1 /*      $Id: man_validate.c,v 1.80 2012/01/03 15:16:24 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_eq0(CHKARGS);
  48 static  int       check_eq2(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 static  void      check_text(CHKARGS);
  56 
  57 static  int       post_AT(CHKARGS);
  58 static  int       post_vs(CHKARGS);
  59 static  int       post_fi(CHKARGS);
  60 static  int       post_ft(CHKARGS);
  61 static  int       post_nf(CHKARGS);
  62 static  int       post_sec(CHKARGS);
  63 static  int       post_TH(CHKARGS);
  64 static  int       post_UC(CHKARGS);
  65 static  int       pre_sec(CHKARGS);
  66 
  67 static  v_check   posts_at[] = { post_AT, NULL };
  68 static  v_check   posts_br[] = { post_vs, check_eq0, NULL };
  69 static  v_check   posts_eq0[] = { check_eq0, NULL };
  70 static  v_check   posts_eq2[] = { check_eq2, NULL };
  71 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  72 static  v_check   posts_ft[] = { post_ft, NULL };
  73 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  74 static  v_check   posts_par[] = { check_par, NULL };
  75 static  v_check   posts_part[] = { check_part, NULL };
  76 static  v_check   posts_sec[] = { post_sec, NULL };
  77 static  v_check   posts_sp[] = { post_vs, check_le1, NULL };
  78 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  79 static  v_check   posts_uc[] = { post_UC, NULL };
  80 static  v_check   pres_sec[] = { pre_sec, NULL };
  81 
  82 static  const struct man_valid man_valids[MAN_MAX] = {
  83         { NULL, posts_br }, /* br */
  84         { NULL, posts_th }, /* TH */
  85         { pres_sec, posts_sec }, /* SH */
  86         { pres_sec, posts_sec }, /* SS */
  87         { NULL, NULL }, /* TP */
  88         { NULL, posts_par }, /* LP */
  89         { NULL, posts_par }, /* PP */
  90         { NULL, posts_par }, /* P */
  91         { NULL, NULL }, /* IP */
  92         { NULL, NULL }, /* HP */
  93         { NULL, NULL }, /* SM */
  94         { NULL, NULL }, /* SB */
  95         { NULL, NULL }, /* BI */
  96         { NULL, NULL }, /* IB */
  97         { NULL, NULL }, /* BR */
  98         { NULL, NULL }, /* RB */
  99         { NULL, NULL }, /* R */
 100         { NULL, NULL }, /* B */
 101         { NULL, NULL }, /* I */
 102         { NULL, NULL }, /* IR */
 103         { NULL, NULL }, /* RI */
 104         { NULL, posts_eq0 }, /* na */
 105         { NULL, posts_sp }, /* sp */
 106         { NULL, posts_nf }, /* nf */
 107         { NULL, posts_fi }, /* fi */
 108         { NULL, NULL }, /* RE */
 109         { NULL, posts_part }, /* RS */
 110         { NULL, NULL }, /* DT */
 111         { NULL, posts_uc }, /* UC */
 112         { NULL, NULL }, /* PD */
 113         { NULL, posts_at }, /* AT */
 114         { NULL, NULL }, /* in */
 115         { NULL, posts_ft }, /* ft */
 116         { NULL, posts_eq2 }, /* OP */
 117 };
 118 
 119 
 120 int
 121 man_valid_pre(struct man *m, struct man_node *n)
 122 {
 123         v_check         *cp;
 124 
 125         switch (n->type) {
 126         case (MAN_TEXT):
 127                 /* FALLTHROUGH */
 128         case (MAN_ROOT):
 129                 /* FALLTHROUGH */
 130         case (MAN_EQN):
 131                 /* FALLTHROUGH */
 132         case (MAN_TBL):
 133                 return(1);
 134         default:
 135                 break;
 136         }
 137 
 138         if (NULL == (cp = man_valids[n->tok].pres))
 139                 return(1);
 140         for ( ; *cp; cp++)
 141                 if ( ! (*cp)(m, n)) 
 142                         return(0);
 143         return(1);
 144 }
 145 
 146 
 147 int
 148 man_valid_post(struct man *m)
 149 {
 150         v_check         *cp;
 151 
 152         if (MAN_VALID & m->last->flags)
 153                 return(1);
 154         m->last->flags |= MAN_VALID;
 155 
 156         switch (m->last->type) {
 157         case (MAN_TEXT): 
 158                 check_text(m, m->last);
 159                 return(1);
 160         case (MAN_ROOT):
 161                 return(check_root(m, m->last));
 162         case (MAN_EQN):
 163                 /* FALLTHROUGH */
 164         case (MAN_TBL):
 165                 return(1);
 166         default:
 167                 break;
 168         }
 169 
 170         if (NULL == (cp = man_valids[m->last->tok].posts))
 171                 return(1);
 172         for ( ; *cp; cp++)
 173                 if ( ! (*cp)(m, m->last))
 174                         return(0);
 175 
 176         return(1);
 177 }
 178 
 179 
 180 static int
 181 check_root(CHKARGS) 
 182 {
 183 
 184         if (MAN_BLINE & m->flags)
 185                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 186         else if (MAN_ELINE & m->flags)
 187                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 188 
 189         m->flags &= ~MAN_BLINE;
 190         m->flags &= ~MAN_ELINE;
 191 
 192         if (NULL == m->first->child) {
 193                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 194                 return(0);
 195         } else if (NULL == m->meta.title) {
 196                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 197 
 198                 /*
 199                  * If a title hasn't been set, do so now (by
 200                  * implication, date and section also aren't set).
 201                  */
 202 
 203                 m->meta.title = mandoc_strdup("unknown");
 204                 m->meta.msec = mandoc_strdup("1");
 205                 m->meta.date = mandoc_normdate
 206                         (m->parse, NULL, n->line, n->pos);
 207         }
 208 
 209         return(1);
 210 }
 211 
 212 static void
 213 check_text(CHKARGS)
 214 {
 215         char            *cp, *p;
 216 
 217         if (MAN_LITERAL & m->flags)
 218                 return;
 219 
 220         cp = n->string;
 221         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 222                 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 223 }
 224 
 225 #define INEQ_DEFINE(x, ineq, name) \
 226 static int \
 227 check_##name(CHKARGS) \
 228 { \
 229         if (n->nchild ineq (x)) \
 230                 return(1); \
 231         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 232                         "line arguments %s %d (have %d)", \
 233                         #ineq, (x), n->nchild); \
 234         return(1); \
 235 }
 236 
 237 INEQ_DEFINE(0, ==, eq0)
 238 INEQ_DEFINE(2, ==, eq2)
 239 INEQ_DEFINE(1, <=, le1)
 240 INEQ_DEFINE(2, >=, ge2)
 241 INEQ_DEFINE(5, <=, le5)
 242 
 243 static int
 244 post_ft(CHKARGS)
 245 {
 246         char    *cp;
 247         int      ok;
 248 
 249         if (0 == n->nchild)
 250                 return(1);
 251 
 252         ok = 0;
 253         cp = n->child->string;
 254         switch (*cp) {
 255         case ('1'):
 256                 /* FALLTHROUGH */
 257         case ('2'):
 258                 /* FALLTHROUGH */
 259         case ('3'):
 260                 /* FALLTHROUGH */
 261         case ('4'):
 262                 /* FALLTHROUGH */
 263         case ('I'):
 264                 /* FALLTHROUGH */
 265         case ('P'):
 266                 /* FALLTHROUGH */
 267         case ('R'):
 268                 if ('\0' == cp[1])
 269                         ok = 1;
 270                 break;
 271         case ('B'):
 272                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 273                         ok = 1;
 274                 break;
 275         case ('C'):
 276                 if ('W' == cp[1] && '\0' == cp[2])
 277                         ok = 1;
 278                 break;
 279         default:
 280                 break;
 281         }
 282 
 283         if (0 == ok) {
 284                 mandoc_vmsg
 285                         (MANDOCERR_BADFONT, m->parse,
 286                          n->line, n->pos, "%s", cp);
 287                 *cp = '\0';
 288         }
 289 
 290         if (1 < n->nchild)
 291                 mandoc_vmsg
 292                         (MANDOCERR_ARGCOUNT, m->parse, n->line, 
 293                          n->pos, "want one child (have %d)", 
 294                          n->nchild);
 295 
 296         return(1);
 297 }
 298 
 299 static int
 300 pre_sec(CHKARGS)
 301 {
 302 
 303         if (MAN_BLOCK == n->type)
 304                 m->flags &= ~MAN_LITERAL;
 305         return(1);
 306 }
 307 
 308 static int
 309 post_sec(CHKARGS)
 310 {
 311 
 312         if ( ! (MAN_HEAD == n->type && 0 == n->nchild)) 
 313                 return(1);
 314 
 315         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 316         return(0);
 317 }
 318 
 319 static int
 320 check_part(CHKARGS)
 321 {
 322 
 323         if (MAN_BODY == n->type && 0 == n->nchild)
 324                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line, 
 325                                 n->pos, "want children (have none)");
 326 
 327         return(1);
 328 }
 329 
 330 
 331 static int
 332 check_par(CHKARGS)
 333 {
 334 
 335         switch (n->type) {
 336         case (MAN_BLOCK):
 337                 if (0 == n->body->nchild)
 338                         man_node_delete(m, n);
 339                 break;
 340         case (MAN_BODY):
 341                 if (0 == n->nchild)
 342                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 343                 break;
 344         case (MAN_HEAD):
 345                 if (n->nchild)
 346                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 347                 break;
 348         default:
 349                 break;
 350         }
 351 
 352         return(1);
 353 }
 354 
 355 
 356 static int
 357 post_TH(CHKARGS)
 358 {
 359         const char      *p;
 360         int              line, pos;
 361 
 362         if (m->meta.title)
 363                 free(m->meta.title);
 364         if (m->meta.vol)
 365                 free(m->meta.vol);
 366         if (m->meta.source)
 367                 free(m->meta.source);
 368         if (m->meta.msec)
 369                 free(m->meta.msec);
 370         if (m->meta.date)
 371                 free(m->meta.date);
 372 
 373         line = n->line;
 374         pos = n->pos;
 375         m->meta.title = m->meta.vol = m->meta.date =
 376                 m->meta.msec = m->meta.source = NULL;
 377 
 378         /* ->TITLE<- MSEC DATE SOURCE VOL */
 379 
 380         n = n->child;
 381         if (n && n->string) {
 382                 for (p = n->string; '\0' != *p; p++) {
 383                         /* Only warn about this once... */
 384                         if (isalpha((unsigned char)*p) && 
 385                                         ! isupper((unsigned char)*p)) {
 386                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 387                                 break;
 388                         }
 389                 }
 390                 m->meta.title = mandoc_strdup(n->string);
 391         } else
 392                 m->meta.title = mandoc_strdup("");
 393 
 394         /* TITLE ->MSEC<- DATE SOURCE VOL */
 395 
 396         if (n)
 397                 n = n->next;
 398         if (n && n->string)
 399                 m->meta.msec = mandoc_strdup(n->string);
 400         else
 401                 m->meta.msec = mandoc_strdup("");
 402 
 403         /* TITLE MSEC ->DATE<- SOURCE VOL */
 404 
 405         if (n)
 406                 n = n->next;
 407         if (n && n->string && '\0' != n->string[0]) {
 408                 pos = n->pos;
 409                 m->meta.date = mandoc_normdate
 410                     (m->parse, n->string, line, pos);
 411         } else
 412                 m->meta.date = mandoc_strdup("");
 413 
 414         /* TITLE MSEC DATE ->SOURCE<- VOL */
 415 
 416         if (n && (n = n->next))
 417                 m->meta.source = mandoc_strdup(n->string);
 418 
 419         /* TITLE MSEC DATE SOURCE ->VOL<- */
 420         /* If missing, use the default VOL name for MSEC. */
 421 
 422         if (n && (n = n->next))
 423                 m->meta.vol = mandoc_strdup(n->string);
 424         else if ('\0' != m->meta.msec[0] &&
 425             (NULL != (p = mandoc_a2msec(m->meta.msec))))
 426                 m->meta.vol = mandoc_strdup(p);
 427 
 428         /*
 429          * Remove the `TH' node after we've processed it for our
 430          * meta-data.
 431          */
 432         man_node_delete(m, m->last);
 433         return(1);
 434 }
 435 
 436 static int
 437 post_nf(CHKARGS)
 438 {
 439 
 440         if (MAN_LITERAL & m->flags)
 441                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 442 
 443         m->flags |= MAN_LITERAL;
 444         return(1);
 445 }
 446 
 447 static int
 448 post_fi(CHKARGS)
 449 {
 450 
 451         if ( ! (MAN_LITERAL & m->flags))
 452                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 453 
 454         m->flags &= ~MAN_LITERAL;
 455         return(1);
 456 }
 457 
 458 static int
 459 post_UC(CHKARGS)
 460 {
 461         static const char * const bsd_versions[] = {
 462             "3rd Berkeley Distribution",
 463             "4th Berkeley Distribution",
 464             "4.2 Berkeley Distribution",
 465             "4.3 Berkeley Distribution",
 466             "4.4 Berkeley Distribution",
 467         };
 468 
 469         const char      *p, *s;
 470 
 471         n = n->child;
 472 
 473         if (NULL == n || MAN_TEXT != n->type)
 474                 p = bsd_versions[0];
 475         else {
 476                 s = n->string;
 477                 if (0 == strcmp(s, "3"))
 478                         p = bsd_versions[0];
 479                 else if (0 == strcmp(s, "4"))
 480                         p = bsd_versions[1];
 481                 else if (0 == strcmp(s, "5"))
 482                         p = bsd_versions[2];
 483                 else if (0 == strcmp(s, "6"))
 484                         p = bsd_versions[3];
 485                 else if (0 == strcmp(s, "7"))
 486                         p = bsd_versions[4];
 487                 else
 488                         p = bsd_versions[0];
 489         }
 490 
 491         if (m->meta.source)
 492                 free(m->meta.source);
 493 
 494         m->meta.source = mandoc_strdup(p);
 495         return(1);
 496 }
 497 
 498 static int
 499 post_AT(CHKARGS)
 500 {
 501         static const char * const unix_versions[] = {
 502             "7th Edition",
 503             "System III",
 504             "System V",
 505             "System V Release 2",
 506         };
 507 
 508         const char      *p, *s;
 509         struct man_node *nn;
 510 
 511         n = n->child;
 512 
 513         if (NULL == n || MAN_TEXT != n->type)
 514                 p = unix_versions[0];
 515         else {
 516                 s = n->string;
 517                 if (0 == strcmp(s, "3"))
 518                         p = unix_versions[0];
 519                 else if (0 == strcmp(s, "4"))
 520                         p = unix_versions[1];
 521                 else if (0 == strcmp(s, "5")) {
 522                         nn = n->next;
 523                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 524                                 p = unix_versions[3];
 525                         else
 526                                 p = unix_versions[2];
 527                 } else
 528                         p = unix_versions[0];
 529         }
 530 
 531         if (m->meta.source)
 532                 free(m->meta.source);
 533 
 534         m->meta.source = mandoc_strdup(p);
 535         return(1);
 536 }
 537 
 538 static int
 539 post_vs(CHKARGS)
 540 {
 541 
 542         /* 
 543          * Don't warn about this because it occurs in pod2man and would
 544          * cause considerable (unfixable) warnage.
 545          */
 546         if (NULL == n->prev && MAN_ROOT == n->parent->type)
 547                 man_node_delete(m, n);
 548 
 549         return(1);
 550 }