The Android Open Source Project | 7790ef5 | 2009-03-03 19:30:40 -0800 | [diff] [blame] | 1 | /* |
| 2 | * Copyright (C) 2007 The Android Open Source Project |
| 3 | * |
| 4 | * Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | * you may not use this file except in compliance with the License. |
| 6 | * You may obtain a copy of the License at |
| 7 | * |
| 8 | * http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | * |
| 10 | * Unless required by applicable law or agreed to in writing, software |
| 11 | * distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | * See the License for the specific language governing permissions and |
| 14 | * limitations under the License. |
| 15 | */ |
| 16 | |
| 17 | #define LOG_TAG "sqlite3_android" |
| 18 | |
| 19 | #include <ctype.h> |
| 20 | #include <stdlib.h> |
| 21 | #include <string.h> |
| 22 | #include <unistd.h> |
| 23 | |
| 24 | #include <unicode/ucol.h> |
| 25 | #include <unicode/ustring.h> |
| 26 | #include <cutils/log.h> |
| 27 | |
| 28 | #include "sqlite3_android.h" |
| 29 | #include "PhoneNumberUtils.h" |
The Android Open Source Project | 455ed29 | 2009-03-13 13:04:22 -0700 | [diff] [blame] | 30 | #include "PhoneticStringUtils.h" |
The Android Open Source Project | 7790ef5 | 2009-03-03 19:30:40 -0800 | [diff] [blame] | 31 | |
| 32 | #define ENABLE_ANDROID_LOG 0 |
| 33 | |
| 34 | static int collate16(void *p, int n1, const void *v1, int n2, const void *v2) |
| 35 | { |
| 36 | UCollator *coll = (UCollator *) p; |
| 37 | UCollationResult result = ucol_strcoll(coll, (const UChar *) v1, n1, |
| 38 | (const UChar *) v2, n2); |
| 39 | |
| 40 | if (result == UCOL_LESS) { |
| 41 | return -1; |
| 42 | } else if (result == UCOL_GREATER) { |
| 43 | return 1; |
| 44 | } else { |
| 45 | return 0; |
| 46 | } |
| 47 | } |
| 48 | |
| 49 | static int collate8(void *p, int n1, const void *v1, int n2, const void *v2) |
| 50 | { |
| 51 | UCollator *coll = (UCollator *) p; |
| 52 | UCharIterator i1, i2; |
| 53 | UErrorCode status = U_ZERO_ERROR; |
| 54 | |
| 55 | uiter_setUTF8(&i1, (const char *) v1, n1); |
| 56 | uiter_setUTF8(&i2, (const char *) v2, n2); |
| 57 | |
| 58 | UCollationResult result = ucol_strcollIter(coll, &i1, &i2, &status); |
| 59 | |
| 60 | if (U_FAILURE(status)) { |
| 61 | // LOGE("Collation iterator error: %d\n", status); |
| 62 | } |
| 63 | |
| 64 | if (result == UCOL_LESS) { |
| 65 | return -1; |
| 66 | } else if (result == UCOL_GREATER) { |
| 67 | return 1; |
| 68 | } else { |
| 69 | return 0; |
| 70 | } |
| 71 | } |
| 72 | |
The Android Open Source Project | 455ed29 | 2009-03-13 13:04:22 -0700 | [diff] [blame] | 73 | static void get_phonetically_sortable_string( |
| 74 | sqlite3_context * context, int argc, sqlite3_value ** argv) |
| 75 | { |
| 76 | if (argc != 1) { |
| 77 | sqlite3_result_null(context); |
| 78 | return; |
| 79 | } |
| 80 | char const * src = (char const *)sqlite3_value_text(argv[0]); |
| 81 | char * ret; |
| 82 | size_t len; |
| 83 | |
| 84 | if (!android::GetPhoneticallySortableString(src, &ret, &len)) { |
Daisuke Miyakawa | 0c45e82 | 2009-03-27 19:41:52 -0700 | [diff] [blame] | 85 | // Put this text at the end of a list. |
| 86 | sqlite3_result_text(context, "\xF0\x9F\xBF\xBD", -1, SQLITE_STATIC); |
| 87 | // sqlite3_result_null(context); |
The Android Open Source Project | 455ed29 | 2009-03-13 13:04:22 -0700 | [diff] [blame] | 88 | } else { |
| 89 | sqlite3_result_text(context, ret, len, free); |
| 90 | } |
| 91 | } |
| 92 | |
The Android Open Source Project | 7790ef5 | 2009-03-03 19:30:40 -0800 | [diff] [blame] | 93 | static void phone_numbers_equal(sqlite3_context * context, int argc, sqlite3_value ** argv) |
| 94 | { |
| 95 | if (argc != 2) { |
| 96 | sqlite3_result_int(context, 0); |
| 97 | return; |
| 98 | } |
| 99 | |
| 100 | char const * num1 = (char const *)sqlite3_value_text(argv[0]); |
| 101 | char const * num2 = (char const *)sqlite3_value_text(argv[1]); |
| 102 | |
| 103 | if (num1 == NULL || num2 == NULL) { |
| 104 | sqlite3_result_null(context); |
| 105 | return; |
| 106 | } |
| 107 | |
| 108 | bool equal = android::phone_number_compare(num1, num2); |
| 109 | |
| 110 | if (equal) { |
| 111 | sqlite3_result_int(context, 1); |
| 112 | } else { |
| 113 | sqlite3_result_int(context, 0); |
| 114 | } |
| 115 | } |
| 116 | |
| 117 | #if ENABLE_ANDROID_LOG |
| 118 | static void android_log(sqlite3_context * context, int argc, sqlite3_value ** argv) |
| 119 | { |
| 120 | char const * tag = "sqlite_trigger"; |
| 121 | char const * msg = ""; |
| 122 | int msgIndex = 0; |
| 123 | |
| 124 | switch (argc) { |
| 125 | case 2: |
| 126 | tag = (char const *)sqlite3_value_text(argv[0]); |
| 127 | if (tag == NULL) { |
| 128 | tag = "sqlite_trigger"; |
| 129 | } |
| 130 | msgIndex = 1; |
| 131 | case 1: |
| 132 | msg = (char const *)sqlite3_value_text(argv[msgIndex]); |
| 133 | if (msg == NULL) { |
| 134 | msg = ""; |
| 135 | } |
| 136 | LOG(LOG_INFO, tag, msg); |
| 137 | sqlite3_result_int(context, 1); |
| 138 | return; |
| 139 | |
| 140 | default: |
| 141 | sqlite3_result_int(context, 0); |
| 142 | return; |
| 143 | } |
| 144 | } |
| 145 | #endif |
| 146 | |
| 147 | static void delete_file(sqlite3_context * context, int argc, sqlite3_value ** argv) |
| 148 | { |
| 149 | if (argc != 1) { |
| 150 | sqlite3_result_int(context, 0); |
| 151 | return; |
| 152 | } |
| 153 | |
| 154 | char const * path = (char const *)sqlite3_value_text(argv[0]); |
| 155 | if (path == NULL) { |
| 156 | sqlite3_result_null(context); |
| 157 | return; |
| 158 | } |
| 159 | |
| 160 | if (strncmp("/sdcard/", path, 8) != 0) { |
| 161 | sqlite3_result_null(context); |
| 162 | return; |
| 163 | } |
Marco Nelissen | 2da78c0 | 2009-05-06 11:08:08 -0700 | [diff] [blame^] | 164 | if (strstr(path, "/../") != NULL) { |
| 165 | sqlite3_result_null(context); |
| 166 | return; |
| 167 | } |
| 168 | |
The Android Open Source Project | 7790ef5 | 2009-03-03 19:30:40 -0800 | [diff] [blame] | 169 | int err = unlink(path); |
| 170 | if (err != -1) { |
| 171 | // No error occured, return true |
| 172 | sqlite3_result_int(context, 1); |
| 173 | } else { |
| 174 | // An error occured, return false |
| 175 | sqlite3_result_int(context, 0); |
| 176 | } |
| 177 | } |
| 178 | |
| 179 | static void tokenize_auxdata_delete(void * data) |
| 180 | { |
| 181 | sqlite3_stmt * statement = (sqlite3_stmt *)data; |
| 182 | sqlite3_finalize(statement); |
| 183 | } |
| 184 | |
| 185 | static void base16Encode(char* dest, const char* src, uint32_t size) |
| 186 | { |
| 187 | static const char * BASE16_TABLE = "0123456789abcdef"; |
| 188 | for (uint32_t i = 0; i < size; i++) { |
| 189 | char ch = *src++; |
| 190 | *dest++ = BASE16_TABLE[ (ch & 0xf0) >> 4 ]; |
| 191 | *dest++ = BASE16_TABLE[ (ch & 0x0f) ]; |
| 192 | } |
| 193 | } |
| 194 | |
| 195 | struct SqliteUserData { |
| 196 | sqlite3 * handle; |
| 197 | UCollator* collator; |
| 198 | }; |
| 199 | |
| 200 | /** |
| 201 | * This function is invoked as: |
| 202 | * |
| 203 | * _TOKENIZE('<token_table>', <data_row_id>, <data>, <delimiter>) |
| 204 | * |
| 205 | * It will then split data on each instance of delimiter and insert each token |
| 206 | * into token_table's 'token' column with data_row_id in the 'source' column. |
| 207 | * The function returns the number of tokens generated. |
| 208 | */ |
| 209 | static void tokenize(sqlite3_context * context, int argc, sqlite3_value ** argv) |
| 210 | { |
| 211 | //LOGD("enter tokenize"); |
| 212 | int err; |
| 213 | |
| 214 | if (argc != 4) { |
| 215 | LOGE("Tokenize requires 4 arguments"); |
| 216 | sqlite3_result_null(context); |
| 217 | return; |
| 218 | } |
| 219 | |
| 220 | sqlite3 * handle = sqlite3_context_db_handle(context); |
| 221 | UCollator* collator = (UCollator*)sqlite3_user_data(context); |
| 222 | char const * tokenTable = (char const *)sqlite3_value_text(argv[0]); |
| 223 | if (tokenTable == NULL) { |
| 224 | LOGE("tokenTable null"); |
| 225 | sqlite3_result_null(context); |
| 226 | return; |
| 227 | } |
| 228 | |
| 229 | // Get or create the prepared statement for the insertions |
| 230 | sqlite3_stmt * statement = (sqlite3_stmt *)sqlite3_get_auxdata(context, 0); |
| 231 | if (!statement) { |
| 232 | char * sql = sqlite3_mprintf("INSERT INTO %s (token, source) VALUES (?, ?);", tokenTable); |
| 233 | err = sqlite3_prepare_v2(handle, sql, -1, &statement, NULL); |
| 234 | sqlite3_free(sql); |
| 235 | if (err) { |
| 236 | LOGE("prepare failed"); |
| 237 | sqlite3_result_null(context); |
| 238 | return; |
| 239 | } |
| 240 | // This binds the statement to the table it was compiled against, which is argv[0]. |
| 241 | // If this function is ever called with a different table the finalizer will be called |
| 242 | // and sqlite3_get_auxdata() will return null above, forcing a recompile for the new table. |
| 243 | sqlite3_set_auxdata(context, 0, statement, tokenize_auxdata_delete); |
| 244 | } else { |
| 245 | // Reset the cached statement so that binding the row ID will work properly |
| 246 | sqlite3_reset(statement); |
| 247 | } |
| 248 | |
| 249 | // Bind the row ID of the source row |
| 250 | int64_t rowID = sqlite3_value_int64(argv[1]); |
| 251 | err = sqlite3_bind_int64(statement, 2, rowID); |
| 252 | if (err != SQLITE_OK) { |
| 253 | LOGE("bind failed"); |
| 254 | sqlite3_result_null(context); |
| 255 | return; |
| 256 | } |
| 257 | |
| 258 | // Get the raw bytes for the string to tokenize |
| 259 | // the string will be modified by following code |
| 260 | // however, sqlite did not reuse the string, so it is safe to not dup it |
| 261 | UChar * origData = (UChar *)sqlite3_value_text16(argv[2]); |
| 262 | if (origData == NULL) { |
| 263 | sqlite3_result_null(context); |
| 264 | return; |
| 265 | } |
| 266 | |
| 267 | // Get the raw bytes for the delimiter |
| 268 | const UChar * delim = (const UChar *)sqlite3_value_text16(argv[3]); |
| 269 | if (delim == NULL) { |
| 270 | LOGE("can't get delimiter"); |
| 271 | sqlite3_result_null(context); |
| 272 | return; |
| 273 | } |
| 274 | |
| 275 | UChar * token = NULL; |
| 276 | UChar *state; |
| 277 | int numTokens = 0; |
| 278 | |
| 279 | do { |
| 280 | if (numTokens == 0) { |
| 281 | token = origData; |
| 282 | } |
| 283 | |
| 284 | // Reset the program so we can use it to perform the insert |
| 285 | sqlite3_reset(statement); |
| 286 | UErrorCode status = U_ZERO_ERROR; |
| 287 | char keybuf[1024]; |
| 288 | uint32_t result = ucol_getSortKey(collator, token, -1, (uint8_t*)keybuf, sizeof(keybuf)-1); |
| 289 | if (result > sizeof(keybuf)) { |
| 290 | // TODO allocate memory for this super big string |
| 291 | LOGE("ucol_getSortKey needs bigger buffer %d", result); |
| 292 | break; |
| 293 | } |
| 294 | uint32_t keysize = result-1; |
| 295 | uint32_t base16Size = keysize*2; |
| 296 | char *base16buf = (char*)malloc(base16Size); |
| 297 | base16Encode(base16buf, keybuf, keysize); |
| 298 | err = sqlite3_bind_text(statement, 1, base16buf, base16Size, SQLITE_STATIC); |
| 299 | |
| 300 | if (err != SQLITE_OK) { |
| 301 | LOGE(" sqlite3_bind_text16 error %d", err); |
| 302 | free(base16buf); |
| 303 | break; |
| 304 | } |
| 305 | |
| 306 | err = sqlite3_step(statement); |
| 307 | free(base16buf); |
| 308 | |
| 309 | if (err != SQLITE_DONE) { |
| 310 | LOGE(" sqlite3_step error %d", err); |
| 311 | break; |
| 312 | } |
| 313 | numTokens++; |
| 314 | if (numTokens == 1) { |
| 315 | // first call |
| 316 | u_strtok_r(origData, delim, &state); |
| 317 | } |
| 318 | } while ((token = u_strtok_r(NULL, delim, &state)) != NULL); |
| 319 | sqlite3_result_int(context, numTokens); |
| 320 | } |
| 321 | |
| 322 | static void localized_collator_dtor(UCollator* collator) |
| 323 | { |
| 324 | ucol_close(collator); |
| 325 | } |
| 326 | |
| 327 | #define LOCALIZED_COLLATOR_NAME "LOCALIZED" |
| 328 | |
| 329 | extern "C" int register_localized_collators(sqlite3* handle, const char* systemLocale, int utf16Storage) |
| 330 | { |
| 331 | int err; |
| 332 | UErrorCode status = U_ZERO_ERROR; |
| 333 | void* icudata; |
| 334 | |
| 335 | UCollator* collator = ucol_open(systemLocale, &status); |
| 336 | if (U_FAILURE(status)) { |
| 337 | return -1; |
| 338 | } |
| 339 | |
| 340 | ucol_setAttribute(collator, UCOL_STRENGTH, UCOL_PRIMARY, &status); |
| 341 | if (U_FAILURE(status)) { |
| 342 | return -1; |
| 343 | } |
| 344 | |
| 345 | status = U_ZERO_ERROR; |
| 346 | char buf[1024]; |
| 347 | int n = ucol_getShortDefinitionString(collator, NULL, buf, 1024, &status); |
| 348 | |
| 349 | if (utf16Storage) { |
| 350 | err = sqlite3_create_collation_v2(handle, LOCALIZED_COLLATOR_NAME, SQLITE_UTF16, collator, |
| 351 | collate16, (void(*)(void*))localized_collator_dtor); |
| 352 | } else { |
| 353 | err = sqlite3_create_collation_v2(handle, LOCALIZED_COLLATOR_NAME, SQLITE_UTF8, collator, |
| 354 | collate8, (void(*)(void*))localized_collator_dtor); |
| 355 | } |
| 356 | if (err != SQLITE_OK) { |
| 357 | return err; |
| 358 | } |
| 359 | |
| 360 | // Register the _TOKENIZE function |
| 361 | err = sqlite3_create_function(handle, "_TOKENIZE", 4, SQLITE_UTF16, collator, tokenize, NULL, NULL); |
| 362 | if (err != SQLITE_OK) { |
| 363 | return err; |
| 364 | } |
| 365 | |
| 366 | return SQLITE_OK; |
| 367 | } |
| 368 | |
| 369 | |
| 370 | extern "C" int register_android_functions(sqlite3 * handle, int utf16Storage) |
| 371 | { |
| 372 | int err; |
| 373 | UErrorCode status = U_ZERO_ERROR; |
| 374 | |
| 375 | UCollator * collator = ucol_open(NULL, &status); |
| 376 | if (U_FAILURE(status)) { |
| 377 | return -1; |
| 378 | } |
| 379 | |
| 380 | if (utf16Storage) { |
| 381 | // Note that text should be stored as UTF-16 |
| 382 | err = sqlite3_exec(handle, "PRAGMA encoding = 'UTF-16'", 0, 0, 0); |
| 383 | if (err != SQLITE_OK) { |
| 384 | return err; |
| 385 | } |
| 386 | |
| 387 | // Register the UNICODE collation |
| 388 | err = sqlite3_create_collation_v2(handle, "UNICODE", SQLITE_UTF16, collator, collate16, |
| 389 | (void(*)(void*))localized_collator_dtor); |
| 390 | } else { |
| 391 | err = sqlite3_create_collation_v2(handle, "UNICODE", SQLITE_UTF8, collator, collate8, |
| 392 | (void(*)(void*))localized_collator_dtor); |
| 393 | } |
| 394 | |
| 395 | if (err != SQLITE_OK) { |
| 396 | return err; |
| 397 | } |
| 398 | |
| 399 | // Register the PHONE_NUM_EQUALS function |
| 400 | err = sqlite3_create_function(handle, "PHONE_NUMBERS_EQUAL", 2, SQLITE_UTF8, NULL, phone_numbers_equal, NULL, NULL); |
| 401 | if (err != SQLITE_OK) { |
| 402 | return err; |
| 403 | } |
| 404 | |
| 405 | // Register the _DELETE_FILE function |
| 406 | err = sqlite3_create_function(handle, "_DELETE_FILE", 1, SQLITE_UTF8, NULL, delete_file, NULL, NULL); |
| 407 | if (err != SQLITE_OK) { |
| 408 | return err; |
| 409 | } |
| 410 | |
| 411 | #if ENABLE_ANDROID_LOG |
| 412 | // Register the _LOG function |
| 413 | err = sqlite3_create_function(handle, "_LOG", 1, SQLITE_UTF8, NULL, android_log, NULL, NULL); |
| 414 | if (err != SQLITE_OK) { |
| 415 | return err; |
| 416 | } |
| 417 | #endif |
| 418 | |
The Android Open Source Project | 455ed29 | 2009-03-13 13:04:22 -0700 | [diff] [blame] | 419 | // Register the GET_PHONETICALLY_SORTABLE_STRING function |
| 420 | err = sqlite3_create_function(handle, |
| 421 | "GET_PHONETICALLY_SORTABLE_STRING", |
| 422 | 1, SQLITE_UTF8, NULL, |
| 423 | get_phonetically_sortable_string, |
| 424 | NULL, NULL); |
| 425 | if (err != SQLITE_OK) { |
| 426 | return err; |
| 427 | } |
| 428 | |
The Android Open Source Project | 7790ef5 | 2009-03-03 19:30:40 -0800 | [diff] [blame] | 429 | return SQLITE_OK; |
| 430 | } |