keyword-gen.c revision 1.3.16.4
1 /* $NetBSD: keyword-gen.c,v 1.3.16.4 2017/05/04 06:01:01 snj Exp $ */ 2 3 /* 4 * keyword-gen.c -- generate keyword scanner finite state machine and 5 * keyword_text array. 6 * 7 * This program is run to generate ntp_keyword.h 8 * After making a change here, two output files should be committed at 9 * the same time as keyword-gen.c: 10 * ntp_keyword.h 11 * keyword-gen-utd 12 * 13 * keyword-gen-utd is a sentinel used by Makefile.am to avoid compiling 14 * keyword_gen.c and generating ntp_keyword.h if the input keyword-gen.c 15 * has not changed. This is not solely an optimization, it also breaks 16 * a dependency chain that otherwise would cause programs to be compiled 17 * when running "make dist" or "make distdir". We want these to package 18 * the existing source without building anything but a tarball. See 19 * [Bug 1470]. 20 */ 21 #include <config.h> 22 #include <stdio.h> 23 #include <stdlib.h> 24 #include <time.h> 25 26 #include <ntp_stdlib.h> 27 #include <ntp_config.h> 28 #include <lib_strbuf.h> 29 #include "ntp_scanner.h" 30 #include "ntp_parser.h" 31 32 33 /* Define a structure to hold a (keyword, token) pair */ 34 struct key_tok { 35 char * key; /* Keyword */ 36 u_short token; /* Associated Token */ 37 follby followedby; /* nonzero indicates the next token(s) 38 forced to be string(s) */ 39 }; 40 41 struct key_tok ntp_keywords[] = { 42 { "...", T_Ellipsis, FOLLBY_TOKEN }, 43 { "allpeers", T_Allpeers, FOLLBY_TOKEN }, 44 { "automax", T_Automax, FOLLBY_TOKEN }, 45 { "broadcast", T_Broadcast, FOLLBY_STRING }, 46 { "broadcastclient", T_Broadcastclient, FOLLBY_TOKEN }, 47 { "broadcastdelay", T_Broadcastdelay, FOLLBY_TOKEN }, 48 { "ctl", T_Ctl, FOLLBY_TOKEN }, 49 { "disable", T_Disable, FOLLBY_TOKEN }, 50 { "driftfile", T_Driftfile, FOLLBY_STRING }, 51 { "dscp", T_Dscp, FOLLBY_TOKEN }, 52 { "enable", T_Enable, FOLLBY_TOKEN }, 53 { "end", T_End, FOLLBY_TOKEN }, 54 { "filegen", T_Filegen, FOLLBY_TOKEN }, 55 { "fudge", T_Fudge, FOLLBY_STRING }, 56 { "io", T_Io, FOLLBY_TOKEN }, 57 { "includefile", T_Includefile, FOLLBY_STRING }, 58 { "leapfile", T_Leapfile, FOLLBY_STRING }, 59 { "leapsmearinterval", T_Leapsmearinterval, FOLLBY_TOKEN }, 60 { "logconfig", T_Logconfig, FOLLBY_STRINGS_TO_EOC }, 61 { "logfile", T_Logfile, FOLLBY_STRING }, 62 { "manycastclient", T_Manycastclient, FOLLBY_STRING }, 63 { "manycastserver", T_Manycastserver, FOLLBY_STRINGS_TO_EOC }, 64 { "mem", T_Mem, FOLLBY_TOKEN }, 65 { "multicastclient", T_Multicastclient, FOLLBY_STRINGS_TO_EOC }, 66 { "peer", T_Peer, FOLLBY_STRING }, 67 { "phone", T_Phone, FOLLBY_STRINGS_TO_EOC }, 68 { "pidfile", T_Pidfile, FOLLBY_STRING }, 69 { "pool", T_Pool, FOLLBY_STRING }, 70 { "discard", T_Discard, FOLLBY_TOKEN }, 71 { "reset", T_Reset, FOLLBY_TOKEN }, 72 { "restrict", T_Restrict, FOLLBY_TOKEN }, 73 { "rlimit", T_Rlimit, FOLLBY_TOKEN }, 74 { "server", T_Server, FOLLBY_STRING }, 75 { "setvar", T_Setvar, FOLLBY_STRING }, 76 { "statistics", T_Statistics, FOLLBY_TOKEN }, 77 { "statsdir", T_Statsdir, FOLLBY_STRING }, 78 { "sys", T_Sys, FOLLBY_TOKEN }, 79 { "tick", T_Tick, FOLLBY_TOKEN }, 80 { "timer", T_Timer, FOLLBY_TOKEN }, 81 { "tinker", T_Tinker, FOLLBY_TOKEN }, 82 { "tos", T_Tos, FOLLBY_TOKEN }, 83 { "trap", T_Trap, FOLLBY_STRING }, 84 { "unconfig", T_Unconfig, FOLLBY_STRING }, 85 { "unpeer", T_Unpeer, FOLLBY_STRING }, 86 /* authentication_command */ 87 { "controlkey", T_ControlKey, FOLLBY_TOKEN }, 88 { "crypto", T_Crypto, FOLLBY_TOKEN }, 89 { "keys", T_Keys, FOLLBY_STRING }, 90 { "keysdir", T_Keysdir, FOLLBY_STRING }, 91 { "ntpsigndsocket", T_NtpSignDsocket, FOLLBY_STRING }, 92 { "requestkey", T_Requestkey, FOLLBY_TOKEN }, 93 { "revoke", T_Revoke, FOLLBY_TOKEN }, 94 { "trustedkey", T_Trustedkey, FOLLBY_TOKEN }, 95 /* IPv4/IPv6 protocol override flag */ 96 { "-4", T_Ipv4_flag, FOLLBY_TOKEN }, 97 { "-6", T_Ipv6_flag, FOLLBY_TOKEN }, 98 /* option */ 99 { "autokey", T_Autokey, FOLLBY_TOKEN }, 100 { "burst", T_Burst, FOLLBY_TOKEN }, 101 { "iburst", T_Iburst, FOLLBY_TOKEN }, 102 { "key", T_Key, FOLLBY_TOKEN }, 103 { "maxpoll", T_Maxpoll, FOLLBY_TOKEN }, 104 { "mdnstries", T_Mdnstries, FOLLBY_TOKEN }, 105 { "minpoll", T_Minpoll, FOLLBY_TOKEN }, 106 { "mode", T_Mode, FOLLBY_TOKEN }, 107 { "noselect", T_Noselect, FOLLBY_TOKEN }, 108 { "preempt", T_Preempt, FOLLBY_TOKEN }, 109 { "true", T_True, FOLLBY_TOKEN }, 110 { "prefer", T_Prefer, FOLLBY_TOKEN }, 111 { "ttl", T_Ttl, FOLLBY_TOKEN }, 112 { "version", T_Version, FOLLBY_TOKEN }, 113 { "xleave", T_Xleave, FOLLBY_TOKEN }, 114 /* crypto_command */ 115 { "host", T_Host, FOLLBY_STRING }, 116 { "ident", T_Ident, FOLLBY_STRING }, 117 { "pw", T_Pw, FOLLBY_STRING }, 118 { "randfile", T_Randfile, FOLLBY_STRING }, 119 { "digest", T_Digest, FOLLBY_STRING }, 120 /*** MONITORING COMMANDS ***/ 121 /* stat */ 122 { "clockstats", T_Clockstats, FOLLBY_TOKEN }, 123 { "cryptostats", T_Cryptostats, FOLLBY_TOKEN }, 124 { "loopstats", T_Loopstats, FOLLBY_TOKEN }, 125 { "peerstats", T_Peerstats, FOLLBY_TOKEN }, 126 { "rawstats", T_Rawstats, FOLLBY_TOKEN }, 127 { "sysstats", T_Sysstats, FOLLBY_TOKEN }, 128 { "protostats", T_Protostats, FOLLBY_TOKEN }, 129 { "timingstats", T_Timingstats, FOLLBY_TOKEN }, 130 /* filegen_option */ 131 { "file", T_File, FOLLBY_STRING }, 132 { "link", T_Link, FOLLBY_TOKEN }, 133 { "nolink", T_Nolink, FOLLBY_TOKEN }, 134 { "type", T_Type, FOLLBY_TOKEN }, 135 /* filegen_type */ 136 { "age", T_Age, FOLLBY_TOKEN }, 137 { "day", T_Day, FOLLBY_TOKEN }, 138 { "month", T_Month, FOLLBY_TOKEN }, 139 { "none", T_None, FOLLBY_TOKEN }, 140 { "pid", T_Pid, FOLLBY_TOKEN }, 141 { "week", T_Week, FOLLBY_TOKEN }, 142 { "year", T_Year, FOLLBY_TOKEN }, 143 /*** ORPHAN MODE COMMANDS ***/ 144 /* tos_option */ 145 { "minclock", T_Minclock, FOLLBY_TOKEN }, 146 { "maxclock", T_Maxclock, FOLLBY_TOKEN }, 147 { "minsane", T_Minsane, FOLLBY_TOKEN }, 148 { "floor", T_Floor, FOLLBY_TOKEN }, 149 { "ceiling", T_Ceiling, FOLLBY_TOKEN }, 150 { "cohort", T_Cohort, FOLLBY_TOKEN }, 151 { "mindist", T_Mindist, FOLLBY_TOKEN }, 152 { "maxdist", T_Maxdist, FOLLBY_TOKEN }, 153 { "bcpollbstep", T_Bcpollbstep, FOLLBY_TOKEN }, 154 { "beacon", T_Beacon, FOLLBY_TOKEN }, 155 { "orphan", T_Orphan, FOLLBY_TOKEN }, 156 { "orphanwait", T_Orphanwait, FOLLBY_TOKEN }, 157 { "nonvolatile", T_Nonvolatile, FOLLBY_TOKEN }, 158 /* access_control_flag */ 159 { "default", T_Default, FOLLBY_TOKEN }, 160 { "source", T_Source, FOLLBY_TOKEN }, 161 { "flake", T_Flake, FOLLBY_TOKEN }, 162 { "ignore", T_Ignore, FOLLBY_TOKEN }, 163 { "limited", T_Limited, FOLLBY_TOKEN }, 164 { "mssntp", T_Mssntp, FOLLBY_TOKEN }, 165 { "kod", T_Kod, FOLLBY_TOKEN }, 166 { "lowpriotrap", T_Lowpriotrap, FOLLBY_TOKEN }, 167 { "mask", T_Mask, FOLLBY_TOKEN }, 168 { "nomodify", T_Nomodify, FOLLBY_TOKEN }, 169 { "nomrulist", T_Nomrulist, FOLLBY_TOKEN }, 170 { "nopeer", T_Nopeer, FOLLBY_TOKEN }, 171 { "noquery", T_Noquery, FOLLBY_TOKEN }, 172 { "noserve", T_Noserve, FOLLBY_TOKEN }, 173 { "notrap", T_Notrap, FOLLBY_TOKEN }, 174 { "notrust", T_Notrust, FOLLBY_TOKEN }, 175 { "ntpport", T_Ntpport, FOLLBY_TOKEN }, 176 /* discard_option */ 177 { "average", T_Average, FOLLBY_TOKEN }, 178 { "minimum", T_Minimum, FOLLBY_TOKEN }, 179 { "monitor", T_Monitor, FOLLBY_TOKEN }, 180 /* mru_option */ 181 { "incalloc", T_Incalloc, FOLLBY_TOKEN }, 182 { "incmem", T_Incmem, FOLLBY_TOKEN }, 183 { "initalloc", T_Initalloc, FOLLBY_TOKEN }, 184 { "initmem", T_Initmem, FOLLBY_TOKEN }, 185 { "mindepth", T_Mindepth, FOLLBY_TOKEN }, 186 { "maxage", T_Maxage, FOLLBY_TOKEN }, 187 { "maxdepth", T_Maxdepth, FOLLBY_TOKEN }, 188 { "maxmem", T_Maxmem, FOLLBY_TOKEN }, 189 { "mru", T_Mru, FOLLBY_TOKEN }, 190 /* fudge_factor */ 191 { "abbrev", T_Abbrev, FOLLBY_STRING }, 192 { "flag1", T_Flag1, FOLLBY_TOKEN }, 193 { "flag2", T_Flag2, FOLLBY_TOKEN }, 194 { "flag3", T_Flag3, FOLLBY_TOKEN }, 195 { "flag4", T_Flag4, FOLLBY_TOKEN }, 196 { "refid", T_Refid, FOLLBY_STRING }, 197 { "stratum", T_Stratum, FOLLBY_TOKEN }, 198 { "time1", T_Time1, FOLLBY_TOKEN }, 199 { "time2", T_Time2, FOLLBY_TOKEN }, 200 /* system_option */ 201 { "auth", T_Auth, FOLLBY_TOKEN }, 202 { "bclient", T_Bclient, FOLLBY_TOKEN }, 203 { "calibrate", T_Calibrate, FOLLBY_TOKEN }, 204 { "kernel", T_Kernel, FOLLBY_TOKEN }, 205 { "mode7", T_Mode7, FOLLBY_TOKEN }, 206 { "ntp", T_Ntp, FOLLBY_TOKEN }, 207 { "peer_clear_digest_early", T_PCEdigest, FOLLBY_TOKEN }, 208 { "stats", T_Stats, FOLLBY_TOKEN }, 209 { "unpeer_crypto_early", T_UEcrypto, FOLLBY_TOKEN }, 210 { "unpeer_crypto_nak_early", T_UEcryptonak, FOLLBY_TOKEN }, 211 { "unpeer_digest_early", T_UEdigest, FOLLBY_TOKEN }, 212 /* rlimit_option */ 213 { "memlock", T_Memlock, FOLLBY_TOKEN }, 214 { "stacksize", T_Stacksize, FOLLBY_TOKEN }, 215 { "filenum", T_Filenum, FOLLBY_TOKEN }, 216 /* tinker_option */ 217 { "step", T_Step, FOLLBY_TOKEN }, 218 { "stepback", T_Stepback, FOLLBY_TOKEN }, 219 { "stepfwd", T_Stepfwd, FOLLBY_TOKEN }, 220 { "panic", T_Panic, FOLLBY_TOKEN }, 221 { "dispersion", T_Dispersion, FOLLBY_TOKEN }, 222 { "stepout", T_Stepout, FOLLBY_TOKEN }, 223 { "allan", T_Allan, FOLLBY_TOKEN }, 224 { "huffpuff", T_Huffpuff, FOLLBY_TOKEN }, 225 { "freq", T_Freq, FOLLBY_TOKEN }, 226 /* miscellaneous_command */ 227 { "port", T_Port, FOLLBY_TOKEN }, 228 { "interface", T_Interface, FOLLBY_TOKEN }, 229 { "saveconfigdir", T_Saveconfigdir, FOLLBY_STRING }, 230 /* interface_command (ignore and interface already defined) */ 231 { "nic", T_Nic, FOLLBY_TOKEN }, 232 { "all", T_All, FOLLBY_TOKEN }, 233 { "ipv4", T_Ipv4, FOLLBY_TOKEN }, 234 { "ipv6", T_Ipv6, FOLLBY_TOKEN }, 235 { "wildcard", T_Wildcard, FOLLBY_TOKEN }, 236 { "listen", T_Listen, FOLLBY_TOKEN }, 237 { "drop", T_Drop, FOLLBY_TOKEN }, 238 /* simulator commands */ 239 { "simulate", T_Simulate, FOLLBY_TOKEN }, 240 { "simulation_duration",T_Sim_Duration, FOLLBY_TOKEN }, 241 { "beep_delay", T_Beep_Delay, FOLLBY_TOKEN }, 242 { "duration", T_Duration, FOLLBY_TOKEN }, 243 { "server_offset", T_Server_Offset, FOLLBY_TOKEN }, 244 { "freq_offset", T_Freq_Offset, FOLLBY_TOKEN }, 245 { "wander", T_Wander, FOLLBY_TOKEN }, 246 { "jitter", T_Jitter, FOLLBY_TOKEN }, 247 { "prop_delay", T_Prop_Delay, FOLLBY_TOKEN }, 248 { "proc_delay", T_Proc_Delay, FOLLBY_TOKEN }, 249 }; 250 251 typedef struct big_scan_state_tag { 252 char ch; /* Character this state matches on */ 253 char followedby; /* Forces next token(s) to T_String */ 254 u_short finishes_token; /* nonzero ID if last keyword char */ 255 u_short match_next_s; /* next state to check matching ch */ 256 u_short other_next_s; /* next state to check if not ch */ 257 } big_scan_state; 258 259 /* 260 * Note: to increase MAXSTATES beyond 2048, be aware it is currently 261 * crammed into 11 bits in scan_state form. Raising to 4096 would be 262 * relatively easy by storing the followedby value in a separate 263 * array with one entry per token, and shrinking the char value to 264 * 7 bits to free a bit for accepting/non-accepting. More than 4096 265 * states will require expanding scan_state beyond 32 bits each. 266 */ 267 #define MAXSTATES 2048 268 #define MAX_TOK_LEN 63 269 270 const char * current_keyword;/* for error reporting */ 271 big_scan_state sst[MAXSTATES]; /* scanner FSM state entries */ 272 u_short sst_highwater; /* next entry index to consider */ 273 char * symb[1024]; /* map token ID to symbolic name */ 274 275 /* for libntp */ 276 const char * progname = "keyword-gen"; 277 278 int main (int, char **); 279 static void generate_preamble (void); 280 static void generate_fsm (void); 281 static void generate_token_text (void); 282 static u_short create_keyword_scanner (void); 283 static u_short create_scan_states (char *, u_short, follby, u_short); 284 int compare_key_tok_id (const void *, const void *); 285 int compare_key_tok_text (const void *, const void *); 286 void populate_symb (char *); 287 const char * symbname (u_short); 288 289 290 int main(int argc, char **argv) 291 { 292 if (argc < 2) { 293 fprintf(stderr, "Usage:\n%s t_header.h\n", argv[0]); 294 exit(1); 295 } 296 debug = 1; 297 298 populate_symb(argv[1]); 299 300 generate_preamble(); 301 generate_token_text(); 302 generate_fsm(); 303 304 return 0; 305 } 306 307 308 static void 309 generate_preamble(void) 310 { 311 time_t now; 312 char timestamp[128]; 313 char preamble[] = 314 "/*\n" 315 " * ntp_keyword.h\n" 316 " * \n" 317 " * NOTE: edit this file with caution, it is generated by keyword-gen.c\n" 318 " *\t Generated %s UTC diff_ignore_line\n" 319 " *\n" 320 " */\n" 321 "#include \"ntp_scanner.h\"\n" 322 "#include \"ntp_parser.h\"\n" 323 "\n"; 324 325 time(&now); 326 if (!strftime(timestamp, sizeof(timestamp), 327 "%Y-%m-%d %H:%M:%S", gmtime(&now))) 328 timestamp[0] = '\0'; 329 330 printf(preamble, timestamp); 331 } 332 333 334 static void 335 generate_fsm(void) 336 { 337 char rprefix[MAX_TOK_LEN + 1]; 338 char prefix[MAX_TOK_LEN + 1]; 339 char token_id_comment[16 + MAX_TOK_LEN + 1]; 340 size_t prefix_len; 341 char *p; 342 char *r; 343 u_short initial_state; 344 u_short this_state; 345 u_short state; 346 u_short i; 347 u_short token; 348 349 /* 350 * Sort ntp_keywords in alphabetical keyword order. This is 351 * not necessary, but minimizes nonfunctional changes in the 352 * generated finite state machine when keywords are modified. 353 */ 354 qsort(ntp_keywords, COUNTOF(ntp_keywords), 355 sizeof(ntp_keywords[0]), compare_key_tok_text); 356 357 /* 358 * To save space, reserve the state array entry matching each 359 * token number for its terminal state, so the token identifier 360 * does not need to be stored in each state, but can be 361 * recovered trivially. To mark the entry reserved, 362 * finishes_token is nonzero. 363 */ 364 365 for (i = 0; i < COUNTOF(ntp_keywords); i++) { 366 token = ntp_keywords[i].token; 367 if (1 > token || token >= COUNTOF(sst)) { 368 fprintf(stderr, 369 "keyword-gen sst[%u] too small " 370 "for keyword '%s' id %d\n", 371 (int)COUNTOF(sst), 372 ntp_keywords[i].key, 373 token); 374 exit(4); 375 } 376 sst[token].finishes_token = token; 377 } 378 379 initial_state = create_keyword_scanner(); 380 381 fprintf(stderr, 382 "%d keywords consumed %d states of %d max.\n", 383 (int)COUNTOF(ntp_keywords), 384 sst_highwater - 1, 385 (int)COUNTOF(sst) - 1); 386 387 printf("#define SCANNER_INIT_S %d\n\n", initial_state); 388 389 printf("const scan_state sst[%d] = {\n" 390 "/*SS_T( ch,\tf-by, match, other ),\t\t\t\t */\n" 391 " 0,\t\t\t\t /* %5d %-17s */\n", 392 sst_highwater, 393 0, ""); 394 395 for (i = 1; i < sst_highwater; i++) { 396 397 /* verify fields will fit */ 398 if (sst[i].followedby & ~0x3) { 399 fprintf(stderr, 400 "keyword-gen internal error " 401 "sst[%d].followedby %d too big\n", 402 i, sst[i].followedby); 403 exit(7); 404 } 405 406 if (sst_highwater <= sst[i].match_next_s 407 || sst[i].match_next_s & ~0x7ff) { 408 fprintf(stderr, 409 "keyword-gen internal error " 410 "sst[%d].match_next_s %d too big\n", 411 i, sst[i].match_next_s); 412 exit(8); 413 } 414 415 if (sst_highwater <= sst[i].other_next_s 416 || sst[i].other_next_s & ~0x7ff) { 417 fprintf(stderr, 418 "keyword-gen internal error " 419 "sst[%d].other_next_s %d too big\n", 420 i, sst[i].other_next_s); 421 exit(9); 422 } 423 424 if (sst[i].finishes_token) { 425 snprintf(token_id_comment, 426 sizeof(token_id_comment), "%5d %-17s", 427 i, symbname(sst[i].finishes_token)); 428 if (i != sst[i].finishes_token) { 429 fprintf(stderr, 430 "keyword-gen internal error " 431 "entry %d finishes token %d\n", 432 i, sst[i].finishes_token); 433 exit(5); 434 } 435 } else { 436 /* 437 * Determine the keyword prefix that leads to this 438 * state. This is expensive but keyword-gen is run 439 * only when it changes. Distributing keyword-gen-utd 440 * achieves that, which is why it must be committed 441 * at the same time as keyword-gen.c and ntp_keyword.h. 442 * 443 * Scan the state array iteratively looking for a state 444 * which leads to the current one, collecting matching 445 * characters along the way. There is only one such 446 * path back to the starting state given the way our 447 * scanner state machine is built and the practice of 448 * using the spelling of the keyword as its T_* token 449 * identifier, which results in never having two 450 * spellings result in the same T_* value. 451 */ 452 prefix_len = 0; 453 this_state = i; 454 do { 455 for (state = 1; state < sst_highwater; state++) 456 if (sst[state].other_next_s == this_state) { 457 this_state = state; 458 break; 459 } else if (sst[state].match_next_s == this_state) { 460 this_state = state; 461 rprefix[prefix_len] = sst[state].ch; 462 prefix_len++; 463 break; 464 } 465 } while (this_state != initial_state); 466 467 if (prefix_len) { 468 /* reverse rprefix into prefix */ 469 p = prefix + prefix_len; 470 r = rprefix; 471 while (r < rprefix + prefix_len) 472 *--p = *r++; 473 } 474 prefix[prefix_len] = '\0'; 475 476 snprintf(token_id_comment, 477 sizeof(token_id_comment), "%5d %-17s", 478 i, (initial_state == i) 479 ? "[initial state]" 480 : prefix); 481 } 482 483 printf(" S_ST( '%c',\t%d, %5u, %5u )%s /* %s */\n", 484 sst[i].ch, 485 sst[i].followedby, 486 sst[i].match_next_s, 487 sst[i].other_next_s, 488 (i + 1 < sst_highwater) 489 ? "," 490 : " ", 491 token_id_comment); 492 } 493 494 printf("};\n\n"); 495 } 496 497 498 /* Define a function to create the states of the scanner. This function 499 * is used by the create_keyword_scanner function below. 500 * 501 * This function takes a suffix of a keyword, the token to be returned on 502 * recognizing the complete keyword, and any pre-existing state that exists 503 * for some other keyword that has the same prefix as the current one. 504 */ 505 static u_short 506 create_scan_states( 507 char * text, 508 u_short token, 509 follby followedby, 510 u_short prev_state 511 ) 512 { 513 u_short my_state; 514 u_short return_state; 515 u_short prev_char_s; 516 u_short curr_char_s; 517 518 return_state = prev_state; 519 curr_char_s = prev_state; 520 prev_char_s = 0; 521 522 /* Find the correct position to insert the state. 523 * All states should be in alphabetical order 524 */ 525 while (curr_char_s && (text[0] < sst[curr_char_s].ch)) { 526 prev_char_s = curr_char_s; 527 curr_char_s = sst[curr_char_s].other_next_s; 528 } 529 530 /* 531 * Check if a previously seen keyword has the same prefix as 532 * the current keyword. If so, simply use the state for that 533 * keyword as my_state, otherwise, allocate a new state. 534 */ 535 if (curr_char_s && (text[0] == sst[curr_char_s].ch)) { 536 my_state = curr_char_s; 537 if ('\0' == text[1]) { 538 fprintf(stderr, 539 "Duplicate entries for keyword '%s' in" 540 " keyword_gen.c ntp_keywords[].\n", 541 current_keyword); 542 exit(2); 543 } 544 } else { 545 do 546 my_state = sst_highwater++; 547 while (my_state < COUNTOF(sst) 548 && sst[my_state].finishes_token); 549 if (my_state >= COUNTOF(sst)) { 550 fprintf(stderr, 551 "fatal, keyword scanner state array " 552 "sst[%d] is too small, modify\n" 553 "keyword-gen.c to increase.\n", 554 (int)COUNTOF(sst)); 555 exit(3); 556 } 557 /* Store the next character of the keyword */ 558 sst[my_state].ch = text[0]; 559 sst[my_state].other_next_s = curr_char_s; 560 sst[my_state].followedby = FOLLBY_NON_ACCEPTING; 561 562 if (prev_char_s) 563 sst[prev_char_s].other_next_s = my_state; 564 else 565 return_state = my_state; 566 } 567 568 /* Check if the next character is '\0'. 569 * If yes, we are done with the recognition and this is an accepting 570 * state. 571 * If not, we need to continue scanning 572 */ 573 if ('\0' == text[1]) { 574 sst[my_state].finishes_token = (u_short)token; 575 sst[my_state].followedby = (char)followedby; 576 577 if (sst[token].finishes_token != (u_short)token) { 578 fprintf(stderr, 579 "fatal, sst[%d] not reserved for %s.\n", 580 token, symbname(token)); 581 exit(6); 582 } 583 /* relocate so token id is sst[] index */ 584 if (my_state != token) { 585 sst[token] = sst[my_state]; 586 ZERO(sst[my_state]); 587 do 588 sst_highwater--; 589 while (sst[sst_highwater].finishes_token); 590 my_state = token; 591 if (prev_char_s) 592 sst[prev_char_s].other_next_s = my_state; 593 else 594 return_state = my_state; 595 } 596 } else 597 sst[my_state].match_next_s = 598 create_scan_states( 599 &text[1], 600 token, 601 followedby, 602 sst[my_state].match_next_s); 603 604 return return_state; 605 } 606 607 608 /* Define a function that takes a list of (keyword, token) values and 609 * creates a keywords scanner out of it. 610 */ 611 612 static u_short 613 create_keyword_scanner(void) 614 { 615 u_short scanner; 616 u_short i; 617 618 sst_highwater = 1; /* index 0 invalid, unused */ 619 scanner = 0; 620 621 for (i = 0; i < COUNTOF(ntp_keywords); i++) { 622 current_keyword = ntp_keywords[i].key; 623 scanner = 624 create_scan_states( 625 ntp_keywords[i].key, 626 ntp_keywords[i].token, 627 ntp_keywords[i].followedby, 628 scanner); 629 } 630 631 return scanner; 632 } 633 634 635 static void 636 generate_token_text(void) 637 { 638 u_short lowest_id; 639 u_short highest_id; 640 u_short id_count; 641 u_short id; 642 u_short i; 643 644 /* sort ntp_keywords in token ID order */ 645 qsort(ntp_keywords, COUNTOF(ntp_keywords), 646 sizeof(ntp_keywords[0]), compare_key_tok_id); 647 648 lowest_id = ntp_keywords[0].token; 649 highest_id = ntp_keywords[COUNTOF(ntp_keywords) - 1].token; 650 id_count = highest_id - lowest_id + 1; 651 652 printf("#define LOWEST_KEYWORD_ID %d\n\n", lowest_id); 653 654 printf("const char * const keyword_text[%d] = {", id_count); 655 656 id = lowest_id; 657 i = 0; 658 while (i < COUNTOF(ntp_keywords)) { 659 while (id < ntp_keywords[i].token) { 660 printf(",\n\t/* %-5d %5d %20s */\tNULL", 661 id - lowest_id, id, symbname(id)); 662 id++; 663 } 664 if (i > 0) 665 printf(","); 666 printf("\n\t/* %-5d %5d %20s */\t\"%s\"", 667 id - lowest_id, id, symbname(id), 668 ntp_keywords[i].key); 669 i++; 670 id++; 671 } 672 673 printf("\n};\n\n"); 674 } 675 676 677 int 678 compare_key_tok_id( 679 const void *a1, 680 const void *a2 681 ) 682 { 683 const struct key_tok *p1 = a1; 684 const struct key_tok *p2 = a2; 685 686 if (p1->token == p2->token) 687 return 0; 688 689 if (p1->token < p2->token) 690 return -1; 691 else 692 return 1; 693 } 694 695 696 int 697 compare_key_tok_text( 698 const void *a1, 699 const void *a2 700 ) 701 { 702 const struct key_tok *p1 = a1; 703 const struct key_tok *p2 = a2; 704 705 return strcmp(p1->key, p2->key); 706 } 707 708 709 /* 710 * populate_symb() - populate symb[] lookup array with symbolic token 711 * names such that symb[T_Age] == "T_Age", etc. 712 */ 713 void 714 populate_symb( 715 char *header_file 716 ) 717 { 718 FILE * yh; 719 char line[2 * MAX_TOK_LEN]; 720 char name[2 * MAX_TOK_LEN]; 721 int token; 722 723 yh = fopen(header_file, "r"); 724 if (NULL == yh) { 725 perror("unable to open yacc/bison header file"); 726 exit(4); 727 } 728 729 while (NULL != fgets(line, sizeof(line), yh)) 730 if (2 == sscanf(line, "#define %s %d", name, &token) 731 && 'T' == name[0] && '_' == name[1] && token >= 0 732 && token < COUNTOF(symb)) { 733 734 symb[token] = estrdup(name); 735 if (strlen(name) > MAX_TOK_LEN) { 736 fprintf(stderr, 737 "MAX_TOK_LEN %d too small for '%s'\n" 738 "Edit keyword-gen.c to raise.\n", 739 MAX_TOK_LEN, name); 740 exit(10); 741 } 742 } 743 fclose(yh); 744 } 745 746 747 const char * 748 symbname( 749 u_short token 750 ) 751 { 752 char *name; 753 754 if (token < COUNTOF(symb) && symb[token] != NULL) { 755 name = symb[token]; 756 } else { 757 LIB_GETBUF(name); 758 snprintf(name, LIB_BUFLENGTH, "%d", token); 759 } 760 761 return name; 762 } 763
Indexes created Wed May 06 00:24:28 UTC 2026