root / tags / 1.0.8 / xapian-core / backends / flint / flint_utils.h
| Revision 10730, 10.9 kB (checked in by olly, 7 months ago) | |
|---|---|
|
|
| Line | |
|---|---|
| 1 | /* flint_utils.h: Generic functions for flint |
| 2 | * |
| 3 | * Copyright 1999,2000,2001 BrightStation PLC |
| 4 | * Copyright 2002 Ananova Ltd |
| 5 | * Copyright 2002,2003,2004,2006 Olly Betts |
| 6 | * |
| 7 | * This program is free software; you can redistribute it and/or |
| 8 | * modify it under the terms of the GNU General Public License as |
| 9 | * published by the Free Software Foundation; either version 2 of the |
| 10 | * License, or (at your option) any later version. |
| 11 | * |
| 12 | * This program is distributed in the hope that it will be useful, |
| 13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 15 | * GNU General Public License for more details. |
| 16 | * |
| 17 | * You should have received a copy of the GNU General Public License |
| 18 | * along with this program; if not, write to the Free Software |
| 19 | * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 |
| 20 | * USA |
| 21 | */ |
| 22 | |
| 23 | #ifndef OM_HGUARD_FLINT_UTILS_H |
| 24 | #define OM_HGUARD_FLINT_UTILS_H |
| 25 | |
| 26 | #include <xapian/types.h> |
| 27 | |
| 28 | #include <string> |
| 29 | |
| 30 | using namespace std; |
| 31 | |
| 32 | /// Compile time assert a condition. |
| 33 | #define CASSERT(a) {char assert[(a) ? 1 : -1];(void)assert;} |
| 34 | |
| 35 | /// Compile time assert that type T is unsigned. |
| 36 | #define CASSERT_TYPE_UNSIGNED(T) CASSERT(static_cast<T>(-1) > 0) |
| 37 | |
| 38 | typedef unsigned char om_byte; |
| 39 | typedef unsigned int om_uint32; |
| 40 | typedef int om_int32; |
| 41 | |
| 42 | /** FIXME: the pack and unpack int methods store in low-byte-first order |
| 43 | * - it might be easier to implement efficient specialisations with |
| 44 | * high-byte-first order. |
| 45 | */ |
| 46 | |
| 47 | /** Reads an unsigned integer from a string starting at a given position. |
| 48 | * |
| 49 | * @param src A pointer to a pointer to the data to read. The |
| 50 | * character pointer will be updated to point to the |
| 51 | * next character to read, or 0 if the method ran out of |
| 52 | * data. (It is only set to 0 in case of an error). |
| 53 | * @param src_end A pointer to the byte after the end of the data to |
| 54 | * read the integer from. |
| 55 | * @param resultptr A pointer to a place to store the result. If an |
| 56 | * error occurs, the value stored in this location is |
| 57 | * undefined. If this pointer is 0, the result is not |
| 58 | * stored, and the method simply skips over the result. |
| 59 | * |
| 60 | * @result True if an integer was successfully read. False if the read |
| 61 | * failed. Failure may either be due to the data running out (in |
| 62 | * which case *src will equal 0), or due to the value read |
| 63 | * overflowing the size of result (in which case *src will point |
| 64 | * to wherever the value ends, despite the overflow). |
| 65 | */ |
| 66 | template<class T> |
| 67 | bool |
| 68 | unpack_uint(const char ** src, |
| 69 | const char * src_end, |
| 70 | T * resultptr) |
| 71 | { |
| 72 | // Check unsigned |
| 73 | CASSERT_TYPE_UNSIGNED(T); |
| 74 | |
| 75 | // Check byte is what it's meant to be |
| 76 | CASSERT(sizeof(om_byte) == 1); |
| 77 | |
| 78 | unsigned int shift = 0; |
| 79 | T result = 0; |
| 80 | |
| 81 | while (true) { |
| 82 | if ((*src) == src_end) { |
| 83 | *src = 0; |
| 84 | return false; |
| 85 | } |
| 86 | |
| 87 | om_byte part = static_cast<om_byte>(**src); |
| 88 | (*src)++; |
| 89 | |
| 90 | // if new byte might cause overflow, and it does |
| 91 | if (((shift > (sizeof(T) - 1) * 8 + 1) && |
| 92 | ((part & 0x7f) << (shift % 8)) >= 0x100) || |
| 93 | (shift >= sizeof(T) * 8)) { |
| 94 | // Overflowed - move to end of this integer |
| 95 | while (true) { |
| 96 | if ((part & 0x80) == 0) return false; |
| 97 | if ((*src) == src_end) { |
| 98 | *src = 0; |
| 99 | return false; |
| 100 | } |
| 101 | part = static_cast<om_byte>(**src); |
| 102 | (*src)++; |
| 103 | } |
| 104 | } |
| 105 | |
| 106 | result += T(part & 0x7f) << shift; |
| 107 | shift += 7; |
| 108 | |
| 109 | if ((part & 0x80) == 0) { |
| 110 | if (resultptr) *resultptr = result; |
| 111 | return true; |
| 112 | } |
| 113 | } |
| 114 | } |
| 115 | |
| 116 | |
| 117 | /** Generates a packed representation of an integer. |
| 118 | * |
| 119 | * @param value The integer to represent. |
| 120 | * |
| 121 | * @result A string containing the representation of the integer. |
| 122 | */ |
| 123 | template<class T> |
| 124 | string |
| 125 | pack_uint(T value) |
| 126 | { |
| 127 | // Check unsigned |
| 128 | CASSERT_TYPE_UNSIGNED(T); |
| 129 | |
| 130 | if (value == 0) return string("", 1u); |
| 131 | string result; |
| 132 | |
| 133 | while (value != 0) { |
| 134 | om_byte part = static_cast<om_byte>(value & 0x7f); |
| 135 | value = value >> 7; |
| 136 | if (value) part |= 0x80; |
| 137 | result.append(1u, char(part)); |
| 138 | } |
| 139 | |
| 140 | return result; |
| 141 | } |
| 142 | |
| 143 | /** Generates a packed representation of a bool. |
| 144 | * |
| 145 | * This is a specialisation of the template above. |
| 146 | * |
| 147 | * @param value The bool to represent. |
| 148 | * |
| 149 | * @result A string containing the representation of the bool. |
| 150 | */ |
| 151 | template<> |
| 152 | inline string |
| 153 | pack_uint<bool>(bool value) |
| 154 | { |
| 155 | return string(1, static_cast<char>(value)); |
| 156 | } |
| 157 | |
| 158 | /** Reads an unsigned integer from a string starting at a given position. |
| 159 | * This encoding requires that we know the encoded length from out-of-band |
| 160 | * information (so is suitable when only one integer is encoded, or for |
| 161 | * the last integer encoded). |
| 162 | * |
| 163 | * @param src A pointer to a pointer to the data to read. |
| 164 | * @param src_end A pointer to the byte after the end of the data to |
| 165 | * read the integer from. |
| 166 | * @param resultptr A pointer to a place to store the result. If an |
| 167 | * error occurs, the value stored in this location is |
| 168 | * undefined. If this pointer is 0, the result is not |
| 169 | * stored, and the method simply skips over the result. |
| 170 | * |
| 171 | * @result True if an integer was successfully read. False if the read |
| 172 | * failed. Failure can hapen if the value read overflows |
| 173 | * the size of result. |
| 174 | */ |
| 175 | template<class T> |
| 176 | bool |
| 177 | unpack_uint_last(const char ** src, const char * src_end, T * resultptr) |
| 178 | { |
| 179 | // Check unsigned |
| 180 | CASSERT_TYPE_UNSIGNED(T); |
| 181 | // Check byte is what it's meant to be |
| 182 | CASSERT(sizeof(om_byte) == 1); |
| 183 | |
| 184 | if (src_end - *src > int(sizeof(T))) { |
| 185 | // Would overflow |
| 186 | *src = src_end; |
| 187 | return false; |
| 188 | } |
| 189 | |
| 190 | T result = 0; |
| 191 | int shift = 0; |
| 192 | while (*src != src_end) { |
| 193 | result |= static_cast<T>(static_cast<om_byte>(**src)) << shift; |
| 194 | ++(*src); |
| 195 | shift += 8; |
| 196 | } |
| 197 | *resultptr = result; |
| 198 | return true; |
| 199 | } |
| 200 | |
| 201 | /** Generates a packed representation of an integer. |
| 202 | * This encoding requires that we know the encoded length from out-of-band |
| 203 | * information (so is suitable when only one integer is encoded, or for |
| 204 | * the last integer encoded). |
| 205 | * |
| 206 | * @param value The integer to represent. |
| 207 | * |
| 208 | * @result A string containing the representation of the integer. |
| 209 | */ |
| 210 | template<class T> |
| 211 | string |
| 212 | pack_uint_last(T value) |
| 213 | { |
| 214 | // Check unsigned |
| 215 | CASSERT_TYPE_UNSIGNED(T); |
| 216 | |
| 217 | string result; |
| 218 | while (value) { |
| 219 | result += char(value); |
| 220 | value >>= 8; |
| 221 | } |
| 222 | return result; |
| 223 | } |
| 224 | |
| 225 | /** Generate a packed representation of an integer, preserving sort order. |
| 226 | * |
| 227 | * This representation is less compact than the usual one, and has a limit |
| 228 | * of 256 bytes on the length of the integer. However, this is unlikely to |
| 229 | * ever be a problem. |
| 230 | * |
| 231 | * @param value The integer to represent. |
| 232 | * |
| 233 | * @result A string containing the representation of the integer. |
| 234 | */ |
| 235 | template<class T> |
| 236 | string |
| 237 | pack_uint_preserving_sort(T value) |
| 238 | { |
| 239 | // Check unsigned |
| 240 | CASSERT_TYPE_UNSIGNED(T); |
| 241 | |
| 242 | string result; |
| 243 | while (value != 0) { |
| 244 | om_byte part = static_cast<om_byte>(value & 0xff); |
| 245 | value = value >> 8; |
| 246 | result.insert(string::size_type(0), 1u, char(part)); |
| 247 | } |
| 248 | result.insert(string::size_type(0), 1u, char(result.size())); |
| 249 | return result; |
| 250 | } |
| 251 | |
| 252 | /** Unpack a unsigned integer, store in sort preserving order. |
| 253 | * |
| 254 | * @param src A pointer to a pointer to the data to read. The |
| 255 | * character pointer will be updated to point to the |
| 256 | * next character to read, or 0 if the method ran out of |
| 257 | * data. (It is only set to 0 in case of an error). |
| 258 | * @param src_end A pointer to the byte after the end of the data to |
| 259 | * read the integer from. |
| 260 | * @param resultptr A pointer to a place to store the result. If an |
| 261 | * error occurs, the value stored in this location is |
| 262 | * undefined. If this pointer is 0, the result is not |
| 263 | * stored, and the method simply skips over the result. |
| 264 | * |
| 265 | * @result True if an integer was successfully read. False if the read |
| 266 | * failed. Failure may either be due to the data running out (in |
| 267 | * which case *src will equal 0), or due to the value read |
| 268 | * overflowing the size of result (in which case *src will point |
| 269 | * to wherever the value ends, despite the overflow). |
| 270 | */ |
| 271 | template<class T> |
| 272 | bool |
| 273 | unpack_uint_preserving_sort(const char ** src, |
| 274 | const char * src_end, |
| 275 | T * resultptr) |
| 276 | { |
| 277 | if (*src == src_end) { |
| 278 | *src = 0; |
| 279 | return false; |
| 280 | } |
| 281 | |
| 282 | unsigned int length = static_cast<om_byte>(**src); |
| 283 | (*src)++; |
| 284 | |
| 285 | if (length > sizeof(T)) { |
| 286 | *src += length; |
| 287 | if (*src > src_end) { |
| 288 | *src = 0; |
| 289 | } |
| 290 | return false; |
| 291 | } |
| 292 | |
| 293 | // Can't be overflow now. |
| 294 | T result = 0; |
| 295 | while (length > 0) { |
| 296 | result = result << 8; |
| 297 | result += static_cast<om_byte>(**src); |
| 298 | (*src)++; |
| 299 | length--; |
| 300 | } |
| 301 | *resultptr = result; |
| 302 | |
| 303 | return true; |
| 304 | } |
| 305 | |
| 306 | inline bool |
| 307 | unpack_string(const char ** src, |
| 308 | const char * src_end, |
| 309 | string & result) |
| 310 | { |
| 311 | string::size_type length; |
| 312 | if (!unpack_uint(src, src_end, &length)) { |
| 313 | return false; |
| 314 | } |
| 315 | |
| 316 | if (src_end - *src < 0 || |
| 317 | string::size_type(src_end - *src) < length) { |
| 318 | src = 0; |
| 319 | return false; |
| 320 | } |
| 321 | |
| 322 | result = string(*src, length); |
| 323 | *src += length; |
| 324 | return true; |
| 325 | } |
| 326 | |
| 327 | inline string |
| 328 | pack_string(string value) |
| 329 | { |
| 330 | return pack_uint(value.size()) + value; |
| 331 | } |
| 332 | |
| 333 | /** Pack a string into a representation which preserves sort order. |
| 334 | * We do this by replacing zero bytes in the string with a zero byte |
| 335 | * followed by byte value 0xff, and then appending two zero bytes to |
| 336 | * the end. |
| 337 | */ |
| 338 | inline string |
| 339 | pack_string_preserving_sort(string value) |
| 340 | { |
| 341 | string::size_type i = 0, j; |
| 342 | while ((j = value.find('\0', i)) != string::npos) { |
| 343 | value.replace(j, 1, "\0\xff", 2); |
| 344 | i = j + 2; |
| 345 | } |
| 346 | value += '\0'; // FIXME temp... |
| 347 | return value + '\0'; // Note - next byte mustn't be '\xff'... |
| 348 | } |
| 349 | |
| 350 | inline bool |
| 351 | unpack_string_preserving_sort(const char ** src, |
| 352 | const char * src_end, |
| 353 | string & result) |
| 354 | { |
| 355 | result = ""; |
| 356 | while (*src < src_end) { |
| 357 | const char *begin = *src; |
| 358 | while (**src) { |
| 359 | ++(*src); |
| 360 | if (*src == src_end) return false; |
| 361 | } |
| 362 | result += string(begin, *src - begin); |
| 363 | ++(*src); |
| 364 | if (*src == src_end) return false; |
| 365 | if (**src != '\xff') { |
| 366 | ++(*src); // FIXME temp |
| 367 | return true; |
| 368 | } |
| 369 | result += '\0'; |
| 370 | ++(*src); |
| 371 | } |
| 372 | return false; |
| 373 | } |
| 374 | |
| 375 | inline bool |
| 376 | unpack_bool(const char ** src, |
| 377 | const char * src_end, |
| 378 | bool * resultptr) |
| 379 | { |
| 380 | if (*src == src_end) { |
| 381 | *src = 0; |
| 382 | return false; |
| 383 | } |
| 384 | switch (*((*src)++)) { |
| 385 | case '0': |
| 386 | if (resultptr) *resultptr = false; |
| 387 | return true; |
| 388 | case '1': |
| 389 | if (resultptr) *resultptr = true; |
| 390 | return true; |
| 391 | } |
| 392 | *src = 0; |
| 393 | return false; |
| 394 | } |
| 395 | |
| 396 | inline string |
| 397 | pack_bool(bool value) |
| 398 | { |
| 399 | return value ? "1" : "0"; |
| 400 | } |
| 401 | |
| 402 | /** Convert a document id to a key (suitable when the docid is the only |
| 403 | * component of the key). |
| 404 | */ |
| 405 | inline string |
| 406 | flint_docid_to_key(Xapian::docid did) |
| 407 | { |
| 408 | return pack_uint_preserving_sort(did); |
| 409 | } |
| 410 | |
| 411 | #endif /* OM_HGUARD_FLINT_UTILS_H */ |
Note: See TracBrowser
for help on using the browser.
