contrib/mdocml/man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2016 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  v_check man_valids[MAN_MAX] = {
  58         post_vs,    /* br */
  59         post_TH,    /* TH */
  60         NULL,       /* SH */
  61         NULL,       /* SS */
  62         NULL,       /* TP */
  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         post_vs,    /* sp */
  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         NULL,       /* in */
  89         post_ft,    /* ft */
  90         post_OP,    /* OP */
  91         NULL,       /* EX */
  92         NULL,       /* EE */
  93         post_UR,    /* UR */
  94         NULL,       /* UE */
  95         NULL,       /* ll */
  96 };
  97
  98
  99 void
 100 man_node_validate(struct roff_man *man)
 101 {
 102         struct roff_node *n;
 103         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                 cp = man_valids + n->tok;
 129                 if (*cp)
 130                         (*cp)(man, n);
 131                 if (man->last == n)
 132                         man_state(man, n);
 133                 break;
 134         }
 135 }
 136
 137 static void
 138 check_root(CHKARGS)
 139 {
 140
 141         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 142
 143         if (NULL == man->first->child)
 144                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 145                     n->line, n->pos, NULL);
 146         else
 147                 man->meta.hasbody = 1;
 148
 149         if (NULL == man->meta.title) {
 150                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 151                     n->line, n->pos, NULL);
 152
 153                 /*
 154                  * If a title hasn't been set, do so now (by
 155                  * implication, date and section also aren't set).
 156                  */
 157
 158                 man->meta.title = mandoc_strdup("");
 159                 man->meta.msec = mandoc_strdup("");
 160                 man->meta.date = man->quick ? mandoc_strdup("") :
 161                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 162         }
 163 }
 164
 165 static void
 166 check_text(CHKARGS)
 167 {
 168         char            *cp, *p;
 169
 170         if (MAN_LITERAL & man->flags)
 171                 return;
 172
 173         cp = n->string;
 174         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 175                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 176                     n->line, n->pos + (p - cp), NULL);
 177 }
 178
 179 static void
 180 post_OP(CHKARGS)
 181 {
 182
 183         if (n->child == NULL)
 184                 mandoc_msg(MANDOCERR_OP_EMPTY, man->parse,
 185                     n->line, n->pos, "OP");
 186         else if (n->child->next != NULL && n->child->next->next != NULL) {
 187                 n = n->child->next->next;
 188                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 189                     n->line, n->pos, "OP ... %s", n->string);
 190         }
 191 }
 192
 193 static void
 194 post_UR(CHKARGS)
 195 {
 196
 197         if (n->type == ROFFT_HEAD && n->child == NULL)
 198                 mandoc_vmsg(MANDOCERR_UR_NOHEAD, man->parse,
 199                     n->line, n->pos, "UR");
 200         check_part(man, n);
 201 }
 202
 203 static void
 204 post_ft(CHKARGS)
 205 {
 206         char    *cp;
 207         int      ok;
 208
 209         if (n->child == NULL)
 210                 return;
 211
 212         ok = 0;
 213         cp = n->child->string;
 214         switch (*cp) {
 215         case '1':
 216         case '2':
 217         case '3':
 218         case '4':
 219         case 'I':
 220         case 'P':
 221         case 'R':
 222                 if ('\0' == cp[1])
 223                         ok = 1;
 224                 break;
 225         case 'B':
 226                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 227                         ok = 1;
 228                 break;
 229         case 'C':
 230                 if ('W' == cp[1] && '\0' == cp[2])
 231                         ok = 1;
 232                 break;
 233         default:
 234                 break;
 235         }
 236
 237         if (0 == ok) {
 238                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 239                     n->line, n->pos, "ft %s", cp);
 240                 *cp = '\0';
 241         }
 242 }
 243
 244 static void
 245 check_part(CHKARGS)
 246 {
 247
 248         if (n->type == ROFFT_BODY && n->child == NULL)
 249                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 250                     n->line, n->pos, man_macronames[n->tok]);
 251 }
 252
 253 static void
 254 check_par(CHKARGS)
 255 {
 256
 257         switch (n->type) {
 258         case ROFFT_BLOCK:
 259                 if (n->body->child == NULL)
 260                         roff_node_delete(man, n);
 261                 break;
 262         case ROFFT_BODY:
 263                 if (n->child == NULL)
 264                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 265                             man->parse, n->line, n->pos,
 266                             "%s empty", man_macronames[n->tok]);
 267                 break;
 268         case ROFFT_HEAD:
 269                 if (n->child != NULL)
 270                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 271                             man->parse, n->line, n->pos,
 272                             "%s %s%s", man_macronames[n->tok],
 273                             n->child->string,
 274                             n->child->next != NULL ? " ..." : "");
 275                 break;
 276         default:
 277                 break;
 278         }
 279 }
 280
 281 static void
 282 post_IP(CHKARGS)
 283 {
 284
 285         switch (n->type) {
 286         case ROFFT_BLOCK:
 287                 if (n->head->child == NULL && n->body->child == NULL)
 288                         roff_node_delete(man, n);
 289                 break;
 290         case ROFFT_BODY:
 291                 if (n->parent->head->child == NULL && n->child == NULL)
 292                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 293                             man->parse, n->line, n->pos,
 294                             "%s empty", man_macronames[n->tok]);
 295                 break;
 296         default:
 297                 break;
 298         }
 299 }
 300
 301 static void
 302 post_TH(CHKARGS)
 303 {
 304         struct roff_node *nb;
 305         const char      *p;
 306
 307         free(man->meta.title);
 308         free(man->meta.vol);
 309         free(man->meta.os);
 310         free(man->meta.msec);
 311         free(man->meta.date);
 312
 313         man->meta.title = man->meta.vol = man->meta.date =
 314             man->meta.msec = man->meta.os = NULL;
 315
 316         nb = n;
 317
 318         /* ->TITLE<- MSEC DATE OS VOL */
 319
 320         n = n->child;
 321         if (n && n->string) {
 322                 for (p = n->string; '\0' != *p; p++) {
 323                         /* Only warn about this once... */
 324                         if (isalpha((unsigned char)*p) &&
 325                             ! isupper((unsigned char)*p)) {
 326                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 327                                     man->parse, n->line,
 328                                     n->pos + (p - n->string),
 329                                     "TH %s", n->string);
 330                                 break;
 331                         }
 332                 }
 333                 man->meta.title = mandoc_strdup(n->string);
 334         } else {
 335                 man->meta.title = mandoc_strdup("");
 336                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 337                     nb->line, nb->pos, "TH");
 338         }
 339
 340         /* TITLE ->MSEC<- DATE OS VOL */
 341
 342         if (n)
 343                 n = n->next;
 344         if (n && n->string)
 345                 man->meta.msec = mandoc_strdup(n->string);
 346         else {
 347                 man->meta.msec = mandoc_strdup("");
 348                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 349                     nb->line, nb->pos, "TH %s", man->meta.title);
 350         }
 351
 352         /* TITLE MSEC ->DATE<- OS VOL */
 353
 354         if (n)
 355                 n = n->next;
 356         if (n && n->string && '\0' != n->string[0]) {
 357                 man->meta.date = man->quick ?
 358                     mandoc_strdup(n->string) :
 359                     mandoc_normdate(man->parse, n->string,
 360                         n->line, n->pos);
 361         } else {
 362                 man->meta.date = mandoc_strdup("");
 363                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 364                     n ? n->line : nb->line,
 365                     n ? n->pos : nb->pos, "TH");
 366         }
 367
 368         /* TITLE MSEC DATE ->OS<- VOL */
 369
 370         if (n && (n = n->next))
 371                 man->meta.os = mandoc_strdup(n->string);
 372         else if (man->defos != NULL)
 373                 man->meta.os = mandoc_strdup(man->defos);
 374
 375         /* TITLE MSEC DATE OS ->VOL<- */
 376         /* If missing, use the default VOL name for MSEC. */
 377
 378         if (n && (n = n->next))
 379                 man->meta.vol = mandoc_strdup(n->string);
 380         else if ('\0' != man->meta.msec[0] &&
 381             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 382                 man->meta.vol = mandoc_strdup(p);
 383
 384         if (n != NULL && (n = n->next) != NULL)
 385                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 386                     n->line, n->pos, "TH ... %s", n->string);
 387
 388         /*
 389          * Remove the `TH' node after we've processed it for our
 390          * meta-data.
 391          */
 392         roff_node_delete(man, man->last);
 393 }
 394
 395 static void
 396 post_UC(CHKARGS)
 397 {
 398         static const char * const bsd_versions[] = {
 399             "3rd Berkeley Distribution",
 400             "4th Berkeley Distribution",
 401             "4.2 Berkeley Distribution",
 402             "4.3 Berkeley Distribution",
 403             "4.4 Berkeley Distribution",
 404         };
 405
 406         const char      *p, *s;
 407
 408         n = n->child;
 409
 410         if (n == NULL || n->type != ROFFT_TEXT)
 411                 p = bsd_versions[0];
 412         else {
 413                 s = n->string;
 414                 if (0 == strcmp(s, "3"))
 415                         p = bsd_versions[0];
 416                 else if (0 == strcmp(s, "4"))
 417                         p = bsd_versions[1];
 418                 else if (0 == strcmp(s, "5"))
 419                         p = bsd_versions[2];
 420                 else if (0 == strcmp(s, "6"))
 421                         p = bsd_versions[3];
 422                 else if (0 == strcmp(s, "7"))
 423                         p = bsd_versions[4];
 424                 else
 425                         p = bsd_versions[0];
 426         }
 427
 428         free(man->meta.os);
 429         man->meta.os = mandoc_strdup(p);
 430 }
 431
 432 static void
 433 post_AT(CHKARGS)
 434 {
 435         static const char * const unix_versions[] = {
 436             "7th Edition",
 437             "System III",
 438             "System V",
 439             "System V Release 2",
 440         };
 441
 442         struct roff_node *nn;
 443         const char      *p, *s;
 444
 445         n = n->child;
 446
 447         if (n == NULL || n->type != ROFFT_TEXT)
 448                 p = unix_versions[0];
 449         else {
 450                 s = n->string;
 451                 if (0 == strcmp(s, "3"))
 452                         p = unix_versions[0];
 453                 else if (0 == strcmp(s, "4"))
 454                         p = unix_versions[1];
 455                 else if (0 == strcmp(s, "5")) {
 456                         nn = n->next;
 457                         if (nn != NULL &&
 458                             nn->type == ROFFT_TEXT &&
 459                             nn->string[0] != '\0')
 460                                 p = unix_versions[3];
 461                         else
 462                                 p = unix_versions[2];
 463                 } else
 464                         p = unix_versions[0];
 465         }
 466
 467         free(man->meta.os);
 468         man->meta.os = mandoc_strdup(p);
 469 }
 470
 471 static void
 472 post_vs(CHKARGS)
 473 {
 474
 475         if (NULL != n->prev)
 476                 return;
 477
 478         switch (n->parent->tok) {
 479         case MAN_SH:
 480         case MAN_SS:
 481                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 482                     "%s after %s", man_macronames[n->tok],
 483                     man_macronames[n->parent->tok]);
 484                 /* FALLTHROUGH */
 485         case TOKEN_NONE:
 486                 /*
 487                  * Don't warn about this because it occurs in pod2man
 488                  * and would cause considerable (unfixable) warnage.
 489                  */
 490                 roff_node_delete(man, n);
 491                 break;
 492         default:
 493                 break;
 494         }
 495 }