man_validate.c

   1 /*      $Id: man_validate.c,v 1.62 2011/02/09 09:18:15 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <errno.h>
  26 #include <limits.h>
  27 #include <stdarg.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <time.h>
  31
  32 #include "mandoc.h"
  33 #include "libman.h"
  34 #include "libmandoc.h"
  35
  36 #define CHKARGS   struct man *m, struct man_node *n
  37
  38 typedef int     (*v_check)(CHKARGS);
  39
  40 struct  man_valid {
  41         v_check  *pres;
  42         v_check  *posts;
  43 };
  44
  45 static  int       check_bline(CHKARGS);
  46 static  int       check_eq0(CHKARGS);
  47 static  int       check_ft(CHKARGS);
  48 static  int       check_le1(CHKARGS);
  49 static  int       check_ge2(CHKARGS);
  50 static  int       check_le5(CHKARGS);
  51 static  int       check_par(CHKARGS);
  52 static  int       check_part(CHKARGS);
  53 static  int       check_root(CHKARGS);
  54 static  int       check_sec(CHKARGS);
  55 static  int       check_text(CHKARGS);
  56
  57 static  int       post_AT(CHKARGS);
  58 static  int       post_fi(CHKARGS);
  59 static  int       post_nf(CHKARGS);
  60 static  int       post_TH(CHKARGS);
  61 static  int       post_UC(CHKARGS);
  62
  63 static  v_check   posts_at[] = { post_AT, NULL };
  64 static  v_check   posts_eq0[] = { check_eq0, NULL };
  65 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  66 static  v_check   posts_le1[] = { check_le1, NULL };
  67 static  v_check   posts_ft[] = { check_ft, NULL };
  68 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  69 static  v_check   posts_par[] = { check_par, NULL };
  70 static  v_check   posts_part[] = { check_part, NULL };
  71 static  v_check   posts_sec[] = { check_sec, NULL };
  72 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  73 static  v_check   posts_uc[] = { post_UC, NULL };
  74 static  v_check   pres_bline[] = { check_bline, NULL };
  75
  76
  77 static  const struct man_valid man_valids[MAN_MAX] = {
  78         { NULL, posts_eq0 }, /* br */
  79         { pres_bline, posts_th }, /* TH */
  80         { pres_bline, posts_sec }, /* SH */
  81         { pres_bline, posts_sec }, /* SS */
  82         { pres_bline, NULL }, /* TP */
  83         { pres_bline, posts_par }, /* LP */
  84         { pres_bline, posts_par }, /* PP */
  85         { pres_bline, posts_par }, /* P */
  86         { pres_bline, NULL }, /* IP */
  87         { pres_bline, NULL }, /* HP */
  88         { NULL, NULL }, /* SM */
  89         { NULL, NULL }, /* SB */
  90         { NULL, NULL }, /* BI */
  91         { NULL, NULL }, /* IB */
  92         { NULL, NULL }, /* BR */
  93         { NULL, NULL }, /* RB */
  94         { NULL, NULL }, /* R */
  95         { NULL, NULL }, /* B */
  96         { NULL, NULL }, /* I */
  97         { NULL, NULL }, /* IR */
  98         { NULL, NULL }, /* RI */
  99         { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
 100         { NULL, posts_le1 }, /* sp */ /* FIXME: should warn only. */
 101         { pres_bline, posts_nf }, /* nf */
 102         { pres_bline, posts_fi }, /* fi */
 103         { NULL, NULL }, /* RE */
 104         { NULL, posts_part }, /* RS */
 105         { NULL, NULL }, /* DT */
 106         { NULL, posts_uc }, /* UC */
 107         { NULL, NULL }, /* PD */
 108         { NULL, posts_at }, /* AT */
 109         { NULL, NULL }, /* in */
 110         { NULL, posts_ft }, /* ft */
 111 };
 112
 113
 114 int
 115 man_valid_pre(struct man *m, struct man_node *n)
 116 {
 117         v_check         *cp;
 118
 119         switch (n->type) {
 120         case (MAN_TEXT):
 121                 /* FALLTHROUGH */
 122         case (MAN_ROOT):
 123                 /* FALLTHROUGH */
 124         case (MAN_EQN):
 125                 /* FALLTHROUGH */
 126         case (MAN_TBL):
 127                 return(1);
 128         default:
 129                 break;
 130         }
 131
 132         if (NULL == (cp = man_valids[n->tok].pres))
 133                 return(1);
 134         for ( ; *cp; cp++)
 135                 if ( ! (*cp)(m, n))
 136                         return(0);
 137         return(1);
 138 }
 139
 140
 141 int
 142 man_valid_post(struct man *m)
 143 {
 144         v_check         *cp;
 145
 146         if (MAN_VALID & m->last->flags)
 147                 return(1);
 148         m->last->flags |= MAN_VALID;
 149
 150         switch (m->last->type) {
 151         case (MAN_TEXT):
 152                 return(check_text(m, m->last));
 153         case (MAN_ROOT):
 154                 return(check_root(m, m->last));
 155         case (MAN_EQN):
 156                 /* FALLTHROUGH */
 157         case (MAN_TBL):
 158                 return(1);
 159         default:
 160                 break;
 161         }
 162
 163         if (NULL == (cp = man_valids[m->last->tok].posts))
 164                 return(1);
 165         for ( ; *cp; cp++)
 166                 if ( ! (*cp)(m, m->last))
 167                         return(0);
 168
 169         return(1);
 170 }
 171
 172
 173 static int
 174 check_root(CHKARGS)
 175 {
 176
 177         if (MAN_BLINE & m->flags)
 178                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 179         else if (MAN_ELINE & m->flags)
 180                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 181
 182         m->flags &= ~MAN_BLINE;
 183         m->flags &= ~MAN_ELINE;
 184
 185         if (NULL == m->first->child) {
 186                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 187                 return(0);
 188         } else if (NULL == m->meta.title) {
 189                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 190
 191                 /*
 192                  * If a title hasn't been set, do so now (by
 193                  * implication, date and section also aren't set).
 194                  */
 195
 196                 m->meta.title = mandoc_strdup("unknown");
 197                 m->meta.date = time(NULL);
 198                 m->meta.msec = mandoc_strdup("1");
 199         }
 200
 201         return(1);
 202 }
 203
 204
 205 static int
 206 check_text(CHKARGS)
 207 {
 208         char            *p;
 209         int              pos, c;
 210         size_t           sz;
 211
 212         for (p = n->string, pos = n->pos + 1; *p; p++, pos++) {
 213                 sz = strcspn(p, "\t\\");
 214                 p += (int)sz;
 215
 216                 if ('\0' == *p)
 217                         break;
 218
 219                 pos += (int)sz;
 220
 221                 if ('\t' == *p) {
 222                         if (MAN_LITERAL & m->flags)
 223                                 continue;
 224                         if (man_pmsg(m, n->line, pos, MANDOCERR_BADTAB))
 225                                 continue;
 226                         return(0);
 227                 }
 228
 229                 /* Check the special character. */
 230
 231                 c = mandoc_special(p);
 232                 if (c) {
 233                         p += c - 1;
 234                         pos += c - 1;
 235                 } else
 236                         man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
 237         }
 238
 239         return(1);
 240 }
 241
 242
 243 #define INEQ_DEFINE(x, ineq, name) \
 244 static int \
 245 check_##name(CHKARGS) \
 246 { \
 247         if (n->nchild ineq (x)) \
 248                 return(1); \
 249         man_vmsg(m, MANDOCERR_ARGCOUNT, n->line, n->pos, \
 250                         "line arguments %s %d (have %d)", \
 251                         #ineq, (x), n->nchild); \
 252         return(1); \
 253 }
 254
 255 INEQ_DEFINE(0, ==, eq0)
 256 INEQ_DEFINE(1, <=, le1)
 257 INEQ_DEFINE(2, >=, ge2)
 258 INEQ_DEFINE(5, <=, le5)
 259
 260 static int
 261 check_ft(CHKARGS)
 262 {
 263         char    *cp;
 264         int      ok;
 265
 266         if (0 == n->nchild)
 267                 return(1);
 268
 269         ok = 0;
 270         cp = n->child->string;
 271         switch (*cp) {
 272         case ('1'):
 273                 /* FALLTHROUGH */
 274         case ('2'):
 275                 /* FALLTHROUGH */
 276         case ('3'):
 277                 /* FALLTHROUGH */
 278         case ('4'):
 279                 /* FALLTHROUGH */
 280         case ('I'):
 281                 /* FALLTHROUGH */
 282         case ('P'):
 283                 /* FALLTHROUGH */
 284         case ('R'):
 285                 if ('\0' == cp[1])
 286                         ok = 1;
 287                 break;
 288         case ('B'):
 289                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 290                         ok = 1;
 291                 break;
 292         case ('C'):
 293                 if ('W' == cp[1] && '\0' == cp[2])
 294                         ok = 1;
 295                 break;
 296         default:
 297                 break;
 298         }
 299
 300         if (0 == ok) {
 301                 man_vmsg(m, MANDOCERR_BADFONT,
 302                                 n->line, n->pos, "%s", cp);
 303                 *cp = '\0';
 304         }
 305
 306         if (1 < n->nchild)
 307                 man_vmsg(m, MANDOCERR_ARGCOUNT, n->line, n->pos,
 308                                 "want one child (have %d)", n->nchild);
 309
 310         return(1);
 311 }
 312
 313 static int
 314 check_sec(CHKARGS)
 315 {
 316
 317         if (MAN_HEAD == n->type && 0 == n->nchild) {
 318                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 319                 return(0);
 320         } else if (MAN_BODY == n->type && 0 == n->nchild)
 321                 man_nmsg(m, n, MANDOCERR_NOBODY);
 322
 323         return(1);
 324 }
 325
 326
 327 static int
 328 check_part(CHKARGS)
 329 {
 330
 331         if (MAN_BODY == n->type && 0 == n->nchild)
 332                 man_nmsg(m, n, MANDOCERR_NOBODY);
 333
 334         return(1);
 335 }
 336
 337
 338 static int
 339 check_par(CHKARGS)
 340 {
 341
 342         switch (n->type) {
 343         case (MAN_BLOCK):
 344                 if (0 == n->body->nchild)
 345                         man_node_delete(m, n);
 346                 break;
 347         case (MAN_BODY):
 348                 if (0 == n->nchild)
 349                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 350                 break;
 351         case (MAN_HEAD):
 352                 if (n->nchild)
 353                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 354                 break;
 355         default:
 356                 break;
 357         }
 358
 359         return(1);
 360 }
 361
 362
 363 static int
 364 check_bline(CHKARGS)
 365 {
 366
 367         assert( ! (MAN_ELINE & m->flags));
 368         if (MAN_BLINE & m->flags) {
 369                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 370                 return(0);
 371         }
 372
 373         return(1);
 374 }
 375
 376 static int
 377 post_TH(CHKARGS)
 378 {
 379         const char      *p;
 380
 381         if (m->meta.title)
 382                 free(m->meta.title);
 383         if (m->meta.vol)
 384                 free(m->meta.vol);
 385         if (m->meta.source)
 386                 free(m->meta.source);
 387         if (m->meta.msec)
 388                 free(m->meta.msec);
 389         if (m->meta.rawdate)
 390                 free(m->meta.rawdate);
 391
 392         m->meta.title = m->meta.vol = m->meta.rawdate =
 393                 m->meta.msec = m->meta.source = NULL;
 394         m->meta.date = 0;
 395
 396         /* ->TITLE<- MSEC DATE SOURCE VOL */
 397
 398         n = n->child;
 399         if (n && n->string) {
 400                 for (p = n->string; '\0' != *p; p++) {
 401                         /* Only warn about this once... */
 402                         if (isalpha((u_char)*p) && ! isupper((u_char)*p)) {
 403                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 404                                 break;
 405                         }
 406                 }
 407                 m->meta.title = mandoc_strdup(n->string);
 408         } else
 409                 m->meta.title = mandoc_strdup("");
 410
 411         /* TITLE ->MSEC<- DATE SOURCE VOL */
 412
 413         if (n)
 414                 n = n->next;
 415         if (n && n->string)
 416                 m->meta.msec = mandoc_strdup(n->string);
 417         else
 418                 m->meta.msec = mandoc_strdup("");
 419
 420         /* TITLE MSEC ->DATE<- SOURCE VOL */
 421
 422         /*
 423          * Try to parse the date.  If this works, stash the epoch (this
 424          * is optimal because we can reformat it in the canonical form).
 425          * If it doesn't parse, isn't specified at all, or is an empty
 426          * string, then use the current date.
 427          */
 428
 429         if (n)
 430                 n = n->next;
 431         if (n && n->string && *n->string) {
 432                 m->meta.date = mandoc_a2time
 433                         (MTIME_ISO_8601, n->string);
 434                 if (0 == m->meta.date) {
 435                         man_nmsg(m, n, MANDOCERR_BADDATE);
 436                         m->meta.rawdate = mandoc_strdup(n->string);
 437                 }
 438         } else
 439                 m->meta.date = time(NULL);
 440
 441         /* TITLE MSEC DATE ->SOURCE<- VOL */
 442
 443         if (n && (n = n->next))
 444                 m->meta.source = mandoc_strdup(n->string);
 445
 446         /* TITLE MSEC DATE SOURCE ->VOL<- */
 447
 448         if (n && (n = n->next))
 449                 m->meta.vol = mandoc_strdup(n->string);
 450
 451         /*
 452          * Remove the `TH' node after we've processed it for our
 453          * meta-data.
 454          */
 455         man_node_delete(m, m->last);
 456         return(1);
 457 }
 458
 459 static int
 460 post_nf(CHKARGS)
 461 {
 462
 463         if (MAN_LITERAL & m->flags)
 464                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 465
 466         m->flags |= MAN_LITERAL;
 467         return(1);
 468 }
 469
 470 static int
 471 post_fi(CHKARGS)
 472 {
 473
 474         if ( ! (MAN_LITERAL & m->flags))
 475                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 476
 477         m->flags &= ~MAN_LITERAL;
 478         return(1);
 479 }
 480
 481 static int
 482 post_UC(CHKARGS)
 483 {
 484         static const char * const bsd_versions[] = {
 485             "3rd Berkeley Distribution",
 486             "4th Berkeley Distribution",
 487             "4.2 Berkeley Distribution",
 488             "4.3 Berkeley Distribution",
 489             "4.4 Berkeley Distribution",
 490         };
 491
 492         const char      *p, *s;
 493
 494         n = n->child;
 495         n = m->last->child;
 496
 497         if (NULL == n || MAN_TEXT != n->type)
 498                 p = bsd_versions[0];
 499         else {
 500                 s = n->string;
 501                 if (0 == strcmp(s, "3"))
 502                         p = bsd_versions[0];
 503                 else if (0 == strcmp(s, "4"))
 504                         p = bsd_versions[1];
 505                 else if (0 == strcmp(s, "5"))
 506                         p = bsd_versions[2];
 507                 else if (0 == strcmp(s, "6"))
 508                         p = bsd_versions[3];
 509                 else if (0 == strcmp(s, "7"))
 510                         p = bsd_versions[4];
 511                 else
 512                         p = bsd_versions[0];
 513         }
 514
 515         if (m->meta.source)
 516                 free(m->meta.source);
 517
 518         m->meta.source = mandoc_strdup(p);
 519         return(1);
 520 }
 521
 522 static int
 523 post_AT(CHKARGS)
 524 {
 525         static const char * const unix_versions[] = {
 526             "7th Edition",
 527             "System III",
 528             "System V",
 529             "System V Release 2",
 530         };
 531
 532         const char      *p, *s;
 533         struct man_node *nn;
 534
 535         n = n->child;
 536
 537         if (NULL == n || MAN_TEXT != n->type)
 538                 p = unix_versions[0];
 539         else {
 540                 s = n->string;
 541                 if (0 == strcmp(s, "3"))
 542                         p = unix_versions[0];
 543                 else if (0 == strcmp(s, "4"))
 544                         p = unix_versions[1];
 545                 else if (0 == strcmp(s, "5")) {
 546                         nn = n->next;
 547                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 548                                 p = unix_versions[3];
 549                         else
 550                                 p = unix_versions[2];
 551                 } else
 552                         p = unix_versions[0];
 553         }
 554
 555         if (m->meta.source)
 556                 free(m->meta.source);
 557
 558         m->meta.source = mandoc_strdup(p);
 559         return(1);
 560 }