1 /* $Id: man_validate.c,v 1.52 2012/02/26 19:41:27 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 }; 114 115 116 int 117 man_valid_pre(struct man *m, struct man_node *n) 118 { 119 v_check *cp; 120 121 switch (n->type) { 122 case (MAN_TEXT): 123 /* FALLTHROUGH */ 124 case (MAN_ROOT): 125 /* FALLTHROUGH */ 126 case (MAN_EQN): 127 /* FALLTHROUGH */ 128 case (MAN_TBL): 129 return(1); 130 default: 131 break; 132 } 133 134 if (NULL == (cp = man_valids[n->tok].pres)) 135 return(1); 136 for ( ; *cp; cp++) 137 if ( ! (*cp)(m, n)) 138 return(0); 139 return(1); 140 } 141 142 143 int 144 man_valid_post(struct man *m) 145 { 146 v_check *cp; 147 148 if (MAN_VALID & m->last->flags) 149 return(1); 150 m->last->flags |= MAN_VALID; 151 152 switch (m->last->type) { 153 case (MAN_TEXT): 154 check_text(m, m->last); 155 return(1); 156 case (MAN_ROOT): 157 return(check_root(m, m->last)); 158 case (MAN_EQN): 159 /* FALLTHROUGH */ 160 case (MAN_TBL): 161 return(1); 162 default: 163 break; 164 } 165 166 if (NULL == (cp = man_valids[m->last->tok].posts)) 167 return(1); 168 for ( ; *cp; cp++) 169 if ( ! (*cp)(m, m->last)) 170 return(0); 171 172 return(1); 173 } 174 175 176 static int 177 check_root(CHKARGS) 178 { 179 180 if (MAN_BLINE & m->flags) 181 man_nmsg(m, n, MANDOCERR_SCOPEEXIT); 182 else if (MAN_ELINE & m->flags) 183 man_nmsg(m, n, MANDOCERR_SCOPEEXIT); 184 185 m->flags &= ~MAN_BLINE; 186 m->flags &= ~MAN_ELINE; 187 188 if (NULL == m->first->child) { 189 man_nmsg(m, n, MANDOCERR_NODOCBODY); 190 return(0); 191 } else if (NULL == m->meta.title) { 192 man_nmsg(m, n, MANDOCERR_NOTITLE); 193 194 /* 195 * If a title hasn't been set, do so now (by 196 * implication, date and section also aren't set). 197 */ 198 199 m->meta.title = mandoc_strdup("unknown"); 200 m->meta.msec = mandoc_strdup("1"); 201 m->meta.date = mandoc_normdate 202 (m->parse, NULL, n->line, n->pos); 203 } 204 205 return(1); 206 } 207 208 static void 209 check_text(CHKARGS) 210 { 211 char *cp, *p; 212 213 if (MAN_LITERAL & m->flags) 214 return; 215 216 cp = n->string; 217 for (p = cp; NULL != (p = strchr(p, '\t')); p++) 218 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB); 219 } 220 221 #define INEQ_DEFINE(x, ineq, name) \ 222 static int \ 223 check_##name(CHKARGS) \ 224 { \ 225 if (n->nchild ineq (x)) \ 226 return(1); \ 227 mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \ 228 "line arguments %s %d (have %d)", \ 229 #ineq, (x), n->nchild); \ 230 return(1); \ 231 } 232 233 INEQ_DEFINE(0, ==, eq0) 234 INEQ_DEFINE(2, ==, eq2) 235 INEQ_DEFINE(1, <=, le1) 236 INEQ_DEFINE(2, >=, ge2) 237 INEQ_DEFINE(5, <=, le5) 238 239 static int 240 post_ft(CHKARGS) 241 { 242 char *cp; 243 int ok; 244 245 if (0 == n->nchild) 246 return(1); 247 248 ok = 0; 249 cp = n->child->string; 250 switch (*cp) { 251 case ('1'): 252 /* FALLTHROUGH */ 253 case ('2'): 254 /* FALLTHROUGH */ 255 case ('3'): 256 /* FALLTHROUGH */ 257 case ('4'): 258 /* FALLTHROUGH */ 259 case ('I'): 260 /* FALLTHROUGH */ 261 case ('P'): 262 /* FALLTHROUGH */ 263 case ('R'): 264 if ('\0' == cp[1]) 265 ok = 1; 266 break; 267 case ('B'): 268 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2])) 269 ok = 1; 270 break; 271 case ('C'): 272 if ('W' == cp[1] && '\0' == cp[2]) 273 ok = 1; 274 break; 275 default: 276 break; 277 } 278 279 if (0 == ok) { 280 mandoc_vmsg 281 (MANDOCERR_BADFONT, m->parse, 282 n->line, n->pos, "%s", cp); 283 *cp = '\0'; 284 } 285 286 if (1 < n->nchild) 287 mandoc_vmsg 288 (MANDOCERR_ARGCOUNT, m->parse, n->line, 289 n->pos, "want one child (have %d)", 290 n->nchild); 291 292 return(1); 293 } 294 295 static int 296 pre_sec(CHKARGS) 297 { 298 299 if (MAN_BLOCK == n->type) 300 m->flags &= ~MAN_LITERAL; 301 return(1); 302 } 303 304 static int 305 post_sec(CHKARGS) 306 { 307 308 if ( ! (MAN_HEAD == n->type && 0 == n->nchild)) 309 return(1); 310 311 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT); 312 return(0); 313 } 314 315 static int 316 check_part(CHKARGS) 317 { 318 319 if (MAN_BODY == n->type && 0 == n->nchild) 320 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line, 321 n->pos, "want children (have none)"); 322 323 return(1); 324 } 325 326 327 static int 328 check_par(CHKARGS) 329 { 330 331 switch (n->type) { 332 case (MAN_BLOCK): 333 if (0 == n->body->nchild) 334 man_node_delete(m, n); 335 break; 336 case (MAN_BODY): 337 if (0 == n->nchild) 338 man_nmsg(m, n, MANDOCERR_IGNPAR); 339 break; 340 case (MAN_HEAD): 341 if (n->nchild) 342 man_nmsg(m, n, MANDOCERR_ARGSLOST); 343 break; 344 default: 345 break; 346 } 347 348 return(1); 349 } 350 351 352 static int 353 post_TH(CHKARGS) 354 { 355 const char *p; 356 int line, pos; 357 358 if (m->meta.title) 359 free(m->meta.title); 360 if (m->meta.vol) 361 free(m->meta.vol); 362 if (m->meta.source) 363 free(m->meta.source); 364 if (m->meta.msec) 365 free(m->meta.msec); 366 if (m->meta.date) 367 free(m->meta.date); 368 369 line = n->line; 370 pos = n->pos; 371 m->meta.title = m->meta.vol = m->meta.date = 372 m->meta.msec = m->meta.source = NULL; 373 374 /* ->TITLE<- MSEC DATE SOURCE VOL */ 375 376 n = n->child; 377 if (n && n->string) { 378 for (p = n->string; '\0' != *p; p++) { 379 /* Only warn about this once... */ 380 if (isalpha((unsigned char)*p) && 381 ! isupper((unsigned char)*p)) { 382 man_nmsg(m, n, MANDOCERR_UPPERCASE); 383 break; 384 } 385 } 386 m->meta.title = mandoc_strdup(n->string); 387 } else 388 m->meta.title = mandoc_strdup(""); 389 390 /* TITLE ->MSEC<- DATE SOURCE VOL */ 391 392 if (n) 393 n = n->next; 394 if (n && n->string) 395 m->meta.msec = mandoc_strdup(n->string); 396 else 397 m->meta.msec = mandoc_strdup(""); 398 399 /* TITLE MSEC ->DATE<- SOURCE VOL */ 400 401 if (n) 402 n = n->next; 403 if (n && n->string && '\0' != n->string[0]) { 404 pos = n->pos; 405 m->meta.date = mandoc_normdate 406 (m->parse, n->string, line, pos); 407 } else 408 m->meta.date = mandoc_strdup(""); 409 410 /* TITLE MSEC DATE ->SOURCE<- VOL */ 411 412 if (n && (n = n->next)) 413 m->meta.source = mandoc_strdup(n->string); 414 415 /* TITLE MSEC DATE SOURCE ->VOL<- */ 416 /* If missing, use the default VOL name for MSEC. */ 417 418 if (n && (n = n->next)) 419 m->meta.vol = mandoc_strdup(n->string); 420 else if ('\0' != m->meta.msec[0] && 421 (NULL != (p = mandoc_a2msec(m->meta.msec)))) 422 m->meta.vol = mandoc_strdup(p); 423 424 /* 425 * Remove the `TH' node after we've processed it for our 426 * meta-data. 427 */ 428 man_node_delete(m, m->last); 429 return(1); 430 } 431 432 static int 433 post_nf(CHKARGS) 434 { 435 436 if (MAN_LITERAL & m->flags) 437 man_nmsg(m, n, MANDOCERR_SCOPEREP); 438 439 m->flags |= MAN_LITERAL; 440 return(1); 441 } 442 443 static int 444 post_fi(CHKARGS) 445 { 446 447 if ( ! (MAN_LITERAL & m->flags)) 448 man_nmsg(m, n, MANDOCERR_WNOSCOPE); 449 450 m->flags &= ~MAN_LITERAL; 451 return(1); 452 } 453 454 static int 455 post_UC(CHKARGS) 456 { 457 static const char * const bsd_versions[] = { 458 "3rd Berkeley Distribution", 459 "4th Berkeley Distribution", 460 "4.2 Berkeley Distribution", 461 "4.3 Berkeley Distribution", 462 "4.4 Berkeley Distribution", 463 }; 464 465 const char *p, *s; 466 467 n = n->child; 468 469 if (NULL == n || MAN_TEXT != n->type) 470 p = bsd_versions[0]; 471 else { 472 s = n->string; 473 if (0 == strcmp(s, "3")) 474 p = bsd_versions[0]; 475 else if (0 == strcmp(s, "4")) 476 p = bsd_versions[1]; 477 else if (0 == strcmp(s, "5")) 478 p = bsd_versions[2]; 479 else if (0 == strcmp(s, "6")) 480 p = bsd_versions[3]; 481 else if (0 == strcmp(s, "7")) 482 p = bsd_versions[4]; 483 else 484 p = bsd_versions[0]; 485 } 486 487 if (m->meta.source) 488 free(m->meta.source); 489 490 m->meta.source = mandoc_strdup(p); 491 return(1); 492 } 493 494 static int 495 post_AT(CHKARGS) 496 { 497 static const char * const unix_versions[] = { 498 "7th Edition", 499 "System III", 500 "System V", 501 "System V Release 2", 502 }; 503 504 const char *p, *s; 505 struct man_node *nn; 506 507 n = n->child; 508 509 if (NULL == n || MAN_TEXT != n->type) 510 p = unix_versions[0]; 511 else { 512 s = n->string; 513 if (0 == strcmp(s, "3")) 514 p = unix_versions[0]; 515 else if (0 == strcmp(s, "4")) 516 p = unix_versions[1]; 517 else if (0 == strcmp(s, "5")) { 518 nn = n->next; 519 if (nn && MAN_TEXT == nn->type && nn->string[0]) 520 p = unix_versions[3]; 521 else 522 p = unix_versions[2]; 523 } else 524 p = unix_versions[0]; 525 } 526 527 if (m->meta.source) 528 free(m->meta.source); 529 530 m->meta.source = mandoc_strdup(p); 531 return(1); 532 } 533 534 static int 535 post_vs(CHKARGS) 536 { 537 538 /* 539 * Don't warn about this because it occurs in pod2man and would 540 * cause considerable (unfixable) warnage. 541 */ 542 if (NULL == n->prev && MAN_ROOT == n->parent->type) 543 man_node_delete(m, n); 544 545 return(1); 546 } 547