man_validate.c

   1 /*      $Id: man_validate.c,v 1.67 2011/03/22 15:30:30 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010 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 "libman.h"
  36 #include "libmandoc.h"
  37
  38 #define CHKARGS   struct man *m, struct man_node *n
  39
  40 typedef int     (*v_check)(CHKARGS);
  41
  42 struct  man_valid {
  43         v_check  *pres;
  44         v_check  *posts;
  45 };
  46
  47 static  int       check_bline(CHKARGS);
  48 static  int       check_eq0(CHKARGS);
  49 static  int       check_ft(CHKARGS);
  50 static  int       check_le1(CHKARGS);
  51 static  int       check_ge2(CHKARGS);
  52 static  int       check_le5(CHKARGS);
  53 static  int       check_par(CHKARGS);
  54 static  int       check_part(CHKARGS);
  55 static  int       check_root(CHKARGS);
  56 static  int       check_sec(CHKARGS);
  57 static  int       check_text(CHKARGS);
  58
  59 static  int       post_AT(CHKARGS);
  60 static  int       post_fi(CHKARGS);
  61 static  int       post_nf(CHKARGS);
  62 static  int       post_TH(CHKARGS);
  63 static  int       post_UC(CHKARGS);
  64
  65 static  v_check   posts_at[] = { post_AT, NULL };
  66 static  v_check   posts_eq0[] = { check_eq0, NULL };
  67 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  68 static  v_check   posts_le1[] = { check_le1, NULL };
  69 static  v_check   posts_ft[] = { check_ft, NULL };
  70 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  71 static  v_check   posts_par[] = { check_par, NULL };
  72 static  v_check   posts_part[] = { check_part, NULL };
  73 static  v_check   posts_sec[] = { check_sec, NULL };
  74 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  75 static  v_check   posts_uc[] = { post_UC, NULL };
  76 static  v_check   pres_bline[] = { check_bline, NULL };
  77
  78
  79 static  const struct man_valid man_valids[MAN_MAX] = {
  80         { NULL, posts_eq0 }, /* br */
  81         { pres_bline, posts_th }, /* TH */
  82         { pres_bline, posts_sec }, /* SH */
  83         { pres_bline, posts_sec }, /* SS */
  84         { pres_bline, NULL }, /* TP */
  85         { pres_bline, posts_par }, /* LP */
  86         { pres_bline, posts_par }, /* PP */
  87         { pres_bline, posts_par }, /* P */
  88         { pres_bline, NULL }, /* IP */
  89         { pres_bline, NULL }, /* HP */
  90         { NULL, NULL }, /* SM */
  91         { NULL, NULL }, /* SB */
  92         { NULL, NULL }, /* BI */
  93         { NULL, NULL }, /* IB */
  94         { NULL, NULL }, /* BR */
  95         { NULL, NULL }, /* RB */
  96         { NULL, NULL }, /* R */
  97         { NULL, NULL }, /* B */
  98         { NULL, NULL }, /* I */
  99         { NULL, NULL }, /* IR */
 100         { NULL, NULL }, /* RI */
 101         { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
 102         { NULL, posts_le1 }, /* sp */ /* FIXME: should warn only. */
 103         { pres_bline, posts_nf }, /* nf */
 104         { pres_bline, posts_fi }, /* fi */
 105         { NULL, NULL }, /* RE */
 106         { NULL, posts_part }, /* RS */
 107         { NULL, NULL }, /* DT */
 108         { NULL, posts_uc }, /* UC */
 109         { NULL, NULL }, /* PD */
 110         { NULL, posts_at }, /* AT */
 111         { NULL, NULL }, /* in */
 112         { NULL, posts_ft }, /* ft */
 113 };
 114
 115
 116 int
 117 man_valid_pre(struct man *m, struct man_node *n)
 118 {
 119         v_check         *cp;
 120
 121         switch (n->type) {
 122         case (MAN_TEXT):
 123                 /* FALLTHROUGH */
 124         case (MAN_ROOT):
 125                 /* FALLTHROUGH */
 126         case (MAN_EQN):
 127                 /* FALLTHROUGH */
 128         case (MAN_TBL):
 129                 return(1);
 130         default:
 131                 break;
 132         }
 133
 134         if (NULL == (cp = man_valids[n->tok].pres))
 135                 return(1);
 136         for ( ; *cp; cp++)
 137                 if ( ! (*cp)(m, n))
 138                         return(0);
 139         return(1);
 140 }
 141
 142
 143 int
 144 man_valid_post(struct man *m)
 145 {
 146         v_check         *cp;
 147
 148         if (MAN_VALID & m->last->flags)
 149                 return(1);
 150         m->last->flags |= MAN_VALID;
 151
 152         switch (m->last->type) {
 153         case (MAN_TEXT):
 154                 return(check_text(m, m->last));
 155         case (MAN_ROOT):
 156                 return(check_root(m, m->last));
 157         case (MAN_EQN):
 158                 /* FALLTHROUGH */
 159         case (MAN_TBL):
 160                 return(1);
 161         default:
 162                 break;
 163         }
 164
 165         if (NULL == (cp = man_valids[m->last->tok].posts))
 166                 return(1);
 167         for ( ; *cp; cp++)
 168                 if ( ! (*cp)(m, m->last))
 169                         return(0);
 170
 171         return(1);
 172 }
 173
 174
 175 static int
 176 check_root(CHKARGS)
 177 {
 178
 179         if (MAN_BLINE & m->flags)
 180                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 181         else if (MAN_ELINE & m->flags)
 182                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 183
 184         m->flags &= ~MAN_BLINE;
 185         m->flags &= ~MAN_ELINE;
 186
 187         if (NULL == m->first->child) {
 188                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 189                 return(0);
 190         } else if (NULL == m->meta.title) {
 191                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 192
 193                 /*
 194                  * If a title hasn't been set, do so now (by
 195                  * implication, date and section also aren't set).
 196                  */
 197
 198                 m->meta.title = mandoc_strdup("unknown");
 199                 m->meta.msec = mandoc_strdup("1");
 200                 m->meta.date = mandoc_normdate
 201                         (m->parse, NULL, n->line, n->pos);
 202         }
 203
 204         return(1);
 205 }
 206
 207
 208 static int
 209 check_text(CHKARGS)
 210 {
 211         char            *p;
 212         int              pos, c;
 213         size_t           sz;
 214
 215         for (p = n->string, pos = n->pos + 1; *p; p++, pos++) {
 216                 sz = strcspn(p, "\t\\");
 217                 p += (int)sz;
 218
 219                 if ('\0' == *p)
 220                         break;
 221
 222                 pos += (int)sz;
 223
 224                 if ('\t' == *p) {
 225                         if (MAN_LITERAL & m->flags)
 226                                 continue;
 227                         man_pmsg(m, n->line, pos, MANDOCERR_BADTAB);
 228                         continue;
 229                 }
 230
 231                 /* Check the special character. */
 232
 233                 c = mandoc_special(p);
 234                 if (c) {
 235                         p += c - 1;
 236                         pos += c - 1;
 237                 } else
 238                         man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
 239         }
 240
 241         return(1);
 242 }
 243
 244
 245 #define INEQ_DEFINE(x, ineq, name) \
 246 static int \
 247 check_##name(CHKARGS) \
 248 { \
 249         if (n->nchild ineq (x)) \
 250                 return(1); \
 251         mandoc_vmsg(MANDOCERR_ARGCOUNT, m->parse, n->line, n->pos, \
 252                         "line arguments %s %d (have %d)", \
 253                         #ineq, (x), n->nchild); \
 254         return(1); \
 255 }
 256
 257 INEQ_DEFINE(0, ==, eq0)
 258 INEQ_DEFINE(1, <=, le1)
 259 INEQ_DEFINE(2, >=, ge2)
 260 INEQ_DEFINE(5, <=, le5)
 261
 262 static int
 263 check_ft(CHKARGS)
 264 {
 265         char    *cp;
 266         int      ok;
 267
 268         if (0 == n->nchild)
 269                 return(1);
 270
 271         ok = 0;
 272         cp = n->child->string;
 273         switch (*cp) {
 274         case ('1'):
 275                 /* FALLTHROUGH */
 276         case ('2'):
 277                 /* FALLTHROUGH */
 278         case ('3'):
 279                 /* FALLTHROUGH */
 280         case ('4'):
 281                 /* FALLTHROUGH */
 282         case ('I'):
 283                 /* FALLTHROUGH */
 284         case ('P'):
 285                 /* FALLTHROUGH */
 286         case ('R'):
 287                 if ('\0' == cp[1])
 288                         ok = 1;
 289                 break;
 290         case ('B'):
 291                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 292                         ok = 1;
 293                 break;
 294         case ('C'):
 295                 if ('W' == cp[1] && '\0' == cp[2])
 296                         ok = 1;
 297                 break;
 298         default:
 299                 break;
 300         }
 301
 302         if (0 == ok) {
 303                 mandoc_vmsg
 304                         (MANDOCERR_BADFONT, m->parse,
 305                          n->line, n->pos, "%s", cp);
 306                 *cp = '\0';
 307         }
 308
 309         if (1 < n->nchild)
 310                 mandoc_vmsg
 311                         (MANDOCERR_ARGCOUNT, m->parse, n->line,
 312                          n->pos, "want one child (have %d)",
 313                          n->nchild);
 314
 315         return(1);
 316 }
 317
 318 static int
 319 check_sec(CHKARGS)
 320 {
 321
 322         if (MAN_HEAD == n->type && 0 == n->nchild) {
 323                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 324                 return(0);
 325         } else if (MAN_BODY == n->type && 0 == n->nchild)
 326                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 327                                 n->pos, "want children (have none)");
 328
 329         return(1);
 330 }
 331
 332
 333 static int
 334 check_part(CHKARGS)
 335 {
 336
 337         if (MAN_BODY == n->type && 0 == n->nchild)
 338                 mandoc_msg(MANDOCERR_ARGCWARN, m->parse, n->line,
 339                                 n->pos, "want children (have none)");
 340
 341         return(1);
 342 }
 343
 344
 345 static int
 346 check_par(CHKARGS)
 347 {
 348
 349         switch (n->type) {
 350         case (MAN_BLOCK):
 351                 if (0 == n->body->nchild)
 352                         man_node_delete(m, n);
 353                 break;
 354         case (MAN_BODY):
 355                 if (0 == n->nchild)
 356                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 357                 break;
 358         case (MAN_HEAD):
 359                 if (n->nchild)
 360                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 361                 break;
 362         default:
 363                 break;
 364         }
 365
 366         return(1);
 367 }
 368
 369
 370 static int
 371 check_bline(CHKARGS)
 372 {
 373
 374         assert( ! (MAN_ELINE & m->flags));
 375         if (MAN_BLINE & m->flags) {
 376                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 377                 return(0);
 378         }
 379
 380         return(1);
 381 }
 382
 383 static int
 384 post_TH(CHKARGS)
 385 {
 386         const char      *p;
 387         int              line, pos;
 388
 389         if (m->meta.title)
 390                 free(m->meta.title);
 391         if (m->meta.vol)
 392                 free(m->meta.vol);
 393         if (m->meta.source)
 394                 free(m->meta.source);
 395         if (m->meta.msec)
 396                 free(m->meta.msec);
 397         if (m->meta.date)
 398                 free(m->meta.date);
 399
 400         line = n->line;
 401         pos = n->pos;
 402         m->meta.title = m->meta.vol = m->meta.date =
 403                 m->meta.msec = m->meta.source = NULL;
 404
 405         /* ->TITLE<- MSEC DATE SOURCE VOL */
 406
 407         n = n->child;
 408         if (n && n->string) {
 409                 for (p = n->string; '\0' != *p; p++) {
 410                         /* Only warn about this once... */
 411                         if (isalpha((u_char)*p) && ! isupper((u_char)*p)) {
 412                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 413                                 break;
 414                         }
 415                 }
 416                 m->meta.title = mandoc_strdup(n->string);
 417         } else
 418                 m->meta.title = mandoc_strdup("");
 419
 420         /* TITLE ->MSEC<- DATE SOURCE VOL */
 421
 422         if (n)
 423                 n = n->next;
 424         if (n && n->string)
 425                 m->meta.msec = mandoc_strdup(n->string);
 426         else
 427                 m->meta.msec = mandoc_strdup("");
 428
 429         /* TITLE MSEC ->DATE<- SOURCE VOL */
 430
 431         if (n)
 432                 n = n->next;
 433         if (n)
 434                 pos = n->pos;
 435         m->meta.date = mandoc_normdate
 436                 (m->parse, n ? n->string : NULL, line, pos);
 437
 438         /* TITLE MSEC DATE ->SOURCE<- VOL */
 439
 440         if (n && (n = n->next))
 441                 m->meta.source = mandoc_strdup(n->string);
 442
 443         /* TITLE MSEC DATE SOURCE ->VOL<- */
 444
 445         if (n && (n = n->next))
 446                 m->meta.vol = mandoc_strdup(n->string);
 447
 448         /*
 449          * Remove the `TH' node after we've processed it for our
 450          * meta-data.
 451          */
 452         man_node_delete(m, m->last);
 453         return(1);
 454 }
 455
 456 static int
 457 post_nf(CHKARGS)
 458 {
 459
 460         if (MAN_LITERAL & m->flags)
 461                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 462
 463         m->flags |= MAN_LITERAL;
 464         return(1);
 465 }
 466
 467 static int
 468 post_fi(CHKARGS)
 469 {
 470
 471         if ( ! (MAN_LITERAL & m->flags))
 472                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 473
 474         m->flags &= ~MAN_LITERAL;
 475         return(1);
 476 }
 477
 478 static int
 479 post_UC(CHKARGS)
 480 {
 481         static const char * const bsd_versions[] = {
 482             "3rd Berkeley Distribution",
 483             "4th Berkeley Distribution",
 484             "4.2 Berkeley Distribution",
 485             "4.3 Berkeley Distribution",
 486             "4.4 Berkeley Distribution",
 487         };
 488
 489         const char      *p, *s;
 490
 491         n = n->child;
 492         n = m->last->child;
 493
 494         if (NULL == n || MAN_TEXT != n->type)
 495                 p = bsd_versions[0];
 496         else {
 497                 s = n->string;
 498                 if (0 == strcmp(s, "3"))
 499                         p = bsd_versions[0];
 500                 else if (0 == strcmp(s, "4"))
 501                         p = bsd_versions[1];
 502                 else if (0 == strcmp(s, "5"))
 503                         p = bsd_versions[2];
 504                 else if (0 == strcmp(s, "6"))
 505                         p = bsd_versions[3];
 506                 else if (0 == strcmp(s, "7"))
 507                         p = bsd_versions[4];
 508                 else
 509                         p = bsd_versions[0];
 510         }
 511
 512         if (m->meta.source)
 513                 free(m->meta.source);
 514
 515         m->meta.source = mandoc_strdup(p);
 516         return(1);
 517 }
 518
 519 static int
 520 post_AT(CHKARGS)
 521 {
 522         static const char * const unix_versions[] = {
 523             "7th Edition",
 524             "System III",
 525             "System V",
 526             "System V Release 2",
 527         };
 528
 529         const char      *p, *s;
 530         struct man_node *nn;
 531
 532         n = n->child;
 533
 534         if (NULL == n || MAN_TEXT != n->type)
 535                 p = unix_versions[0];
 536         else {
 537                 s = n->string;
 538                 if (0 == strcmp(s, "3"))
 539                         p = unix_versions[0];
 540                 else if (0 == strcmp(s, "4"))
 541                         p = unix_versions[1];
 542                 else if (0 == strcmp(s, "5")) {
 543                         nn = n->next;
 544                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 545                                 p = unix_versions[3];
 546                         else
 547                                 p = unix_versions[2];
 548                 } else
 549                         p = unix_versions[0];
 550         }
 551
 552         if (m->meta.source)
 553                 free(m->meta.source);
 554
 555         m->meta.source = mandoc_strdup(p);
 556         return(1);
 557 }