man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012, 2013, 2014 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 AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR 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 "man.h"
  32 #include "mandoc.h"
  33 #include "mandoc_aux.h"
  34 #include "libman.h"
  35 #include "libmandoc.h"
  36
  37 #define CHKARGS   struct man *man, struct man_node *n
  38
  39 typedef void    (*v_check)(CHKARGS);
  40
  41 static  void      check_eq2(CHKARGS);
  42 static  void      check_le1(CHKARGS);
  43 static  void      check_le5(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_vs(CHKARGS);
  52 static  void      post_fi(CHKARGS);
  53 static  void      post_ft(CHKARGS);
  54 static  void      post_nf(CHKARGS);
  55 static  void      post_TH(CHKARGS);
  56 static  void      post_UC(CHKARGS);
  57 static  void      post_UR(CHKARGS);
  58
  59 static  v_check man_valids[MAN_MAX] = {
  60         post_vs,    /* br */
  61         post_TH,    /* TH */
  62         NULL,       /* SH */
  63         NULL,       /* SS */
  64         NULL,       /* TP */
  65         check_par,  /* LP */
  66         check_par,  /* PP */
  67         check_par,  /* P */
  68         post_IP,    /* IP */
  69         NULL,       /* HP */
  70         NULL,       /* SM */
  71         NULL,       /* SB */
  72         NULL,       /* BI */
  73         NULL,       /* IB */
  74         NULL,       /* BR */
  75         NULL,       /* RB */
  76         NULL,       /* R */
  77         NULL,       /* B */
  78         NULL,       /* I */
  79         NULL,       /* IR */
  80         NULL,       /* RI */
  81         post_vs,    /* sp */
  82         post_nf,    /* nf */
  83         post_fi,    /* fi */
  84         NULL,       /* RE */
  85         check_part, /* RS */
  86         NULL,       /* DT */
  87         post_UC,    /* UC */
  88         check_le1,  /* PD */
  89         post_AT,    /* AT */
  90         NULL,       /* in */
  91         post_ft,    /* ft */
  92         check_eq2,  /* OP */
  93         post_nf,    /* EX */
  94         post_fi,    /* EE */
  95         post_UR,    /* UR */
  96         NULL,       /* UE */
  97         NULL,       /* ll */
  98 };
  99
 100
 101 void
 102 man_valid_post(struct man *man)
 103 {
 104         struct man_node *n;
 105         v_check         *cp;
 106
 107         n = man->last;
 108         if (n->flags & MAN_VALID)
 109                 return;
 110         n->flags |= MAN_VALID;
 111
 112         switch (n->type) {
 113         case MAN_TEXT:
 114                 check_text(man, n);
 115                 break;
 116         case MAN_ROOT:
 117                 check_root(man, n);
 118                 break;
 119         case MAN_EQN:
 120                 /* FALLTHROUGH */
 121         case MAN_TBL:
 122                 break;
 123         default:
 124                 cp = man_valids + n->tok;
 125                 if (*cp)
 126                         (*cp)(man, n);
 127                 break;
 128         }
 129 }
 130
 131 static void
 132 check_root(CHKARGS)
 133 {
 134
 135         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 136
 137         if (NULL == man->first->child)
 138                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 139                     n->line, n->pos, NULL);
 140         else
 141                 man->meta.hasbody = 1;
 142
 143         if (NULL == man->meta.title) {
 144                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 145                     n->line, n->pos, NULL);
 146
 147                 /*
 148                  * If a title hasn't been set, do so now (by
 149                  * implication, date and section also aren't set).
 150                  */
 151
 152                 man->meta.title = mandoc_strdup("");
 153                 man->meta.msec = mandoc_strdup("");
 154                 man->meta.date = man->quick ? mandoc_strdup("") :
 155                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 156         }
 157 }
 158
 159 static void
 160 check_text(CHKARGS)
 161 {
 162         char            *cp, *p;
 163
 164         if (MAN_LITERAL & man->flags)
 165                 return;
 166
 167         cp = n->string;
 168         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 169                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 170                     n->line, n->pos + (p - cp), NULL);
 171 }
 172
 173 #define INEQ_DEFINE(x, ineq, name) \
 174 static void \
 175 check_##name(CHKARGS) \
 176 { \
 177         if (n->nchild ineq (x)) \
 178                 return; \
 179         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 180             "line arguments %s %d (have %d)", \
 181             #ineq, (x), n->nchild); \
 182 }
 183
 184 INEQ_DEFINE(2, ==, eq2)
 185 INEQ_DEFINE(1, <=, le1)
 186 INEQ_DEFINE(5, <=, le5)
 187
 188 static void
 189 post_UR(CHKARGS)
 190 {
 191
 192         if (MAN_HEAD == n->type && 1 != n->nchild)
 193                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 194                     n->pos, "line arguments eq 1 (have %d)", n->nchild);
 195         check_part(man, n);
 196 }
 197
 198 static void
 199 post_ft(CHKARGS)
 200 {
 201         char    *cp;
 202         int      ok;
 203
 204         if (0 == n->nchild)
 205                 return;
 206
 207         ok = 0;
 208         cp = n->child->string;
 209         switch (*cp) {
 210         case '1':
 211                 /* FALLTHROUGH */
 212         case '2':
 213                 /* FALLTHROUGH */
 214         case '3':
 215                 /* FALLTHROUGH */
 216         case '4':
 217                 /* FALLTHROUGH */
 218         case 'I':
 219                 /* FALLTHROUGH */
 220         case 'P':
 221                 /* FALLTHROUGH */
 222         case 'R':
 223                 if ('\0' == cp[1])
 224                         ok = 1;
 225                 break;
 226         case 'B':
 227                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 228                         ok = 1;
 229                 break;
 230         case 'C':
 231                 if ('W' == cp[1] && '\0' == cp[2])
 232                         ok = 1;
 233                 break;
 234         default:
 235                 break;
 236         }
 237
 238         if (0 == ok) {
 239                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 240                     n->line, n->pos, "ft %s", cp);
 241                 *cp = '\0';
 242         }
 243
 244         if (1 < n->nchild)
 245                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 246                     n->pos, "want one child (have %d)", n->nchild);
 247 }
 248
 249 static void
 250 check_part(CHKARGS)
 251 {
 252
 253         if (n->type == MAN_BODY && n->child == NULL)
 254                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 255                     n->line, n->pos, man_macronames[n->tok]);
 256 }
 257
 258 static void
 259 check_par(CHKARGS)
 260 {
 261
 262         switch (n->type) {
 263         case MAN_BLOCK:
 264                 if (0 == n->body->nchild)
 265                         man_node_delete(man, n);
 266                 break;
 267         case MAN_BODY:
 268                 if (0 == n->nchild)
 269                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 270                             man->parse, n->line, n->pos,
 271                             "%s empty", man_macronames[n->tok]);
 272                 break;
 273         case MAN_HEAD:
 274                 if (n->nchild)
 275                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 276                             man->parse, n->line, n->pos,
 277                             "%s %s%s", man_macronames[n->tok],
 278                             n->child->string,
 279                             n->nchild > 1 ? " ..." : "");
 280                 break;
 281         default:
 282                 break;
 283         }
 284 }
 285
 286 static void
 287 post_IP(CHKARGS)
 288 {
 289
 290         switch (n->type) {
 291         case MAN_BLOCK:
 292                 if (0 == n->head->nchild && 0 == n->body->nchild)
 293                         man_node_delete(man, n);
 294                 break;
 295         case MAN_BODY:
 296                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 297                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 298                             man->parse, n->line, n->pos,
 299                             "%s empty", man_macronames[n->tok]);
 300                 break;
 301         default:
 302                 break;
 303         }
 304 }
 305
 306 static void
 307 post_TH(CHKARGS)
 308 {
 309         struct man_node *nb;
 310         const char      *p;
 311
 312         check_le5(man, n);
 313
 314         free(man->meta.title);
 315         free(man->meta.vol);
 316         free(man->meta.source);
 317         free(man->meta.msec);
 318         free(man->meta.date);
 319
 320         man->meta.title = man->meta.vol = man->meta.date =
 321             man->meta.msec = man->meta.source = NULL;
 322
 323         nb = n;
 324
 325         /* ->TITLE<- MSEC DATE SOURCE VOL */
 326
 327         n = n->child;
 328         if (n && n->string) {
 329                 for (p = n->string; '\0' != *p; p++) {
 330                         /* Only warn about this once... */
 331                         if (isalpha((unsigned char)*p) &&
 332                             ! isupper((unsigned char)*p)) {
 333                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 334                                     man->parse, n->line,
 335                                     n->pos + (p - n->string),
 336                                     "TH %s", n->string);
 337                                 break;
 338                         }
 339                 }
 340                 man->meta.title = mandoc_strdup(n->string);
 341         } else {
 342                 man->meta.title = mandoc_strdup("");
 343                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 344                     nb->line, nb->pos, "TH");
 345         }
 346
 347         /* TITLE ->MSEC<- DATE SOURCE VOL */
 348
 349         if (n)
 350                 n = n->next;
 351         if (n && n->string)
 352                 man->meta.msec = mandoc_strdup(n->string);
 353         else {
 354                 man->meta.msec = mandoc_strdup("");
 355                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 356                     nb->line, nb->pos, "TH %s", man->meta.title);
 357         }
 358
 359         /* TITLE MSEC ->DATE<- SOURCE VOL */
 360
 361         if (n)
 362                 n = n->next;
 363         if (n && n->string && '\0' != n->string[0]) {
 364                 man->meta.date = man->quick ?
 365                     mandoc_strdup(n->string) :
 366                     mandoc_normdate(man->parse, n->string,
 367                         n->line, n->pos);
 368         } else {
 369                 man->meta.date = mandoc_strdup("");
 370                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 371                     n ? n->line : nb->line,
 372                     n ? n->pos : nb->pos, "TH");
 373         }
 374
 375         /* TITLE MSEC DATE ->SOURCE<- VOL */
 376
 377         if (n && (n = n->next))
 378                 man->meta.source = mandoc_strdup(n->string);
 379         else if (man->defos != NULL)
 380                 man->meta.source = mandoc_strdup(man->defos);
 381
 382         /* TITLE MSEC DATE SOURCE ->VOL<- */
 383         /* If missing, use the default VOL name for MSEC. */
 384
 385         if (n && (n = n->next))
 386                 man->meta.vol = mandoc_strdup(n->string);
 387         else if ('\0' != man->meta.msec[0] &&
 388             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 389                 man->meta.vol = mandoc_strdup(p);
 390
 391         /*
 392          * Remove the `TH' node after we've processed it for our
 393          * meta-data.
 394          */
 395         man_node_delete(man, man->last);
 396 }
 397
 398 static void
 399 post_nf(CHKARGS)
 400 {
 401
 402         if (man->flags & MAN_LITERAL)
 403                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 404                     n->line, n->pos, "nf");
 405
 406         man->flags |= MAN_LITERAL;
 407 }
 408
 409 static void
 410 post_fi(CHKARGS)
 411 {
 412
 413         if ( ! (MAN_LITERAL & man->flags))
 414                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 415                     n->line, n->pos, "fi");
 416
 417         man->flags &= ~MAN_LITERAL;
 418 }
 419
 420 static void
 421 post_UC(CHKARGS)
 422 {
 423         static const char * const bsd_versions[] = {
 424             "3rd Berkeley Distribution",
 425             "4th Berkeley Distribution",
 426             "4.2 Berkeley Distribution",
 427             "4.3 Berkeley Distribution",
 428             "4.4 Berkeley Distribution",
 429         };
 430
 431         const char      *p, *s;
 432
 433         n = n->child;
 434
 435         if (NULL == n || MAN_TEXT != n->type)
 436                 p = bsd_versions[0];
 437         else {
 438                 s = n->string;
 439                 if (0 == strcmp(s, "3"))
 440                         p = bsd_versions[0];
 441                 else if (0 == strcmp(s, "4"))
 442                         p = bsd_versions[1];
 443                 else if (0 == strcmp(s, "5"))
 444                         p = bsd_versions[2];
 445                 else if (0 == strcmp(s, "6"))
 446                         p = bsd_versions[3];
 447                 else if (0 == strcmp(s, "7"))
 448                         p = bsd_versions[4];
 449                 else
 450                         p = bsd_versions[0];
 451         }
 452
 453         free(man->meta.source);
 454         man->meta.source = mandoc_strdup(p);
 455 }
 456
 457 static void
 458 post_AT(CHKARGS)
 459 {
 460         static const char * const unix_versions[] = {
 461             "7th Edition",
 462             "System III",
 463             "System V",
 464             "System V Release 2",
 465         };
 466
 467         const char      *p, *s;
 468         struct man_node *nn;
 469
 470         n = n->child;
 471
 472         if (NULL == n || MAN_TEXT != n->type)
 473                 p = unix_versions[0];
 474         else {
 475                 s = n->string;
 476                 if (0 == strcmp(s, "3"))
 477                         p = unix_versions[0];
 478                 else if (0 == strcmp(s, "4"))
 479                         p = unix_versions[1];
 480                 else if (0 == strcmp(s, "5")) {
 481                         nn = n->next;
 482                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 483                                 p = unix_versions[3];
 484                         else
 485                                 p = unix_versions[2];
 486                 } else
 487                         p = unix_versions[0];
 488         }
 489
 490         free(man->meta.source);
 491         man->meta.source = mandoc_strdup(p);
 492 }
 493
 494 static void
 495 post_vs(CHKARGS)
 496 {
 497
 498         if (n->tok == MAN_sp)
 499                 check_le1(man, n);
 500
 501         if (NULL != n->prev)
 502                 return;
 503
 504         switch (n->parent->tok) {
 505         case MAN_SH:
 506                 /* FALLTHROUGH */
 507         case MAN_SS:
 508                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 509                     "%s after %s", man_macronames[n->tok],
 510                     man_macronames[n->parent->tok]);
 511                 /* FALLTHROUGH */
 512         case MAN_MAX:
 513                 /*
 514                  * Don't warn about this because it occurs in pod2man
 515                  * and would cause considerable (unfixable) warnage.
 516                  */
 517                 man_node_delete(man, n);
 518                 break;
 519         default:
 520                 break;
 521         }
 522 }