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