[15429] | 1 | /**********************************************************************
|
---|
| 2 | *
|
---|
| 3 | * gdbmclass.cpp --
|
---|
| 4 | * Copyright (C) 1999-2008 The New Zealand Digital Library Project
|
---|
| 5 | *
|
---|
| 6 | * A component of the Greenstone digital library software
|
---|
| 7 | * from the New Zealand Digital Library Project at the
|
---|
| 8 | * University of Waikato, New Zealand.
|
---|
| 9 | *
|
---|
| 10 | * This program is free software; you can redistribute it and/or modify
|
---|
| 11 | * it under the terms of the GNU General Public License as published by
|
---|
| 12 | * the Free Software Foundation; either version 2 of the License, or
|
---|
| 13 | * (at your option) any later version.
|
---|
| 14 | *
|
---|
| 15 | * This program is distributed in the hope that it will be useful,
|
---|
| 16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 18 | * GNU General Public License for more details.
|
---|
| 19 | *
|
---|
| 20 | * You should have received a copy of the GNU General Public License
|
---|
| 21 | * along with this program; if not, write to the Free Software
|
---|
| 22 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 23 | *
|
---|
| 24 | *********************************************************************/
|
---|
| 25 |
|
---|
| 26 | #include "gdbmclass.h"
|
---|
[15679] | 27 | #include "gsdltools.h"
|
---|
[15429] | 28 | #include "gsdlunicode.h"
|
---|
[16895] | 29 | #include "fileutil.h"
|
---|
[17701] | 30 | #include "stdlib.h"
|
---|
[18880] | 31 | #include <cstring>
|
---|
[15429] | 32 |
|
---|
| 33 |
|
---|
[22067] | 34 | gdbmclass::gdbmclass(const text_t& gsdlhome)
|
---|
| 35 | : dbclass(gsdlhome)
|
---|
| 36 | {
|
---|
| 37 | gdbmfile = NULL;
|
---|
| 38 | }
|
---|
| 39 |
|
---|
[15429] | 40 | gdbmclass::~gdbmclass()
|
---|
| 41 | {
|
---|
| 42 | closedatabase();
|
---|
| 43 | }
|
---|
| 44 |
|
---|
[15644] | 45 |
|
---|
[15429] | 46 | // returns true if opened
|
---|
| 47 | bool gdbmclass::opendatabase (const text_t &filename, int mode, int num_retrys,
|
---|
[26677] | 48 | #ifdef _MSC_VER
|
---|
[15429] | 49 | bool need_filelock
|
---|
| 50 | #else
|
---|
| 51 | bool
|
---|
| 52 | #endif
|
---|
| 53 | )
|
---|
| 54 | {
|
---|
[16895] | 55 |
|
---|
[15429] | 56 | text_t data_location;
|
---|
| 57 | int block_size = 512;
|
---|
| 58 |
|
---|
| 59 | if (gdbmfile != NULL) {
|
---|
| 60 | if (openfile == filename) return true;
|
---|
| 61 | else closedatabase ();
|
---|
| 62 | }
|
---|
| 63 |
|
---|
[15557] | 64 | // Map the DB mode values into GDBM mode values
|
---|
| 65 | int gdbm_mode = GDBM_READER;
|
---|
| 66 | if (mode == DB_WRITER)
|
---|
| 67 | {
|
---|
| 68 | gdbm_mode = GDBM_WRITER;
|
---|
| 69 | }
|
---|
| 70 | else if (mode == DB_WRITER_CREATE)
|
---|
| 71 | {
|
---|
| 72 | gdbm_mode = GDBM_WRCREAT;
|
---|
| 73 | }
|
---|
| 74 |
|
---|
[18051] | 75 | text_t gdbm_filename = filename;
|
---|
[18050] | 76 | if (gdbm_mode == GDBM_READER)
|
---|
| 77 | {
|
---|
[19061] | 78 | // make sure we have the right file extension. Should be all gdb now.
|
---|
| 79 | // But need to handle old style ldb/bdb collection dbs
|
---|
[18656] | 80 |
|
---|
| 81 | // If the specified GDBM file doesn't exist, (should be gdb) try the other extensions (we can now read both)
|
---|
[18051] | 82 | if (!file_exists(gdbm_filename))
|
---|
[18050] | 83 | {
|
---|
[19061] | 84 | // should be all gdb now
|
---|
[18656] | 85 | if (ends_with(gdbm_filename, ".gdb")) {
|
---|
| 86 | gdbm_filename = substr(gdbm_filename.begin(), gdbm_filename.end() - ((text_t) ".gdb").size()) + ".ldb";
|
---|
[19061] | 87 |
|
---|
[18656] | 88 | if (!file_exists(gdbm_filename)) {
|
---|
| 89 | // try bdb as well
|
---|
| 90 | gdbm_filename = substr(gdbm_filename.begin(), gdbm_filename.end() - ((text_t) ".ldb").size()) + ".bdb";
|
---|
| 91 | }
|
---|
[18051] | 92 | }
|
---|
[16895] | 93 | }
|
---|
[18051] | 94 |
|
---|
[21405] | 95 | // The following fall-back is less likely to be needed now that GDBM
|
---|
| 96 | // library reads both little-endian and big-endian files
|
---|
[18051] | 97 | if (!file_exists(gdbm_filename))
|
---|
| 98 | {
|
---|
| 99 | // We're desperate, so try generating the desired GDBM file from a txtgz file
|
---|
| 100 | gdbm_filename = filename;
|
---|
| 101 | generate_from_txtgz(gdbm_filename);
|
---|
| 102 | }
|
---|
[16895] | 103 | }
|
---|
| 104 |
|
---|
[18051] | 105 | char *namebuffer = gdbm_filename.getcstr();
|
---|
[15429] | 106 | do {
|
---|
[26677] | 107 | #ifdef _MSC_VER
|
---|
[15557] | 108 | gdbmfile = gdbm_open (namebuffer, block_size, gdbm_mode, 00664, NULL, (need_filelock) ? 1 : 0);
|
---|
[15429] | 109 | #else
|
---|
[15557] | 110 | gdbmfile = gdbm_open (namebuffer, block_size, gdbm_mode, 00664, NULL);
|
---|
[15429] | 111 | #endif
|
---|
| 112 | --num_retrys;
|
---|
| 113 | } while (num_retrys>0 && gdbmfile==NULL &&
|
---|
| 114 | (gdbm_errno==GDBM_CANT_BE_READER || gdbm_errno==GDBM_CANT_BE_WRITER));
|
---|
| 115 | delete []namebuffer;
|
---|
| 116 |
|
---|
| 117 | if (gdbmfile == NULL && logout != NULL) {
|
---|
| 118 | outconvertclass text_t2ascii;
|
---|
[18051] | 119 | (*logout) << text_t2ascii << "database open failed on: " << gdbm_filename << "\n";
|
---|
[15429] | 120 | }
|
---|
| 121 |
|
---|
[21796] | 122 | openfile = filename;
|
---|
| 123 |
|
---|
[15429] | 124 | return (gdbmfile != NULL);
|
---|
| 125 | }
|
---|
| 126 |
|
---|
| 127 |
|
---|
| 128 | void gdbmclass::closedatabase ()
|
---|
| 129 | {
|
---|
| 130 | if (gdbmfile == NULL) return;
|
---|
| 131 |
|
---|
| 132 | gdbm_close (gdbmfile);
|
---|
| 133 | gdbmfile = NULL;
|
---|
| 134 | openfile.clear();
|
---|
| 135 | }
|
---|
| 136 |
|
---|
| 137 |
|
---|
[15644] | 138 | void gdbmclass::deletekey (const text_t &key)
|
---|
| 139 | {
|
---|
| 140 | if (gdbmfile == NULL) return;
|
---|
| 141 |
|
---|
| 142 | // get a utf-8 encoded c string of the unicode key
|
---|
| 143 | datum key_data;
|
---|
| 144 | key_data.dptr = (to_utf8(key)).getcstr();
|
---|
| 145 | if (key_data.dptr == NULL) return;
|
---|
| 146 | key_data.dsize = strlen (key_data.dptr);
|
---|
| 147 |
|
---|
| 148 | // delete the key
|
---|
| 149 | gdbm_delete (gdbmfile, key_data);
|
---|
| 150 |
|
---|
| 151 | // free up the key memory
|
---|
| 152 | delete []key_data.dptr;
|
---|
| 153 | }
|
---|
| 154 |
|
---|
| 155 |
|
---|
[15679] | 156 | // returns file extension string
|
---|
| 157 | text_t gdbmclass::getfileextension ()
|
---|
| 158 | {
|
---|
[18656] | 159 | // now we always use gdb for gdbm files. later on, if we can't find it, we'll try ldb and bdb for backwards compatibility
|
---|
| 160 | return ".gdb";
|
---|
| 161 | //if (littleEndian()) return ".ldb";
|
---|
| 162 | //return ".bdb";
|
---|
[15679] | 163 | }
|
---|
| 164 |
|
---|
| 165 |
|
---|
[15429] | 166 | // returns true on success
|
---|
[15644] | 167 | bool gdbmclass::getkeydata (const text_t& key, text_t &data)
|
---|
| 168 | {
|
---|
| 169 | datum key_data;
|
---|
| 170 | datum return_data;
|
---|
| 171 |
|
---|
| 172 | if (gdbmfile == NULL) return false;
|
---|
| 173 |
|
---|
| 174 | // get a utf-8 encoded c string of the unicode key
|
---|
| 175 | key_data.dptr = (to_utf8(key)).getcstr();
|
---|
| 176 | if (key_data.dptr == NULL) {
|
---|
[17719] | 177 | if (logout != NULL) (*logout) << "gdbmclass: out of memory" << endl;
|
---|
[15644] | 178 | return false;
|
---|
| 179 | }
|
---|
| 180 | key_data.dsize = strlen (key_data.dptr);
|
---|
| 181 |
|
---|
| 182 | // fetch the result
|
---|
| 183 | return_data = gdbm_fetch (gdbmfile, key_data);
|
---|
| 184 | delete []key_data.dptr;
|
---|
| 185 |
|
---|
| 186 | if (return_data.dptr == NULL) return false;
|
---|
| 187 |
|
---|
| 188 | data.setcarr (return_data.dptr, return_data.dsize);
|
---|
| 189 | free (return_data.dptr);
|
---|
| 190 | data = to_uni(data); // convert to unicode
|
---|
| 191 |
|
---|
| 192 | return true;
|
---|
| 193 | }
|
---|
| 194 |
|
---|
| 195 |
|
---|
| 196 | // returns array of keys
|
---|
| 197 | text_tarray gdbmclass::getkeys ()
|
---|
| 198 | {
|
---|
| 199 | text_tarray keys;
|
---|
| 200 |
|
---|
| 201 | text_t key = getfirstkey();
|
---|
| 202 | while (!key.empty())
|
---|
| 203 | {
|
---|
| 204 | keys.push_back(key);
|
---|
| 205 | key = getnextkey(key);
|
---|
| 206 | }
|
---|
| 207 |
|
---|
| 208 | return keys;
|
---|
| 209 | }
|
---|
| 210 |
|
---|
| 211 |
|
---|
| 212 | // returns true on success
|
---|
[15649] | 213 | bool gdbmclass::setkeydata (const text_t &key, const text_t &data)
|
---|
[15429] | 214 | {
|
---|
| 215 | if (gdbmfile == NULL) return false;
|
---|
| 216 |
|
---|
| 217 | // store the value
|
---|
| 218 | datum key_data;
|
---|
| 219 | datum data_data;
|
---|
| 220 |
|
---|
| 221 | // get a utf-8 encoded c string of the unicode key
|
---|
| 222 | key_data.dptr = (to_utf8(key)).getcstr();
|
---|
| 223 | if (key_data.dptr == NULL) {
|
---|
[17719] | 224 | if (logout != NULL) (*logout) << "gdbmclass: out of memory" << endl;
|
---|
[15429] | 225 | return false;
|
---|
| 226 | }
|
---|
| 227 | key_data.dsize = strlen (key_data.dptr);
|
---|
| 228 |
|
---|
| 229 | data_data.dptr = (to_utf8(data)).getcstr();
|
---|
| 230 | if (data_data.dptr == NULL) {
|
---|
[17719] | 231 | if (logout != NULL) (*logout) << "gdbmclass: out of memory" << endl;
|
---|
[15429] | 232 | delete []key_data.dptr;
|
---|
| 233 | return false;
|
---|
| 234 | }
|
---|
| 235 | data_data.dsize = strlen (data_data.dptr);
|
---|
| 236 |
|
---|
| 237 | int ret = gdbm_store (gdbmfile, key_data, data_data, GDBM_REPLACE);
|
---|
| 238 | delete []key_data.dptr;
|
---|
| 239 | delete []data_data.dptr;
|
---|
| 240 |
|
---|
| 241 | return (ret == 0);
|
---|
| 242 | }
|
---|
| 243 |
|
---|
| 244 |
|
---|
[15644] | 245 | // ----------------------------------------------------------------------------------------
|
---|
| 246 | // GDBM-ONLY FUNCTIONS
|
---|
| 247 | // ----------------------------------------------------------------------------------------
|
---|
[15429] | 248 |
|
---|
| 249 | // getfirstkey and getnextkey are used for traversing the database
|
---|
| 250 | // no insertions or deletions should be carried out while traversing
|
---|
| 251 | // the database. when there are no keys left to visit in the database
|
---|
| 252 | // an empty string is returned.
|
---|
| 253 | text_t gdbmclass::getfirstkey ()
|
---|
| 254 | {
|
---|
| 255 | if (gdbmfile == NULL) return g_EmptyText;
|
---|
| 256 |
|
---|
| 257 | // get the first key
|
---|
| 258 | datum firstkey_data = gdbm_firstkey (gdbmfile);
|
---|
| 259 | if (firstkey_data.dptr == NULL) return g_EmptyText;
|
---|
| 260 |
|
---|
| 261 | // convert it to text_t
|
---|
| 262 | text_t firstkey;
|
---|
| 263 | firstkey.setcarr (firstkey_data.dptr, firstkey_data.dsize);
|
---|
| 264 | free (firstkey_data.dptr);
|
---|
| 265 | return to_uni(firstkey); // convert to unicode
|
---|
| 266 | }
|
---|
| 267 |
|
---|
| 268 |
|
---|
| 269 | text_t gdbmclass::getnextkey (const text_t &key)
|
---|
| 270 | {
|
---|
| 271 | if (gdbmfile == NULL || key.empty()) return g_EmptyText;
|
---|
| 272 |
|
---|
| 273 | // get a utf-8 encoded c string of the unicode key
|
---|
| 274 | datum key_data;
|
---|
| 275 | key_data.dptr = (to_utf8(key)).getcstr();
|
---|
| 276 | if (key_data.dptr == NULL) return g_EmptyText;
|
---|
| 277 | key_data.dsize = strlen (key_data.dptr);
|
---|
| 278 |
|
---|
| 279 | // get the next key
|
---|
| 280 | datum nextkey_data = gdbm_nextkey (gdbmfile, key_data);
|
---|
| 281 | if (nextkey_data.dptr == NULL) {
|
---|
| 282 | delete []key_data.dptr;
|
---|
| 283 | return g_EmptyText;
|
---|
| 284 | }
|
---|
| 285 |
|
---|
| 286 | // convert it to text_t
|
---|
| 287 | text_t nextkey;
|
---|
| 288 | nextkey.setcarr (nextkey_data.dptr, nextkey_data.dsize);
|
---|
| 289 | free (nextkey_data.dptr);
|
---|
| 290 | delete []key_data.dptr;
|
---|
| 291 | return to_uni(nextkey); // convert to unicode
|
---|
| 292 | }
|
---|
[18050] | 293 |
|
---|
| 294 |
|
---|
[21396] | 295 | // The following routine was written before the GDBM library was upgraded
|
---|
| 296 | // to read both little-endian and big-endian files, and so the reason for
|
---|
| 297 | // it being developed is not so important now. It's useful to keep around
|
---|
| 298 | // however because the file format for GDBM is not guaranteed to be the
|
---|
| 299 | // same from one computer architecture to the next. Even if they are both
|
---|
| 300 | // the same endian-ness. We encountered this issue on the Mac at one stage.
|
---|
| 301 | // the size of the header data-structure for GDBM was simply a different
|
---|
| 302 | // size to that produced on Linux. The Mac case resolved to be binary
|
---|
| 303 | // compatabile with Linux, but shows that we can't rely on this always being
|
---|
| 304 | // the case. Using this method provides a contingency plan. The databases
|
---|
| 305 | // can be shipped at .txt.gz (i.e. portable), which are then converted on
|
---|
| 306 | // on the host's machine to a native GDBM database that is meaningly to
|
---|
| 307 | // that computer.
|
---|
| 308 |
|
---|
[18050] | 309 | void gdbmclass::generate_from_txtgz (text_t filename)
|
---|
| 310 | {
|
---|
| 311 | // Looking to read in the database
|
---|
| 312 | // => check to see if .ldb/.bdb file already there
|
---|
| 313 | // if not (first time) then generate using txt2db
|
---|
| 314 |
|
---|
| 315 | // need to generate architecture native GDBM file using txt2db
|
---|
| 316 |
|
---|
| 317 | // replace sought after gdbm filename ext with ".txt.gz"
|
---|
| 318 |
|
---|
| 319 | text_t::const_iterator begin = filename.begin();
|
---|
| 320 | text_t::const_iterator end= filename.end();
|
---|
[19061] | 321 |
|
---|
| 322 | if (begin != end) {
|
---|
| 323 | end = end - 1;
|
---|
| 324 | }
|
---|
[18050] | 325 |
|
---|
| 326 | text_t::const_iterator here = end;
|
---|
| 327 |
|
---|
| 328 | bool found_ext = false;
|
---|
| 329 |
|
---|
| 330 | while (here != begin) {
|
---|
| 331 | if (*here == '.') {
|
---|
| 332 | found_ext = true;
|
---|
| 333 | break;
|
---|
| 334 | }
|
---|
| 335 | here--;
|
---|
| 336 | }
|
---|
| 337 |
|
---|
| 338 | text_t filename_root;
|
---|
| 339 |
|
---|
| 340 | if (found_ext) {
|
---|
| 341 | filename_root = substr(begin,here);
|
---|
| 342 | }
|
---|
| 343 | else {
|
---|
| 344 | filename_root = filename;
|
---|
| 345 | }
|
---|
| 346 |
|
---|
| 347 | text_t txtgz_filename = filename_root + ".txt.gz";
|
---|
| 348 | if (file_exists(txtgz_filename))
|
---|
| 349 | {
|
---|
| 350 | //text_t cmd = "gzip --decompress --to-stdout \"" + txtgz_filename + "\"";
|
---|
| 351 | //cmd += " | txt2db \"" + filename + "\"";
|
---|
| 352 |
|
---|
| 353 | // Test to make sure Perl is on the path
|
---|
| 354 | // On Linux, the output of the test goes to STDOUT so redirect it to STDERR
|
---|
| 355 | text_t cmd_test = "perl -v 1>&2";
|
---|
| 356 | int rv_test = gsdl_system(cmd_test, true, cerr);
|
---|
| 357 | if (rv_test != 0) {
|
---|
| 358 | cerr << "Tried to find Perl. Return exit value of running "
|
---|
| 359 | << cmd_test <<": "<< rv_test << ", (expected this to be 0)" << endl;
|
---|
| 360 | cerr << "Check that Perl is set in your environment variable PATH." << endl;
|
---|
| 361 | cerr << "At present, PATH=" << getenv("PATH") << endl;
|
---|
| 362 | }
|
---|
| 363 |
|
---|
| 364 | text_t cmd = "perl -S txtgz-to-gdbm.pl \"" + txtgz_filename + "\" \"" + filename + "\"";
|
---|
| 365 | int rv = gsdl_system(cmd, true, cerr);
|
---|
| 366 | // For some reason, launching this command with gsdl_system() still returns 1
|
---|
| 367 | // even when it returns 0 when run from the command-line. We can check whether
|
---|
| 368 | // we succeeded by looking at whether the output database file was created.
|
---|
| 369 | if (rv != 0) {
|
---|
| 370 | cerr << "Warning, non-zero return value on running command \""
|
---|
| 371 | << cmd << "\": " << rv << endl;
|
---|
| 372 | if (!file_exists(filename)) {
|
---|
| 373 | cerr << "Tried to run command \""<<cmd<<"\", but it failed" << endl;
|
---|
| 374 | }
|
---|
| 375 | }
|
---|
| 376 | }
|
---|
| 377 | }
|
---|