man_validate.c

   1 /*      $Id: man_validate.c,v 1.100 2014/07/07 21:36:20 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_ELINE) & man->flags)
 194                 mandoc_msg(MANDOCERR_BLK_LINE, man->parse,
 195                     0, 0, "at end of file");
 196
 197         man->flags &= ~MAN_BLINE;
 198         man->flags &= ~MAN_ELINE;
 199
 200         if (NULL == man->first->child)
 201                 man_nmsg(man, n, MANDOCERR_DOC_EMPTY);
 202         else
 203                 man->meta.hasbody = 1;
 204
 205         if (NULL == man->meta.title) {
 206                 man_nmsg(man, n, MANDOCERR_TH_MISSING);
 207
 208                 /*
 209                  * If a title hasn't been set, do so now (by
 210                  * implication, date and section also aren't set).
 211                  */
 212
 213                 man->meta.title = mandoc_strdup("unknown");
 214                 man->meta.msec = mandoc_strdup("1");
 215                 man->meta.date = man->quick ? mandoc_strdup("") :
 216                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 217         }
 218
 219         return(1);
 220 }
 221
 222 static void
 223 check_text(CHKARGS)
 224 {
 225         char            *cp, *p;
 226
 227         if (MAN_LITERAL & man->flags)
 228                 return;
 229
 230         cp = n->string;
 231         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 232                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 233                     n->line, n->pos + (p - cp), NULL);
 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_FT_BAD, man->parse,
 307                     n->line, n->pos, "ft %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                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 361                             man->parse, n->line, n->pos,
 362                             "%s empty", man_macronames[n->tok]);
 363                 break;
 364         case MAN_HEAD:
 365                 if (n->nchild)
 366                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 367                             man->parse, n->line, n->pos,
 368                             "%s %s%s", man_macronames[n->tok],
 369                             n->child->string,
 370                             n->nchild > 1 ? " ..." : "");
 371                 break;
 372         default:
 373                 break;
 374         }
 375
 376         return(1);
 377 }
 378
 379 static int
 380 post_IP(CHKARGS)
 381 {
 382
 383         switch (n->type) {
 384         case MAN_BLOCK:
 385                 if (0 == n->head->nchild && 0 == n->body->nchild)
 386                         man_node_delete(man, n);
 387                 break;
 388         case MAN_BODY:
 389                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 390                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 391                             man->parse, n->line, n->pos,
 392                             "%s empty", man_macronames[n->tok]);
 393                 break;
 394         default:
 395                 break;
 396         }
 397         return(1);
 398 }
 399
 400 static int
 401 post_TH(CHKARGS)
 402 {
 403         struct man_node *nb;
 404         const char      *p;
 405
 406         free(man->meta.title);
 407         free(man->meta.vol);
 408         free(man->meta.source);
 409         free(man->meta.msec);
 410         free(man->meta.date);
 411
 412         man->meta.title = man->meta.vol = man->meta.date =
 413             man->meta.msec = man->meta.source = NULL;
 414
 415         nb = n;
 416
 417         /* ->TITLE<- MSEC DATE SOURCE VOL */
 418
 419         n = n->child;
 420         if (n && n->string) {
 421                 for (p = n->string; '\0' != *p; p++) {
 422                         /* Only warn about this once... */
 423                         if (isalpha((unsigned char)*p) &&
 424                             ! isupper((unsigned char)*p)) {
 425                                 mandoc_msg(MANDOCERR_TITLE_CASE,
 426                                     man->parse, n->line,
 427                                     n->pos + (p - n->string),
 428                                     n->string);
 429                                 break;
 430                         }
 431                 }
 432                 man->meta.title = mandoc_strdup(n->string);
 433         } else
 434                 man->meta.title = mandoc_strdup("");
 435
 436         /* TITLE ->MSEC<- DATE SOURCE VOL */
 437
 438         if (n)
 439                 n = n->next;
 440         if (n && n->string)
 441                 man->meta.msec = mandoc_strdup(n->string);
 442         else
 443                 man->meta.msec = mandoc_strdup("");
 444
 445         /* TITLE MSEC ->DATE<- SOURCE VOL */
 446
 447         if (n)
 448                 n = n->next;
 449         if (n && n->string && '\0' != n->string[0]) {
 450                 man->meta.date = man->quick ?
 451                     mandoc_strdup(n->string) :
 452                     mandoc_normdate(man->parse, n->string,
 453                         n->line, n->pos);
 454         } else {
 455                 man->meta.date = mandoc_strdup("");
 456                 man_nmsg(man, n ? n : nb, MANDOCERR_DATE_MISSING);
 457         }
 458
 459         /* TITLE MSEC DATE ->SOURCE<- VOL */
 460
 461         if (n && (n = n->next))
 462                 man->meta.source = mandoc_strdup(n->string);
 463
 464         /* TITLE MSEC DATE SOURCE ->VOL<- */
 465         /* If missing, use the default VOL name for MSEC. */
 466
 467         if (n && (n = n->next))
 468                 man->meta.vol = mandoc_strdup(n->string);
 469         else if ('\0' != man->meta.msec[0] &&
 470             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 471                 man->meta.vol = mandoc_strdup(p);
 472
 473         /*
 474          * Remove the `TH' node after we've processed it for our
 475          * meta-data.
 476          */
 477         man_node_delete(man, man->last);
 478         return(1);
 479 }
 480
 481 static int
 482 post_nf(CHKARGS)
 483 {
 484
 485         if (MAN_LITERAL & man->flags)
 486                 man_nmsg(man, n, MANDOCERR_NF_SKIP);
 487
 488         man->flags |= MAN_LITERAL;
 489         return(1);
 490 }
 491
 492 static int
 493 post_fi(CHKARGS)
 494 {
 495
 496         if ( ! (MAN_LITERAL & man->flags))
 497                 man_nmsg(man, n, MANDOCERR_FI_SKIP);
 498
 499         man->flags &= ~MAN_LITERAL;
 500         return(1);
 501 }
 502
 503 static int
 504 post_UC(CHKARGS)
 505 {
 506         static const char * const bsd_versions[] = {
 507             "3rd Berkeley Distribution",
 508             "4th Berkeley Distribution",
 509             "4.2 Berkeley Distribution",
 510             "4.3 Berkeley Distribution",
 511             "4.4 Berkeley Distribution",
 512         };
 513
 514         const char      *p, *s;
 515
 516         n = n->child;
 517
 518         if (NULL == n || MAN_TEXT != n->type)
 519                 p = bsd_versions[0];
 520         else {
 521                 s = n->string;
 522                 if (0 == strcmp(s, "3"))
 523                         p = bsd_versions[0];
 524                 else if (0 == strcmp(s, "4"))
 525                         p = bsd_versions[1];
 526                 else if (0 == strcmp(s, "5"))
 527                         p = bsd_versions[2];
 528                 else if (0 == strcmp(s, "6"))
 529                         p = bsd_versions[3];
 530                 else if (0 == strcmp(s, "7"))
 531                         p = bsd_versions[4];
 532                 else
 533                         p = bsd_versions[0];
 534         }
 535
 536         free(man->meta.source);
 537         man->meta.source = mandoc_strdup(p);
 538         return(1);
 539 }
 540
 541 static int
 542 post_AT(CHKARGS)
 543 {
 544         static const char * const unix_versions[] = {
 545             "7th Edition",
 546             "System III",
 547             "System V",
 548             "System V Release 2",
 549         };
 550
 551         const char      *p, *s;
 552         struct man_node *nn;
 553
 554         n = n->child;
 555
 556         if (NULL == n || MAN_TEXT != n->type)
 557                 p = unix_versions[0];
 558         else {
 559                 s = n->string;
 560                 if (0 == strcmp(s, "3"))
 561                         p = unix_versions[0];
 562                 else if (0 == strcmp(s, "4"))
 563                         p = unix_versions[1];
 564                 else if (0 == strcmp(s, "5")) {
 565                         nn = n->next;
 566                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 567                                 p = unix_versions[3];
 568                         else
 569                                 p = unix_versions[2];
 570                 } else
 571                         p = unix_versions[0];
 572         }
 573
 574         free(man->meta.source);
 575         man->meta.source = mandoc_strdup(p);
 576         return(1);
 577 }
 578
 579 static int
 580 post_vs(CHKARGS)
 581 {
 582
 583         if (NULL != n->prev)
 584                 return(1);
 585
 586         switch (n->parent->tok) {
 587         case MAN_SH:
 588                 /* FALLTHROUGH */
 589         case MAN_SS:
 590                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 591                     "%s after %s", man_macronames[n->tok],
 592                     man_macronames[n->parent->tok]);
 593                 /* FALLTHROUGH */
 594         case MAN_MAX:
 595                 /*
 596                  * Don't warn about this because it occurs in pod2man
 597                  * and would cause considerable (unfixable) warnage.
 598                  */
 599                 man_node_delete(man, n);
 600                 break;
 601         default:
 602                 break;
 603         }
 604
 605         return(1);
 606 }