| 1 | /*============================================================================= |
| 2 | Copyright (c) 2001-2003 Daniel Nuffer |
| 3 | Copyright (c) 2002-2003 Hartmut Kaiser |
| 4 | http://spirit.sourceforge.net/ |
| 5 | |
| 6 | Use, modification and distribution is subject to the Boost Software |
| 7 | License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at |
| 8 | http://www.boost.org/LICENSE_1_0.txt) |
| 9 | =============================================================================*/ |
| 10 | #ifndef BOOST_SPIRIT_ESCAPE_CHAR_IPP |
| 11 | #define BOOST_SPIRIT_ESCAPE_CHAR_IPP |
| 12 | |
| 13 | #include <boost/spirit/home/classic/core/parser.hpp> |
| 14 | #include <boost/spirit/home/classic/core/primitives/numerics.hpp> |
| 15 | #include <boost/spirit/home/classic/core/composite/difference.hpp> |
| 16 | #include <boost/spirit/home/classic/core/composite/sequence.hpp> |
| 17 | |
| 18 | /////////////////////////////////////////////////////////////////////////////// |
| 19 | namespace boost { namespace spirit { |
| 20 | |
| 21 | BOOST_SPIRIT_CLASSIC_NAMESPACE_BEGIN |
| 22 | |
| 23 | /////////////////////////////////////////////////////////////////////////////// |
| 24 | // |
| 25 | // escape_char_parser class |
| 26 | // |
| 27 | /////////////////////////////////////////////////////////////////////////////// |
| 28 | |
| 29 | const unsigned long c_escapes = 1; |
| 30 | const unsigned long lex_escapes = c_escapes << 1; |
| 31 | |
| 32 | ////////////////////////////////// |
| 33 | namespace impl { |
| 34 | |
| 35 | ////////////////////////////////// |
| 36 | #if defined(BOOST_MSVC) |
| 37 | #pragma warning(push) |
| 38 | #pragma warning(disable:4127) |
| 39 | #endif |
| 40 | template <unsigned long Flags, typename CharT> |
| 41 | struct escape_char_action_parse { |
| 42 | |
| 43 | template <typename ParserT, typename ScannerT> |
| 44 | static typename parser_result<ParserT, ScannerT>::type |
| 45 | parse(ScannerT const& scan, ParserT const &p) |
| 46 | { |
| 47 | // Actually decode the escape char. |
| 48 | typedef CharT char_t; |
| 49 | typedef typename ScannerT::iterator_t iterator_t; |
| 50 | typedef typename parser_result<ParserT, ScannerT>::type result_t; |
| 51 | |
| 52 | if (scan.first != scan.last) { |
| 53 | |
| 54 | iterator_t save = scan.first; |
| 55 | if (result_t hit = p.subject().parse(scan)) { |
| 56 | |
| 57 | char_t unescaped; |
| 58 | |
| 59 | scan.first = save; |
| 60 | if (*scan.first == '\\') { |
| 61 | |
| 62 | ++scan.first; |
| 63 | switch (*scan.first) { |
| 64 | case 'b': unescaped = '\b'; ++scan.first; break; |
| 65 | case 't': unescaped = '\t'; ++scan.first; break; |
| 66 | case 'n': unescaped = '\n'; ++scan.first; break; |
| 67 | case 'f': unescaped = '\f'; ++scan.first; break; |
| 68 | case 'r': unescaped = '\r'; ++scan.first; break; |
| 69 | case '"': unescaped = '"'; ++scan.first; break; |
| 70 | case '\'': unescaped = '\''; ++scan.first; break; |
| 71 | case '\\': unescaped = '\\'; ++scan.first; break; |
| 72 | |
| 73 | case 'x': case 'X': |
| 74 | { |
| 75 | char_t hex = 0; |
| 76 | char_t const lim = |
| 77 | (std::numeric_limits<char_t>::max)() >> 4; |
| 78 | |
| 79 | ++scan.first; |
| 80 | while (scan.first != scan.last) |
| 81 | { |
| 82 | char_t c = *scan.first; |
| 83 | if (hex > lim && impl::isxdigit_(c)) |
| 84 | { |
| 85 | // overflow detected |
| 86 | scan.first = save; |
| 87 | return scan.no_match(); |
| 88 | } |
| 89 | if (impl::isdigit_(c)) |
| 90 | { |
| 91 | hex <<= 4; |
| 92 | hex |= c - '0'; |
| 93 | ++scan.first; |
| 94 | } |
| 95 | else if (impl::isxdigit_(c)) |
| 96 | { |
| 97 | hex <<= 4; |
| 98 | c = impl::toupper_(c); |
| 99 | hex |= c - 'A' + 0xA; |
| 100 | ++scan.first; |
| 101 | } |
| 102 | else |
| 103 | { |
| 104 | break; // reached the end of the number |
| 105 | } |
| 106 | } |
| 107 | unescaped = hex; |
| 108 | } |
| 109 | break; |
| 110 | |
| 111 | case '0': case '1': case '2': case '3': |
| 112 | case '4': case '5': case '6': case '7': |
| 113 | { |
| 114 | char_t oct = 0; |
| 115 | char_t const lim = |
| 116 | (std::numeric_limits<char_t>::max)() >> 3; |
| 117 | while (scan.first != scan.last) |
| 118 | { |
| 119 | char_t c = *scan.first; |
| 120 | if (oct > lim && (c >= '0' && c <= '7')) |
| 121 | { |
| 122 | // overflow detected |
| 123 | scan.first = save; |
| 124 | return scan.no_match(); |
| 125 | } |
| 126 | |
| 127 | if (c >= '0' && c <= '7') |
| 128 | { |
| 129 | oct <<= 3; |
| 130 | oct |= c - '0'; |
| 131 | ++scan.first; |
| 132 | } |
| 133 | else |
| 134 | { |
| 135 | break; // reached end of digits |
| 136 | } |
| 137 | } |
| 138 | unescaped = oct; |
| 139 | } |
| 140 | break; |
| 141 | |
| 142 | default: |
| 143 | if (Flags & c_escapes) |
| 144 | { |
| 145 | // illegal C escape sequence |
| 146 | scan.first = save; |
| 147 | return scan.no_match(); |
| 148 | } |
| 149 | else |
| 150 | { |
| 151 | unescaped = *scan.first; |
| 152 | ++scan.first; |
| 153 | } |
| 154 | break; |
| 155 | } |
| 156 | } |
| 157 | else { |
| 158 | unescaped = *scan.first; |
| 159 | ++scan.first; |
| 160 | } |
| 161 | |
| 162 | scan.do_action(p.predicate(), unescaped, save, scan.first); |
| 163 | return hit; |
| 164 | } |
| 165 | } |
| 166 | return scan.no_match(); // overflow detected |
| 167 | } |
| 168 | }; |
| 169 | #if defined(BOOST_MSVC) |
| 170 | #pragma warning(pop) |
| 171 | #endif |
| 172 | |
| 173 | ////////////////////////////////// |
| 174 | template <typename CharT> |
| 175 | struct escape_char_parse { |
| 176 | |
| 177 | template <typename ScannerT, typename ParserT> |
| 178 | static typename parser_result<ParserT, ScannerT>::type |
| 179 | parse(ScannerT const &scan, ParserT const &/*p*/) |
| 180 | { |
| 181 | typedef |
| 182 | uint_parser<CharT, 8, 1, |
| 183 | std::numeric_limits<CharT>::digits / 3 + 1 |
| 184 | > |
| 185 | oct_parser_t; |
| 186 | typedef |
| 187 | uint_parser<CharT, 16, 1, |
| 188 | std::numeric_limits<CharT>::digits / 4 + 1 |
| 189 | > |
| 190 | hex_parser_t; |
| 191 | |
| 192 | typedef alternative<difference<anychar_parser, chlit<CharT> >, |
| 193 | sequence<chlit<CharT>, alternative<alternative<oct_parser_t, |
| 194 | sequence<inhibit_case<chlit<CharT> >, hex_parser_t > >, |
| 195 | difference<difference<anychar_parser, |
| 196 | inhibit_case<chlit<CharT> > >, oct_parser_t > > > > |
| 197 | parser_t; |
| 198 | |
| 199 | static parser_t p = |
| 200 | ( (anychar_p - chlit<CharT>(CharT('\\'))) |
| 201 | | (chlit<CharT>(CharT('\\')) >> |
| 202 | ( oct_parser_t() |
| 203 | | as_lower_d[chlit<CharT>(CharT('x'))] >> hex_parser_t() |
| 204 | | (anychar_p - as_lower_d[chlit<CharT>(CharT('x'))] - oct_parser_t()) |
| 205 | ) |
| 206 | )); |
| 207 | |
| 208 | BOOST_SPIRIT_DEBUG_TRACE_NODE(p, |
| 209 | (BOOST_SPIRIT_DEBUG_FLAGS & BOOST_SPIRIT_DEBUG_FLAGS_ESCAPE_CHAR) != 0); |
| 210 | |
| 211 | return p.parse(scan); |
| 212 | } |
| 213 | }; |
| 214 | |
| 215 | /////////////////////////////////////////////////////////////////////////////// |
| 216 | } // namespace impl |
| 217 | |
| 218 | /////////////////////////////////////////////////////////////////////////////// |
| 219 | BOOST_SPIRIT_CLASSIC_NAMESPACE_END |
| 220 | |
| 221 | }} // namespace boost::spirit |
| 222 | |
| 223 | #endif |
| 224 | |
| 225 | |