man_validate.c

   1 /*      $Id: man_validate.c,v 1.91 2014/04/20 16:46:04 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_NODOCBODY);
 203                 return(0);
 204         } else if (NULL == man->meta.title) {
 205                 man_nmsg(man, n, MANDOCERR_NOTITLE);
 206
 207                 /*
 208                  * If a title hasn't been set, do so now (by
 209                  * implication, date and section also aren't set).
 210                  */
 211
 212                 man->meta.title = mandoc_strdup("unknown");
 213                 man->meta.msec = mandoc_strdup("1");
 214                 man->meta.date = man->quick ? mandoc_strdup("") :
 215                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 216         }
 217
 218         return(1);
 219 }
 220
 221 static void
 222 check_text(CHKARGS)
 223 {
 224         char            *cp, *p;
 225
 226         if (MAN_LITERAL & man->flags)
 227                 return;
 228
 229         cp = n->string;
 230         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 231                 man_pmsg(man, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 232 }
 233
 234 #define INEQ_DEFINE(x, ineq, name) \
 235 static int \
 236 check_##name(CHKARGS) \
 237 { \
 238         if (n->nchild ineq (x)) \
 239                 return(1); \
 240         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 241             "line arguments %s %d (have %d)", \
 242             #ineq, (x), n->nchild); \
 243         return(1); \
 244 }
 245
 246 INEQ_DEFINE(0, ==, eq0)
 247 INEQ_DEFINE(2, ==, eq2)
 248 INEQ_DEFINE(1, <=, le1)
 249 INEQ_DEFINE(2, >=, ge2)
 250 INEQ_DEFINE(5, <=, le5)
 251
 252 static int
 253 check_head1(CHKARGS)
 254 {
 255
 256         if (MAN_HEAD == n->type && 1 != n->nchild)
 257                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 258                     n->pos, "line arguments eq 1 (have %d)", n->nchild);
 259
 260         return(1);
 261 }
 262
 263 static int
 264 post_ft(CHKARGS)
 265 {
 266         char    *cp;
 267         int      ok;
 268
 269         if (0 == n->nchild)
 270                 return(1);
 271
 272         ok = 0;
 273         cp = n->child->string;
 274         switch (*cp) {
 275         case '1':
 276                 /* FALLTHROUGH */
 277         case '2':
 278                 /* FALLTHROUGH */
 279         case '3':
 280                 /* FALLTHROUGH */
 281         case '4':
 282                 /* FALLTHROUGH */
 283         case 'I':
 284                 /* FALLTHROUGH */
 285         case 'P':
 286                 /* FALLTHROUGH */
 287         case 'R':
 288                 if ('\0' == cp[1])
 289                         ok = 1;
 290                 break;
 291         case 'B':
 292                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 293                         ok = 1;
 294                 break;
 295         case 'C':
 296                 if ('W' == cp[1] && '\0' == cp[2])
 297                         ok = 1;
 298                 break;
 299         default:
 300                 break;
 301         }
 302
 303         if (0 == ok) {
 304                 mandoc_vmsg(MANDOCERR_BADFONT, man->parse, n->line,
 305                     n->pos, "%s", cp);
 306                 *cp = '\0';
 307         }
 308
 309         if (1 < n->nchild)
 310                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 311                     n->pos, "want one child (have %d)", n->nchild);
 312
 313         return(1);
 314 }
 315
 316 static int
 317 pre_sec(CHKARGS)
 318 {
 319
 320         if (MAN_BLOCK == n->type)
 321                 man->flags &= ~MAN_LITERAL;
 322         return(1);
 323 }
 324
 325 static int
 326 post_sec(CHKARGS)
 327 {
 328
 329         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 330                 return(1);
 331
 332         man_nmsg(man, n, MANDOCERR_SYNTARGCOUNT);
 333         return(0);
 334 }
 335
 336 static int
 337 check_part(CHKARGS)
 338 {
 339
 340         if (MAN_BODY == n->type && 0 == n->nchild)
 341                 mandoc_msg(MANDOCERR_ARGCWARN, man->parse, n->line,
 342                     n->pos, "want children (have none)");
 343
 344         return(1);
 345 }
 346
 347 static int
 348 check_par(CHKARGS)
 349 {
 350
 351         switch (n->type) {
 352         case MAN_BLOCK:
 353                 if (0 == n->body->nchild)
 354                         man_node_delete(man, n);
 355                 break;
 356         case MAN_BODY:
 357                 if (0 == n->nchild)
 358                         man_nmsg(man, n, MANDOCERR_IGNPAR);
 359                 break;
 360         case MAN_HEAD:
 361                 if (n->nchild)
 362                         man_nmsg(man, n, MANDOCERR_ARGSLOST);
 363                 break;
 364         default:
 365                 break;
 366         }
 367
 368         return(1);
 369 }
 370
 371 static int
 372 post_IP(CHKARGS)
 373 {
 374
 375         switch (n->type) {
 376         case MAN_BLOCK:
 377                 if (0 == n->head->nchild && 0 == n->body->nchild)
 378                         man_node_delete(man, n);
 379                 break;
 380         case MAN_BODY:
 381                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 382                         man_nmsg(man, n, MANDOCERR_IGNPAR);
 383                 break;
 384         default:
 385                 break;
 386         }
 387         return(1);
 388 }
 389
 390 static int
 391 post_TH(CHKARGS)
 392 {
 393         const char      *p;
 394
 395         free(man->meta.title);
 396         free(man->meta.vol);
 397         free(man->meta.source);
 398         free(man->meta.msec);
 399         free(man->meta.date);
 400
 401         man->meta.title = man->meta.vol = man->meta.date =
 402             man->meta.msec = man->meta.source = NULL;
 403
 404         /* ->TITLE<- MSEC DATE SOURCE VOL */
 405
 406         n = n->child;
 407         if (n && n->string) {
 408                 for (p = n->string; '\0' != *p; p++) {
 409                         /* Only warn about this once... */
 410                         if (isalpha((unsigned char)*p) &&
 411                             ! isupper((unsigned char)*p)) {
 412                                 man_nmsg(man, n, MANDOCERR_UPPERCASE);
 413                                 break;
 414                         }
 415                 }
 416                 man->meta.title = mandoc_strdup(n->string);
 417         } else
 418                 man->meta.title = mandoc_strdup("");
 419
 420         /* TITLE ->MSEC<- DATE SOURCE VOL */
 421
 422         if (n)
 423                 n = n->next;
 424         if (n && n->string)
 425                 man->meta.msec = mandoc_strdup(n->string);
 426         else
 427                 man->meta.msec = mandoc_strdup("");
 428
 429         /* TITLE MSEC ->DATE<- SOURCE VOL */
 430
 431         if (n)
 432                 n = n->next;
 433         if (n && n->string && '\0' != n->string[0]) {
 434                 man->meta.date = man->quick ?
 435                     mandoc_strdup(n->string) :
 436                     mandoc_normdate(man->parse, n->string,
 437                         n->line, n->pos);
 438         } else
 439                 man->meta.date = mandoc_strdup("");
 440
 441         /* TITLE MSEC DATE ->SOURCE<- VOL */
 442
 443         if (n && (n = n->next))
 444                 man->meta.source = mandoc_strdup(n->string);
 445
 446         /* TITLE MSEC DATE SOURCE ->VOL<- */
 447         /* If missing, use the default VOL name for MSEC. */
 448
 449         if (n && (n = n->next))
 450                 man->meta.vol = mandoc_strdup(n->string);
 451         else if ('\0' != man->meta.msec[0] &&
 452             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 453                 man->meta.vol = mandoc_strdup(p);
 454
 455         /*
 456          * Remove the `TH' node after we've processed it for our
 457          * meta-data.
 458          */
 459         man_node_delete(man, man->last);
 460         return(1);
 461 }
 462
 463 static int
 464 post_nf(CHKARGS)
 465 {
 466
 467         if (MAN_LITERAL & man->flags)
 468                 man_nmsg(man, n, MANDOCERR_SCOPEREP);
 469
 470         man->flags |= MAN_LITERAL;
 471         return(1);
 472 }
 473
 474 static int
 475 post_fi(CHKARGS)
 476 {
 477
 478         if ( ! (MAN_LITERAL & man->flags))
 479                 man_nmsg(man, n, MANDOCERR_WNOSCOPE);
 480
 481         man->flags &= ~MAN_LITERAL;
 482         return(1);
 483 }
 484
 485 static int
 486 post_UC(CHKARGS)
 487 {
 488         static const char * const bsd_versions[] = {
 489             "3rd Berkeley Distribution",
 490             "4th Berkeley Distribution",
 491             "4.2 Berkeley Distribution",
 492             "4.3 Berkeley Distribution",
 493             "4.4 Berkeley Distribution",
 494         };
 495
 496         const char      *p, *s;
 497
 498         n = n->child;
 499
 500         if (NULL == n || MAN_TEXT != n->type)
 501                 p = bsd_versions[0];
 502         else {
 503                 s = n->string;
 504                 if (0 == strcmp(s, "3"))
 505                         p = bsd_versions[0];
 506                 else if (0 == strcmp(s, "4"))
 507                         p = bsd_versions[1];
 508                 else if (0 == strcmp(s, "5"))
 509                         p = bsd_versions[2];
 510                 else if (0 == strcmp(s, "6"))
 511                         p = bsd_versions[3];
 512                 else if (0 == strcmp(s, "7"))
 513                         p = bsd_versions[4];
 514                 else
 515                         p = bsd_versions[0];
 516         }
 517
 518         free(man->meta.source);
 519         man->meta.source = mandoc_strdup(p);
 520         return(1);
 521 }
 522
 523 static int
 524 post_AT(CHKARGS)
 525 {
 526         static const char * const unix_versions[] = {
 527             "7th Edition",
 528             "System III",
 529             "System V",
 530             "System V Release 2",
 531         };
 532
 533         const char      *p, *s;
 534         struct man_node *nn;
 535
 536         n = n->child;
 537
 538         if (NULL == n || MAN_TEXT != n->type)
 539                 p = unix_versions[0];
 540         else {
 541                 s = n->string;
 542                 if (0 == strcmp(s, "3"))
 543                         p = unix_versions[0];
 544                 else if (0 == strcmp(s, "4"))
 545                         p = unix_versions[1];
 546                 else if (0 == strcmp(s, "5")) {
 547                         nn = n->next;
 548                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 549                                 p = unix_versions[3];
 550                         else
 551                                 p = unix_versions[2];
 552                 } else
 553                         p = unix_versions[0];
 554         }
 555
 556         free(man->meta.source);
 557         man->meta.source = mandoc_strdup(p);
 558         return(1);
 559 }
 560
 561 static int
 562 post_vs(CHKARGS)
 563 {
 564
 565         if (NULL != n->prev)
 566                 return(1);
 567
 568         switch (n->parent->tok) {
 569         case MAN_SH:
 570                 /* FALLTHROUGH */
 571         case MAN_SS:
 572                 man_nmsg(man, n, MANDOCERR_IGNPAR);
 573                 /* FALLTHROUGH */
 574         case MAN_MAX:
 575                 /*
 576                  * Don't warn about this because it occurs in pod2man
 577                  * and would cause considerable (unfixable) warnage.
 578                  */
 579                 man_node_delete(man, n);
 580                 break;
 581         default:
 582                 break;
 583         }
 584
 585         return(1);
 586 }