contrib/mdocml/man_validate.c

   1 /*      $Id: man_validate.c,v 1.105 2014/08/06 15:09:05 schwarze Exp $ */
   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 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <sys/types.h>
  23
  24 #include <assert.h>
  25 #include <ctype.h>
  26 #include <errno.h>
  27 #include <limits.h>
  28 #include <stdarg.h>
  29 #include <stdlib.h>
  30 #include <string.h>
  31 #include <time.h>
  32
  33 #include "man.h"
  34 #include "mandoc.h"
  35 #include "mandoc_aux.h"
  36 #include "libman.h"
  37 #include "libmandoc.h"
  38
  39 #define CHKARGS   struct man *man, struct man_node *n
  40
  41 typedef int     (*v_check)(CHKARGS);
  42
  43 static  int       check_eq0(CHKARGS);
  44 static  int       check_eq2(CHKARGS);
  45 static  int       check_le1(CHKARGS);
  46 static  int       check_le5(CHKARGS);
  47 static  int       check_par(CHKARGS);
  48 static  int       check_part(CHKARGS);
  49 static  int       check_root(CHKARGS);
  50 static  int       check_text(CHKARGS);
  51
  52 static  int       post_AT(CHKARGS);
  53 static  int       post_IP(CHKARGS);
  54 static  int       post_vs(CHKARGS);
  55 static  int       post_fi(CHKARGS);
  56 static  int       post_ft(CHKARGS);
  57 static  int       post_nf(CHKARGS);
  58 static  int       post_TH(CHKARGS);
  59 static  int       post_UC(CHKARGS);
  60 static  int       post_UR(CHKARGS);
  61
  62 static  v_check man_valids[MAN_MAX] = {
  63         post_vs,    /* br */
  64         post_TH,    /* TH */
  65         NULL,       /* SH */
  66         NULL,       /* SS */
  67         NULL,       /* TP */
  68         check_par,  /* LP */
  69         check_par,  /* PP */
  70         check_par,  /* P */
  71         post_IP,    /* IP */
  72         NULL,       /* HP */
  73         NULL,       /* SM */
  74         NULL,       /* SB */
  75         NULL,       /* BI */
  76         NULL,       /* IB */
  77         NULL,       /* BR */
  78         NULL,       /* RB */
  79         NULL,       /* R */
  80         NULL,       /* B */
  81         NULL,       /* I */
  82         NULL,       /* IR */
  83         NULL,       /* RI */
  84         check_eq0,  /* na */
  85         post_vs,    /* sp */
  86         post_nf,    /* nf */
  87         post_fi,    /* fi */
  88         NULL,       /* RE */
  89         check_part, /* RS */
  90         NULL,       /* DT */
  91         post_UC,    /* UC */
  92         check_le1,  /* PD */
  93         post_AT,    /* AT */
  94         NULL,       /* in */
  95         post_ft,    /* ft */
  96         check_eq2,  /* OP */
  97         post_nf,    /* EX */
  98         post_fi,    /* EE */
  99         post_UR,    /* UR */
 100         NULL,       /* UE */
 101         NULL,       /* ll */
 102 };
 103
 104
 105 int
 106 man_valid_post(struct man *man)
 107 {
 108         struct man_node *n;
 109         v_check         *cp;
 110
 111         n = man->last;
 112         if (n->flags & MAN_VALID)
 113                 return(1);
 114         n->flags |= MAN_VALID;
 115
 116         switch (n->type) {
 117         case MAN_TEXT:
 118                 return(check_text(man, n));
 119         case MAN_ROOT:
 120                 return(check_root(man, n));
 121         case MAN_EQN:
 122                 /* FALLTHROUGH */
 123         case MAN_TBL:
 124                 return(1);
 125         default:
 126                 cp = man_valids + n->tok;
 127                 return(*cp ? (*cp)(man, n) : 1);
 128         }
 129 }
 130
 131 static int
 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         return(1);
 159 }
 160
 161 static int
 162 check_text(CHKARGS)
 163 {
 164         char            *cp, *p;
 165
 166         if (MAN_LITERAL & man->flags)
 167                 return(1);
 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         return(1);
 174 }
 175
 176 #define INEQ_DEFINE(x, ineq, name) \
 177 static int \
 178 check_##name(CHKARGS) \
 179 { \
 180         if (n->nchild ineq (x)) \
 181                 return(1); \
 182         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 183             "line arguments %s %d (have %d)", \
 184             #ineq, (x), n->nchild); \
 185         return(1); \
 186 }
 187
 188 INEQ_DEFINE(0, ==, eq0)
 189 INEQ_DEFINE(2, ==, eq2)
 190 INEQ_DEFINE(1, <=, le1)
 191 INEQ_DEFINE(5, <=, le5)
 192
 193 static int
 194 post_UR(CHKARGS)
 195 {
 196
 197         if (MAN_HEAD == n->type && 1 != n->nchild)
 198                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 199                     n->pos, "line arguments eq 1 (have %d)", n->nchild);
 200
 201         return(check_part(man, n));
 202 }
 203
 204 static int
 205 post_ft(CHKARGS)
 206 {
 207         char    *cp;
 208         int      ok;
 209
 210         if (0 == n->nchild)
 211                 return(1);
 212
 213         ok = 0;
 214         cp = n->child->string;
 215         switch (*cp) {
 216         case '1':
 217                 /* FALLTHROUGH */
 218         case '2':
 219                 /* FALLTHROUGH */
 220         case '3':
 221                 /* FALLTHROUGH */
 222         case '4':
 223                 /* FALLTHROUGH */
 224         case 'I':
 225                 /* FALLTHROUGH */
 226         case 'P':
 227                 /* FALLTHROUGH */
 228         case 'R':
 229                 if ('\0' == cp[1])
 230                         ok = 1;
 231                 break;
 232         case 'B':
 233                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 234                         ok = 1;
 235                 break;
 236         case 'C':
 237                 if ('W' == cp[1] && '\0' == cp[2])
 238                         ok = 1;
 239                 break;
 240         default:
 241                 break;
 242         }
 243
 244         if (0 == ok) {
 245                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 246                     n->line, n->pos, "ft %s", cp);
 247                 *cp = '\0';
 248         }
 249
 250         if (1 < n->nchild)
 251                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 252                     n->pos, "want one child (have %d)", n->nchild);
 253
 254         return(1);
 255 }
 256
 257 static int
 258 check_part(CHKARGS)
 259 {
 260
 261         if (MAN_BODY == n->type && 0 == n->nchild)
 262                 mandoc_msg(MANDOCERR_ARGCWARN, man->parse, n->line,
 263                     n->pos, "want children (have none)");
 264
 265         return(1);
 266 }
 267
 268 static int
 269 check_par(CHKARGS)
 270 {
 271
 272         switch (n->type) {
 273         case MAN_BLOCK:
 274                 if (0 == n->body->nchild)
 275                         man_node_delete(man, n);
 276                 break;
 277         case MAN_BODY:
 278                 if (0 == n->nchild)
 279                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 280                             man->parse, n->line, n->pos,
 281                             "%s empty", man_macronames[n->tok]);
 282                 break;
 283         case MAN_HEAD:
 284                 if (n->nchild)
 285                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 286                             man->parse, n->line, n->pos,
 287                             "%s %s%s", man_macronames[n->tok],
 288                             n->child->string,
 289                             n->nchild > 1 ? " ..." : "");
 290                 break;
 291         default:
 292                 break;
 293         }
 294
 295         return(1);
 296 }
 297
 298 static int
 299 post_IP(CHKARGS)
 300 {
 301
 302         switch (n->type) {
 303         case MAN_BLOCK:
 304                 if (0 == n->head->nchild && 0 == n->body->nchild)
 305                         man_node_delete(man, n);
 306                 break;
 307         case MAN_BODY:
 308                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 309                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 310                             man->parse, n->line, n->pos,
 311                             "%s empty", man_macronames[n->tok]);
 312                 break;
 313         default:
 314                 break;
 315         }
 316         return(1);
 317 }
 318
 319 static int
 320 post_TH(CHKARGS)
 321 {
 322         struct man_node *nb;
 323         const char      *p;
 324
 325         check_le5(man, n);
 326
 327         free(man->meta.title);
 328         free(man->meta.vol);
 329         free(man->meta.source);
 330         free(man->meta.msec);
 331         free(man->meta.date);
 332
 333         man->meta.title = man->meta.vol = man->meta.date =
 334             man->meta.msec = man->meta.source = NULL;
 335
 336         nb = n;
 337
 338         /* ->TITLE<- MSEC DATE SOURCE VOL */
 339
 340         n = n->child;
 341         if (n && n->string) {
 342                 for (p = n->string; '\0' != *p; p++) {
 343                         /* Only warn about this once... */
 344                         if (isalpha((unsigned char)*p) &&
 345                             ! isupper((unsigned char)*p)) {
 346                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 347                                     man->parse, n->line,
 348                                     n->pos + (p - n->string),
 349                                     "TH %s", n->string);
 350                                 break;
 351                         }
 352                 }
 353                 man->meta.title = mandoc_strdup(n->string);
 354         } else {
 355                 man->meta.title = mandoc_strdup("");
 356                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 357                     nb->line, nb->pos, "TH");
 358         }
 359
 360         /* TITLE ->MSEC<- DATE SOURCE VOL */
 361
 362         if (n)
 363                 n = n->next;
 364         if (n && n->string)
 365                 man->meta.msec = mandoc_strdup(n->string);
 366         else {
 367                 man->meta.msec = mandoc_strdup("");
 368                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 369                     nb->line, nb->pos, "TH %s", man->meta.title);
 370         }
 371
 372         /* TITLE MSEC ->DATE<- SOURCE VOL */
 373
 374         if (n)
 375                 n = n->next;
 376         if (n && n->string && '\0' != n->string[0]) {
 377                 man->meta.date = man->quick ?
 378                     mandoc_strdup(n->string) :
 379                     mandoc_normdate(man->parse, n->string,
 380                         n->line, n->pos);
 381         } else {
 382                 man->meta.date = mandoc_strdup("");
 383                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 384                     n ? n->line : nb->line,
 385                     n ? n->pos : nb->pos, "TH");
 386         }
 387
 388         /* TITLE MSEC DATE ->SOURCE<- VOL */
 389
 390         if (n && (n = n->next))
 391                 man->meta.source = mandoc_strdup(n->string);
 392
 393         /* TITLE MSEC DATE SOURCE ->VOL<- */
 394         /* If missing, use the default VOL name for MSEC. */
 395
 396         if (n && (n = n->next))
 397                 man->meta.vol = mandoc_strdup(n->string);
 398         else if ('\0' != man->meta.msec[0] &&
 399             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 400                 man->meta.vol = mandoc_strdup(p);
 401
 402         /*
 403          * Remove the `TH' node after we've processed it for our
 404          * meta-data.
 405          */
 406         man_node_delete(man, man->last);
 407         return(1);
 408 }
 409
 410 static int
 411 post_nf(CHKARGS)
 412 {
 413
 414         check_eq0(man, n);
 415
 416         if (MAN_LITERAL & man->flags)
 417                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 418                     n->line, n->pos, "nf");
 419
 420         man->flags |= MAN_LITERAL;
 421         return(1);
 422 }
 423
 424 static int
 425 post_fi(CHKARGS)
 426 {
 427
 428         check_eq0(man, n);
 429
 430         if ( ! (MAN_LITERAL & man->flags))
 431                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 432                     n->line, n->pos, "fi");
 433
 434         man->flags &= ~MAN_LITERAL;
 435         return(1);
 436 }
 437
 438 static int
 439 post_UC(CHKARGS)
 440 {
 441         static const char * const bsd_versions[] = {
 442             "3rd Berkeley Distribution",
 443             "4th Berkeley Distribution",
 444             "4.2 Berkeley Distribution",
 445             "4.3 Berkeley Distribution",
 446             "4.4 Berkeley Distribution",
 447         };
 448
 449         const char      *p, *s;
 450
 451         n = n->child;
 452
 453         if (NULL == n || MAN_TEXT != n->type)
 454                 p = bsd_versions[0];
 455         else {
 456                 s = n->string;
 457                 if (0 == strcmp(s, "3"))
 458                         p = bsd_versions[0];
 459                 else if (0 == strcmp(s, "4"))
 460                         p = bsd_versions[1];
 461                 else if (0 == strcmp(s, "5"))
 462                         p = bsd_versions[2];
 463                 else if (0 == strcmp(s, "6"))
 464                         p = bsd_versions[3];
 465                 else if (0 == strcmp(s, "7"))
 466                         p = bsd_versions[4];
 467                 else
 468                         p = bsd_versions[0];
 469         }
 470
 471         free(man->meta.source);
 472         man->meta.source = mandoc_strdup(p);
 473         return(1);
 474 }
 475
 476 static int
 477 post_AT(CHKARGS)
 478 {
 479         static const char * const unix_versions[] = {
 480             "7th Edition",
 481             "System III",
 482             "System V",
 483             "System V Release 2",
 484         };
 485
 486         const char      *p, *s;
 487         struct man_node *nn;
 488
 489         n = n->child;
 490
 491         if (NULL == n || MAN_TEXT != n->type)
 492                 p = unix_versions[0];
 493         else {
 494                 s = n->string;
 495                 if (0 == strcmp(s, "3"))
 496                         p = unix_versions[0];
 497                 else if (0 == strcmp(s, "4"))
 498                         p = unix_versions[1];
 499                 else if (0 == strcmp(s, "5")) {
 500                         nn = n->next;
 501                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 502                                 p = unix_versions[3];
 503                         else
 504                                 p = unix_versions[2];
 505                 } else
 506                         p = unix_versions[0];
 507         }
 508
 509         free(man->meta.source);
 510         man->meta.source = mandoc_strdup(p);
 511         return(1);
 512 }
 513
 514 static int
 515 post_vs(CHKARGS)
 516 {
 517
 518         if (n->tok == MAN_br)
 519                 check_eq0(man, n);
 520         else
 521                 check_le1(man, n);
 522
 523         if (NULL != n->prev)
 524                 return(1);
 525
 526         switch (n->parent->tok) {
 527         case MAN_SH:
 528                 /* FALLTHROUGH */
 529         case MAN_SS:
 530                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 531                     "%s after %s", man_macronames[n->tok],
 532                     man_macronames[n->parent->tok]);
 533                 /* FALLTHROUGH */
 534         case MAN_MAX:
 535                 /*
 536                  * Don't warn about this because it occurs in pod2man
 537                  * and would cause considerable (unfixable) warnage.
 538                  */
 539                 man_node_delete(man, n);
 540                 break;
 541         default:
 542                 break;
 543         }
 544
 545         return(1);
 546 }