man_validate.c

   1 /*      $Id: man_validate.c,v 1.72 2011/07/26 14:09:01 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_ft(CHKARGS);
  50 static  int       check_le1(CHKARGS);
  51 static  int       check_ge2(CHKARGS);
  52 static  int       check_le5(CHKARGS);
  53 static  int       check_par(CHKARGS);
  54 static  int       check_part(CHKARGS);
  55 static  int       check_root(CHKARGS);
  56 static  int       check_sec(CHKARGS);
  57 static  void      check_text(CHKARGS);
  58
  59 static  int       post_AT(CHKARGS);
  60 static  int       post_vs(CHKARGS);
  61 static  int       post_fi(CHKARGS);
  62 static  int       post_nf(CHKARGS);
  63 static  int       post_TH(CHKARGS);
  64 static  int       post_UC(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[] = { check_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[] = { check_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
  80 static  const struct man_valid man_valids[MAN_MAX] = {
  81         { NULL, posts_br }, /* br */
  82         { pres_bline, posts_th }, /* TH */
  83         { pres_bline, posts_sec }, /* SH */
  84         { pres_bline, posts_sec }, /* SS */
  85         { pres_bline, NULL }, /* TP */
  86         { pres_bline, posts_par }, /* LP */
  87         { pres_bline, posts_par }, /* PP */
  88         { pres_bline, posts_par }, /* P */
  89         { pres_bline, NULL }, /* IP */
  90         { pres_bline, NULL }, /* HP */
  91         { NULL, NULL }, /* SM */
  92         { NULL, NULL }, /* SB */
  93         { NULL, NULL }, /* BI */
  94         { NULL, NULL }, /* IB */
  95         { NULL, NULL }, /* BR */
  96         { NULL, NULL }, /* RB */
  97         { NULL, NULL }, /* R */
  98         { NULL, NULL }, /* B */
  99         { NULL, NULL }, /* I */
 100         { NULL, NULL }, /* IR */
 101         { NULL, NULL }, /* RI */
 102         { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
 103         { NULL, posts_sp }, /* sp */ /* FIXME: should warn only. */
 104         { pres_bline, posts_nf }, /* nf */
 105         { pres_bline, posts_fi }, /* fi */
 106         { NULL, NULL }, /* RE */
 107         { NULL, posts_part }, /* RS */
 108         { NULL, NULL }, /* DT */
 109         { NULL, posts_uc }, /* UC */
 110         { NULL, NULL }, /* PD */
 111         { NULL, posts_at }, /* AT */
 112         { NULL, NULL }, /* in */
 113         { NULL, posts_ft }, /* ft */
 114 };
 115
 116
 117 int
 118 man_valid_pre(struct man *m, struct man_node *n)
 119 {
 120         v_check         *cp;
 121
 122         switch (n->type) {
 123         case (MAN_TEXT):
 124                 /* FALLTHROUGH */
 125         case (MAN_ROOT):
 126                 /* FALLTHROUGH */
 127         case (MAN_EQN):
 128                 /* FALLTHROUGH */
 129         case (MAN_TBL):
 130                 return(1);
 131         default:
 132                 break;
 133         }
 134
 135         if (NULL == (cp = man_valids[n->tok].pres))
 136                 return(1);
 137         for ( ; *cp; cp++)
 138                 if ( ! (*cp)(m, n))
 139                         return(0);
 140         return(1);
 141 }
 142
 143
 144 int
 145 man_valid_post(struct man *m)
 146 {
 147         v_check         *cp;
 148
 149         if (MAN_VALID & m->last->flags)
 150                 return(1);
 151         m->last->flags |= MAN_VALID;
 152
 153         switch (m->last->type) {
 154         case (MAN_TEXT):
 155                 check_text(m, m->last);
 156                 return(1);
 157         case (MAN_ROOT):
 158                 return(check_root(m, m->last));
 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 static void
 210 check_text(CHKARGS)
 211 {
 212         char            *cp, *p;
 213
 214         cp = p = n->string;
 215         for (cp = p; NULL != (p = strchr(p, '\t')); p++) {
 216                 if (MAN_LITERAL & m->flags)
 217                         continue;
 218                 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 219         }
 220 }
 221
 222 #define INEQ_DEFINE(x, ineq, name) \
 223 static int \
 224 check_##name(CHKARGS) \
 225 { \
 226         if (n->nchild ineq (x)) \
 227                 return(1); \
 228         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 229                         "line arguments %s %d (have %d)", \
 230                         #ineq, (x), n->nchild); \
 231         return(1); \
 232 }
 233
 234 INEQ_DEFINE(0, ==, eq0)
 235 INEQ_DEFINE(1, <=, le1)
 236 INEQ_DEFINE(2, >=, ge2)
 237 INEQ_DEFINE(5, <=, le5)
 238
 239 static int
 240 check_ft(CHKARGS)
 241 {
 242         char    *cp;
 243         int      ok;
 244
 245         if (0 == n->nchild)
 246                 return(1);
 247
 248         ok = 0;
 249         cp = n->child->string;
 250         switch (*cp) {
 251         case ('1'):
 252                 /* FALLTHROUGH */
 253         case ('2'):
 254                 /* FALLTHROUGH */
 255         case ('3'):
 256                 /* FALLTHROUGH */
 257         case ('4'):
 258                 /* FALLTHROUGH */
 259         case ('I'):
 260                 /* FALLTHROUGH */
 261         case ('P'):
 262                 /* FALLTHROUGH */
 263         case ('R'):
 264                 if ('\0' == cp[1])
 265                         ok = 1;
 266                 break;
 267         case ('B'):
 268                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 269                         ok = 1;
 270                 break;
 271         case ('C'):
 272                 if ('W' == cp[1] && '\0' == cp[2])
 273                         ok = 1;
 274                 break;
 275         default:
 276                 break;
 277         }
 278
 279         if (0 == ok) {
 280                 mandoc_vmsg
 281                         (MANDOCERR_BADFONT, m->parse,
 282                          n->line, n->pos, "%s", cp);
 283                 *cp = '\0';
 284         }
 285
 286         if (1 < n->nchild)
 287                 mandoc_vmsg
 288                         (MANDOCERR_ARGCOUNT, m->parse, n->line,
 289                          n->pos, "want one child (have %d)",
 290                          n->nchild);
 291
 292         return(1);
 293 }
 294
 295 static int
 296 check_sec(CHKARGS)
 297 {
 298
 299         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 300                 return(1);
 301
 302         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 303         return(0);
 304 }
 305
 306
 307 static int
 308 check_part(CHKARGS)
 309 {
 310
 311         if (MAN_BODY == n->type && 0 == n->nchild)
 312                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 313                                 n->pos, "want children (have none)");
 314
 315         return(1);
 316 }
 317
 318
 319 static int
 320 check_par(CHKARGS)
 321 {
 322
 323         switch (n->type) {
 324         case (MAN_BLOCK):
 325                 if (0 == n->body->nchild)
 326                         man_node_delete(m, n);
 327                 break;
 328         case (MAN_BODY):
 329                 if (0 == n->nchild)
 330                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 331                 break;
 332         case (MAN_HEAD):
 333                 if (n->nchild)
 334                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 335                 break;
 336         default:
 337                 break;
 338         }
 339
 340         return(1);
 341 }
 342
 343
 344 static int
 345 check_bline(CHKARGS)
 346 {
 347
 348         assert( ! (MAN_ELINE & m->flags));
 349         if (MAN_BLINE & m->flags) {
 350                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 351                 return(0);
 352         }
 353
 354         return(1);
 355 }
 356
 357 static int
 358 post_TH(CHKARGS)
 359 {
 360         const char      *p;
 361         int              line, pos;
 362
 363         if (m->meta.title)
 364                 free(m->meta.title);
 365         if (m->meta.vol)
 366                 free(m->meta.vol);
 367         if (m->meta.source)
 368                 free(m->meta.source);
 369         if (m->meta.msec)
 370                 free(m->meta.msec);
 371         if (m->meta.date)
 372                 free(m->meta.date);
 373
 374         line = n->line;
 375         pos = n->pos;
 376         m->meta.title = m->meta.vol = m->meta.date =
 377                 m->meta.msec = m->meta.source = NULL;
 378
 379         /* ->TITLE<- MSEC DATE SOURCE VOL */
 380
 381         n = n->child;
 382         if (n && n->string) {
 383                 for (p = n->string; '\0' != *p; p++) {
 384                         /* Only warn about this once... */
 385                         if (isalpha((unsigned char)*p) &&
 386                                         ! isupper((unsigned char)*p)) {
 387                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 388                                 break;
 389                         }
 390                 }
 391                 m->meta.title = mandoc_strdup(n->string);
 392         } else
 393                 m->meta.title = mandoc_strdup("");
 394
 395         /* TITLE ->MSEC<- DATE SOURCE VOL */
 396
 397         if (n)
 398                 n = n->next;
 399         if (n && n->string)
 400                 m->meta.msec = mandoc_strdup(n->string);
 401         else
 402                 m->meta.msec = mandoc_strdup("");
 403
 404         /* TITLE MSEC ->DATE<- SOURCE VOL */
 405
 406         if (n)
 407                 n = n->next;
 408         if (n)
 409                 pos = n->pos;
 410         m->meta.date = mandoc_normdate
 411                 (m->parse, n ? n->string : NULL, line, pos);
 412
 413         /* TITLE MSEC DATE ->SOURCE<- VOL */
 414
 415         if (n && (n = n->next))
 416                 m->meta.source = mandoc_strdup(n->string);
 417
 418         /* TITLE MSEC DATE SOURCE ->VOL<- */
 419
 420         if (n && (n = n->next))
 421                 m->meta.vol = mandoc_strdup(n->string);
 422
 423         /*
 424          * Remove the `TH' node after we've processed it for our
 425          * meta-data.
 426          */
 427         man_node_delete(m, m->last);
 428         return(1);
 429 }
 430
 431 static int
 432 post_nf(CHKARGS)
 433 {
 434
 435         if (MAN_LITERAL & m->flags)
 436                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 437
 438         m->flags |= MAN_LITERAL;
 439         return(1);
 440 }
 441
 442 static int
 443 post_fi(CHKARGS)
 444 {
 445
 446         if ( ! (MAN_LITERAL & m->flags))
 447                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 448
 449         m->flags &= ~MAN_LITERAL;
 450         return(1);
 451 }
 452
 453 static int
 454 post_UC(CHKARGS)
 455 {
 456         static const char * const bsd_versions[] = {
 457             "3rd Berkeley Distribution",
 458             "4th Berkeley Distribution",
 459             "4.2 Berkeley Distribution",
 460             "4.3 Berkeley Distribution",
 461             "4.4 Berkeley Distribution",
 462         };
 463
 464         const char      *p, *s;
 465
 466         n = n->child;
 467         n = m->last->child;
 468
 469         if (NULL == n || MAN_TEXT != n->type)
 470                 p = bsd_versions[0];
 471         else {
 472                 s = n->string;
 473                 if (0 == strcmp(s, "3"))
 474                         p = bsd_versions[0];
 475                 else if (0 == strcmp(s, "4"))
 476                         p = bsd_versions[1];
 477                 else if (0 == strcmp(s, "5"))
 478                         p = bsd_versions[2];
 479                 else if (0 == strcmp(s, "6"))
 480                         p = bsd_versions[3];
 481                 else if (0 == strcmp(s, "7"))
 482                         p = bsd_versions[4];
 483                 else
 484                         p = bsd_versions[0];
 485         }
 486
 487         if (m->meta.source)
 488                 free(m->meta.source);
 489
 490         m->meta.source = mandoc_strdup(p);
 491         return(1);
 492 }
 493
 494 static int
 495 post_AT(CHKARGS)
 496 {
 497         static const char * const unix_versions[] = {
 498             "7th Edition",
 499             "System III",
 500             "System V",
 501             "System V Release 2",
 502         };
 503
 504         const char      *p, *s;
 505         struct man_node *nn;
 506
 507         n = n->child;
 508
 509         if (NULL == n || MAN_TEXT != n->type)
 510                 p = unix_versions[0];
 511         else {
 512                 s = n->string;
 513                 if (0 == strcmp(s, "3"))
 514                         p = unix_versions[0];
 515                 else if (0 == strcmp(s, "4"))
 516                         p = unix_versions[1];
 517                 else if (0 == strcmp(s, "5")) {
 518                         nn = n->next;
 519                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 520                                 p = unix_versions[3];
 521                         else
 522                                 p = unix_versions[2];
 523                 } else
 524                         p = unix_versions[0];
 525         }
 526
 527         if (m->meta.source)
 528                 free(m->meta.source);
 529
 530         m->meta.source = mandoc_strdup(p);
 531         return(1);
 532 }
 533
 534 static int
 535 post_vs(CHKARGS)
 536 {
 537
 538         /*
 539          * Don't warn about this because it occurs in pod2man and would
 540          * cause considerable (unfixable) warnage.
 541          */
 542         if (NULL == n->prev && MAN_ROOT == n->parent->type)
 543                 man_node_delete(m, n);
 544
 545         return(1);
 546 }