projects / ceph.git / blame
| Commit | Line | Data |
|---|---|---|
| 7c673cae FG |
1 | <html> |
| 2 | ||
| 3 | <head> | |
| 4 | <meta http-equiv="Content-Language" content="en-us"> | |
| 5 | <meta name="GENERATOR" content="Microsoft FrontPage 5.0"> | |
| 6 | <meta name="ProgId" content="FrontPage.Editor.Document"> | |
| 7 | <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> | |
| 8 | <title>Endian Buffer Types</title> | |
| 9 | <link href="styles.css" rel="stylesheet"> | |
| 10 | </style> | |
| 11 | </head> | |
| 12 | ||
| 13 | <body> | |
| 14 | ||
| 15 | ||
| 16 | <table border="0" cellpadding="5" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%"> | |
| 17 | <tr> | |
| 18 | <td> | |
| 19 | <a href="../../../index.html"> | |
| 20 | <img src="../../../boost.png" alt="Boost logo" align="middle" border="0" width="277" height="86"></a></td> | |
| 21 | <td align="middle"> | |
| 22 | <b> | |
| 23 | <font size="6">Endian Buffer Types</font> </b> | |
| 24 | </td> | |
| 25 | </tr> | |
| 26 | </table> | |
| 27 | ||
| 28 | <table border="0" cellpadding="5" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" bgcolor="#D7EEFF" width="100%"> | |
| 29 | <tr> | |
| 30 | <td><b> | |
| 31 | <a href="index.html">Endian Home</a> | |
| 32 | <a href="conversion.html">Conversion Functions</a> | |
| 33 | <a href="arithmetic.html">Arithmetic Types</a> | |
| 34 | <a href="buffers.html">Buffer Types</a> | |
| 35 | <a href="choosing_approach.html">Choosing Approach</a></b></td> | |
| 36 | </tr> | |
| 37 | </table> | |
| 38 | ||
| 39 | <p></p> | |
| 40 | ||
| 41 | <table border="1" cellpadding="5" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" align="right"> | |
| 42 | <tr> | |
| 43 | <td width="100%" bgcolor="#D7EEFF" align="center"> | |
| 44 | <i><b>Contents</b></i></td> | |
| 45 | </tr> | |
| 46 | <tr> | |
| 47 | <td width="100%" bgcolor="#E8F5FF"> | |
| 48 | <a href="#Introduction">Introduction</a><br> | |
| 49 | <a href="#Example">Example</a><br> | |
| 50 | <a href="#Limitations">Limitations</a><br> | |
| 51 | <a href="#Feature-set">Feature set</a><br> | |
| 52 | <a href="#Types">Enums and typedefs</a><br> | |
| 53 | <a href="#Class_template_endian">Class template <code>endian_buffer</code></a><br> | |
| 54 | | |
| 55 | <a href="#Synopsis">Synopsis</a><br> | |
| 56 | <a href="#Members">Members</a><br> | |
| 57 | <a href="#Non-member-functions">Non-Members</a><br> | |
| 58 | <a href="#FAQ">FAQ</a><br> | |
| 59 | <a href="#Design">Design</a><br> | |
| 60 | <a href="#C++0x">C++11</a><br> | |
| 61 | <a href="#Compilation">Compilation</a></td> | |
| 62 | </tr> | |
| 63 | </table> | |
| 64 | <h2><a name="Introduction">Introduction</a></h2> | |
| 65 | <p>The internal byte order of arithmetic types is traditionally called <b><i>endianness</i></b>. See | |
| 66 | the | |
| 67 | <a href="http://en.wikipedia.org/wiki/Endian" name="endianness">Wikipedia</a> for | |
| 68 | a full | |
| 69 | exploration of <b><i>endianness</i></b>, including definitions of <i><b>big | |
| 70 | endian</b></i> and <i><b>little endian</b></i>.</p> | |
| 71 | <p>Header <b><code>boost/endian/buffers.hpp</code></b> | |
| 72 | provides <code>endian_buffer</code>, a portable endian integer binary buffer | |
| 73 | class template with control over | |
| 74 | byte order, value type, size, and alignment independent of the platform's native | |
| 75 | endianness. Typedefs provide easy-to-use names | |
| 76 | for common configurations.</p> | |
| 77 | <p>Use cases primarily involve data portability, either via files or network | |
| 78 | connections, but these byte-holders may | |
| 79 | also be used to reduce memory use, file size, or network activity since they | |
| 80 | ||
| 81 | provide binary numeric sizes not otherwise available.</p> | |
| 82 | <p dir="ltr">Class <code>endian_buffer</code> is aimed at users who wish | |
| 83 | explicit control over when endianness conversions occur. It also serves as the | |
| 84 | base class for the <code><a href="arithmetic.html">endian_arithmetic</a></code> | |
| 85 | class template, which is aimed at users who wish fully automatic endianness | |
| 86 | conversion and direct support for all normal arithmetic operations.</p> | |
| 87 | <h2><a name="Example">Example</a></h2> | |
| 88 | <p>The <b><code>example/endian_example.cpp</code></b> program writes a | |
| 89 | binary file containing four-byte, big-endian and little-endian integers:</p> | |
| 90 | <blockquote> | |
| 91 | <pre>#include <iostream> | |
| 92 | #include <cstdio> | |
| 93 | #include <boost/endian/buffers.hpp> // see <a href="#Synopsis">Synopsis</a> below | |
| 94 | #include <boost/static_assert.hpp> | |
| 95 | ||
| 96 | using namespace boost::endian; | |
| 97 | ||
| 98 | namespace | |
| 99 | { | |
| 100 | // This is an extract from a very widely used GIS file format. | |
| 101 | // Why the designer decided to mix big and little endians in | |
| 102 | // the same file is not known. But this is a real-world format | |
| 103 | // and users wishing to write low level code manipulating these | |
| 104 | // files have to deal with the mixed endianness. | |
| 105 | ||
| 106 | struct header | |
| 107 | { | |
| 108 | big_int32_<code>buf_</code>t file_code; | |
| 109 | big_int32_<code>buf_</code>t file_length; | |
| 110 | little_int32_<code>buf_</code>t version; | |
| 111 | little_int32_<code>buf_</code>t shape_type; | |
| 112 | }; | |
| 113 | ||
| 114 | const char* filename = "test.dat"; | |
| 115 | } | |
| 116 | ||
| 117 | int main(int, char* []) | |
| 118 | { | |
| 119 | header h; | |
| 120 | ||
| 121 | BOOST_STATIC_ASSERT(sizeof(h) == 16U); // reality check | |
| 122 | ||
| 123 | h.file_code = 0x01020304; | |
| 124 | h.file_length = sizeof(header); | |
| 125 | h.version = 1; | |
| 126 | h.shape_type = 0x01020304; | |
| 127 | ||
| 128 | // Low-level I/O such as POSIX read/write or <cstdio> | |
| 129 | // fread/fwrite is sometimes used for binary file operations | |
| 130 | // when ultimate efficiency is important. Such I/O is often | |
| 131 | // performed in some C++ wrapper class, but to drive home the | |
| 132 | // point that endian integers are often used in fairly | |
| 133 | // low-level code that does bulk I/O operations, <cstdio> | |
| 134 | // fopen/fwrite is used for I/O in this example. | |
| 135 | ||
| 136 | std::FILE* fi = std::fopen(filename, "wb"); // MUST BE BINARY | |
| 137 | ||
| 138 | if (!fi) | |
| 139 | { | |
| 140 | std::cout << "could not open " << filename << '\n'; | |
| 141 | return 1; | |
| 142 | } | |
| 143 | ||
| 144 | if (std::fwrite(&h, sizeof(header), 1, fi)!= 1) | |
| 145 | { | |
| 146 | std::cout << "write failure for " << filename << '\n'; | |
| 147 | return 1; | |
| 148 | } | |
| 149 | ||
| 150 | std::fclose(fi); | |
| 151 | ||
| 152 | std::cout << "created file " << filename << '\n'; | |
| 153 | ||
| 154 | return 0; | |
| 155 | } | |
| 156 | </pre> | |
| 157 | </blockquote> | |
| 158 | <p>After compiling and executing <b><code>example/endian_example.cpp</code></b>, | |
| 159 | a hex dump of <code>test.dat</code> shows:</p> | |
| 160 | <blockquote> | |
| 161 | <pre>01020304 00000010 01000000 04030201</pre> | |
| 162 | </blockquote> | |
| 163 | <p>Notice that the first two 32-bit integers are big endian while the second two | |
| 164 | are little endian, even though the machine this was compiled and run on was | |
| 165 | little endian.</p> | |
| 166 | <h2><a name="Limitations">Limitations</a></h2> | |
| 167 | <p>Requires <code><climits></code> <code>CHAR_BIT == 8</code>. If <code>CHAR_BIT</code> | |
| 168 | is some other value, compilation will result in an <code>#error</code>. This | |
| 169 | restriction is in place because the design, implementation, testing, and | |
| 170 | documentation has only considered issues related to 8-bit bytes, and there have | |
| 171 | been no real-world use cases presented for other sizes.</p> | |
| 172 | <p>In C++03, <code>endian_buffer</code> does not meet the requirements for POD types | |
| 173 | because it has constructors, private data members, and a base class. This means | |
| 174 | that common use cases are relying on unspecified behavior in that the C++ | |
| 175 | Standard does not guarantee memory layout for non-POD types. This has not been a | |
| 176 | problem in practice since all known C++ compilers lay out memory as if <code> | |
| 177 | endian</code> were a POD type. In C++11, it is possible to specify the | |
| 178 | default constructor as trivial, and private data members and base classes no longer disqualify a type from being a POD | |
| 179 | type. Thus under C++11, <code>endian_buffer</code> | |
| 180 | will no longer be relying on unspecified behavior.</p> | |
| 181 | <h2><a name="Feature-set">Feature set</a></h2> | |
| 182 | <ul> | |
| 183 | <li>Big endian| little endian | native endian byte ordering.</li> | |
| 184 | <li>Signed | unsigned</li> | |
| 185 | <li>Unaligned | aligned</li> | |
| 186 | <li>1-8 byte (unaligned) | 1, 2, 4, 8 byte (aligned)</li> | |
| 187 | <li>Choice of value type</li> | |
| 188 | </ul> | |
| 189 | <h2>Enums and t<a name="Types">ypedefs</a></h2> | |
| 190 | <p>Two scoped enums are provided:</p> | |
| 191 | <blockquote> | |
| 192 | <pre>enum class order {big, little, native}; | |
| 193 | ||
| 194 | enum class align {no, yes}; </pre> | |
| 195 | </blockquote> | |
| 196 | <p>One class template is provided:</p> | |
| 197 | <blockquote> | |
| 198 | <pre>template <order Order, typename T, std::size_t Nbits, | |
| 199 | align Align = align::no> | |
| 200 | class endian_buffer; | |
| 201 | </pre> | |
| 202 | </blockquote> | |
| 203 | <p>Typedefs, such as <code>big_int32_buf_t</code>, provide convenient naming | |
| 204 | conventions for common use cases:</p> | |
| 205 | <blockquote> | |
| 206 | <table border="1" cellpadding="5" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="49%"> | |
| 207 | <tr> | |
| 208 | <td width="18%" align="center"><b><i>Name</i></b></td> | |
| 209 | <td width="49%" align="center"><b><i>Alignment</i></b></td> | |
| 210 | <td width="10%" align="center"><b><i>Endianness</i></b></td> | |
| 211 | <td width="10%" align="center"><b><i>Sign</i></b></td> | |
| 212 | <td width="15%" align="center"><b><i>Sizes in bits (n)</i></b></td> | |
| 213 | </tr> | |
| 214 | <tr> | |
| 215 | <td width="18%" dir="ltr"><code>big_int</code><b><i>n</i></b><code>_buf_t</code></td> | |
| 216 | <td width="49%" align="center" dir="ltr"><code>no</code></td> | |
| 217 | <td width="10%" align="center" dir="ltr"><code>big</code></td> | |
| 218 | <td width="10%" align="center" dir="ltr">signed</td> | |
| 219 | <td width="15%" dir="ltr">8,16,24,32,40,48,56,64</td> | |
| 220 | </tr> | |
| 221 | <tr> | |
| 222 | <td width="18%" dir="ltr"><code>big_uint</code><i><b>n</b></i><code>_buf_t</code></td> | |
| 223 | <td width="49%" align="center" dir="ltr"><code>no</code></td> | |
| 224 | <td width="10%" align="center" dir="ltr"><code>big</code></td> | |
| 225 | <td width="10%" align="center" dir="ltr">unsigned</td> | |
| 226 | <td width="15%" dir="ltr">8,16,24,32,40,48,56,64</td> | |
| 227 | </tr> | |
| 228 | <tr> | |
| 229 | <td width="18%"><code>little_int</code><i><b>n</b></i><code>_buf_t</code></td> | |
| 230 | <td width="49%" align="center"><code>no</code></td> | |
| 231 | <td width="10%" align="center"><code>little</code></td> | |
| 232 | <td width="10%" align="center">signed</td> | |
| 233 | <td width="15%">8,16,24,32,40,48,56,64</td> | |
| 234 | </tr> | |
| 235 | <tr> | |
| 236 | <td width="18%"><code>little_uint</code><i><b>n</b></i><code>_buf_t</code></td> | |
| 237 | <td width="49%" align="center"><code>no</code></td> | |
| 238 | <td width="10%" align="center"><code>little</code></td> | |
| 239 | <td width="10%" align="center">unsigned</td> | |
| 240 | <td width="15%">8,16,24,32,40,48,56,64</td> | |
| 241 | </tr> | |
| 242 | <tr> | |
| 243 | <td width="18%"><code>native_int</code><i><b>n</b></i><code>_buf_t</code></td> | |
| 244 | <td width="49%" align="center"><code>no</code></td> | |
| 245 | <td width="10%" align="center"><code>native</code></td> | |
| 246 | <td width="10%" align="center">signed</td> | |
| 247 | <td width="15%">8,16,24,32,40,48,56,64</td> | |
| 248 | </tr> | |
| 249 | <tr> | |
| 250 | <td width="18%"><code>native_uint</code><i><b>n</b></i><code>_buf_t</code></td> | |
| 251 | <td width="49%" align="center"><code>no</code></td> | |
| 252 | <td width="10%" align="center"><code>native</code></td> | |
| 253 | <td width="10%" align="center">unsigned</td> | |
| 254 | <td width="15%">8,16,24,32,40,48,56,64</td> | |
| 255 | </tr> | |
| 256 | <tr> | |
| 257 | <td width="18%"><code>big_int</code><i><b>n</b></i><code>_buf_at</code></td> | |
| 258 | <td width="49%" align="center"><code>yes</code></td> | |
| 259 | <td width="10%" align="center"><code>big</code></td> | |
| 260 | <td width="10%" align="center">signed</td> | |
| 261 | <td width="15%">8,16,32,64</td> | |
| 262 | </tr> | |
| 263 | <tr> | |
| 264 | <td width="18%"><code>big_uint</code><i><b>n</b></i><code>_</code><code>buf_at</code></td> | |
| 265 | <td width="49%" align="center"><code>yes</code></td> | |
| 266 | <td width="10%" align="center"><code>big</code></td> | |
| 267 | <td width="10%" align="center">unsigned</td> | |
| 268 | <td width="15%">8,16,32,64</td> | |
| 269 | </tr> | |
| 270 | <tr> | |
| 271 | <td width="18%"><code>little_int</code><i><b>n</b></i><code>_</code><code>buf_at</code></td> | |
| 272 | <td width="49%" align="center"><code>yes</code></td> | |
| 273 | <td width="10%" align="center"><code>little</code></td> | |
| 274 | <td width="10%" align="center">signed</td> | |
| 275 | <td width="15%">8,16,32,64</td> | |
| 276 | </tr> | |
| 277 | <tr> | |
| 278 | <td width="18%"><code>little_uint</code><i><b>n</b></i><code>_</code><code>buf_at</code></td> | |
| 279 | <td width="49%" align="center"><code>yes</code></td> | |
| 280 | <td width="10%" align="center"><code>little</code></td> | |
| 281 | <td width="10%" align="center">unsigned</td> | |
| 282 | <td width="15%">8,16,32,64</td> | |
| 283 | </tr> | |
| 284 | </table> | |
| 285 | </blockquote> | |
| 286 | <p>The unaligned types do not cause compilers to insert padding bytes in classes | |
| 287 | and structs. This is an important characteristic that can be exploited to minimize wasted space in | |
| 288 | memory, files, and network transmissions. </p> | |
| 289 | <p><font color="#FF0000"><b><i><span style="background-color: #FFFFFF">Warning:</span></i></b></font><span style="background-color: #FFFFFF"> | |
| 290 | Code that uses a</span>ligned types is possibly non-portable because alignment | |
| 291 | requirements vary between hardware architectures and because alignment may be | |
| 292 | affected by compiler switches or pragmas. For example, alignment of an 64-bit | |
| 293 | integer may be to a 32-bit boundary on a 32-bit machine and to a 64-bit boundary | |
| 294 | on a 64-bit machine. Furthermore, aligned types | |
| 295 | are only available on architectures with 8, 16, 32, and 64-bit integer types. </p> | |
| 296 | <p><i><b>Recommendation:</b></i> Prefer unaligned buffer types.</p> | |
| 297 | <p><i><b>Recommendation:</b></i> Protect yourself against alignment ills. For | |
| 298 | example:</p> | |
| 299 | <blockquote> | |
| 300 | <pre>static_assert(sizeof(containing_struct) == 12, "sizeof(containing_struct) is wrong"); </pre> | |
| 301 | </blockquote> | |
| 302 | <p><b><i>Note:</i></b> One-byte big and little buffer types | |
| 303 | have identical layout on all platforms, so they never actually reverse endianness. They are provided to enable generic code, and | |
| 304 | to improve code readability and searchability.</p> | |
| 305 | <h2><a name="Class_template_endian">Class template <code>endian</code></a><code>_buffer</code></h2> | |
| 306 | <p>An <code>endian_buffer</code> is a byte-holder for arithmetic types with user-specified <a href="#endianness"> | |
| 307 | endianness</a>, value type, size, and <a href="#alignment">alignment</a>.</p> | |
| 308 | <h3><a name="Synopsis">Synopsis</a></h3> | |
| 309 | <pre>#include <boost/endian/conversion.hpp | |
| 310 | ||
| 311 | namespace boost | |
| 312 | { | |
| 313 | namespace endian | |
| 314 | { | |
| 315 | // C++11 features emulated if not available | |
| 316 | ||
| 317 | enum class <a name="alignment">align</a> {no, yes}; | |
| 318 | ||
| 319 | template <order Order, class T, std::size_t Nbits, | |
| 320 | align Align = align::no> | |
| 321 | class endian_buffer | |
| 322 | { | |
| 323 | public: | |
| 324 | typedef T value_type; | |
| 325 | ||
| 326 | <a href="#endian">endian_buffer</a>() noexcept = default; | |
| 327 | explicit <a href="#explicit-endian">endian_buffer</a>(T v) noexcept; | |
| 328 | ||
| 329 | endian_buffer& <a href="#operator-eq">operator=</a>(T v) noexcept; | |
| 330 | value_type <a href="#value">value</a>() const noexcept; | |
| 331 | const char* <a href="#data">data</a>() const noexcept; | |
| 332 | protected: | |
| 333 | <b><i>implementaton-defined</i></b> endian_value; // for exposition only | |
| 334 | }; | |
| 335 | ||
| 336 | // stream inserter | |
| 337 | template <class charT, class traits, order Order, class T, | |
| 338 | std::size_t n_bits, align Align> | |
| 339 | std::basic_ostream<charT, traits>& | |
| 340 | <a href="#inserter">operator<<</a>(std::basic_ostream<charT, traits>& os, | |
| 341 | const endian_buffer<Order, T, n_bits, Align>& x); | |
| 342 | ||
| 343 | // stream extractor | |
| 344 | template <class charT, class traits, order Order, class T, | |
| 345 | std::size_t n_bits, align A> | |
| 346 | std::basic_istream<charT, traits>& | |
| 347 | <a href="#extractor">operator>></a>(std::basic_istream<charT, traits>& is, | |
| 348 | endian_buffer<Order, T, n_bits, Align>& x); | |
| 349 | ||
| 350 | // typedefs | |
| 351 | ||
| 352 | // unaligned big endian signed integer buffers | |
| 353 | typedef endian_buffer<order::big, int_least8_t, 8> big_int8_buf_t; | |
| 354 | typedef endian_buffer<order::big, int_least16_t, 16> big_int16_buf_t; | |
| 355 | typedef endian_buffer<order::big, int_least32_t, 24> big_int24_buf_t; | |
| 356 | typedef endian_buffer<order::big, int_least32_t, 32> big_int32_buf_t; | |
| 357 | typedef endian_buffer<order::big, int_least64_t, 40> big_int40_buf_t; | |
| 358 | typedef endian_buffer<order::big, int_least64_t, 48> big_int48_buf_t; | |
| 359 | typedef endian_buffer<order::big, int_least64_t, 56> big_int56_buf_t; | |
| 360 | typedef endian_buffer<order::big, int_least64_t, 64> big_int64_buf_t; | |
| 361 | ||
| 362 | // unaligned big endian unsigned integer buffers | |
| 363 | typedef endian_buffer<order::big, uint_least8_t, 8> big_uint8_buf_t; | |
| 364 | typedef endian_buffer<order::big, uint_least16_t, 16> big_uint16_buf_t; | |
| 365 | typedef endian_buffer<order::big, uint_least32_t, 24> big_uint24_buf_t; | |
| 366 | typedef endian_buffer<order::big, uint_least32_t, 32> big_uint32_buf_t; | |
| 367 | typedef endian_buffer<order::big, uint_least64_t, 40> big_uint40_buf_t; | |
| 368 | typedef endian_buffer<order::big, uint_least64_t, 48> big_uint48_buf_t; | |
| 369 | typedef endian_buffer<order::big, uint_least64_t, 56> big_uint56_buf_t; | |
| 370 | typedef endian_buffer<order::big, uint_least64_t, 64> big_uint64_buf_t; | |
| 371 | ||
| 372 | // unaligned little endian signed integer buffers | |
| 373 | typedef endian_buffer<order::little, int_least8_t, 8> little_int8_buf_t; | |
| 374 | typedef endian_buffer<order::little, int_least16_t, 16> little_int16_buf_t; | |
| 375 | typedef endian_buffer<order::little, int_least32_t, 24> little_int24_buf_t; | |
| 376 | typedef endian_buffer<order::little, int_least32_t, 32> little_int32_buf_t; | |
| 377 | typedef endian_buffer<order::little, int_least64_t, 40> little_int40_buf_t; | |
| 378 | typedef endian_buffer<order::little, int_least64_t, 48> little_int48_buf_t; | |
| 379 | typedef endian_buffer<order::little, int_least64_t, 56> little_int56_buf_t; | |
| 380 | typedef endian_buffer<order::little, int_least64_t, 64> little_int64_buf_t; | |
| 381 | ||
| 382 | // unaligned little endian unsigned integer buffers | |
| 383 | typedef endian_buffer<order::little, uint_least8_t, 8> little_uint8_buf_t; | |
| 384 | typedef endian_buffer<order::little, uint_least16_t, 16> little_uint16_buf_t; | |
| 385 | typedef endian_buffer<order::little, uint_least32_t, 24> little_uint24_buf_t; | |
| 386 | typedef endian_buffer<order::little, uint_least32_t, 32> little_uint32_buf_t; | |
| 387 | typedef endian_buffer<order::little, uint_least64_t, 40> little_uint40_buf_t; | |
| 388 | typedef endian_buffer<order::little, uint_least64_t, 48> little_uint48_buf_t; | |
| 389 | typedef endian_buffer<order::little, uint_least64_t, 56> little_uint56_buf_t; | |
| 390 | typedef endian_buffer<order::little, uint_least64_t, 64> little_uint64_buf_t; | |
| 391 | ||
| 392 | // unaligned native endian signed integer types | |
| 393 | typedef <b><i>implementation-defined</i></b>_int8_buf_t native_int8_buf_t; | |
| 394 | typedef <b><i>implementation-defined</i></b>_int16_buf_t native_int16_buf_t; | |
| 395 | typedef <b><i>implementation-defined</i></b>_int24_buf_t native_int24_buf_t; | |
| 396 | typedef <b><i>implementation-defined</i></b>_int32_buf_t native_int32_buf_t; | |
| 397 | typedef <b><i>implementation-defined</i></b>_int40_buf_t native_int40_buf_t; | |
| 398 | typedef <b><i>implementation-defined</i></b>_int48_buf_t native_int48_buf_t; | |
| 399 | typedef <b><i>implementation-defined</i></b>_int56_buf_t native_int56_buf_t; | |
| 400 | typedef <b><i>implementation-defined</i></b>_int64_buf_t native_int64_buf_t; | |
| 401 | ||
| 402 | // unaligned native endian unsigned integer types | |
| 403 | typedef <b><i>implementation-defined</i></b>_uint8_buf_t native_uint8_buf_t; | |
| 404 | typedef <b><i>implementation-defined</i></b>_uint16_buf_t native_uint16_buf_t; | |
| 405 | typedef <b><i>implementation-defined</i></b>_uint24_buf_t native_uint24_buf_t; | |
| 406 | typedef <b><i>implementation-defined</i></b>_uint32_buf_t native_uint32_buf_t; | |
| 407 | typedef <b><i>implementation-defined</i></b>_uint40_buf_t native_uint40_buf_t; | |
| 408 | typedef <b><i>implementation-defined</i></b>_uint48_buf_t native_uint48_buf_t; | |
| 409 | typedef <b><i>implementation-defined</i></b>_uint56_buf_t native_uint56_buf_t; | |
| 410 | typedef <b><i>implementation-defined</i></b>_uint64_buf_t native_uint64_buf_t; | |
| 411 | ||
| 412 | // aligned big endian signed integer buffers | |
| 413 | typedef endian_buffer<order::big, int8_t, 8, align::yes> big_int8_buf_at; | |
| 414 | typedef endian_buffer<order::big, int16_t, 16, align::yes> big_int16_buf_at; | |
| 415 | typedef endian_buffer<order::big, int32_t, 32, align::yes> big_int32_buf_at; | |
| 416 | typedef endian_buffer<order::big, int64_t, 64, align::yes> big_int64_buf_at; | |
| 417 | ||
| 418 | // aligned big endian unsigned integer buffers | |
| 419 | typedef endian_buffer<order::big, uint8_t, 8, align::yes> big_uint8_buf_at; | |
| 420 | typedef endian_buffer<order::big, uint16_t, 16, align::yes> big_uint16_buf_at; | |
| 421 | typedef endian_buffer<order::big, uint32_t, 32, align::yes> big_uint32_buf_at; | |
| 422 | typedef endian_buffer<order::big, uint64_t, 64, align::yes> big_uint64_buf_at; | |
| 423 | ||
| 424 | // aligned little endian signed integer buffers | |
| 425 | typedef endian_buffer<order::little, int8_t, 8, align::yes> little_int8_buf_at; | |
| 426 | typedef endian_buffer<order::little, int16_t, 16, align::yes> little_int16_buf_at; | |
| 427 | typedef endian_buffer<order::little, int32_t, 32, align::yes> little_int32_buf_at; | |
| 428 | typedef endian_buffer<order::little, int64_t, 64, align::yes> little_int64_buf_at; | |
| 429 | ||
| 430 | // aligned little endian unsigned integer buffers | |
| 431 | typedef endian_buffer<order::little, uint8_t, 8, align::yes> little_uint8_buf_at; | |
| 432 | typedef endian_buffer<order::little, uint16_t, 16, align::yes> little_uint16_buf_at; | |
| 433 | typedef endian_buffer<order::little, uint32_t, 32, align::yes> little_uint32_buf_at; | |
| 434 | typedef endian_buffer<order::little, uint64_t, 64, align::yes> little_uint64_buf_at; | |
| 435 | ||
| 436 | // aligned native endian typedefs are not provided because | |
| 437 | // <cstdint> types are superior for this use case | |
| 438 | ||
| 439 | } // namespace endian | |
| 440 | } // namespace boost</pre> | |
| 441 | <p>The <i><b><code>implementation-defined</code></b></i> text in typedefs above is either | |
| 442 | <code>big</code> or <code>little</code> according to the native endianness of the | |
| 443 | platform.</p> | |
| 444 | <p>The expository data member <code>endian_value</code> stores the current value | |
| 445 | of an <code>endian_value</code> object as a sequence of bytes ordered as | |
| 446 | specified by the <code>Order</code> template parameter. The <i><b><code> | |
| 447 | implementation-defined</code></b></i> type of <code>endian_value</code> is a | |
| 448 | type such as <code><span style="font-size: 85%">char[Nbits/CHAR_BIT]</span></code> | |
| 449 | or <code><span style="font-size: 85%">T</span></code> that meets the | |
| 450 | requirements imposed by the <code>Nbits</code> and <code>Align</code> template | |
| 451 | parameters. The <code><span style="font-size: 85%">CHAR_BIT</span></code> | |
| 452 | macro is defined in <code><span style="font-size: 85%"><climits></span></code>. | |
| 453 | The only value of <code><span style="font-size: 85%">CHAR_BIT</span></code> that | |
| 454 | is required to be supported is 8.</p> | |
| 455 | <p>Template parameter <code><span style="font-size: 85%">T</span></code> is | |
| 456 | required to be a standard integer type (C++std, 3.9.1) and <code> | |
| 457 | <span style="font-size: 85%">sizeof(T)*CHAR_BIT</span></code> is required to be | |
| 458 | greater or equal to <span style="font-size: 85%"> <code>Nbits</code>.</span></p> | |
| 459 | <h3><a name="Members">Members</a></h3> | |
| 460 | <pre><code><a name="endian">endian</a>_buffer() noexcept = default;</code></pre> | |
| 461 | <blockquote> | |
| 462 | <p><i>Effects:</i> Constructs an uninitialized object of type <code>endian_buffer<Order, T, | |
| 463 | Nbits, Align></code>.</p> | |
| 464 | </blockquote> | |
| 465 | <pre><code>explicit <a name="explicit-endian">endian</a>_buffer(T v) noexcept;</code></pre> | |
| 466 | <blockquote> | |
| 467 | <p><i>Effects:</i> Constructs an object of type <code>endian_buffer<Order, T, | |
| 468 | Nbits, Align></code>.</p> | |
| 469 | <p><i>Postcondition:</i> <code>value() == v & mask</code>, where <code>mask</code> | |
| 470 | is a constant of type <code>value_type</code> with <code>Nbits</code> low-order | |
| 471 | bits set to one.</p> | |
| 472 | <p><i>Remarks:</i> If <code>Align</code> is <code>align::yes</code> then | |
| 473 | endianness conversion, if required, is performed by <code> | |
| 474 | boost::endian::endian_reverse</code>.</p> | |
| 475 | </blockquote> | |
| 476 | <pre><code>endian_buffer& <a name="operator-eq">operator=</a>(T v) noexcept;</code></pre> | |
| 477 | <blockquote> | |
| 478 | <p><i>Postcondition:</i> <code>value() == v & mask</code>, where <code>mask</code> | |
| 479 | is a constant of type <code>value_type</code> with <code>Nbits</code> | |
| 480 | low-order bits set to one.</p> | |
| 481 | <p><i>Returns:</i> <code>*this</code>.</p> | |
| 482 | <p><i>Remarks:</i> If <code>Align</code> is <code>align::yes</code> then | |
| 483 | endianness conversion, if required, is performed by <code> | |
| 484 | boost::endian::endian_reverse</code>.</p> | |
| 485 | </blockquote> | |
| 486 | <pre>value_type <a name="value">value</a>()<code> const noexcept;</code></pre> | |
| 487 | <blockquote> | |
| 488 | <p><i>Returns:</i> <code>endian_value</code>, converted to <code>value_type</code>, | |
| 489 | if required, and having the endianness of the native platform.</p> | |
| 490 | <p><i>Remarks:</i> If <code>Align</code> is <code>align::yes</code> then | |
| 491 | endianness conversion, if required, is performed by <code> | |
| 492 | boost::endian::endian_reverse</code>.</p> | |
| 493 | </blockquote> | |
| 494 | <pre><code>const char* <a name="data">data</a>() const noexcept;</code></pre> | |
| 495 | <blockquote> | |
| 496 | <p><i>Returns:</i> A pointer to the first byte of <code>endian_value</code>.</p> | |
| 497 | </blockquote> | |
| 498 | <h3><a name="Non-member-functions">Non-member functions</a></h3> | |
| 499 | <pre>template <class charT, class traits, order Order, class T, | |
| 500 | std::size_t n_bits, align Align> | |
| 501 | std::basic_ostream<charT, traits>& <a name="inserter">operator<<</a>(std::basic_ostream<charT, traits>& os, | |
| 502 | const endian_buffer<Order, T, n_bits, Align>& x); | |
| 503 | </pre> | |
| 504 | <blockquote> | |
| 505 | <p><i>Returns:</i> <code>os << x.value()</code>.</p> | |
| 506 | </blockquote> | |
| 507 | <pre>template <class charT, class traits, order Order, class T, | |
| 508 | std::size_t n_bits, align A> | |
| 509 | std::basic_istream<charT, traits>& <a name="extractor">operator>></a>(std::basic_istream<charT, traits>& is, | |
| 510 | endian_buffer<Order, T, n_bits, Align>& x); | |
| 511 | </pre> | |
| 512 | <blockquote> | |
| 513 | <p><i>Effects: </i>As if:</p> | |
| 514 | <blockquote> | |
| 515 | <pre>T i; | |
| 516 | if (is >> i) | |
| 517 | x = i; | |
| 518 | </pre> | |
| 519 | </blockquote> | |
| 520 | <p><i>Returns:</i> <code>is</code>.</p> | |
| 521 | </blockquote> | |
| 522 | <h2><a name="FAQ">FAQ</a></h2> | |
| 523 | ||
| 524 | <p>See the <a href="index.html#FAQ">Endian home page</a> FAQ for a library-wide | |
| 525 | FAQ.</p> | |
| 526 | ||
| 527 | <p><b>Why not just use Boost.Serialization?</b> Serialization involves a | |
| 528 | conversion for every object involved in I/O. Endian integers require no | |
| 529 | conversion or copying. They are already in the desired format for binary I/O. | |
| 530 | Thus they can be read or written in bulk.</p> | |
| 531 | <p><b>Are endian types PODs?</b> Yes for C++11. No for C++03, although several | |
| 532 | <a href="#Compilation">macros</a> are available to force PODness in all cases.</p> | |
| 533 | <p><b>What are the implications of endian integer types not being PODs with C++03 | |
| 534 | compilers?</b> They | |
| 535 | can't be used in unions. Also, compilers aren't required to align or lay | |
| 536 | out storage in portable ways, although this potential problem hasn't prevented | |
| 537 | use of Boost.Endian with | |
| 538 | real compilers.</p> | |
| 539 | <p><b>What good is <i>native </i>endianness?</b> It provides alignment and | |
| 540 | size guarantees not available from the built-in types. It eases generic | |
| 541 | programming.</p> | |
| 542 | <p><b>Why bother with the aligned endian types?</b> Aligned integer operations | |
| 543 | may be faster (as much as 10 to 20 times faster) if the endianness and alignment of | |
| 544 | the type matches the endianness and alignment requirements of the machine. The code, | |
| 545 | however, is | |
| 546 | likely to be somewhat less portable than with the unaligned types.</p> | |
| 547 | <h2><a name="Design">Design</a> considerations for Boost.Endian buffers</h2> | |
| 548 | <ul> | |
| 549 | <li>Must be suitable for I/O - in other words, must be memcpyable.</li> | |
| 550 | <li>Must provide exactly the size and internal byte ordering specified.</li> | |
| 551 | <li>Must work correctly when the internal integer representation has more bits | |
| 552 | that the sum of the bits in the external byte representation. Sign extension | |
| 553 | must work correctly when the internal integer representation type has more | |
| 554 | bits than the sum of the bits in the external bytes. For example, using | |
| 555 | a 64-bit integer internally to represent 40-bit (5 byte) numbers must work for | |
| 556 | both positive and negative values.</li> | |
| 557 | <li>Must work correctly (including using the same defined external | |
| 558 | representation) regardless of whether a compiler treats char as signed or | |
| 559 | unsigned.</li> | |
| 560 | <li>Unaligned types must not cause compilers to insert padding bytes.</li> | |
| 561 | <li>The implementation should supply optimizations with great care. Experience has shown that optimizations of endian | |
| 562 | integers often become pessimizations when changing | |
| 563 | machines or compilers. Pessimizations can also happen when changing compiler switches, | |
| 564 | compiler versions, or CPU models of the same architecture.</li> | |
| 565 | </ul> | |
| 566 | <h2><a name="C++0x">C++11</a></h2> | |
| 567 | <p>The availability of the C++11 | |
| 568 | <a href="http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2346.htm"> | |
| 569 | Defaulted Functions</a> feature is detected automatically, and will be used if | |
| 570 | present to ensure that objects of <code>class endian_buffer</code> are trivial, and | |
| 571 | thus PODs.</p> | |
| 572 | <h2><a name="Compilation">Compilation</a></h2> | |
| 573 | <p>Boost.Endian is implemented entirely within headers, with no need to link to | |
| 574 | any Boost object libraries.</p> | |
| 575 | <p>Several macros allow user control over features:</p> | |
| 576 | <ul> | |
| 577 | <li>BOOST_ENDIAN_NO_CTORS causes <code>class endian_buffer</code> to have no | |
| 578 | constructors. The intended use is for compiling user code that must be | |
| 579 | portable between compilers regardless of C++11 | |
| 580 | <a href="http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2346.htm"> | |
| 581 | Defaulted Functions</a> support. Use of constructors will always fail, <br> | |
| 582 | </li> | |
| 583 | <li>BOOST_ENDIAN_FORCE_PODNESS causes BOOST_ENDIAN_NO_CTORS to be defined if | |
| 584 | the compiler does not support C++11 | |
| 585 | <a href="http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2346.htm"> | |
| 586 | Defaulted Functions</a>. This is ensures that objects of <code>class endian_buffer</code> | |
| 587 | are PODs, and so can be used in C++03 unions. | |
| 588 | In C++11, <code>class endian_buffer</code> objects are PODs, even though they have | |
| 589 | constructors, so can always be used in unions.</li> | |
| 590 | </ul> | |
| 591 | <hr> | |
| 592 | <p>Last revised: | |
| 593 | <!--webbot bot="Timestamp" s-type="EDITED" s-format="%d %B, %Y" startspan -->14 October, 2015<!--webbot bot="Timestamp" endspan i-checksum="38874" --></p> | |
| 594 | <p>© Copyright Beman Dawes, 2006-2009, 2013</p> | |
| 595 | <p>Distributed under the Boost Software License, Version 1.0. See | |
| 596 | <a href="http://www.boost.org/LICENSE_1_0.txt">www.boost.org/ LICENSE_1_0.txt</a></p> | |
| 597 | ||
| 598 | </body> | |
| 599 | ||
| 600 | </html> |