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