src/recdb.c

   1 /* recdb.c - recursive/record database implementation
   2  * Copyright 2000-2004 srvx Development Team
   3  *
   4  * This file is part of x3.
   5  *
   6  * srvx is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with srvx; if not, write to the Free Software Foundation,
  18  * Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA.
  19  */
  20
  21 #include "recdb.h"
  22 #include "log.h"
  23
  24 #ifdef HAVE_FCNTL_H
  25 #include <fcntl.h>
  26 #endif
  27 #ifdef HAVE_SYS_MMAN_H
  28 #include <sys/mman.h>
  29 #endif
  30 #ifdef HAVE_SYS_STAT_H
  31 #include <sys/stat.h>
  32 #endif
  33
  34 /* 4 MiB on x86 */
  35 #define MMAP_MAP_LENGTH (getpagesize()*1024)
  36
  37 /* file format (grammar in Backus-Naur Form):
  38  *
  39  * database := record*
  40  * record := qstring [ '=' ] ( qstring | object | stringlist ) ';'
  41  * qstring := '"' ( [^\\] | ('\\' [\\n]) )* '"'
  42  * object := '{' record* '}'
  43  * stringlist := '(' [ qstring [',' qstring]* ] ')'
  44  *
  45  */
  46
  47 /* when a database or object is read from disk, it is represented as
  48  * a dictionary object, keys are names (what's left of the '=') and
  49  * values are 'struct record_data's
  50  */
  51
  52 struct recdb_context {
  53     int line;
  54     int col;
  55 };
  56
  57 enum recdb_filetype {
  58     RECDB_FILE,
  59     RECDB_STRING,
  60     RECDB_MMAP
  61 };
  62
  63 typedef struct recdb_file {
  64     const char *source;
  65     FILE *f; /* For RECDB_FILE, RECDB_MMAP */
  66     char *s; /* For RECDB_STRING, RECDB_MMAP */
  67     enum recdb_filetype type;
  68     size_t length;
  69     off_t pos;
  70     struct recdb_context ctx;
  71     jmp_buf env;
  72 } RECDB;
  73
  74 typedef struct recdb_outfile {
  75     FILE *f; /* For RECDB_FILE, RECDB_MMAP */
  76     char *s; /* For RECDB_STRING, RECDB_MMAP */
  77     union {
  78         struct { /* For RECDB_STRING */
  79             size_t chunksize;
  80             size_t alloc_length;
  81         } s;
  82         struct { /* For RECDB_MMAP */
  83             off_t mmap_begin;
  84             size_t mmap_length;
  85         } m;
  86     } state;
  87     enum recdb_filetype type;
  88     off_t pos;
  89     int tablvl;
  90 #ifdef NDEBUG
  91     int need_tab;
  92 #endif
  93 } RECDB_OUT;
  94
  95 #ifdef HAVE_MMAP
  96 static int mmap_error=0;
  97 #endif
  98
  99 #define EOL '\n'
 100
 101 #if 1
 102 #define ABORT(recdb, code, ch) longjmp((recdb)->env, ((code) << 8) | (ch))
 103 #else
 104 static void
 105 ABORT(RECDB *recdb, int code, unsigned char ch) {
 106     longjmp(recdb->env, code << 8 | ch);
 107 }
 108 #endif
 109
 110 enum fail_codes {
 111     UNTERMINATED_STRING,
 112     EXPECTED_OPEN_QUOTE,
 113     EXPECTED_OPEN_BRACE,
 114     EXPECTED_OPEN_PAREN,
 115     EXPECTED_COMMA,
 116     EXPECTED_START_RECORD_DATA,
 117     EXPECTED_SEMICOLON,
 118     EXPECTED_RECORD_DATA
 119 };
 120
 121 static void parse_record_int(RECDB *recdb, char **pname, struct record_data **prd);
 122
 123 /* allocation functions */
 124
 125 #define alloc_record_data_int() malloc(sizeof(struct record_data))
 126
 127 struct record_data *
 128 alloc_record_data_qstring(const char *string)
 129 {
 130     struct record_data *rd;
 131     rd = alloc_record_data_int();
 132     SET_RECORD_QSTRING(rd, string);
 133     return rd;
 134 }
 135
 136 struct record_data *
 137 alloc_record_data_object(dict_t obj)
 138 {
 139     struct record_data *rd;
 140     rd = alloc_record_data_int();
 141     SET_RECORD_OBJECT(rd, obj);
 142     return rd;
 143 }
 144
 145 struct record_data *
 146 alloc_record_data_string_list(struct string_list *slist)
 147 {
 148     struct record_data *rd;
 149     rd = alloc_record_data_int();
 150     SET_RECORD_STRING_LIST(rd, slist);
 151     return rd;
 152 }
 153
 154 struct string_list*
 155 alloc_string_list(int size)
 156 {
 157     struct string_list *slist;
 158     slist = malloc(sizeof(struct string_list));
 159     slist->used = 0;
 160     slist->size = size;
 161     slist->list = slist->size ? malloc(size*sizeof(char*)) : NULL;
 162     return slist;
 163 }
 164
 165 dict_t
 166 alloc_database(void)
 167 {
 168     dict_t db = dict_new();
 169     dict_set_free_data(db, free_record_data);
 170     return db;
 171 }
 172
 173 /* misc. operations */
 174
 175 void
 176 string_list_append(struct string_list *slist, char *string)
 177 {
 178     if (slist->used == slist->size) {
 179         if (slist->size) {
 180             slist->size <<= 1;
 181             slist->list = realloc(slist->list, slist->size*sizeof(char*));
 182         } else {
 183             slist->size = 4;
 184             slist->list = malloc(slist->size*sizeof(char*));
 185         }
 186     }
 187     slist->list[slist->used++] = string;
 188 }
 189
 190 struct string_list *
 191 string_list_copy(struct string_list *slist)
 192 {
 193     struct string_list *new_list;
 194     unsigned int i;
 195     new_list = alloc_string_list(slist->size);
 196     new_list->used = slist->used;
 197     for (i=0; i<new_list->used; i++) {
 198         new_list->list[i] = strdup(slist->list[i]);
 199     }
 200     return new_list;
 201 }
 202
 203 int slist_compare_two(const void *pa, const void *pb)
 204 {
 205     return irccasecmp(*(const char**)pa, *(const char **)pb);
 206 }
 207
 208 void
 209 string_list_sort(struct string_list *slist)
 210 {
 211     qsort(slist->list, slist->used, sizeof(slist->list[0]), slist_compare_two);
 212 }
 213
 214 struct record_data*
 215 database_get_path(dict_t db, const char *path)
 216 {
 217     char *new_path = strdup(path), *orig_path = new_path;
 218     char *part;
 219     struct record_data *rd;
 220
 221     for (part=new_path; *new_path; new_path++) {
 222         if (*new_path != '/') continue;
 223         *new_path = 0;
 224
 225         rd = dict_find(db, part, NULL);
 226         if (!rd || rd->type != RECDB_OBJECT) {
 227             free(orig_path);
 228             return NULL;
 229         }
 230
 231         db = rd->d.object;
 232         part = new_path+1;
 233     }
 234
 235     rd = dict_find(db, part, NULL);
 236     free(orig_path);
 237     return rd;
 238 }
 239
 240 void*
 241 database_get_data(dict_t db, const char *path, enum recdb_type type)
 242 {
 243     struct record_data *rd = database_get_path(db, path);
 244     return (rd && rd->type == type) ? rd->d.whatever : NULL;
 245 }
 246
 247 /* free functions */
 248
 249 void
 250 free_string_list(struct string_list *slist)
 251 {
 252     unsigned int i;
 253     if (!slist)
 254         return;
 255     for (i=0; i<slist->used; i++)
 256         free(slist->list[i]);
 257     free(slist->list);
 258     free(slist);
 259 }
 260
 261 void
 262 free_record_data(void *rdata)
 263 {
 264     struct record_data *r = rdata;
 265     switch (r->type) {
 266     case RECDB_INVALID: break;
 267     case RECDB_QSTRING: free(r->d.qstring); break;
 268     case RECDB_OBJECT: dict_delete(r->d.object); break;
 269     case RECDB_STRING_LIST: free_string_list(r->d.slist); break;
 270     }
 271     free(r);
 272 }
 273
 274 /* parse functions */
 275
 276 static int
 277 dbeof(RECDB *recdb)
 278 {
 279     switch (recdb->type) {
 280         case RECDB_FILE:
 281             return feof(recdb->f);
 282             break;
 283         case RECDB_STRING:
 284             return !*recdb->s;
 285             break;
 286         case RECDB_MMAP:
 287             return ((size_t)recdb->pos >= recdb->length);
 288             break;
 289         default:
 290             return 1;
 291             break;
 292     }
 293 }
 294
 295 static int
 296 dbgetc(RECDB *recdb)
 297 {
 298     int res;
 299     switch (recdb->type) {
 300         case RECDB_FILE:
 301             res = fgetc(recdb->f);
 302             break;
 303         case RECDB_STRING:
 304         case RECDB_MMAP:
 305             res = dbeof(recdb) ? EOF : recdb->s[recdb->pos++];
 306             break;
 307         default:
 308             res = EOF;
 309             break;
 310     }
 311     if (res == EOL) recdb->ctx.line++, recdb->ctx.col=1;
 312     else if (res != EOF) recdb->ctx.col++;
 313     return res;
 314 }
 315
 316 static void
 317 dbungetc(int c, RECDB *recdb)
 318 {
 319     switch (recdb->type) {
 320         case RECDB_FILE:
 321             ungetc(c, recdb->f);
 322             break;
 323         case RECDB_STRING:
 324         case RECDB_MMAP:
 325             recdb->s[--recdb->pos] = c;
 326             break;
 327     }
 328     if (c == EOL) recdb->ctx.line--, recdb->ctx.col=-1;
 329     else recdb->ctx.col--;
 330 }
 331
 332 /* returns first non-whitespace, non-comment character (-1 for EOF found) */
 333 int
 334 parse_skip_ws(RECDB *recdb)
 335 {
 336     int c, d, in_comment = 0;
 337     while (!dbeof(recdb)) {
 338         c = dbgetc(recdb);
 339         if (c == EOF) return EOF;
 340         if (isspace(c)) continue;
 341         if (c != '/') return c;
 342         if ((d = dbgetc(recdb)) == '*') {
 343             /* C style comment, with slash star comment star slash */
 344             in_comment = 1;
 345             do {
 346                 do {
 347                     c = dbgetc(recdb);
 348                 } while (c != '*' && c != EOF);
 349                 if ((c = dbgetc(recdb)) == '/') in_comment = 0;
 350             } while (in_comment);
 351         } else if (d == '/') {
 352             /* C++ style comment, with slash slash comment newline */
 353             do {
 354                 c = dbgetc(recdb);
 355             } while (c != EOF && c != EOL);
 356         } else {
 357             if (d != EOF) dbungetc(d, recdb);
 358             return c;
 359         }
 360     }
 361     return -1;
 362 }
 363
 364 char *
 365 parse_qstring(RECDB *recdb)
 366 {
 367     char *buff;
 368     int used=0, size=8, c;
 369     struct recdb_context start_ctx;
 370     unsigned int i;
 371
 372     if ((c = parse_skip_ws(recdb)) == EOF) return NULL;
 373     start_ctx = recdb->ctx;
 374     if (c != '"') ABORT(recdb, EXPECTED_OPEN_QUOTE, c);
 375     buff = malloc(size);
 376     while (!dbeof(recdb) && (c = dbgetc(recdb)) != '"') {
 377         if (c != '\\') {
 378             /* There should never be a literal newline, as it is saved as a \n */
 379             if (c == EOL) {
 380                 dbungetc(c, recdb);
 381                 ABORT(recdb, UNTERMINATED_STRING, ' ');
 382             }
 383             buff[used++] = c;
 384         } else {
 385             switch (c = dbgetc(recdb)) {
 386                 case '0': /* \<octal>, 000 through 377 */
 387                 case '1':
 388                 case '2':
 389                 case '3':
 390                 case '4':
 391                 case '5':
 392                 case '6':
 393                 case '7':
 394                     {
 395                         char digits[3] = { (char)c, '\0', '\0' };
 396                         for (i=1; i < 3; i++) {
 397                             /* Maximum of \377, so there's a max of 2 digits
 398                              * if digits[0] > '3' (no \400, but \40 is fine) */
 399                             if (i == 2 && digits[0] > '3') {
 400                                 break;
 401                             }
 402                             if ((c = dbgetc(recdb)) == EOF) {
 403                                 break;
 404                             }
 405                             if ((c < '0') || (c > '7')) {
 406                                 dbungetc(c, recdb);
 407                                 break;
 408                             }
 409                             digits[i] = (char)c;
 410                         }
 411                         if (i) {
 412                             c = (int)strtol(digits, NULL, 8);
 413                             buff[used++] = c;
 414                         } else {
 415                             buff[used++] = '\0';
 416                         }
 417                     }
 418                     break;
 419                 case 'x': /* Hex */
 420                     {
 421                         char digits[3] = { '\0', '\0', '\0' };
 422                         for (i=0; i < 2; i++) {
 423                             if ((c = dbgetc(recdb)) == EOF) {
 424                                 break;
 425                             }
 426                             if (!isxdigit(c)) {
 427                                 dbungetc(c, recdb);
 428                                 break;
 429                             }
 430                             digits[i] = (char)c;
 431                         }
 432                         if (i) {
 433                             c = (int)strtol(digits, NULL, 16);
 434                             buff[used++] = c;
 435                         } else {
 436                             buff[used++] = '\\';
 437                             buff[used++] = 'x';
 438                         }
 439                     }
 440                     break;
 441                 case 'a': buff[used++] = '\a'; break;
 442                 case 'b': buff[used++] = '\b'; break;
 443                 case 't': buff[used++] = '\t'; break;
 444                 case 'n': buff[used++] = EOL; break;
 445                 case 'v': buff[used++] = '\v'; break;
 446                 case 'f': buff[used++] = '\f'; break;
 447                 case 'r': buff[used++] = '\r'; break;
 448                 case '\\': buff[used++] = '\\'; break;
 449                 case '"': buff[used++] = '"'; break;
 450                 default: buff[used++] = '\\'; buff[used++] = c; break;
 451             }
 452         }
 453         if (used == size) {
 454             size <<= 1;
 455             buff = realloc(buff, size);
 456         }
 457     }
 458     if (c != '"' && dbeof(recdb)) {
 459         free(buff);
 460         recdb->ctx = start_ctx;
 461         ABORT(recdb, UNTERMINATED_STRING, EOF);
 462     }
 463     buff[used] = 0;
 464     return buff;
 465 }
 466
 467 dict_t
 468 parse_object(RECDB *recdb)
 469 {
 470     dict_t obj;
 471     char *name;
 472     struct record_data *rd;
 473     int c;
 474     if ((c = parse_skip_ws(recdb)) == EOF) return NULL;
 475     if (c != '{') ABORT(recdb, EXPECTED_OPEN_BRACE, c);
 476     obj = alloc_object();
 477     dict_set_free_keys(obj, free);
 478     while (!dbeof(recdb)) {
 479         if ((c = parse_skip_ws(recdb)) == '}') break;
 480         if (c == EOF) break;
 481         dbungetc(c, recdb);
 482         parse_record_int(recdb, &name, &rd);
 483         dict_insert(obj, name, rd);
 484     }
 485     return obj;
 486 }
 487
 488 struct string_list *
 489 parse_string_list(RECDB *recdb)
 490 {
 491     struct string_list *slist;
 492     int c;
 493     if ((c = parse_skip_ws(recdb)) == EOF) return NULL;
 494     if (c != '(') ABORT(recdb, EXPECTED_OPEN_PAREN, c);
 495     slist = alloc_string_list(4);
 496     while (true) {
 497         c = parse_skip_ws(recdb);
 498         if (c == EOF || c == ')') break;
 499         dbungetc(c, recdb);
 500         string_list_append(slist, parse_qstring(recdb));
 501         c = parse_skip_ws(recdb);
 502         if (c == EOF || c == ')') break;
 503         if (c != ',') ABORT(recdb, EXPECTED_COMMA, c);
 504     }
 505     return slist;
 506 }
 507
 508 static void
 509 parse_record_int(RECDB *recdb, char **pname, struct record_data **prd)
 510 {
 511     int c;
 512     *pname = parse_qstring(recdb);
 513     c = parse_skip_ws(recdb);
 514     if (c == EOF) {
 515         if (!*pname) return;
 516         free(*pname);
 517         ABORT(recdb, EXPECTED_RECORD_DATA, EOF);
 518     }
 519     if (c == '=') c = parse_skip_ws(recdb);
 520     dbungetc(c, recdb);
 521     *prd = malloc(sizeof(**prd));
 522     switch (c) {
 523     case '"':
 524         /* Don't use SET_RECORD_QSTRING, since that does an extra strdup() of the string. */
 525         (*prd)->type = RECDB_QSTRING;
 526         (*prd)->d.qstring = parse_qstring(recdb);
 527         break;
 528     case '{': SET_RECORD_OBJECT(*prd, parse_object(recdb)); break;
 529     case '(': SET_RECORD_STRING_LIST(*prd, parse_string_list(recdb)); break;
 530     default: ABORT(recdb, EXPECTED_START_RECORD_DATA, c);
 531     }
 532     if ((c = parse_skip_ws(recdb)) != ';') ABORT(recdb, EXPECTED_SEMICOLON, c);
 533 }
 534
 535 static dict_t
 536 parse_database_int(RECDB *recdb)
 537 {
 538     char *name;
 539     struct record_data *rd;
 540     dict_t db = alloc_database();
 541     dict_set_free_keys(db, free);
 542     while (!dbeof(recdb)) {
 543         parse_record_int(recdb, &name, &rd);
 544         if (name) dict_insert(db, name, rd);
 545     }
 546     return db;
 547 }
 548
 549 const char *
 550 failure_reason(int code)
 551 {
 552     const char *reason;
 553     switch (code >> 8) {
 554     case UNTERMINATED_STRING: reason = "Unterminated string"; break;
 555     case EXPECTED_OPEN_QUOTE: reason = "Expected '\"'"; break;
 556     case EXPECTED_OPEN_BRACE: reason = "Expected '{'"; break;
 557     case EXPECTED_OPEN_PAREN: reason = "Expected '('"; break;
 558     case EXPECTED_COMMA: reason = "Expected ','"; break;
 559     case EXPECTED_START_RECORD_DATA: reason = "Expected start of some record data"; break;
 560     case EXPECTED_SEMICOLON: reason = "Expected ';'"; break;
 561     case EXPECTED_RECORD_DATA: reason = "Expected record data"; break;
 562     default: reason = "Unknown error";
 563     }
 564     if (code == -1) reason = "Premature end of file";
 565     return reason;
 566 }
 567
 568 void
 569 explain_failure(RECDB *recdb, int code)
 570 {
 571     log_module(MAIN_LOG, LOG_ERROR, "%s (got '%c') at %s line %d column %d.",
 572                failure_reason(code), code & 255,
 573                recdb->source, recdb->ctx.line, recdb->ctx.col);
 574 }
 575
 576 const char *
 577 parse_record(const char *text, char **pname, struct record_data **prd)
 578 {
 579     RECDB recdb;
 580     int res;
 581     *pname = NULL;
 582     *prd = NULL;
 583     recdb.source = "<user-supplied text>";
 584     recdb.f = NULL;
 585     recdb.s = strdup(text);
 586     recdb.length = strlen(text);
 587     recdb.pos = 0;
 588     recdb.type = RECDB_STRING;
 589     recdb.ctx.line = recdb.ctx.col = 1;
 590     if ((res = setjmp(recdb.env)) == 0) {
 591         parse_record_int(&recdb, pname, prd);
 592         return 0;
 593     } else {
 594         free(*pname);
 595         free(*prd);
 596         return failure_reason(res);
 597     }
 598 }
 599
 600 dict_t
 601 parse_database(const char *filename)
 602 {
 603     RECDB recdb;
 604     int res;
 605     dict_t db;
 606     struct stat statinfo;
 607
 608     recdb.source = filename;
 609     if (!(recdb.f = fopen(filename, "r"))) {
 610         log_module(MAIN_LOG, LOG_ERROR, "Unable to open database file '%s' for reading: %s", filename, strerror(errno));
 611         return NULL;
 612     }
 613
 614     if (fstat(fileno(recdb.f), &statinfo)) {
 615         log_module(MAIN_LOG, LOG_ERROR, "Unable to fstat database file '%s': %s", filename, strerror(errno));
 616         return NULL;
 617     }
 618     recdb.length = (size_t)statinfo.st_size;
 619
 620 #ifdef HAVE_MMAP
 621     /* Try mmap */
 622     if (!mmap_error && (recdb.s = mmap(NULL, recdb.length, PROT_READ|PROT_WRITE, MAP_PRIVATE, fileno(recdb.f), 0)) != MAP_FAILED) {
 623         recdb.type = RECDB_MMAP;
 624     } else {
 625         /* Fall back to stdio */
 626         if (!mmap_error) {
 627             log_module(MAIN_LOG, LOG_WARNING, "Unable to mmap database file '%s' (falling back to stdio): %s", filename, strerror(errno));
 628             mmap_error = 1;
 629         }
 630 #else
 631     if (1) {
 632 #endif
 633         recdb.s = NULL;
 634         recdb.type = RECDB_FILE;
 635     }
 636
 637     recdb.ctx.line = recdb.ctx.col = 1;
 638     recdb.pos = 0;
 639
 640     if ((res = setjmp(recdb.env)) == 0) {
 641         db = parse_database_int(&recdb);
 642     } else {
 643         explain_failure(&recdb, res);
 644         _exit(1);
 645     }
 646
 647     switch (recdb.type) {
 648         case RECDB_MMAP:
 649 #ifdef HAVE_MMAP
 650             munmap(recdb.s, recdb.length);
 651 #endif
 652         case RECDB_FILE:
 653             fclose(recdb.f);
 654             break;
 655         /* Appease gcc */
 656         default:
 657             break;
 658     }
 659     return db;
 660 }