perl_matcher_common.hpp

来自「Boost provides free peer-reviewed portab」· HPP 代码 · 共 972 行 · 第 1/2 页

         if((*m_presult)[i].matched)
            ((*m_presult)[i]).get_captures().push_back((*m_presult)[i]);
   }
#endif
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_word_boundary()
{
   bool b; // indcates whether next character is a word character
   if(position != last)
   {
      // prev and this character must be opposites:
   #if defined(BOOST_REGEX_USE_C_LOCALE) && defined(__GNUC__) && (__GNUC__ == 2) && (__GNUC_MINOR__ < 95)
      b = traits::isctype(*position, m_word_mask);
   #else
      b = traits_inst.isctype(*position, m_word_mask);
   #endif
   }
   else
   {
      b = (m_match_flags & match_not_eow) ? true : false;
   }
   if((position == backstop) && ((m_match_flags & match_prev_avail) == 0))
   {
      if(m_match_flags & match_not_bow)
         b ^= true;
      else
         b ^= false;
   }
   else
   {
      --position;
      b ^= traits_inst.isctype(*position, m_word_mask);
      ++position;
   }
   if(b)
   {
      pstate = pstate->next.p;
      return true;
   }
   return false; // no match if we get to here...
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_within_word()
{
   if(position == last)
      return false;
   // both prev and this character must be m_word_mask:
   bool prev = traits_inst.isctype(*position, m_word_mask);
   {
      bool b;
      if((position == backstop) && ((m_match_flags & match_prev_avail) == 0)) 
         return false;
      else
      {
         --position;
         b = traits_inst.isctype(*position, m_word_mask);
         ++position;
      }
      if(b == prev)
      {
         pstate = pstate->next.p;
         return true;
      }
   }
   return false;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_word_start()
{
   if(position == last)
      return false; // can't be starting a word if we're already at the end of input
   if(!traits_inst.isctype(*position, m_word_mask))
      return false; // next character isn't a word character
   if((position == backstop) && ((m_match_flags & match_prev_avail) == 0))
   {
      if(m_match_flags & match_not_bow)
         return false; // no previous input
   }
   else
   {
      // otherwise inside buffer:
      BidiIterator t(position);
      --t;
      if(traits_inst.isctype(*t, m_word_mask))
         return false; // previous character not non-word
   }
   // OK we have a match:
   pstate = pstate->next.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_word_end()
{
   if((position == backstop) && ((m_match_flags & match_prev_avail) == 0))
      return false;  // start of buffer can't be end of word
   BidiIterator t(position);
   --t;
   if(traits_inst.isctype(*t, m_word_mask) == false)
      return false;  // previous character wasn't a word character

   if(position == last)
   {
      if(m_match_flags & match_not_eow)
         return false; // end of buffer but not end of word
   }
   else
   {
      // otherwise inside buffer:
      if(traits_inst.isctype(*position, m_word_mask))
         return false; // next character is a word character
   }
   pstate = pstate->next.p;
   return true;      // if we fall through to here then we've succeeded
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_buffer_start()
{
   if((position != backstop) || (m_match_flags & match_not_bob))
      return false;
   // OK match:
   pstate = pstate->next.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_buffer_end()
{
   if((position != last) || (m_match_flags & match_not_eob))
      return false;
   // OK match:
   pstate = pstate->next.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_backref()
{
   //
   // Compare with what we previously matched.
   // Note that this succeeds if the backref did not partisipate
   // in the match, this is in line with ECMAScript, but not Perl
   // or PCRE.
   //
   BidiIterator i = (*m_presult)[static_cast<const re_brace*>(pstate)->index].first;
   BidiIterator j = (*m_presult)[static_cast<const re_brace*>(pstate)->index].second;
   while(i != j)
   {
      if((position == last) || (traits_inst.translate(*position, icase) != traits_inst.translate(*i, icase)))
         return false;
      ++i;
      ++position;
   }
   pstate = pstate->next.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_long_set()
{
   typedef typename traits::char_class_type char_class_type;
   // let the traits class do the work:
   if(position == last)
      return false;
   BidiIterator t = re_is_set_member(position, last, static_cast<const re_set_long<char_class_type>*>(pstate), re.get_data(), icase);
   if(t != position)
   {
      pstate = pstate->next.p;
      position = t;
      return true;
   }
   return false;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_set()
{
   if(position == last)
      return false;
   if(static_cast<const re_set*>(pstate)->_map[static_cast<unsigned char>(traits_inst.translate(*position, icase))])
   {
      pstate = pstate->next.p;
      ++position;
      return true;
   }
   return false;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_jump()
{
   pstate = static_cast<const re_jump*>(pstate)->alt.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_combining()
{
   if(position == last)
      return false;
   if(is_combining(traits_inst.translate(*position, icase)))
      return false;
   ++position;
   while((position != last) && is_combining(traits_inst.translate(*position, icase)))
      ++position;
   pstate = pstate->next.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_soft_buffer_end()
{
   if(m_match_flags & match_not_eob)
      return false;
   BidiIterator p(position);
   while((p != last) && is_separator(traits_inst.translate(*p, icase)))++p;
   if(p != last)
      return false;
   pstate = pstate->next.p;
   return true;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_restart_continue()
{
   if(position == search_base)
   {
      pstate = pstate->next.p;
      return true;
   }
   return false;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_backstep()
{
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable:4127)
#endif
   if( ::boost::is_random_access_iterator<BidiIterator>::value)
   {
      std::ptrdiff_t maxlen = ::boost::re_detail::distance(backstop, position);
      if(maxlen < static_cast<const re_brace*>(pstate)->index)
         return false;
      std::advance(position, -static_cast<const re_brace*>(pstate)->index);
   }
   else
   {
      int c = static_cast<const re_brace*>(pstate)->index;
      while(c--)
      {
         if(position == backstop)
            return false;
         --position;
      }
   }
   pstate = pstate->next.p;
   return true;
#ifdef BOOST_MSVC
#pragma warning(pop)
#endif
}

template <class BidiIterator, class Allocator, class traits>
inline bool perl_matcher<BidiIterator, Allocator, traits>::match_assert_backref()
{
   // return true if marked sub-expression N has been matched:
   bool result = (*m_presult)[static_cast<const re_brace*>(pstate)->index].matched;
   pstate = pstate->next.p;
   return result;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::match_toggle_case()
{
   // change our case sensitivity:
   this->icase = static_cast<const re_case*>(pstate)->icase;
   pstate = pstate->next.p;
   return true;
}


template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_any()
{
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable:4127)
#endif
   const unsigned char* _map = re.get_map();
   while(true)
   {
      // skip everything we can't match:
      while((position != last) && !can_start(*position, _map, (unsigned char)mask_any) )
         ++position;
      if(position == last)
      {
         // run out of characters, try a null match if possible:
         if(re.can_be_null())
            return match_prefix();
         break;
      }
      // now try and obtain a match:
      if(match_prefix())
         return true;
      if(position == last)
         return false;
      ++position;
   }
   return false;
#ifdef BOOST_MSVC
#pragma warning(pop)
#endif
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_word()
{
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable:4127)
#endif
   // do search optimised for word starts:
   const unsigned char* _map = re.get_map();
   if((m_match_flags & match_prev_avail) || (position != base))
      --position;
   else if(match_prefix())
      return true;
   do
   {
      while((position != last) && traits_inst.isctype(*position, m_word_mask))
         ++position;
      while((position != last) && !traits_inst.isctype(*position, m_word_mask))
         ++position;
      if(position == last)
         break;

      if(can_start(*position, _map, (unsigned char)mask_any) )
      {
         if(match_prefix())
            return true;
      }
      if(position == last)
         break;
   } while(true);
   return false;
#ifdef BOOST_MSVC
#pragma warning(pop)
#endif
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_line()
{
   // do search optimised for line starts:
   const unsigned char* _map = re.get_map();
   if(match_prefix())
      return true;
   while(position != last)
   {
      while((position != last) && !is_separator(*position))
         ++position;
      if(position == last)
         return false;
      ++position;
      if(position == last)
      {
         if(re.can_be_null() && match_prefix())
            return true;
         return false;
      }

      if( can_start(*position, _map, (unsigned char)mask_any) )
      {
         if(match_prefix())
            return true;
      }
      if(position == last)
         return false;
      //++position;
   }
   return false;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_buf()
{
   if((position == base) && ((m_match_flags & match_not_bob) == 0))
      return match_prefix();
   return false;
}

template <class BidiIterator, class Allocator, class traits>
bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_lit()
{
#if 0
   if(position == last)
      return false; // can't possibly match if we're at the end already

   unsigned type = (m_match_flags & match_continuous) ? 
      static_cast<unsigned int>(regbase::restart_continue) 
         : static_cast<unsigned int>(re.get_restart_type());

   const kmp_info<char_type>* info = access::get_kmp(re);
   int len = info->len;
   const char_type* x = info->pstr;
   int j = 0; 
   while (position != last) 
   {
      while((j > -1) && (x[j] != traits_inst.translate(*position, icase))) 
         j = info->kmp_next[j];
      ++position;
      ++j;
      if(j >= len) 
      {
         if(type == regbase::restart_fixed_lit)
         {
            std::advance(position, -j);
            restart = position;
            std::advance(restart, len);
            m_result.set_first(position);
            m_result.set_second(restart);
            position = restart;
            return true;
         }
         else
         {
            restart = position;
            std::advance(position, -j);
            if(match_prefix())
               return true;
            else
            {
               for(int k = 0; (restart != position) && (k < j); ++k, --restart)
                     {} // dwa 10/20/2000 - warning suppression for MWCW
               if(restart != last)
                  ++restart;
               position = restart;
               j = 0;  //we could do better than this...
            }
         }
      }
   }
   if((m_match_flags & match_partial) && (position == last) && j)
   {
      // we need to check for a partial match:
      restart = position;
      std::advance(position, -j);
      return match_prefix();
   }
#endif
   return false;
}

} // namespace re_detail

} // namespace boost

#ifdef BOOST_MSVC
#  pragma warning(pop)
#endif

#ifdef __BORLANDC__
#  pragma option pop
#endif
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable: 4103)
#endif
#ifdef BOOST_HAS_ABI_HEADERS
#  include BOOST_ABI_SUFFIX
#endif
#ifdef BOOST_MSVC
#pragma warning(pop)
#endif

#endif