man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2015 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 "libman.h"
  37
  38 #define CHKARGS   struct man *man, struct roff_node *n
  39
  40 typedef void    (*v_check)(CHKARGS);
  41
  42 static  void      check_par(CHKARGS);
  43 static  void      check_part(CHKARGS);
  44 static  void      check_root(CHKARGS);
  45 static  void      check_text(CHKARGS);
  46
  47 static  void      post_AT(CHKARGS);
  48 static  void      post_IP(CHKARGS);
  49 static  void      post_vs(CHKARGS);
  50 static  void      post_fi(CHKARGS);
  51 static  void      post_ft(CHKARGS);
  52 static  void      post_nf(CHKARGS);
  53 static  void      post_OP(CHKARGS);
  54 static  void      post_TH(CHKARGS);
  55 static  void      post_UC(CHKARGS);
  56 static  void      post_UR(CHKARGS);
  57
  58 static  v_check man_valids[MAN_MAX] = {
  59         post_vs,    /* br */
  60         post_TH,    /* TH */
  61         NULL,       /* SH */
  62         NULL,       /* SS */
  63         NULL,       /* TP */
  64         check_par,  /* LP */
  65         check_par,  /* PP */
  66         check_par,  /* P */
  67         post_IP,    /* IP */
  68         NULL,       /* HP */
  69         NULL,       /* SM */
  70         NULL,       /* SB */
  71         NULL,       /* BI */
  72         NULL,       /* IB */
  73         NULL,       /* BR */
  74         NULL,       /* RB */
  75         NULL,       /* R */
  76         NULL,       /* B */
  77         NULL,       /* I */
  78         NULL,       /* IR */
  79         NULL,       /* RI */
  80         post_vs,    /* sp */
  81         post_nf,    /* nf */
  82         post_fi,    /* fi */
  83         NULL,       /* RE */
  84         check_part, /* RS */
  85         NULL,       /* DT */
  86         post_UC,    /* UC */
  87         NULL,       /* PD */
  88         post_AT,    /* AT */
  89         NULL,       /* in */
  90         post_ft,    /* ft */
  91         post_OP,    /* OP */
  92         post_nf,    /* EX */
  93         post_fi,    /* EE */
  94         post_UR,    /* UR */
  95         NULL,       /* UE */
  96         NULL,       /* ll */
  97 };
  98
  99
 100 void
 101 man_valid_post(struct man *man)
 102 {
 103         struct roff_node *n;
 104         v_check         *cp;
 105
 106         n = man->last;
 107         if (n->flags & MAN_VALID)
 108                 return;
 109         n->flags |= MAN_VALID;
 110
 111         switch (n->type) {
 112         case ROFFT_TEXT:
 113                 check_text(man, n);
 114                 break;
 115         case ROFFT_ROOT:
 116                 check_root(man, n);
 117                 break;
 118         case ROFFT_EQN:
 119                 /* FALLTHROUGH */
 120         case ROFFT_TBL:
 121                 break;
 122         default:
 123                 cp = man_valids + n->tok;
 124                 if (*cp)
 125                         (*cp)(man, n);
 126                 break;
 127         }
 128 }
 129
 130 static void
 131 check_root(CHKARGS)
 132 {
 133
 134         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 135
 136         if (NULL == man->first->child)
 137                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 138                     n->line, n->pos, NULL);
 139         else
 140                 man->meta.hasbody = 1;
 141
 142         if (NULL == man->meta.title) {
 143                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 144                     n->line, n->pos, NULL);
 145
 146                 /*
 147                  * If a title hasn't been set, do so now (by
 148                  * implication, date and section also aren't set).
 149                  */
 150
 151                 man->meta.title = mandoc_strdup("");
 152                 man->meta.msec = mandoc_strdup("");
 153                 man->meta.date = man->quick ? mandoc_strdup("") :
 154                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 155         }
 156 }
 157
 158 static void
 159 check_text(CHKARGS)
 160 {
 161         char            *cp, *p;
 162
 163         if (MAN_LITERAL & man->flags)
 164                 return;
 165
 166         cp = n->string;
 167         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 168                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 169                     n->line, n->pos + (p - cp), NULL);
 170 }
 171
 172 static void
 173 post_OP(CHKARGS)
 174 {
 175
 176         if (n->nchild == 0)
 177                 mandoc_msg(MANDOCERR_OP_EMPTY, man->parse,
 178                     n->line, n->pos, "OP");
 179         else if (n->nchild > 2) {
 180                 n = n->child->next->next;
 181                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 182                     n->line, n->pos, "OP ... %s", n->string);
 183         }
 184 }
 185
 186 static void
 187 post_UR(CHKARGS)
 188 {
 189
 190         if (n->type == ROFFT_HEAD && n->child == NULL)
 191                 mandoc_vmsg(MANDOCERR_UR_NOHEAD, man->parse,
 192                     n->line, n->pos, "UR");
 193         check_part(man, n);
 194 }
 195
 196 static void
 197 post_ft(CHKARGS)
 198 {
 199         char    *cp;
 200         int      ok;
 201
 202         if (0 == n->nchild)
 203                 return;
 204
 205         ok = 0;
 206         cp = n->child->string;
 207         switch (*cp) {
 208         case '1':
 209                 /* FALLTHROUGH */
 210         case '2':
 211                 /* FALLTHROUGH */
 212         case '3':
 213                 /* FALLTHROUGH */
 214         case '4':
 215                 /* FALLTHROUGH */
 216         case 'I':
 217                 /* FALLTHROUGH */
 218         case 'P':
 219                 /* FALLTHROUGH */
 220         case 'R':
 221                 if ('\0' == cp[1])
 222                         ok = 1;
 223                 break;
 224         case 'B':
 225                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 226                         ok = 1;
 227                 break;
 228         case 'C':
 229                 if ('W' == cp[1] && '\0' == cp[2])
 230                         ok = 1;
 231                 break;
 232         default:
 233                 break;
 234         }
 235
 236         if (0 == ok) {
 237                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 238                     n->line, n->pos, "ft %s", cp);
 239                 *cp = '\0';
 240         }
 241 }
 242
 243 static void
 244 check_part(CHKARGS)
 245 {
 246
 247         if (n->type == ROFFT_BODY && n->child == NULL)
 248                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 249                     n->line, n->pos, man_macronames[n->tok]);
 250 }
 251
 252 static void
 253 check_par(CHKARGS)
 254 {
 255
 256         switch (n->type) {
 257         case ROFFT_BLOCK:
 258                 if (0 == n->body->nchild)
 259                         man_node_delete(man, n);
 260                 break;
 261         case ROFFT_BODY:
 262                 if (0 == n->nchild)
 263                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 264                             man->parse, n->line, n->pos,
 265                             "%s empty", man_macronames[n->tok]);
 266                 break;
 267         case ROFFT_HEAD:
 268                 if (n->nchild)
 269                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 270                             man->parse, n->line, n->pos,
 271                             "%s %s%s", man_macronames[n->tok],
 272                             n->child->string,
 273                             n->nchild > 1 ? " ..." : "");
 274                 break;
 275         default:
 276                 break;
 277         }
 278 }
 279
 280 static void
 281 post_IP(CHKARGS)
 282 {
 283
 284         switch (n->type) {
 285         case ROFFT_BLOCK:
 286                 if (0 == n->head->nchild && 0 == n->body->nchild)
 287                         man_node_delete(man, n);
 288                 break;
 289         case ROFFT_BODY:
 290                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 291                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 292                             man->parse, n->line, n->pos,
 293                             "%s empty", man_macronames[n->tok]);
 294                 break;
 295         default:
 296                 break;
 297         }
 298 }
 299
 300 static void
 301 post_TH(CHKARGS)
 302 {
 303         struct roff_node *nb;
 304         const char      *p;
 305
 306         free(man->meta.title);
 307         free(man->meta.vol);
 308         free(man->meta.source);
 309         free(man->meta.msec);
 310         free(man->meta.date);
 311
 312         man->meta.title = man->meta.vol = man->meta.date =
 313             man->meta.msec = man->meta.source = NULL;
 314
 315         nb = n;
 316
 317         /* ->TITLE<- MSEC DATE SOURCE VOL */
 318
 319         n = n->child;
 320         if (n && n->string) {
 321                 for (p = n->string; '\0' != *p; p++) {
 322                         /* Only warn about this once... */
 323                         if (isalpha((unsigned char)*p) &&
 324                             ! isupper((unsigned char)*p)) {
 325                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 326                                     man->parse, n->line,
 327                                     n->pos + (p - n->string),
 328                                     "TH %s", n->string);
 329                                 break;
 330                         }
 331                 }
 332                 man->meta.title = mandoc_strdup(n->string);
 333         } else {
 334                 man->meta.title = mandoc_strdup("");
 335                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 336                     nb->line, nb->pos, "TH");
 337         }
 338
 339         /* TITLE ->MSEC<- DATE SOURCE VOL */
 340
 341         if (n)
 342                 n = n->next;
 343         if (n && n->string)
 344                 man->meta.msec = mandoc_strdup(n->string);
 345         else {
 346                 man->meta.msec = mandoc_strdup("");
 347                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 348                     nb->line, nb->pos, "TH %s", man->meta.title);
 349         }
 350
 351         /* TITLE MSEC ->DATE<- SOURCE VOL */
 352
 353         if (n)
 354                 n = n->next;
 355         if (n && n->string && '\0' != n->string[0]) {
 356                 man->meta.date = man->quick ?
 357                     mandoc_strdup(n->string) :
 358                     mandoc_normdate(man->parse, n->string,
 359                         n->line, n->pos);
 360         } else {
 361                 man->meta.date = mandoc_strdup("");
 362                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 363                     n ? n->line : nb->line,
 364                     n ? n->pos : nb->pos, "TH");
 365         }
 366
 367         /* TITLE MSEC DATE ->SOURCE<- VOL */
 368
 369         if (n && (n = n->next))
 370                 man->meta.source = mandoc_strdup(n->string);
 371         else if (man->defos != NULL)
 372                 man->meta.source = mandoc_strdup(man->defos);
 373
 374         /* TITLE MSEC DATE SOURCE ->VOL<- */
 375         /* If missing, use the default VOL name for MSEC. */
 376
 377         if (n && (n = n->next))
 378                 man->meta.vol = mandoc_strdup(n->string);
 379         else if ('\0' != man->meta.msec[0] &&
 380             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 381                 man->meta.vol = mandoc_strdup(p);
 382
 383         if (n != NULL && (n = n->next) != NULL)
 384                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 385                     n->line, n->pos, "TH ... %s", n->string);
 386
 387         /*
 388          * Remove the `TH' node after we've processed it for our
 389          * meta-data.
 390          */
 391         man_node_delete(man, man->last);
 392 }
 393
 394 static void
 395 post_nf(CHKARGS)
 396 {
 397
 398         if (man->flags & MAN_LITERAL)
 399                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 400                     n->line, n->pos, "nf");
 401
 402         man->flags |= MAN_LITERAL;
 403 }
 404
 405 static void
 406 post_fi(CHKARGS)
 407 {
 408
 409         if ( ! (MAN_LITERAL & man->flags))
 410                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 411                     n->line, n->pos, "fi");
 412
 413         man->flags &= ~MAN_LITERAL;
 414 }
 415
 416 static void
 417 post_UC(CHKARGS)
 418 {
 419         static const char * const bsd_versions[] = {
 420             "3rd Berkeley Distribution",
 421             "4th Berkeley Distribution",
 422             "4.2 Berkeley Distribution",
 423             "4.3 Berkeley Distribution",
 424             "4.4 Berkeley Distribution",
 425         };
 426
 427         const char      *p, *s;
 428
 429         n = n->child;
 430
 431         if (n == NULL || n->type != ROFFT_TEXT)
 432                 p = bsd_versions[0];
 433         else {
 434                 s = n->string;
 435                 if (0 == strcmp(s, "3"))
 436                         p = bsd_versions[0];
 437                 else if (0 == strcmp(s, "4"))
 438                         p = bsd_versions[1];
 439                 else if (0 == strcmp(s, "5"))
 440                         p = bsd_versions[2];
 441                 else if (0 == strcmp(s, "6"))
 442                         p = bsd_versions[3];
 443                 else if (0 == strcmp(s, "7"))
 444                         p = bsd_versions[4];
 445                 else
 446                         p = bsd_versions[0];
 447         }
 448
 449         free(man->meta.source);
 450         man->meta.source = mandoc_strdup(p);
 451 }
 452
 453 static void
 454 post_AT(CHKARGS)
 455 {
 456         static const char * const unix_versions[] = {
 457             "7th Edition",
 458             "System III",
 459             "System V",
 460             "System V Release 2",
 461         };
 462
 463         struct roff_node *nn;
 464         const char      *p, *s;
 465
 466         n = n->child;
 467
 468         if (n == NULL || n->type != ROFFT_TEXT)
 469                 p = unix_versions[0];
 470         else {
 471                 s = n->string;
 472                 if (0 == strcmp(s, "3"))
 473                         p = unix_versions[0];
 474                 else if (0 == strcmp(s, "4"))
 475                         p = unix_versions[1];
 476                 else if (0 == strcmp(s, "5")) {
 477                         nn = n->next;
 478                         if (nn != NULL &&
 479                             nn->type == ROFFT_TEXT &&
 480                             nn->string[0] != '\0')
 481                                 p = unix_versions[3];
 482                         else
 483                                 p = unix_versions[2];
 484                 } else
 485                         p = unix_versions[0];
 486         }
 487
 488         free(man->meta.source);
 489         man->meta.source = mandoc_strdup(p);
 490 }
 491
 492 static void
 493 post_vs(CHKARGS)
 494 {
 495
 496         if (NULL != n->prev)
 497                 return;
 498
 499         switch (n->parent->tok) {
 500         case MAN_SH:
 501                 /* FALLTHROUGH */
 502         case MAN_SS:
 503                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 504                     "%s after %s", man_macronames[n->tok],
 505                     man_macronames[n->parent->tok]);
 506                 /* FALLTHROUGH */
 507         case MAN_MAX:
 508                 /*
 509                  * Don't warn about this because it occurs in pod2man
 510                  * and would cause considerable (unfixable) warnage.
 511                  */
 512                 man_node_delete(man, n);
 513                 break;
 514         default:
 515                 break;
 516         }
 517 }