man_validate.c

   1 /*      $Id: man_validate.c,v 1.81 2012/06/02 20:16:23 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_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         { NULL, posts_nf }, /* EX */
 118         { NULL, posts_fi }, /* EE */
 119 };
 120
 121
 122 int
 123 man_valid_pre(struct man *m, struct man_node *n)
 124 {
 125         v_check         *cp;
 126
 127         switch (n->type) {
 128         case (MAN_TEXT):
 129                 /* FALLTHROUGH */
 130         case (MAN_ROOT):
 131                 /* FALLTHROUGH */
 132         case (MAN_EQN):
 133                 /* FALLTHROUGH */
 134         case (MAN_TBL):
 135                 return(1);
 136         default:
 137                 break;
 138         }
 139
 140         if (NULL == (cp = man_valids[n->tok].pres))
 141                 return(1);
 142         for ( ; *cp; cp++)
 143                 if ( ! (*cp)(m, n))
 144                         return(0);
 145         return(1);
 146 }
 147
 148
 149 int
 150 man_valid_post(struct man *m)
 151 {
 152         v_check         *cp;
 153
 154         if (MAN_VALID & m->last->flags)
 155                 return(1);
 156         m->last->flags |= MAN_VALID;
 157
 158         switch (m->last->type) {
 159         case (MAN_TEXT):
 160                 check_text(m, m->last);
 161                 return(1);
 162         case (MAN_ROOT):
 163                 return(check_root(m, m->last));
 164         case (MAN_EQN):
 165                 /* FALLTHROUGH */
 166         case (MAN_TBL):
 167                 return(1);
 168         default:
 169                 break;
 170         }
 171
 172         if (NULL == (cp = man_valids[m->last->tok].posts))
 173                 return(1);
 174         for ( ; *cp; cp++)
 175                 if ( ! (*cp)(m, m->last))
 176                         return(0);
 177
 178         return(1);
 179 }
 180
 181
 182 static int
 183 check_root(CHKARGS)
 184 {
 185
 186         if (MAN_BLINE & m->flags)
 187                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 188         else if (MAN_ELINE & m->flags)
 189                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 190
 191         m->flags &= ~MAN_BLINE;
 192         m->flags &= ~MAN_ELINE;
 193
 194         if (NULL == m->first->child) {
 195                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 196                 return(0);
 197         } else if (NULL == m->meta.title) {
 198                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 199
 200                 /*
 201                  * If a title hasn't been set, do so now (by
 202                  * implication, date and section also aren't set).
 203                  */
 204
 205                 m->meta.title = mandoc_strdup("unknown");
 206                 m->meta.msec = mandoc_strdup("1");
 207                 m->meta.date = mandoc_normdate
 208                         (m->parse, NULL, n->line, n->pos);
 209         }
 210
 211         return(1);
 212 }
 213
 214 static void
 215 check_text(CHKARGS)
 216 {
 217         char            *cp, *p;
 218
 219         if (MAN_LITERAL & m->flags)
 220                 return;
 221
 222         cp = n->string;
 223         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 224                 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 225 }
 226
 227 #define INEQ_DEFINE(x, ineq, name) \
 228 static int \
 229 check_##name(CHKARGS) \
 230 { \
 231         if (n->nchild ineq (x)) \
 232                 return(1); \
 233         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 234                         "line arguments %s %d (have %d)", \
 235                         #ineq, (x), n->nchild); \
 236         return(1); \
 237 }
 238
 239 INEQ_DEFINE(0, ==, eq0)
 240 INEQ_DEFINE(2, ==, eq2)
 241 INEQ_DEFINE(1, <=, le1)
 242 INEQ_DEFINE(2, >=, ge2)
 243 INEQ_DEFINE(5, <=, le5)
 244
 245 static int
 246 post_ft(CHKARGS)
 247 {
 248         char    *cp;
 249         int      ok;
 250
 251         if (0 == n->nchild)
 252                 return(1);
 253
 254         ok = 0;
 255         cp = n->child->string;
 256         switch (*cp) {
 257         case ('1'):
 258                 /* FALLTHROUGH */
 259         case ('2'):
 260                 /* FALLTHROUGH */
 261         case ('3'):
 262                 /* FALLTHROUGH */
 263         case ('4'):
 264                 /* FALLTHROUGH */
 265         case ('I'):
 266                 /* FALLTHROUGH */
 267         case ('P'):
 268                 /* FALLTHROUGH */
 269         case ('R'):
 270                 if ('\0' == cp[1])
 271                         ok = 1;
 272                 break;
 273         case ('B'):
 274                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 275                         ok = 1;
 276                 break;
 277         case ('C'):
 278                 if ('W' == cp[1] && '\0' == cp[2])
 279                         ok = 1;
 280                 break;
 281         default:
 282                 break;
 283         }
 284
 285         if (0 == ok) {
 286                 mandoc_vmsg
 287                         (MANDOCERR_BADFONT, m->parse,
 288                          n->line, n->pos, "%s", cp);
 289                 *cp = '\0';
 290         }
 291
 292         if (1 < n->nchild)
 293                 mandoc_vmsg
 294                         (MANDOCERR_ARGCOUNT, m->parse, n->line,
 295                          n->pos, "want one child (have %d)",
 296                          n->nchild);
 297
 298         return(1);
 299 }
 300
 301 static int
 302 pre_sec(CHKARGS)
 303 {
 304
 305         if (MAN_BLOCK == n->type)
 306                 m->flags &= ~MAN_LITERAL;
 307         return(1);
 308 }
 309
 310 static int
 311 post_sec(CHKARGS)
 312 {
 313
 314         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 315                 return(1);
 316
 317         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 318         return(0);
 319 }
 320
 321 static int
 322 check_part(CHKARGS)
 323 {
 324
 325         if (MAN_BODY == n->type && 0 == n->nchild)
 326                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 327                                 n->pos, "want children (have none)");
 328
 329         return(1);
 330 }
 331
 332
 333 static int
 334 check_par(CHKARGS)
 335 {
 336
 337         switch (n->type) {
 338         case (MAN_BLOCK):
 339                 if (0 == n->body->nchild)
 340                         man_node_delete(m, n);
 341                 break;
 342         case (MAN_BODY):
 343                 if (0 == n->nchild)
 344                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 345                 break;
 346         case (MAN_HEAD):
 347                 if (n->nchild)
 348                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 349                 break;
 350         default:
 351                 break;
 352         }
 353
 354         return(1);
 355 }
 356
 357
 358 static int
 359 post_TH(CHKARGS)
 360 {
 361         const char      *p;
 362         int              line, pos;
 363
 364         if (m->meta.title)
 365                 free(m->meta.title);
 366         if (m->meta.vol)
 367                 free(m->meta.vol);
 368         if (m->meta.source)
 369                 free(m->meta.source);
 370         if (m->meta.msec)
 371                 free(m->meta.msec);
 372         if (m->meta.date)
 373                 free(m->meta.date);
 374
 375         line = n->line;
 376         pos = n->pos;
 377         m->meta.title = m->meta.vol = m->meta.date =
 378                 m->meta.msec = m->meta.source = NULL;
 379
 380         /* ->TITLE<- MSEC DATE SOURCE VOL */
 381
 382         n = n->child;
 383         if (n && n->string) {
 384                 for (p = n->string; '\0' != *p; p++) {
 385                         /* Only warn about this once... */
 386                         if (isalpha((unsigned char)*p) &&
 387                                         ! isupper((unsigned char)*p)) {
 388                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 389                                 break;
 390                         }
 391                 }
 392                 m->meta.title = mandoc_strdup(n->string);
 393         } else
 394                 m->meta.title = mandoc_strdup("");
 395
 396         /* TITLE ->MSEC<- DATE SOURCE VOL */
 397
 398         if (n)
 399                 n = n->next;
 400         if (n && n->string)
 401                 m->meta.msec = mandoc_strdup(n->string);
 402         else
 403                 m->meta.msec = mandoc_strdup("");
 404
 405         /* TITLE MSEC ->DATE<- SOURCE VOL */
 406
 407         if (n)
 408                 n = n->next;
 409         if (n && n->string && '\0' != n->string[0]) {
 410                 pos = n->pos;
 411                 m->meta.date = mandoc_normdate
 412                     (m->parse, n->string, line, pos);
 413         } else
 414                 m->meta.date = mandoc_strdup("");
 415
 416         /* TITLE MSEC DATE ->SOURCE<- VOL */
 417
 418         if (n && (n = n->next))
 419                 m->meta.source = mandoc_strdup(n->string);
 420
 421         /* TITLE MSEC DATE SOURCE ->VOL<- */
 422         /* If missing, use the default VOL name for MSEC. */
 423
 424         if (n && (n = n->next))
 425                 m->meta.vol = mandoc_strdup(n->string);
 426         else if ('\0' != m->meta.msec[0] &&
 427             (NULL != (p = mandoc_a2msec(m->meta.msec))))
 428                 m->meta.vol = mandoc_strdup(p);
 429
 430         /*
 431          * Remove the `TH' node after we've processed it for our
 432          * meta-data.
 433          */
 434         man_node_delete(m, m->last);
 435         return(1);
 436 }
 437
 438 static int
 439 post_nf(CHKARGS)
 440 {
 441
 442         if (MAN_LITERAL & m->flags)
 443                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 444
 445         m->flags |= MAN_LITERAL;
 446         return(1);
 447 }
 448
 449 static int
 450 post_fi(CHKARGS)
 451 {
 452
 453         if ( ! (MAN_LITERAL & m->flags))
 454                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 455
 456         m->flags &= ~MAN_LITERAL;
 457         return(1);
 458 }
 459
 460 static int
 461 post_UC(CHKARGS)
 462 {
 463         static const char * const bsd_versions[] = {
 464             "3rd Berkeley Distribution",
 465             "4th Berkeley Distribution",
 466             "4.2 Berkeley Distribution",
 467             "4.3 Berkeley Distribution",
 468             "4.4 Berkeley Distribution",
 469         };
 470
 471         const char      *p, *s;
 472
 473         n = n->child;
 474
 475         if (NULL == n || MAN_TEXT != n->type)
 476                 p = bsd_versions[0];
 477         else {
 478                 s = n->string;
 479                 if (0 == strcmp(s, "3"))
 480                         p = bsd_versions[0];
 481                 else if (0 == strcmp(s, "4"))
 482                         p = bsd_versions[1];
 483                 else if (0 == strcmp(s, "5"))
 484                         p = bsd_versions[2];
 485                 else if (0 == strcmp(s, "6"))
 486                         p = bsd_versions[3];
 487                 else if (0 == strcmp(s, "7"))
 488                         p = bsd_versions[4];
 489                 else
 490                         p = bsd_versions[0];
 491         }
 492
 493         if (m->meta.source)
 494                 free(m->meta.source);
 495
 496         m->meta.source = mandoc_strdup(p);
 497         return(1);
 498 }
 499
 500 static int
 501 post_AT(CHKARGS)
 502 {
 503         static const char * const unix_versions[] = {
 504             "7th Edition",
 505             "System III",
 506             "System V",
 507             "System V Release 2",
 508         };
 509
 510         const char      *p, *s;
 511         struct man_node *nn;
 512
 513         n = n->child;
 514
 515         if (NULL == n || MAN_TEXT != n->type)
 516                 p = unix_versions[0];
 517         else {
 518                 s = n->string;
 519                 if (0 == strcmp(s, "3"))
 520                         p = unix_versions[0];
 521                 else if (0 == strcmp(s, "4"))
 522                         p = unix_versions[1];
 523                 else if (0 == strcmp(s, "5")) {
 524                         nn = n->next;
 525                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 526                                 p = unix_versions[3];
 527                         else
 528                                 p = unix_versions[2];
 529                 } else
 530                         p = unix_versions[0];
 531         }
 532
 533         if (m->meta.source)
 534                 free(m->meta.source);
 535
 536         m->meta.source = mandoc_strdup(p);
 537         return(1);
 538 }
 539
 540 static int
 541 post_vs(CHKARGS)
 542 {
 543
 544         /*
 545          * Don't warn about this because it occurs in pod2man and would
 546          * cause considerable (unfixable) warnage.
 547          */
 548         if (NULL == n->prev && MAN_ROOT == n->parent->type)
 549                 man_node_delete(m, n);
 550
 551         return(1);
 552 }