ircproxy  The Ultimate Cyborg

---

match.cc

#ifndef USE_PCH
#include "sys.h"
#endif

#include "match.h"
#include "ircd_chattr.h"

/*
 * mmatch()
 *
 * Written by Run (run@alinoe.com), 25-10-96
 */

/*
 * From: Carlo Wood <carlo@runaway.xs4all.nl>
 * Message-Id: <199609021026.MAA02393@runaway.xs4all.nl>
 * Subject: [C-Com] Analysis for `mmatch' (was: gline4 problem)
 * To: coder-com@mail.undernet.org (coder committee)
 * Date: Mon, 2 Sep 1996 12:26:01 +0200 (MET DST)
 *
 * We need a new function `mmatch(char const* old_mask, char const* new_mask)'
 * which returns `true' likewise the current `match' (start with copying it),
 * but which treats '*' and '?' in `new_mask' differently (not "\*" and "\?" !)
 * as follows:  a '*' in `new_mask' does not match a '?' in `old_mask' and
 * a '?' in `new_mask' does not match a '\?' in `old_mask'.
 * And ofcourse... a '*' in `new_mask' does not match a '\*' in `old_mask'...
 * And last but not least, '\?' and '\*' in `new_mask' now become one character.
 */

int mmatch(char const* old_mask, char const* new_mask)
{
  register char const* m = old_mask;
  register char const* n = new_mask;
  char const* ma = m;
  char const* na = n;
  int wild = 0;
  int mq = 0, nq = 0;

  while (1)
  {
    if (*m == '*')
    {
      while (*m == '*')
        m++;
      wild = 1;
      ma = m;
      na = n;
    }

    if (!*m)
    {
      if (!*n)
        return 0;
      for (m--; (m > old_mask) && (*m == '?'); m--)
        ;
      if ((*m == '*') && (m > old_mask) && (m[-1] != '\\'))
        return 0;
      if (!wild)
        return 1;
      m = ma;

      /* Added to `mmatch' : Because '\?' and '\*' now is one character: */
      if ((*na == '\\') && ((na[1] == '*') || (na[1] == '?')))
        ++na;

      n = ++na;
    }
    else if (!*n)
    {
      while (*m == '*')
        m++;
      return (*m != 0);
    }
    if ((*m == '\\') && ((m[1] == '*') || (m[1] == '?')))
    {
      m++;
      mq = 1;
    }
    else
      mq = 0;

    /* Added to `mmatch' : Because '\?' and '\*' now is one character: */
    if ((*n == '\\') && ((n[1] == '*') || (n[1] == '?')))
    {
      n++;
      nq = 1;
    }
    else
      nq = 0;

/*
 * This `if' has been changed compared to match() to do the following:
 * Match when:
 *   old (m)         new (n)         boolean expression
 *    *               any             (*m == '*' && !mq) ||
 *    ?               any except '*'  (*m == '?' && !mq && (*n != '*' || nq)) ||
 * any except * or ?  same as m       (!((*m == '*' || *m == '?') && !mq) &&
 *                                      ToLower(*m) == ToLower(*n) &&
 *                                        !((mq && !nq) || (!mq && nq)))
 *
 * Here `any' also includes \* and \? !
 *
 * After reworking the boolean expressions, we get:
 * (Optimized to use boolean shortcircuits, with most frequently occuring
 *  cases upfront (which took 2 hours!)).
 */
    if ((*m == '*' && !mq) ||
        ((!mq || nq) && ToLower(*m) == ToLower(*n)) ||
        (*m == '?' && !mq && (*n != '*' || nq)))
    {
      if (*m)
        m++;
      if (*n)
        n++;
    }
    else
    {
      if (!wild)
        return 1;
      m = ma;

      /* Added to `mmatch' : Because '\?' and '\*' now is one character: */
      if ((*na == '\\') && ((na[1] == '*') || (na[1] == '?')))
        ++na;

      n = ++na;
    }
  }
}

/*
 * Compare if a given string (name) matches the given
 * mask (which can contain wild cards: '*' - match any
 * number of chars, '?' - match any single character.
 *
 * return  0, if match
 *         1, if no match
 */

/*
 * match
 *
 * Rewritten by Andrea Cocito (Nemesi), November 1998.
 *
 */

/****************** Nemesi's match() ***************/

int match(char const* mask, char const* string)
{
  register char const* m = mask, *s = string;
  register char ch;
  char const* bm, *bs;          /* Will be reg anyway on a decent CPU/compiler */

  /* Process the "head" of the mask, if any */
  while ((ch = *m++) && (ch != '*'))
    switch (ch)
    {
      case '\\':
        if (*m == '?' || *m == '*')
          ch = *m++;
      default:
        if (ToLower(*s) != ToLower(ch))
          return 1;
      case '?':
        if (!*s++)
          return 1;
    };
  if (!ch)
    return *s;

  /* We got a star: quickly find if/where we match the next char */
got_star:
  bm = m;                       /* Next try rollback here */
  while ((ch = *m++))
    switch (ch)
    {
      case '?':
        if (!*s++)
          return 1;
      case '*':
        bm = m;
        continue;               /* while */
      case '\\':
        if (*m == '?' || *m == '*')
          ch = *m++;
      default:
        goto break_while;       /* C is structured ? */
    };
break_while:
  if (!ch)
    return 0;                   /* mask ends with '*', we got it */
  ch = ToLower(ch);
  while (ToLower(*s++) != ch)
    if (!*s)
      return 1;
  bs = s;                       /* Next try start from here */

  /* Check the rest of the "chunk" */
  while ((ch = *m++))
  {
    switch (ch)
    {
      case '*':
        goto got_star;
      case '\\':
        if (*m == '?' || *m == '*')
          ch = *m++;
      default:
        if (ToLower(*s) != ToLower(ch))
        {
          m = bm;
          s = bs;
          goto got_star;
        };
      case '?':
        if (!*s++)
          return 1;
    };
  };
  if (*s)
  {
    m = bm;
    s = bs;
    goto got_star;
  };
  return 0;
}

/*
 * collapse()
 * Collapse a pattern string into minimal components.
 * This particular version is "in place", so that it changes the pattern
 * which is to be reduced to a "minimal" size.
 *
 * (C) Carlo Wood - 6 Oct 1998
 * Speedup rewrite by Andrea Cocito, December 1998.
 * Note that this new optimized alghoritm can *only* work in place.
 */

char* collapse(char* mask)
{
  register int star = 0;
  register char* m = mask;
  register char* b;

  if (m)
  {
    do
    {
      if ((*m == '*') && ((m[1] == '*') || (m[1] == '?')))
      {
        b = m;
        do
        {
          if (*m == '*')
            star = 1;
          else
          {
            if (star && (*m != '?'))
            {
              *b++ = '*';
              star = 0;
            };
            *b++ = *m;
            if ((*m == '\\') && ((m[1] == '*') || (m[1] == '?')))
              *b++ = *++m;
          };
        }
        while (*m++);
        break;
      }
      else
      {
        if ((*m == '\\') && ((m[1] == '*') || (m[1] == '?')))
          m++;
      };
    }
    while (*m++);
  };
  return mask;
}

/*
 ***************** Nemesi's matchcomp() / matchexec() **************
 */

/* These functions allow the use of "compiled" masks, you compile a mask
 * by means of matchcomp() that gets the plain text mask as input and writes
 * its result in the memory locations addressed by the 3 parameters:
 * - *cmask will contain the text of the compiled mask
 * - *minlen will contain the lenght of the shortest string that can match 
 *   the mask
 * - *charset will contain the minimal set of chars needed to match the mask
 * You can pass NULL as *charset and it will be simply not returned, but you
 * MUST pass valid pointers for *minlen and *cmask (wich must be big enough 
 * to contain the compiled mask text that is in the worst case as long as the 
 * text of the mask itself in plaintext format) and the return value of 
 * matchcomp() will be the number of chars actually written there (excluded 
 * the trailing zero). cmask can be == mask, matchcomp() can work in place.
 * The {cmask, minlen} couple of values make the real compiled mask and
 * need to be passed to the functions that use the compiled mask, if you pass
 * the wrong minlen or something wrong in cmask to one of these expect a
 * coredump. This means that when you record a compiled mask you must store
 * *both* these values.
 * Once compiled the mask can be used to match a string by means of 
 * matchexec(), it can be printed back to human-readable format by means
 * of sprintmatch() or it can be compared to another compiled mask by means
 * of mmexec() that will tell if it completely overrides that mask (a lot like
 * what mmatch() does for plain text masks).
 * You can gain a lot of speed in many situations avoiding to matchexec() when:
 * - The maximum lenght of the field you are about to match() the mask to is
 *   shorter than minlen, in example when matching abc*def*ghil with a nick:
 *   It just cannot match since a nick is at most 9 chars long and the mask
 *   needs at least 10 chars (10 will be the value returned in minlen).
 * - The charset allowed for the field you are about to match to doesn't
 *   "contain" the charset returned by matchcomp(), in example when you
 *   have *.* as mask it makes no sense to try to match it against a nick
 *   because, again, a nick can't contain a '.', you can check this with
 *   a simple (charset & NTL_IRCNK) in this case.
 * - As a special case, since compiled masks are forced to lowercase,
 *   it would make no sense to use the NTL_LOWER and NTL_UPPER on a compiled
 *   mask, thus they are reused as follows: if the NTL_LOWER bit of charset
 *   is set it means that the mask contains only non-wilds chars (i.e. you can
 *   use strCasecmp() to match it or a direct hash lookup), if the NTL_UPPER
 *   bit is set it means that it contains only wild chars (and you can
 *   match it with strlen(field)>=minlen).
 * Do these optimizations ONLY when the data you are about to pass to
 * matchexec() are *known* to be invalid in advance, using strChattr() 
 * or strlen() on the text would be slower than calling matchexec() directly
 * and let it fail.
 * Internally a compiled mask contain in the *cmask area the text of
 * the plain text form of the mask itself with applied the following hacks:
 * - All characters are forced to lowercase (so that uppercase letters and
 *   specifically the symbols 'A' and 'Z' are reserved for special use)
 * - All non-escaped stars '*' are replaced by the letter 'Z'
 * - All non-escaped question marks '?' are replaced by the letter 'A' 
 * - All escape characters are removed, the wilds escaped by them are
 *   then passed by without the escape since they don't collide anymore
 *   with the real wilds (encoded as A/Z) 
 * - Finally the part of the mask that follows the last asterisk is
 *   reversed (byte order mirroring) and moved right after the first
 *   asterisk.
 * After all this a mask like:   Head*CHUNK1*chu\*nK2*ch??k3*TaIl 
 *               .... becomes:   headZliatZchunk1Zchu*nk2ZchAAk3
 * This can still be printed on a console, more or less understood by an
 * human and handled with the usual str*() library functions.
 * When you store somewhere the compiled mask you can avoid storing the
 * textform of it since it can be "decompiled" by means of sprintmatch(),
 * but at that time the following things are changed in the mask:
 * - All chars have been forced to lowercase.
 * - The mask is collapsed.
 * The balance point of using compiled masks in terms of CPU is when you expect
 * to use matchexec() instead of match() at least 20 times on the same mask
 * or when you expect to use mmexec() instead of mmatch() 3 times.
 */

 /* 
  * matchcomp()
  *
  * Compiles a mask into a form suitable for using in matchexec().
  */

int matchcomp(char* cmask, int* minlen, int* charset, char const* mask)
{
  char const* m = mask;
  char* b = cmask;
  char* fs = NULL;
  char* ls = NULL;
  char* x1, *x2;
  int l1, l2, lmin, loop, sign;
  int star = 0;
  int cnt = 0;
  char ch;
  int chset = ~0;
  int chset2 = (NTL_LOWER | NTL_UPPER);

  if (m)
    while ((ch = *m++))
      switch (ch)
      {
        case '*':
          star = 1;
          break;
        case '?':
          cnt++;
          *b++ = 'A';
          chset2 &= ~NTL_LOWER;
          break;
        case '\\':
          if ((*m == '?') || (*m == '*'))
            ch = *m++;
        default:
          if (star)
          {
            ls = b;
            fs = fs ? fs : b;
            *b++ = 'Z';
            chset2 &= ~NTL_LOWER;
            star = 0;
          };
          cnt++;
          chset &= IRCD_CharAttrTab[((*b++ = ToLower(ch))) - std::numeric_limits<char>::min()];
          chset2 &= ~NTL_UPPER;
      };

  if (charset)
    *charset = (chset | chset2);

  if (star)
  {
    ls = b;
    fs = (fs ? fs : b);
    *b++ = 'Z';
  };

  if (ls)
  {
    for (x1 = ls + 1, x2 = (b - 1); x1 < x2; x1++, x2--)
    {
      ch = *x1;
      *x1 = *x2;
      *x2 = ch;
    };
    l1 = (ls - fs);
    l2 = (b - ls);
    x1 = fs;
    while ((lmin = (l1 < l2) ? l1 : l2))
    {
      x2 = x1 + l1;
      for (loop = 0; loop < lmin; loop++)
      {
        ch = x1[loop];
        x1[loop] = x2[loop];
        x2[loop] = ch;
      };
      x1 += lmin;
      sign = l1 - l2;
      l1 -= (sign < 0) ? 0 : lmin;
      l2 -= (sign > 0) ? 0 : lmin;
    };
  };

  *b = '\000';
  *minlen = cnt;
  return (b - cmask);
}

/*
 * matchexec()
 *
 * Executes a match with a mask previosuly compiled with matchcomp()
 * Note 1: If the mask isn't correctly produced by matchcomp() I will core
 * Note 2: 'min' MUST be the value returned by matchcomp on that mask,
 *         or.... I will core even faster :-)
 * Note 3: This piece of code is not intended to be nice but efficient.
 *
 * 25 May 2006:
 *
 * Adapted by Carlo Wood for a non- zero-terminated string.
 * Instead, the a parameter 'len' is passed, being the number of valid characters in 'string'.
 */

int matchexec(char const* string, size_t len, char const* cmask, int minlen)
{
  char const* bs = string - 1;
  char const* s = string + len;
  char const* b = cmask - 1;
  int trash;
  char const* bb;
  char ch;

tryhead:
  while ((ToLower(*++bs) == *++b) && bs != s);
  if (bs == s)
    return ((*b != '\000') && ((*b++ != 'Z') || (*b != '\000')));
  if (*b != 'Z')
  {
    if (*b == 'A')
      goto tryhead;
    return 1;
  };

  if ((trash = (len - minlen)) < 0)
    return 2;

trytail:
  while ((ToLower(*--s) == *++b) && *b && (ToLower(*--s) == *++b) && *b
      && (ToLower(*--s) == *++b) && *b && (ToLower(*--s) == *++b) && *b);
  if (*b != 'Z')
  {
    if (*b == 'A')
      goto trytail;
    return (*b != '\000');
  };

  s = --bs;
  bb = b;

  while ((ch = *++b))
  {
    while ((ToLower(*++s) != ch))
      if (--trash < 0)
        return 4;
    bs = s;

  trychunk:
    while ((ToLower(*++s) == *++b) && *b);
    if (!*b)
      return 0;
    if (*b == 'Z')
    {
      bs = --s;
      bb = b;
      continue;
    };
    if (*b == 'A')
      goto trychunk;

    b = bb;
    s = bs;
    if (--trash < 0)
      return 5;
  };

  return 0;
}

/*
 * matchdecomp()
 * Prints the human readable version of *cmask into *mask, (decompiles
 * cmask).
 * The area pointed by *mask MUST be big enough (the mask might be up to
 * twice the size of its compiled form if it's made all of \? or \*, and
 * this function can NOT work in place since it might enflate the mask)
 * The printed mask is not identical to the one that was compiled to cmask,
 * infact it is 1) forced to all lowercase, 2) collapsed, both things
 * are supposed to NOT change it's meaning.
 * It returns the number of chars actually written to *mask;
 */

int matchdecomp(char* mask, char const* cmask)
{
  register char* rtb = mask;
  register char const* rcm = cmask;
  register char const* begtail, *endtail;

  if (rtb == NULL)
    return (-1);

  if (rcm == NULL)
    return (-2);

  for (; (*rcm != 'Z'); rcm++, rtb++)
  {
    if ((*rcm == '?') || (*rcm == '*'))
      *rtb++ = '\\';
    if (!((*rtb = ((*rcm == 'A') ? '?' : *rcm))))
      return (rtb - mask);
  };

  begtail = rcm++;
  *rtb++ = '*';

  while (*rcm && (*rcm != 'Z'))
    rcm++;

  endtail = rcm;

  if (*rcm)
  {
    while (*++rcm)
      switch (*rcm)
      {
        case 'A':
          *rtb++ = '?';
          break;
        case 'Z':
          *rtb++ = '*';
          break;
        case '*':
        case '?':
          *rtb++ = '\\';
        default:
          *rtb++ = *rcm;
      };
    *rtb++ = '*';
  };

  for (rcm = endtail; (--rcm) > begtail; *rtb++ = ((*rcm == 'A') ? '?' : *rcm))
    if ((*rcm == '?') || (*rcm == '*'))
      *rtb++ = '\\';

  *rtb = '\000';
  return (rtb - mask);
}

/*
 * mmexec()
 * Checks if a wider compiled mask (wcm/wminlen) completely overrides
 * a more restrict one (rcm/rminlen), basically what mmatch() does for
 * non-compiled masks, returns 0 if the override is true (like mmatch()).
 * "the wider overrides the restrict" means that any string that matches
 * the restrict one _will_ also match the wider one, always. 
 * In this we behave differently from mmatch() because in example we return 
 * true for " a?*cd overrides a*bcd " for wich the override happens for how 
 * we literally defined it, here mmatch() would have returned false.
 * The original concepts and the base alghoritm are copied from mmatch() 
 * written by Run (Carlo Wood), this function is written by
 * Nemesi (Andrea Cocito)
 */

int mmexec(char const* wcm, int wminlen, char const* rcm, int rminlen)
{
  register char const* w, *r, *br, *bw, *rx, *rz;
  register int eat, trash;

  /* First of all rm must have enough non-stars to 'contain' wm */
  if ((trash = rminlen - wminlen) < 0)
    return 1;
  w = wcm;
  r = rcm;
  eat = 0;

  /* Let's start the game, remember that '*' is mapped to 'Z', '?'
     is mapped to 'A' and that head?*??*?chunk*???*tail becomes
     headAAAAZliatAAAZchunk for compiled masks */

  /* Match the head of wm with the head of rm */
  for (; (*r) && (*r != 'Z') && ((*w == *r) || (*w == 'A')); r++, w++);
  if (*r == 'Z')
    while (*w == 'A')           /* Eat extra '?' before '*' in wm if got '*' in rm */
      w++, eat++;
  if (*w != 'Z')                /* head1<any>.. can't match head2<any>.. */
    return ((*w) || (*r)) ? 1 : 0;      /* and head<nul> matches only head<nul> */
  if (!*++w)
    return 0;                   /* headZ<nul> matches head<anything>    */

  /* Does rm have any stars in it ? let's check */
  for (rx = r; *r && (*r != 'Z'); r++);
  if (!*r)
  {
    /* rm has no stars and thus isn't a mask but it's just a flat
       string: special handling occurs here, note that eat must be 0 here */

    /* match the tail */
    if (*w != 'Z')
    {
      for (; r--, (*w) && ((*w == *r) || (*w == 'A')); w++);
      if (*w != 'Z')            /* headZliat1<any> fails on head<any>2tail  */
        return (*w) ? 1 : 0;    /* but headZliat<nul> matches head<any>tail */
    };

    /* match the chunks */
    while (1)
    {                           /* This loop can't break but only return   */

      for (bw = w++; (*w != *rx); rx++) /* Seek the 1st char of the chunk */
        if (--trash < 0)        /* See if we can trash one more char of rm */
          return 1;             /* If not we can only fail of course       */
      for (r = ++rx, w++; (*w) && ((*w == *r) || (*w == 'A')); r++, w++);
      if (!*w)                  /* Did last loop match the rest of chunk ? */
        return 0;               /* ... Yes, end of wm, matched !           */
      if (*w != 'Z')
      {                         /* ... No, hitted non-star                 */
        w = bw;                 /* Rollback at beginning of chunk          */
        if (--trash < 0)        /* Trashed the char where this try started */
          return 1;             /* if we can't trash more chars fail       */
      }
      else
      {
        rx = r;                 /* Successfully matched a chunk, move rx   */
      };                        /* and go on with the next one             */
    };
  };

  /* rm has at least one '*' and thus is a 'real' mask */
  rz = r++;                     /* rx = unused of head, rz = beg-tail */

  /* Match the tail of wm (if any) against the tail of rm */
  if (*w != 'Z')
  {
    for (; (*w) && (*r != 'Z') && ((*w == *r) || (*w == 'A')); w++, r++);
    if (*r == 'Z')              /* extra '?' before tail are fluff, just flush 'em */
      while (*w == 'A')
        w++;
    if (*w != 'Z')              /* We aren't matching a chunk, can't rollback      */
      return (*w) ? 1 : 0;
  };

  /* Match the chunks of wm against what remains of the head of rm */
  while (1)
  {
    bw = w;
    for (bw++; (rx < rz) && (*bw != *rx); rx++) /* Seek the first           */
      if (--trash < 0)          /* waste some trash reserve */
        return 1;
    if (!(rx < rz))             /* head finished            */
      break;
    for (bw++, (br = ++rx);
        (br < rz) && (*bw) && ((*bw == *br) || (*bw == 'A')); br++, bw++);
    if (!(br < rz))             /* Note that we didn't use any 'eat' char yet, if  */
      while (*bw == 'A')        /* there were eat-en chars the head would be over  */
        bw++, eat++;            /* Happens only at end of head, and eat is still 0 */
    if (!*bw)
      return 0;
    if (*bw != 'Z')
    {
      eat = 0;
      if (!(br < rz))
      {                         /* If we failed because we got the end of head */
        trash -= (br - rx);     /* it makes no sense to rollback, just trash   */
        if (--trash < 0)        /* all the rest of the head wich isn't long    */
          return 1;             /* enough for this chunk and go out of this    */
        break;                  /* loop, then we try with the chunks of rm     */
      };
      if (--trash < 0)
        return 1;
    }
    else
    {
      w = bw;
      rx = br;
    };
  };

  /* Match the unused chunks of wm against the chunks of rm */
  rx = r;
  for (; *r && (*r != 'Z'); r++);
  rz = r;
  if (*r++)
  {
    while (*r)
    {
      bw = w;
      while (eat && *r)         /* the '?' we had eated make us skip as many chars */
        if (*r++ != 'Z')        /* here, but can't skip stars or trailing zero     */
          eat--;
      for (bw++; (*r) && (*bw != *r); r++)
        if ((*r != 'Z') && (--trash < 0))
          return 1;
      if (!*r)
        break;
      for ((br = ++r), bw++;
          (*br) && (*br != 'Z') && ((*bw == *br) || (*bw == 'A')); br++, bw++);
      if (*br == 'Z')
        while (*bw == 'A')
          bw++, eat++;
      if (!*bw)
        return 0;
      if (*bw != 'Z')
      {
        eat = 0;
        if ((!*br) || (*r == 'Z'))
        {                       /* If we hit the end of rm or a star in it */
          trash -= (br - r);    /* makes no sense to rollback within this  */
          if (trash < 0)        /* same chunk of br, skip it all and then  */
            return 1;           /* either rollback or break this loop if   */
          if (!*br)             /* it was the end of rm                    */
            break;
          r = br;
        };
        if (--trash < 0)
          return 1;
      }
      else
      {
        r = br;
        w = bw;
      };
    };
  };

  /* match the remaining chunks of wm against what remains of the tail of rm */
  r = rz - eat - 1;             /* can't have <nul> or 'Z'within the tail, so just move r */
  while (r >= rx)
  {
    bw = w;
    for (bw++; (*bw != *r); r--)
      if (--trash < 0)
        return 1;
    if (!(r >= rx))
      return 1;
    for ((br = --r), bw++;
        (*bw) && (br >= rx) && ((*bw == *br) || (*bw == 'A')); br--, bw++);
    if (!*bw)
      return 0;
    if (!(br >= rx))
      return 1;
    if (*bw != 'Z')
    {
      if (--trash < 0)
        return 1;
    }
    else
    {
      r = br;
      w = bw;
    };
  };
  return 1;                     /* Auch... something left out ? Fail */
}

/*
 * matchcompIP()
 * Compiles an IP mask into an in_mask structure
 * The given <mask> can either be:
 * - An usual irc type mask, containing * and or ?
 * - An ip number plus a /bitnumber part, that will only consider
 *   the first "bitnumber" bits of the IP (bitnumber must be in 0-31 range)
 * - An ip number plus a /ip.bit.mask.values that will consider
 *   only the bits marked as 1 in the ip.bit.mask.values
 * In the last two cases both the ip number and the bitmask can specify
 * less than 4 bytes, the missing bytes then default to zero, note that
 * this is *different* from the way inet_aton() does and that this does
 * NOT happen for normal IPmasks (not containing '/')
 * If the returned value is zero the produced in_mask might match some IP,
 * if it's nonzero it will never match anything (and the imask struct is
 * set so that always fails).
 *
 * The returned structure contains 3 fields whose meaning is the following:
 * im.mask = The bits considered significative in the IP
 * im.bits = What these bits should look like to have a match
 * im.fall = If zero means that the above information used as 
 *           ((IP & im.mask) == im.bits) is enough to tell if the compiled
 *           mask matches the given IP, nonzero means that it is needed,
 *           in case they did match, to call also the usual text match
 *           functions, because the mask wasn't "completely compiled"
 *
 * They should be used like:
 * matchcompIP(&im, mask);
 * if ( ((IP & im.mask)!=im.bits)) || (im.fall&&match(mask,inet_ntoa(IP))) )
 *    { handle_non_match } else { handle_match };
 * instead of:
 * if ( match(mask, inet_ntoa(IP)) )
 *    { handle_non_match } else { handle_match };
 * 
 * Note: This function could be smarter when dealing with complex masks,
 *       this implementation is quite lazy and understands only very simple
 *       cases, whatever contains a ? anywhere or contains a '*' that isn't
 *       part of a trailing '.*' will fallback to text-match, this could be 
 *       avoided for masks like 12?3.5.6 12.*.3.4 1.*.*.2 72?72?72?72 and
 *       so on that "could" be completely compiled to IP masks.
 *       If you try to improve this be aware of the fact that ? and *
 *       could match both dots and digits and we _must_ always reject
 *       what doesn't match in textform (like leading zeros and so on),
 *       so it's a LOT more tricky than it might seem. By now most common
 *       cases are optimized.
 */

int matchcompIP(in_mask* imask, char const* mask)
{
  register char const* m = mask;
  register unsigned int bits = 0;
  register unsigned int filt = 0;
  register int unco = 0;
  register int digits = 0;
  register int shift = 24;
  register int tmp = 0;

  do
  {
    switch (*m)
    {
      case '\\':
        if ((m[1] == '\\') || (m[1] == '*') || (m[1] == '?')
            || (m[1] == '\000'))
          break;
        continue;
      case '0':
      case '1':
      case '2':
      case '3':
      case '4':
      case '5':
      case '6':
      case '7':
      case '8':
      case '9':
        if (digits && !tmp)     /* Leading zeros */
          break;
        digits++;
        tmp *= 10;
        tmp += (*m - '0');      /* Can't overflow, INT_MAX > 2559 */
        if (tmp > 255)
          break;
        continue;
      case '\000':
        filt = 0xFFFFFFFF;
        /* Intentional fallthrough */
      case '.':
        if ((!shift) != (!*m))
          break;
        /* Intentional fallthrough */
      case '/':
        bits |= (tmp << shift);
        shift -= 8;
        digits = 0;
        tmp = 0;
        if (*m != '/')
          continue;
        shift = 24;
        do
        {
          m++;
          if (IsDigit(*m))
          {
            if (digits && !tmp) /* Leading zeros */
              break;
            digits++;
            tmp *= 10;
            tmp += (*m - '0');  /* Can't overflow, INT_MAX > 2559 */
            if (tmp > 255)
              break;
          }
          else
          {
            switch (*m)
            {
              case '.':
              case '\000':
                if ((!shift) && (*m))
                  break;
                filt |= (tmp << shift);
                shift -= 8;
                tmp = 0;
                digits = 0;
                continue;
              default:
                break;
            }
            break;
          }
        }
        while (*m);
        if (*m)
          break;
        if (filt && (!(shift < 16)) && (!(filt & 0xE0FFFFFF)))
          filt = 0xFFFFFFFF << (32 - ((filt >> 24)));
        bits &= filt;
        continue;
      case '?':
        unco = 1;
        /* Intentional fallthrough */
      case '*':
        if (digits)
          unco = 1;
        filt = (0xFFFFFFFF << (shift)) << 8;
        while (*++m)
        {
          if (IsDigit(*m))
            unco = 1;
          else
          {
            switch (*m)
            {
              case '.':
                if (m[1] != '*')
                  unco = 1;
                if (!shift)
                  break;
                shift -= 8;
                continue;
              case '?':
                unco = 1;
              case '*':
                continue;
              default:
                break;
            }
            break;
          }
        }
        if (*m)
          break;
        continue;
      default:
        break;
    }

    /* If we get here there is some error and this can't ever match */
    filt = 0;
    bits = ~0;
    unco = 0;
    break;                      /* This time break the loop :) */

  }
  while (*m++);

  imask->bits.s_addr = htonl(bits);
  imask->mask.s_addr = htonl(filt);
  imask->fall = unco;
  return ((bits & ~filt) ? -1 : 0);
}

---