man_validate.c

   1 /*      $Id: man_validate.c,v 1.77 2011/11/03 20:48:52 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_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 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_fi[] = { check_eq0, post_fi, NULL };
  71 static  v_check   posts_ft[] = { post_ft, NULL };
  72 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  73 static  v_check   posts_par[] = { check_par, NULL };
  74 static  v_check   posts_part[] = { check_part, NULL };
  75 static  v_check   posts_sec[] = { post_sec, NULL };
  76 static  v_check   posts_sp[] = { post_vs, check_le1, NULL };
  77 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  78 static  v_check   posts_uc[] = { post_UC, NULL };
  79 static  v_check   pres_bline[] = { check_bline, NULL };
  80 static  v_check   pres_sec[] = { check_bline, pre_sec, NULL};
  81
  82 static  const struct man_valid man_valids[MAN_MAX] = {
  83         { NULL, posts_br }, /* br */
  84         { pres_bline, posts_th }, /* TH */
  85         { pres_sec, posts_sec }, /* SH */
  86         { pres_sec, posts_sec }, /* SS */
  87         { pres_bline, NULL }, /* TP */
  88         { pres_bline, posts_par }, /* LP */
  89         { pres_bline, posts_par }, /* PP */
  90         { pres_bline, posts_par }, /* P */
  91         { pres_bline, NULL }, /* IP */
  92         { pres_bline, 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 */ /* FIXME: should warn only. */
 105         { NULL, posts_sp }, /* sp */ /* FIXME: should warn only. */
 106         { pres_bline, posts_nf }, /* nf */
 107         { pres_bline, 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 };
 117
 118
 119 int
 120 man_valid_pre(struct man *m, struct man_node *n)
 121 {
 122         v_check         *cp;
 123
 124         switch (n->type) {
 125         case (MAN_TEXT):
 126                 /* FALLTHROUGH */
 127         case (MAN_ROOT):
 128                 /* FALLTHROUGH */
 129         case (MAN_EQN):
 130                 /* FALLTHROUGH */
 131         case (MAN_TBL):
 132                 return(1);
 133         default:
 134                 break;
 135         }
 136
 137         if (NULL == (cp = man_valids[n->tok].pres))
 138                 return(1);
 139         for ( ; *cp; cp++)
 140                 if ( ! (*cp)(m, n))
 141                         return(0);
 142         return(1);
 143 }
 144
 145
 146 int
 147 man_valid_post(struct man *m)
 148 {
 149         v_check         *cp;
 150
 151         if (MAN_VALID & m->last->flags)
 152                 return(1);
 153         m->last->flags |= MAN_VALID;
 154
 155         switch (m->last->type) {
 156         case (MAN_TEXT):
 157                 check_text(m, m->last);
 158                 return(1);
 159         case (MAN_ROOT):
 160                 return(check_root(m, m->last));
 161         case (MAN_EQN):
 162                 /* FALLTHROUGH */
 163         case (MAN_TBL):
 164                 return(1);
 165         default:
 166                 break;
 167         }
 168
 169         if (NULL == (cp = man_valids[m->last->tok].posts))
 170                 return(1);
 171         for ( ; *cp; cp++)
 172                 if ( ! (*cp)(m, m->last))
 173                         return(0);
 174
 175         return(1);
 176 }
 177
 178
 179 static int
 180 check_root(CHKARGS)
 181 {
 182
 183         if (MAN_BLINE & m->flags)
 184                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 185         else if (MAN_ELINE & m->flags)
 186                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 187
 188         m->flags &= ~MAN_BLINE;
 189         m->flags &= ~MAN_ELINE;
 190
 191         if (NULL == m->first->child) {
 192                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 193                 return(0);
 194         } else if (NULL == m->meta.title) {
 195                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 196
 197                 /*
 198                  * If a title hasn't been set, do so now (by
 199                  * implication, date and section also aren't set).
 200                  */
 201
 202                 m->meta.title = mandoc_strdup("unknown");
 203                 m->meta.msec = mandoc_strdup("1");
 204                 m->meta.date = mandoc_normdate
 205                         (m->parse, NULL, n->line, n->pos);
 206         }
 207
 208         return(1);
 209 }
 210
 211 static void
 212 check_text(CHKARGS)
 213 {
 214         char            *cp, *p;
 215
 216         if (MAN_LITERAL & m->flags)
 217                 return;
 218
 219         cp = n->string;
 220         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 221                 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 222 }
 223
 224 #define INEQ_DEFINE(x, ineq, name) \
 225 static int \
 226 check_##name(CHKARGS) \
 227 { \
 228         if (n->nchild ineq (x)) \
 229                 return(1); \
 230         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 231                         "line arguments %s %d (have %d)", \
 232                         #ineq, (x), n->nchild); \
 233         return(1); \
 234 }
 235
 236 INEQ_DEFINE(0, ==, eq0)
 237 INEQ_DEFINE(1, <=, le1)
 238 INEQ_DEFINE(2, >=, ge2)
 239 INEQ_DEFINE(5, <=, le5)
 240
 241 static int
 242 post_ft(CHKARGS)
 243 {
 244         char    *cp;
 245         int      ok;
 246
 247         if (0 == n->nchild)
 248                 return(1);
 249
 250         ok = 0;
 251         cp = n->child->string;
 252         switch (*cp) {
 253         case ('1'):
 254                 /* FALLTHROUGH */
 255         case ('2'):
 256                 /* FALLTHROUGH */
 257         case ('3'):
 258                 /* FALLTHROUGH */
 259         case ('4'):
 260                 /* FALLTHROUGH */
 261         case ('I'):
 262                 /* FALLTHROUGH */
 263         case ('P'):
 264                 /* FALLTHROUGH */
 265         case ('R'):
 266                 if ('\0' == cp[1])
 267                         ok = 1;
 268                 break;
 269         case ('B'):
 270                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 271                         ok = 1;
 272                 break;
 273         case ('C'):
 274                 if ('W' == cp[1] && '\0' == cp[2])
 275                         ok = 1;
 276                 break;
 277         default:
 278                 break;
 279         }
 280
 281         if (0 == ok) {
 282                 mandoc_vmsg
 283                         (MANDOCERR_BADFONT, m->parse,
 284                          n->line, n->pos, "%s", cp);
 285                 *cp = '\0';
 286         }
 287
 288         if (1 < n->nchild)
 289                 mandoc_vmsg
 290                         (MANDOCERR_ARGCOUNT, m->parse, n->line,
 291                          n->pos, "want one child (have %d)",
 292                          n->nchild);
 293
 294         return(1);
 295 }
 296
 297 static int
 298 pre_sec(CHKARGS)
 299 {
 300
 301         if (MAN_BLOCK == n->type)
 302                 m->flags &= ~MAN_LITERAL;
 303         return(1);
 304 }
 305
 306 static int
 307 post_sec(CHKARGS)
 308 {
 309
 310         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 311                 return(1);
 312
 313         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 314         return(0);
 315 }
 316
 317 static int
 318 check_part(CHKARGS)
 319 {
 320
 321         if (MAN_BODY == n->type && 0 == n->nchild)
 322                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 323                                 n->pos, "want children (have none)");
 324
 325         return(1);
 326 }
 327
 328
 329 static int
 330 check_par(CHKARGS)
 331 {
 332
 333         switch (n->type) {
 334         case (MAN_BLOCK):
 335                 if (0 == n->body->nchild)
 336                         man_node_delete(m, n);
 337                 break;
 338         case (MAN_BODY):
 339                 if (0 == n->nchild)
 340                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 341                 break;
 342         case (MAN_HEAD):
 343                 if (n->nchild)
 344                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 345                 break;
 346         default:
 347                 break;
 348         }
 349
 350         return(1);
 351 }
 352
 353
 354 static int
 355 check_bline(CHKARGS)
 356 {
 357
 358         assert( ! (MAN_ELINE & m->flags));
 359         if (MAN_BLINE & m->flags) {
 360                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 361                 return(0);
 362         }
 363
 364         return(1);
 365 }
 366
 367 static int
 368 post_TH(CHKARGS)
 369 {
 370         const char      *p;
 371         int              line, pos;
 372
 373         if (m->meta.title)
 374                 free(m->meta.title);
 375         if (m->meta.vol)
 376                 free(m->meta.vol);
 377         if (m->meta.source)
 378                 free(m->meta.source);
 379         if (m->meta.msec)
 380                 free(m->meta.msec);
 381         if (m->meta.date)
 382                 free(m->meta.date);
 383
 384         line = n->line;
 385         pos = n->pos;
 386         m->meta.title = m->meta.vol = m->meta.date =
 387                 m->meta.msec = m->meta.source = NULL;
 388
 389         /* ->TITLE<- MSEC DATE SOURCE VOL */
 390
 391         n = n->child;
 392         if (n && n->string) {
 393                 for (p = n->string; '\0' != *p; p++) {
 394                         /* Only warn about this once... */
 395                         if (isalpha((unsigned char)*p) &&
 396                                         ! isupper((unsigned char)*p)) {
 397                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 398                                 break;
 399                         }
 400                 }
 401                 m->meta.title = mandoc_strdup(n->string);
 402         } else
 403                 m->meta.title = mandoc_strdup("");
 404
 405         /* TITLE ->MSEC<- DATE SOURCE VOL */
 406
 407         if (n)
 408                 n = n->next;
 409         if (n && n->string)
 410                 m->meta.msec = mandoc_strdup(n->string);
 411         else
 412                 m->meta.msec = mandoc_strdup("");
 413
 414         /* TITLE MSEC ->DATE<- SOURCE VOL */
 415
 416         if (n)
 417                 n = n->next;
 418         if (n && n->string && '\0' != n->string[0]) {
 419                 pos = n->pos;
 420                 m->meta.date = mandoc_normdate
 421                     (m->parse, n->string, line, pos);
 422         } else
 423                 m->meta.date = mandoc_strdup("");
 424
 425         /* TITLE MSEC DATE ->SOURCE<- VOL */
 426
 427         if (n && (n = n->next))
 428                 m->meta.source = mandoc_strdup(n->string);
 429
 430         /* TITLE MSEC DATE SOURCE ->VOL<- */
 431
 432         if (n && (n = n->next))
 433                 m->meta.vol = mandoc_strdup(n->string);
 434
 435         /*
 436          * Remove the `TH' node after we've processed it for our
 437          * meta-data.
 438          */
 439         man_node_delete(m, m->last);
 440         return(1);
 441 }
 442
 443 static int
 444 post_nf(CHKARGS)
 445 {
 446
 447         if (MAN_LITERAL & m->flags)
 448                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 449
 450         m->flags |= MAN_LITERAL;
 451         return(1);
 452 }
 453
 454 static int
 455 post_fi(CHKARGS)
 456 {
 457
 458         if ( ! (MAN_LITERAL & m->flags))
 459                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 460
 461         m->flags &= ~MAN_LITERAL;
 462         return(1);
 463 }
 464
 465 static int
 466 post_UC(CHKARGS)
 467 {
 468         static const char * const bsd_versions[] = {
 469             "3rd Berkeley Distribution",
 470             "4th Berkeley Distribution",
 471             "4.2 Berkeley Distribution",
 472             "4.3 Berkeley Distribution",
 473             "4.4 Berkeley Distribution",
 474         };
 475
 476         const char      *p, *s;
 477
 478         n = n->child;
 479
 480         if (NULL == n || MAN_TEXT != n->type)
 481                 p = bsd_versions[0];
 482         else {
 483                 s = n->string;
 484                 if (0 == strcmp(s, "3"))
 485                         p = bsd_versions[0];
 486                 else if (0 == strcmp(s, "4"))
 487                         p = bsd_versions[1];
 488                 else if (0 == strcmp(s, "5"))
 489                         p = bsd_versions[2];
 490                 else if (0 == strcmp(s, "6"))
 491                         p = bsd_versions[3];
 492                 else if (0 == strcmp(s, "7"))
 493                         p = bsd_versions[4];
 494                 else
 495                         p = bsd_versions[0];
 496         }
 497
 498         if (m->meta.source)
 499                 free(m->meta.source);
 500
 501         m->meta.source = mandoc_strdup(p);
 502         return(1);
 503 }
 504
 505 static int
 506 post_AT(CHKARGS)
 507 {
 508         static const char * const unix_versions[] = {
 509             "7th Edition",
 510             "System III",
 511             "System V",
 512             "System V Release 2",
 513         };
 514
 515         const char      *p, *s;
 516         struct man_node *nn;
 517
 518         n = n->child;
 519
 520         if (NULL == n || MAN_TEXT != n->type)
 521                 p = unix_versions[0];
 522         else {
 523                 s = n->string;
 524                 if (0 == strcmp(s, "3"))
 525                         p = unix_versions[0];
 526                 else if (0 == strcmp(s, "4"))
 527                         p = unix_versions[1];
 528                 else if (0 == strcmp(s, "5")) {
 529                         nn = n->next;
 530                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 531                                 p = unix_versions[3];
 532                         else
 533                                 p = unix_versions[2];
 534                 } else
 535                         p = unix_versions[0];
 536         }
 537
 538         if (m->meta.source)
 539                 free(m->meta.source);
 540
 541         m->meta.source = mandoc_strdup(p);
 542         return(1);
 543 }
 544
 545 static int
 546 post_vs(CHKARGS)
 547 {
 548
 549         /*
 550          * Don't warn about this because it occurs in pod2man and would
 551          * cause considerable (unfixable) warnage.
 552          */
 553         if (NULL == n->prev && MAN_ROOT == n->parent->type)
 554                 man_node_delete(m, n);
 555
 556         return(1);
 557 }