| 1 | /* |
| 2 | * This file is part of DisOrder |
| 3 | * Copyright (C) 2005-2008 Richard Kettlewell |
| 4 | * |
| 5 | * This program is free software; you can redistribute it and/or modify |
| 6 | * it under the terms of the GNU General Public License as published by |
| 7 | * the Free Software Foundation; either version 2 of the License, or |
| 8 | * (at your option) any later version. |
| 9 | * |
| 10 | * This program is distributed in the hope that it will be useful, but |
| 11 | * WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 13 | * General Public License for more details. |
| 14 | * |
| 15 | * You should have received a copy of the GNU General Public License |
| 16 | * along with this program; if not, write to the Free Software |
| 17 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 |
| 18 | * USA |
| 19 | */ |
| 20 | |
| 21 | #include <config.h> |
| 22 | #include "types.h" |
| 23 | |
| 24 | #include <getopt.h> |
| 25 | #include <stdio.h> |
| 26 | #include <stdlib.h> |
| 27 | #include <db.h> |
| 28 | #include <locale.h> |
| 29 | #include <errno.h> |
| 30 | #include <sys/types.h> |
| 31 | #include <unistd.h> |
| 32 | #include <pcre.h> |
| 33 | #include <fnmatch.h> |
| 34 | #include <sys/wait.h> |
| 35 | #include <string.h> |
| 36 | #include <syslog.h> |
| 37 | #include <time.h> |
| 38 | |
| 39 | #include "configuration.h" |
| 40 | #include "syscalls.h" |
| 41 | #include "log.h" |
| 42 | #include "defs.h" |
| 43 | #include "mem.h" |
| 44 | #include "plugin.h" |
| 45 | #include "inputline.h" |
| 46 | #include "charset.h" |
| 47 | #include "wstat.h" |
| 48 | #include "kvp.h" |
| 49 | #include "printf.h" |
| 50 | #include "rights.h" |
| 51 | #include "trackdb.h" |
| 52 | #include "trackdb-int.h" |
| 53 | #include "trackname.h" |
| 54 | #include "unicode.h" |
| 55 | #include "version.h" |
| 56 | |
| 57 | static DB_TXN *global_tid; |
| 58 | |
| 59 | static const struct option options[] = { |
| 60 | { "help", no_argument, 0, 'h' }, |
| 61 | { "version", no_argument, 0, 'V' }, |
| 62 | { "config", required_argument, 0, 'c' }, |
| 63 | { "debug", no_argument, 0, 'd' }, |
| 64 | { "no-debug", no_argument, 0, 'D' }, |
| 65 | { "syslog", no_argument, 0, 's' }, |
| 66 | { "no-syslog", no_argument, 0, 'S' }, |
| 67 | { "check", no_argument, 0, 'K' }, |
| 68 | { "no-check", no_argument, 0, 'C' }, |
| 69 | { 0, 0, 0, 0 } |
| 70 | }; |
| 71 | |
| 72 | /* display usage message and terminate */ |
| 73 | static void help(void) { |
| 74 | xprintf("Usage:\n" |
| 75 | " disorder-rescan [OPTIONS] [PATH...]\n" |
| 76 | "Options:\n" |
| 77 | " --help, -h Display usage message\n" |
| 78 | " --version, -V Display version number\n" |
| 79 | " --config PATH, -c PATH Set configuration file\n" |
| 80 | " --debug, -d Turn on debugging\n" |
| 81 | " --[no-]syslog Enable/disable logging to syslog\n" |
| 82 | " --[no-]check Enable/disable track length check\n" |
| 83 | "\n" |
| 84 | "Rescanner for DisOrder. Not intended to be run\n" |
| 85 | "directly.\n"); |
| 86 | xfclose(stdout); |
| 87 | exit(0); |
| 88 | } |
| 89 | |
| 90 | static volatile sig_atomic_t signalled; |
| 91 | |
| 92 | static void signal_handler(int sig) { |
| 93 | if(sig == 0) _exit(-1); /* "Cannot happen" */ |
| 94 | signalled = sig; |
| 95 | } |
| 96 | |
| 97 | static int aborted(void) { |
| 98 | return signalled || getppid() == 1; |
| 99 | } |
| 100 | |
| 101 | /* Exit if our parent has gone away or we have been told to stop. */ |
| 102 | static void checkabort(void) { |
| 103 | if(getppid() == 1) { |
| 104 | info("parent has terminated"); |
| 105 | trackdb_abort_transaction(global_tid); |
| 106 | exit(0); |
| 107 | } |
| 108 | if(signalled) { |
| 109 | info("received signal %d", signalled); |
| 110 | trackdb_abort_transaction(global_tid); |
| 111 | exit(0); |
| 112 | } |
| 113 | } |
| 114 | |
| 115 | /* rescan a collection */ |
| 116 | static void rescan_collection(const struct collection *c) { |
| 117 | pid_t pid, r; |
| 118 | int p[2], n, w; |
| 119 | FILE *fp = 0; |
| 120 | char *path, *track; |
| 121 | long ntracks = 0, nnew = 0; |
| 122 | |
| 123 | checkabort(); |
| 124 | info("rescanning %s with %s", c->root, c->module); |
| 125 | /* plugin runs in a subprocess */ |
| 126 | xpipe(p); |
| 127 | if(!(pid = xfork())) { |
| 128 | exitfn = _exit; |
| 129 | xclose(p[0]); |
| 130 | xdup2(p[1], 1); |
| 131 | xclose(p[1]); |
| 132 | scan(c->module, c->root); |
| 133 | if(fflush(stdout) < 0) |
| 134 | fatal(errno, "error writing to scanner pipe"); |
| 135 | _exit(0); |
| 136 | } |
| 137 | xclose(p[1]); |
| 138 | if(!(fp = fdopen(p[0], "r"))) |
| 139 | fatal(errno, "error calling fdopen"); |
| 140 | /* read tracks from the plugin */ |
| 141 | while(!inputline("rescanner", fp, &path, 0)) { |
| 142 | checkabort(); |
| 143 | /* actually we can cope relatively well within the server, but they'll go |
| 144 | * wrong in track listings */ |
| 145 | if(strchr(path, '\n')) { |
| 146 | error(0, "cannot cope with tracks with newlines in the name"); |
| 147 | continue; |
| 148 | } |
| 149 | if(!(track = any2utf8(c->encoding, path))) { |
| 150 | error(0, "cannot convert track path to UTF-8: %s", path); |
| 151 | continue; |
| 152 | } |
| 153 | if(config->dbversion > 1) { |
| 154 | /* We use NFC track names */ |
| 155 | if(!(track = utf8_compose_canon(track, strlen(track), 0))) { |
| 156 | error(0, "cannot convert track path to NFC: %s", path); |
| 157 | continue; |
| 158 | } |
| 159 | } |
| 160 | D(("track %s", track)); |
| 161 | /* only tracks with a known player are admitted */ |
| 162 | for(n = 0; (n < config->player.n |
| 163 | && fnmatch(config->player.s[n].s[0], track, 0) != 0); ++n) |
| 164 | ; |
| 165 | if(n < config->player.n) { |
| 166 | nnew += !!trackdb_notice(track, path); |
| 167 | ++ntracks; |
| 168 | if(ntracks % 1000 == 0) |
| 169 | info("rescanning %s, %ld tracks so far", c->root, ntracks); |
| 170 | } |
| 171 | } |
| 172 | /* tidy up */ |
| 173 | if(ferror(fp)) { |
| 174 | error(errno, "error reading from scanner pipe"); |
| 175 | goto done; |
| 176 | } |
| 177 | xfclose(fp); |
| 178 | fp = 0; |
| 179 | while((r = waitpid(pid, &w, 0)) == -1 && errno == EINTR) |
| 180 | ; |
| 181 | if(r < 0) fatal(errno, "error calling waitpid"); |
| 182 | pid = 0; |
| 183 | if(w) { |
| 184 | error(0, "scanner subprocess: %s", wstat(w)); |
| 185 | goto done; |
| 186 | } |
| 187 | info("rescanned %s, %ld tracks, %ld new", c->root, ntracks, nnew); |
| 188 | done: |
| 189 | if(fp) |
| 190 | xfclose(fp); |
| 191 | if(pid) |
| 192 | while((r = waitpid(pid, &w, 0)) == -1 && errno == EINTR) |
| 193 | ; |
| 194 | } |
| 195 | |
| 196 | struct recheck_state { |
| 197 | const struct collection *c; |
| 198 | long nobsolete, nnocollection, nlength; |
| 199 | struct recheck_track *tracks; |
| 200 | }; |
| 201 | |
| 202 | struct recheck_track { |
| 203 | struct recheck_track *next; |
| 204 | const char *track; |
| 205 | }; |
| 206 | |
| 207 | /* called for each non-alias track */ |
| 208 | static int recheck_list_callback(const char *track, |
| 209 | struct kvp attribute((unused)) *data, |
| 210 | struct kvp attribute((unused)) *prefs, |
| 211 | void *u, |
| 212 | DB_TXN attribute((unused)) *tid) { |
| 213 | struct recheck_state *cs = u; |
| 214 | struct recheck_track *t = xmalloc(sizeof *t); |
| 215 | |
| 216 | t->next = cs->tracks; |
| 217 | t->track = track; |
| 218 | cs->tracks = t; |
| 219 | return 0; |
| 220 | } |
| 221 | |
| 222 | static int recheck_track_tid(struct recheck_state *cs, |
| 223 | const struct recheck_track *t, |
| 224 | DB_TXN *tid) { |
| 225 | const struct collection *c = cs->c; |
| 226 | const char *path; |
| 227 | char buffer[20]; |
| 228 | int err, n; |
| 229 | long length; |
| 230 | struct kvp *data; |
| 231 | |
| 232 | if((err = trackdb_getdata(trackdb_tracksdb, t->track, &data, tid))) |
| 233 | return err; |
| 234 | path = kvp_get(data, "_path"); |
| 235 | D(("rechecking %s", t->track)); |
| 236 | /* if we're not checking a specific collection, find the right collection */ |
| 237 | if(!c) { |
| 238 | if(!(c = find_track_collection(t->track))) { |
| 239 | D(("obsoleting %s", t->track)); |
| 240 | if((err = trackdb_obsolete(t->track, tid))) |
| 241 | return err; |
| 242 | ++cs->nnocollection; |
| 243 | return 0; |
| 244 | } |
| 245 | } |
| 246 | /* see if the track has evaporated */ |
| 247 | if(check(c->module, c->root, path) == 0) { |
| 248 | D(("obsoleting %s", t->track)); |
| 249 | if((err = trackdb_obsolete(t->track, tid))) |
| 250 | return err; |
| 251 | ++cs->nobsolete; |
| 252 | return 0; |
| 253 | } |
| 254 | /* make sure we know the length */ |
| 255 | if(!kvp_get(data, "_length")) { |
| 256 | D(("recalculating length of %s", t->track)); |
| 257 | for(n = 0; n < config->tracklength.n; ++n) |
| 258 | if(fnmatch(config->tracklength.s[n].s[0], t->track, 0) == 0) |
| 259 | break; |
| 260 | if(n >= config->tracklength.n) |
| 261 | error(0, "no tracklength plugin found for %s", t->track); |
| 262 | else { |
| 263 | length = tracklength(config->tracklength.s[n].s[1], t->track, path); |
| 264 | if(length > 0) { |
| 265 | byte_snprintf(buffer, sizeof buffer, "%ld", length); |
| 266 | kvp_set(&data, "_length", buffer); |
| 267 | if((err = trackdb_putdata(trackdb_tracksdb, t->track, data, tid, 0))) |
| 268 | return err; |
| 269 | ++cs->nlength; |
| 270 | } |
| 271 | } |
| 272 | } |
| 273 | return 0; |
| 274 | } |
| 275 | |
| 276 | static int recheck_track(struct recheck_state *cs, |
| 277 | const struct recheck_track *t) { |
| 278 | int e; |
| 279 | |
| 280 | WITH_TRANSACTION(recheck_track_tid(cs, t, tid)); |
| 281 | return e; |
| 282 | } |
| 283 | |
| 284 | /* recheck a collection */ |
| 285 | static void recheck_collection(const struct collection *c) { |
| 286 | struct recheck_state cs; |
| 287 | const struct recheck_track *t; |
| 288 | long nrc; |
| 289 | |
| 290 | if(c) |
| 291 | info("rechecking %s", c->root); |
| 292 | else |
| 293 | info("rechecking all tracks"); |
| 294 | /* Doing the checking inside a transaction locks up the server for much too |
| 295 | * long (because it spends lots of time thinking about each track). So we |
| 296 | * pull the full track list into memory and work from that. |
| 297 | * |
| 298 | * 100,000 tracks at, say, 80 bytes per track name, gives 8MB, which is quite |
| 299 | * reasonable. |
| 300 | */ |
| 301 | for(;;) { |
| 302 | checkabort(); |
| 303 | info("getting track list"); |
| 304 | global_tid = trackdb_begin_transaction(); |
| 305 | memset(&cs, 0, sizeof cs); |
| 306 | cs.c = c; |
| 307 | if(trackdb_scan(c ? c->root : 0, recheck_list_callback, &cs, global_tid)) |
| 308 | goto fail; |
| 309 | break; |
| 310 | fail: |
| 311 | /* Maybe we need to shut down */ |
| 312 | checkabort(); |
| 313 | /* Abort the transaction and try again in a bit. */ |
| 314 | trackdb_abort_transaction(global_tid); |
| 315 | global_tid = 0; |
| 316 | /* Let anything else that is going on get out of the way. */ |
| 317 | sleep(10); |
| 318 | checkabort(); |
| 319 | if(c) |
| 320 | info("resuming recheck of %s", c->root); |
| 321 | else |
| 322 | info("resuming global recheck"); |
| 323 | } |
| 324 | trackdb_commit_transaction(global_tid); |
| 325 | global_tid = 0; |
| 326 | nrc = 0; |
| 327 | for(t = cs.tracks; t; t = t->next) { |
| 328 | if(aborted()) |
| 329 | return; |
| 330 | recheck_track(&cs, t); |
| 331 | ++nrc; |
| 332 | if(nrc % 100 == 0) { |
| 333 | if(c) |
| 334 | info("rechecking %s, %ld tracks so far", c->root, nrc); |
| 335 | else |
| 336 | info("rechecking all tracks, %ld tracks so far", nrc); |
| 337 | } |
| 338 | } |
| 339 | if(c) |
| 340 | info("rechecked %s, %ld obsoleted, %ld lengths calculated", |
| 341 | c->root, cs.nobsolete, cs.nlength); |
| 342 | else |
| 343 | info("rechecked all tracks, %ld no collection, %ld obsoleted, %ld lengths calculated", |
| 344 | cs.nnocollection, cs.nobsolete, cs.nlength); |
| 345 | } |
| 346 | |
| 347 | /* rescan/recheck a collection by name */ |
| 348 | static void do_directory(const char *s, |
| 349 | void (*fn)(const struct collection *c)) { |
| 350 | int n; |
| 351 | |
| 352 | for(n = 0; (n < config->collection.n |
| 353 | && strcmp(config->collection.s[n].root, s)); ++n) |
| 354 | ; |
| 355 | if(n < config->collection.n) |
| 356 | fn(&config->collection.s[n]); |
| 357 | else |
| 358 | error(0, "no collection has root '%s'", s); |
| 359 | } |
| 360 | |
| 361 | /* rescan/recheck all collections */ |
| 362 | static void do_all(void (*fn)(const struct collection *c)) { |
| 363 | int n; |
| 364 | |
| 365 | for(n = 0; n < config->collection.n; ++n) |
| 366 | fn(&config->collection.s[n]); |
| 367 | /* TODO: we need to tidy up tracks from collections now removed. We could do |
| 368 | * this two ways: either remember collections we think there are and spot |
| 369 | * their disappearance, or iterate over all tracks and gc any that don't fit |
| 370 | * into some collection. |
| 371 | * |
| 372 | * Having a way to rename collections would be rather convenient too but |
| 373 | * that's another kettle of monkeys. |
| 374 | */ |
| 375 | } |
| 376 | |
| 377 | /** @brief Expire noticed.db */ |
| 378 | static void expire_noticed(void) { |
| 379 | time_t now; |
| 380 | |
| 381 | time(&now); |
| 382 | trackdb_expire_noticed(now - config->noticed_history * 86400); |
| 383 | } |
| 384 | |
| 385 | int main(int argc, char **argv) { |
| 386 | int n, logsyslog = !isatty(2); |
| 387 | struct sigaction sa; |
| 388 | int do_check = 1; |
| 389 | |
| 390 | set_progname(argv); |
| 391 | mem_init(); |
| 392 | if(!setlocale(LC_CTYPE, "")) fatal(errno, "error calling setlocale"); |
| 393 | while((n = getopt_long(argc, argv, "hVc:dDSsKC", options, 0)) >= 0) { |
| 394 | switch(n) { |
| 395 | case 'h': help(); |
| 396 | case 'V': version("disorder-rescan"); |
| 397 | case 'c': configfile = optarg; break; |
| 398 | case 'd': debugging = 1; break; |
| 399 | case 'D': debugging = 0; break; |
| 400 | case 'S': logsyslog = 0; break; |
| 401 | case 's': logsyslog = 1; break; |
| 402 | case 'K': do_check = 1; break; |
| 403 | case 'C': do_check = 0; break; |
| 404 | default: fatal(0, "invalid option"); |
| 405 | } |
| 406 | } |
| 407 | if(logsyslog) { |
| 408 | openlog(progname, LOG_PID, LOG_DAEMON); |
| 409 | log_default = &log_syslog; |
| 410 | } |
| 411 | if(config_read(0)) fatal(0, "cannot read configuration"); |
| 412 | xnice(config->nice_rescan); |
| 413 | sa.sa_handler = signal_handler; |
| 414 | sa.sa_flags = SA_RESTART; |
| 415 | sigemptyset(&sa.sa_mask); |
| 416 | xsigaction(SIGTERM, &sa, 0); |
| 417 | xsigaction(SIGINT, &sa, 0); |
| 418 | info("started"); |
| 419 | trackdb_init(TRACKDB_NO_RECOVER); |
| 420 | trackdb_open(TRACKDB_NO_UPGRADE); |
| 421 | if(optind == argc) { |
| 422 | /* Rescan all collections */ |
| 423 | do_all(rescan_collection); |
| 424 | /* Check that every track still exists */ |
| 425 | if(do_check) |
| 426 | recheck_collection(0); |
| 427 | /* Expire noticed.db */ |
| 428 | expire_noticed(); |
| 429 | } |
| 430 | else { |
| 431 | /* Rescan specified collections */ |
| 432 | for(n = optind; n < argc; ++n) |
| 433 | do_directory(argv[n], rescan_collection); |
| 434 | /* Check specified collections for tracks that have gone */ |
| 435 | if(do_check) |
| 436 | for(n = optind; n < argc; ++n) |
| 437 | do_directory(argv[n], recheck_collection); |
| 438 | } |
| 439 | trackdb_close(); |
| 440 | trackdb_deinit(); |
| 441 | info("completed"); |
| 442 | return 0; |
| 443 | } |
| 444 | |
| 445 | /* |
| 446 | Local Variables: |
| 447 | c-basic-offset:2 |
| 448 | comment-column:40 |
| 449 | fill-column:79 |
| 450 | indent-tabs-mode:nil |
| 451 | End: |
| 452 | */ |