regularexp.c

nedit 是一款linux下的开发源码的功能强大的编辑器

          *   I~ B~ N~ B~ (...)~ C~ T~x T~m K~ K~ N~
          *    \_____\__\_|          \   \__4__|  |
          *      10   \  \            \_7_________|
          *            \  \_________6_____________|
          *             \_______5_________________|
          */

         tail (ret_val, emit_special (INC_COUNT, 0UL, Num_Braces)); /* 1 */

         next = emit_special (TEST_COUNT, min_max [1], Num_Braces); /* 2,7 */

         tail (ret_val, next);                                      /* 2 */

         next = emit_special (TEST_COUNT, min_max [0], Num_Braces); /* 4 */

         tail (emit_node (BACK), ret_val);                          /* 3 */
         tail (next, emit_node (BACK));                             /* 4 */
         (void) insert (BRANCH, ret_val, 0UL, 0UL, 0);              /* 6,8 */
         (void) insert (NOTHING, ret_val, 0UL, 0UL, 0);             /* 5 */
         (void) insert (BRANCH,  ret_val, 0UL, 0UL, 0);             /* 8,9 */

         next = emit_node (NOTHING);                                /* 5,6,7 */

         offset_tail (ret_val, NODE_SIZE, next);                    /* 5 */
         offset_tail (ret_val, 2 * NODE_SIZE, next);                /* 6 */
         offset_tail (ret_val, 3 * NODE_SIZE, next);                /* 7 */
         tail (ret_val, ret_val + (2 * NODE_SIZE));                 /* 8 */
         offset_tail (next, -NODE_SIZE, ret_val);                   /* 9 */
         insert (INIT_COUNT, ret_val, 0UL, 0UL, Num_Braces);        /* 10 */
         tail (ret_val, ret_val + INDEX_SIZE + (4 * NODE_SIZE));    /* 10 */
      }

      Num_Braces++;
   } else if (op_code == '{') {
      if (min_max [0] == REG_ZERO && min_max [1] != REG_INFINITY) {
         /* Node structure for (x){0,n} or (x){,n} construct.
          *
          *       ___3____________
          *      |       1_  _2   \   5_
          *      |       / |/ |    \  / |
          *   I~ B~ (...)~ C~ T~x K~ B~ N~
          *    \_|\            \_6___|__|
          *    7   \________4________|
          */

         tail (ret_val, emit_special (INC_COUNT, 0UL, Num_Braces)); /* 1 */

         next = emit_special (TEST_COUNT, min_max [1], Num_Braces); /* 2,6 */

         tail (ret_val, next);                                      /* 2 */
         (void) insert (BRANCH, ret_val, 0UL, 0UL, 0);              /* 3,4,7 */
         tail (emit_node (BACK), ret_val);                          /* 3 */

         next = emit_node (BRANCH);                                 /* 4,5 */

         tail (ret_val, next);                                      /* 4 */
         tail (next, emit_node (NOTHING));                          /* 5,6 */
         offset_tail (ret_val, NODE_SIZE, next);                    /* 6 */

         next = insert (INIT_COUNT, ret_val, 0UL, 0UL, Num_Braces); /* 7 */

         tail (ret_val, next);                                      /* 7 */

      } else if (min_max [0] > REG_ZERO && min_max [1] == REG_INFINITY) {
         /* Node structure for (x){m,} construct.
          *       __________4________
          *      |    __3__________  \
          *     _|___|    1_  _2   \  \    _7
          *    / | 8 |    / |/ |    \  \  / |
          *   I~ B~  (...)~ C~ T~m K~ K~ B~ N~
          *       \             \_5___|  |
          *        \__________6__________|
          */

         tail (ret_val, emit_special (INC_COUNT, 0UL, Num_Braces)); /* 1 */

         next = emit_special (TEST_COUNT, min_max [0], Num_Braces); /* 2 */

         tail (ret_val, next);                                      /* 2 */
         tail (emit_node (BACK), ret_val);                          /* 3 */
         (void) insert (BRANCH, ret_val, 0UL, 0UL, 0);              /* 4,6 */

         next = emit_node (BACK);                                   /* 4 */

         tail        (next, ret_val);                               /* 4 */
         offset_tail (ret_val, NODE_SIZE, next);                    /* 5 */
         tail        (ret_val, emit_node (BRANCH));                 /* 6 */
         tail        (ret_val, emit_node (NOTHING));                /* 7 */

         insert (INIT_COUNT, ret_val, 0UL, 0UL, Num_Braces);        /* 8 */

         tail (ret_val, ret_val + INDEX_SIZE + (2 * NODE_SIZE));    /* 8 */

      } else {
         /* Node structure for (x){m,n} construct.
          *       _____6________________
          *      |   _____________3___  \
          *    9_|__|    1_  _2       \  \    _8
          *    / |  |    / |/ |        \  \  / |
          *   I~ B~ (...)~ C~ T~x T~m K~ K~ B~ N~
          *       \            \   \__4__|  |  |
          *        \            \_7_________|__|
          *         \_________5_____________|
          */

         tail (ret_val, emit_special (INC_COUNT, 0UL, Num_Braces)); /* 1 */

         next = emit_special (TEST_COUNT, min_max [1], Num_Braces); /* 2,4 */

         tail (ret_val, next);                                      /* 2 */

         next = emit_special (TEST_COUNT, min_max [0], Num_Braces); /* 4 */

         tail (emit_node (BACK), ret_val);                          /* 3 */
         tail (next, emit_node (BACK));                             /* 4 */
         (void) insert (BRANCH, ret_val, 0UL, 0UL, 0);              /* 5,6 */

         next = emit_node (BRANCH);                                 /* 5,8 */

         tail        (ret_val, next);                               /* 5 */
         offset_tail (next, -NODE_SIZE, ret_val);                   /* 6 */

         next = emit_node (NOTHING);                                /* 7,8 */

         offset_tail (ret_val, NODE_SIZE, next);                    /* 7 */

         offset_tail (next, -NODE_SIZE, next);                      /* 8 */
         (void) insert (INIT_COUNT, ret_val, 0UL, 0UL, Num_Braces); /* 9 */
         tail (ret_val, ret_val + INDEX_SIZE + (2 * NODE_SIZE));    /* 9 */
      }

      Num_Braces++;
   } else {
      /* We get here if the IS_QUANTIFIER macro is not coordinated properly
         with this function. */

      REG_FAIL ("internal error #2, `piece\'");
   }

   if (IS_QUANTIFIER (*Reg_Parse)) {
      if (op_code == '{') {
         sprintf (Error_Text, "nested quantifiers, {m,n}%c", *Reg_Parse);
      } else {
         sprintf (Error_Text, "nested quantifiers, %c%c", op_code, *Reg_Parse);
      }

      REG_FAIL (Error_Text);
   }

   return (ret_val);
}

/*----------------------------------------------------------------------*
 * atom
 *
 * Process one regex item at the lowest level
 *
 * OPTIMIZATION:  Lumps a continuous sequence of ordinary characters
 * together so that it can turn them into a single EXACTLY node, which
 * is smaller to store and faster to run.
 *----------------------------------------------------------------------*/

static unsigned char * atom (int *flag_param, len_range *range_param) {

   register unsigned char *ret_val;
            unsigned char  test;
            int            flags_local;
	    len_range      range_local;

   *flag_param = WORST;  /* Tentatively. */
   range_param->lower = 0; /* Idem */
   range_param->upper = 0;

   /* Process any regex comments, e.g. `(?# match next token->)'.  The
      terminating right parenthesis can not be escaped.  The comment stops at
      the first right parenthesis encountered (or the end of the regex
      string)... period.  Handles multiple sequential comments,
      e.g. `(?# one)(?# two)...'  */

   while (*Reg_Parse      == '(' &&
         *(Reg_Parse + 1) == '?' &&
         *(Reg_Parse + 2) == '#') {

      Reg_Parse += 3;

      while (*Reg_Parse != ')' && *Reg_Parse != '\0') {
         Reg_Parse++;
      }

      if (*Reg_Parse == ')') {
         Reg_Parse++;
      }

      if (*Reg_Parse == ')' || *Reg_Parse == '|' || *Reg_Parse == '\0') {
         /* Hit end of regex string or end of parenthesized regex; have to
          return "something" (i.e. a NOTHING node) to avoid generating an
          error. */

         ret_val = emit_node (NOTHING);

         return (ret_val);
      }
   }

   switch (*Reg_Parse++) {
      case '^':
         ret_val = emit_node (BOL);
         break;

      case '$':
         ret_val = emit_node (EOL);
         break;

      case '<':
         ret_val = emit_node (BOWORD);
         break;

      case '>':
         ret_val = emit_node (EOWORD);
         break;

      case '.':
         if (Match_Newline) {
            ret_val = emit_node (EVERY);
         } else {
            ret_val = emit_node (ANY);
         }

         *flag_param |= (HAS_WIDTH | SIMPLE); 
	 range_param->lower = 1;
	 range_param->upper = 1;
	 break;

      case '(':
         if (*Reg_Parse == '?') {  /* Special parenthetical expression */
            Reg_Parse++;
	    range_local.lower = 0; /* Make sure it is always used */
	    range_local.upper = 0;

            if (*Reg_Parse == ':') {
               Reg_Parse++;
               ret_val = chunk (NO_CAPTURE, &flags_local, &range_local);
            } else if (*Reg_Parse == '=') {
               Reg_Parse++;
               ret_val = chunk (POS_AHEAD_OPEN, &flags_local, &range_local);
            } else if (*Reg_Parse == '!') {
               Reg_Parse++;
               ret_val = chunk (NEG_AHEAD_OPEN, &flags_local, &range_local);
            } else if (*Reg_Parse == 'i') {
               Reg_Parse++;
               ret_val = chunk (INSENSITIVE, &flags_local, &range_local);
            } else if (*Reg_Parse == 'I') {
               Reg_Parse++;
               ret_val = chunk (SENSITIVE, &flags_local, &range_local);
            } else if (*Reg_Parse == 'n') {
               Reg_Parse++;
               ret_val = chunk (NEWLINE, &flags_local, &range_local);
            } else if (*Reg_Parse == 'N') {
               Reg_Parse++;
               ret_val = chunk (NO_NEWLINE, &flags_local, &range_local);
            } else if (*Reg_Parse == '<') {
               Reg_Parse++;
	       if (*Reg_Parse == '=') {
	          Reg_Parse++;
                  ret_val = chunk (POS_BEHIND_OPEN, &flags_local, &range_local);
	       } else if (*Reg_Parse == '!') {
	          Reg_Parse++;
                  ret_val = chunk (NEG_BEHIND_OPEN, &flags_local, &range_local);
	       } else {
                  sprintf (Error_Text,
                           "invalid look-behind syntax, \"(?<%c...)\"",
                           *Reg_Parse);

                  REG_FAIL (Error_Text);
	       }
            } else {
               sprintf (Error_Text,
                        "invalid grouping syntax, \"(?%c...)\"",
                        *Reg_Parse);

               REG_FAIL (Error_Text);
            }
         } else { /* Normal capturing parentheses */
            ret_val = chunk (PAREN, &flags_local, &range_local);
         }

         if (ret_val == NULL) return (NULL);  /* Something went wrong. */

         /* Add HAS_WIDTH flag if it was set by call to chunk. */

         *flag_param |= flags_local & HAS_WIDTH;
	 *range_param = range_local;

         break;

      case '\0':
      case '|':
      case ')':
         REG_FAIL ("internal error #3, `atom\'");  /* Supposed to be  */
                                                   /* caught earlier. */
      case '?':
      case '+':
      case '*':
         sprintf (Error_Text, "%c follows nothing", *(Reg_Parse - 1));
         REG_FAIL (Error_Text);

      case '{':
         if (Enable_Counting_Quantifier) {
            REG_FAIL ("{m,n} follows nothing");
         } else {
            ret_val = emit_node (EXACTLY); /* Treat braces as literals. */
            emit_byte ('{');
            emit_byte ('\0');
	    range_param->lower = 1;
	    range_param->upper = 1;
         }

         break;

      case '[':
         {
            register unsigned int  second_value;
            register unsigned int  last_value;
                     unsigned char