1 /* $Id: man_validate.c,v 1.53 2012/06/02 20:07:09 schwarze Exp $ */ 2 /* 3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> 4 * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org> 5 * 6 * Permission to use, copy, modify, and distribute this software for any 7 * purpose with or without fee is hereby granted, provided that the above 8 * copyright notice and this permission notice appear in all copies. 9 * 10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES 11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF 12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR 13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN 15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF 16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. 17 */ 18 #include <sys/types.h> 19 20 #include <assert.h> 21 #include <ctype.h> 22 #include <errno.h> 23 #include <limits.h> 24 #include <stdarg.h> 25 #include <stdlib.h> 26 #include <string.h> 27 #include <time.h> 28 29 #include "man.h" 30 #include "mandoc.h" 31 #include "libman.h" 32 #include "libmandoc.h" 33 34 #define CHKARGS struct man *m, struct man_node *n 35 36 typedef int (*v_check)(CHKARGS); 37 38 struct man_valid { 39 v_check *pres; 40 v_check *posts; 41 }; 42 43 static int check_eq0(CHKARGS); 44 static int check_eq2(CHKARGS); 45 static int check_le1(CHKARGS); 46 static int check_ge2(CHKARGS); 47 static int check_le5(CHKARGS); 48 static int check_par(CHKARGS); 49 static int check_part(CHKARGS); 50 static int check_root(CHKARGS); 51 static void check_text(CHKARGS); 52 53 static int post_AT(CHKARGS); 54 static int post_vs(CHKARGS); 55 static int post_fi(CHKARGS); 56 static int post_ft(CHKARGS); 57 static int post_nf(CHKARGS); 58 static int post_sec(CHKARGS); 59 static int post_TH(CHKARGS); 60 static int post_UC(CHKARGS); 61 static int pre_sec(CHKARGS); 62 63 static v_check posts_at[] = { post_AT, NULL }; 64 static v_check posts_br[] = { post_vs, check_eq0, NULL }; 65 static v_check posts_eq0[] = { check_eq0, NULL }; 66 static v_check posts_eq2[] = { check_eq2, NULL }; 67 static v_check posts_fi[] = { check_eq0, post_fi, NULL }; 68 static v_check posts_ft[] = { post_ft, NULL }; 69 static v_check posts_nf[] = { check_eq0, post_nf, NULL }; 70 static v_check posts_par[] = { check_par, NULL }; 71 static v_check posts_part[] = { check_part, NULL }; 72 static v_check posts_sec[] = { post_sec, NULL }; 73 static v_check posts_sp[] = { post_vs, check_le1, NULL }; 74 static v_check posts_th[] = { check_ge2, check_le5, post_TH, NULL }; 75 static v_check posts_uc[] = { post_UC, NULL }; 76 static v_check pres_sec[] = { pre_sec, NULL }; 77 78 static const struct man_valid man_valids[MAN_MAX] = { 79 { NULL, posts_br }, /* br */ 80 { NULL, posts_th }, /* TH */ 81 { pres_sec, posts_sec }, /* SH */ 82 { pres_sec, posts_sec }, /* SS */ 83 { NULL, NULL }, /* TP */ 84 { NULL, posts_par }, /* LP */ 85 { NULL, posts_par }, /* PP */ 86 { NULL, posts_par }, /* P */ 87 { NULL, NULL }, /* IP */ 88 { NULL, NULL }, /* HP */ 89 { NULL, NULL }, /* SM */ 90 { NULL, NULL }, /* SB */ 91 { NULL, NULL }, /* BI */ 92 { NULL, NULL }, /* IB */ 93 { NULL, NULL }, /* BR */ 94 { NULL, NULL }, /* RB */ 95 { NULL, NULL }, /* R */ 96 { NULL, NULL }, /* B */ 97 { NULL, NULL }, /* I */ 98 { NULL, NULL }, /* IR */ 99 { NULL, NULL }, /* RI */ 100 { NULL, posts_eq0 }, /* na */ 101 { NULL, posts_sp }, /* sp */ 102 { NULL, posts_nf }, /* nf */ 103 { NULL, posts_fi }, /* fi */ 104 { NULL, NULL }, /* RE */ 105 { NULL, posts_part }, /* RS */ 106 { NULL, NULL }, /* DT */ 107 { NULL, posts_uc }, /* UC */ 108 { NULL, NULL }, /* PD */ 109 { NULL, posts_at }, /* AT */ 110 { NULL, NULL }, /* in */ 111 { NULL, posts_ft }, /* ft */ 112 { NULL, posts_eq2 }, /* OP */ 113 { NULL, posts_nf }, /* EX */ 114 { NULL, posts_fi }, /* EE */ 115 }; 116 117 118 int 119 man_valid_pre(struct man *m, struct man_node *n) 120 { 121 v_check *cp; 122 123 switch (n->type) { 124 case (MAN_TEXT): 125 /* FALLTHROUGH */ 126 case (MAN_ROOT): 127 /* FALLTHROUGH */ 128 case (MAN_EQN): 129 /* FALLTHROUGH */ 130 case (MAN_TBL): 131 return(1); 132 default: 133 break; 134 } 135 136 if (NULL == (cp = man_valids[n->tok].pres)) 137 return(1); 138 for ( ; *cp; cp++) 139 if ( ! (*cp)(m, n)) 140 return(0); 141 return(1); 142 } 143 144 145 int 146 man_valid_post(struct man *m) 147 { 148 v_check *cp; 149 150 if (MAN_VALID & m->last->flags) 151 return(1); 152 m->last->flags |= MAN_VALID; 153 154 switch (m->last->type) { 155 case (MAN_TEXT): 156 check_text(m, m->last); 157 return(1); 158 case (MAN_ROOT): 159 return(check_root(m, m->last)); 160 case (MAN_EQN): 161 /* FALLTHROUGH */ 162 case (MAN_TBL): 163 return(1); 164 default: 165 break; 166 } 167 168 if (NULL == (cp = man_valids[m->last->tok].posts)) 169 return(1); 170 for ( ; *cp; cp++) 171 if ( ! (*cp)(m, m->last)) 172 return(0); 173 174 return(1); 175 } 176 177 178 static int 179 check_root(CHKARGS) 180 { 181 182 if (MAN_BLINE & m->flags) 183 man_nmsg(m, n, MANDOCERR_SCOPEEXIT); 184 else if (MAN_ELINE & m->flags) 185 man_nmsg(m, n, MANDOCERR_SCOPEEXIT); 186 187 m->flags &= ~MAN_BLINE; 188 m->flags &= ~MAN_ELINE; 189 190 if (NULL == m->first->child) { 191 man_nmsg(m, n, MANDOCERR_NODOCBODY); 192 return(0); 193 } else if (NULL == m->meta.title) { 194 man_nmsg(m, n, MANDOCERR_NOTITLE); 195 196 /* 197 * If a title hasn't been set, do so now (by 198 * implication, date and section also aren't set). 199 */ 200 201 m->meta.title = mandoc_strdup("unknown"); 202 m->meta.msec = mandoc_strdup("1"); 203 m->meta.date = mandoc_normdate 204 (m->parse, NULL, n->line, n->pos); 205 } 206 207 return(1); 208 } 209 210 static void 211 check_text(CHKARGS) 212 { 213 char *cp, *p; 214 215 if (MAN_LITERAL & m->flags) 216 return; 217 218 cp = n->string; 219 for (p = cp; NULL != (p = strchr(p, '\t')); p++) 220 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB); 221 } 222 223 #define INEQ_DEFINE(x, ineq, name) \ 224 static int \ 225 check_##name(CHKARGS) \ 226 { \ 227 if (n->nchild ineq (x)) \ 228 return(1); \ 229 mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \ 230 "line arguments %s %d (have %d)", \ 231 #ineq, (x), n->nchild); \ 232 return(1); \ 233 } 234 235 INEQ_DEFINE(0, ==, eq0) 236 INEQ_DEFINE(2, ==, eq2) 237 INEQ_DEFINE(1, <=, le1) 238 INEQ_DEFINE(2, >=, ge2) 239 INEQ_DEFINE(5, <=, le5) 240 241 static int 242 post_ft(CHKARGS) 243 { 244 char *cp; 245 int ok; 246 247 if (0 == n->nchild) 248 return(1); 249 250 ok = 0; 251 cp = n->child->string; 252 switch (*cp) { 253 case ('1'): 254 /* FALLTHROUGH */ 255 case ('2'): 256 /* FALLTHROUGH */ 257 case ('3'): 258 /* FALLTHROUGH */ 259 case ('4'): 260 /* FALLTHROUGH */ 261 case ('I'): 262 /* FALLTHROUGH */ 263 case ('P'): 264 /* FALLTHROUGH */ 265 case ('R'): 266 if ('\0' == cp[1]) 267 ok = 1; 268 break; 269 case ('B'): 270 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2])) 271 ok = 1; 272 break; 273 case ('C'): 274 if ('W' == cp[1] && '\0' == cp[2]) 275 ok = 1; 276 break; 277 default: 278 break; 279 } 280 281 if (0 == ok) { 282 mandoc_vmsg 283 (MANDOCERR_BADFONT, m->parse, 284 n->line, n->pos, "%s", cp); 285 *cp = '\0'; 286 } 287 288 if (1 < n->nchild) 289 mandoc_vmsg 290 (MANDOCERR_ARGCOUNT, m->parse, n->line, 291 n->pos, "want one child (have %d)", 292 n->nchild); 293 294 return(1); 295 } 296 297 static int 298 pre_sec(CHKARGS) 299 { 300 301 if (MAN_BLOCK == n->type) 302 m->flags &= ~MAN_LITERAL; 303 return(1); 304 } 305 306 static int 307 post_sec(CHKARGS) 308 { 309 310 if ( ! (MAN_HEAD == n->type && 0 == n->nchild)) 311 return(1); 312 313 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT); 314 return(0); 315 } 316 317 static int 318 check_part(CHKARGS) 319 { 320 321 if (MAN_BODY == n->type && 0 == n->nchild) 322 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line, 323 n->pos, "want children (have none)"); 324 325 return(1); 326 } 327 328 329 static int 330 check_par(CHKARGS) 331 { 332 333 switch (n->type) { 334 case (MAN_BLOCK): 335 if (0 == n->body->nchild) 336 man_node_delete(m, n); 337 break; 338 case (MAN_BODY): 339 if (0 == n->nchild) 340 man_nmsg(m, n, MANDOCERR_IGNPAR); 341 break; 342 case (MAN_HEAD): 343 if (n->nchild) 344 man_nmsg(m, n, MANDOCERR_ARGSLOST); 345 break; 346 default: 347 break; 348 } 349 350 return(1); 351 } 352 353 354 static int 355 post_TH(CHKARGS) 356 { 357 const char *p; 358 int line, pos; 359 360 if (m->meta.title) 361 free(m->meta.title); 362 if (m->meta.vol) 363 free(m->meta.vol); 364 if (m->meta.source) 365 free(m->meta.source); 366 if (m->meta.msec) 367 free(m->meta.msec); 368 if (m->meta.date) 369 free(m->meta.date); 370 371 line = n->line; 372 pos = n->pos; 373 m->meta.title = m->meta.vol = m->meta.date = 374 m->meta.msec = m->meta.source = NULL; 375 376 /* ->TITLE<- MSEC DATE SOURCE VOL */ 377 378 n = n->child; 379 if (n && n->string) { 380 for (p = n->string; '\0' != *p; p++) { 381 /* Only warn about this once... */ 382 if (isalpha((unsigned char)*p) && 383 ! isupper((unsigned char)*p)) { 384 man_nmsg(m, n, MANDOCERR_UPPERCASE); 385 break; 386 } 387 } 388 m->meta.title = mandoc_strdup(n->string); 389 } else 390 m->meta.title = mandoc_strdup(""); 391 392 /* TITLE ->MSEC<- DATE SOURCE VOL */ 393 394 if (n) 395 n = n->next; 396 if (n && n->string) 397 m->meta.msec = mandoc_strdup(n->string); 398 else 399 m->meta.msec = mandoc_strdup(""); 400 401 /* TITLE MSEC ->DATE<- SOURCE VOL */ 402 403 if (n) 404 n = n->next; 405 if (n && n->string && '\0' != n->string[0]) { 406 pos = n->pos; 407 m->meta.date = mandoc_normdate 408 (m->parse, n->string, line, pos); 409 } else 410 m->meta.date = mandoc_strdup(""); 411 412 /* TITLE MSEC DATE ->SOURCE<- VOL */ 413 414 if (n && (n = n->next)) 415 m->meta.source = mandoc_strdup(n->string); 416 417 /* TITLE MSEC DATE SOURCE ->VOL<- */ 418 /* If missing, use the default VOL name for MSEC. */ 419 420 if (n && (n = n->next)) 421 m->meta.vol = mandoc_strdup(n->string); 422 else if ('\0' != m->meta.msec[0] && 423 (NULL != (p = mandoc_a2msec(m->meta.msec)))) 424 m->meta.vol = mandoc_strdup(p); 425 426 /* 427 * Remove the `TH' node after we've processed it for our 428 * meta-data. 429 */ 430 man_node_delete(m, m->last); 431 return(1); 432 } 433 434 static int 435 post_nf(CHKARGS) 436 { 437 438 if (MAN_LITERAL & m->flags) 439 man_nmsg(m, n, MANDOCERR_SCOPEREP); 440 441 m->flags |= MAN_LITERAL; 442 return(1); 443 } 444 445 static int 446 post_fi(CHKARGS) 447 { 448 449 if ( ! (MAN_LITERAL & m->flags)) 450 man_nmsg(m, n, MANDOCERR_WNOSCOPE); 451 452 m->flags &= ~MAN_LITERAL; 453 return(1); 454 } 455 456 static int 457 post_UC(CHKARGS) 458 { 459 static const char * const bsd_versions[] = { 460 "3rd Berkeley Distribution", 461 "4th Berkeley Distribution", 462 "4.2 Berkeley Distribution", 463 "4.3 Berkeley Distribution", 464 "4.4 Berkeley Distribution", 465 }; 466 467 const char *p, *s; 468 469 n = n->child; 470 471 if (NULL == n || MAN_TEXT != n->type) 472 p = bsd_versions[0]; 473 else { 474 s = n->string; 475 if (0 == strcmp(s, "3")) 476 p = bsd_versions[0]; 477 else if (0 == strcmp(s, "4")) 478 p = bsd_versions[1]; 479 else if (0 == strcmp(s, "5")) 480 p = bsd_versions[2]; 481 else if (0 == strcmp(s, "6")) 482 p = bsd_versions[3]; 483 else if (0 == strcmp(s, "7")) 484 p = bsd_versions[4]; 485 else 486 p = bsd_versions[0]; 487 } 488 489 if (m->meta.source) 490 free(m->meta.source); 491 492 m->meta.source = mandoc_strdup(p); 493 return(1); 494 } 495 496 static int 497 post_AT(CHKARGS) 498 { 499 static const char * const unix_versions[] = { 500 "7th Edition", 501 "System III", 502 "System V", 503 "System V Release 2", 504 }; 505 506 const char *p, *s; 507 struct man_node *nn; 508 509 n = n->child; 510 511 if (NULL == n || MAN_TEXT != n->type) 512 p = unix_versions[0]; 513 else { 514 s = n->string; 515 if (0 == strcmp(s, "3")) 516 p = unix_versions[0]; 517 else if (0 == strcmp(s, "4")) 518 p = unix_versions[1]; 519 else if (0 == strcmp(s, "5")) { 520 nn = n->next; 521 if (nn && MAN_TEXT == nn->type && nn->string[0]) 522 p = unix_versions[3]; 523 else 524 p = unix_versions[2]; 525 } else 526 p = unix_versions[0]; 527 } 528 529 if (m->meta.source) 530 free(m->meta.source); 531 532 m->meta.source = mandoc_strdup(p); 533 return(1); 534 } 535 536 static int 537 post_vs(CHKARGS) 538 { 539 540 /* 541 * Don't warn about this because it occurs in pod2man and would 542 * cause considerable (unfixable) warnage. 543 */ 544 if (NULL == n->prev && MAN_ROOT == n->parent->type) 545 man_node_delete(m, n); 546 547 return(1); 548 } 549