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 "roff_int.h"
  37 #include "libman.h"
  38
  39 #define CHKARGS   struct roff_man *man, struct roff_node *n
  40
  41 typedef void    (*v_check)(CHKARGS);
  42
  43 static  void      check_par(CHKARGS);
  44 static  void      check_part(CHKARGS);
  45 static  void      check_root(CHKARGS);
  46 static  void      check_text(CHKARGS);
  47
  48 static  void      post_AT(CHKARGS);
  49 static  void      post_IP(CHKARGS);
  50 static  void      post_vs(CHKARGS);
  51 static  void      post_fi(CHKARGS);
  52 static  void      post_ft(CHKARGS);
  53 static  void      post_nf(CHKARGS);
  54 static  void      post_OP(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         NULL,       /* PD */
  89         post_AT,    /* AT */
  90         NULL,       /* in */
  91         post_ft,    /* ft */
  92         post_OP,    /* 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 roff_man *man)
 103 {
 104         struct roff_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 ROFFT_TEXT:
 114                 check_text(man, n);
 115                 break;
 116         case ROFFT_ROOT:
 117                 check_root(man, n);
 118                 break;
 119         case ROFFT_EQN:
 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         case '2':
 210         case '3':
 211         case '4':
 212         case 'I':
 213         case 'P':
 214         case 'R':
 215                 if ('\0' == cp[1])
 216                         ok = 1;
 217                 break;
 218         case 'B':
 219                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 220                         ok = 1;
 221                 break;
 222         case 'C':
 223                 if ('W' == cp[1] && '\0' == cp[2])
 224                         ok = 1;
 225                 break;
 226         default:
 227                 break;
 228         }
 229
 230         if (0 == ok) {
 231                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 232                     n->line, n->pos, "ft %s", cp);
 233                 *cp = '\0';
 234         }
 235 }
 236
 237 static void
 238 check_part(CHKARGS)
 239 {
 240
 241         if (n->type == ROFFT_BODY && n->child == NULL)
 242                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 243                     n->line, n->pos, man_macronames[n->tok]);
 244 }
 245
 246 static void
 247 check_par(CHKARGS)
 248 {
 249
 250         switch (n->type) {
 251         case ROFFT_BLOCK:
 252                 if (0 == n->body->nchild)
 253                         roff_node_delete(man, n);
 254                 break;
 255         case ROFFT_BODY:
 256                 if (0 == n->nchild)
 257                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 258                             man->parse, n->line, n->pos,
 259                             "%s empty", man_macronames[n->tok]);
 260                 break;
 261         case ROFFT_HEAD:
 262                 if (n->nchild)
 263                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 264                             man->parse, n->line, n->pos,
 265                             "%s %s%s", man_macronames[n->tok],
 266                             n->child->string,
 267                             n->nchild > 1 ? " ..." : "");
 268                 break;
 269         default:
 270                 break;
 271         }
 272 }
 273
 274 static void
 275 post_IP(CHKARGS)
 276 {
 277
 278         switch (n->type) {
 279         case ROFFT_BLOCK:
 280                 if (0 == n->head->nchild && 0 == n->body->nchild)
 281                         roff_node_delete(man, n);
 282                 break;
 283         case ROFFT_BODY:
 284                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 285                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 286                             man->parse, n->line, n->pos,
 287                             "%s empty", man_macronames[n->tok]);
 288                 break;
 289         default:
 290                 break;
 291         }
 292 }
 293
 294 static void
 295 post_TH(CHKARGS)
 296 {
 297         struct roff_node *nb;
 298         const char      *p;
 299
 300         free(man->meta.title);
 301         free(man->meta.vol);
 302         free(man->meta.os);
 303         free(man->meta.msec);
 304         free(man->meta.date);
 305
 306         man->meta.title = man->meta.vol = man->meta.date =
 307             man->meta.msec = man->meta.os = NULL;
 308
 309         nb = n;
 310
 311         /* ->TITLE<- MSEC DATE OS VOL */
 312
 313         n = n->child;
 314         if (n && n->string) {
 315                 for (p = n->string; '\0' != *p; p++) {
 316                         /* Only warn about this once... */
 317                         if (isalpha((unsigned char)*p) &&
 318                             ! isupper((unsigned char)*p)) {
 319                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 320                                     man->parse, n->line,
 321                                     n->pos + (p - n->string),
 322                                     "TH %s", n->string);
 323                                 break;
 324                         }
 325                 }
 326                 man->meta.title = mandoc_strdup(n->string);
 327         } else {
 328                 man->meta.title = mandoc_strdup("");
 329                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 330                     nb->line, nb->pos, "TH");
 331         }
 332
 333         /* TITLE ->MSEC<- DATE OS VOL */
 334
 335         if (n)
 336                 n = n->next;
 337         if (n && n->string)
 338                 man->meta.msec = mandoc_strdup(n->string);
 339         else {
 340                 man->meta.msec = mandoc_strdup("");
 341                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 342                     nb->line, nb->pos, "TH %s", man->meta.title);
 343         }
 344
 345         /* TITLE MSEC ->DATE<- OS VOL */
 346
 347         if (n)
 348                 n = n->next;
 349         if (n && n->string && '\0' != n->string[0]) {
 350                 man->meta.date = man->quick ?
 351                     mandoc_strdup(n->string) :
 352                     mandoc_normdate(man->parse, n->string,
 353                         n->line, n->pos);
 354         } else {
 355                 man->meta.date = mandoc_strdup("");
 356                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 357                     n ? n->line : nb->line,
 358                     n ? n->pos : nb->pos, "TH");
 359         }
 360
 361         /* TITLE MSEC DATE ->OS<- VOL */
 362
 363         if (n && (n = n->next))
 364                 man->meta.os = mandoc_strdup(n->string);
 365         else if (man->defos != NULL)
 366                 man->meta.os = mandoc_strdup(man->defos);
 367
 368         /* TITLE MSEC DATE OS ->VOL<- */
 369         /* If missing, use the default VOL name for MSEC. */
 370
 371         if (n && (n = n->next))
 372                 man->meta.vol = mandoc_strdup(n->string);
 373         else if ('\0' != man->meta.msec[0] &&
 374             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 375                 man->meta.vol = mandoc_strdup(p);
 376
 377         if (n != NULL && (n = n->next) != NULL)
 378                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 379                     n->line, n->pos, "TH ... %s", n->string);
 380
 381         /*
 382          * Remove the `TH' node after we've processed it for our
 383          * meta-data.
 384          */
 385         roff_node_delete(man, man->last);
 386 }
 387
 388 static void
 389 post_nf(CHKARGS)
 390 {
 391
 392         if (man->flags & MAN_LITERAL)
 393                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 394                     n->line, n->pos, "nf");
 395
 396         man->flags |= MAN_LITERAL;
 397 }
 398
 399 static void
 400 post_fi(CHKARGS)
 401 {
 402
 403         if ( ! (MAN_LITERAL & man->flags))
 404                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 405                     n->line, n->pos, "fi");
 406
 407         man->flags &= ~MAN_LITERAL;
 408 }
 409
 410 static void
 411 post_UC(CHKARGS)
 412 {
 413         static const char * const bsd_versions[] = {
 414             "3rd Berkeley Distribution",
 415             "4th Berkeley Distribution",
 416             "4.2 Berkeley Distribution",
 417             "4.3 Berkeley Distribution",
 418             "4.4 Berkeley Distribution",
 419         };
 420
 421         const char      *p, *s;
 422
 423         n = n->child;
 424
 425         if (n == NULL || n->type != ROFFT_TEXT)
 426                 p = bsd_versions[0];
 427         else {
 428                 s = n->string;
 429                 if (0 == strcmp(s, "3"))
 430                         p = bsd_versions[0];
 431                 else if (0 == strcmp(s, "4"))
 432                         p = bsd_versions[1];
 433                 else if (0 == strcmp(s, "5"))
 434                         p = bsd_versions[2];
 435                 else if (0 == strcmp(s, "6"))
 436                         p = bsd_versions[3];
 437                 else if (0 == strcmp(s, "7"))
 438                         p = bsd_versions[4];
 439                 else
 440                         p = bsd_versions[0];
 441         }
 442
 443         free(man->meta.os);
 444         man->meta.os = mandoc_strdup(p);
 445 }
 446
 447 static void
 448 post_AT(CHKARGS)
 449 {
 450         static const char * const unix_versions[] = {
 451             "7th Edition",
 452             "System III",
 453             "System V",
 454             "System V Release 2",
 455         };
 456
 457         struct roff_node *nn;
 458         const char      *p, *s;
 459
 460         n = n->child;
 461
 462         if (n == NULL || n->type != ROFFT_TEXT)
 463                 p = unix_versions[0];
 464         else {
 465                 s = n->string;
 466                 if (0 == strcmp(s, "3"))
 467                         p = unix_versions[0];
 468                 else if (0 == strcmp(s, "4"))
 469                         p = unix_versions[1];
 470                 else if (0 == strcmp(s, "5")) {
 471                         nn = n->next;
 472                         if (nn != NULL &&
 473                             nn->type == ROFFT_TEXT &&
 474                             nn->string[0] != '\0')
 475                                 p = unix_versions[3];
 476                         else
 477                                 p = unix_versions[2];
 478                 } else
 479                         p = unix_versions[0];
 480         }
 481
 482         free(man->meta.os);
 483         man->meta.os = mandoc_strdup(p);
 484 }
 485
 486 static void
 487 post_vs(CHKARGS)
 488 {
 489
 490         if (NULL != n->prev)
 491                 return;
 492
 493         switch (n->parent->tok) {
 494         case MAN_SH:
 495         case MAN_SS:
 496                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 497                     "%s after %s", man_macronames[n->tok],
 498                     man_macronames[n->parent->tok]);
 499                 /* FALLTHROUGH */
 500         case TOKEN_NONE:
 501                 /*
 502                  * Don't warn about this because it occurs in pod2man
 503                  * and would cause considerable (unfixable) warnage.
 504                  */
 505                 roff_node_delete(man, n);
 506                 break;
 507         default:
 508                 break;
 509         }
 510 }