man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2017 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 AUTHORS DISCLAIM ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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 #include "config.h"
  19
  20 #include <sys/types.h>
  21
  22 #include <assert.h>
  23 #include <ctype.h>
  24 #include <errno.h>
  25 #include <limits.h>
  26 #include <stdarg.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mandoc_aux.h"
  32 #include "mandoc.h"
  33 #include "roff.h"
  34 #include "man.h"
  35 #include "libmandoc.h"
  36 #include "roff_int.h"
  37 #include "libman.h"
  38
  39 #define CHKARGS   struct roff_man *man, struct roff_node *n
  40
  41 typedef void    (*v_check)(CHKARGS);
  42
  43 static  void      check_par(CHKARGS);
  44 static  void      check_part(CHKARGS);
  45 static  void      check_root(CHKARGS);
  46 static  void      check_text(CHKARGS);
  47
  48 static  void      post_AT(CHKARGS);
  49 static  void      post_IP(CHKARGS);
  50 static  void      post_vs(CHKARGS);
  51 static  void      post_ft(CHKARGS);
  52 static  void      post_OP(CHKARGS);
  53 static  void      post_TH(CHKARGS);
  54 static  void      post_UC(CHKARGS);
  55 static  void      post_UR(CHKARGS);
  56
  57 static  const v_check __man_valids[MAN_MAX - MAN_TH] = {
  58         post_TH,    /* TH */
  59         NULL,       /* SH */
  60         NULL,       /* SS */
  61         NULL,       /* TP */
  62         check_par,  /* LP */
  63         check_par,  /* PP */
  64         check_par,  /* P */
  65         post_IP,    /* IP */
  66         NULL,       /* HP */
  67         NULL,       /* SM */
  68         NULL,       /* SB */
  69         NULL,       /* BI */
  70         NULL,       /* IB */
  71         NULL,       /* BR */
  72         NULL,       /* RB */
  73         NULL,       /* R */
  74         NULL,       /* B */
  75         NULL,       /* I */
  76         NULL,       /* IR */
  77         NULL,       /* RI */
  78         post_vs,    /* sp */
  79         NULL,       /* nf */
  80         NULL,       /* fi */
  81         NULL,       /* RE */
  82         check_part, /* RS */
  83         NULL,       /* DT */
  84         post_UC,    /* UC */
  85         NULL,       /* PD */
  86         post_AT,    /* AT */
  87         NULL,       /* in */
  88         post_ft,    /* ft */
  89         post_OP,    /* OP */
  90         NULL,       /* EX */
  91         NULL,       /* EE */
  92         post_UR,    /* UR */
  93         NULL,       /* UE */
  94         NULL,       /* ll */
  95 };
  96 static  const v_check *man_valids = __man_valids - MAN_TH;
  97
  98
  99 void
 100 man_node_validate(struct roff_man *man)
 101 {
 102         struct roff_node *n;
 103         const v_check    *cp;
 104
 105         n = man->last;
 106         man->last = man->last->child;
 107         while (man->last != NULL) {
 108                 man_node_validate(man);
 109                 if (man->last == n)
 110                         man->last = man->last->child;
 111                 else
 112                         man->last = man->last->next;
 113         }
 114
 115         man->last = n;
 116         man->next = ROFF_NEXT_SIBLING;
 117         switch (n->type) {
 118         case ROFFT_TEXT:
 119                 check_text(man, n);
 120                 break;
 121         case ROFFT_ROOT:
 122                 check_root(man, n);
 123                 break;
 124         case ROFFT_EQN:
 125         case ROFFT_TBL:
 126                 break;
 127         default:
 128                 if (n->tok < ROFF_MAX) {
 129                         switch (n->tok) {
 130                         case ROFF_br:
 131                                 post_vs(man, n);
 132                                 break;
 133                         default:
 134                                 abort();
 135                         }
 136                         break;
 137                 }
 138                 assert(n->tok >= MAN_TH && n->tok < MAN_MAX);
 139                 cp = man_valids + n->tok;
 140                 if (*cp)
 141                         (*cp)(man, n);
 142                 if (man->last == n)
 143                         man_state(man, n);
 144                 break;
 145         }
 146 }
 147
 148 static void
 149 check_root(CHKARGS)
 150 {
 151
 152         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 153
 154         if (NULL == man->first->child)
 155                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 156                     n->line, n->pos, NULL);
 157         else
 158                 man->meta.hasbody = 1;
 159
 160         if (NULL == man->meta.title) {
 161                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 162                     n->line, n->pos, NULL);
 163
 164                 /*
 165                  * If a title hasn't been set, do so now (by
 166                  * implication, date and section also aren't set).
 167                  */
 168
 169                 man->meta.title = mandoc_strdup("");
 170                 man->meta.msec = mandoc_strdup("");
 171                 man->meta.date = man->quick ? mandoc_strdup("") :
 172                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 173         }
 174 }
 175
 176 static void
 177 check_text(CHKARGS)
 178 {
 179         char            *cp, *p;
 180
 181         if (MAN_LITERAL & man->flags)
 182                 return;
 183
 184         cp = n->string;
 185         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 186                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 187                     n->line, n->pos + (p - cp), NULL);
 188 }
 189
 190 static void
 191 post_OP(CHKARGS)
 192 {
 193
 194         if (n->child == NULL)
 195                 mandoc_msg(MANDOCERR_OP_EMPTY, man->parse,
 196                     n->line, n->pos, "OP");
 197         else if (n->child->next != NULL && n->child->next->next != NULL) {
 198                 n = n->child->next->next;
 199                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 200                     n->line, n->pos, "OP ... %s", n->string);
 201         }
 202 }
 203
 204 static void
 205 post_UR(CHKARGS)
 206 {
 207
 208         if (n->type == ROFFT_HEAD && n->child == NULL)
 209                 mandoc_vmsg(MANDOCERR_UR_NOHEAD, man->parse,
 210                     n->line, n->pos, "UR");
 211         check_part(man, n);
 212 }
 213
 214 static void
 215 post_ft(CHKARGS)
 216 {
 217         char    *cp;
 218         int      ok;
 219
 220         if (n->child == NULL)
 221                 return;
 222
 223         ok = 0;
 224         cp = n->child->string;
 225         switch (*cp) {
 226         case '1':
 227         case '2':
 228         case '3':
 229         case '4':
 230         case 'I':
 231         case 'P':
 232         case 'R':
 233                 if ('\0' == cp[1])
 234                         ok = 1;
 235                 break;
 236         case 'B':
 237                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 238                         ok = 1;
 239                 break;
 240         case 'C':
 241                 if ('W' == cp[1] && '\0' == cp[2])
 242                         ok = 1;
 243                 break;
 244         default:
 245                 break;
 246         }
 247
 248         if (0 == ok) {
 249                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 250                     n->line, n->pos, "ft %s", cp);
 251                 *cp = '\0';
 252         }
 253 }
 254
 255 static void
 256 check_part(CHKARGS)
 257 {
 258
 259         if (n->type == ROFFT_BODY && n->child == NULL)
 260                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 261                     n->line, n->pos, roff_name[n->tok]);
 262 }
 263
 264 static void
 265 check_par(CHKARGS)
 266 {
 267
 268         switch (n->type) {
 269         case ROFFT_BLOCK:
 270                 if (n->body->child == NULL)
 271                         roff_node_delete(man, n);
 272                 break;
 273         case ROFFT_BODY:
 274                 if (n->child == NULL)
 275                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 276                             man->parse, n->line, n->pos,
 277                             "%s empty", roff_name[n->tok]);
 278                 break;
 279         case ROFFT_HEAD:
 280                 if (n->child != NULL)
 281                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 282                             man->parse, n->line, n->pos, "%s %s%s",
 283                             roff_name[n->tok], n->child->string,
 284                             n->child->next != NULL ? " ..." : "");
 285                 break;
 286         default:
 287                 break;
 288         }
 289 }
 290
 291 static void
 292 post_IP(CHKARGS)
 293 {
 294
 295         switch (n->type) {
 296         case ROFFT_BLOCK:
 297                 if (n->head->child == NULL && n->body->child == NULL)
 298                         roff_node_delete(man, n);
 299                 break;
 300         case ROFFT_BODY:
 301                 if (n->parent->head->child == NULL && n->child == NULL)
 302                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 303                             man->parse, n->line, n->pos,
 304                             "%s empty", roff_name[n->tok]);
 305                 break;
 306         default:
 307                 break;
 308         }
 309 }
 310
 311 static void
 312 post_TH(CHKARGS)
 313 {
 314         struct roff_node *nb;
 315         const char      *p;
 316
 317         free(man->meta.title);
 318         free(man->meta.vol);
 319         free(man->meta.os);
 320         free(man->meta.msec);
 321         free(man->meta.date);
 322
 323         man->meta.title = man->meta.vol = man->meta.date =
 324             man->meta.msec = man->meta.os = NULL;
 325
 326         nb = n;
 327
 328         /* ->TITLE<- MSEC DATE OS VOL */
 329
 330         n = n->child;
 331         if (n && n->string) {
 332                 for (p = n->string; '\0' != *p; p++) {
 333                         /* Only warn about this once... */
 334                         if (isalpha((unsigned char)*p) &&
 335                             ! isupper((unsigned char)*p)) {
 336                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 337                                     man->parse, n->line,
 338                                     n->pos + (p - n->string),
 339                                     "TH %s", n->string);
 340                                 break;
 341                         }
 342                 }
 343                 man->meta.title = mandoc_strdup(n->string);
 344         } else {
 345                 man->meta.title = mandoc_strdup("");
 346                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 347                     nb->line, nb->pos, "TH");
 348         }
 349
 350         /* TITLE ->MSEC<- DATE OS VOL */
 351
 352         if (n)
 353                 n = n->next;
 354         if (n && n->string)
 355                 man->meta.msec = mandoc_strdup(n->string);
 356         else {
 357                 man->meta.msec = mandoc_strdup("");
 358                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 359                     nb->line, nb->pos, "TH %s", man->meta.title);
 360         }
 361
 362         /* TITLE MSEC ->DATE<- OS VOL */
 363
 364         if (n)
 365                 n = n->next;
 366         if (n && n->string && '\0' != n->string[0]) {
 367                 man->meta.date = man->quick ?
 368                     mandoc_strdup(n->string) :
 369                     mandoc_normdate(man->parse, n->string,
 370                         n->line, n->pos);
 371         } else {
 372                 man->meta.date = mandoc_strdup("");
 373                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 374                     n ? n->line : nb->line,
 375                     n ? n->pos : nb->pos, "TH");
 376         }
 377
 378         /* TITLE MSEC DATE ->OS<- VOL */
 379
 380         if (n && (n = n->next))
 381                 man->meta.os = mandoc_strdup(n->string);
 382         else if (man->defos != NULL)
 383                 man->meta.os = mandoc_strdup(man->defos);
 384
 385         /* TITLE MSEC DATE OS ->VOL<- */
 386         /* If missing, use the default VOL name for MSEC. */
 387
 388         if (n && (n = n->next))
 389                 man->meta.vol = mandoc_strdup(n->string);
 390         else if ('\0' != man->meta.msec[0] &&
 391             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 392                 man->meta.vol = mandoc_strdup(p);
 393
 394         if (n != NULL && (n = n->next) != NULL)
 395                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 396                     n->line, n->pos, "TH ... %s", n->string);
 397
 398         /*
 399          * Remove the `TH' node after we've processed it for our
 400          * meta-data.
 401          */
 402         roff_node_delete(man, man->last);
 403 }
 404
 405 static void
 406 post_UC(CHKARGS)
 407 {
 408         static const char * const bsd_versions[] = {
 409             "3rd Berkeley Distribution",
 410             "4th Berkeley Distribution",
 411             "4.2 Berkeley Distribution",
 412             "4.3 Berkeley Distribution",
 413             "4.4 Berkeley Distribution",
 414         };
 415
 416         const char      *p, *s;
 417
 418         n = n->child;
 419
 420         if (n == NULL || n->type != ROFFT_TEXT)
 421                 p = bsd_versions[0];
 422         else {
 423                 s = n->string;
 424                 if (0 == strcmp(s, "3"))
 425                         p = bsd_versions[0];
 426                 else if (0 == strcmp(s, "4"))
 427                         p = bsd_versions[1];
 428                 else if (0 == strcmp(s, "5"))
 429                         p = bsd_versions[2];
 430                 else if (0 == strcmp(s, "6"))
 431                         p = bsd_versions[3];
 432                 else if (0 == strcmp(s, "7"))
 433                         p = bsd_versions[4];
 434                 else
 435                         p = bsd_versions[0];
 436         }
 437
 438         free(man->meta.os);
 439         man->meta.os = mandoc_strdup(p);
 440 }
 441
 442 static void
 443 post_AT(CHKARGS)
 444 {
 445         static const char * const unix_versions[] = {
 446             "7th Edition",
 447             "System III",
 448             "System V",
 449             "System V Release 2",
 450         };
 451
 452         struct roff_node *nn;
 453         const char      *p, *s;
 454
 455         n = n->child;
 456
 457         if (n == NULL || n->type != ROFFT_TEXT)
 458                 p = unix_versions[0];
 459         else {
 460                 s = n->string;
 461                 if (0 == strcmp(s, "3"))
 462                         p = unix_versions[0];
 463                 else if (0 == strcmp(s, "4"))
 464                         p = unix_versions[1];
 465                 else if (0 == strcmp(s, "5")) {
 466                         nn = n->next;
 467                         if (nn != NULL &&
 468                             nn->type == ROFFT_TEXT &&
 469                             nn->string[0] != '\0')
 470                                 p = unix_versions[3];
 471                         else
 472                                 p = unix_versions[2];
 473                 } else
 474                         p = unix_versions[0];
 475         }
 476
 477         free(man->meta.os);
 478         man->meta.os = mandoc_strdup(p);
 479 }
 480
 481 static void
 482 post_vs(CHKARGS)
 483 {
 484
 485         if (NULL != n->prev)
 486                 return;
 487
 488         switch (n->parent->tok) {
 489         case MAN_SH:
 490         case MAN_SS:
 491                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 492                     "%s after %s", roff_name[n->tok],
 493                     roff_name[n->parent->tok]);
 494                 /* FALLTHROUGH */
 495         case TOKEN_NONE:
 496                 /*
 497                  * Don't warn about this because it occurs in pod2man
 498                  * and would cause considerable (unfixable) warnage.
 499                  */
 500                 roff_node_delete(man, n);
 501                 break;
 502         default:
 503                 break;
 504         }
 505 }