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_eq0(CHKARGS);
  42 static  void      check_eq2(CHKARGS);
  43 static  void      check_le1(CHKARGS);
  44 static  void      check_le5(CHKARGS);
  45 static  void      check_par(CHKARGS);
  46 static  void      check_part(CHKARGS);
  47 static  void      check_root(CHKARGS);
  48 static  void      check_text(CHKARGS);
  49
  50 static  void      post_AT(CHKARGS);
  51 static  void      post_IP(CHKARGS);
  52 static  void      post_vs(CHKARGS);
  53 static  void      post_fi(CHKARGS);
  54 static  void      post_ft(CHKARGS);
  55 static  void      post_nf(CHKARGS);
  56 static  void      post_TH(CHKARGS);
  57 static  void      post_UC(CHKARGS);
  58 static  void      post_UR(CHKARGS);
  59
  60 static  v_check man_valids[MAN_MAX] = {
  61         post_vs,    /* br */
  62         post_TH,    /* TH */
  63         NULL,       /* SH */
  64         NULL,       /* SS */
  65         NULL,       /* TP */
  66         check_par,  /* LP */
  67         check_par,  /* PP */
  68         check_par,  /* P */
  69         post_IP,    /* IP */
  70         NULL,       /* HP */
  71         NULL,       /* SM */
  72         NULL,       /* SB */
  73         NULL,       /* BI */
  74         NULL,       /* IB */
  75         NULL,       /* BR */
  76         NULL,       /* RB */
  77         NULL,       /* R */
  78         NULL,       /* B */
  79         NULL,       /* I */
  80         NULL,       /* IR */
  81         NULL,       /* RI */
  82         check_eq0,  /* na */
  83         post_vs,    /* sp */
  84         post_nf,    /* nf */
  85         post_fi,    /* fi */
  86         NULL,       /* RE */
  87         check_part, /* RS */
  88         NULL,       /* DT */
  89         post_UC,    /* UC */
  90         check_le1,  /* PD */
  91         post_AT,    /* AT */
  92         NULL,       /* in */
  93         post_ft,    /* ft */
  94         check_eq2,  /* OP */
  95         post_nf,    /* EX */
  96         post_fi,    /* EE */
  97         post_UR,    /* UR */
  98         NULL,       /* UE */
  99         NULL,       /* ll */
 100 };
 101
 102
 103 void
 104 man_valid_post(struct man *man)
 105 {
 106         struct man_node *n;
 107         v_check         *cp;
 108
 109         n = man->last;
 110         if (n->flags & MAN_VALID)
 111                 return;
 112         n->flags |= MAN_VALID;
 113
 114         switch (n->type) {
 115         case MAN_TEXT:
 116                 check_text(man, n);
 117                 break;
 118         case MAN_ROOT:
 119                 check_root(man, n);
 120                 break;
 121         case MAN_EQN:
 122                 /* FALLTHROUGH */
 123         case MAN_TBL:
 124                 break;
 125         default:
 126                 cp = man_valids + n->tok;
 127                 if (*cp)
 128                         (*cp)(man, n);
 129                 break;
 130         }
 131 }
 132
 133 static void
 134 check_root(CHKARGS)
 135 {
 136
 137         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 138
 139         if (NULL == man->first->child)
 140                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 141                     n->line, n->pos, NULL);
 142         else
 143                 man->meta.hasbody = 1;
 144
 145         if (NULL == man->meta.title) {
 146                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 147                     n->line, n->pos, NULL);
 148
 149                 /*
 150                  * If a title hasn't been set, do so now (by
 151                  * implication, date and section also aren't set).
 152                  */
 153
 154                 man->meta.title = mandoc_strdup("");
 155                 man->meta.msec = mandoc_strdup("");
 156                 man->meta.date = man->quick ? mandoc_strdup("") :
 157                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 158         }
 159 }
 160
 161 static void
 162 check_text(CHKARGS)
 163 {
 164         char            *cp, *p;
 165
 166         if (MAN_LITERAL & man->flags)
 167                 return;
 168
 169         cp = n->string;
 170         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 171                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 172                     n->line, n->pos + (p - cp), NULL);
 173 }
 174
 175 #define INEQ_DEFINE(x, ineq, name) \
 176 static void \
 177 check_##name(CHKARGS) \
 178 { \
 179         if (n->nchild ineq (x)) \
 180                 return; \
 181         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 182             "line arguments %s %d (have %d)", \
 183             #ineq, (x), n->nchild); \
 184 }
 185
 186 INEQ_DEFINE(0, ==, eq0)
 187 INEQ_DEFINE(2, ==, eq2)
 188 INEQ_DEFINE(1, <=, le1)
 189 INEQ_DEFINE(5, <=, le5)
 190
 191 static void
 192 post_UR(CHKARGS)
 193 {
 194
 195         if (MAN_HEAD == n->type && 1 != n->nchild)
 196                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 197                     n->pos, "line arguments eq 1 (have %d)", n->nchild);
 198         check_part(man, n);
 199 }
 200
 201 static void
 202 post_ft(CHKARGS)
 203 {
 204         char    *cp;
 205         int      ok;
 206
 207         if (0 == n->nchild)
 208                 return;
 209
 210         ok = 0;
 211         cp = n->child->string;
 212         switch (*cp) {
 213         case '1':
 214                 /* FALLTHROUGH */
 215         case '2':
 216                 /* FALLTHROUGH */
 217         case '3':
 218                 /* FALLTHROUGH */
 219         case '4':
 220                 /* FALLTHROUGH */
 221         case 'I':
 222                 /* FALLTHROUGH */
 223         case 'P':
 224                 /* FALLTHROUGH */
 225         case 'R':
 226                 if ('\0' == cp[1])
 227                         ok = 1;
 228                 break;
 229         case 'B':
 230                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 231                         ok = 1;
 232                 break;
 233         case 'C':
 234                 if ('W' == cp[1] && '\0' == cp[2])
 235                         ok = 1;
 236                 break;
 237         default:
 238                 break;
 239         }
 240
 241         if (0 == ok) {
 242                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 243                     n->line, n->pos, "ft %s", cp);
 244                 *cp = '\0';
 245         }
 246
 247         if (1 < n->nchild)
 248                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 249                     n->pos, "want one child (have %d)", n->nchild);
 250 }
 251
 252 static void
 253 check_part(CHKARGS)
 254 {
 255
 256         if (MAN_BODY == n->type && 0 == n->nchild)
 257                 mandoc_msg(MANDOCERR_ARGCWARN, man->parse, n->line,
 258                     n->pos, "want children (have none)");
 259 }
 260
 261 static void
 262 check_par(CHKARGS)
 263 {
 264
 265         switch (n->type) {
 266         case MAN_BLOCK:
 267                 if (0 == n->body->nchild)
 268                         man_node_delete(man, n);
 269                 break;
 270         case MAN_BODY:
 271                 if (0 == n->nchild)
 272                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 273                             man->parse, n->line, n->pos,
 274                             "%s empty", man_macronames[n->tok]);
 275                 break;
 276         case MAN_HEAD:
 277                 if (n->nchild)
 278                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 279                             man->parse, n->line, n->pos,
 280                             "%s %s%s", man_macronames[n->tok],
 281                             n->child->string,
 282                             n->nchild > 1 ? " ..." : "");
 283                 break;
 284         default:
 285                 break;
 286         }
 287 }
 288
 289 static void
 290 post_IP(CHKARGS)
 291 {
 292
 293         switch (n->type) {
 294         case MAN_BLOCK:
 295                 if (0 == n->head->nchild && 0 == n->body->nchild)
 296                         man_node_delete(man, n);
 297                 break;
 298         case MAN_BODY:
 299                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 300                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 301                             man->parse, n->line, n->pos,
 302                             "%s empty", man_macronames[n->tok]);
 303                 break;
 304         default:
 305                 break;
 306         }
 307 }
 308
 309 static void
 310 post_TH(CHKARGS)
 311 {
 312         struct man_node *nb;
 313         const char      *p;
 314
 315         check_le5(man, n);
 316
 317         free(man->meta.title);
 318         free(man->meta.vol);
 319         free(man->meta.source);
 320         free(man->meta.msec);
 321         free(man->meta.date);
 322
 323         man->meta.title = man->meta.vol = man->meta.date =
 324             man->meta.msec = man->meta.source = NULL;
 325
 326         nb = n;
 327
 328         /* ->TITLE<- MSEC DATE SOURCE VOL */
 329
 330         n = n->child;
 331         if (n && n->string) {
 332                 for (p = n->string; '\0' != *p; p++) {
 333                         /* Only warn about this once... */
 334                         if (isalpha((unsigned char)*p) &&
 335                             ! isupper((unsigned char)*p)) {
 336                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 337                                     man->parse, n->line,
 338                                     n->pos + (p - n->string),
 339                                     "TH %s", n->string);
 340                                 break;
 341                         }
 342                 }
 343                 man->meta.title = mandoc_strdup(n->string);
 344         } else {
 345                 man->meta.title = mandoc_strdup("");
 346                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 347                     nb->line, nb->pos, "TH");
 348         }
 349
 350         /* TITLE ->MSEC<- DATE SOURCE VOL */
 351
 352         if (n)
 353                 n = n->next;
 354         if (n && n->string)
 355                 man->meta.msec = mandoc_strdup(n->string);
 356         else {
 357                 man->meta.msec = mandoc_strdup("");
 358                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 359                     nb->line, nb->pos, "TH %s", man->meta.title);
 360         }
 361
 362         /* TITLE MSEC ->DATE<- SOURCE VOL */
 363
 364         if (n)
 365                 n = n->next;
 366         if (n && n->string && '\0' != n->string[0]) {
 367                 man->meta.date = man->quick ?
 368                     mandoc_strdup(n->string) :
 369                     mandoc_normdate(man->parse, n->string,
 370                         n->line, n->pos);
 371         } else {
 372                 man->meta.date = mandoc_strdup("");
 373                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 374                     n ? n->line : nb->line,
 375                     n ? n->pos : nb->pos, "TH");
 376         }
 377
 378         /* TITLE MSEC DATE ->SOURCE<- VOL */
 379
 380         if (n && (n = n->next))
 381                 man->meta.source = mandoc_strdup(n->string);
 382         else if (man->defos != NULL)
 383                 man->meta.source = mandoc_strdup(man->defos);
 384
 385         /* TITLE MSEC DATE SOURCE ->VOL<- */
 386         /* If missing, use the default VOL name for MSEC. */
 387
 388         if (n && (n = n->next))
 389                 man->meta.vol = mandoc_strdup(n->string);
 390         else if ('\0' != man->meta.msec[0] &&
 391             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 392                 man->meta.vol = mandoc_strdup(p);
 393
 394         /*
 395          * Remove the `TH' node after we've processed it for our
 396          * meta-data.
 397          */
 398         man_node_delete(man, man->last);
 399 }
 400
 401 static void
 402 post_nf(CHKARGS)
 403 {
 404
 405         check_eq0(man, n);
 406
 407         if (MAN_LITERAL & man->flags)
 408                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 409                     n->line, n->pos, "nf");
 410
 411         man->flags |= MAN_LITERAL;
 412 }
 413
 414 static void
 415 post_fi(CHKARGS)
 416 {
 417
 418         check_eq0(man, n);
 419
 420         if ( ! (MAN_LITERAL & man->flags))
 421                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 422                     n->line, n->pos, "fi");
 423
 424         man->flags &= ~MAN_LITERAL;
 425 }
 426
 427 static void
 428 post_UC(CHKARGS)
 429 {
 430         static const char * const bsd_versions[] = {
 431             "3rd Berkeley Distribution",
 432             "4th Berkeley Distribution",
 433             "4.2 Berkeley Distribution",
 434             "4.3 Berkeley Distribution",
 435             "4.4 Berkeley Distribution",
 436         };
 437
 438         const char      *p, *s;
 439
 440         n = n->child;
 441
 442         if (NULL == n || MAN_TEXT != n->type)
 443                 p = bsd_versions[0];
 444         else {
 445                 s = n->string;
 446                 if (0 == strcmp(s, "3"))
 447                         p = bsd_versions[0];
 448                 else if (0 == strcmp(s, "4"))
 449                         p = bsd_versions[1];
 450                 else if (0 == strcmp(s, "5"))
 451                         p = bsd_versions[2];
 452                 else if (0 == strcmp(s, "6"))
 453                         p = bsd_versions[3];
 454                 else if (0 == strcmp(s, "7"))
 455                         p = bsd_versions[4];
 456                 else
 457                         p = bsd_versions[0];
 458         }
 459
 460         free(man->meta.source);
 461         man->meta.source = mandoc_strdup(p);
 462 }
 463
 464 static void
 465 post_AT(CHKARGS)
 466 {
 467         static const char * const unix_versions[] = {
 468             "7th Edition",
 469             "System III",
 470             "System V",
 471             "System V Release 2",
 472         };
 473
 474         const char      *p, *s;
 475         struct man_node *nn;
 476
 477         n = n->child;
 478
 479         if (NULL == n || MAN_TEXT != n->type)
 480                 p = unix_versions[0];
 481         else {
 482                 s = n->string;
 483                 if (0 == strcmp(s, "3"))
 484                         p = unix_versions[0];
 485                 else if (0 == strcmp(s, "4"))
 486                         p = unix_versions[1];
 487                 else if (0 == strcmp(s, "5")) {
 488                         nn = n->next;
 489                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 490                                 p = unix_versions[3];
 491                         else
 492                                 p = unix_versions[2];
 493                 } else
 494                         p = unix_versions[0];
 495         }
 496
 497         free(man->meta.source);
 498         man->meta.source = mandoc_strdup(p);
 499 }
 500
 501 static void
 502 post_vs(CHKARGS)
 503 {
 504
 505         if (n->tok == MAN_br)
 506                 check_eq0(man, n);
 507         else
 508                 check_le1(man, n);
 509
 510         if (NULL != n->prev)
 511                 return;
 512
 513         switch (n->parent->tok) {
 514         case MAN_SH:
 515                 /* FALLTHROUGH */
 516         case MAN_SS:
 517                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 518                     "%s after %s", man_macronames[n->tok],
 519                     man_macronames[n->parent->tok]);
 520                 /* FALLTHROUGH */
 521         case MAN_MAX:
 522                 /*
 523                  * Don't warn about this because it occurs in pod2man
 524                  * and would cause considerable (unfixable) warnage.
 525                  */
 526                 man_node_delete(man, n);
 527                 break;
 528         default:
 529                 break;
 530         }
 531 }