boost/regex/v4/regex_split.hpp

   1 /*
   2  *
   3  * Copyright (c) 1998-2002
   4  * John Maddock
   5  *
   6  * Use, modification and distribution are subject to the
   7  * Boost Software License, Version 1.0. (See accompanying file
   8  * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
   9  *
  10  */
  11
  12  /*
  13   *   LOCATION:    see http://www.boost.org for most recent version.
  14   *   FILE         regex_split.hpp
  15   *   VERSION      see <boost/version.hpp>
  16   *   DESCRIPTION: Implements regex_split and associated functions.
  17   *                Note this is an internal header file included
  18   *                by regex.hpp, do not include on its own.
  19   */
  20
  21 #ifndef BOOST_REGEX_SPLIT_HPP
  22 #define BOOST_REGEX_SPLIT_HPP
  23
  24 namespace boost{
  25
  26 #ifdef BOOST_MSVC
  27 #pragma warning(push)
  28 #pragma warning(disable: 4103)
  29 #endif
  30 #ifdef BOOST_HAS_ABI_HEADERS
  31 #  include BOOST_ABI_PREFIX
  32 #endif
  33 #ifdef BOOST_MSVC
  34 #pragma warning(pop)
  35 #endif
  36
  37 #ifdef BOOST_MSVC
  38 #  pragma warning(push)
  39 #  pragma warning(disable: 4800)
  40 #endif
  41
  42 namespace re_detail{
  43
  44 template <class charT>
  45 const basic_regex<charT>& get_default_expression(charT)
  46 {
  47    static const charT expression_text[4] = { '\\', 's', '+', '\00', };
  48    static const basic_regex<charT> e(expression_text);
  49    return e;
  50 }
  51
  52 template <class OutputIterator, class charT, class Traits1, class Alloc1>
  53 class split_pred
  54 {
  55    typedef std::basic_string<charT, Traits1, Alloc1> string_type;
  56    typedef typename string_type::const_iterator iterator_type;
  57    iterator_type* p_last;
  58    OutputIterator* p_out;
  59    std::size_t* p_max;
  60    std::size_t initial_max;
  61 public:
  62    split_pred(iterator_type* a, OutputIterator* b, std::size_t* c)
  63       : p_last(a), p_out(b), p_max(c), initial_max(*c) {}
  64
  65    bool operator()(const match_results<iterator_type>& what);
  66 };
  67
  68 template <class OutputIterator, class charT, class Traits1, class Alloc1>
  69 bool split_pred<OutputIterator, charT, Traits1, Alloc1>::operator()
  70    (const match_results<iterator_type>& what)
  71 {
  72    *p_last = what[0].second;
  73    if(what.size() > 1)
  74    {
  75       // output sub-expressions only:
  76       for(unsigned i = 1; i < what.size(); ++i)
  77       {
  78          *(*p_out) = what.str(i);
  79          ++(*p_out);
  80          if(0 == --*p_max) return false;
  81       }
  82       return *p_max != 0;
  83    }
  84    else
  85    {
  86       // output $` only if it's not-null or not at the start of the input:
  87       const sub_match<iterator_type>& sub = what[-1];
  88       if((sub.first != sub.second) || (*p_max != initial_max))
  89       {
  90          *(*p_out) = sub.str();
  91          ++(*p_out);
  92          return --*p_max;
  93       }
  94    }
  95    //
  96    // initial null, do nothing:
  97    return true;
  98 }
  99
 100 } // namespace re_detail
 101
 102 template <class OutputIterator, class charT, class Traits1, class Alloc1, class Traits2>
 103 std::size_t regex_split(OutputIterator out,
 104                    std::basic_string<charT, Traits1, Alloc1>& s,
 105                    const basic_regex<charT, Traits2>& e,
 106                    match_flag_type flags,
 107                    std::size_t max_split)
 108 {
 109    typedef typename std::basic_string<charT, Traits1, Alloc1>::const_iterator  ci_t;
 110    typedef typename match_results<ci_t>::allocator_type                        match_allocator;
 111    ci_t last = s.begin();
 112    std::size_t init_size = max_split;
 113    re_detail::split_pred<OutputIterator, charT, Traits1, Alloc1> pred(&last, &out, &max_split);
 114    ci_t i, j;
 115    i = s.begin();
 116    j = s.end();
 117    regex_grep(pred, i, j, e, flags);
 118    //
 119    // if there is still input left, do a final push as long as max_split
 120    // is not exhausted, and we're not splitting sub-expressions rather
 121    // than whitespace:
 122    if(max_split && (last != s.end()) && (e.mark_count() == 1))
 123    {
 124       *out = std::basic_string<charT, Traits1, Alloc1>((ci_t)last, (ci_t)s.end());
 125       ++out;
 126       last = s.end();
 127       --max_split;
 128    }
 129    //
 130    // delete from the string everything that has been processed so far:
 131    s.erase(0, last - s.begin());
 132    //
 133    // return the number of new records pushed:
 134    return init_size - max_split;
 135 }
 136
 137 template <class OutputIterator, class charT, class Traits1, class Alloc1, class Traits2>
 138 inline std::size_t regex_split(OutputIterator out,
 139                    std::basic_string<charT, Traits1, Alloc1>& s,
 140                    const basic_regex<charT, Traits2>& e,
 141                    match_flag_type flags = match_default)
 142 {
 143    return regex_split(out, s, e, flags, UINT_MAX);
 144 }
 145
 146 template <class OutputIterator, class charT, class Traits1, class Alloc1>
 147 inline std::size_t regex_split(OutputIterator out,
 148                    std::basic_string<charT, Traits1, Alloc1>& s)
 149 {
 150    return regex_split(out, s, re_detail::get_default_expression(charT(0)), match_default, UINT_MAX);
 151 }
 152
 153 #ifdef BOOST_MSVC
 154 #  pragma warning(pop)
 155 #endif
 156
 157 #ifdef BOOST_MSVC
 158 #pragma warning(push)
 159 #pragma warning(disable: 4103)
 160 #endif
 161 #ifdef BOOST_HAS_ABI_HEADERS
 162 #  include BOOST_ABI_SUFFIX
 163 #endif
 164 #ifdef BOOST_MSVC
 165 #pragma warning(pop)
 166 #endif
 167
 168 } // namespace boost
 169
 170 #endif
 171
 172