man_validate.c

   1 /*      $Id: man_validate.c,v 1.78 2011/11/07 01:24:40 schwarze 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_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  void      check_text(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_sec[] = { pre_sec, NULL };
  79
  80 static  const struct man_valid man_valids[MAN_MAX] = {
  81         { NULL, posts_br }, /* br */
  82         { NULL, posts_th }, /* TH */
  83         { pres_sec, posts_sec }, /* SH */
  84         { pres_sec, posts_sec }, /* SS */
  85         { NULL, NULL }, /* TP */
  86         { NULL, posts_par }, /* LP */
  87         { NULL, posts_par }, /* PP */
  88         { NULL, posts_par }, /* P */
  89         { NULL, NULL }, /* IP */
  90         { NULL, 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         { NULL, posts_nf }, /* nf */
 105         { NULL, 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         if (MAN_LITERAL & m->flags)
 215                 return;
 216
 217         cp = n->string;
 218         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 219                 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 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 post_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 pre_sec(CHKARGS)
 297 {
 298
 299         if (MAN_BLOCK == n->type)
 300                 m->flags &= ~MAN_LITERAL;
 301         return(1);
 302 }
 303
 304 static int
 305 post_sec(CHKARGS)
 306 {
 307
 308         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 309                 return(1);
 310
 311         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 312         return(0);
 313 }
 314
 315 static int
 316 check_part(CHKARGS)
 317 {
 318
 319         if (MAN_BODY == n->type && 0 == n->nchild)
 320                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 321                                 n->pos, "want children (have none)");
 322
 323         return(1);
 324 }
 325
 326
 327 static int
 328 check_par(CHKARGS)
 329 {
 330
 331         switch (n->type) {
 332         case (MAN_BLOCK):
 333                 if (0 == n->body->nchild)
 334                         man_node_delete(m, n);
 335                 break;
 336         case (MAN_BODY):
 337                 if (0 == n->nchild)
 338                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 339                 break;
 340         case (MAN_HEAD):
 341                 if (n->nchild)
 342                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 343                 break;
 344         default:
 345                 break;
 346         }
 347
 348         return(1);
 349 }
 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 && n->string && '\0' != n->string[0]) {
 404                 pos = n->pos;
 405                 m->meta.date = mandoc_normdate
 406                     (m->parse, n->string, line, pos);
 407         } else
 408                 m->meta.date = mandoc_strdup("");
 409
 410         /* TITLE MSEC DATE ->SOURCE<- VOL */
 411
 412         if (n && (n = n->next))
 413                 m->meta.source = mandoc_strdup(n->string);
 414
 415         /* TITLE MSEC DATE SOURCE ->VOL<- */
 416
 417         if (n && (n = n->next))
 418                 m->meta.vol = mandoc_strdup(n->string);
 419
 420         /*
 421          * Remove the `TH' node after we've processed it for our
 422          * meta-data.
 423          */
 424         man_node_delete(m, m->last);
 425         return(1);
 426 }
 427
 428 static int
 429 post_nf(CHKARGS)
 430 {
 431
 432         if (MAN_LITERAL & m->flags)
 433                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 434
 435         m->flags |= MAN_LITERAL;
 436         return(1);
 437 }
 438
 439 static int
 440 post_fi(CHKARGS)
 441 {
 442
 443         if ( ! (MAN_LITERAL & m->flags))
 444                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 445
 446         m->flags &= ~MAN_LITERAL;
 447         return(1);
 448 }
 449
 450 static int
 451 post_UC(CHKARGS)
 452 {
 453         static const char * const bsd_versions[] = {
 454             "3rd Berkeley Distribution",
 455             "4th Berkeley Distribution",
 456             "4.2 Berkeley Distribution",
 457             "4.3 Berkeley Distribution",
 458             "4.4 Berkeley Distribution",
 459         };
 460
 461         const char      *p, *s;
 462
 463         n = n->child;
 464
 465         if (NULL == n || MAN_TEXT != n->type)
 466                 p = bsd_versions[0];
 467         else {
 468                 s = n->string;
 469                 if (0 == strcmp(s, "3"))
 470                         p = bsd_versions[0];
 471                 else if (0 == strcmp(s, "4"))
 472                         p = bsd_versions[1];
 473                 else if (0 == strcmp(s, "5"))
 474                         p = bsd_versions[2];
 475                 else if (0 == strcmp(s, "6"))
 476                         p = bsd_versions[3];
 477                 else if (0 == strcmp(s, "7"))
 478                         p = bsd_versions[4];
 479                 else
 480                         p = bsd_versions[0];
 481         }
 482
 483         if (m->meta.source)
 484                 free(m->meta.source);
 485
 486         m->meta.source = mandoc_strdup(p);
 487         return(1);
 488 }
 489
 490 static int
 491 post_AT(CHKARGS)
 492 {
 493         static const char * const unix_versions[] = {
 494             "7th Edition",
 495             "System III",
 496             "System V",
 497             "System V Release 2",
 498         };
 499
 500         const char      *p, *s;
 501         struct man_node *nn;
 502
 503         n = n->child;
 504
 505         if (NULL == n || MAN_TEXT != n->type)
 506                 p = unix_versions[0];
 507         else {
 508                 s = n->string;
 509                 if (0 == strcmp(s, "3"))
 510                         p = unix_versions[0];
 511                 else if (0 == strcmp(s, "4"))
 512                         p = unix_versions[1];
 513                 else if (0 == strcmp(s, "5")) {
 514                         nn = n->next;
 515                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 516                                 p = unix_versions[3];
 517                         else
 518                                 p = unix_versions[2];
 519                 } else
 520                         p = unix_versions[0];
 521         }
 522
 523         if (m->meta.source)
 524                 free(m->meta.source);
 525
 526         m->meta.source = mandoc_strdup(p);
 527         return(1);
 528 }
 529
 530 static int
 531 post_vs(CHKARGS)
 532 {
 533
 534         /*
 535          * Don't warn about this because it occurs in pod2man and would
 536          * cause considerable (unfixable) warnage.
 537          */
 538         if (NULL == n->prev && MAN_ROOT == n->parent->type)
 539                 man_node_delete(m, n);
 540
 541         return(1);
 542 }