man_validate.c

   1 /*      $Id: man_validate.c,v 1.94 2014/07/01 22:37:15 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012, 2013, 2014 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 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <sys/types.h>
  23
  24 #include <assert.h>
  25 #include <ctype.h>
  26 #include <errno.h>
  27 #include <limits.h>
  28 #include <stdarg.h>
  29 #include <stdlib.h>
  30 #include <string.h>
  31 #include <time.h>
  32
  33 #include "man.h"
  34 #include "mandoc.h"
  35 #include "mandoc_aux.h"
  36 #include "libman.h"
  37 #include "libmandoc.h"
  38
  39 #define CHKARGS   struct man *man, struct man_node *n
  40
  41 typedef int     (*v_check)(CHKARGS);
  42
  43 struct  man_valid {
  44         v_check  *pres;
  45         v_check  *posts;
  46 };
  47
  48 static  int       check_eq0(CHKARGS);
  49 static  int       check_eq2(CHKARGS);
  50 static  int       check_le1(CHKARGS);
  51 static  int       check_ge2(CHKARGS);
  52 static  int       check_le5(CHKARGS);
  53 static  int       check_head1(CHKARGS);
  54 static  int       check_par(CHKARGS);
  55 static  int       check_part(CHKARGS);
  56 static  int       check_root(CHKARGS);
  57 static  void      check_text(CHKARGS);
  58
  59 static  int       post_AT(CHKARGS);
  60 static  int       post_IP(CHKARGS);
  61 static  int       post_vs(CHKARGS);
  62 static  int       post_fi(CHKARGS);
  63 static  int       post_ft(CHKARGS);
  64 static  int       post_nf(CHKARGS);
  65 static  int       post_sec(CHKARGS);
  66 static  int       post_TH(CHKARGS);
  67 static  int       post_UC(CHKARGS);
  68 static  int       pre_sec(CHKARGS);
  69
  70 static  v_check   posts_at[] = { post_AT, NULL };
  71 static  v_check   posts_br[] = { post_vs, check_eq0, NULL };
  72 static  v_check   posts_eq0[] = { check_eq0, NULL };
  73 static  v_check   posts_eq2[] = { check_eq2, NULL };
  74 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  75 static  v_check   posts_ft[] = { post_ft, NULL };
  76 static  v_check   posts_ip[] = { post_IP, NULL };
  77 static  v_check   posts_le1[] = { check_le1, NULL };
  78 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  79 static  v_check   posts_par[] = { check_par, NULL };
  80 static  v_check   posts_part[] = { check_part, NULL };
  81 static  v_check   posts_sec[] = { post_sec, NULL };
  82 static  v_check   posts_sp[] = { post_vs, check_le1, NULL };
  83 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  84 static  v_check   posts_uc[] = { post_UC, NULL };
  85 static  v_check   posts_ur[] = { check_head1, check_part, NULL };
  86 static  v_check   pres_sec[] = { pre_sec, NULL };
  87
  88 static  const struct man_valid man_valids[MAN_MAX] = {
  89         { NULL, posts_br }, /* br */
  90         { NULL, posts_th }, /* TH */
  91         { pres_sec, posts_sec }, /* SH */
  92         { pres_sec, posts_sec }, /* SS */
  93         { NULL, NULL }, /* TP */
  94         { NULL, posts_par }, /* LP */
  95         { NULL, posts_par }, /* PP */
  96         { NULL, posts_par }, /* P */
  97         { NULL, posts_ip }, /* IP */
  98         { NULL, NULL }, /* HP */
  99         { NULL, NULL }, /* SM */
 100         { NULL, NULL }, /* SB */
 101         { NULL, NULL }, /* BI */
 102         { NULL, NULL }, /* IB */
 103         { NULL, NULL }, /* BR */
 104         { NULL, NULL }, /* RB */
 105         { NULL, NULL }, /* R */
 106         { NULL, NULL }, /* B */
 107         { NULL, NULL }, /* I */
 108         { NULL, NULL }, /* IR */
 109         { NULL, NULL }, /* RI */
 110         { NULL, posts_eq0 }, /* na */
 111         { NULL, posts_sp }, /* sp */
 112         { NULL, posts_nf }, /* nf */
 113         { NULL, posts_fi }, /* fi */
 114         { NULL, NULL }, /* RE */
 115         { NULL, posts_part }, /* RS */
 116         { NULL, NULL }, /* DT */
 117         { NULL, posts_uc }, /* UC */
 118         { NULL, posts_le1 }, /* PD */
 119         { NULL, posts_at }, /* AT */
 120         { NULL, NULL }, /* in */
 121         { NULL, posts_ft }, /* ft */
 122         { NULL, posts_eq2 }, /* OP */
 123         { NULL, posts_nf }, /* EX */
 124         { NULL, posts_fi }, /* EE */
 125         { NULL, posts_ur }, /* UR */
 126         { NULL, NULL }, /* UE */
 127         { NULL, NULL }, /* ll */
 128 };
 129
 130
 131 int
 132 man_valid_pre(struct man *man, struct man_node *n)
 133 {
 134         v_check         *cp;
 135
 136         switch (n->type) {
 137         case MAN_TEXT:
 138                 /* FALLTHROUGH */
 139         case MAN_ROOT:
 140                 /* FALLTHROUGH */
 141         case MAN_EQN:
 142                 /* FALLTHROUGH */
 143         case MAN_TBL:
 144                 return(1);
 145         default:
 146                 break;
 147         }
 148
 149         if (NULL == (cp = man_valids[n->tok].pres))
 150                 return(1);
 151         for ( ; *cp; cp++)
 152                 if ( ! (*cp)(man, n))
 153                         return(0);
 154         return(1);
 155 }
 156
 157 int
 158 man_valid_post(struct man *man)
 159 {
 160         v_check         *cp;
 161
 162         if (MAN_VALID & man->last->flags)
 163                 return(1);
 164         man->last->flags |= MAN_VALID;
 165
 166         switch (man->last->type) {
 167         case MAN_TEXT:
 168                 check_text(man, man->last);
 169                 return(1);
 170         case MAN_ROOT:
 171                 return(check_root(man, man->last));
 172         case MAN_EQN:
 173                 /* FALLTHROUGH */
 174         case MAN_TBL:
 175                 return(1);
 176         default:
 177                 break;
 178         }
 179
 180         if (NULL == (cp = man_valids[man->last->tok].posts))
 181                 return(1);
 182         for ( ; *cp; cp++)
 183                 if ( ! (*cp)(man, man->last))
 184                         return(0);
 185
 186         return(1);
 187 }
 188
 189 static int
 190 check_root(CHKARGS)
 191 {
 192
 193         if (MAN_BLINE & man->flags)
 194                 man_nmsg(man, n, MANDOCERR_SCOPEEXIT);
 195         else if (MAN_ELINE & man->flags)
 196                 man_nmsg(man, n, MANDOCERR_SCOPEEXIT);
 197
 198         man->flags &= ~MAN_BLINE;
 199         man->flags &= ~MAN_ELINE;
 200
 201         if (NULL == man->first->child)
 202                 man_nmsg(man, n, MANDOCERR_DOC_EMPTY);
 203         else
 204                 man->meta.hasbody = 1;
 205
 206         if (NULL == man->meta.title) {
 207                 man_nmsg(man, n, MANDOCERR_TH_MISSING);
 208
 209                 /*
 210                  * If a title hasn't been set, do so now (by
 211                  * implication, date and section also aren't set).
 212                  */
 213
 214                 man->meta.title = mandoc_strdup("unknown");
 215                 man->meta.msec = mandoc_strdup("1");
 216                 man->meta.date = man->quick ? mandoc_strdup("") :
 217                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 218         }
 219
 220         return(1);
 221 }
 222
 223 static void
 224 check_text(CHKARGS)
 225 {
 226         char            *cp, *p;
 227
 228         if (MAN_LITERAL & man->flags)
 229                 return;
 230
 231         cp = n->string;
 232         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 233                 man_pmsg(man, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 234 }
 235
 236 #define INEQ_DEFINE(x, ineq, name) \
 237 static int \
 238 check_##name(CHKARGS) \
 239 { \
 240         if (n->nchild ineq (x)) \
 241                 return(1); \
 242         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 243             "line arguments %s %d (have %d)", \
 244             #ineq, (x), n->nchild); \
 245         return(1); \
 246 }
 247
 248 INEQ_DEFINE(0, ==, eq0)
 249 INEQ_DEFINE(2, ==, eq2)
 250 INEQ_DEFINE(1, <=, le1)
 251 INEQ_DEFINE(2, >=, ge2)
 252 INEQ_DEFINE(5, <=, le5)
 253
 254 static int
 255 check_head1(CHKARGS)
 256 {
 257
 258         if (MAN_HEAD == n->type && 1 != n->nchild)
 259                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 260                     n->pos, "line arguments eq 1 (have %d)", n->nchild);
 261
 262         return(1);
 263 }
 264
 265 static int
 266 post_ft(CHKARGS)
 267 {
 268         char    *cp;
 269         int      ok;
 270
 271         if (0 == n->nchild)
 272                 return(1);
 273
 274         ok = 0;
 275         cp = n->child->string;
 276         switch (*cp) {
 277         case '1':
 278                 /* FALLTHROUGH */
 279         case '2':
 280                 /* FALLTHROUGH */
 281         case '3':
 282                 /* FALLTHROUGH */
 283         case '4':
 284                 /* FALLTHROUGH */
 285         case 'I':
 286                 /* FALLTHROUGH */
 287         case 'P':
 288                 /* FALLTHROUGH */
 289         case 'R':
 290                 if ('\0' == cp[1])
 291                         ok = 1;
 292                 break;
 293         case 'B':
 294                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 295                         ok = 1;
 296                 break;
 297         case 'C':
 298                 if ('W' == cp[1] && '\0' == cp[2])
 299                         ok = 1;
 300                 break;
 301         default:
 302                 break;
 303         }
 304
 305         if (0 == ok) {
 306                 mandoc_vmsg(MANDOCERR_BADFONT, man->parse, n->line,
 307                     n->pos, "%s", cp);
 308                 *cp = '\0';
 309         }
 310
 311         if (1 < n->nchild)
 312                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 313                     n->pos, "want one child (have %d)", n->nchild);
 314
 315         return(1);
 316 }
 317
 318 static int
 319 pre_sec(CHKARGS)
 320 {
 321
 322         if (MAN_BLOCK == n->type)
 323                 man->flags &= ~MAN_LITERAL;
 324         return(1);
 325 }
 326
 327 static int
 328 post_sec(CHKARGS)
 329 {
 330
 331         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 332                 return(1);
 333
 334         man_nmsg(man, n, MANDOCERR_SYNTARGCOUNT);
 335         return(0);
 336 }
 337
 338 static int
 339 check_part(CHKARGS)
 340 {
 341
 342         if (MAN_BODY == n->type && 0 == n->nchild)
 343                 mandoc_msg(MANDOCERR_ARGCWARN, man->parse, n->line,
 344                     n->pos, "want children (have none)");
 345
 346         return(1);
 347 }
 348
 349 static int
 350 check_par(CHKARGS)
 351 {
 352
 353         switch (n->type) {
 354         case MAN_BLOCK:
 355                 if (0 == n->body->nchild)
 356                         man_node_delete(man, n);
 357                 break;
 358         case MAN_BODY:
 359                 if (0 == n->nchild)
 360                         man_nmsg(man, n, MANDOCERR_IGNPAR);
 361                 break;
 362         case MAN_HEAD:
 363                 if (n->nchild)
 364                         man_nmsg(man, n, MANDOCERR_ARGSLOST);
 365                 break;
 366         default:
 367                 break;
 368         }
 369
 370         return(1);
 371 }
 372
 373 static int
 374 post_IP(CHKARGS)
 375 {
 376
 377         switch (n->type) {
 378         case MAN_BLOCK:
 379                 if (0 == n->head->nchild && 0 == n->body->nchild)
 380                         man_node_delete(man, n);
 381                 break;
 382         case MAN_BODY:
 383                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 384                         man_nmsg(man, n, MANDOCERR_IGNPAR);
 385                 break;
 386         default:
 387                 break;
 388         }
 389         return(1);
 390 }
 391
 392 static int
 393 post_TH(CHKARGS)
 394 {
 395         struct man_node *nb;
 396         const char      *p;
 397
 398         free(man->meta.title);
 399         free(man->meta.vol);
 400         free(man->meta.source);
 401         free(man->meta.msec);
 402         free(man->meta.date);
 403
 404         man->meta.title = man->meta.vol = man->meta.date =
 405             man->meta.msec = man->meta.source = NULL;
 406
 407         nb = n;
 408
 409         /* ->TITLE<- MSEC DATE SOURCE VOL */
 410
 411         n = n->child;
 412         if (n && n->string) {
 413                 for (p = n->string; '\0' != *p; p++) {
 414                         /* Only warn about this once... */
 415                         if (isalpha((unsigned char)*p) &&
 416                             ! isupper((unsigned char)*p)) {
 417                                 mandoc_msg(MANDOCERR_TITLE_CASE,
 418                                     man->parse, n->line,
 419                                     n->pos + (p - n->string),
 420                                     n->string);
 421                                 break;
 422                         }
 423                 }
 424                 man->meta.title = mandoc_strdup(n->string);
 425         } else
 426                 man->meta.title = mandoc_strdup("");
 427
 428         /* TITLE ->MSEC<- DATE SOURCE VOL */
 429
 430         if (n)
 431                 n = n->next;
 432         if (n && n->string)
 433                 man->meta.msec = mandoc_strdup(n->string);
 434         else
 435                 man->meta.msec = mandoc_strdup("");
 436
 437         /* TITLE MSEC ->DATE<- SOURCE VOL */
 438
 439         if (n)
 440                 n = n->next;
 441         if (n && n->string && '\0' != n->string[0]) {
 442                 man->meta.date = man->quick ?
 443                     mandoc_strdup(n->string) :
 444                     mandoc_normdate(man->parse, n->string,
 445                         n->line, n->pos);
 446         } else {
 447                 man->meta.date = mandoc_strdup("");
 448                 man_nmsg(man, n ? n : nb, MANDOCERR_DATE_MISSING);
 449         }
 450
 451         /* TITLE MSEC DATE ->SOURCE<- VOL */
 452
 453         if (n && (n = n->next))
 454                 man->meta.source = mandoc_strdup(n->string);
 455
 456         /* TITLE MSEC DATE SOURCE ->VOL<- */
 457         /* If missing, use the default VOL name for MSEC. */
 458
 459         if (n && (n = n->next))
 460                 man->meta.vol = mandoc_strdup(n->string);
 461         else if ('\0' != man->meta.msec[0] &&
 462             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 463                 man->meta.vol = mandoc_strdup(p);
 464
 465         /*
 466          * Remove the `TH' node after we've processed it for our
 467          * meta-data.
 468          */
 469         man_node_delete(man, man->last);
 470         return(1);
 471 }
 472
 473 static int
 474 post_nf(CHKARGS)
 475 {
 476
 477         if (MAN_LITERAL & man->flags)
 478                 man_nmsg(man, n, MANDOCERR_SCOPEREP);
 479
 480         man->flags |= MAN_LITERAL;
 481         return(1);
 482 }
 483
 484 static int
 485 post_fi(CHKARGS)
 486 {
 487
 488         if ( ! (MAN_LITERAL & man->flags))
 489                 man_nmsg(man, n, MANDOCERR_WNOSCOPE);
 490
 491         man->flags &= ~MAN_LITERAL;
 492         return(1);
 493 }
 494
 495 static int
 496 post_UC(CHKARGS)
 497 {
 498         static const char * const bsd_versions[] = {
 499             "3rd Berkeley Distribution",
 500             "4th Berkeley Distribution",
 501             "4.2 Berkeley Distribution",
 502             "4.3 Berkeley Distribution",
 503             "4.4 Berkeley Distribution",
 504         };
 505
 506         const char      *p, *s;
 507
 508         n = n->child;
 509
 510         if (NULL == n || MAN_TEXT != n->type)
 511                 p = bsd_versions[0];
 512         else {
 513                 s = n->string;
 514                 if (0 == strcmp(s, "3"))
 515                         p = bsd_versions[0];
 516                 else if (0 == strcmp(s, "4"))
 517                         p = bsd_versions[1];
 518                 else if (0 == strcmp(s, "5"))
 519                         p = bsd_versions[2];
 520                 else if (0 == strcmp(s, "6"))
 521                         p = bsd_versions[3];
 522                 else if (0 == strcmp(s, "7"))
 523                         p = bsd_versions[4];
 524                 else
 525                         p = bsd_versions[0];
 526         }
 527
 528         free(man->meta.source);
 529         man->meta.source = mandoc_strdup(p);
 530         return(1);
 531 }
 532
 533 static int
 534 post_AT(CHKARGS)
 535 {
 536         static const char * const unix_versions[] = {
 537             "7th Edition",
 538             "System III",
 539             "System V",
 540             "System V Release 2",
 541         };
 542
 543         const char      *p, *s;
 544         struct man_node *nn;
 545
 546         n = n->child;
 547
 548         if (NULL == n || MAN_TEXT != n->type)
 549                 p = unix_versions[0];
 550         else {
 551                 s = n->string;
 552                 if (0 == strcmp(s, "3"))
 553                         p = unix_versions[0];
 554                 else if (0 == strcmp(s, "4"))
 555                         p = unix_versions[1];
 556                 else if (0 == strcmp(s, "5")) {
 557                         nn = n->next;
 558                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 559                                 p = unix_versions[3];
 560                         else
 561                                 p = unix_versions[2];
 562                 } else
 563                         p = unix_versions[0];
 564         }
 565
 566         free(man->meta.source);
 567         man->meta.source = mandoc_strdup(p);
 568         return(1);
 569 }
 570
 571 static int
 572 post_vs(CHKARGS)
 573 {
 574
 575         if (NULL != n->prev)
 576                 return(1);
 577
 578         switch (n->parent->tok) {
 579         case MAN_SH:
 580                 /* FALLTHROUGH */
 581         case MAN_SS:
 582                 man_nmsg(man, n, MANDOCERR_IGNPAR);
 583                 /* FALLTHROUGH */
 584         case MAN_MAX:
 585                 /*
 586                  * Don't warn about this because it occurs in pod2man
 587                  * and would cause considerable (unfixable) warnage.
 588                  */
 589                 man_node_delete(man, n);
 590                 break;
 591         default:
 592                 break;
 593         }
 594
 595         return(1);
 596 }