haystack, characters, haystack_end, characters_end, false);
}
/* }}} */

/* {{{ Finds first occurrence of a string within another, case insensitive */
PHP_FUNCTION(stristr)
{
  zend_string *haystack, *needle;
  const char *found = NULL;
  size_t  found_offset;
  bool part = 0;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_BOOL(part)
  ZEND_PARSE_PARAMETERS_END();

  found = php_stristr(ZSTR_VAL(haystack), ZSTR_VAL(needle), ZSTR_LEN(haystack), ZSTR_LEN(needle));

  if (UNEXPECTED(!found)) {
    RETURN_FALSE;
  }
  found_offset = found - ZSTR_VAL(haystack);
  if (part) {
    RETURN_STRINGL(ZSTR_VAL(haystack), found_offset);
  }
  RETURN_STRINGL(found, ZSTR_LEN(haystack) - found_offset);
}
/* }}} */

static zend_always_inline void _zend_strstr(zval *return_value, zend_string *haystack, zend_string *needle, bool part)
{
  const char *found = NULL;
  zend_long found_offset;

  found = php_memnstr(ZSTR_VAL(haystack), ZSTR_VAL(needle), ZSTR_LEN(needle), ZSTR_VAL(haystack) + ZSTR_LEN(haystack));

  if (UNEXPECTED(!found)) {
    RETURN_FALSE;
  }
  found_offset = found - ZSTR_VAL(haystack);
  if (part) {
    RETURN_STRINGL(ZSTR_VAL(haystack), found_offset);
  }
  RETURN_STRINGL(found, ZSTR_LEN(haystack) - found_offset);
}

/* {{{ Finds first occurrence of a string within another */
PHP_FUNCTION(strstr)
{
  zend_string *haystack, *needle;
  bool part = 0;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_BOOL(part)
  ZEND_PARSE_PARAMETERS_END();

  _zend_strstr(return_value, haystack, needle, part);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(strstr, 2)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);

  _zend_strstr(return_value, haystack, needle, /* part */ false);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

ZEND_FRAMELESS_FUNCTION(strstr, 3)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;
  bool part;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);
  Z_FLF_PARAM_BOOL(3, part);

  _zend_strstr(return_value, haystack, needle, part);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

/* {{{ Checks if a string contains another */
PHP_FUNCTION(str_contains)
{
  zend_string *haystack, *needle;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_BOOL(php_memnstr(ZSTR_VAL(haystack), ZSTR_VAL(needle), ZSTR_LEN(needle), ZSTR_VAL(haystack) + ZSTR_LEN(haystack)));
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(str_contains, 2)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);

  RETVAL_BOOL(php_memnstr(ZSTR_VAL(haystack), ZSTR_VAL(needle), ZSTR_LEN(needle), ZSTR_VAL(haystack) + ZSTR_LEN(haystack)));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

/* {{{ Checks if haystack starts with needle */
PHP_FUNCTION(str_starts_with)
{
  zend_string *haystack, *needle;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_BOOL(zend_string_starts_with(haystack, needle));
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(str_starts_with, 2)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);

  RETVAL_BOOL(zend_string_starts_with(haystack, needle));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

/* {{{ Checks if haystack ends with needle */
PHP_FUNCTION(str_ends_with)
{
  zend_string *haystack, *needle;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_BOOL(zend_string_ends_with(haystack, needle));
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(str_ends_with, 2)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);

  RETVAL_BOOL(zend_string_ends_with(haystack, needle));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

static zend_always_inline void _zend_strpos(zval *return_value, zend_string *haystack, zend_string *needle, zend_long offset)
{
  const char *found = NULL;

  if (offset < 0) {
    offset += (zend_long)ZSTR_LEN(haystack);
  }
  if (offset < 0 || (size_t)offset > ZSTR_LEN(haystack)) {
    zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
    RETURN_THROWS();
  }

  found = (char*)php_memnstr(ZSTR_VAL(haystack) + offset,
            ZSTR_VAL(needle), ZSTR_LEN(needle),
            ZSTR_VAL(haystack) + ZSTR_LEN(haystack));

  if (UNEXPECTED(!found)) {
    RETURN_FALSE;
  }
  RETURN_LONG(found - ZSTR_VAL(haystack));
}

/* {{{ Finds position of first occurrence of a string within another */
PHP_FUNCTION(strpos)
{
  zend_string *haystack, *needle;
  zend_long offset = 0;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(offset)
  ZEND_PARSE_PARAMETERS_END();

  _zend_strpos(return_value, haystack, needle, offset);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(strpos, 2)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);

  _zend_strpos(return_value, haystack, needle, 0);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

ZEND_FRAMELESS_FUNCTION(strpos, 3)
{
  zval haystack_tmp, needle_tmp;
  zend_string *haystack, *needle;
  zend_long offset;

  Z_FLF_PARAM_STR(1, haystack, haystack_tmp);
  Z_FLF_PARAM_STR(2, needle, needle_tmp);
  Z_FLF_PARAM_LONG(3, offset);

  _zend_strpos(return_value, haystack, needle, offset);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, haystack_tmp);
  Z_FLF_PARAM_FREE_STR(2, needle_tmp);
}

/* {{{ Finds position of first occurrence of a string within another, case insensitive */
PHP_FUNCTION(stripos)
{
  const char *found = NULL;
  zend_string *haystack, *needle;
  zend_long offset = 0;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(offset)
  ZEND_PARSE_PARAMETERS_END();

  if (offset < 0) {
    offset += (zend_long)ZSTR_LEN(haystack);
  }
  if (offset < 0 || (size_t)offset > ZSTR_LEN(haystack)) {
    zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
    RETURN_THROWS();
  }

  found = (char*)php_memnistr(ZSTR_VAL(haystack) + offset,
      ZSTR_VAL(needle), ZSTR_LEN(needle), ZSTR_VAL(haystack) + ZSTR_LEN(haystack));

  if (UNEXPECTED(!found)) {
    RETURN_FALSE;
  }
  RETURN_LONG(found - ZSTR_VAL(haystack));
}
/* }}} */

/* {{{ Finds position of last occurrence of a string within another string */
PHP_FUNCTION(strrpos)
{
  zend_string *needle;
  zend_string *haystack;
  zend_long offset = 0;
  const char *p, *e, *found;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(offset)
  ZEND_PARSE_PARAMETERS_END();

  if (offset >= 0) {
    if ((size_t)offset > ZSTR_LEN(haystack)) {
      zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
      RETURN_THROWS();
    }
    p = ZSTR_VAL(haystack) + (size_t)offset;
    e = ZSTR_VAL(haystack) + ZSTR_LEN(haystack);
  } else {
    if (offset < -ZEND_LONG_MAX || (size_t)(-offset) > ZSTR_LEN(haystack)) {
      zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
      RETURN_THROWS();
    }

    p = ZSTR_VAL(haystack);
    if ((size_t)-offset < ZSTR_LEN(needle)) {
      e = ZSTR_VAL(haystack) + ZSTR_LEN(haystack);
    } else {
      e = ZSTR_VAL(haystack) + ZSTR_LEN(haystack) + offset + ZSTR_LEN(needle);
    }
  }

  found = zend_memnrstr(p, ZSTR_VAL(needle), ZSTR_LEN(needle), e);

  if (UNEXPECTED(!found)) {
    RETURN_FALSE;
  }
  RETURN_LONG(found - ZSTR_VAL(haystack));
}
/* }}} */

/* {{{ Finds position of last occurrence of a string within another string */
PHP_FUNCTION(strripos)
{
  zend_string *needle;
  zend_string *haystack;
  zend_long offset = 0;
  const char *p, *e, *found;
  zend_string *needle_dup, *haystack_dup;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(offset)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(needle) == 1) {
    /* Single character search can shortcut memcmps
       Can also avoid tolower emallocs */
    char lowered;
    if (offset >= 0) {
      if ((size_t)offset > ZSTR_LEN(haystack)) {
        zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
        RETURN_THROWS();
      }
      p = ZSTR_VAL(haystack) + (size_t)offset;
      e = ZSTR_VAL(haystack) + ZSTR_LEN(haystack) - 1;
    } else {
      p = ZSTR_VAL(haystack);
      if (offset < -ZEND_LONG_MAX || (size_t)(-offset) > ZSTR_LEN(haystack)) {
        zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
        RETURN_THROWS();
      }
      e = ZSTR_VAL(haystack) + (ZSTR_LEN(haystack) + (size_t)offset);
    }
    lowered = zend_tolower_ascii(*ZSTR_VAL(needle));
    while (e >= p) {
      if (zend_tolower_ascii(*e) == lowered) {
        RETURN_LONG(e - p + (offset > 0 ? offset : 0));
      }
      e--;
    }
    RETURN_FALSE;
  }

  haystack_dup = zend_string_tolower(haystack);
  if (offset >= 0) {
    if ((size_t)offset > ZSTR_LEN(haystack)) {
      zend_string_release_ex(haystack_dup, 0);
      zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
      RETURN_THROWS();
    }
    p = ZSTR_VAL(haystack_dup) + offset;
    e = ZSTR_VAL(haystack_dup) + ZSTR_LEN(haystack);
  } else {
    if (offset < -ZEND_LONG_MAX || (size_t)(-offset) > ZSTR_LEN(haystack)) {
      zend_string_release_ex(haystack_dup, 0);
      zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
      RETURN_THROWS();
    }

    p = ZSTR_VAL(haystack_dup);
    if ((size_t)-offset < ZSTR_LEN(needle)) {
      e = ZSTR_VAL(haystack_dup) + ZSTR_LEN(haystack);
    } else {
      e = ZSTR_VAL(haystack_dup) + ZSTR_LEN(haystack) + offset + ZSTR_LEN(needle);
    }
  }

  needle_dup = zend_string_tolower(needle);
  if ((found = (char *)zend_memnrstr(p, ZSTR_VAL(needle_dup), ZSTR_LEN(needle_dup), e))) {
    RETVAL_LONG(found - ZSTR_VAL(haystack_dup));
  } else {
    RETVAL_FALSE;
  }
  zend_string_release_ex(needle_dup, false);
  zend_string_release_ex(haystack_dup, false);
}
/* }}} */

/* {{{ Finds the last occurrence of a character in a string within another */
PHP_FUNCTION(strrchr)
{
  zend_string *haystack, *needle;
  const char *found = NULL;
  zend_long found_offset;
  bool part = 0;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(needle)
    Z_PARAM_OPTIONAL
    Z_PARAM_BOOL(part)
  ZEND_PARSE_PARAMETERS_END();

  found = zend_memrchr(ZSTR_VAL(haystack), *ZSTR_VAL(needle), ZSTR_LEN(haystack));
  if (UNEXPECTED(!found)) {
    RETURN_FALSE;
  }
  found_offset = found - ZSTR_VAL(haystack);
  if (part) {
    RETURN_STRINGL(ZSTR_VAL(haystack), found_offset);
  }
  RETURN_STRINGL(found, ZSTR_LEN(haystack) - found_offset);
}
/* }}} */

/* {{{ php_chunk_split */
static zend_string *php_chunk_split(const char *src, size_t srclen, const char *end, size_t endlen, size_t chunklen)
{
  char *q;
  const char *p;
  size_t chunks;
  size_t restlen;
  zend_string *dest;

  chunks = srclen / chunklen;
  restlen = srclen - chunks * chunklen; /* srclen % chunklen */
  if (restlen) {
    /* We want chunks to be rounded up rather than rounded down.
     * Increment can't overflow because chunks <= SIZE_MAX/2 at this point. */
    chunks++;
  }

  dest = zend_string_safe_falloc(chunks, endlen, srclen, 0, /* fallible */ true);
  if (UNEXPECTED(!dest)) {
    return NULL;
  }

  for (p = src, q = ZSTR_VAL(dest); p < (src + srclen - chunklen + 1); ) {
    q = zend_mempcpy(q, p, chunklen);
    q = zend_mempcpy(q, end, endlen);
    p += chunklen;
  }

  if (restlen) {
    q = zend_mempcpy(q, p, restlen);
    q = zend_mempcpy(q, end, endlen);
  }

  *q = '\0';
  ZEND_ASSERT(q - ZSTR_VAL(dest) == ZSTR_LEN(dest));

  return dest;
}
/* }}} */

/* {{{ Returns split line */
PHP_FUNCTION(chunk_split)
{
  zend_string *str;
  char *end    = "\r\n";
  size_t endlen   = 2;
  zend_long chunklen = 76;
  zend_string *result;

  ZEND_PARSE_PARAMETERS_START(1, 3)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(chunklen)
    Z_PARAM_STRING(end, endlen)
  ZEND_PARSE_PARAMETERS_END();

  if (chunklen <= 0) {
    zend_argument_value_error(2, "must be greater than 0");
    RETURN_THROWS();
  }

  if ((size_t)chunklen > ZSTR_LEN(str)) {
    /* to maintain BC, we must return original string + ending */
    result = zend_string_safe_falloc(ZSTR_LEN(str), 1, endlen, 0, /* fallible */ true);
    if (UNEXPECTED(!result)) {
      RETURN_OOM();
    }
    memcpy(ZSTR_VAL(result), ZSTR_VAL(str), ZSTR_LEN(str));
    memcpy(ZSTR_VAL(result) + ZSTR_LEN(str), end, endlen);
    ZSTR_VAL(result)[ZSTR_LEN(result)] = '\0';
    RETURN_NEW_STR(result);
  }

  if (!ZSTR_LEN(str)) {
    RETURN_EMPTY_STRING();
  }

  result = php_chunk_split(ZSTR_VAL(str), ZSTR_LEN(str), end, endlen, (size_t)chunklen);
  if (UNEXPECTED(!result)) {
    RETURN_OOM();
  }

  RETURN_STR(result);
}
/* }}} */

static inline void _zend_substr(zval *return_value, zend_string *str, zend_long f, bool len_is_null, zend_long l)
{
  if (f < 0) {
    /* if "from" position is negative, count start position from the end
     * of the string
     */
    if (-(size_t)f > ZSTR_LEN(str)) {
      f = 0;
    } else {
      f = (zend_long)ZSTR_LEN(str) + f;
    }
  } else if ((size_t)f > ZSTR_LEN(str)) {
    RETURN_EMPTY_STRING();
  }

  if (!len_is_null) {
    if (l < 0) {
      /* if "length" position is negative, set it to the length
       * needed to stop that many chars from the end of the string
       */
      if (-(size_t)l > ZSTR_LEN(str) - (size_t)f) {
        l = 0;
      } else {
        l = (zend_long)ZSTR_LEN(str) - f + l;
      }
    } else if ((size_t)l > ZSTR_LEN(str) - (size_t)f) {
      l = (zend_long)ZSTR_LEN(str) - f;
    }
  } else {
    l = (zend_long)ZSTR_LEN(str) - f;
  }

  if (l == ZSTR_LEN(str)) {
    RETURN_STR_COPY(str);
  } else {
    RETURN_STRINGL_FAST(ZSTR_VAL(str) + f, l);
  }
}

/* {{{ Returns part of a string */
PHP_FUNCTION(substr)
{
  zend_string *str;
  zend_long l = 0, f;
  bool len_is_null = 1;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(str)
    Z_PARAM_LONG(f)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG_OR_NULL(l, len_is_null)
  ZEND_PARSE_PARAMETERS_END();

  _zend_substr(return_value, str, f, len_is_null, l);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(substr, 2)
{
  zval str_tmp;
  zend_string *str;
  zend_long f;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_LONG(2, f);

  _zend_substr(return_value, str, f, /* len_is_null */ true, 0);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

ZEND_FRAMELESS_FUNCTION(substr, 3)
{
  zval str_tmp;
  zend_string *str;
  zend_long f, l;
  bool len_is_null;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_LONG(2, f);
  Z_FLF_PARAM_LONG_OR_NULL(3, len_is_null, l);

  _zend_substr(return_value, str, f, len_is_null, l);

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

/* {{{ Replaces part of a string with another string */
PHP_FUNCTION(substr_replace)
{
  zend_string *str, *repl_str;
  HashTable *str_ht, *repl_ht;
  HashTable *from_ht;
  zend_long from_long;
  HashTable *len_ht = NULL;
  zend_long len_long;
  bool len_is_null = 1;
  zend_long l = 0;
  zend_long f;
  zend_string *result;
  HashPosition from_idx, repl_idx, len_idx;
  zval *tmp_str = NULL, *tmp_repl, *tmp_from = NULL, *tmp_len= NULL;

  ZEND_PARSE_PARAMETERS_START(3, 4)
    Z_PARAM_ARRAY_HT_OR_STR(str_ht, str)
    Z_PARAM_ARRAY_HT_OR_STR(repl_ht, repl_str)
    Z_PARAM_ARRAY_HT_OR_LONG(from_ht, from_long)
    Z_PARAM_OPTIONAL
    Z_PARAM_ARRAY_HT_OR_LONG_OR_NULL(len_ht, len_long, len_is_null)
  ZEND_PARSE_PARAMETERS_END();

  if (len_is_null) {
    if (str) {
      l = ZSTR_LEN(str);
    }
  } else if (!len_ht) {
    l = len_long;
  }

  if (str) {
    if (from_ht) {
      zend_argument_type_error(3, "cannot be an array when working on a single string");
      RETURN_THROWS();
    }
    if (len_ht) {
      zend_argument_type_error(4, "cannot be an array when working on a single string");
      RETURN_THROWS();
    }

    f = from_long;

    /* if "from" position is negative, count start position from the end
     * of the string
     */
    if (f < 0) {
      f = (zend_long)ZSTR_LEN(str) + f;
      if (f < 0) {
        f = 0;
      }
    } else if ((size_t)f > ZSTR_LEN(str)) {
      f = ZSTR_LEN(str);
    }
    /* if "length" position is negative, set it to the length
     * needed to stop that many chars from the end of the string
     */
    if (l < 0) {
      l = ((zend_long)ZSTR_LEN(str) - f) + l;
      if (l < 0) {
        l = 0;
      }
    }

    if ((size_t)l > ZSTR_LEN(str)) {
      l = ZSTR_LEN(str);
    }

    if ((f + l) > (zend_long)ZSTR_LEN(str)) {
      l = ZSTR_LEN(str) - f;
    }

    zend_string *tmp_repl_str = NULL;
    if (repl_ht) {
      repl_idx = 0;
      if (HT_IS_PACKED(repl_ht)) {
        while (repl_idx < repl_ht->nNumUsed) {
          tmp_repl = &repl_ht->arPacked[repl_idx];
          if (Z_TYPE_P(tmp_repl) != IS_UNDEF) {
            break;
          }
          repl_idx++;
        }
      } else {
        while (repl_idx < repl_ht->nNumUsed) {
          tmp_repl = &repl_ht->arData[repl_idx].val;
          if (Z_TYPE_P(tmp_repl) != IS_UNDEF) {
            break;
          }
          repl_idx++;
        }
      }
      if (repl_idx < repl_ht->nNumUsed) {
        repl_str = zval_get_tmp_string(tmp_repl, &tmp_repl_str);
      } else {
        repl_str = ZSTR_EMPTY_ALLOC();
      }
    }

    result = zend_string_safe_alloc(1, ZSTR_LEN(str) - l + ZSTR_LEN(repl_str), 0, 0);

    memcpy(ZSTR_VAL(result), ZSTR_VAL(str), f);
    if (ZSTR_LEN(repl_str)) {
      memcpy((ZSTR_VAL(result) + f), ZSTR_VAL(repl_str), ZSTR_LEN(repl_str));
    }
    memcpy((ZSTR_VAL(result) + f + ZSTR_LEN(repl_str)), ZSTR_VAL(str) + f + l, ZSTR_LEN(str) - f - l);
    ZSTR_VAL(result)[ZSTR_LEN(result)] = '\0';
    zend_tmp_string_release(tmp_repl_str);
    RETURN_NEW_STR(result);
  } else { /* str is array of strings */
    zend_string *str_index = NULL;
    size_t result_len;
    zend_ulong num_index;

    /* TODO
    if (!len_is_null && from_ht) {
      if (zend_hash_num_elements(from_ht) != zend_hash_num_elements(len_ht)) {
        php_error_docref(NULL, E_WARNING, "'start' and 'length' should have the same number of elements");
        RETURN_STR_COPY(str);
      }
    }
    */

    array_init(return_value);

    from_idx = len_idx = repl_idx = 0;

    ZEND_HASH_FOREACH_KEY_VAL(str_ht, num_index, str_index, tmp_str) {
      zend_string *tmp_orig_str;
      zend_string *orig_str;

      /* The replacement is repeated once per element, so stop once over the limit. */
      if (ZEND_HAS_OOM()) {
        RETURN_OOM();
      }
      orig_str = zval_get_tmp_string(tmp_str, &tmp_orig_str);

      if (from_ht) {
        if (HT_IS_PACKED(from_ht)) {
          while (from_idx < from_ht->nNumUsed) {
            tmp_from = &from_ht->arPacked[from_idx];
            if (Z_TYPE_P(tmp_from) != IS_UNDEF) {
              break;
            }
            from_idx++;
          }
        } else {
          while (from_idx < from_ht->nNumUsed) {
            tmp_from = &from_ht->arData[from_idx].val;
            if (Z_TYPE_P(tmp_from) != IS_UNDEF) {
              break;
            }
            from_idx++;
          }
        }
        if (from_idx < from_ht->nNumUsed) {
          f = zval_get_long(tmp_from);

          if (f < 0) {
            f = (zend_long)ZSTR_LEN(orig_str) + f;
            if (f < 0) {
              f = 0;
            }
          } else if (f > (zend_long)ZSTR_LEN(orig_str)) {
            f = ZSTR_LEN(orig_str);
          }
          from_idx++;
        } else {
          f = 0;
        }
      } else {
        f = from_long;
        if (f < 0) {
          f = (zend_long)ZSTR_LEN(orig_str) + f;
          if (f < 0) {
            f = 0;
          }
        } else if (f > (zend_long)ZSTR_LEN(orig_str)) {
          f = ZSTR_LEN(orig_str);
        }
      }

      if (len_ht) {
        if (HT_IS_PACKED(len_ht)) {
          while (len_idx < len_ht->nNumUsed) {
            tmp_len = &len_ht->arPacked[len_idx];
            if (Z_TYPE_P(tmp_len) != IS_UNDEF) {
              break;
            }
            len_idx++;
          }
        } else {
          while (len_idx < len_ht->nNumUsed) {
            tmp_len = &len_ht->arData[len_idx].val;
            if (Z_TYPE_P(tmp_len) != IS_UNDEF) {
              break;
            }
            len_idx++;
          }
        }
        if (len_idx < len_ht->nNumUsed) {
          l = zval_get_long(tmp_len);
          len_idx++;
        } else {
          l = ZSTR_LEN(orig_str);
        }
      } else if (!len_is_null) {
        l = len_long;
      } else {
        l = ZSTR_LEN(orig_str);
      }

      if (l < 0) {
        l = (ZSTR_LEN(orig_str) - f) + l;
        if (l < 0) {
          l = 0;
        }
      }

      ZEND_ASSERT(0 <= f && f <= ZEND_LONG_MAX);
      ZEND_ASSERT(0 <= l && l <= ZEND_LONG_MAX);
      if (((size_t) f + l) > ZSTR_LEN(orig_str)) {
        l = ZSTR_LEN(orig_str) - f;
      }

      result_len = ZSTR_LEN(orig_str) - l;

      if (repl_ht) {
        if (HT_IS_PACKED(repl_ht)) {
          while (repl_idx < repl_ht->nNumUsed) {
            tmp_repl = &repl_ht->arPacked[repl_idx];
            if (Z_TYPE_P(tmp_repl) != IS_UNDEF) {
              break;
            }
            repl_idx++;
          }
        } else {
          while (repl_idx < repl_ht->nNumUsed) {
            tmp_repl = &repl_ht->arData[repl_idx].val;
            if (Z_TYPE_P(tmp_repl) != IS_UNDEF) {
              break;
            }
            repl_idx++;
          }
        }
        if (repl_idx < repl_ht->nNumUsed) {
          zend_string *tmp_repl_str;
          zend_string *repl_str = zval_get_tmp_string(tmp_repl, &tmp_repl_str);

          result_len += ZSTR_LEN(repl_str);
          repl_idx++;
          result = zend_string_safe_alloc(1, result_len, 0, 0);

          memcpy(ZSTR_VAL(result), ZSTR_VAL(orig_str), f);
          memcpy((ZSTR_VAL(result) + f), ZSTR_VAL(repl_str), ZSTR_LEN(repl_str));
          memcpy((ZSTR_VAL(result) + f + ZSTR_LEN(repl_str)), ZSTR_VAL(orig_str) + f + l, ZSTR_LEN(orig_str) - f - l);
          zend_tmp_string_release(tmp_repl_str);
        } else {
          result = zend_string_safe_alloc(1, result_len, 0, 0);

          memcpy(ZSTR_VAL(result), ZSTR_VAL(orig_str), f);
          memcpy((ZSTR_VAL(result) + f), ZSTR_VAL(orig_str) + f + l, ZSTR_LEN(orig_str) - f - l);
        }
      } else {
        result_len += ZSTR_LEN(repl_str);

        result = zend_string_safe_alloc(1, result_len, 0, 0);

        memcpy(ZSTR_VAL(result), ZSTR_VAL(orig_str), f);
        memcpy((ZSTR_VAL(result) + f), ZSTR_VAL(repl_str), ZSTR_LEN(repl_str));
        memcpy((ZSTR_VAL(result) + f + ZSTR_LEN(repl_str)), ZSTR_VAL(orig_str) + f + l, ZSTR_LEN(orig_str) - f - l);
      }

      ZSTR_VAL(result)[ZSTR_LEN(result)] = '\0';

      if (str_index) {
        zval tmp;

        ZVAL_NEW_STR(&tmp, result);
        zend_symtable_update(Z_ARRVAL_P(return_value), str_index, &tmp);
      } else {
        add_index_str(return_value, num_index, result);
      }

      zend_tmp_string_release(tmp_orig_str);
    } ZEND_HASH_FOREACH_END();
  } /* if */
}
/* }}} */

/* {{{ Quotes meta characters */
PHP_FUNCTION(quotemeta)
{
  zend_string *old;
  const char *old_end, *p;
  char *q;
  char c;
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(old)
  ZEND_PARSE_PARAMETERS_END();

  old_end = ZSTR_VAL(old) + ZSTR_LEN(old);

  if (ZSTR_LEN(old) == 0) {
    RETURN_EMPTY_STRING();
  }

  str = zend_string_safe_alloc(2, ZSTR_LEN(old), 0, 0);

  for (p = ZSTR_VAL(old), q = ZSTR_VAL(str); p != old_end; p++) {
    c = *p;
    switch (c) {
      case '.':
      case '\\':
      case '+':
      case '*':
      case '?':
      case '[':
      case '^':
      case ']':
      case '$':
      case '(':
      case ')':
        *q++ = '\\';
        ZEND_FALLTHROUGH;
      default:
        *q++ = c;
    }
  }

  *q = '\0';

  RETURN_NEW_STR(zend_string_truncate(str, q - ZSTR_VAL(str), 0));
}
/* }}} */

/* {{{ Returns ASCII value of character
   Warning: This function is special-cased by zend_compile.c and so is bypassed for constant string argument */
PHP_FUNCTION(ord)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  if (UNEXPECTED(ZSTR_LEN(str) != 1)) {
    if (ZSTR_LEN(str) == 0) {
      php_error_docref(NULL, E_DEPRECATED,
        "Providing an empty string is deprecated");
    } else {
      php_error_docref(NULL, E_DEPRECATED,
        "Providing a string that is not one byte long is deprecated. Use ord($str[0]) instead");
    }
  }
  RETURN_LONG((unsigned char) ZSTR_VAL(str)[0]);
}
/* }}} */

/* {{{ Converts ASCII code to a character
   Warning: This function is special-cased by zend_compile.c and so is bypassed for constant integer argument */
PHP_FUNCTION(chr)
{
  zend_long c;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_LONG(c)
  ZEND_PARSE_PARAMETERS_END();

  if (UNEXPECTED(c < 0 || c > 255)) {
    php_error_docref(NULL, E_DEPRECATED,
      "Providing a value not in-between 0 and 255 is deprecated,"
      " this is because a byte value must be in the [0, 255] interval."
      " The value used will be constrained using %% 256");
  }
  c &= 0xff;
  RETURN_CHAR(c);
}
/* }}} */

/* {{{ php_ucfirst
   Uppercase the first character of the word in a native string */
static zend_string* php_ucfirst(zend_string *str)
{
  const unsigned char ch = ZSTR_VAL(str)[0];
  unsigned char r = zend_toupper_ascii(ch);
  if (r == ch) {
    return zend_string_copy(str);
  } else {
    zend_string *s = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), 0);
    ZSTR_VAL(s)[0] = r;
    return s;
  }
}
/* }}} */

/* {{{ Makes a string's first character uppercase */
PHP_FUNCTION(ucfirst)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  if (!ZSTR_LEN(str)) {
    RETURN_EMPTY_STRING();
  }

  RETURN_STR(php_ucfirst(str));
}
/* }}} */

/* {{{
   Lowercase the first character of the word in a native string */
static zend_string* php_lcfirst(zend_string *str)
{
  unsigned char r = zend_tolower_ascii(ZSTR_VAL(str)[0]);
  if (r == ZSTR_VAL(str)[0]) {
    return zend_string_copy(str);
  } else {
    zend_string *s = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), 0);
    ZSTR_VAL(s)[0] = r;
    return s;
  }
}
/* }}} */

/* {{{ Make a string's first character lowercase */
PHP_FUNCTION(lcfirst)
{
  zend_string  *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  if (!ZSTR_LEN(str)) {
    RETURN_EMPTY_STRING();
  }

  RETURN_STR(php_lcfirst(str));
}
/* }}} */

/* {{{ Uppercase the first character of every word in a string */
PHP_FUNCTION(ucwords)
{
  zend_string *str;
  char *delims = " \t\r\n\f\v";
  char *r;
  const char *r_end;
  size_t delims_len = 6;
  char mask[256];

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_STRING(delims, delims_len)
  ZEND_PARSE_PARAMETERS_END();

  if (!ZSTR_LEN(str)) {
    RETURN_EMPTY_STRING();
  }

  php_charmask((const unsigned char *) delims, delims_len, mask);

  ZVAL_STRINGL(return_value, ZSTR_VAL(str), ZSTR_LEN(str));
  r = Z_STRVAL_P(return_value);

  *r = zend_toupper_ascii((unsigned char) *r);
  for (r_end = r + Z_STRLEN_P(return_value) - 1; r < r_end; ) {
    if (mask[(unsigned char)*r++]) {
      *r = zend_toupper_ascii((unsigned char) *r);
    }
  }
}
/* }}} */

/* {{{ php_strtr */
PHPAPI char *php_strtr(char *str, size_t len, const char *str_from, const char *str_to, size_t trlen)
{
  size_t i;

  if (UNEXPECTED(trlen < 1)) {
    return str;
  } else if (trlen == 1) {
    char ch_from = *str_from;
    char ch_to = *str_to;

    for (i = 0; i < len; i++) {
      if (str[i] == ch_from) {
        str[i] = ch_to;
      }
    }
  } else {
    unsigned char xlat[256];

    memset(xlat, 0, sizeof(xlat));

    for (i = 0; i < trlen; i++) {
      xlat[(size_t)(unsigned char) str_from[i]] = str_to[i] - str_from[i];
    }

    for (i = 0; i < len; i++) {
      str[i] += xlat[(size_t)(unsigned char) str[i]];
    }
  }

  return str;
}
/* }}} */

/* {{{ php_strtr_ex */
static zend_string *php_strtr_ex(zend_string *str, const char *str_from, const char *str_to, size_t trlen)
{
  zend_string *new_str = NULL;
  size_t i;

  if (UNEXPECTED(trlen < 1)) {
    return zend_string_copy(str);
  } else if (trlen == 1) {
    char ch_from = *str_from;
    char ch_to = *str_to;
    char *output;
    char *input = ZSTR_VAL(str);
    size_t len = ZSTR_LEN(str);

#ifdef XSSE2
    if (ZSTR_LEN(str) >= sizeof(__m128i)) {
      __m128i search = _mm_set1_epi8(ch_from);
      __m128i delta = _mm_set1_epi8(ch_to - ch_from);

      do {
        __m128i src = _mm_loadu_si128((__m128i*)(input));
        __m128i mask = _mm_cmpeq_epi8(src, search);
        if (_mm_movemask_epi8(mask)) {
          new_str = zend_string_alloc(ZSTR_LEN(str), 0);
          memcpy(ZSTR_VAL(new_str), ZSTR_VAL(str), input - ZSTR_VAL(str));
          output = ZSTR_VAL(new_str) + (input - ZSTR_VAL(str));
          _mm_storeu_si128((__m128i *)(output),
            _mm_add_epi8(src,
              _mm_and_si128(mask, delta)));
          input += sizeof(__m128i);
          output += sizeof(__m128i);
          len -= sizeof(__m128i);
          for (; len >= sizeof(__m128i); input += sizeof(__m128i), output += sizeof(__m128i), len -= sizeof(__m128i)) {
            src = _mm_loadu_si128((__m128i*)(input));
            mask = _mm_cmpeq_epi8(src, search);
            _mm_storeu_si128((__m128i *)(output),
              _mm_add_epi8(src,
                _mm_and_si128(mask, delta)));
          }
          for (; len > 0; input++, output++, len--) {
            *output = (*input == ch_from) ? ch_to : *input;
          }
          *output = 0;
          return new_str;
        }
        input += sizeof(__m128i);
        len -= sizeof(__m128i);
      } while (len >= sizeof(__m128i));
    }
#endif
    for (; len > 0; input++, len--) {
      if (*input == ch_from) {
        new_str = zend_string_alloc(ZSTR_LEN(str), 0);
        memcpy(ZSTR_VAL(new_str), ZSTR_VAL(str), input - ZSTR_VAL(str));
        output = ZSTR_VAL(new_str) + (input - ZSTR_VAL(str));
        *output = ch_to;
        input++;
        output++;
        len--;
        for (; len > 0; input++, output++, len--) {
          *output = (*input == ch_from) ? ch_to : *input;
        }
        *output = 0;
        return new_str;
      }
    }
  } else {
    unsigned char xlat[256];

    memset(xlat, 0, sizeof(xlat));;

    for (i = 0; i < trlen; i++) {
      xlat[(size_t)(unsigned char) str_from[i]] = str_to[i] - str_from[i];
    }

    for (i = 0; i < ZSTR_LEN(str); i++) {
      if (xlat[(size_t)(unsigned char) ZSTR_VAL(str)[i]]) {
        new_str = zend_string_alloc(ZSTR_LEN(str), 0);
        memcpy(ZSTR_VAL(new_str), ZSTR_VAL(str), i);
        do {
          ZSTR_VAL(new_str)[i] = ZSTR_VAL(str)[i] + xlat[(size_t)(unsigned char) ZSTR_VAL(str)[i]];
          i++;
        } while (i < ZSTR_LEN(str));
        ZSTR_VAL(new_str)[i] = 0;
        return new_str;
      }
    }
  }

  return zend_string_copy(str);
}
/* }}} */

static void php_strtr_array_ex(zval *return_value, zend_string *input, HashTable *pats)
{
  const char *str = ZSTR_VAL(input);
  size_t slen = ZSTR_LEN(input);
  zend_ulong num_key;
  zend_string *str_key;
  size_t len, pos, old_pos;
  bool has_num_keys = false;
  size_t minlen = 128*1024;
  size_t maxlen = 0;
  HashTable str_hash;
  zval *entry;
  const char *key;
  smart_str result = {0};
  zend_ulong bitset[256/sizeof(zend_ulong)];
  zend_ulong *num_bitset;

  /* we will collect all possible key lengths */
  num_bitset = ecalloc((slen + sizeof(zend_ulong)) / sizeof(zend_ulong), sizeof(zend_ulong));
  memset(bitset, 0, sizeof(bitset));

  /* check if original array has numeric keys */
  ZEND_HASH_FOREACH_STR_KEY(pats, str_key) {
    if (UNEXPECTED(!str_key)) {
      has_num_keys = true;
    } else {
      len = ZSTR_LEN(str_key);
      if (UNEXPECTED(len == 0)) {
        php_error_docref(NULL, E_WARNING, "Ignoring replacement of empty string");
        continue;
      } else if (UNEXPECTED(len > slen)) {
        /* skip long patterns */
        continue;
      }
      if (len > maxlen) {
        maxlen = len;
      }
      if (len < minlen) {
        minlen = len;
      }
      /* remember possible key length */
      num_bitset[len / sizeof(zend_ulong)] |= Z_UL(1) << (len % sizeof(zend_ulong));
      bitset[((unsigned char)ZSTR_VAL(str_key)[0]) / sizeof(zend_ulong)] |= Z_UL(1) << (((unsigned char)ZSTR_VAL(str_key)[0]) % sizeof(zend_ulong));
    }
  } ZEND_HASH_FOREACH_END();

  if (UNEXPECTED(has_num_keys)) {
    zend_string *key_used;
    /* we have to rebuild HashTable with numeric keys */
    zend_hash_init(&str_hash, zend_hash_num_elements(pats), NULL, NULL, 0);
    ZEND_HASH_FOREACH_KEY_VAL(pats, num_key, str_key, entry) {
      if (UNEXPECTED(!str_key)) {
        key_used = zend_long_to_str(num_key);
        len = ZSTR_LEN(key_used);
        if (UNEXPECTED(len > slen)) {
          /* skip long patterns */
          zend_string_release_ex(key_used, false);
          continue;
        }
        if (len > maxlen) {
          maxlen = len;
        }
        if (len < minlen) {
          minlen = len;
        }
        /* remember possible key length */
        num_bitset[len / sizeof(zend_ulong)] |= Z_UL(1) << (len % sizeof(zend_ulong));
        bitset[((unsigned char)ZSTR_VAL(key_used)[0]) / sizeof(zend_ulong)] |= Z_UL(1) << (((unsigned char)ZSTR_VAL(key_used)[0]) % sizeof(zend_ulong));
      } else {
        key_used = str_key;
        len = ZSTR_LEN(key_used);
        if (UNEXPECTED(len > slen)) {
          /* skip long patterns */
          continue;
        }
      }
      zend_hash_add(&str_hash, key_used, entry);
      if (UNEXPECTED(!str_key)) {
        zend_string_release_ex(key_used, 0);
      }
    } ZEND_HASH_FOREACH_END();
    pats = &str_hash;
  }

  if (UNEXPECTED(minlen > maxlen)) {
    /* return the original string */
    if (pats == &str_hash) {
      zend_hash_destroy(&str_hash);
    }
    efree(num_bitset);
    RETURN_STR_COPY(input);
  }

  old_pos = pos = 0;
  while (pos <= slen - minlen) {
    /* One replacement is appended per match, so stop once over the limit. */
    if (ZEND_HAS_OOM()) {
      smart_str_free(&result);
      if (pats == &str_hash) {
        zend_hash_destroy(&str_hash);
      }
      efree(num_bitset);
      RETURN_OOM();
    }
    key = str + pos;
    if (bitset[((unsigned char)key[0]) / sizeof(zend_ulong)] & (Z_UL(1) << (((unsigned char)key[0]) % sizeof(zend_ulong)))) {
      len = maxlen;
      if (len > slen - pos) {
        len = slen - pos;
      }
      while (len >= minlen) {
        if ((num_bitset[len / sizeof(zend_ulong)] & (Z_UL(1) << (len % sizeof(zend_ulong))))) {
          entry = zend_hash_str_find(pats, key, len);
          if (entry != NULL) {
            zend_string *tmp;
            zend_string *s = zval_get_tmp_string(entry, &tmp);
            smart_str_appendl(&result, str + old_pos, pos - old_pos);
            smart_str_append(&result, s);
            old_pos = pos + len;
            pos = old_pos - 1;
            zend_tmp_string_release(tmp);
            break;
          }
        }
        len--;
      }
    }
    pos++;
  }

  if (result.s) {
    smart_str_appendl(&result, str + old_pos, slen - old_pos);
    RETVAL_STR(smart_str_extract(&result));
  } else {
    smart_str_free(&result);
    RETVAL_STR_COPY(input);
  }

  if (pats == &str_hash) {
    zend_hash_destroy(&str_hash);
  }
  efree(num_bitset);
}

/* {{{ count_chars */
static zend_always_inline zend_long count_chars(const char *p, zend_long length, char ch)
{
  zend_long count = 0;
  const char *endp;

#ifdef XSSE2
  if (length >= sizeof(__m128i)) {
    __m128i search = _mm_set1_epi8(ch);

    do {
      __m128i src = _mm_loadu_si128((__m128i*)(p));
      uint32_t mask = _mm_movemask_epi8(_mm_cmpeq_epi8(src, search));
      // TODO: It would be great to use POPCNT, but it's available only with SSE4.1
#if 1
      while (mask != 0) {
        count++;
        mask = mask & (mask - 1);
      }
#else
      if (mask) {
        mask = mask - ((mask >> 1) & 0x5555);
        mask = (mask & 0x3333) + ((mask >> 2) & 0x3333);
        mask = (mask + (mask >> 4)) & 0x0F0F;
        mask = (mask + (mask >> 8)) & 0x00ff;
        count += mask;
      }
#endif
      p += sizeof(__m128i);
      length -= sizeof(__m128i);
    } while (length >= sizeof(__m128i));
  }
  endp = p + length;
  while (p != endp) {
    count += (*p == ch);
    p++;
  }
#else
  endp = p + length;
  while ((p = memchr(p, ch, endp-p))) {
    count++;
    p++;
  }
#endif
  return count;
}
/* }}} */

/* {{{ php_char_to_str_ex */
static zend_string* php_char_to_str_ex(zend_string *str, char from, char *to, size_t to_len, bool case_sensitivity, zend_long *replace_count)
{
  zend_string *result;
  size_t char_count;
  int lc_from = 0;
  const char *source, *source_end;
  char *target;

  if (case_sensitivity) {
    char_count = count_chars(ZSTR_VAL(str), ZSTR_LEN(str), from);
  } else {
    char_count = 0;
    lc_from = zend_tolower_ascii(from);
    source_end = ZSTR_VAL(str) + ZSTR_LEN(str);
    for (source = ZSTR_VAL(str); source < source_end; source++) {
      if (zend_tolower_ascii(*source) == lc_from) {
        char_count++;
      }
    }
  }

  if (char_count == 0) {
    return zend_string_copy(str);
  }

  if (replace_count) {
    *replace_count += char_count;
  }

  if (to_len > 0) {
    /* The replacement is repeated once per match, so the result can dwarf the input. */
    result = zend_string_safe_falloc(char_count, to_len - 1, ZSTR_LEN(str), 0, /* fallible */ true);
    if (UNEXPECTED(!result)) {
      return NULL;
    }
  } else {
    result = zend_string_alloc(ZSTR_LEN(str) - char_count, 0);
  }
  target = ZSTR_VAL(result);

  if (case_sensitivity) {
    char *p = ZSTR_VAL(str), *e = p + ZSTR_LEN(str), *s = ZSTR_VAL(str);

    while ((p = memchr(p, from, (e - p)))) {
      target = zend_mempcpy(target, s, (p - s));
      target = zend_mempcpy(target, to, to_len);
      p++;
      s = p;
      if (--char_count == 0) break;
    }
    if (s < e) {
      target = zend_mempcpy(target, s, e - s);
    }
  } else {
    source_end = ZSTR_VAL(str) + ZSTR_LEN(str);
    for (source = ZSTR_VAL(str); source < source_end; source++) {
      if (zend_tolower_ascii(*source) == lc_from) {
        target = zend_mempcpy(target, to, to_len);
      } else {
        *target = *source;
        target++;
      }
    }
  }
  *target = 0;
  return result;
}
/* }}} */

/* {{{ php_str_to_str_ex */
static zend_string *php_str_to_str_ex(zend_string *haystack,
  const char *needle, size_t needle_len, const char *str, size_t str_len, zend_long *replace_count)
{

  if (needle_len < ZSTR_LEN(haystack)) {
    zend_string *new_str;
    const char *end;
    const char *p, *r;
    char *e;

    if (needle_len == str_len) {
      new_str = NULL;
      end = ZSTR_VAL(haystack) + ZSTR_LEN(haystack);
      for (p = ZSTR_VAL(haystack); (r = (char*)php_memnstr(p, needle, needle_len, end)); p = r + needle_len) {
        if (!new_str) {
          new_str = zend_string_init(ZSTR_VAL(haystack), ZSTR_LEN(haystack), 0);
        }
        memcpy(ZSTR_VAL(new_str) + (r - ZSTR_VAL(haystack)), str, str_len);
        (*replace_count)++;
      }
      if (!new_str) {
        goto nothing_todo;
      }
      return new_str;
    } else {
      size_t count = 0;
      const char *o = ZSTR_VAL(haystack);
      const char *n = needle;
      const char *endp = o + ZSTR_LEN(haystack);

      while ((o = (char*)php_memnstr(o, n, needle_len, endp))) {
        o += needle_len;
        count++;
      }
      if (count == 0) {
        /* Needle doesn't occur, shortcircuit the actual replacement. */
        goto nothing_todo;
      }
      if (str_len > needle_len) {
        new_str = zend_string_safe_falloc(count, str_len - needle_len, ZSTR_LEN(haystack), 0, /* fallible */ true);
        if (UNEXPECTED(!new_str)) {
          return NULL;
        }
      } else {
        new_str = zend_string_alloc(count * (str_len - needle_len) + ZSTR_LEN(haystack), 0);
      }

      e = ZSTR_VAL(new_str);
      end = ZSTR_VAL(haystack) + ZSTR_LEN(haystack);
      for (p = ZSTR_VAL(haystack); (r = (char*)php_memnstr(p, needle, needle_len, end)); p = r + needle_len) {
        e = zend_mempcpy(e, p, r - p);
        e = zend_mempcpy(e, str, str_len);
        (*replace_count)++;
      }

      if (p < end) {
        e = zend_mempcpy(e, p, end - p);
      }

      *e = '\0';
      return new_str;
    }
  } else if (needle_len > ZSTR_LEN(haystack) || memcmp(ZSTR_VAL(haystack), needle, ZSTR_LEN(haystack))) {
nothing_todo:
    return zend_string_copy(haystack);
  } else {
    (*replace_count)++;
    return zend_string_init_fast(str, str_len);
  }
}
/* }}} */

/* {{{ php_str_to_str_i_ex */
static zend_string *php_str_to_str_i_ex(zend_string *haystack, const char *lc_haystack,
  zend_string *needle, const char *str, size_t str_len, zend_long *replace_count)
{
  zend_string *new_str = NULL;
  zend_string *lc_needle;

  if (ZSTR_LEN(needle) < ZSTR_LEN(haystack)) {
    const char *end;
    const char *p, *r;
    char *e;

    if (ZSTR_LEN(needle) == str_len) {
      lc_needle = zend_string_tolower(needle);
      end = lc_haystack + ZSTR_LEN(haystack);
      for (p = lc_haystack; (r = (char*)php_memnstr(p, ZSTR_VAL(lc_needle), ZSTR_LEN(lc_needle), end)); p = r + ZSTR_LEN(lc_needle)) {
        if (!new_str) {
          new_str = zend_string_init(ZSTR_VAL(haystack), ZSTR_LEN(haystack), 0);
        }
        memcpy(ZSTR_VAL(new_str) + (r - lc_haystack), str, str_len);
        (*replace_count)++;
      }
      zend_string_release_ex(lc_needle, 0);

      if (!new_str) {
        goto nothing_todo;
      }
      return new_str;
    } else {
      size_t count = 0;
      const char *o = lc_haystack;
      const char *n;
      const char *endp = o + ZSTR_LEN(haystack);

      lc_needle = zend_string_tolower(needle);
      n = ZSTR_VAL(lc_needle);

      while ((o = (char*)php_memnstr(o, n, ZSTR_LEN(lc_needle), endp))) {
        o += ZSTR_LEN(lc_needle);
        count++;
      }
      if (count == 0) {
        /* Needle doesn't occur, shortcircuit the actual replacement. */
        zend_string_release_ex(lc_needle, 0);
        goto nothing_todo;
      }

      if (str_len > ZSTR_LEN(lc_needle)) {
        new_str = zend_string_safe_falloc(count, str_len - ZSTR_LEN(lc_needle), ZSTR_LEN(haystack), 0, /* fallible */ true);
        if (UNEXPECTED(!new_str)) {
          return NULL;
        }
      } else {
        new_str = zend_string_alloc(count * (str_len - ZSTR_LEN(lc_needle)) + ZSTR_LEN(haystack), 0);
      }

      e = ZSTR_VAL(new_str);
      end = lc_haystack + ZSTR_LEN(haystack);

      for (p = lc_haystack; (r = (char*)php_memnstr(p, ZSTR_VAL(lc_needle), ZSTR_LEN(lc_needle), end)); p = r + ZSTR_LEN(lc_needle)) {
        e = zend_mempcpy(e, ZSTR_VAL(haystack) + (p - lc_haystack), r - p);
        e = zend_mempcpy(e, str, str_len);
        (*replace_count)++;
      }

      if (p < end) {
        e = zend_mempcpy(e, ZSTR_VAL(haystack) + (p - lc_haystack), end - p);
      }
      *e = '\0';

      zend_string_release_ex(lc_needle, 0);

      return new_str;
    }
  } else if (ZSTR_LEN(needle) > ZSTR_LEN(haystack)) {
nothing_todo:
    return zend_string_copy(haystack);
  } else {
    lc_needle = zend_string_tolower(needle);

    if (memcmp(lc_haystack, ZSTR_VAL(lc_needle), ZSTR_LEN(lc_needle))) {
      zend_string_release_ex(lc_needle, 0);
      goto nothing_todo;
    }
    zend_string_release_ex(lc_needle, 0);

    new_str = zend_string_init(str, str_len, 0);

    (*replace_count)++;
    return new_str;
  }
}
/* }}} */

/* {{{ php_str_to_str */
PHPAPI zend_string *php_str_to_str(const char *haystack, size_t length, const char *needle, size_t needle_len, const char *str, size_t str_len)
{
  zend_string *new_str;

  if (needle_len < length) {
    const char *end;
    const char *s, *p;
    char *e, *r;

    if (needle_len == str_len) {
      new_str = zend_string_init(haystack, length, 0);
      end = ZSTR_VAL(new_str) + length;
      for (p = ZSTR_VAL(new_str); (r = (char*)php_memnstr(p, needle, needle_len, end)); p = r + needle_len) {
        memcpy(r, str, str_len);
      }
      return new_str;
    } else {
      if (str_len < needle_len) {
        new_str = zend_string_alloc(length, 0);
      } else {
        size_t count = 0;
        const char *o = haystack;
        const char *n = needle;
        const char *endp = o + length;

        while ((o = (char*)php_memnstr(o, n, needle_len, endp))) {
          o += needle_len;
          count++;
        }
        if (count == 0) {
          /* Needle doesn't occur, shortcircuit the actual replacement. */
          new_str = zend_string_init(haystack, length, 0);
          return new_str;
        } else {
          if (str_len > needle_len) {
            new_str = zend_string_safe_alloc(count, str_len - needle_len, length, 0);
          } else {
            new_str = zend_string_alloc(count * (str_len - needle_len) + length, 0);
          }
        }
      }

      s = e = ZSTR_VAL(new_str);
      end = haystack + length;
      for (p = haystack; (r = (char*)php_memnstr(p, needle, needle_len, end)); p = r + needle_len) {
        e = zend_mempcpy(e, p, r - p);
        e = zend_mempcpy(e, str, str_len);
      }

      if (p < end) {
        e = zend_mempcpy(e, p, end - p);
      }

      *e = '\0';
      new_str = zend_string_truncate(new_str, e - s, 0);
      return new_str;
    }
  } else if (needle_len > length || memcmp(haystack, needle, length)) {
    new_str = zend_string_init(haystack, length, 0);
    return new_str;
  } else {
    new_str = zend_string_init(str, str_len, 0);

    return new_str;
  }
}
/* }}} */

static void php_strtr_array(zval *return_value, zend_string *str, HashTable *from_ht)
{
  if (zend_hash_num_elements(from_ht) < 1) {
    RETURN_STR_COPY(str);
  }

  /* Converting a replacement may call __toString(), which can destroy from_ht. */
  GC_TRY_ADDREF(from_ht);

  if (zend_hash_num_elements(from_ht) == 1) {
    zend_long num_key;
    zend_string *str_key, *tmp_str, *replace, *tmp_replace;
    zval *entry;

    ZEND_HASH_FOREACH_KEY_VAL(from_ht, num_key, str_key, entry) {
      tmp_str = NULL;
      if (UNEXPECTED(!str_key)) {
        str_key = tmp_str = zend_long_to_str(num_key);
      }
      replace = zval_get_tmp_string(entry, &tmp_replace);
      if (ZSTR_LEN(str_key) < 1) {
        php_error_docref(NULL, E_WARNING, "Ignoring replacement of empty string");
        RETVAL_STR_COPY(str);
      } else if (ZSTR_LEN(str_key) == 1) {
        zend_string *res = php_char_to_str_ex(str,
              ZSTR_VAL(str_key)[0],
              ZSTR_VAL(replace),
              ZSTR_LEN(replace),
              /* case_sensitive */ true,
              NULL);
        if (UNEXPECTED(!res)) {
          goto strtr_oom;
        }
        RETVAL_STR(res);
      } else {
        zend_long dummy = 0;
        zend_string *res = php_str_to_str_ex(str,
              ZSTR_VAL(str_key), ZSTR_LEN(str_key),
              ZSTR_VAL(replace), ZSTR_LEN(replace), &dummy);
        if (UNEXPECTED(!res)) {
          goto strtr_oom;
        }
        RETVAL_STR(res);
      }
      zend_tmp_string_release(tmp_str);
      zend_tmp_string_release(tmp_replace);
      break;

strtr_oom:
      zend_tmp_string_release(tmp_str);
      zend_tmp_string_release(tmp_replace);
      RETURN_OOM();
    } ZEND_HASH_FOREACH_END();
  } else {
    php_strtr_array_ex(return_value, str, from_ht);
  }

  GC_TRY_DTOR_NO_REF(from_ht);
}

/* {{{ Translates characters in str using given translation tables */
PHP_FUNCTION(strtr)
{
  zend_string *str, *from_str = NULL;
  HashTable *from_ht = NULL;
  char *to = NULL;
  size_t to_len = 0;

  if (ZEND_NUM_ARGS() <= 2) {
    ZEND_PARSE_PARAMETERS_START(2, 2)
      Z_PARAM_STR(str)
      Z_PARAM_ARRAY_HT(from_ht)
    ZEND_PARSE_PARAMETERS_END();
  } else {
    ZEND_PARSE_PARAMETERS_START(3, 3)
      Z_PARAM_STR(str)
      Z_PARAM_STR(from_str)
      Z_PARAM_STRING(to, to_len)
    ZEND_PARSE_PARAMETERS_END();
  }

  /* shortcut for empty string */
  if (ZSTR_LEN(str) == 0) {
    RETURN_EMPTY_STRING();
  }

  if (!to) {
    php_strtr_array(return_value, str, from_ht);
  } else {
    RETURN_STR(php_strtr_ex(str,
          ZSTR_VAL(from_str),
          to,
          MIN(ZSTR_LEN(from_str), to_len)));
  }
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(strtr, 2)
{
  zval str_tmp;
  zend_string *str;
  zval *from;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_ARRAY(2, from);

  if (ZSTR_LEN(str) == 0) {
    RETVAL_EMPTY_STRING();
    goto flf_clean;
  }

  php_strtr_array(return_value, str, Z_ARR_P(from));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
}

ZEND_FRAMELESS_FUNCTION(strtr, 3)
{
  zval str_tmp, from_tmp, to_tmp;
  zend_string *str, *from, *to;

  Z_FLF_PARAM_STR(1, str, str_tmp);
  Z_FLF_PARAM_STR(2, from, from_tmp);
  Z_FLF_PARAM_STR(3, to, to_tmp);

  if (ZSTR_LEN(str) == 0) {
    RETVAL_EMPTY_STRING();
    goto flf_clean;
  }

  RETVAL_STR(php_strtr_ex(str, ZSTR_VAL(from), ZSTR_VAL(to), MIN(ZSTR_LEN(from), ZSTR_LEN(to))));

flf_clean:
  Z_FLF_PARAM_FREE_STR(1, str_tmp);
  Z_FLF_PARAM_FREE_STR(2, from_tmp);
  Z_FLF_PARAM_FREE_STR(3, to_tmp);
}

/* {{{ Reverse a string */
#ifdef ZEND_INTRIN_SSSE3_NATIVE
#include <tmmintrin.h>
#elif defined(__aarch64__) || defined(_M_ARM64)
#include <arm_neon.h>
#endif
PHP_FUNCTION(strrev)
{
  zend_string *str;
  const char *s, *e;
  char *p;
  zend_string *n;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  n = zend_string_alloc(ZSTR_LEN(str), 0);
  p = ZSTR_VAL(n);

  s = ZSTR_VAL(str);
  e = s + ZSTR_LEN(str);
  --e;
#ifdef ZEND_INTRIN_SSSE3_NATIVE
  if (e - s > 15) {
    const __m128i map = _mm_set_epi8(
        0, 1, 2, 3,
        4, 5, 6, 7,
        8, 9, 10, 11,
        12, 13, 14, 15);
    do {
      const __m128i str = _mm_loadu_si128((__m128i *)(e - 15));
      _mm_storeu_si128((__m128i *)p, _mm_shuffle_epi8(str, map));
      p += 16;
      e -= 16;
    } while (e - s > 15);
  }
#elif defined(__aarch64__)
  if (e - s > 15) {
    do {
      const uint8x16_t str = vld1q_u8((uint8_t *)(e - 15));
      /* Synthesize rev128 with a rev64 + ext. */
      const uint8x16_t rev = vrev64q_u8(str);
      const uint8x16_t ext = (uint8x16_t)
        vextq_u64((uint64x2_t)rev, (uint64x2_t)rev, 1);
      vst1q_u8((uint8_t *)p, ext);
      p += 16;
      e -= 16;
    } while (e - s > 15);
  }
#elif defined(_M_ARM64)
  if (e - s > 15) {
    do {
      const __n128 str = vld1q_u8((uint8_t *)(e - 15));
      /* Synthesize rev128 with a rev64 + ext. */
      /* strange force cast limit on windows: you cannot convert anything */
      const __n128 rev = vrev64q_u8(str);
      const __n128 ext = vextq_u64(rev, rev, 1);
      vst1q_u8((uint8_t *)p, ext);
      p += 16;
      e -= 16;
    } while (e - s > 15);
  }
#endif
  while (e >= s) {
    *p++ = *e--;
  }

  *p = '\0';

  RETVAL_NEW_STR(n);
}
/* }}} */

/* {{{ php_similar_str */
static void php_similar_str(const char *txt1, size_t len1, const char *txt2, size_t len2, size_t *pos1, size_t *pos2, size_t *max, size_t *count)
{
  const char *p, *q;
  const char *end1 = (char *) txt1 + len1;
  const char *end2 = (char *) txt2 + len2;
  size_t l;

  *max = 0;
  *count = 0;
  for (p = (char *) txt1; p < end1; p++) {
    for (q = (char *) txt2; q < end2; q++) {
      for (l = 0; (p + l < end1) && (q + l < end2) && (p[l] == q[l]); l++);
      if (l > *max) {
        *max = l;
        *count += 1;
        *pos1 = p - txt1;
        *pos2 = q - txt2;
      }
    }
  }
}
/* }}} */

/* {{{ php_similar_char */
static size_t php_similar_char(const char *txt1, size_t len1, const char *txt2, size_t len2)
{
  size_t sum;
  size_t pos1 = 0, pos2 = 0, max, count;

  php_similar_str(txt1, len1, txt2, len2, &pos1, &pos2, &max, &count);
  if ((sum = max)) {
    if (pos1 && pos2 && count > 1) {
      sum += php_similar_char(txt1, pos1,
                  txt2, pos2);
    }
    if ((pos1 + max < len1) && (pos2 + max < len2)) {
      sum += php_similar_char(txt1 + pos1 + max, len1 - pos1 - max,
                  txt2 + pos2 + max, len2 - pos2 - max);
    }
  }

  return sum;
}
/* }}} */

/* {{{ Calculates the similarity between two strings */
PHP_FUNCTION(similar_text)
{
  zend_string *t1, *t2;
  zval *percent = NULL;
  bool compute_percentage = ZEND_NUM_ARGS() >= 3;
  size_t sim;

  ZEND_PARSE_PARAMETERS_START(2, 3)
    Z_PARAM_STR(t1)
    Z_PARAM_STR(t2)
    Z_PARAM_OPTIONAL
    Z_PARAM_ZVAL(percent)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(t1) + ZSTR_LEN(t2) == 0) {
    if (compute_percentage) {
      ZEND_TRY_ASSIGN_REF_DOUBLE(percent, 0);
    }

    RETURN_LONG(0);
  }

  sim = php_similar_char(ZSTR_VAL(t1), ZSTR_LEN(t1), ZSTR_VAL(t2), ZSTR_LEN(t2));

  if (compute_percentage) {
    ZEND_TRY_ASSIGN_REF_DOUBLE(percent, sim * 200.0 / (ZSTR_LEN(t1) + ZSTR_LEN(t2)));
  }

  RETURN_LONG(sim);
}
/* }}} */

/* {{{ Escapes all chars mentioned in charlist with backslash. It creates octal representations if asked to backslash characters with 8th bit set or with ASCII<32 (except '\n', '\r', '\t' etc...) */
PHP_FUNCTION(addcslashes)
{
  zend_string *str, *what;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(str)
    Z_PARAM_STR(what)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(str) == 0) {
    RETURN_EMPTY_STRING();
  }

  if (ZSTR_LEN(what) == 0) {
    RETURN_STR_COPY(str);
  }

  RETURN_STR(php_addcslashes_str(ZSTR_VAL(str), ZSTR_LEN(str), ZSTR_VAL(what), ZSTR_LEN(what)));
}
/* }}} */

/* {{{ Escapes single quote, double quotes and backslash characters in a string with backslashes */
PHP_FUNCTION(addslashes)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  if (ZSTR_LEN(str) == 0) {
    RETURN_EMPTY_STRING();
  }

  RETURN_STR(php_addslashes(str));
}
/* }}} */

/* {{{ Strips backslashes from a string. Uses C-style conventions */
PHP_FUNCTION(stripcslashes)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  ZVAL_STRINGL(return_value, ZSTR_VAL(str), ZSTR_LEN(str));
  php_stripcslashes(Z_STR_P(return_value));
}
/* }}} */

/* {{{ Strips backslashes from a string */
PHP_FUNCTION(stripslashes)
{
  zend_string *str;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(str)
  ZEND_PARSE_PARAMETERS_END();

  ZVAL_STRINGL(return_value, ZSTR_VAL(str), ZSTR_LEN(str));
  php_stripslashes(Z_STR_P(return_value));
}
/* }}} */

/* {{{ php_stripcslashes */
PHPAPI void php_stripcslashes(zend_string *str)
{
  const char *source, *end;
  char *target;
  size_t  nlen = ZSTR_LEN(str), i;
  char numtmp[4];

  for (source = (char*)ZSTR_VAL(str), end = source + ZSTR_LEN(str), target = ZSTR_VAL(str); source < end; source++) {
    if (*source == '\\' && source + 1 < end) {
      source++;
      switch (*source) {
        case 'n':  *target++='\n'; nlen--; break;
        case 'r':  *target++='\r'; nlen--; break;
        case 'a':  *target++='\a'; nlen--; break;
        case 't':  *target++='\t'; nlen--; break;
        case 'v':  *target++='\v'; nlen--; break;
        case 'b':  *target++='\b'; nlen--; break;
        case 'f':  *target++='\f'; nlen--; break;
        case '\\': *target++='\\'; nlen--; break;
        case 'x':
          if (source+1 < end && isxdigit((unsigned char)source[1])) {
            numtmp[0] = *++source;
            if (source+1 < end && isxdigit((unsigned char)source[1])) {
              numtmp[1] = *++source;
              numtmp[2] = '\0';
              nlen-=3;
            } else {
              numtmp[1] = '\0';
              nlen-=2;
            }
            *target++=(char)strtol(numtmp, NULL, 16);
            break;
          }
          ZEND_FALLTHROUGH;
        default:
          i=0;
          while (source < end && *source >= '0' && *source <= '7' && i<3) {
            numtmp[i++] = *source++;
          }
          if (i) {
            numtmp[i]='\0';
            *target++=(char)strtol(numtmp, NULL, 8);
            nlen-=i;
            source--;
          } else {
            *target++=*source;
            nlen--;
          }
      }
    } else {
      *target++=*source;
    }
  }

  if (nlen != 0) {
    *target='\0';
  }

  ZSTR_LEN(str) = nlen;
}
/* }}} */

/* {{{ php_addcslashes_str */
PHPAPI zend_string *php_addcslashes_str(const char *str, size_t len, const char *what, size_t wlength)
{
  char flags[256];
  char *target;
  const char *source, *end;
  char c;
  size_t  newlen;
  zend_string *new_str = zend_string_safe_alloc(4, len, 0, 0);

  php_charmask((const unsigned char *) what, wlength, flags);

  for (source = str, end = source + len, target = ZSTR_VAL(new_str); source < end; source++) {
    c = *source;
    if (flags[(unsigned char)c]) {
      if ((unsigned char) c < 32 || (unsigned char) c > 126) {
        *target++ = '\\';
        switch (c) {
          case '\n': *target++ = 'n'; break;
          case '\t': *target++ = 't'; break;
          case '\r': *target++ = 'r'; break;
          case '\a': *target++ = 'a'; break;
          case '\v': *target++ = 'v'; break;
          case '\b': *target++ = 'b'; break;
          case '\f': *target++ = 'f'; break;
          default:
            /* Write the byte as three octal digits, including leading zeros. */
            *target++ = ((unsigned char) c >> 6) + '0';
            *target++ = (((unsigned char) c >> 3) & 7) + '0';
            *target++ = ((unsigned char) c & 7) + '0';
        }
        continue;
      }
      *target++ = '\\';
    }
    *target++ = c;
  }
  *target = 0;
  newlen = target - ZSTR_VAL(new_str);
  if (newlen < len * 4) {
    new_str = zend_string_truncate(new_str, newlen, 0);
  }
  return new_str;
}
/* }}} */

/* {{{ php_addcslashes */
PHPAPI zend_string *php_addcslashes(zend_string *str, const char *what, size_t wlength)
{
  return php_addcslashes_str(ZSTR_VAL(str), ZSTR_LEN(str), what, wlength);
}
/* }}} */

/* {{{ php_addslashes */

#ifdef ZEND_INTRIN_SSE4_2_NATIVE
# include <nmmintrin.h>
# include "Zend/zend_bitset.h"
#elif defined(ZEND_INTRIN_SSE4_2_RESOLVER)
# include <nmmintrin.h>
# include "Zend/zend_bitset.h"
# include "Zend/zend_cpuinfo.h"

ZEND_INTRIN_SSE4_2_FUNC_DECL(zend_string *php_addslashes_sse42(zend_string *str));
zend_string *php_addslashes_default(zend_string *str);

# ifdef ZEND_INTRIN_SSE4_2_FUNC_PROTO
PHPAPI zend_string *php_addslashes(zend_string *str) __attribute__((ifunc("resolve_addslashes")));

typedef zend_string *(*php_addslashes_func_t)(zend_string *);

ZEND_NO_SANITIZE_ADDRESS
ZEND_ATTRIBUTE_UNUSED /* clang mistakenly warns about this */
static php_addslashes_func_t resolve_addslashes(void) {
  if (zend_cpu_supports_sse42()) {
    return php_addslashes_sse42;
  }
  return php_addslashes_default;
}
# else /* ZEND_INTRIN_SSE4_2_FUNC_PTR */

static zend_string *(*php_addslashes_ptr)(zend_string *str) = NULL;

PHPAPI zend_string *php_addslashes(zend_string *str) {
  return php_addslashes_ptr(str);
}

/* {{{ PHP_MINIT_FUNCTION */
PHP_MINIT_FUNCTION(string_intrin)
{
  if (zend_cpu_supports_sse42()) {
    php_addslashes_ptr = php_addslashes_sse42;
  } else {
    php_addslashes_ptr = php_addslashes_default;
  }
  return SUCCESS;
}
/* }}} */
# endif
#endif

#if defined(ZEND_INTRIN_SSE4_2_NATIVE) || defined(ZEND_INTRIN_SSE4_2_RESOLVER)
# ifdef ZEND_INTRIN_SSE4_2_NATIVE
PHPAPI zend_string *php_addslashes(zend_string *str) /* {{{ */
# elif defined(ZEND_INTRIN_SSE4_2_RESOLVER)
zend_string *php_addslashes_sse42(zend_string *str)
# endif
{
  ZEND_SET_ALIGNED(16, static const char slashchars[16]) = "\'\"\\\0";
  __m128i w128, s128;
  uint32_t res = 0;
  /* maximum string length, worst case situation */
  char *target;
  const char *source, *end;
  size_t offset;
  zend_string *new_str;

  if (!str) {
    return ZSTR_EMPTY_ALLOC();
  }

  source = ZSTR_VAL(str);
  end = source + ZSTR_LEN(str);

  if (ZSTR_LEN(str) > 15) {
    w128 = _mm_load_si128((__m128i *)slashchars);
    do {
      s128 = _mm_loadu_si128((__m128i *)source);
      res = _mm_cvtsi128_si32(_mm_cmpestrm(w128, 4, s128, 16, _SIDD_UBYTE_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK));
      if (res) {
        goto do_escape;
      }
      source += 16;
    } while ((end - source) > 15);
  }

  while (source < end) {
    switch (*source) {
      case '\0':
      case '\'':
      case '\"':
      case '\\':
        goto do_escape;
      default:
        source++;
        break;
    }
  }

  return zend_string_copy(str);

do_escape:
  offset = source - (char *)ZSTR_VAL(str);
  new_str = zend_string_safe_alloc(2, ZSTR_LEN(str) - offset, offset, 0);
  memcpy(ZSTR_VAL(new_str), ZSTR_VAL(str), offset);
  target = ZSTR_VAL(new_str) + offset;

  if (res) {
    int pos = 0;
    do {
      int i, n = zend_ulong_ntz(res);
      for (i = 0; i < n; i++) {
        *target++ = source[pos + i];
      }
      pos += n;
      *target++ = '\\';
      if (source[pos] == '\0') {
        *target++ = '0';
      } else {
        *target++ = source[pos];
      }
      pos++;
      res = res >> (n + 1);
    } while (res);

    for (; pos < 16; pos++) {
      *target++ = source[pos];
    }
    source += 16;
  } else if (end - source > 15) {
    w128 = _mm_load_si128((__m128i *)slashchars);
  }

  for (; end - source > 15; source += 16) {
    int pos = 0;
    s128 = _mm_loadu_si128((__m128i *)source);
    res = _mm_cvtsi128_si32(_mm_cmpestrm(w128, 4, s128, 16, _SIDD_UBYTE_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK));
    if (res) {
      do {
        int i, n = zend_ulong_ntz(res);
        for (i = 0; i < n; i++) {
          *target++ = source[pos + i];
        }
        pos += n;
        *target++ = '\\';
        if (source[pos] == '\0') {
          *target++ = '0';
        } else {
          *target++ = source[pos];
        }
        pos++;
        res = res >> (n + 1);
      } while (res);
      for (; pos < 16; pos++) {
        *target++ = source[pos];
      }
    } else {
      _mm_storeu_si128((__m128i*)target, s128);
      target += 16;
    }
  }

  while (source < end) {
    switch (*source) {
      case '\0':
        *target++ = '\\';
        *target++ = '0';
        break;
      case '\'':
      case '\"':
      case '\\':
        *target++ = '\\';
        ZEND_FALLTHROUGH;
      default:
        *target++ = *source;
        break;
    }
    source++;
  }

  *target = '\0';

  if (ZSTR_LEN(new_str) - (target - ZSTR_VAL(new_str)) > 16) {
    new_str = zend_string_truncate(new_str, target - ZSTR_VAL(new_str), 0);
  } else {
    ZSTR_LEN(new_str) = target - ZSTR_VAL(new_str);
  }

  return new_str;
}
/* }}} */
#endif

#if defined(__aarch64__) || defined(_M_ARM64)
typedef union {
  uint8_t mem[16];
  uint64_t dw[2];
} quad_word;

static zend_always_inline quad_word aarch64_contains_slash_chars(uint8x16_t x) {
  uint8x16_t s0 = vceqq_u8(x, vdupq_n_u8('\0'));
  uint8x16_t s1 = vceqq_u8(x, vdupq_n_u8('\''));
  uint8x16_t s2 = vceqq_u8(x, vdupq_n_u8('\"'));
  uint8x16_t s3 = vceqq_u8(x, vdupq_n_u8('\\'));
  uint8x16_t s01 = vorrq_u8(s0, s1);
  uint8x16_t s23 = vorrq_u8(s2, s3);
  uint8x16_t s0123 = vorrq_u8(s01, s23);
  quad_word qw;
  vst1q_u8(qw.mem, s0123);
  return qw;
}

static zend_always_inline char *aarch64_add_slashes(quad_word res, const char *source, char *target)
{
  for (int i = 0; i < 16; i++) {
    char s = source[i];
    if (res.mem[i] == 0)
      *target++ = s;
    else {
      *target++ = '\\';
      if (s == '\0')
        *target++ = '0';
      else
        *target++ = s;
    }
  }
  return target;
}
#endif /* defined(__aarch64__) || defined(_M_ARM64) */

#ifndef ZEND_INTRIN_SSE4_2_NATIVE
# ifdef ZEND_INTRIN_SSE4_2_RESOLVER
zend_string *php_addslashes_default(zend_string *str) /* {{{ */
# else
PHPAPI zend_string *php_addslashes(zend_string *str)
# endif
{
  /* maximum string length, worst case situation */
  char *target;
  const char *source, *end;
  size_t offset;
  zend_string *new_str;

  if (!str) {
    return ZSTR_EMPTY_ALLOC();
  }

  source = ZSTR_VAL(str);
  end = source + ZSTR_LEN(str);

# if defined(__aarch64__) || defined(_M_ARM64)
  quad_word res = {0};
  if (ZSTR_LEN(str) > 15) {
    do {
      res = aarch64_contains_slash_chars(vld1q_u8((uint8_t *)source));
      if (res.dw[0] | res.dw[1])
        goto do_escape;
      source += 16;
    } while ((end - source) > 15);
  }
  /* Finish the last 15 bytes or less with the scalar loop. */
# endif /* defined(__aarch64__) || defined(_M_ARM64) */

  while (source < end) {
    switch (*source) {
      case '\0':
      case '\'':
      case '\"':
      case '\\':
        goto do_escape;
      default:
        source++;
        break;
    }
  }

  return zend_string_copy(str);

do_escape:
  offset = source - (char *)ZSTR_VAL(str);
  new_str = zend_string_safe_alloc(2, ZSTR_LEN(str) - offset, offset, 0);
  memcpy(ZSTR_VAL(new_str), ZSTR_VAL(str), offset);
  target = ZSTR_VAL(new_str) + offset;

# if defined(__aarch64__) || defined(_M_ARM64)
  if (res.dw[0] | res.dw[1]) {
    target = aarch64_add_slashes(res, source, target);
    source += 16;
  }
  for (; end - source > 15; source += 16) {
    uint8x16_t x = vld1q_u8((uint8_t *)source);
    res = aarch64_contains_slash_chars(x);
    if (res.dw[0] | res.dw[1]) {
      target = aarch64_add_slashes(res, source, target);
    } else {
      vst1q_u8((uint8_t*)target, x);
      target += 16;
    }
  }
  /* Finish the last 15 bytes or less with the scalar loop. */
# endif /* defined(__aarch64__) || defined(_M_ARM64) */

  while (source < end) {
    switch (*source) {
      case '\0':
        *target++ = '\\';
        *target++ = '0';
        break;
      case '\'':
      case '\"':
      case '\\':
        *target++ = '\\';
        ZEND_FALLTHROUGH;
      default:
        *target++ = *source;
        break;
    }
    source++;
  }

  *target = '\0';

  if (ZSTR_LEN(new_str) - (target - ZSTR_VAL(new_str)) > 16) {
    new_str = zend_string_truncate(new_str, target - ZSTR_VAL(new_str), 0);
  } else {
    ZSTR_LEN(new_str) = target - ZSTR_VAL(new_str);
  }

  return new_str;
}
#endif
/* }}} */
/* }}} */

/* {{{ php_stripslashes
 *
 * be careful, this edits the string in-place */
static zend_always_inline char *php_stripslashes_impl(const char *str, char *out, size_t len)
{
#if defined(__aarch64__) || defined(_M_ARM64)
  while (len > 15) {
    uint8x16_t x = vld1q_u8((uint8_t *)str);
    quad_word q;
    vst1q_u8(q.mem, vceqq_u8(x, vdupq_n_u8('\\')));
    if (q.dw[0] | q.dw[1]) {
      unsigned int i = 0;
      while (i < 16) {
        if (q.mem[i] == 0) {
          *out++ = str[i];
          i++;
          continue;
        }

        i++;      /* skip the slash */
        if (i < len) {
          char s = str[i];
          if (s == '0')
            *out++ = '\0';
          else
            *out++ = s; /* preserve the next character */
          i++;
        }
      }
      str += i;
      len -= i;
    } else {
      vst1q_u8((uint8_t*)out, x);
      out += 16;
      str += 16;
      len -= 16;
    }
  }
  /* Finish the last 15 bytes or less with the scalar loop. */
#endif /* defined(__aarch64__) || defined(_M_ARM64) */
  while (len > 0) {
    if (*str == '\\') {
      str++;        /* skip the slash */
      len--;
      if (len > 0) {
        if (*str == '0') {
          *out++='\0';
          str++;
        } else {
          *out++ = *str++;  /* preserve the next character */
        }
        len--;
      }
    } else {
      *out++ = *str++;
      len--;
    }
  }

  return out;
}

#ifdef __SSE2__
PHPAPI void php_stripslashes(zend_string *str)
{
  const char *s = ZSTR_VAL(str);
  char *t = ZSTR_VAL(str);
  size_t l = ZSTR_LEN(str);

  if (l > 15) {
    const __m128i slash = _mm_set1_epi8('\\');

    do {
      __m128i in = _mm_loadu_si128((__m128i *)s);
      __m128i any_slash = _mm_cmpeq_epi8(in, slash);
      uint32_t res = _mm_movemask_epi8(any_slash);

      if (res) {
        int i, n = zend_ulong_ntz(res);
        const char *e = s + 15;
        l -= n;
        for (i = 0; i < n; i++) {
          *t++ = *s++;
        }
        for (; s < e; s++) {
          if (*s == '\\') {
            s++;
            l--;
            if (*s == '0') {
              *t = '\0';
            } else {
              *t = *s;
            }
          } else {
            *t = *s;
          }
          t++;
          l--;
        }
      } else {
        _mm_storeu_si128((__m128i *)t, in);
        s += 16;
        t += 16;
        l -= 16;
      }
    } while (l > 15);
  }

  t = php_stripslashes_impl(s, t, l);
  if (t != (ZSTR_VAL(str) + ZSTR_LEN(str))) {
    ZSTR_LEN(str) = t - ZSTR_VAL(str);
    ZSTR_VAL(str)[ZSTR_LEN(str)] = '\0';
  }
}
#else
PHPAPI void php_stripslashes(zend_string *str)
{
  const char *t = php_stripslashes_impl(ZSTR_VAL(str), ZSTR_VAL(str), ZSTR_LEN(str));
  if (t != (ZSTR_VAL(str) + ZSTR_LEN(str))) {
    ZSTR_LEN(str) = t - ZSTR_VAL(str);
    ZSTR_VAL(str)[ZSTR_LEN(str)] = '\0';
  }
}
#endif
/* }}} */

#define _HEB_BLOCK_TYPE_ENG 1
#define _HEB_BLOCK_TYPE_HEB 2
#define isheb(c)      (((((unsigned char) c) >= 224) && (((unsigned char) c) <= 250)) ? 1 : 0)
#define _isblank(c)   (((((unsigned char) c) == ' '  || ((unsigned char) c) == '\t')) ? 1 : 0)
#define _isnewline(c) (((((unsigned char) c) == '\n' || ((unsigned char) c) == '\r')) ? 1 : 0)

/* {{{ php_str_replace_in_subject */
static zend_long php_str_replace_in_subject(
  zend_string *search_str, HashTable *search_ht, zend_string *replace_str, HashTable *replace_ht,
  zend_string *subject_str, zval *result, bool case_sensitivity
) {
  zval    *search_entry;
  zend_string *tmp_result;
  char    *replace_value = NULL;
  size_t     replace_len = 0;
  zend_long  replace_count = 0;
  zend_string *lc_subject_str = NULL;
  uint32_t     replace_idx;

  if (ZSTR_LEN(subject_str) == 0) {
    ZVAL_EMPTY_STRING(result);
    return 0;
  }

  /* If search is an array */
  if (search_ht) {
    /* Duplicate subject string for repeated replacement */
    zend_string_addref(subject_str);

    if (replace_ht) {
      replace_idx = 0;
    } else {
      /* Set replacement value to the passed one */
      replace_value = ZSTR_VAL(replace_str);
      replace_len = ZSTR_LEN(replace_str);
    }

    /* For each entry in the search array, get the entry */
    ZEND_HASH_FOREACH_VAL(search_ht, search_entry) {
      /* Make sure we're dealing with strings. */
      zend_string *tmp_search_str;
      zend_string *search_str = zval_get_tmp_string(search_entry, &tmp_search_str);
      zend_string *replace_entry_str, *tmp_replace_entry_str = NULL;

      /* If replace is an array. */
      if (replace_ht) {
        /* Get current entry */
        zval *replace_entry = NULL;
        if (HT_IS_PACKED(replace_ht)) {
          while (replace_idx < replace_ht->nNumUsed) {
            replace_entry = &replace_ht->arPacked[replace_idx];
            if (Z_TYPE_P(replace_entry) != IS_UNDEF) {
              break;
            }
            replace_idx++;
          }
        } else {
          while (replace_idx < replace_ht->nNumUsed) {
            replace_entry = &replace_ht->arData[replace_idx].val;
            if (Z_TYPE_P(replace_entry) != IS_UNDEF) {
              break;
            }
            replace_idx++;
          }
        }
        if (replace_idx < replace_ht->nNumUsed) {
          /* Make sure we're dealing with strings. */
          replace_entry_str = zval_get_tmp_string(replace_entry, &tmp_replace_entry_str);

          /* Set replacement value to the one we got from array */
          replace_value = ZSTR_VAL(replace_entry_str);
          replace_len = ZSTR_LEN(replace_entry_str);

          replace_idx++;
        } else {
          /* We've run out of replacement strings, so use an empty one. */
          replace_value = "";
          replace_len = 0;
        }
      }

      if (ZSTR_LEN(search_str) == 1) {
        zend_long old_replace_count = replace_count;

        tmp_result = php_char_to_str_ex(subject_str,
                ZSTR_VAL(search_str)[0],
                replace_value,
                replace_len,
                case_sensitivity,
                &replace_count);
        if (lc_subject_str && replace_count != old_replace_count) {
          zend_string_release_ex(lc_subject_str, 0);
          lc_subject_str = NULL;
        }
      } else if (ZSTR_LEN(search_str) > 1) {
        if (case_sensitivity) {
          tmp_result = php_str_to_str_ex(subject_str,
              ZSTR_VAL(search_str), ZSTR_LEN(search_str),
              replace_value, replace_len, &replace_count);
        } else {
          zend_long old_replace_count = replace_count;

          if (!lc_subject_str) {
            lc_subject_str = zend_string_tolower(subject_str);
          }
          tmp_result = php_str_to_str_i_ex(subject_str, ZSTR_VAL(lc_subject_str),
              search_str, replace_value, replace_len, &replace_count);
          if (replace_count != old_replace_count) {
            zend_string_release_ex(lc_subject_str, 0);
            lc_subject_str = NULL;
          }
        }
      } else {
        zend_tmp_string_release(tmp_search_str);
        zend_tmp_string_release(tmp_replace_entry_str);
        continue;
      }

      zend_tmp_string_release(tmp_search_str);
      zend_tmp_string_release(tmp_replace_entry_str);

      if (UNEXPECTED(!tmp_result)) {
        zend_string_release_ex(subject_str, 0);
        ZVAL_EMPTY_STRING(result);
        if (lc_subject_str) {
          zend_string_release_ex(lc_subject_str, 0);
        }
        return replace_count;
      }

      if (subject_str == tmp_result) {
        zend_string_delref(subject_str);
      } else {
        zend_string_release_ex(subject_str, 0);
        subject_str = tmp_result;
        if (ZSTR_LEN(subject_str) == 0) {
          zend_string_release_ex(subject_str, 0);
          ZVAL_EMPTY_STRING(result);
          if (lc_subject_str) {
            zend_string_release_ex(lc_subject_str, 0);
          }
          return replace_count;
        }
      }
    } ZEND_HASH_FOREACH_END();
    ZVAL_STR(result, subject_str);
    if (lc_subject_str) {
      zend_string_release_ex(lc_subject_str, 0);
    }
  } else {
    ZEND_ASSERT(search_str);
    if (ZSTR_LEN(search_str) == 1) {
      tmp_result = php_char_to_str_ex(subject_str,
              ZSTR_VAL(search_str)[0],
              ZSTR_VAL(replace_str),
              ZSTR_LEN(replace_str),
              case_sensitivity,
              &replace_count);
      if (UNEXPECTED(!tmp_result)) {
        ZVAL_EMPTY_STRING(result);
        return replace_count;
      }
      ZVAL_STR(result, tmp_result);
    } else if (ZSTR_LEN(search_str) > 1) {
      if (case_sensitivity) {
        tmp_result = php_str_to_str_ex(subject_str,
            ZSTR_VAL(search_str), ZSTR_LEN(search_str),
            ZSTR_VAL(replace_str), ZSTR_LEN(replace_str), &replace_count);
      } else {
        lc_subject_str = zend_string_tolower(subject_str);
        tmp_result = php_str_to_str_i_ex(subject_str, ZSTR_VAL(lc_subject_str),
            search_str, ZSTR_VAL(replace_str), ZSTR_LEN(replace_str), &replace_count);
        zend_string_release_ex(lc_subject_str, 0);
      }
      if (UNEXPECTED(!tmp_result)) {
        ZVAL_EMPTY_STRING(result);
        return replace_count;
      }
      ZVAL_STR(result, tmp_result);
    } else {
      ZVAL_STR_COPY(result, subject_str);
    }
  }
  return replace_count;
}
/* }}} */

static void _php_str_replace_common(
  zval *return_value,
  HashTable *search_ht, zend_string *search_str,
  HashTable *replace_ht, zend_string *replace_str,
  HashTable *subject_ht, zend_string *subject_str,
  zval *zcount,
  bool case_sensitivity
) {
  zval *subject_entry;
  zval result;
  zend_string *string_key;
  zend_ulong num_key;
  zend_long count = 0;

  /* Make sure we're dealing with strings and do the replacement. */
  if (search_str && replace_ht) {
    zend_argument_type_error(2, "must be of type string when argument #1 ($search) is a string");
    RETURN_THROWS();
  }

  /* Converting an element may call __toString(), which can destroy the arrays. */
  if (search_ht) {
    GC_TRY_ADDREF(search_ht);
  }
  if (replace_ht) {
    GC_TRY_ADDREF(replace_ht);
  }
  if (subject_ht) {
    GC_TRY_ADDREF(subject_ht);
  }

  /* if subject is an array */
  if (subject_ht) {
    array_init(return_value);

    /* For each subject entry, convert it to string, then perform replacement
       and add the result to the return_value array. */
    ZEND_HASH_FOREACH_KEY_VAL(subject_ht, num_key, string_key, subject_entry) {
      zend_string *tmp_subject_str;
      ZVAL_DEREF(subject_entry);
      subject_str = zval_get_tmp_string(subject_entry, &tmp_subject_str);
      count += php_str_replace_in_subject(search_str, search_ht, replace_str, replace_ht, subject_str, &result, case_sensitivity);
      zend_tmp_string_release(tmp_subject_str);

      /* Add to return array */
      if (string_key) {
        zend_hash_add_new(Z_ARRVAL_P(return_value), string_key, &result);
      } else {
        zend_hash_index_add_new(Z_ARRVAL_P(return_value), num_key, &result);
      }
    } ZEND_HASH_FOREACH_END();
  } else { /* if subject is not an array */
    count = php_str_replace_in_subject(search_str, search_ht, replace_str, replace_ht, subject_str, return_value, case_sensitivity);
  }
  if (zcount) {
    ZEND_TRY_ASSIGN_REF_LONG(zcount, count);
  }

  if (search_ht) {
    GC_TRY_DTOR_NO_REF(search_ht);
  }
  if (replace_ht) {
    GC_TRY_DTOR_NO_REF(replace_ht);
  }
  if (subject_ht) {
    GC_TRY_DTOR_NO_REF(subject_ht);
  }
}

/* {{{ php_str_replace_common */
static void php_str_replace_common(INTERNAL_FUNCTION_PARAMETERS, bool case_sensitivity)
{
  zend_string *search_str;
  HashTable *search_ht;
  zend_string *replace_str;
  HashTable *replace_ht;
  zend_string *subject_str;
  HashTable *subject_ht;
  zval *zcount = NULL;

  ZEND_PARSE_PARAMETERS_START(3, 4)
    Z_PARAM_ARRAY_HT_OR_STR(search_ht, search_str)
    Z_PARAM_ARRAY_HT_OR_STR(replace_ht, replace_str)
    Z_PARAM_ARRAY_HT_OR_STR(subject_ht, subject_str)
    Z_PARAM_OPTIONAL
    Z_PARAM_ZVAL(zcount)
  ZEND_PARSE_PARAMETERS_END();

  _php_str_replace_common(return_value, search_ht, search_str, replace_ht, replace_str, subject_ht, subject_str, zcount, case_sensitivity);
}
/* }}} */

/* {{{ Replaces all occurrences of search in haystack with replace */
PHP_FUNCTION(str_replace)
{
  php_str_replace_common(INTERNAL_FUNCTION_PARAM_PASSTHRU, 1);
}
/* }}} */

ZEND_FRAMELESS_FUNCTION(str_replace, 3)
{
  zend_string *search_str, *replace_str, *subject_str;
  HashTable *search_ht, *replace_ht, *subject_ht;
  zval search_tmp, replace_tmp, subject_tmp;

  Z_FLF_PARAM_ARRAY_HT_OR_STR(1, search_ht, search_str, search_tmp);
  Z_FLF_PARAM_ARRAY_HT_OR_STR(2, replace_ht, replace_str, replace_tmp);
  Z_FLF_PARAM_ARRAY_HT_OR_STR(3, subject_ht, subject_str, subject_tmp);

  _php_str_replace_common(return_value, search_ht, search_str, replace_ht, replace_str, subject_ht, subject_str, /* zcount */ NULL, /* case_sensitivity */ true);

flf_clean:;
  Z_FLF_PARAM_FREE_STR(1, search_tmp);
  Z_FLF_PARAM_FREE_STR(2, replace_tmp);
  Z_FLF_PARAM_FREE_STR(3, subject_tmp);
}

/* {{{ Replaces all occurrences of search in haystack with replace / case-insensitive */
PHP_FUNCTION(str_ireplace)
{
  php_str_replace_common(INTERNAL_FUNCTION_PARAM_PASSTHRU, 0);
}
/* }}} */

/* {{{ Converts logical Hebrew text to visual text */
PHP_FUNCTION(hebrev)
{
  char *str, *heb_str, *target;
  const char *tmp;
  size_t block_start, block_end, block_type, i;
  zend_long max_chars=0, char_count;
  size_t begin, end, orig_begin;
  size_t str_len;
  zend_string *broken_str;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STRING(str, str_len)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(max_chars)
  ZEND_PARSE_PARAMETERS_END();

  if (str_len == 0) {
    RETURN_EMPTY_STRING();
  }

  tmp = str;
  block_start=block_end=0;

  heb_str = (char *) emalloc(str_len+1);
  target = heb_str+str_len;
  *target = 0;
  target--;

  if (isheb(*tmp)) {
    block_type = _HEB_BLOCK_TYPE_HEB;
  } else {
    block_type = _HEB_BLOCK_TYPE_ENG;
  }

  do {
    if (block_type == _HEB_BLOCK_TYPE_HEB) {
      while ((isheb((int)*(tmp+1)) || _isblank((int)*(tmp+1)) || ispunct((unsigned char)tmp[1]) || (int)*(tmp+1)=='\n' ) && block_end<str_len-1) {
        tmp++;
        block_end++;
      }
      for (i = block_start+1; i<= block_end+1; i++) {
        *target = str[i-1];
        switch (*target) {
          case '(':
            *target = ')';
            break;
          case ')':
            *target = '(';
            break;
          case '[':
            *target = ']';
            break;
          case ']':
            *target = '[';
            break;
          case '{':
            *target = '}';
            break;
          case '}':
            *target = '{';
            break;
          case '<':
            *target = '>';
            break;
          case '>':
            *target = '<';
            break;
          case '\\':
            *target = '/';
            break;
          case '/':
            *target = '\\';
            break;
          default:
            break;
        }
        target--;
      }
      block_type = _HEB_BLOCK_TYPE_ENG;
    } else {
      while (!isheb(*(tmp+1)) && (int)*(tmp+1)!='\n' && block_end < str_len-1) {
        tmp++;
        block_end++;
      }
      while ((_isblank((int)*tmp) || ispunct((unsigned char)*tmp)) && *tmp!='/' && *tmp!='-' && block_end > block_start) {
        tmp--;
        block_end--;
      }
      for (i = block_end+1; i >= block_start+1; i--) {
        *target = str[i-1];
        target--;
      }
      block_type = _HEB_BLOCK_TYPE_HEB;
    }
    block_start=block_end+1;
  } while (block_end < str_len-1);


  broken_str = zend_string_alloc(str_len, 0);
  begin = end = str_len-1;
  target = ZSTR_VAL(broken_str);

  while (1) {
    char_count=0;
    while ((!max_chars || (max_chars > 0 && char_count < max_chars)) && begin > 0) {
      char_count++;
      begin--;
      if (_isnewline(heb_str[begin])) {
        while (begin > 0 && _isnewline(heb_str[begin-1])) {
          begin--;
          char_count++;
        }
        break;
      }
    }
    if (max_chars >= 0 && char_count == max_chars) { /* try to avoid breaking words */
      size_t new_char_count=char_count, new_begin=begin;

      while (new_char_count > 0) {
        if (_isblank(heb_str[new_begin]) || _isnewline(heb_str[new_begin])) {
          break;
        }
        new_begin++;
        new_char_count--;
      }
      if (new_char_count > 0) {
        begin=new_begin;
      }
    }
    orig_begin=begin;

    if (_isblank(heb_str[begin])) {
      heb_str[begin]='\n';
    }
    while (begin <= end && _isnewline(heb_str[begin])) { /* skip leading newlines */
      begin++;
    }
    for (i = begin; i <= end; i++) { /* copy content */
      *target = heb_str[i];
      target++;
    }
    for (i = orig_begin; i <= end && _isnewline(heb_str[i]); i++) {
      *target = heb_str[i];
      target++;
    }
    begin=orig_begin;

    if (begin == 0) {
      *target = 0;
      break;
    }
    begin--;
    end=begin;
  }
  efree(heb_str);

  RETURN_NEW_STR(broken_str);
}
/* }}} */

/* {{{ Converts newlines to HTML line breaks */
PHP_FUNCTION(nl2br)
{
  /* in brief this inserts <br /> or <br> before matched regexp \n\r?|\r\n? */
  const char  *tmp, *end;
  zend_string *str;
  char *target;
  size_t  repl_cnt = 0;
  bool  is_xhtml = 1;
  zend_string *result;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_BOOL(is_xhtml)
  ZEND_PARSE_PARAMETERS_END();

  tmp = ZSTR_VAL(str);
  end = ZSTR_VAL(str) + ZSTR_LEN(str);

  /* it is really faster to scan twice and allocate mem once instead of scanning once
     and constantly reallocing */
  while (tmp < end) {
    if (*tmp == '\r') {
      if (*(tmp+1) == '\n') {
        tmp++;
      }
      repl_cnt++;
    } else if (*tmp == '\n') {
      if (*(tmp+1) == '\r') {
        tmp++;
      }
      repl_cnt++;
    }

    tmp++;
  }

  if (repl_cnt == 0) {
    RETURN_STR_COPY(str);
  }

  {
    size_t repl_len = is_xhtml ? (sizeof("<br />") - 1) : (sizeof("<br>") - 1);

    result = zend_string_safe_falloc(repl_cnt, repl_len, ZSTR_LEN(str), 0, /* fallible */ true);
    if (UNEXPECTED(!result)) {
      RETURN_OOM();
    }
    target = ZSTR_VAL(result);
  }

  tmp = ZSTR_VAL(str);
  while (tmp < end) {
    switch (*tmp) {
      case '\r':
      case '\n':
        *target++ = '<';
        *target++ = 'b';
        *target++ = 'r';

        if (is_xhtml) {
          *target++ = ' ';
          *target++ = '/';
        }

        *target++ = '>';

        if ((*tmp == '\r' && *(tmp+1) == '\n') || (*tmp == '\n' && *(tmp+1) == '\r')) {
          *target++ = *tmp++;
        }
        ZEND_FALLTHROUGH;
      default:
        *target++ = *tmp;
    }

    tmp++;
  }

  *target = '\0';

  RETURN_NEW_STR(result);
}
/* }}} */

/* {{{ Strips HTML and PHP tags from a string */
PHP_FUNCTION(strip_tags)
{
  zend_string *buf;
  zend_string *str;
  zend_string *allow_str = NULL;
  HashTable *allow_ht = NULL;
  const char *allowed_tags=NULL;
  size_t allowed_tags_len=0;
  smart_str tags_ss = {0};

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_ARRAY_HT_OR_STR_OR_NULL(allow_ht, allow_str)
  ZEND_PARSE_PARAMETERS_END();

  if (allow_ht) {
    zval *tmp;
    zend_string *tag;

    ZEND_HASH_FOREACH_VAL(allow_ht, tmp) {
      tag = zval_get_string(tmp);
      smart_str_appendc(&tags_ss, '<');
      smart_str_append(&tags_ss, tag);
      smart_str_appendc(&tags_ss, '>');
      zend_string_release(tag);
    } ZEND_HASH_FOREACH_END();
    if (tags_ss.s) {
      smart_str_0(&tags_ss);
      allowed_tags = ZSTR_VAL(tags_ss.s);
      allowed_tags_len = ZSTR_LEN(tags_ss.s);
    }
  } else if (allow_str) {
    allowed_tags = ZSTR_VAL(allow_str);
    allowed_tags_len = ZSTR_LEN(allow_str);
  }

  buf = zend_string_init(ZSTR_VAL(str), ZSTR_LEN(str), 0);
  ZSTR_LEN(buf) = php_strip_tags_ex(ZSTR_VAL(buf), ZSTR_LEN(str), allowed_tags, allowed_tags_len, 0);
  smart_str_free(&tags_ss);
  RETURN_NEW_STR(buf);
}
/* }}} */

static zend_string *try_setlocale_str(zend_long cat, zend_string *loc) {
  const char *retval;

  if (zend_string_equals_literal(loc, "0")) {
    loc = NULL;
  } else {
    if (ZSTR_LEN(loc) >= 255) {
      php_error_docref(NULL, E_WARNING, "Specified locale name is too long");
      return NULL;
    }
  }

# ifndef PHP_WIN32
  retval = setlocale(cat, loc ? ZSTR_VAL(loc) : NULL);
# else
  if (loc) {
    /* BC: don't try /^[a-z]{2}_[A-Z]{2}($|\..*)/ except for /^u[ks]_U[KS]$/ */
    char *locp = ZSTR_VAL(loc);
    if (ZSTR_LEN(loc) >= 5 && locp[2] == '_'
      && locp[0] >= 'a' && locp[0] <= 'z' && locp[1] >= 'a' && locp[1] <= 'z'
      && locp[3] >= 'A' && locp[3] <= 'Z' && locp[4] >= 'A' && locp[4] <= 'Z'
      && (locp[5] == '\0' || locp[5] == '.')
      && !(locp[0] == 'u' && (locp[1] == 'k' || locp[1] == 's')
        && locp[3] == 'U' && (locp[4] == 'K' || locp[4] == 'S')
        && locp[5] == '\0')
    ) {
      retval = NULL;
    } else {
      retval = setlocale(cat, ZSTR_VAL(loc));
    }
  } else {
    retval = setlocale(cat, NULL);
  }
# endif
  if (!retval) {
    return NULL;
  }

  if (loc) {
    /* Remember if locale was changed */
    size_t len = strlen(retval);

    BG(locale_changed) = 1;
    if (cat == LC_CTYPE || cat == LC_ALL) {
      zend_update_current_locale();
      if (BG(ctype_string)) {
        zend_string_release_ex(BG(ctype_string), 0);
      }
      if (len == 1 && *retval == 'C') {
        /* C locale is represented as NULL. */
        BG(ctype_string) = NULL;
        return ZSTR_CHAR('C');
      } else if (zend_string_equals_cstr(loc, retval, len)) {
        BG(ctype_string) = zend_string_copy(loc);
        return zend_string_copy(BG(ctype_string));
      } else {
        BG(ctype_string) = zend_string_init(retval, len, 0);
        return zend_string_copy(BG(ctype_string));
      }
    } else if (zend_string_equals_cstr(loc, retval, len)) {
      return zend_string_copy(loc);
    }
  }
  return zend_string_init(retval, strlen(retval), 0);
}

static zend_string *try_setlocale_zval(zend_long cat, zval *loc_zv) {
  zend_string *tmp_loc_str;
  zend_string *loc_str = zval_try_get_tmp_string(loc_zv, &tmp_loc_str);
  if (UNEXPECTED(loc_str == NULL)) {
    return NULL;
  }
  if (zend_str_has_nul_byte(loc_str)) {
    zend_argument_value_error(2, "must not contain any null bytes");
    zend_tmp_string_release(tmp_loc_str);
    return NULL;
  }
  zend_string *result = try_setlocale_str(cat, loc_str);
  zend_tmp_string_release(tmp_loc_str);
  return result;
}

/* {{{ Set locale information */
PHP_FUNCTION(setlocale)
{
  zend_long cat;
  zval *args = NULL;
  uint32_t num_args;
  ALLOCA_FLAG(use_heap);

  ZEND_PARSE_PARAMETERS_START(2, -1)
    Z_PARAM_LONG(cat)
    Z_PARAM_VARIADIC('+', args, num_args)
  ZEND_PARSE_PARAMETERS_END();

  zend_string **strings = do_alloca(sizeof(zend_string *) * num_args, use_heap);

  for (uint32_t i = 0; i < num_args; i++) {
    if (Z_TYPE(args[i]) == IS_ARRAY) {
      if (UNEXPECTED(i != 0)) {
        zend_wrong_parameter_type_error(i + 2, Z_EXPECTED_STRING_OR_NULL, &args[i]);
        goto out;
      }
      if (UNEXPECTED(num_args > 1)) {
        zend_argument_count_error(
          "setlocale() expects exactly 2 arguments when argument #2 ($locales) is an array, %d given",
          ZEND_NUM_ARGS());
        goto out;
      }
      break;
    }
    if (UNEXPECTED(!zend_parse_arg_path_str(&args[i], &strings[i], true, i + 2))) {
      zend_wrong_parameter_type_error(
        i + 2,
        Z_TYPE(args[i]) == IS_STRING
          ? Z_EXPECTED_PATH
          : (i == 0 ? Z_EXPECTED_ARRAY_OR_STRING_OR_NULL : Z_EXPECTED_STRING_OR_NULL),
        &args[i]);
      goto out;
    }
  }

  for (uint32_t i = 0; i < num_args; i++) {
    zend_string *result;
    if (Z_TYPE(args[i]) == IS_ARRAY) {
      zval *elem;
      ZEND_HASH_FOREACH_VAL(Z_ARRVAL(args[i]), elem) {
        result = try_setlocale_zval(cat, elem);
        if (EG(exception)) {
          goto out;
        }
        if (result) {
          RETVAL_STR(result);
          goto out;
        }
      } ZEND_HASH_FOREACH_END();
      continue;
    } else if (Z_ISNULL(args[i])) {
      result = try_setlocale_str(cat, ZSTR_EMPTY_ALLOC());
    } else {
      result = try_setlocale_str(cat, strings[i]);
    }
    if (EG(exception)) {
      goto out;
    }
    if (result) {
      RETVAL_STR(result);
      goto out;
    }
  }

  RETVAL_FALSE;

out:
  free_alloca(strings, use_heap);
}
/* }}} */

/* {{{ Parses GET/POST/COOKIE data and sets global variables */
PHP_FUNCTION(parse_str)
{
  char *arg;
  zval *arrayArg = NULL;
  char *res = NULL;
  size_t arglen;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_PATH(arg, arglen)
    Z_PARAM_ZVAL(arrayArg)
  ZEND_PARSE_PARAMETERS_END();

  arrayArg = zend_try_array_init(arrayArg);
  if (!arrayArg) {
    RETURN_THROWS();
  }

  res = estrndup(arg, arglen);
  sapi_module.treat_data(PARSE_STRING, res, arrayArg);
}
/* }}} */

#define PHP_TAG_BUF_SIZE 1023

/* {{{ php_tag_find
 *
 * Check if tag is in a set of tags
 *
 * states:
 *
 * 0 start tag
 * 1 first non-whitespace char seen
 */
static bool php_tag_find(char *tag, size_t len, const char *set) {
  char c, *n;
  const char *t;
  int state = 0;
  bool done = false;
  char *norm;

  if (len == 0) {
    return false;
  }

  norm = emalloc(len+1);

  n = norm;
  t = tag;
  c = zend_tolower_ascii(*t);
  /*
     normalize the tag removing leading and trailing whitespace
     and turn any <a whatever...> into just <a> and any </tag>
     into <tag>
  */
  while (!done) {
    switch (c) {
      case '<':
        *(n++) = c;
        break;
      case '>':
        done = true;
        break;
      default:
        if (!isspace((unsigned char)c)) {
          if (state == 0) {
            state=1;
          }
          if (c != '/' || (*(t-1) != '<' && *(t+1) != '>')) {
            *(n++) = c;
          }
        } else {
          if (state == 1)
            done = true;
        }
        break;
    }
    c = zend_tolower_ascii(*(++t));
  }
  *(n++) = '>';
  *n = '\0';
  if (strstr(set, norm)) {
    done = true;
  } else {
    done = false;
  }
  efree(norm);
  return done;
}
/* }}} */

PHPAPI size_t php_strip_tags(char *rbuf, size_t len, const char *allow, size_t allow_len) /* {{{ */
{
  return php_strip_tags_ex(rbuf, len, allow, allow_len, false);
}
/* }}} */

/* {{{ php_strip_tags

  A simple little state-machine to strip out html and php tags

  State 0 is the output state, State 1 means we are inside a
  normal html tag and state 2 means we are inside a php tag.

  The state variable is passed in to allow a function like fgetss
  to maintain state across calls to the function.

  lc holds the last significant character read and br is a bracket
  counter.

  When an allow string is passed in we keep track of the string
  in state 1 and when the tag is closed check it against the
  allow string to see if we should allow it.

  swm: Added ability to strip <?xml tags without assuming it PHP
  code.
*/
PHPAPI size_t php_strip_tags_ex(char *rbuf, size_t len, const char *allow, size_t allow_len, bool allow_tag_spaces)
{
  char *tbuf, *tp, *rp, c, lc;
  const char *buf, *p, *end;
  int br, depth=0, in_q = 0;
  uint8_t state = 0;
  size_t pos;
  char *allow_free = NULL;
  char is_xml = 0;

  buf = estrndup(rbuf, len);
  end = buf + len;
  lc = '\0';
  p = buf;
  rp = rbuf;
  br = 0;
  if (allow) {
    allow_free = zend_str_tolower_dup_ex(allow, allow_len);
    allow = allow_free ? allow_free : allow;
    tbuf = emalloc(PHP_TAG_BUF_SIZE + 1);
    tp = tbuf;
  } else {
    tbuf = tp = NULL;
  }

state_0:
  if (p >= end) {
    goto finish;
  }
  c = *p;
  switch (c) {
    case '\0':
      break;
    case '<':
      if (in_q) {
        break;
      }
      if (isspace((unsigned char)p[1]) && !allow_tag_spaces) {
        *(rp++) = c;
        break;
      }
      lc = '<';
      state = 1;
      if (allow) {
        if (tp - tbuf >= PHP_TAG_BUF_SIZE) {
          pos = tp - tbuf;
          tbuf = erealloc(tbuf, (tp - tbuf) + PHP_TAG_BUF_SIZE + 1);
          tp = tbuf + pos;
        }
        *(tp++) = '<';
      }
      p++;
      goto state_1;
    case '>':
      if (depth) {
        depth--;
        break;
      }

      if (in_q) {
        break;
      }

      *(rp++) = c;
      break;
    default:
      *(rp++) = c;
      break;
  }
  p++;
  goto state_0;

state_1:
  if (p >= end) {
    goto finish;
  }
  c = *p;
  switch (c) {
    case '\0':
      break;
    case '<':
      if (in_q) {
        break;
      }
      if (isspace((unsigned char)p[1]) && !allow_tag_spaces) {
        goto reg_char_1;
      }
      depth++;
      break;
    case '>':
      if (depth) {
        depth--;
        break;
      }
      if (in_q) {
        break;
      }

      lc = '>';
      if (is_xml && p >= buf + 1 && *(p -1) == '-') {
        break;
      }
      in_q = state = is_xml = 0;
      if (allow) {
        if (tp - tbuf >= PHP_TAG_BUF_SIZE) {
          pos = tp - tbuf;
          tbuf = erealloc(tbuf, (tp - tbuf) + PHP_TAG_BUF_SIZE + 1);
          tp = tbuf + pos;
        }
        *(tp++) = '>';
        *tp='\0';
        if (php_tag_find(tbuf, tp-tbuf, allow)) {
          rp = zend_mempcpy(rp, tbuf, tp - tbuf);
        }
        tp = tbuf;
      }
      p++;
      goto state_0;
    case '"':
    case '\'':
      if (p != buf && (!in_q || *p == in_q)) {
        if (in_q) {
          in_q = 0;
        } else {
          in_q = *p;
        }
      }
      goto reg_char_1;
    case '!':
      /* JavaScript & Other HTML scripting languages */
      if (p >= buf + 1 && *(p-1) == '<') {
        state = 3;
        lc = c;
        p++;
        goto state_3;
      } else {
        goto reg_char_1;
      }
      break;
    case '?':
      if (p >= buf + 1 && *(p-1) == '<') {
        br=0;
        state = 2;
        p++;
        goto state_2;
      } else {
        goto reg_char_1;
      }
      break;
    default:
reg_char_1:
      if (allow) {
        if (tp - tbuf >= PHP_TAG_BUF_SIZE) {
          pos = tp - tbuf;
          tbuf = erealloc(tbuf, (tp - tbuf) + PHP_TAG_BUF_SIZE + 1);
          tp = tbuf + pos;
        }
        *(tp++) = c;
      }
      break;
  }
  p++;
  goto state_1;

state_2:
  if (p >= end) {
    goto finish;
  }
  c = *p;
  switch (c) {
    case '(':
      if (lc != '"' && lc != '\'') {
        lc = '(';
        br++;
      }
      break;
    case ')':
      if (lc != '"' && lc != '\'') {
        lc = ')';
        br--;
      }
      break;
    case '>':
      if (depth) {
        depth--;
        break;
      }
      if (in_q) {
        break;
      }

      if (!br && p >= buf + 1 && lc != '\"' && *(p-1) == '?') {
        in_q = state = 0;
        tp = tbuf;
        p++;
        goto state_0;
      }
      break;
    case '"':
    case '\'':
      if (p >= buf + 1 && *(p-1) != '\\') {
        if (lc == c) {
          lc = '\0';
        } else if (lc != '\\') {
          lc = c;
        }
        if (p != buf && (!in_q || *p == in_q)) {
          if (in_q) {
            in_q = 0;
          } else {
            in_q = *p;
          }
        }
      }
      break;
    case 'l':
    case 'L':
      /* swm: If we encounter '<?xml' then we shouldn't be in
       * state == 2 (PHP). Switch back to HTML.
       */
      if (state == 2 && p > buf+4
             && (*(p-1) == 'm' || *(p-1) == 'M')
             && (*(p-2) == 'x' || *(p-2) == 'X')
             && *(p-3) == '?'
             && *(p-4) == '<') {
        state = 1; is_xml=1;
        p++;
        goto state_1;
      }
      break;
    default:
      break;
  }
  p++;
  goto state_2;

state_3:
  if (p >= end) {
    goto finish;
  }
  c = *p;
  switch (c) {
    case '>':
      if (depth) {
        depth--;
        break;
      }
      if (in_q) {
        break;
      }
      in_q = state = 0;
      tp = tbuf;
      p++;
      goto state_0;
    case '"':
    case '\'':
      if (p != buf && *(p-1) != '\\' && (!in_q || *p == in_q)) {
        if (in_q) {
          in_q = 0;
        } else {
          in_q = *p;
        }
      }
      break;
    case '-':
      if (p >= buf + 2 && *(p-1) == '-' && *(p-2) == '!') {
        state = 4;
        p++;
        goto state_4;
      }
      break;
    case 'E':
    case 'e':
      /* !DOCTYPE exception */
      if (p > buf+6
           && (*(p-1) == 'p' || *(p-1) == 'P')
           && (*(p-2) == 'y' || *(p-2) == 'Y')
           && (*(p-3) == 't' || *(p-3) == 'T')
           && (*(p-4) == 'c' || *(p-4) == 'C')
           && (*(p-5) == 'o' || *(p-5) == 'O')
           && (*(p-6) == 'd' || *(p-6) == 'D')) {
        state = 1;
        p++;
        goto state_1;
      }
      break;
    default:
      break;
  }
  p++;
  goto state_3;

state_4:
  while (p < end) {
    c = *p;
    if (c == '>' && !in_q) {
      if (p >= buf + 2 && *(p-1) == '-' && *(p-2) == '-') {
        in_q = state = 0;
        tp = tbuf;
        p++;
        goto state_0;
      }
    }
    p++;
  }

finish:
  if (rp < rbuf + len) {
    *rp = '\0';
  }
  efree((void *)buf);
  if (tbuf) {
    efree(tbuf);
  }
  if (allow_free) {
    efree(allow_free);
  }

  return (size_t)(rp - rbuf);
}
/* }}} */

/* {{{ Parse a CSV string into an array */
PHP_FUNCTION(str_getcsv)
{
  zend_string *str;
  char delimiter = ',', enclosure = '"';
  char *delimiter_str = NULL, *enclosure_str = NULL;
  size_t delimiter_str_len = 0, enclosure_str_len = 0;
  zend_string *escape_str = NULL;

  ZEND_PARSE_PARAMETERS_START(1, 4)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_STRING(delimiter_str, delimiter_str_len)
    Z_PARAM_STRING(enclosure_str, enclosure_str_len)
    Z_PARAM_STR(escape_str)
  ZEND_PARSE_PARAMETERS_END();

  if (delimiter_str != NULL) {
    /* Make sure that there is at least one character in string */
    if (delimiter_str_len != 1) {
      zend_argument_value_error(2, "must be a single character");
      RETURN_THROWS();
    }
    /* use first character from string */
    delimiter = delimiter_str[0];
  }
  if (enclosure_str != NULL) {
    if (enclosure_str_len != 1) {
      zend_argument_value_error(3, "must be a single character");
      RETURN_THROWS();
    }
    /* use first character from string */
    enclosure = enclosure_str[0];
  }

  int escape_char = php_csv_handle_escape_argument(escape_str, 4);
  if (escape_char == PHP_CSV_ESCAPE_ERROR) {
    RETURN_THROWS();
  }

  HashTable *values = php_fgetcsv(NULL, delimiter, enclosure, escape_char, ZSTR_LEN(str), ZSTR_VAL(str));
  if (values == NULL) {
    values = php_bc_fgetcsv_empty_line();
  }
  RETURN_ARR(values);
}
/* }}} */

/* {{{ Returns the input string repeat mult times */
PHP_FUNCTION(str_repeat)
{
  zend_string   *input_str;   /* Input string */
  zend_long     mult;     /* Multiplier */
  zend_string *result;    /* Resulting string */
  size_t    result_len;   /* Length of the resulting string */

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(input_str)
    Z_PARAM_LONG(mult)
  ZEND_PARSE_PARAMETERS_END();

  if (mult < 0) {
    zend_argument_value_error(2, "must be greater than or equal to 0");
    RETURN_THROWS();
  }

  /* Don't waste our time if it's empty */
  /* ... or if the multiplier is zero */
  if (ZSTR_LEN(input_str) == 0 || mult == 0)
    RETURN_EMPTY_STRING();

  if (mult == 1) {
    RETURN_STR_COPY(input_str);
  }

  /* Initialize the result string */
  result = zend_string_safe_falloc(ZSTR_LEN(input_str), mult, 0, /*persistent */ false, /* fallible */ true);
  if (UNEXPECTED(!result)) {
    RETURN_OOM();
  }

  result_len = ZSTR_LEN(input_str) * mult;
  ZSTR_COPY_CONCAT_PROPERTIES(result, input_str);

  /* Heavy optimization for situations where input string is 1 byte long */
  if (ZSTR_LEN(input_str) == 1) {
    memset(ZSTR_VAL(result), *ZSTR_VAL(input_str), mult);
  } else {
    const char *s, *ee;
    char *e;
    ptrdiff_t l=0;
    memcpy(ZSTR_VAL(result), ZSTR_VAL(input_str), ZSTR_LEN(input_str));
    s = ZSTR_VAL(result);
    e = ZSTR_VAL(result) + ZSTR_LEN(input_str);
    ee = ZSTR_VAL(result) + result_len;

    while (e<ee) {
      l = (e-s) < (ee-e) ? (e-s) : (ee-e);
      /* src [s, s+l) and dst [e, e+l) cannot overlap: l <= e-s */
      memcpy(e, s, l);
      e += l;
    }
  }

  ZSTR_VAL(result)[result_len] = '\0';

  RETURN_NEW_STR(result);
}
/* }}} */

/* {{{ Returns info about what characters are used in input */
PHP_FUNCTION(count_chars)
{
  zend_string *input;
  int chars[256];
  zend_long mymode=0;
  const unsigned char *buf;
  int inx;
  char retstr[256];
  size_t retlen=0;
  size_t tmp = 0;

  ZEND_PARSE_PARAMETERS_START(1, 2)
    Z_PARAM_STR(input)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(mymode)
  ZEND_PARSE_PARAMETERS_END();

  if (mymode < 0 || mymode > 4) {
    zend_argument_value_error(2, "must be between 0 and 4 (inclusive)");
    RETURN_THROWS();
  }

  buf = (const unsigned char *) ZSTR_VAL(input);
  memset((void*) chars, 0, sizeof(chars));

  while (tmp < ZSTR_LEN(input)) {
    chars[*buf]++;
    buf++;
    tmp++;
  }

  if (mymode < 3) {
    array_init(return_value);
  }

  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 (mymode == 3 || mymode == 4) {
    RETURN_STRINGL(retstr, retlen);
  }
}
/* }}} */

/* {{{ php_strnatcmp */
static void php_strnatcmp(INTERNAL_FUNCTION_PARAMETERS, bool is_case_insensitive)
{
  zend_string *s1, *s2;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(s1)
    Z_PARAM_STR(s2)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_LONG(strnatcmp_ex(ZSTR_VAL(s1), ZSTR_LEN(s1),
               ZSTR_VAL(s2), ZSTR_LEN(s2),
               is_case_insensitive));
}
/* }}} */

/* {{{ Returns the result of string comparison using 'natural' algorithm */
PHP_FUNCTION(strnatcmp)
{
  php_strnatcmp(INTERNAL_FUNCTION_PARAM_PASSTHRU, 0);
}
/* }}} */

/* {{{ Returns the result of case-insensitive string comparison using 'natural' algorithm */
PHP_FUNCTION(strnatcasecmp)
{
  php_strnatcmp(INTERNAL_FUNCTION_PARAM_PASSTHRU, 1);
}
/* }}} */

/* {{{ Returns numeric formatting information based on the current locale */
PHP_FUNCTION(localeconv)
{
  zval grouping, mon_grouping;
  size_t len, i;

  ZEND_PARSE_PARAMETERS_NONE();

  array_init(return_value);
  array_init(&grouping);
  array_init(&mon_grouping);

  {
    struct lconv currlocdata;

    localeconv_r( &currlocdata );

    /* Grab the grouping data out of the array */
    len = strlen(currlocdata.grouping);

    for (i = 0; i < len; i++) {
      add_index_long(&grouping, i, currlocdata.grouping[i]);
    }

    /* Grab the monetary grouping data out of the array */
    len = strlen(currlocdata.mon_grouping);

    for (i = 0; i < len; i++) {
      add_index_long(&mon_grouping, i, currlocdata.mon_grouping[i]);
    }

    add_assoc_string(return_value, "decimal_point",     currlocdata.decimal_point);
    add_assoc_string(return_value, "thousands_sep",     currlocdata.thousands_sep);
    add_assoc_string(return_value, "int_curr_symbol",   currlocdata.int_curr_symbol);
    add_assoc_string(return_value, "currency_symbol",   currlocdata.currency_symbol);
    add_assoc_string(return_value, "mon_decimal_point", currlocdata.mon_decimal_point);
    add_assoc_string(return_value, "mon_thousands_sep", currlocdata.mon_thousands_sep);
    add_assoc_string(return_value, "positive_sign",     currlocdata.positive_sign);
    add_assoc_string(return_value, "negative_sign",     currlocdata.negative_sign);
    add_assoc_long(  return_value, "int_frac_digits",   currlocdata.int_frac_digits);
    add_assoc_long(  return_value, "frac_digits",       currlocdata.frac_digits);
    add_assoc_long(  return_value, "p_cs_precedes",     currlocdata.p_cs_precedes);
    add_assoc_long(  return_value, "p_sep_by_space",    currlocdata.p_sep_by_space);
    add_assoc_long(  return_value, "n_cs_precedes",     currlocdata.n_cs_precedes);
    add_assoc_long(  return_value, "n_sep_by_space",    currlocdata.n_sep_by_space);
    add_assoc_long(  return_value, "p_sign_posn",       currlocdata.p_sign_posn);
    add_assoc_long(  return_value, "n_sign_posn",       currlocdata.n_sign_posn);
  }

  zend_hash_str_update(Z_ARRVAL_P(return_value), "grouping", sizeof("grouping")-1, &grouping);
  zend_hash_str_update(Z_ARRVAL_P(return_value), "mon_grouping", sizeof("mon_grouping")-1, &mon_grouping);
}
/* }}} */

/* {{{ Returns the number of times a substring occurs in the string */
PHP_FUNCTION(substr_count)
{
  char *haystack, *needle;
  zend_long offset = 0, length = 0;
  bool length_is_null = 1;
  zend_long count;
  size_t haystack_len, needle_len;
  const char *p, *endp;

  ZEND_PARSE_PARAMETERS_START(2, 4)
    Z_PARAM_STRING(haystack, haystack_len)
    Z_PARAM_STRING(needle, needle_len)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(offset)
    Z_PARAM_LONG_OR_NULL(length, length_is_null)
  ZEND_PARSE_PARAMETERS_END();

  if (needle_len == 0) {
    zend_argument_must_not_be_empty_error(2);
    RETURN_THROWS();
  }

  p = haystack;

  if (offset) {
    if (offset < 0) {
      offset += (zend_long)haystack_len;
    }
    if ((offset < 0) || ((size_t)offset > haystack_len)) {
      zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
      RETURN_THROWS();
    }
    p += offset;
    haystack_len -= offset;
  }

  if (!length_is_null) {
    if (length < 0) {
      length += haystack_len;
    }
    if (length < 0 || ((size_t)length > haystack_len)) {
      zend_argument_value_error(4, "must be contained in argument #1 ($haystack)");
      RETURN_THROWS();
    }
  } else {
    length = haystack_len;
  }

  if (needle_len == 1) {
    count = count_chars(p, length, needle[0]);
  } else {
    count = 0;
    endp = p + length;
    while ((p = (char*)php_memnstr(p, needle, needle_len, endp))) {
      p += needle_len;
      count++;
    }
  }

  RETURN_LONG(count);
}
/* }}} */

static void php_str_pad_fill(zend_string *result, size_t pad_chars, const char *pad_str, size_t pad_str_len) {
  char *p = ZSTR_VAL(result) + ZSTR_LEN(result);

  if (pad_str_len == 1) {
    memset(p, pad_str[0], pad_chars);
    ZSTR_LEN(result) += pad_chars;
    return;
  }

  const char *start = p;
  const char *end = p + pad_chars;
  size_t len = MIN(pad_str_len, pad_chars);
  p = zend_mempcpy(p, pad_str, len);

  /* Double the filled area on each iteration. */
  while (p < end) {
    len = MIN(p - start, end - p);
    p = zend_mempcpy(p, start, len);
  }

  ZSTR_LEN(result) += pad_chars;
}

/* {{{ Returns input string padded on the left or right to specified length with pad_string */
PHP_FUNCTION(str_pad)
{
  /* Input arguments */
  zend_string *input;       /* Input string */
  zend_long pad_length;     /* Length to pad to */

  /* Helper variables */
  size_t num_pad_chars;   /* Number of padding characters (total - input size) */
  char *pad_str = " "; /* Pointer to padding string */
  size_t pad_str_len = 1;
  zend_long   pad_type_val = PHP_STR_PAD_RIGHT; /* The padding type value */
  size_t left_pad=0, right_pad=0;
  zend_string *result = NULL; /* Resulting string */

  ZEND_PARSE_PARAMETERS_START(2, 4)
    Z_PARAM_STR(input)
    Z_PARAM_LONG(pad_length)
    Z_PARAM_OPTIONAL
    Z_PARAM_STRING(pad_str, pad_str_len)
    Z_PARAM_LONG(pad_type_val)
  ZEND_PARSE_PARAMETERS_END();

  /* If resulting string turns out to be shorter than input string,
     we simply copy the input and return. */
  if (pad_length < 0  || (size_t)pad_length <= ZSTR_LEN(input)) {
    RETURN_STR_COPY(input);
  }

  if (pad_str_len == 0) {
    zend_argument_must_not_be_empty_error(3);
    RETURN_THROWS();
  }

  if (pad_type_val < PHP_STR_PAD_LEFT || pad_type_val > PHP_STR_PAD_BOTH) {
    zend_argument_value_error(4, "must be STR_PAD_LEFT, STR_PAD_RIGHT, or STR_PAD_BOTH");
    RETURN_THROWS();
  }

  num_pad_chars = pad_length - ZSTR_LEN(input);
  result = zend_string_safe_falloc(1, ZSTR_LEN(input), num_pad_chars, 0, /* falllible */ true);
  if (UNEXPECTED(!result)) {
    RETURN_OOM();
  }
  ZSTR_LEN(result) = 0;

  /* We need to figure out the left/right padding lengths. */
  switch (pad_type_val) {
    case PHP_STR_PAD_RIGHT:
      left_pad = 0;
      right_pad = num_pad_chars;
      break;

    case PHP_STR_PAD_LEFT:
      left_pad = num_pad_chars;
      right_pad = 0;
      break;

    case PHP_STR_PAD_BOTH:
      left_pad = num_pad_chars / 2;
      right_pad = num_pad_chars - left_pad;
      break;
  }

  /* First we pad on the left. */
  if (left_pad > 0) {
    php_str_pad_fill(result, left_pad, pad_str, pad_str_len);
  }

  /* Then we copy the input string. */
  memcpy(ZSTR_VAL(result) + ZSTR_LEN(result), ZSTR_VAL(input), ZSTR_LEN(input));
  ZSTR_LEN(result) += ZSTR_LEN(input);

  /* Finally, we pad on the right. */
  if (right_pad > 0) {
    php_str_pad_fill(result, right_pad, pad_str, pad_str_len);
  }

  ZSTR_VAL(result)[ZSTR_LEN(result)] = '\0';

  RETURN_NEW_STR(result);
}
/* }}} */

/* {{{ Implements an ANSI C compatible sscanf */
PHP_FUNCTION(sscanf)
{
  zval *args = NULL;
  char *str, *format;
  size_t str_len, format_len;
  int result, num_args = 0;

  ZEND_PARSE_PARAMETERS_START(2, -1)
    Z_PARAM_STRING(str, str_len)
    Z_PARAM_STRING(format, format_len)
    Z_PARAM_VARIADIC('*', args, num_args)
  ZEND_PARSE_PARAMETERS_END();

  result = php_sscanf_internal(str, format, num_args, args, 0, return_value);

  if (SCAN_ERROR_WRONG_PARAM_COUNT == result) {
    zend_wrong_param_count();
    RETURN_THROWS();
  }
}
/* }}} */

/* static zend_string *php_str_rot13(zend_string *str) {{{ */
static zend_string *php_str_rot13(zend_string *str)
{
  zend_string *ret;
  const char *p, *e;
  char *target;

  if (UNEXPECTED(ZSTR_LEN(str) == 0)) {
    return ZSTR_EMPTY_ALLOC();
  }

  ret = zend_string_alloc(ZSTR_LEN(str), 0);

  p = ZSTR_VAL(str);
  e = p + ZSTR_LEN(str);
  target = ZSTR_VAL(ret);

#ifdef XSSE2
  if (e - p > 15) {
    const __m128i a_minus_1 = _mm_set1_epi8('a' - 1);
    const __m128i m_plus_1 = _mm_set1_epi8('m' + 1);
    const __m128i n_minus_1 = _mm_set1_epi8('n' - 1);
    const __m128i z_plus_1 = _mm_set1_epi8('z' + 1);
    const __m128i A_minus_1 = _mm_set1_epi8('A' - 1);
    const __m128i M_plus_1 = _mm_set1_epi8('M' + 1);
    const __m128i N_minus_1 = _mm_set1_epi8('N' - 1);
    const __m128i Z_plus_1 = _mm_set1_epi8('Z' + 1);
    const __m128i add = _mm_set1_epi8(13);
    const __m128i sub = _mm_set1_epi8(-13);

    do {
      __m128i in, gt, lt, cmp, delta;

      delta = _mm_setzero_si128();
      in = _mm_loadu_si128((__m128i *)p);

      gt = _mm_cmpgt_epi8(in, a_minus_1);
      lt = _mm_cmplt_epi8(in, m_plus_1);
      cmp = _mm_and_si128(lt, gt);
      cmp = _mm_and_si128(cmp, add);
      delta = _mm_or_si128(delta, cmp);

      gt = _mm_cmpgt_epi8(in, n_minus_1);
      lt = _mm_cmplt_epi8(in, z_plus_1);
      cmp = _mm_and_si128(lt, gt);
      cmp = _mm_and_si128(cmp, sub);
      delta = _mm_or_si128(delta, cmp);

      gt = _mm_cmpgt_epi8(in, A_minus_1);
      lt = _mm_cmplt_epi8(in, M_plus_1);
      cmp = _mm_and_si128(lt, gt);
      cmp = _mm_and_si128(cmp, add);
      delta = _mm_or_si128(delta, cmp);

      gt = _mm_cmpgt_epi8(in, N_minus_1);
      lt = _mm_cmplt_epi8(in, Z_plus_1);
      cmp = _mm_and_si128(lt, gt);
      cmp = _mm_and_si128(cmp, sub);
      delta = _mm_or_si128(delta, cmp);

      in = _mm_add_epi8(in, delta);
      _mm_storeu_si128((__m128i *)target, in);

      p += 16;
      target += 16;
    } while (e - p > 15);
  }
#endif

  while (p < e) {
    if (*p >= 'a' && *p <= 'z') {
      *target++ = 'a' + (((*p++ - 'a') + 13) % 26);
    } else if (*p >= 'A' && *p <= 'Z') {
      *target++ = 'A' + (((*p++ - 'A') + 13) % 26);
    } else {
      *target++ = *p++;
    }
  }

  *target = '\0';

  return ret;
}
/* }}} */

/* {{{ Perform the rot13 transform on a string */
PHP_FUNCTION(str_rot13)
{
  zend_string *arg;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(arg)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_STR(php_str_rot13(arg));
}
/* }}} */

/* {{{ php_binary_string_shuffle */
PHPAPI bool php_binary_string_shuffle(php_random_algo_with_state engine, char *str, zend_long len) /* {{{ */
{
  const php_random_algo *algo = engine.algo;
  void *state = engine.state;

  int64_t n_elems, rnd_idx, n_left;
  char temp;

  /* The implementation is stolen from array_data_shuffle       */
  /* Thus the characteristics of the randomization are the same */
  n_elems = len;

  if (n_elems <= 1) {
    return true;
  }

  n_left = n_elems;

  while (--n_left) {
    rnd_idx = algo->range(state, 0, n_left);
    if (EG(exception)) {
      return false;
    }
    if (rnd_idx != n_left) {
      temp = str[n_left];
      str[n_left] = str[rnd_idx];
      str[rnd_idx] = temp;
    }
  }

  return true;
}
/* }}} */

/* {{{ Shuffles string. One permutation of all possible is created */
PHP_FUNCTION(str_shuffle)
{
  zend_string *arg;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STR(arg)
  ZEND_PARSE_PARAMETERS_END();

  RETVAL_STRINGL(ZSTR_VAL(arg), ZSTR_LEN(arg));
  if (Z_STRLEN_P(return_value) > 1) {
    php_binary_string_shuffle(
      php_random_default_engine(),
      Z_STRVAL_P(return_value),
      Z_STRLEN_P(return_value)
    );
  }
}
/* }}} */

/* {{{ Counts the number of words inside a string. If format of 1 is specified,
    then the function will return an array containing all the words
    found inside the string. If format of 2 is specified, then the function
    will return an associated array where the position of the word is the key
    and the word itself is the value.
    For the purpose of this function, 'word' is defined as a locale dependent
    string containing alphabetic characters, which also may contain, but not start
    with "'" and "-" characters.
*/
PHP_FUNCTION(str_word_count)
{
  zend_string *str;
  char *char_list = NULL, ch[256];
  const char *p, *e, *s;
  size_t char_list_len = 0, word_count = 0;
  zend_long type = 0;

  ZEND_PARSE_PARAMETERS_START(1, 3)
    Z_PARAM_STR(str)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG(type)
    Z_PARAM_STRING_OR_NULL(char_list, char_list_len)
  ZEND_PARSE_PARAMETERS_END();

  switch(type) {
    case 1:
    case 2:
      array_init(return_value);
      if (!ZSTR_LEN(str)) {
        return;
      }
      break;
    case 0:
      if (!ZSTR_LEN(str)) {
        RETURN_LONG(0);
      }
      /* nothing to be done */
      break;
    default:
      zend_argument_value_error(2, "must be a valid format value");
      RETURN_THROWS();
  }

  if (char_list) {
    php_charmask((const unsigned char *) char_list, char_list_len, ch);
  }

  p = ZSTR_VAL(str);
  e = ZSTR_VAL(str) + ZSTR_LEN(str);

  /* first character cannot be ' or -, unless explicitly allowed by the user */
  if ((*p == '\'' && (!char_list || !ch['\''])) || (*p == '-' && (!char_list || !ch['-']))) {
    p++;
  }
  /* last character cannot be -, unless explicitly allowed by the user */
  if (*(e - 1) == '-' && (!char_list || !ch['-'])) {
    e--;
  }

  while (p < e) {
    /* Each word costs a zval plus a zend_string, so stop once over the limit. */
    if (type && ZEND_HAS_OOM()) {
      RETURN_OOM();
    }
    s = p;
    while (p < e && (isalpha((unsigned char)*p) || (char_list && ch[(unsigned char)*p]) || *p == '\'' || *p == '-')) {
      p++;
    }
    if (p > s) {
      switch (type)
      {
        case 1:
          add_next_index_stringl(return_value, s, p - s);
          break;
        case 2:
          add_index_stringl(return_value, (s - ZSTR_VAL(str)), s, p - s);
          break;
        default:
          word_count++;
          break;
      }
    }
    p++;
  }

  if (!type) {
    RETURN_LONG(word_count);
  }
}

/* }}} */

/* {{{ Convert a string to an array. If split_length is specified, break the string down into chunks each split_length characters long. */
PHP_FUNCTION(str_split)
{
  zend_string *str;
  zend_long split_length = 1;
  const char *p;
  size_t n_reg_segments;

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

  if (split_length <= 0) {
    zend_argument_value_error(2, "must be greater than 0");
    RETURN_THROWS();
  }

  if ((size_t)split_length >= ZSTR_LEN(str)) {
    if (0 == ZSTR_LEN(str)) {
      RETURN_EMPTY_ARRAY();
    }

    array_init_size(return_value, 1);
    GC_TRY_ADDREF(str);
    add_next_index_str(return_value, str);
    return;
  }

  if (UNEXPECTED(!array_finit_size_packed(return_value, (uint32_t)(((ZSTR_LEN(str) - 1) / split_length) + 1)))) {
    RETURN_OOM();
  }

  n_reg_segments = ZSTR_LEN(str) / split_length;
  p = ZSTR_VAL(str);

  ZEND_HASH_FILL_PACKED(Z_ARRVAL_P(return_value)) {
    zval zv;
    while (n_reg_segments-- > 0) {
      /* Each segment costs a zend_string, so stop once over the limit. */
      if (ZEND_HAS_OOM()) {
        RETURN_OOM();
      }
      ZEND_ASSERT(split_length > 0);
      ZVAL_STRINGL_FAST(&zv, p, split_length);
      ZEND_HASH_FILL_ADD(&zv);
      p += split_length;
    }

    if (p != (ZSTR_VAL(str) + ZSTR_LEN(str))) {
      ZVAL_STRINGL_FAST(&zv, p, (ZSTR_VAL(str) + ZSTR_LEN(str) - p));
      ZEND_HASH_FILL_ADD(&zv);
    }
  } ZEND_HASH_FILL_END();
}
/* }}} */

/* {{{ Search a string for any of a set of characters */
PHP_FUNCTION(strpbrk)
{
  zend_string *haystack, *char_list;

  ZEND_PARSE_PARAMETERS_START(2, 2)
    Z_PARAM_STR(haystack)
    Z_PARAM_STR(char_list)
  ZEND_PARSE_PARAMETERS_END();

  if (!ZSTR_LEN(char_list)) {
    zend_argument_value_error(2, "must be a non-empty string");
    RETURN_THROWS();
  }

  size_t shift = php_strcspn(
    ZSTR_VAL(haystack),
    ZSTR_VAL(char_list),
    ZSTR_VAL(haystack) + ZSTR_LEN(haystack),
    ZSTR_VAL(char_list) + ZSTR_LEN(char_list)
  );
  if (shift < ZSTR_LEN(haystack)) {
    RETURN_STRINGL(ZSTR_VAL(haystack) + shift, ZSTR_LEN(haystack) - shift);
  }

  RETURN_FALSE;
}
/* }}} */

/* {{{ Binary safe optionally case insensitive comparison of 2 strings from an offset, up to length characters */
PHP_FUNCTION(substr_compare)
{
  zend_string *s1, *s2;
  zend_long offset, len=0;
  bool len_is_default=1;
  bool cs=0;
  size_t cmp_len;

  ZEND_PARSE_PARAMETERS_START(3, 5)
    Z_PARAM_STR(s1)
    Z_PARAM_STR(s2)
    Z_PARAM_LONG(offset)
    Z_PARAM_OPTIONAL
    Z_PARAM_LONG_OR_NULL(len, len_is_default)
    Z_PARAM_BOOL(cs)
  ZEND_PARSE_PARAMETERS_END();

  if (!len_is_default && len <= 0) {
    if (len == 0) {
      RETURN_LONG(0L);
    } else {
      zend_argument_value_error(4, "must be greater than or equal to 0");
      RETURN_THROWS();
    }
  }

  if (offset < 0) {
    offset = ZSTR_LEN(s1) + offset;
    offset = (offset < 0) ? 0 : offset;
  }

  if ((size_t)offset > ZSTR_LEN(s1)) {
    zend_argument_value_error(3, "must be contained in argument #1 ($haystack)");
    RETURN_THROWS();
  }

  cmp_len = len ? (size_t)len : MAX(ZSTR_LEN(s2), (ZSTR_LEN(s1) - offset));

  if (!cs) {
    RETURN_LONG(zend_binary_strncmp(ZSTR_VAL(s1) + offset, (ZSTR_LEN(s1) - offset), ZSTR_VAL(s2), ZSTR_LEN(s2), cmp_len));
  } else {
    RETURN_LONG(zend_binary_strncasecmp(ZSTR_VAL(s1) + offset, (ZSTR_LEN(s1) - offset), ZSTR_VAL(s2), ZSTR_LEN(s2), cmp_len));
  }
}
/* }}} */

/* {{{ */
static zend_string *php_utf8_encode(const char *s, size_t len)
{
  size_t pos = len;
  zend_string *str;
  unsigned char c;

  str = zend_string_safe_alloc(len, 2, 0, 0);
  ZSTR_LEN(str) = 0;
  while (pos > 0) {
    /* The lower 256 codepoints of Unicode are identical to Latin-1,
     * so we don't need to do any mapping here. */
    c = (unsigned char)(*s);
    if (c < 0x80) {
      ZSTR_VAL(str)[ZSTR_LEN(str)++] = (char) c;
    /* We only account for the single-byte and two-byte cases because
     * we're only dealing with the first 256 Unicode codepoints. */
    } else {
      ZSTR_VAL(str)[ZSTR_LEN(str)++] = (0xc0 | (c >> 6));
      ZSTR_VAL(str)[ZSTR_LEN(str)++] = (0x80 | (c & 0x3f));
    }
    pos--;
    s++;
  }
  ZSTR_VAL(str)[ZSTR_LEN(str)] = '\0';
  str = zend_string_truncate(str, ZSTR_LEN(str), 0);
  return str;
}
/* }}} */

/* {{{ */
static zend_string *php_utf8_decode(const char *s, size_t len)
{
  size_t pos = 0;
  unsigned int c;
  zend_string *str;

  str = zend_string_alloc(len, 0);
  ZSTR_LEN(str) = 0;
  while (pos < len) {
    zend_result status = FAILURE;
    c = php_next_utf8_char((const unsigned char*)s, (size_t) len, &pos, &status);

    /* The lower 256 codepoints of Unicode are identical to Latin-1,
     * so we don't need to do any mapping here beyond replacing non-Latin-1
     * characters. */
    if (status == FAILURE || c > 0xFFU) {
      c = '?';
    }

    ZSTR_VAL(str)[ZSTR_LEN(str)++] = c;
  }
  ZSTR_VAL(str)[ZSTR_LEN(str)] = '\0';
  if (ZSTR_LEN(str) < len) {
    str = zend_string_truncate(str, ZSTR_LEN(str), 0);
  }

  return str;
}
/* }}} */

/* {{{ Encodes an ISO-8859-1 string to UTF-8 */
PHP_FUNCTION(utf8_encode)
{
  char *arg;
  size_t arg_len;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STRING(arg, arg_len)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_STR(php_utf8_encode(arg, arg_len));
}
/* }}} */

/* {{{ Converts a UTF-8 encoded string to ISO-8859-1 */
PHP_FUNCTION(utf8_decode)
{
  char *arg;
  size_t arg_len;

  ZEND_PARSE_PARAMETERS_START(1, 1)
    Z_PARAM_STRING(arg, arg_len)
  ZEND_PARSE_PARAMETERS_END();

  RETURN_STR(php_utf8_decode(arg, arg_len));
}
/* }}} */

Coverage Report

Created: 2026-10-28 06:20