man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2018 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_abort(CHKARGS);
  44 static  void      check_par(CHKARGS);
  45 static  void      check_part(CHKARGS);
  46 static  void      check_root(CHKARGS);
  47 static  void      check_text(CHKARGS);
  48
  49 static  void      post_AT(CHKARGS);
  50 static  void      post_IP(CHKARGS);
  51 static  void      post_OP(CHKARGS);
  52 static  void      post_TH(CHKARGS);
  53 static  void      post_UC(CHKARGS);
  54 static  void      post_UR(CHKARGS);
  55 static  void      post_in(CHKARGS);
  56 static  void      post_vs(CHKARGS);
  57
  58 static  const v_check man_valids[MAN_MAX - MAN_TH] = {
  59         post_TH,    /* TH */
  60         NULL,       /* SH */
  61         NULL,       /* SS */
  62         NULL,       /* TP */
  63         NULL,       /* TQ */
  64         check_abort,/* LP */
  65         check_par,  /* PP */
  66         check_abort,/* P */
  67         post_IP,    /* IP */
  68         NULL,       /* HP */
  69         NULL,       /* SM */
  70         NULL,       /* SB */
  71         NULL,       /* BI */
  72         NULL,       /* IB */
  73         NULL,       /* BR */
  74         NULL,       /* RB */
  75         NULL,       /* R */
  76         NULL,       /* B */
  77         NULL,       /* I */
  78         NULL,       /* IR */
  79         NULL,       /* RI */
  80         NULL,       /* nf */
  81         NULL,       /* fi */
  82         NULL,       /* RE */
  83         check_part, /* RS */
  84         NULL,       /* DT */
  85         post_UC,    /* UC */
  86         NULL,       /* PD */
  87         post_AT,    /* AT */
  88         post_in,    /* in */
  89         NULL,       /* SY */
  90         NULL,       /* YS */
  91         post_OP,    /* OP */
  92         NULL,       /* EX */
  93         NULL,       /* EE */
  94         post_UR,    /* UR */
  95         NULL,       /* UE */
  96         post_UR,    /* MT */
  97         NULL,       /* ME */
  98 };
  99
 100
 101 /* Validate the subtree rooted at man->last. */
 102 void
 103 man_node_validate(struct roff_man *man)
 104 {
 105         struct roff_node *n;
 106         const v_check    *cp;
 107
 108         /*
 109          * Translate obsolete macros such that later code
 110          * does not need to look for them.
 111          */
 112
 113         n = man->last;
 114         switch (n->tok) {
 115         case MAN_LP:
 116         case MAN_P:
 117                 n->tok = MAN_PP;
 118                 break;
 119         default:
 120                 break;
 121         }
 122
 123         /*
 124          * Iterate over all children, recursing into each one
 125          * in turn, depth-first.
 126          */
 127
 128         man->last = man->last->child;
 129         while (man->last != NULL) {
 130                 man_node_validate(man);
 131                 if (man->last == n)
 132                         man->last = man->last->child;
 133                 else
 134                         man->last = man->last->next;
 135         }
 136
 137         /* Finally validate the macro itself. */
 138
 139         man->last = n;
 140         man->next = ROFF_NEXT_SIBLING;
 141         switch (n->type) {
 142         case ROFFT_TEXT:
 143                 check_text(man, n);
 144                 break;
 145         case ROFFT_ROOT:
 146                 check_root(man, n);
 147                 break;
 148         case ROFFT_COMMENT:
 149         case ROFFT_EQN:
 150         case ROFFT_TBL:
 151                 break;
 152         default:
 153                 if (n->tok < ROFF_MAX) {
 154                         switch (n->tok) {
 155                         case ROFF_br:
 156                         case ROFF_sp:
 157                                 post_vs(man, n);
 158                                 break;
 159                         default:
 160                                 roff_validate(man);
 161                                 break;
 162                         }
 163                         break;
 164                 }
 165                 assert(n->tok >= MAN_TH && n->tok < MAN_MAX);
 166                 cp = man_valids + (n->tok - MAN_TH);
 167                 if (*cp)
 168                         (*cp)(man, n);
 169                 if (man->last == n)
 170                         man_state(man, n);
 171                 break;
 172         }
 173 }
 174
 175 static void
 176 check_root(CHKARGS)
 177 {
 178         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 179
 180         if (n->last == NULL || n->last->type == ROFFT_COMMENT)
 181                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 182                     n->line, n->pos, NULL);
 183         else
 184                 man->meta.hasbody = 1;
 185
 186         if (NULL == man->meta.title) {
 187                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 188                     n->line, n->pos, NULL);
 189
 190                 /*
 191                  * If a title hasn't been set, do so now (by
 192                  * implication, date and section also aren't set).
 193                  */
 194
 195                 man->meta.title = mandoc_strdup("");
 196                 man->meta.msec = mandoc_strdup("");
 197                 man->meta.date = man->quick ? mandoc_strdup("") :
 198                     mandoc_normdate(man, NULL, n->line, n->pos);
 199         }
 200
 201         if (man->meta.os_e &&
 202             (man->meta.rcsids & (1 << man->meta.os_e)) == 0)
 203                 mandoc_msg(MANDOCERR_RCS_MISSING, man->parse, 0, 0,
 204                     man->meta.os_e == MANDOC_OS_OPENBSD ?
 205                     "(OpenBSD)" : "(NetBSD)");
 206 }
 207
 208 static void
 209 check_abort(CHKARGS)
 210 {
 211         abort();
 212 }
 213
 214 static void
 215 check_text(CHKARGS)
 216 {
 217         char            *cp, *p;
 218
 219         if (MAN_LITERAL & man->flags)
 220                 return;
 221
 222         cp = n->string;
 223         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 224                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 225                     n->line, n->pos + (p - cp), NULL);
 226 }
 227
 228 static void
 229 post_OP(CHKARGS)
 230 {
 231
 232         if (n->child == NULL)
 233                 mandoc_msg(MANDOCERR_OP_EMPTY, man->parse,
 234                     n->line, n->pos, "OP");
 235         else if (n->child->next != NULL && n->child->next->next != NULL) {
 236                 n = n->child->next->next;
 237                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 238                     n->line, n->pos, "OP ... %s", n->string);
 239         }
 240 }
 241
 242 static void
 243 post_UR(CHKARGS)
 244 {
 245         if (n->type == ROFFT_HEAD && n->child == NULL)
 246                 mandoc_msg(MANDOCERR_UR_NOHEAD, man->parse,
 247                     n->line, n->pos, roff_name[n->tok]);
 248         check_part(man, n);
 249 }
 250
 251 static void
 252 check_part(CHKARGS)
 253 {
 254
 255         if (n->type == ROFFT_BODY && n->child == NULL)
 256                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 257                     n->line, n->pos, roff_name[n->tok]);
 258 }
 259
 260 static void
 261 check_par(CHKARGS)
 262 {
 263
 264         switch (n->type) {
 265         case ROFFT_BLOCK:
 266                 if (n->body->child == NULL)
 267                         roff_node_delete(man, n);
 268                 break;
 269         case ROFFT_BODY:
 270                 if (n->child == NULL)
 271                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 272                             man->parse, n->line, n->pos,
 273                             "%s empty", roff_name[n->tok]);
 274                 break;
 275         case ROFFT_HEAD:
 276                 if (n->child != NULL)
 277                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 278                             man->parse, n->line, n->pos, "%s %s%s",
 279                             roff_name[n->tok], n->child->string,
 280                             n->child->next != NULL ? " ..." : "");
 281                 break;
 282         default:
 283                 break;
 284         }
 285 }
 286
 287 static void
 288 post_IP(CHKARGS)
 289 {
 290
 291         switch (n->type) {
 292         case ROFFT_BLOCK:
 293                 if (n->head->child == NULL && n->body->child == NULL)
 294                         roff_node_delete(man, n);
 295                 break;
 296         case ROFFT_BODY:
 297                 if (n->parent->head->child == NULL && n->child == NULL)
 298                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 299                             man->parse, n->line, n->pos,
 300                             "%s empty", roff_name[n->tok]);
 301                 break;
 302         default:
 303                 break;
 304         }
 305 }
 306
 307 static void
 308 post_TH(CHKARGS)
 309 {
 310         struct roff_node *nb;
 311         const char      *p;
 312
 313         free(man->meta.title);
 314         free(man->meta.vol);
 315         free(man->meta.os);
 316         free(man->meta.msec);
 317         free(man->meta.date);
 318
 319         man->meta.title = man->meta.vol = man->meta.date =
 320             man->meta.msec = man->meta.os = NULL;
 321
 322         nb = n;
 323
 324         /* ->TITLE<- MSEC DATE OS VOL */
 325
 326         n = n->child;
 327         if (n && n->string) {
 328                 for (p = n->string; '\0' != *p; p++) {
 329                         /* Only warn about this once... */
 330                         if (isalpha((unsigned char)*p) &&
 331                             ! isupper((unsigned char)*p)) {
 332                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 333                                     man->parse, n->line,
 334                                     n->pos + (p - n->string),
 335                                     "TH %s", n->string);
 336                                 break;
 337                         }
 338                 }
 339                 man->meta.title = mandoc_strdup(n->string);
 340         } else {
 341                 man->meta.title = mandoc_strdup("");
 342                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 343                     nb->line, nb->pos, "TH");
 344         }
 345
 346         /* TITLE ->MSEC<- DATE OS VOL */
 347
 348         if (n)
 349                 n = n->next;
 350         if (n && n->string)
 351                 man->meta.msec = mandoc_strdup(n->string);
 352         else {
 353                 man->meta.msec = mandoc_strdup("");
 354                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 355                     nb->line, nb->pos, "TH %s", man->meta.title);
 356         }
 357
 358         /* TITLE MSEC ->DATE<- OS VOL */
 359
 360         if (n)
 361                 n = n->next;
 362         if (n && n->string && '\0' != n->string[0]) {
 363                 man->meta.date = man->quick ?
 364                     mandoc_strdup(n->string) :
 365                     mandoc_normdate(man, n->string, n->line, n->pos);
 366         } else {
 367                 man->meta.date = mandoc_strdup("");
 368                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 369                     n ? n->line : nb->line,
 370                     n ? n->pos : nb->pos, "TH");
 371         }
 372
 373         /* TITLE MSEC DATE ->OS<- VOL */
 374
 375         if (n && (n = n->next))
 376                 man->meta.os = mandoc_strdup(n->string);
 377         else if (man->os_s != NULL)
 378                 man->meta.os = mandoc_strdup(man->os_s);
 379         if (man->meta.os_e == MANDOC_OS_OTHER && man->meta.os != NULL) {
 380                 if (strstr(man->meta.os, "OpenBSD") != NULL)
 381                         man->meta.os_e = MANDOC_OS_OPENBSD;
 382                 else if (strstr(man->meta.os, "NetBSD") != NULL)
 383                         man->meta.os_e = MANDOC_OS_NETBSD;
 384         }
 385
 386         /* TITLE MSEC DATE OS ->VOL<- */
 387         /* If missing, use the default VOL name for MSEC. */
 388
 389         if (n && (n = n->next))
 390                 man->meta.vol = mandoc_strdup(n->string);
 391         else if ('\0' != man->meta.msec[0] &&
 392             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 393                 man->meta.vol = mandoc_strdup(p);
 394
 395         if (n != NULL && (n = n->next) != NULL)
 396                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 397                     n->line, n->pos, "TH ... %s", n->string);
 398
 399         /*
 400          * Remove the `TH' node after we've processed it for our
 401          * meta-data.
 402          */
 403         roff_node_delete(man, man->last);
 404 }
 405
 406 static void
 407 post_UC(CHKARGS)
 408 {
 409         static const char * const bsd_versions[] = {
 410             "3rd Berkeley Distribution",
 411             "4th Berkeley Distribution",
 412             "4.2 Berkeley Distribution",
 413             "4.3 Berkeley Distribution",
 414             "4.4 Berkeley Distribution",
 415         };
 416
 417         const char      *p, *s;
 418
 419         n = n->child;
 420
 421         if (n == NULL || n->type != ROFFT_TEXT)
 422                 p = bsd_versions[0];
 423         else {
 424                 s = n->string;
 425                 if (0 == strcmp(s, "3"))
 426                         p = bsd_versions[0];
 427                 else if (0 == strcmp(s, "4"))
 428                         p = bsd_versions[1];
 429                 else if (0 == strcmp(s, "5"))
 430                         p = bsd_versions[2];
 431                 else if (0 == strcmp(s, "6"))
 432                         p = bsd_versions[3];
 433                 else if (0 == strcmp(s, "7"))
 434                         p = bsd_versions[4];
 435                 else
 436                         p = bsd_versions[0];
 437         }
 438
 439         free(man->meta.os);
 440         man->meta.os = mandoc_strdup(p);
 441 }
 442
 443 static void
 444 post_AT(CHKARGS)
 445 {
 446         static const char * const unix_versions[] = {
 447             "7th Edition",
 448             "System III",
 449             "System V",
 450             "System V Release 2",
 451         };
 452
 453         struct roff_node *nn;
 454         const char      *p, *s;
 455
 456         n = n->child;
 457
 458         if (n == NULL || n->type != ROFFT_TEXT)
 459                 p = unix_versions[0];
 460         else {
 461                 s = n->string;
 462                 if (0 == strcmp(s, "3"))
 463                         p = unix_versions[0];
 464                 else if (0 == strcmp(s, "4"))
 465                         p = unix_versions[1];
 466                 else if (0 == strcmp(s, "5")) {
 467                         nn = n->next;
 468                         if (nn != NULL &&
 469                             nn->type == ROFFT_TEXT &&
 470                             nn->string[0] != '\0')
 471                                 p = unix_versions[3];
 472                         else
 473                                 p = unix_versions[2];
 474                 } else
 475                         p = unix_versions[0];
 476         }
 477
 478         free(man->meta.os);
 479         man->meta.os = mandoc_strdup(p);
 480 }
 481
 482 static void
 483 post_in(CHKARGS)
 484 {
 485         char    *s;
 486
 487         if (n->parent->tok != MAN_TP ||
 488             n->parent->type != ROFFT_HEAD ||
 489             n->child == NULL ||
 490             *n->child->string == '+' ||
 491             *n->child->string == '-')
 492                 return;
 493         mandoc_asprintf(&s, "+%s", n->child->string);
 494         free(n->child->string);
 495         n->child->string = s;
 496 }
 497
 498 static void
 499 post_vs(CHKARGS)
 500 {
 501
 502         if (NULL != n->prev)
 503                 return;
 504
 505         switch (n->parent->tok) {
 506         case MAN_SH:
 507         case MAN_SS:
 508         case MAN_PP:
 509                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 510                     "%s after %s", roff_name[n->tok],
 511                     roff_name[n->parent->tok]);
 512                 /* FALLTHROUGH */
 513         case TOKEN_NONE:
 514                 /*
 515                  * Don't warn about this because it occurs in pod2man
 516                  * and would cause considerable (unfixable) warnage.
 517                  */
 518                 roff_node_delete(man, n);
 519                 break;
 520         default:
 521                 break;
 522         }
 523 }