regparse.c

PHP v6.0 For Linux 运行环境：Win9X/ WinME/ WinNT/ Win2K/ WinXP

    PFETCH(c);
    if (c == '>' || c == ')') break;
    if (! ONIGENC_IS_CODE_DIGIT(enc, c))
      r = ONIGERR_INVALID_CHAR_IN_GROUP_NAME;
  }
  if (c != '>') {
    r = ONIGERR_INVALID_GROUP_NAME;
    name_end = end;
  }

  if (r == 0) {
    *rname_end = name_end;
    *src = p;
    return 0;
  }
  else {
  err:
    onig_scan_env_set_error_string(env, r, *src, name_end);
    return r;
  }
}
#endif

static void
CC_ESC_WARN(ScanEnv* env, UChar *c)
{
  if (onig_warn == onig_null_warn) return ;

  if (IS_SYNTAX_BV(env->syntax, ONIG_SYN_WARN_CC_OP_NOT_ESCAPED) &&
      IS_SYNTAX_BV(env->syntax, ONIG_SYN_BACKSLASH_ESCAPE_IN_CC)) {
    char buf[WARN_BUFSIZE];
    onig_snprintf_with_pattern(buf, WARN_BUFSIZE, env->enc,
		env->pattern, env->pattern_end,
		"character class has '%s' without escape", c);
    (*onig_warn)(buf);
  }
}

static void
CCEND_ESC_WARN(ScanEnv* env, UChar* c)
{
  if (onig_warn == onig_null_warn) return ;

  if (IS_SYNTAX_BV((env)->syntax, ONIG_SYN_WARN_CC_OP_NOT_ESCAPED)) {
    char buf[WARN_BUFSIZE];
    onig_snprintf_with_pattern(buf, WARN_BUFSIZE, (env)->enc,
		(env)->pattern, (env)->pattern_end,
		"regular expression has '%s' without escape", c);
    (*onig_warn)(buf);
  }
}

static UChar*
find_str_position(OnigCodePoint s[], int n, UChar* from, UChar* to,
		  UChar **next, OnigEncoding enc)
{
  int i;
  OnigCodePoint x;
  UChar *q;
  UChar *p = from;
  
  while (p < to) {
    x = ONIGENC_MBC_TO_CODE(enc, p, to);
    q = p + enc_len(enc, p);
    if (x == s[0]) {
      for (i = 1; i < n && q < to; i++) {
	x = ONIGENC_MBC_TO_CODE(enc, q, to);
	if (x != s[i]) break;
	q += enc_len(enc, q);
      }
      if (i >= n) {
	if (IS_NOT_NULL(next))
	  *next = q;
	return p;
      }
    }
    p = q;
  }
  return NULL_UCHARP;
}

static int
str_exist_check_with_esc(OnigCodePoint s[], int n, UChar* from, UChar* to,
			 OnigCodePoint bad, OnigEncoding enc)
{
  int i, in_esc;
  OnigCodePoint x;
  UChar *q;
  UChar *p = from;

  in_esc = 0;
  while (p < to) {
    if (in_esc) {
      in_esc = 0;
      p += enc_len(enc, p);
    }
    else {
      x = ONIGENC_MBC_TO_CODE(enc, p, to);
      q = p + enc_len(enc, p);
      if (x == s[0]) {
	for (i = 1; i < n && q < to; i++) {
	  x = ONIGENC_MBC_TO_CODE(enc, q, to);
	  if (x != s[i]) break;
	  q += enc_len(enc, q);
	}
	if (i >= n) return 1;
	p += enc_len(enc, p);
      }
      else {
	x = ONIGENC_MBC_TO_CODE(enc, p, to);
	if (x == bad) return 0;
	else if (x == MC_ESC(enc)) in_esc = 1;
	p = q;
      }
    }
  }
  return 0;
}

static int
fetch_token_in_cc(OnigToken* tok, UChar** src, UChar* end, ScanEnv* env)
{
  int num;
  OnigCodePoint c, c2;
  OnigSyntaxType* syn = env->syntax;
  OnigEncoding enc = env->enc;
  UChar* prev;
  UChar* p = *src;
  PFETCH_READY;

  if (PEND) {
    tok->type = TK_EOT;
    return tok->type;
  }

  PFETCH(c);
  tok->type = TK_CHAR;
  tok->base = 0;
  tok->u.c  = c;
  if (c == ']') {
    tok->type = TK_CC_CLOSE;
  }
  else if (c == '-') {
    tok->type = TK_CC_RANGE;
  }
  else if (c == MC_ESC(enc)) {
    if (! IS_SYNTAX_BV(syn, ONIG_SYN_BACKSLASH_ESCAPE_IN_CC))
      goto end;

    if (PEND) return ONIGERR_END_PATTERN_AT_ESCAPE;

    PFETCH(c);
    tok->escaped = 1;
    tok->u.c = c;
    switch (c) {
    case 'w':
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_WORD;
      break;
    case 'W':
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_WORD;
      break;
    case 'd':
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_DIGIT;
      break;
    case 'D':
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_DIGIT;
      break;
    case 's':
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_WHITE_SPACE;
      break;
    case 'S':
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_WHITE_SPACE;
      break;
    case 'h':
      if (! IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_H_XDIGIT)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_XDIGIT;
      break;
    case 'H':
      if (! IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_H_XDIGIT)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_XDIGIT;
      break;

    case 'p':
    case 'P':
      c2 = PPEEK;
      if (c2 == '{' &&
	  IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY)) {
	PINC;
	tok->type = TK_CHAR_PROPERTY;
	tok->u.prop.not = (c == 'P' ? 1 : 0);

	if (IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT)) {
	  PFETCH(c2);
	  if (c2 == '^') {
	    tok->u.prop.not = (tok->u.prop.not == 0 ? 1 : 0);
	  }
	  else
	    PUNFETCH;
	}
      }
      break;

    case 'x':
      if (PEND) break;

      prev = p;
      if (PPEEK_IS('{') && IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_X_BRACE_HEX8)) {
	PINC;
	num = scan_unsigned_hexadecimal_number(&p, end, 8, enc);
	if (num < 0) return ONIGERR_TOO_BIG_WIDE_CHAR_VALUE;
	if (!PEND) {
          c2 = PPEEK;
          if (ONIGENC_IS_CODE_XDIGIT(enc, c2))
            return ONIGERR_TOO_LONG_WIDE_CHAR_VALUE;
        }

	if (p > prev + enc_len(enc, prev) && !PEND && (PPEEK_IS('}'))) {
	  PINC;
	  tok->type   = TK_CODE_POINT;
	  tok->base   = 16;
	  tok->u.code = (OnigCodePoint )num;
	}
	else {
	  /* can't read nothing or invalid format */
	  p = prev;
	}
      }
      else if (IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_X_HEX2)) {
	num = scan_unsigned_hexadecimal_number(&p, end, 2, enc);
	if (num < 0) return ONIGERR_TOO_BIG_NUMBER;
	if (p == prev) {  /* can't read nothing. */
	  num = 0; /* but, it's not error */
	}
	tok->type = TK_RAW_BYTE;
	tok->base = 16;
	tok->u.c  = num;
      }
      break;

    case 'u':
      if (PEND) break;

      prev = p;
      if (IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_U_HEX4)) {
	num = scan_unsigned_hexadecimal_number(&p, end, 4, enc);
	if (num < 0) return ONIGERR_TOO_BIG_NUMBER;
	if (p == prev) {  /* can't read nothing. */
	  num = 0; /* but, it's not error */
	}
	tok->type   = TK_CODE_POINT;
	tok->base   = 16;
	tok->u.code = (OnigCodePoint )num;
      }
      break;

    case '0':
    case '1': case '2': case '3': case '4': case '5': case '6': case '7':
      if (IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_OCTAL3)) {
	PUNFETCH;
	prev = p;
	num = scan_unsigned_octal_number(&p, end, 3, enc);
	if (num < 0) return ONIGERR_TOO_BIG_NUMBER;
	if (p == prev) {  /* can't read nothing. */
	  num = 0; /* but, it's not error */
	}
	tok->type = TK_RAW_BYTE;
	tok->base = 8;
	tok->u.c  = num;
      }
      break;

    default:
      PUNFETCH;
      num = fetch_escaped_value(&p, end, env);
      if (num < 0) return num;
      if (tok->u.c != num) {
	tok->u.code = (OnigCodePoint )num;
	tok->type   = TK_CODE_POINT;
      }
      break;
    }
  }
  else if (c == '[') {
    if (IS_SYNTAX_OP(syn, ONIG_SYN_OP_POSIX_BRACKET) && (PPEEK_IS(':'))) {
      OnigCodePoint send[] = { (OnigCodePoint )':', (OnigCodePoint )']' };
      tok->backp = p; /* point at '[' is readed */
      PINC;
      if (str_exist_check_with_esc(send, 2, p, end,
                                   (OnigCodePoint )']', enc)) {
	tok->type = TK_POSIX_BRACKET_OPEN;
      }
      else {
	PUNFETCH;
	goto cc_in_cc;
      }
    }
    else {
    cc_in_cc:
      if (IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_CCLASS_SET_OP)) {
	tok->type = TK_CC_CC_OPEN;
      }
      else {
	CC_ESC_WARN(env, "[");
      }
    }
  }
  else if (c == '&') {
    if (IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_CCLASS_SET_OP) &&
	!PEND && (PPEEK_IS('&'))) {
      PINC;
      tok->type = TK_CC_AND;
    }
  }

 end:
  *src = p;
  return tok->type;
}

static int
fetch_token(OnigToken* tok, UChar** src, UChar* end, ScanEnv* env)
{
  int r, num;
  OnigCodePoint c;
  OnigEncoding enc = env->enc;
  OnigSyntaxType* syn = env->syntax;
  UChar* prev;
  UChar* p = *src;
  PFETCH_READY;

 start:
  if (PEND) {
    tok->type = TK_EOT;
    return tok->type;
  }

  tok->type  = TK_STRING;
  tok->base  = 0;
  tok->backp = p;

  PFETCH(c);
  if (c == MC_ESC(enc)) {
    if (PEND) return ONIGERR_END_PATTERN_AT_ESCAPE;

    tok->backp = p;
    PFETCH(c);

    tok->u.c = c;
    tok->escaped = 1;
    switch (c) {
    case '*':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_ASTERISK_ZERO_INF)) break;
      tok->type = TK_OP_REPEAT;
      tok->u.repeat.lower = 0;
      tok->u.repeat.upper = REPEAT_INFINITE;
      goto greedy_check;
      break;

    case '+':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_PLUS_ONE_INF)) break;
      tok->type = TK_OP_REPEAT;
      tok->u.repeat.lower = 1;
      tok->u.repeat.upper = REPEAT_INFINITE;
      goto greedy_check;
      break;

    case '?':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_QMARK_ZERO_ONE)) break;
      tok->type = TK_OP_REPEAT;
      tok->u.repeat.lower = 0;
      tok->u.repeat.upper = 1;
    greedy_check:
      if (!PEND && PPEEK_IS('?') &&
	  IS_SYNTAX_OP(syn, ONIG_SYN_OP_QMARK_NON_GREEDY)) {
	PFETCH(c);
	tok->u.repeat.greedy     = 0;
	tok->u.repeat.possessive = 0;
      }
      else {
      possessive_check:
	if (!PEND && PPEEK_IS('+') &&
	    ((IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT) &&
	      tok->type != TK_INTERVAL)  ||
	     (IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL) &&
	      tok->type == TK_INTERVAL))) {
	  PFETCH(c);
	  tok->u.repeat.greedy     = 1;
	  tok->u.repeat.possessive = 1;
	}
	else {
	  tok->u.repeat.greedy     = 1;
	  tok->u.repeat.possessive = 0;
	}
      }
      break;

    case '{':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_BRACE_INTERVAL)) break;
      r = fetch_range_qualifier(&p, end, tok, env);
      if (r < 0) return r;  /* error */
      if (r == 0) goto greedy_check;
      else if (r == 2) { /* {n} */
	if (IS_SYNTAX_BV(syn, ONIG_SYN_FIXED_INTERVAL_IS_GREEDY_ONLY))
	  goto possessive_check;

	goto greedy_check;
      }
      /* r == 1 : normal char */
      break;

    case '|':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_VBAR_ALT)) break;
      tok->type = TK_ALT;
      break;

    case '(':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_LPAREN_SUBEXP)) break;
      tok->type = TK_SUBEXP_OPEN;
      break;

    case ')':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_LPAREN_SUBEXP)) break;
      tok->type = TK_SUBEXP_CLOSE;
      break;

    case 'w':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_W_WORD)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_WORD;
      break;

    case 'W':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_W_WORD)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_WORD;
      break;

    case 'b':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_B_WORD_BOUND)) break;
      tok->type = TK_ANCHOR;
      tok->u.anchor = ANCHOR_WORD_BOUND;
      break;

    case 'B':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_B_WORD_BOUND)) break;
      tok->type = TK_ANCHOR;
      tok->u.anchor = ANCHOR_NOT_WORD_BOUND;
      break;

#ifdef USE_WORD_BEGIN_END
    case '<':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END)) break;
      tok->type = TK_ANCHOR;
      tok->u.anchor = ANCHOR_WORD_BEGIN;
      break;

    case '>':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END)) break;
      tok->type = TK_ANCHOR;
      tok->u.anchor = ANCHOR_WORD_END;
      break;
#endif

    case 's':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_S_WHITE_SPACE)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_WHITE_SPACE;
      break;

    case 'S':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_S_WHITE_SPACE)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_WHITE_SPACE;
      break;

    case 'd':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_D_DIGIT)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_DIGIT;
      break;

    case 'D':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_D_DIGIT)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_DIGIT;
      break;

    case 'h':
      if (! IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_H_XDIGIT)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_XDIGIT;
      break;

    case 'H':
      if (! IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_H_XDIGIT)) break;
      tok->type = TK_CHAR_TYPE;
      tok->u.subtype = CTYPE_NOT_XDIGIT;
      break;

    case 'A':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR)) break;
    begin_buf:
      tok->type = TK_ANCHOR;
      tok->u.subtype = ANCHOR_BEGIN_BUF;
      break;

    case 'Z':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR)) break;
      tok->type = TK_ANCHOR;
      tok->u.subtype = ANCHOR_SEMI_END_BUF;
      break;

    case 'z':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR)) break;
    end_buf:
      tok->type = TK_ANCHOR;
      tok->u.subtype = ANCHOR_END_BUF;
      break;

    case 'G':
      if (! IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR)) break;
      tok->type = TK_ANCHOR;
      tok->u.subtype = ANCHOR_BEGIN_POSITION;
      break;

    case '`':
      if (! IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR)) break;
      goto begin_buf;
      break;

    case '\'':
      if (! IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR)) break;
      goto end_buf;
      break;

    case 'x':
      if (PEND) break;

      prev = p;
      if (PPEEK_IS('{') && IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_X_BRACE_HEX8)) {
	PINC;
	num = scan_unsigned_hexadecimal_number(&p, end, 8, enc);
	if (num < 0) return ONIGERR_TOO_BIG_WIDE_CHAR_VALUE;
	if (!PEND) {
          if (ONIGENC_IS_CODE_XDIGIT(enc, PPEEK))
            return ONIGERR_TOO_LONG_WIDE_CHAR_VALUE;
        }

	if ((p > prev + enc_len(enc, prev)) && !PEND && PPEEK_IS('}')) {
	  PINC;
	  tok->type   = TK_CODE_POINT;
	  tok->u.code = (OnigCodePoint )num;
	}
	else {
	  /* can't read nothing or invalid format */
	  p = prev;
	}
      }
      else if (IS_SYNTAX_OP(syn, ONIG_SYN_OP_ESC_X_HEX2)) {
	num = scan_unsigned_hexadecimal_number(&p, end, 2, enc);
	if (num < 0) return ONIGERR_TOO_BIG_NUMBER;
	if (p == prev) {  /* can't read nothing. */
	  num = 0; /* but, it's not error */
	}
	tok->type = TK_RAW_BYTE;
	tok->base = 16;
	tok->u.c  = num;
      }
      break;

    case 'u':
      if (PEND) break;

      prev = p;
      if (IS_SYNTAX_OP2(syn, ONIG_SYN_OP2_ESC_U_HEX4)) {
	num = scan_unsigned_hexadecimal_number(&p, end, 4, enc);
	if (num < 0) return ONIGERR_TOO_BIG_NUMBER;
	if (p == prev) {  /* can't read nothing. */
	  num = 0; /* but, it's not error */
	}
	tok->type   = TK_CODE_POINT;
	tok->base   = 16;
	tok->u.code = (OnigCodePoint )num;
      }
      break;

    case '1': case '2': case '3': case '4':
    case '5': case '6': case '7': case '8':