man_validate.c

   1 /*      $Id: man_validate.c,v 1.106 2014/08/10 23:54:41 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 #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 int     (*v_check)(CHKARGS);
  40
  41 static  int       check_eq0(CHKARGS);
  42 static  int       check_eq2(CHKARGS);
  43 static  int       check_le1(CHKARGS);
  44 static  int       check_le5(CHKARGS);
  45 static  int       check_par(CHKARGS);
  46 static  int       check_part(CHKARGS);
  47 static  int       check_root(CHKARGS);
  48 static  int       check_text(CHKARGS);
  49
  50 static  int       post_AT(CHKARGS);
  51 static  int       post_IP(CHKARGS);
  52 static  int       post_vs(CHKARGS);
  53 static  int       post_fi(CHKARGS);
  54 static  int       post_ft(CHKARGS);
  55 static  int       post_nf(CHKARGS);
  56 static  int       post_TH(CHKARGS);
  57 static  int       post_UC(CHKARGS);
  58 static  int       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 int
 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(1);
 112         n->flags |= MAN_VALID;
 113
 114         switch (n->type) {
 115         case MAN_TEXT:
 116                 return(check_text(man, n));
 117         case MAN_ROOT:
 118                 return(check_root(man, n));
 119         case MAN_EQN:
 120                 /* FALLTHROUGH */
 121         case MAN_TBL:
 122                 return(1);
 123         default:
 124                 cp = man_valids + n->tok;
 125                 return(*cp ? (*cp)(man, n) : 1);
 126         }
 127 }
 128
 129 static int
 130 check_root(CHKARGS)
 131 {
 132
 133         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 134
 135         if (NULL == man->first->child)
 136                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 137                     n->line, n->pos, NULL);
 138         else
 139                 man->meta.hasbody = 1;
 140
 141         if (NULL == man->meta.title) {
 142                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 143                     n->line, n->pos, NULL);
 144
 145                 /*
 146                  * If a title hasn't been set, do so now (by
 147                  * implication, date and section also aren't set).
 148                  */
 149
 150                 man->meta.title = mandoc_strdup("");
 151                 man->meta.msec = mandoc_strdup("");
 152                 man->meta.date = man->quick ? mandoc_strdup("") :
 153                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 154         }
 155
 156         return(1);
 157 }
 158
 159 static int
 160 check_text(CHKARGS)
 161 {
 162         char            *cp, *p;
 163
 164         if (MAN_LITERAL & man->flags)
 165                 return(1);
 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         return(1);
 172 }
 173
 174 #define INEQ_DEFINE(x, ineq, name) \
 175 static int \
 176 check_##name(CHKARGS) \
 177 { \
 178         if (n->nchild ineq (x)) \
 179                 return(1); \
 180         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 181             "line arguments %s %d (have %d)", \
 182             #ineq, (x), n->nchild); \
 183         return(1); \
 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 int
 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
 199         return(check_part(man, n));
 200 }
 201
 202 static int
 203 post_ft(CHKARGS)
 204 {
 205         char    *cp;
 206         int      ok;
 207
 208         if (0 == n->nchild)
 209                 return(1);
 210
 211         ok = 0;
 212         cp = n->child->string;
 213         switch (*cp) {
 214         case '1':
 215                 /* FALLTHROUGH */
 216         case '2':
 217                 /* FALLTHROUGH */
 218         case '3':
 219                 /* FALLTHROUGH */
 220         case '4':
 221                 /* FALLTHROUGH */
 222         case 'I':
 223                 /* FALLTHROUGH */
 224         case 'P':
 225                 /* FALLTHROUGH */
 226         case 'R':
 227                 if ('\0' == cp[1])
 228                         ok = 1;
 229                 break;
 230         case 'B':
 231                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 232                         ok = 1;
 233                 break;
 234         case 'C':
 235                 if ('W' == cp[1] && '\0' == cp[2])
 236                         ok = 1;
 237                 break;
 238         default:
 239                 break;
 240         }
 241
 242         if (0 == ok) {
 243                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 244                     n->line, n->pos, "ft %s", cp);
 245                 *cp = '\0';
 246         }
 247
 248         if (1 < n->nchild)
 249                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 250                     n->pos, "want one child (have %d)", n->nchild);
 251
 252         return(1);
 253 }
 254
 255 static int
 256 check_part(CHKARGS)
 257 {
 258
 259         if (MAN_BODY == n->type && 0 == n->nchild)
 260                 mandoc_msg(MANDOCERR_ARGCWARN, man->parse, n->line,
 261                     n->pos, "want children (have none)");
 262
 263         return(1);
 264 }
 265
 266 static int
 267 check_par(CHKARGS)
 268 {
 269
 270         switch (n->type) {
 271         case MAN_BLOCK:
 272                 if (0 == n->body->nchild)
 273                         man_node_delete(man, n);
 274                 break;
 275         case MAN_BODY:
 276                 if (0 == n->nchild)
 277                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 278                             man->parse, n->line, n->pos,
 279                             "%s empty", man_macronames[n->tok]);
 280                 break;
 281         case MAN_HEAD:
 282                 if (n->nchild)
 283                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 284                             man->parse, n->line, n->pos,
 285                             "%s %s%s", man_macronames[n->tok],
 286                             n->child->string,
 287                             n->nchild > 1 ? " ..." : "");
 288                 break;
 289         default:
 290                 break;
 291         }
 292
 293         return(1);
 294 }
 295
 296 static int
 297 post_IP(CHKARGS)
 298 {
 299
 300         switch (n->type) {
 301         case MAN_BLOCK:
 302                 if (0 == n->head->nchild && 0 == n->body->nchild)
 303                         man_node_delete(man, n);
 304                 break;
 305         case MAN_BODY:
 306                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 307                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 308                             man->parse, n->line, n->pos,
 309                             "%s empty", man_macronames[n->tok]);
 310                 break;
 311         default:
 312                 break;
 313         }
 314         return(1);
 315 }
 316
 317 static int
 318 post_TH(CHKARGS)
 319 {
 320         struct man_node *nb;
 321         const char      *p;
 322
 323         check_le5(man, n);
 324
 325         free(man->meta.title);
 326         free(man->meta.vol);
 327         free(man->meta.source);
 328         free(man->meta.msec);
 329         free(man->meta.date);
 330
 331         man->meta.title = man->meta.vol = man->meta.date =
 332             man->meta.msec = man->meta.source = NULL;
 333
 334         nb = n;
 335
 336         /* ->TITLE<- MSEC DATE SOURCE VOL */
 337
 338         n = n->child;
 339         if (n && n->string) {
 340                 for (p = n->string; '\0' != *p; p++) {
 341                         /* Only warn about this once... */
 342                         if (isalpha((unsigned char)*p) &&
 343                             ! isupper((unsigned char)*p)) {
 344                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 345                                     man->parse, n->line,
 346                                     n->pos + (p - n->string),
 347                                     "TH %s", n->string);
 348                                 break;
 349                         }
 350                 }
 351                 man->meta.title = mandoc_strdup(n->string);
 352         } else {
 353                 man->meta.title = mandoc_strdup("");
 354                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 355                     nb->line, nb->pos, "TH");
 356         }
 357
 358         /* TITLE ->MSEC<- DATE SOURCE VOL */
 359
 360         if (n)
 361                 n = n->next;
 362         if (n && n->string)
 363                 man->meta.msec = mandoc_strdup(n->string);
 364         else {
 365                 man->meta.msec = mandoc_strdup("");
 366                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 367                     nb->line, nb->pos, "TH %s", man->meta.title);
 368         }
 369
 370         /* TITLE MSEC ->DATE<- SOURCE VOL */
 371
 372         if (n)
 373                 n = n->next;
 374         if (n && n->string && '\0' != n->string[0]) {
 375                 man->meta.date = man->quick ?
 376                     mandoc_strdup(n->string) :
 377                     mandoc_normdate(man->parse, n->string,
 378                         n->line, n->pos);
 379         } else {
 380                 man->meta.date = mandoc_strdup("");
 381                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 382                     n ? n->line : nb->line,
 383                     n ? n->pos : nb->pos, "TH");
 384         }
 385
 386         /* TITLE MSEC DATE ->SOURCE<- VOL */
 387
 388         if (n && (n = n->next))
 389                 man->meta.source = mandoc_strdup(n->string);
 390
 391         /* TITLE MSEC DATE SOURCE ->VOL<- */
 392         /* If missing, use the default VOL name for MSEC. */
 393
 394         if (n && (n = n->next))
 395                 man->meta.vol = mandoc_strdup(n->string);
 396         else if ('\0' != man->meta.msec[0] &&
 397             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 398                 man->meta.vol = mandoc_strdup(p);
 399
 400         /*
 401          * Remove the `TH' node after we've processed it for our
 402          * meta-data.
 403          */
 404         man_node_delete(man, man->last);
 405         return(1);
 406 }
 407
 408 static int
 409 post_nf(CHKARGS)
 410 {
 411
 412         check_eq0(man, n);
 413
 414         if (MAN_LITERAL & man->flags)
 415                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 416                     n->line, n->pos, "nf");
 417
 418         man->flags |= MAN_LITERAL;
 419         return(1);
 420 }
 421
 422 static int
 423 post_fi(CHKARGS)
 424 {
 425
 426         check_eq0(man, n);
 427
 428         if ( ! (MAN_LITERAL & man->flags))
 429                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 430                     n->line, n->pos, "fi");
 431
 432         man->flags &= ~MAN_LITERAL;
 433         return(1);
 434 }
 435
 436 static int
 437 post_UC(CHKARGS)
 438 {
 439         static const char * const bsd_versions[] = {
 440             "3rd Berkeley Distribution",
 441             "4th Berkeley Distribution",
 442             "4.2 Berkeley Distribution",
 443             "4.3 Berkeley Distribution",
 444             "4.4 Berkeley Distribution",
 445         };
 446
 447         const char      *p, *s;
 448
 449         n = n->child;
 450
 451         if (NULL == n || MAN_TEXT != n->type)
 452                 p = bsd_versions[0];
 453         else {
 454                 s = n->string;
 455                 if (0 == strcmp(s, "3"))
 456                         p = bsd_versions[0];
 457                 else if (0 == strcmp(s, "4"))
 458                         p = bsd_versions[1];
 459                 else if (0 == strcmp(s, "5"))
 460                         p = bsd_versions[2];
 461                 else if (0 == strcmp(s, "6"))
 462                         p = bsd_versions[3];
 463                 else if (0 == strcmp(s, "7"))
 464                         p = bsd_versions[4];
 465                 else
 466                         p = bsd_versions[0];
 467         }
 468
 469         free(man->meta.source);
 470         man->meta.source = mandoc_strdup(p);
 471         return(1);
 472 }
 473
 474 static int
 475 post_AT(CHKARGS)
 476 {
 477         static const char * const unix_versions[] = {
 478             "7th Edition",
 479             "System III",
 480             "System V",
 481             "System V Release 2",
 482         };
 483
 484         const char      *p, *s;
 485         struct man_node *nn;
 486
 487         n = n->child;
 488
 489         if (NULL == n || MAN_TEXT != n->type)
 490                 p = unix_versions[0];
 491         else {
 492                 s = n->string;
 493                 if (0 == strcmp(s, "3"))
 494                         p = unix_versions[0];
 495                 else if (0 == strcmp(s, "4"))
 496                         p = unix_versions[1];
 497                 else if (0 == strcmp(s, "5")) {
 498                         nn = n->next;
 499                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 500                                 p = unix_versions[3];
 501                         else
 502                                 p = unix_versions[2];
 503                 } else
 504                         p = unix_versions[0];
 505         }
 506
 507         free(man->meta.source);
 508         man->meta.source = mandoc_strdup(p);
 509         return(1);
 510 }
 511
 512 static int
 513 post_vs(CHKARGS)
 514 {
 515
 516         if (n->tok == MAN_br)
 517                 check_eq0(man, n);
 518         else
 519                 check_le1(man, n);
 520
 521         if (NULL != n->prev)
 522                 return(1);
 523
 524         switch (n->parent->tok) {
 525         case MAN_SH:
 526                 /* FALLTHROUGH */
 527         case MAN_SS:
 528                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 529                     "%s after %s", man_macronames[n->tok],
 530                     man_macronames[n->parent->tok]);
 531                 /* FALLTHROUGH */
 532         case MAN_MAX:
 533                 /*
 534                  * Don't warn about this because it occurs in pod2man
 535                  * and would cause considerable (unfixable) warnage.
 536                  */
 537                 man_node_delete(man, n);
 538                 break;
 539         default:
 540                 break;
 541         }
 542
 543         return(1);
 544 }