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