man_validate.c

   1 /*      $Id: man_validate.c,v 1.84 2012/07/29 12:35:42 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012 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_IP(CHKARGS);
  59 static  int       post_vs(CHKARGS);
  60 static  int       post_fi(CHKARGS);
  61 static  int       post_ft(CHKARGS);
  62 static  int       post_nf(CHKARGS);
  63 static  int       post_sec(CHKARGS);
  64 static  int       post_TH(CHKARGS);
  65 static  int       post_UC(CHKARGS);
  66 static  int       pre_sec(CHKARGS);
  67
  68 static  v_check   posts_at[] = { post_AT, NULL };
  69 static  v_check   posts_br[] = { post_vs, check_eq0, NULL };
  70 static  v_check   posts_eq0[] = { check_eq0, NULL };
  71 static  v_check   posts_eq2[] = { check_eq2, NULL };
  72 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  73 static  v_check   posts_ft[] = { post_ft, NULL };
  74 static  v_check   posts_ip[] = { post_IP, NULL };
  75 static  v_check   posts_le1[] = { check_le1, NULL };
  76 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  77 static  v_check   posts_par[] = { check_par, NULL };
  78 static  v_check   posts_part[] = { check_part, NULL };
  79 static  v_check   posts_sec[] = { post_sec, NULL };
  80 static  v_check   posts_sp[] = { post_vs, check_le1, NULL };
  81 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  82 static  v_check   posts_uc[] = { post_UC, NULL };
  83 static  v_check   pres_sec[] = { pre_sec, NULL };
  84
  85 static  const struct man_valid man_valids[MAN_MAX] = {
  86         { NULL, posts_br }, /* br */
  87         { NULL, posts_th }, /* TH */
  88         { pres_sec, posts_sec }, /* SH */
  89         { pres_sec, posts_sec }, /* SS */
  90         { NULL, NULL }, /* TP */
  91         { NULL, posts_par }, /* LP */
  92         { NULL, posts_par }, /* PP */
  93         { NULL, posts_par }, /* P */
  94         { NULL, posts_ip }, /* IP */
  95         { NULL, NULL }, /* HP */
  96         { NULL, NULL }, /* SM */
  97         { NULL, NULL }, /* SB */
  98         { NULL, NULL }, /* BI */
  99         { NULL, NULL }, /* IB */
 100         { NULL, NULL }, /* BR */
 101         { NULL, NULL }, /* RB */
 102         { NULL, NULL }, /* R */
 103         { NULL, NULL }, /* B */
 104         { NULL, NULL }, /* I */
 105         { NULL, NULL }, /* IR */
 106         { NULL, NULL }, /* RI */
 107         { NULL, posts_eq0 }, /* na */
 108         { NULL, posts_sp }, /* sp */
 109         { NULL, posts_nf }, /* nf */
 110         { NULL, posts_fi }, /* fi */
 111         { NULL, NULL }, /* RE */
 112         { NULL, posts_part }, /* RS */
 113         { NULL, NULL }, /* DT */
 114         { NULL, posts_uc }, /* UC */
 115         { NULL, posts_le1 }, /* PD */
 116         { NULL, posts_at }, /* AT */
 117         { NULL, NULL }, /* in */
 118         { NULL, posts_ft }, /* ft */
 119         { NULL, posts_eq2 }, /* OP */
 120         { NULL, posts_nf }, /* EX */
 121         { NULL, posts_fi }, /* EE */
 122 };
 123
 124
 125 int
 126 man_valid_pre(struct man *m, struct man_node *n)
 127 {
 128         v_check         *cp;
 129
 130         switch (n->type) {
 131         case (MAN_TEXT):
 132                 /* FALLTHROUGH */
 133         case (MAN_ROOT):
 134                 /* FALLTHROUGH */
 135         case (MAN_EQN):
 136                 /* FALLTHROUGH */
 137         case (MAN_TBL):
 138                 return(1);
 139         default:
 140                 break;
 141         }
 142
 143         if (NULL == (cp = man_valids[n->tok].pres))
 144                 return(1);
 145         for ( ; *cp; cp++)
 146                 if ( ! (*cp)(m, n))
 147                         return(0);
 148         return(1);
 149 }
 150
 151
 152 int
 153 man_valid_post(struct man *m)
 154 {
 155         v_check         *cp;
 156
 157         if (MAN_VALID & m->last->flags)
 158                 return(1);
 159         m->last->flags |= MAN_VALID;
 160
 161         switch (m->last->type) {
 162         case (MAN_TEXT):
 163                 check_text(m, m->last);
 164                 return(1);
 165         case (MAN_ROOT):
 166                 return(check_root(m, m->last));
 167         case (MAN_EQN):
 168                 /* FALLTHROUGH */
 169         case (MAN_TBL):
 170                 return(1);
 171         default:
 172                 break;
 173         }
 174
 175         if (NULL == (cp = man_valids[m->last->tok].posts))
 176                 return(1);
 177         for ( ; *cp; cp++)
 178                 if ( ! (*cp)(m, m->last))
 179                         return(0);
 180
 181         return(1);
 182 }
 183
 184
 185 static int
 186 check_root(CHKARGS)
 187 {
 188
 189         if (MAN_BLINE & m->flags)
 190                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 191         else if (MAN_ELINE & m->flags)
 192                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 193
 194         m->flags &= ~MAN_BLINE;
 195         m->flags &= ~MAN_ELINE;
 196
 197         if (NULL == m->first->child) {
 198                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 199                 return(0);
 200         } else if (NULL == m->meta.title) {
 201                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 202
 203                 /*
 204                  * If a title hasn't been set, do so now (by
 205                  * implication, date and section also aren't set).
 206                  */
 207
 208                 m->meta.title = mandoc_strdup("unknown");
 209                 m->meta.msec = mandoc_strdup("1");
 210                 m->meta.date = mandoc_normdate
 211                         (m->parse, NULL, n->line, n->pos);
 212         }
 213
 214         return(1);
 215 }
 216
 217 static void
 218 check_text(CHKARGS)
 219 {
 220         char            *cp, *p;
 221
 222         if (MAN_LITERAL & m->flags)
 223                 return;
 224
 225         cp = n->string;
 226         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 227                 man_pmsg(m, n->line, (int)(p - cp), MANDOCERR_BADTAB);
 228 }
 229
 230 #define INEQ_DEFINE(x, ineq, name) \
 231 static int \
 232 check_##name(CHKARGS) \
 233 { \
 234         if (n->nchild ineq (x)) \
 235                 return(1); \
 236         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 237                         "line arguments %s %d (have %d)", \
 238                         #ineq, (x), n->nchild); \
 239         return(1); \
 240 }
 241
 242 INEQ_DEFINE(0, ==, eq0)
 243 INEQ_DEFINE(2, ==, eq2)
 244 INEQ_DEFINE(1, <=, le1)
 245 INEQ_DEFINE(2, >=, ge2)
 246 INEQ_DEFINE(5, <=, le5)
 247
 248 static int
 249 post_ft(CHKARGS)
 250 {
 251         char    *cp;
 252         int      ok;
 253
 254         if (0 == n->nchild)
 255                 return(1);
 256
 257         ok = 0;
 258         cp = n->child->string;
 259         switch (*cp) {
 260         case ('1'):
 261                 /* FALLTHROUGH */
 262         case ('2'):
 263                 /* FALLTHROUGH */
 264         case ('3'):
 265                 /* FALLTHROUGH */
 266         case ('4'):
 267                 /* FALLTHROUGH */
 268         case ('I'):
 269                 /* FALLTHROUGH */
 270         case ('P'):
 271                 /* FALLTHROUGH */
 272         case ('R'):
 273                 if ('\0' == cp[1])
 274                         ok = 1;
 275                 break;
 276         case ('B'):
 277                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 278                         ok = 1;
 279                 break;
 280         case ('C'):
 281                 if ('W' == cp[1] && '\0' == cp[2])
 282                         ok = 1;
 283                 break;
 284         default:
 285                 break;
 286         }
 287
 288         if (0 == ok) {
 289                 mandoc_vmsg
 290                         (MANDOCERR_BADFONT, m->parse,
 291                          n->line, n->pos, "%s", cp);
 292                 *cp = '\0';
 293         }
 294
 295         if (1 < n->nchild)
 296                 mandoc_vmsg
 297                         (MANDOCERR_ARGCOUNT, m->parse, n->line,
 298                          n->pos, "want one child (have %d)",
 299                          n->nchild);
 300
 301         return(1);
 302 }
 303
 304 static int
 305 pre_sec(CHKARGS)
 306 {
 307
 308         if (MAN_BLOCK == n->type)
 309                 m->flags &= ~MAN_LITERAL;
 310         return(1);
 311 }
 312
 313 static int
 314 post_sec(CHKARGS)
 315 {
 316
 317         if ( ! (MAN_HEAD == n->type && 0 == n->nchild))
 318                 return(1);
 319
 320         man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 321         return(0);
 322 }
 323
 324 static int
 325 check_part(CHKARGS)
 326 {
 327
 328         if (MAN_BODY == n->type && 0 == n->nchild)
 329                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 330                                 n->pos, "want children (have none)");
 331
 332         return(1);
 333 }
 334
 335
 336 static int
 337 check_par(CHKARGS)
 338 {
 339
 340         switch (n->type) {
 341         case (MAN_BLOCK):
 342                 if (0 == n->body->nchild)
 343                         man_node_delete(m, n);
 344                 break;
 345         case (MAN_BODY):
 346                 if (0 == n->nchild)
 347                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 348                 break;
 349         case (MAN_HEAD):
 350                 if (n->nchild)
 351                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 352                 break;
 353         default:
 354                 break;
 355         }
 356
 357         return(1);
 358 }
 359
 360 static int
 361 post_IP(CHKARGS)
 362 {
 363
 364         switch (n->type) {
 365         case (MAN_BLOCK):
 366                 if (0 == n->head->nchild && 0 == n->body->nchild)
 367                         man_node_delete(m, n);
 368                 break;
 369         case (MAN_BODY):
 370                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 371                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 372                 break;
 373         default:
 374                 break;
 375         }
 376         return(1);
 377 }
 378
 379 static int
 380 post_TH(CHKARGS)
 381 {
 382         const char      *p;
 383         int              line, pos;
 384
 385         if (m->meta.title)
 386                 free(m->meta.title);
 387         if (m->meta.vol)
 388                 free(m->meta.vol);
 389         if (m->meta.source)
 390                 free(m->meta.source);
 391         if (m->meta.msec)
 392                 free(m->meta.msec);
 393         if (m->meta.date)
 394                 free(m->meta.date);
 395
 396         line = n->line;
 397         pos = n->pos;
 398         m->meta.title = m->meta.vol = m->meta.date =
 399                 m->meta.msec = m->meta.source = NULL;
 400
 401         /* ->TITLE<- MSEC DATE SOURCE VOL */
 402
 403         n = n->child;
 404         if (n && n->string) {
 405                 for (p = n->string; '\0' != *p; p++) {
 406                         /* Only warn about this once... */
 407                         if (isalpha((unsigned char)*p) &&
 408                                         ! isupper((unsigned char)*p)) {
 409                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 410                                 break;
 411                         }
 412                 }
 413                 m->meta.title = mandoc_strdup(n->string);
 414         } else
 415                 m->meta.title = mandoc_strdup("");
 416
 417         /* TITLE ->MSEC<- DATE SOURCE VOL */
 418
 419         if (n)
 420                 n = n->next;
 421         if (n && n->string)
 422                 m->meta.msec = mandoc_strdup(n->string);
 423         else
 424                 m->meta.msec = mandoc_strdup("");
 425
 426         /* TITLE MSEC ->DATE<- SOURCE VOL */
 427
 428         if (n)
 429                 n = n->next;
 430         if (n && n->string && '\0' != n->string[0]) {
 431                 pos = n->pos;
 432                 m->meta.date = mandoc_normdate
 433                     (m->parse, n->string, line, pos);
 434         } else
 435                 m->meta.date = mandoc_strdup("");
 436
 437         /* TITLE MSEC DATE ->SOURCE<- VOL */
 438
 439         if (n && (n = n->next))
 440                 m->meta.source = mandoc_strdup(n->string);
 441
 442         /* TITLE MSEC DATE SOURCE ->VOL<- */
 443         /* If missing, use the default VOL name for MSEC. */
 444
 445         if (n && (n = n->next))
 446                 m->meta.vol = mandoc_strdup(n->string);
 447         else if ('\0' != m->meta.msec[0] &&
 448             (NULL != (p = mandoc_a2msec(m->meta.msec))))
 449                 m->meta.vol = mandoc_strdup(p);
 450
 451         /*
 452          * Remove the `TH' node after we've processed it for our
 453          * meta-data.
 454          */
 455         man_node_delete(m, m->last);
 456         return(1);
 457 }
 458
 459 static int
 460 post_nf(CHKARGS)
 461 {
 462
 463         if (MAN_LITERAL & m->flags)
 464                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 465
 466         m->flags |= MAN_LITERAL;
 467         return(1);
 468 }
 469
 470 static int
 471 post_fi(CHKARGS)
 472 {
 473
 474         if ( ! (MAN_LITERAL & m->flags))
 475                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 476
 477         m->flags &= ~MAN_LITERAL;
 478         return(1);
 479 }
 480
 481 static int
 482 post_UC(CHKARGS)
 483 {
 484         static const char * const bsd_versions[] = {
 485             "3rd Berkeley Distribution",
 486             "4th Berkeley Distribution",
 487             "4.2 Berkeley Distribution",
 488             "4.3 Berkeley Distribution",
 489             "4.4 Berkeley Distribution",
 490         };
 491
 492         const char      *p, *s;
 493
 494         n = n->child;
 495
 496         if (NULL == n || MAN_TEXT != n->type)
 497                 p = bsd_versions[0];
 498         else {
 499                 s = n->string;
 500                 if (0 == strcmp(s, "3"))
 501                         p = bsd_versions[0];
 502                 else if (0 == strcmp(s, "4"))
 503                         p = bsd_versions[1];
 504                 else if (0 == strcmp(s, "5"))
 505                         p = bsd_versions[2];
 506                 else if (0 == strcmp(s, "6"))
 507                         p = bsd_versions[3];
 508                 else if (0 == strcmp(s, "7"))
 509                         p = bsd_versions[4];
 510                 else
 511                         p = bsd_versions[0];
 512         }
 513
 514         if (m->meta.source)
 515                 free(m->meta.source);
 516
 517         m->meta.source = mandoc_strdup(p);
 518         return(1);
 519 }
 520
 521 static int
 522 post_AT(CHKARGS)
 523 {
 524         static const char * const unix_versions[] = {
 525             "7th Edition",
 526             "System III",
 527             "System V",
 528             "System V Release 2",
 529         };
 530
 531         const char      *p, *s;
 532         struct man_node *nn;
 533
 534         n = n->child;
 535
 536         if (NULL == n || MAN_TEXT != n->type)
 537                 p = unix_versions[0];
 538         else {
 539                 s = n->string;
 540                 if (0 == strcmp(s, "3"))
 541                         p = unix_versions[0];
 542                 else if (0 == strcmp(s, "4"))
 543                         p = unix_versions[1];
 544                 else if (0 == strcmp(s, "5")) {
 545                         nn = n->next;
 546                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 547                                 p = unix_versions[3];
 548                         else
 549                                 p = unix_versions[2];
 550                 } else
 551                         p = unix_versions[0];
 552         }
 553
 554         if (m->meta.source)
 555                 free(m->meta.source);
 556
 557         m->meta.source = mandoc_strdup(p);
 558         return(1);
 559 }
 560
 561 static int
 562 post_vs(CHKARGS)
 563 {
 564
 565         if (NULL != n->prev)
 566                 return(1);
 567
 568         switch (n->parent->tok) {
 569         case (MAN_SH):
 570                 /* FALLTHROUGH */
 571         case (MAN_SS):
 572                 man_nmsg(m, n, MANDOCERR_IGNPAR);
 573                 /* FALLTHROUGH */
 574         case (MAN_MAX):
 575                 /*
 576                  * Don't warn about this because it occurs in pod2man
 577                  * and would cause considerable (unfixable) warnage.
 578                  */
 579                 man_node_delete(m, n);
 580                 break;
 581         default:
 582                 break;
 583         }
 584
 585         return(1);
 586 }