man_validate.c

   1 /*      $Id: man_validate.c,v 1.74 2011/09/04 09:58:40 kristaps 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 #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 "libman.h"
  36 #include "libmandoc.h"
  37
  38 #define CHKARGS   struct man *m, struct man_node *n
  39
  40 typedef int     (*v_check)(CHKARGS);
  41
  42 struct  man_valid {
  43         v_check  *pres;
  44         v_check  *posts;
  45 };
  46
  47 static  int       check_bline(CHKARGS);
  48 static  int       check_eq0(CHKARGS);
  49 static  int       check_le1(CHKARGS);
  50 static  int       check_ge2(CHKARGS);
  51 static  int       check_le5(CHKARGS);
  52 static  int       check_par(CHKARGS);
  53 static  int       check_part(CHKARGS);
  54 static  int       check_root(CHKARGS);
  55
  56 static  int       post_AT(CHKARGS);
  57 static  int       post_vs(CHKARGS);
  58 static  int       post_fi(CHKARGS);
  59 static  int       post_ft(CHKARGS);
  60 static  int       post_nf(CHKARGS);
  61 static  int       post_sec(CHKARGS);
  62 static  int       post_TH(CHKARGS);
  63 static  int       post_UC(CHKARGS);
  64 static  int       pre_sec(CHKARGS);
  65
  66 static  v_check   posts_at[] = { post_AT, NULL };
  67 static  v_check   posts_br[] = { post_vs, check_eq0, NULL };
  68 static  v_check   posts_eq0[] = { check_eq0, NULL };
  69 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  70 static  v_check   posts_ft[] = { post_ft, NULL };
  71 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  72 static  v_check   posts_par[] = { check_par, NULL };
  73 static  v_check   posts_part[] = { check_part, NULL };
  74 static  v_check   posts_sec[] = { post_sec, NULL };
  75 static  v_check   posts_sp[] = { post_vs, check_le1, NULL };
  76 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  77 static  v_check   posts_uc[] = { post_UC, NULL };
  78 static  v_check   pres_bline[] = { check_bline, NULL };
  79 static  v_check   pres_sec[] = { check_bline, pre_sec, NULL};
  80
  81 static  const struct man_valid man_valids[MAN_MAX] = {
  82         { NULL, posts_br }, /* br */
  83         { pres_bline, posts_th }, /* TH */
  84         { pres_sec, posts_sec }, /* SH */
  85         { pres_sec, posts_sec }, /* SS */
  86         { pres_bline, NULL }, /* TP */
  87         { pres_bline, posts_par }, /* LP */
  88         { pres_bline, posts_par }, /* PP */
  89         { pres_bline, posts_par }, /* P */
  90         { pres_bline, NULL }, /* IP */
  91         { pres_bline, NULL }, /* HP */
  92         { NULL, NULL }, /* SM */
  93         { NULL, NULL }, /* SB */
  94         { NULL, NULL }, /* BI */
  95         { NULL, NULL }, /* IB */
  96         { NULL, NULL }, /* BR */
  97         { NULL, NULL }, /* RB */
  98         { NULL, NULL }, /* R */
  99         { NULL, NULL }, /* B */
 100         { NULL, NULL }, /* I */
 101         { NULL, NULL }, /* IR */
 102         { NULL, NULL }, /* RI */
 103         { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
 104         { NULL, posts_sp }, /* sp */ /* FIXME: should warn only. */
 105         { pres_bline, posts_nf }, /* nf */
 106         { pres_bline, posts_fi }, /* fi */
 107         { NULL, NULL }, /* RE */
 108         { NULL, posts_part }, /* RS */
 109         { NULL, NULL }, /* DT */
 110         { NULL, posts_uc }, /* UC */
 111         { NULL, NULL }, /* PD */
 112         { NULL, posts_at }, /* AT */
 113         { NULL, NULL }, /* in */
 114         { NULL, posts_ft }, /* ft */
 115 };
 116
 117
 118 int
 119 man_valid_pre(struct man *m, struct man_node *n)
 120 {
 121         v_check         *cp;
 122
 123         switch (n->type) {
 124         case (MAN_TEXT):
 125                 /* FALLTHROUGH */
 126         case (MAN_ROOT):
 127                 /* FALLTHROUGH */
 128         case (MAN_EQN):
 129                 /* FALLTHROUGH */
 130         case (MAN_TBL):
 131                 return(1);
 132         default:
 133                 break;
 134         }
 135
 136         if (NULL == (cp = man_valids[n->tok].pres))
 137                 return(1);
 138         for ( ; *cp; cp++)
 139                 if ( ! (*cp)(m, n))
 140                         return(0);
 141         return(1);
 142 }
 143
 144
 145 int
 146 man_valid_post(struct man *m)
 147 {
 148         v_check         *cp;
 149
 150         if (MAN_VALID & m->last->flags)
 151                 return(1);
 152         m->last->flags |= MAN_VALID;
 153
 154         switch (m->last->type) {
 155         case (MAN_ROOT):
 156                 return(check_root(m, m->last));
 157         case (MAN_TEXT):
 158                 /* FALLTHROUGH */
 159         case (MAN_EQN):
 160                 /* FALLTHROUGH */
 161         case (MAN_TBL):
 162                 return(1);
 163         default:
 164                 break;
 165         }
 166
 167         if (NULL == (cp = man_valids[m->last->tok].posts))
 168                 return(1);
 169         for ( ; *cp; cp++)
 170                 if ( ! (*cp)(m, m->last))
 171                         return(0);
 172
 173         return(1);
 174 }
 175
 176
 177 static int
 178 check_root(CHKARGS)
 179 {
 180
 181         if (MAN_BLINE & m->flags)
 182                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 183         else if (MAN_ELINE & m->flags)
 184                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 185
 186         m->flags &= ~MAN_BLINE;
 187         m->flags &= ~MAN_ELINE;
 188
 189         if (NULL == m->first->child) {
 190                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 191                 return(0);
 192         } else if (NULL == m->meta.title) {
 193                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 194
 195                 /*
 196                  * If a title hasn't been set, do so now (by
 197                  * implication, date and section also aren't set).
 198                  */
 199
 200                 m->meta.title = mandoc_strdup("unknown");
 201                 m->meta.msec = mandoc_strdup("1");
 202                 m->meta.date = mandoc_normdate
 203                         (m->parse, NULL, n->line, n->pos);
 204         }
 205
 206         return(1);
 207 }
 208
 209 #define INEQ_DEFINE(x, ineq, name) \
 210 static int \
 211 check_##name(CHKARGS) \
 212 { \
 213         if (n->nchild ineq (x)) \
 214                 return(1); \
 215         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 216                         "line arguments %s %d (have %d)", \
 217                         #ineq, (x), n->nchild); \
 218         return(1); \
 219 }
 220
 221 INEQ_DEFINE(0, ==, eq0)
 222 INEQ_DEFINE(1, <=, le1)
 223 INEQ_DEFINE(2, >=, ge2)
 224 INEQ_DEFINE(5, <=, le5)
 225
 226 static int
 227 post_ft(CHKARGS)
 228 {
 229         char    *cp;
 230         int      ok;
 231
 232         if (0 == n->nchild)
 233                 return(1);
 234
 235         ok = 0;
 236         cp = n->child->string;
 237         switch (*cp) {
 238         case ('1'):
 239                 /* FALLTHROUGH */
 240         case ('2'):
 241                 /* FALLTHROUGH */
 242         case ('3'):
 243                 /* FALLTHROUGH */
 244         case ('4'):
 245                 /* FALLTHROUGH */
 246         case ('I'):
 247                 /* FALLTHROUGH */
 248         case ('P'):
 249                 /* FALLTHROUGH */
 250         case ('R'):
 251                 if ('\0' == cp[1])
 252                         ok = 1;
 253                 break;
 254         case ('B'):
 255                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 256                         ok = 1;
 257                 break;
 258         case ('C'):
 259                 if ('W' == cp[1] && '\0' == cp[2])
 260                         ok = 1;
 261                 break;
 262         default:
 263                 break;
 264         }
 265
 266         if (0 == ok) {
 267                 mandoc_vmsg
 268                         (MANDOCERR_BADFONT, m->parse,
 269                          n->line, n->pos, "%s", cp);
 270                 *cp = '\0';
 271         }
 272
 273         if (1 < n->nchild)
 274                 mandoc_vmsg
 275                         (MANDOCERR_ARGCOUNT, m->parse, n->line,
 276                          n->pos, "want one child (have %d)",
 277                          n->nchild);
 278
 279         return(1);
 280 }
 281
 282 static int
 283 pre_sec(CHKARGS)
 284 {
 285
 286         if (MAN_BLOCK == n->type)
 287                 m->flags &= ~MAN_LITERAL;
 288         return(1);
 289 }
 290
 291 static int
 292 post_sec(CHKARGS)
 293 {
 294
 295         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 296                 return(1);
 297
 298         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 299         return(0);
 300 }
 301
 302 static int
 303 check_part(CHKARGS)
 304 {
 305
 306         if (MAN_BODY == n->type && 0 == n->nchild)
 307                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 308                                 n->pos, "want children (have none)");
 309
 310         return(1);
 311 }
 312
 313
 314 static int
 315 check_par(CHKARGS)
 316 {
 317
 318         switch (n->type) {
 319         case (MAN_BLOCK):
 320                 if (0 == n->body->nchild)
 321                         man_node_delete(m, n);
 322                 break;
 323         case (MAN_BODY):
 324                 if (0 == n->nchild)
 325                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 326                 break;
 327         case (MAN_HEAD):
 328                 if (n->nchild)
 329                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 330                 break;
 331         default:
 332                 break;
 333         }
 334
 335         return(1);
 336 }
 337
 338
 339 static int
 340 check_bline(CHKARGS)
 341 {
 342
 343         assert( ! (MAN_ELINE & m->flags));
 344         if (MAN_BLINE & m->flags) {
 345                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 346                 return(0);
 347         }
 348
 349         return(1);
 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)
 404                 pos = n->pos;
 405         m->meta.date = mandoc_normdate
 406                 (m->parse, n ? n->string : NULL, line, pos);
 407
 408         /* TITLE MSEC DATE ->SOURCE<- VOL */
 409
 410         if (n && (n = n->next))
 411                 m->meta.source = mandoc_strdup(n->string);
 412
 413         /* TITLE MSEC DATE SOURCE ->VOL<- */
 414
 415         if (n && (n = n->next))
 416                 m->meta.vol = mandoc_strdup(n->string);
 417
 418         /*
 419          * Remove the `TH' node after we've processed it for our
 420          * meta-data.
 421          */
 422         man_node_delete(m, m->last);
 423         return(1);
 424 }
 425
 426 static int
 427 post_nf(CHKARGS)
 428 {
 429
 430         if (MAN_LITERAL & m->flags)
 431                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 432
 433         m->flags |= MAN_LITERAL;
 434         return(1);
 435 }
 436
 437 static int
 438 post_fi(CHKARGS)
 439 {
 440
 441         if ( ! (MAN_LITERAL & m->flags))
 442                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 443
 444         m->flags &= ~MAN_LITERAL;
 445         return(1);
 446 }
 447
 448 static int
 449 post_UC(CHKARGS)
 450 {
 451         static const char * const bsd_versions[] = {
 452             "3rd Berkeley Distribution",
 453             "4th Berkeley Distribution",
 454             "4.2 Berkeley Distribution",
 455             "4.3 Berkeley Distribution",
 456             "4.4 Berkeley Distribution",
 457         };
 458
 459         const char      *p, *s;
 460
 461         n = n->child;
 462         n = m->last->child;
 463
 464         if (NULL == n || MAN_TEXT != n->type)
 465                 p = bsd_versions[0];
 466         else {
 467                 s = n->string;
 468                 if (0 == strcmp(s, "3"))
 469                         p = bsd_versions[0];
 470                 else if (0 == strcmp(s, "4"))
 471                         p = bsd_versions[1];
 472                 else if (0 == strcmp(s, "5"))
 473                         p = bsd_versions[2];
 474                 else if (0 == strcmp(s, "6"))
 475                         p = bsd_versions[3];
 476                 else if (0 == strcmp(s, "7"))
 477                         p = bsd_versions[4];
 478                 else
 479                         p = bsd_versions[0];
 480         }
 481
 482         if (m->meta.source)
 483                 free(m->meta.source);
 484
 485         m->meta.source = mandoc_strdup(p);
 486         return(1);
 487 }
 488
 489 static int
 490 post_AT(CHKARGS)
 491 {
 492         static const char * const unix_versions[] = {
 493             "7th Edition",
 494             "System III",
 495             "System V",
 496             "System V Release 2",
 497         };
 498
 499         const char      *p, *s;
 500         struct man_node *nn;
 501
 502         n = n->child;
 503
 504         if (NULL == n || MAN_TEXT != n->type)
 505                 p = unix_versions[0];
 506         else {
 507                 s = n->string;
 508                 if (0 == strcmp(s, "3"))
 509                         p = unix_versions[0];
 510                 else if (0 == strcmp(s, "4"))
 511                         p = unix_versions[1];
 512                 else if (0 == strcmp(s, "5")) {
 513                         nn = n->next;
 514                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 515                                 p = unix_versions[3];
 516                         else
 517                                 p = unix_versions[2];
 518                 } else
 519                         p = unix_versions[0];
 520         }
 521
 522         if (m->meta.source)
 523                 free(m->meta.source);
 524
 525         m->meta.source = mandoc_strdup(p);
 526         return(1);
 527 }
 528
 529 static int
 530 post_vs(CHKARGS)
 531 {
 532
 533         /*
 534          * Don't warn about this because it occurs in pod2man and would
 535          * cause considerable (unfixable) warnage.
 536          */
 537         if (NULL == n->prev && MAN_ROOT == n->parent->type)
 538                 man_node_delete(m, n);
 539
 540         return(1);
 541 }