| 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: primary_transform.hpp |
|---|
| 15 | * VERSION: see <boost/version.hpp> |
|---|
| 16 | * DESCRIPTION: Heuristically determines the sort string format in use |
|---|
| 17 | * by the current locale. |
|---|
| 18 | */ |
|---|
| 19 | |
|---|
| 20 | #ifndef BOOST_REGEX_PRIMARY_TRANSFORM |
|---|
| 21 | #define BOOST_REGEX_PRIMARY_TRANSFORM |
|---|
| 22 | |
|---|
| 23 | #ifdef BOOST_HAS_ABI_HEADERS |
|---|
| 24 | # include BOOST_ABI_PREFIX |
|---|
| 25 | #endif |
|---|
| 26 | |
|---|
| 27 | namespace boost{ |
|---|
| 28 | namespace re_detail{ |
|---|
| 29 | |
|---|
| 30 | |
|---|
| 31 | enum{ |
|---|
| 32 | sort_C, |
|---|
| 33 | sort_fixed, |
|---|
| 34 | sort_delim, |
|---|
| 35 | sort_unknown |
|---|
| 36 | }; |
|---|
| 37 | |
|---|
| 38 | template <class S, class charT> |
|---|
| 39 | unsigned count_chars(const S& s, charT c) |
|---|
| 40 | { |
|---|
| 41 | // |
|---|
| 42 | // Count how many occurances of character c occur |
|---|
| 43 | // in string s: if c is a delimeter between collation |
|---|
| 44 | // fields, then this should be the same value for all |
|---|
| 45 | // sort keys: |
|---|
| 46 | // |
|---|
| 47 | unsigned int count = 0; |
|---|
| 48 | for(unsigned pos = 0; pos < s.size(); ++pos) |
|---|
| 49 | { |
|---|
| 50 | if(s[pos] == c) ++count; |
|---|
| 51 | } |
|---|
| 52 | return count; |
|---|
| 53 | } |
|---|
| 54 | |
|---|
| 55 | |
|---|
| 56 | template <class traits, class charT> |
|---|
| 57 | unsigned find_sort_syntax(const traits* pt, charT* delim) |
|---|
| 58 | { |
|---|
| 59 | // |
|---|
| 60 | // compare 'a' with 'A' to see how similar they are, |
|---|
| 61 | // should really use a-accute but we can't portably do that, |
|---|
| 62 | // |
|---|
| 63 | typedef typename traits::string_type string_type; |
|---|
| 64 | typedef typename traits::char_type char_type; |
|---|
| 65 | |
|---|
| 66 | // Suppress incorrect warning for MSVC |
|---|
| 67 | (void)pt; |
|---|
| 68 | |
|---|
| 69 | char_type a[2] = {'a', '\0', }; |
|---|
| 70 | string_type sa(pt->transform(a, a+1)); |
|---|
| 71 | if(sa == a) |
|---|
| 72 | { |
|---|
| 73 | *delim = 0; |
|---|
| 74 | return sort_C; |
|---|
| 75 | } |
|---|
| 76 | char_type A[2] = { 'A', '\0', }; |
|---|
| 77 | string_type sA(pt->transform(A, A+1)); |
|---|
| 78 | char_type c[2] = { ';', '\0', }; |
|---|
| 79 | string_type sc(pt->transform(c, c+1)); |
|---|
| 80 | |
|---|
| 81 | int pos = 0; |
|---|
| 82 | while((pos <= static_cast<int>(sa.size())) && (pos <= static_cast<int>(sA.size())) && (sa[pos] == sA[pos])) ++pos; |
|---|
| 83 | --pos; |
|---|
| 84 | if(pos < 0) |
|---|
| 85 | { |
|---|
| 86 | *delim = 0; |
|---|
| 87 | return sort_unknown; |
|---|
| 88 | } |
|---|
| 89 | // |
|---|
| 90 | // at this point sa[pos] is either the end of a fixed width field |
|---|
| 91 | // or the character that acts as a delimiter: |
|---|
| 92 | // |
|---|
| 93 | charT maybe_delim = sa[pos]; |
|---|
| 94 | if((pos != 0) && (count_chars(sa, maybe_delim) == count_chars(sA, maybe_delim)) && (count_chars(sa, maybe_delim) == count_chars(sc, maybe_delim))) |
|---|
| 95 | { |
|---|
| 96 | *delim = maybe_delim; |
|---|
| 97 | return sort_delim; |
|---|
| 98 | } |
|---|
| 99 | // |
|---|
| 100 | // OK doen't look like a delimiter, try for fixed width field: |
|---|
| 101 | // |
|---|
| 102 | if((sa.size() == sA.size()) && (sa.size() == sc.size())) |
|---|
| 103 | { |
|---|
| 104 | // note assumes that the fixed width field is less than |
|---|
| 105 | // (numeric_limits<charT>::max)(), should be true for all types |
|---|
| 106 | // I can't imagine 127 character fields... |
|---|
| 107 | *delim = static_cast<charT>(++pos); |
|---|
| 108 | return sort_fixed; |
|---|
| 109 | } |
|---|
| 110 | // |
|---|
| 111 | // don't know what it is: |
|---|
| 112 | // |
|---|
| 113 | *delim = 0; |
|---|
| 114 | return sort_unknown; |
|---|
| 115 | } |
|---|
| 116 | |
|---|
| 117 | |
|---|
| 118 | } // namespace re_detail |
|---|
| 119 | } // namespace boost |
|---|
| 120 | |
|---|
| 121 | #ifdef BOOST_HAS_ABI_HEADERS |
|---|
| 122 | # include BOOST_ABI_SUFFIX |
|---|
| 123 | #endif |
|---|
| 124 | |
|---|
| 125 | #endif |
|---|
| 126 | |
|---|
| 127 | |
|---|
| 128 | |
|---|
| 129 | |
|---|
| 130 | |
|---|
| 131 | |
|---|
| 132 | |
|---|