regex_compile.hpp

来自「CGAL is a collaborative effort of severa」· HPP 代码 · 共 1,861 行 · 第 1/5 页

/*
 *
 * Copyright (c) 1998-2002
 * Dr John Maddock
 *
 * Use, modification and distribution are subject to the 
 * Boost Software License, Version 1.0. (See accompanying file 
 * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
 *
 */
 
 /*
  *   LOCATION:    see http://www.boost.org for most recent version.
  *   FILE         regex_compile.hpp
  *   VERSION      see <boost/version.hpp>
  *   DESCRIPTION: Declares reg_expression<> member functions.  This is
  *                an internal header file, do not include directly.
  */

#ifndef BOOST_REGEX_COMPILE_HPP
#define BOOST_REGEX_COMPILE_HPP

namespace boost{
#ifdef __BORLANDC__
   #pragma option push -a8 -b -Vx -Ve -pc  -w-8004
#endif
   namespace re_detail{


template <class traits>
struct kmp_translator
{
   typedef typename traits::char_type char_type;
   bool icase;
   const traits* pt;
   kmp_translator(bool c, traits* p) : icase(c), pt(p) {}
   char_type operator()(char_type c)
   {
      return pt->translate(c, icase);
   }
};


template <class charT, class traits_type, class Allocator>
bool BOOST_REGEX_CALL re_maybe_set_member(charT c,
                                 const re_set_long* set_,
                                 const reg_expression<charT, traits_type, Allocator>& e)
{
   const charT* p = reinterpret_cast<const charT*>(set_+1);
   bool icase = e.flags() & regbase::icase;
   charT col = e.get_traits().translate(c, icase);
   for(unsigned int i = 0; i < set_->csingles; ++i)
   {
      if(col == *p)
         return set_->isnot ? false : true;

      while(*p)++p;
      ++p;     // skip null
   }
   return set_->isnot ? true : false;
}

} // namespace re_detail


template <class charT, class traits, class Allocator>
inline bool BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::can_start(charT c, const unsigned char* _map, unsigned char mask, const re_detail::_wide_type&)
{
   if((traits_size_type)(traits_uchar_type)c >= 256)
      return true;
   return BOOST_REGEX_MAKE_BOOL(_map[(traits_uchar_type)c] & mask);
}

template <class charT, class traits, class Allocator>
inline bool BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::can_start(charT c, const unsigned char* _map, unsigned char mask, const re_detail::_narrow_type&)
{
   return BOOST_REGEX_MAKE_BOOL(_map[(traits_uchar_type)c] & mask);
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>::reg_expression(const Allocator& a)
    : regbase(), data(a), pkmp(0), error_code_(REG_EMPTY), _expression(0)
{
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>::reg_expression(const charT* p, flag_type f, const Allocator& a)
    : data(a), pkmp(0), error_code_(REG_EMPTY), _expression(0)
{
   set_expression(p, f | regbase::use_except);
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>::reg_expression(const charT* p1, const charT* p2, flag_type f, const Allocator& a)
    : data(a), pkmp(0), error_code_(REG_EMPTY), _expression(0)
{
    set_expression(p1, p2, f | regbase::use_except);
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>::reg_expression(const charT* p, size_type len, flag_type f, const Allocator& a)
    : data(a), pkmp(0), error_code_(REG_EMPTY), _expression(0)
{
    set_expression(p, p + len, f | regbase::use_except);
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>::reg_expression(const reg_expression<charT, traits, Allocator>& e)
   : regbase(e), data(e.allocator()), pkmp(0), error_code_(REG_EMPTY), _expression(0)
{
   //
   // we do a deep copy only if e is a valid expression, otherwise fail.
   //
   if(e.error_code() == 0)
   {
      const charT* pe = e.expression();
      set_expression(pe, pe + e._expression_len, e.flags() | regbase::use_except);
   }
   else
   {
      _flags = e.flags() & ~(regbase::use_except);
      fail(e.error_code());
   }
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>::~reg_expression()
{
   if(pkmp)
      re_detail::kmp_free(pkmp, data.allocator());
}

template <class charT, class traits, class Allocator>
reg_expression<charT, traits, Allocator>& BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::operator=(const reg_expression<charT, traits, Allocator>& e)
{
   //
   // we do a deep copy only if e is a valid expression, otherwise fail.
   //
   if(this == &e) return *this;
   _flags = use_except;
   fail(e.error_code());
   if(error_code() == 0)
      set_expression(e._expression, e._expression + e._expression_len, e.flags() | regbase::use_except);
   return *this;
}

template <class charT, class traits, class Allocator>
inline bool BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::operator==(const reg_expression<charT, traits, Allocator>& e)const
{
   return (_flags == e.flags())
           && (_expression_len == e._expression_len)
           && (std::memcmp(_expression, e._expression, _expression_len * sizeof(charT)) == 0);
}

template <class charT, class traits, class Allocator>
bool BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::operator<(const reg_expression<charT, traits, Allocator>& e)const
{
   //
   // we can't offer a diffinitive ordering, but we can be consistant:
   if(_flags != e.flags()) return _flags < e.flags();
   if(_expression_len != e._expression_len) return _expression_len < e._expression_len;
   return std::memcmp(expression(), e.expression(), _expression_len);
}

template <class charT, class traits, class Allocator>
Allocator BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::allocator()const
{
    return data.allocator();
}

template <class charT, class traits, class Allocator>
Allocator BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::get_allocator()const
{
    return data.allocator();
}

template <class charT, class traits, class Allocator>
unsigned int BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::parse_inner_set(const charT*& first, const charT* last)
{
   //
   // we have an inner [...] construct
   //
   jm_assert(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*first) == traits_type::syntax_open_set);
   const charT* base = first;
   while( (first != last)
      && (traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*first) != traits_type::syntax_close_set) )
         ++first;
   if(first == last)
      return 0;
   ++first;
   if((first-base) < 5)
      return 0;
   if(*(base+1) != *(first-2))
      return 0;
   unsigned int result = traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*(base+1));
   if((result == traits_type::syntax_colon) && ((first-base) == 5))
   {
      unsigned type = traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*(base+2));
      if((type == traits_type::syntax_left_word) || (type == traits_type::syntax_right_word))
         return type;
   }
   return ((result == traits_type::syntax_colon) || (result == traits_type::syntax_dot) || (result == traits_type::syntax_equal)) ? result : 0;
}


template <class charT, class traits, class Allocator>
bool BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::skip_space(const charT*& first, const charT* last)
{
   //
   // returns true if we get to last:
   //
   while((first != last) && (traits_inst.is_class(*first, traits_type::char_class_space) == true))
   {
      ++first;
   }
   return first == last;
}

template <class charT, class traits, class Allocator>
void BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::parse_range(const charT*& ptr, const charT* end, unsigned& min, unsigned& max)
{
   //
   // we have {x} or {x,} or {x,y} NB no spaces inside braces
   // anything else is illegal
   // On input ptr points to "{"
   //
   ++ptr;
   if(skip_space(ptr, end))
   {
      fail(REG_EBRACE);
      return;
   }
   if(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*ptr) != traits_type::syntax_digit)
   {
      fail(REG_BADBR);
      return;
   }
   min = traits_inst.toi(ptr, end, 10);
   if(skip_space(ptr, end))
   {
      fail(REG_EBRACE);
      return;
   }
   if(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*ptr) == traits_type::syntax_comma)
   {
      //we have a second interval:
      ++ptr;
      if(skip_space(ptr, end))
      {
         fail(REG_EBRACE);
         return;
      }
      if(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*ptr) == traits_type::syntax_digit)
         max = traits_inst.toi(ptr, end, 10);
      else
         max = (unsigned)-1;
   }
   else
      max = min;

   // validate input:
   if(skip_space(ptr, end))
   {
      fail(REG_EBRACE);
      return;
   }
   if(max < min)
   {
      fail(REG_ERANGE);
      return;
   }
   if(_flags & bk_braces)
   {
      if(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*ptr) != traits_type::syntax_slash)
      {
         fail(REG_BADBR);
         return;
      }
      else
      {
         // back\ is OK now check the }
         ++ptr;
         if((ptr == end) || (traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*ptr) != traits_type::syntax_close_brace))
         {
            fail(REG_BADBR);
            return;
         }
      }
   }
   else if(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)*ptr) != traits_type::syntax_close_brace)
   {
      fail(REG_BADBR);
      return;
   }
}

template <class charT, class traits, class Allocator>
charT BOOST_REGEX_CALL reg_expression<charT, traits, Allocator>::parse_escape(const charT*& first, const charT* last)
{
   charT c(*first);
   traits_size_type c_unsigned = (traits_size_type)(traits_uchar_type)*first;
   // this is only used for the switch(), but cannot be folded in
   // due to a bug in Comeau 4.2.44beta3
   traits_size_type syntax = traits_inst.syntax_type(c_unsigned);
   switch(syntax)
   {
   case traits_type::syntax_a:
      c = '\a';
      ++first;
      break;
   case traits_type::syntax_f:
      c = '\f';
      ++first;
      break;
   case traits_type::syntax_n:
      c = '\n';
      ++first;
      break;
   case traits_type::syntax_r:
      c = '\r';
      ++first;
      break;
   case traits_type::syntax_t:
      c = '\t';
      ++first;
      break;
   case traits_type::syntax_v:
      c = '\v';
      ++first;
      break;
   case traits_type::syntax_x:
      ++first;
      if(first == last)
      {
         fail(REG_EESCAPE);
         break;
      }
      // maybe have \x{ddd}
      if(traits_inst.syntax_type((traits_size_type)(traits_uchar_type)(*first)) == traits_type::syntax_open_brace)
      {
         ++first;
         if(first == last)
         {
            fail(REG_EESCAPE);
            break;
         }
         if(traits_inst.is_class(*first, traits_type::char_class_xdigit) == false)
         {
            fail(REG_BADBR);
            break;
         }
         c = (charT)traits_inst.toi(first, last, -16);
         if((first == last) || (traits_inst.syntax_type((traits_size_type)(traits_uchar_type)(*first)) != traits_type::syntax_close_brace))
         {
            fail(REG_BADBR);
         }
         ++first;
         break;
      }
      else
      {
         if(traits_inst.is_class(*first, traits_type::char_class_xdigit) == false)
         {
            fail(REG_BADBR);
            break;
         }
         c = (charT)traits_inst.toi(first, last, -16);
      }
      break;
   case traits_type::syntax_c:
      ++first;
      if(first == last)
      {