| 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_HAS_ABI_HEADERS |
|---|
| 27 | # include BOOST_ABI_PREFIX |
|---|
| 28 | #endif |
|---|
| 29 | |
|---|
| 30 | namespace re_detail{ |
|---|
| 31 | |
|---|
| 32 | template <class charT> |
|---|
| 33 | const basic_regex<charT>& get_default_expression(charT) |
|---|
| 34 | { |
|---|
| 35 | static const charT expression_text[4] = { '\\', 's', '+', '\00', }; |
|---|
| 36 | static const basic_regex<charT> e(expression_text); |
|---|
| 37 | return e; |
|---|
| 38 | } |
|---|
| 39 | |
|---|
| 40 | template <class OutputIterator, class charT, class Traits1, class Alloc1> |
|---|
| 41 | class split_pred |
|---|
| 42 | { |
|---|
| 43 | typedef std::basic_string<charT, Traits1, Alloc1> string_type; |
|---|
| 44 | typedef typename string_type::const_iterator iterator_type; |
|---|
| 45 | iterator_type* p_last; |
|---|
| 46 | OutputIterator* p_out; |
|---|
| 47 | std::size_t* p_max; |
|---|
| 48 | std::size_t initial_max; |
|---|
| 49 | public: |
|---|
| 50 | split_pred(iterator_type* a, OutputIterator* b, std::size_t* c) |
|---|
| 51 | : p_last(a), p_out(b), p_max(c), initial_max(*c) {} |
|---|
| 52 | |
|---|
| 53 | bool operator()(const match_results<iterator_type>& what); |
|---|
| 54 | }; |
|---|
| 55 | |
|---|
| 56 | template <class OutputIterator, class charT, class Traits1, class Alloc1> |
|---|
| 57 | bool split_pred<OutputIterator, charT, Traits1, Alloc1>::operator() |
|---|
| 58 | (const match_results<iterator_type>& what) |
|---|
| 59 | { |
|---|
| 60 | *p_last = what[0].second; |
|---|
| 61 | if(what.size() > 1) |
|---|
| 62 | { |
|---|
| 63 | // output sub-expressions only: |
|---|
| 64 | for(unsigned i = 1; i < what.size(); ++i) |
|---|
| 65 | { |
|---|
| 66 | *(*p_out) = what.str(i); |
|---|
| 67 | ++(*p_out); |
|---|
| 68 | if(0 == --*p_max) return false; |
|---|
| 69 | } |
|---|
| 70 | return *p_max != 0; |
|---|
| 71 | } |
|---|
| 72 | else |
|---|
| 73 | { |
|---|
| 74 | // output $` only if it's not-null or not at the start of the input: |
|---|
| 75 | const sub_match<iterator_type>& sub = what[-1]; |
|---|
| 76 | if((sub.first != sub.second) || (*p_max != initial_max)) |
|---|
| 77 | { |
|---|
| 78 | *(*p_out) = sub.str(); |
|---|
| 79 | ++(*p_out); |
|---|
| 80 | return --*p_max; |
|---|
| 81 | } |
|---|
| 82 | } |
|---|
| 83 | // |
|---|
| 84 | // initial null, do nothing: |
|---|
| 85 | return true; |
|---|
| 86 | } |
|---|
| 87 | |
|---|
| 88 | } // namespace re_detail |
|---|
| 89 | |
|---|
| 90 | template <class OutputIterator, class charT, class Traits1, class Alloc1, class Traits2> |
|---|
| 91 | std::size_t regex_split(OutputIterator out, |
|---|
| 92 | std::basic_string<charT, Traits1, Alloc1>& s, |
|---|
| 93 | const basic_regex<charT, Traits2>& e, |
|---|
| 94 | match_flag_type flags, |
|---|
| 95 | std::size_t max_split) |
|---|
| 96 | { |
|---|
| 97 | typedef typename std::basic_string<charT, Traits1, Alloc1>::const_iterator ci_t; |
|---|
| 98 | typedef typename match_results<ci_t>::allocator_type match_allocator; |
|---|
| 99 | ci_t last = s.begin(); |
|---|
| 100 | std::size_t init_size = max_split; |
|---|
| 101 | re_detail::split_pred<OutputIterator, charT, Traits1, Alloc1> pred(&last, &out, &max_split); |
|---|
| 102 | ci_t i, j; |
|---|
| 103 | i = s.begin(); |
|---|
| 104 | j = s.end(); |
|---|
| 105 | regex_grep(pred, i, j, e, flags); |
|---|
| 106 | // |
|---|
| 107 | // if there is still input left, do a final push as long as max_split |
|---|
| 108 | // is not exhausted, and we're not splitting sub-expressions rather |
|---|
| 109 | // than whitespace: |
|---|
| 110 | if(max_split && (last != s.end()) && (e.mark_count() == 1)) |
|---|
| 111 | { |
|---|
| 112 | *out = std::basic_string<charT, Traits1, Alloc1>((ci_t)last, (ci_t)s.end()); |
|---|
| 113 | ++out; |
|---|
| 114 | last = s.end(); |
|---|
| 115 | --max_split; |
|---|
| 116 | } |
|---|
| 117 | // |
|---|
| 118 | // delete from the string everything that has been processed so far: |
|---|
| 119 | s.erase(0, last - s.begin()); |
|---|
| 120 | // |
|---|
| 121 | // return the number of new records pushed: |
|---|
| 122 | return init_size - max_split; |
|---|
| 123 | } |
|---|
| 124 | |
|---|
| 125 | template <class OutputIterator, class charT, class Traits1, class Alloc1, class Traits2> |
|---|
| 126 | inline std::size_t regex_split(OutputIterator out, |
|---|
| 127 | std::basic_string<charT, Traits1, Alloc1>& s, |
|---|
| 128 | const basic_regex<charT, Traits2>& e, |
|---|
| 129 | match_flag_type flags = match_default) |
|---|
| 130 | { |
|---|
| 131 | return regex_split(out, s, e, flags, UINT_MAX); |
|---|
| 132 | } |
|---|
| 133 | |
|---|
| 134 | template <class OutputIterator, class charT, class Traits1, class Alloc1> |
|---|
| 135 | inline std::size_t regex_split(OutputIterator out, |
|---|
| 136 | std::basic_string<charT, Traits1, Alloc1>& s) |
|---|
| 137 | { |
|---|
| 138 | return regex_split(out, s, re_detail::get_default_expression(charT(0)), match_default, UINT_MAX); |
|---|
| 139 | } |
|---|
| 140 | |
|---|
| 141 | #ifdef BOOST_HAS_ABI_HEADERS |
|---|
| 142 | # include BOOST_ABI_SUFFIX |
|---|
| 143 | #endif |
|---|
| 144 | |
|---|
| 145 | } // namespace boost |
|---|
| 146 | |
|---|
| 147 | #endif |
|---|
| 148 | |
|---|
| 149 | |
|---|