man_validate.c

   1 /*      $Id: man_validate.c,v 1.52 2010/12/05 16:14:16 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <errno.h>
  26 #include <limits.h>
  27 #include <stdarg.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <time.h>
  31
  32 #include "mandoc.h"
  33 #include "libman.h"
  34 #include "libmandoc.h"
  35
  36 #define CHKARGS   struct man *m, struct man_node *n
  37
  38 typedef int     (*v_check)(CHKARGS);
  39
  40 struct  man_valid {
  41         v_check  *pres;
  42         v_check  *posts;
  43 };
  44
  45 static  int       check_bline(CHKARGS);
  46 static  int       check_eq0(CHKARGS);
  47 static  int       check_le1(CHKARGS);
  48 static  int       check_ge2(CHKARGS);
  49 static  int       check_le5(CHKARGS);
  50 static  int       check_par(CHKARGS);
  51 static  int       check_part(CHKARGS);
  52 static  int       check_root(CHKARGS);
  53 static  int       check_sec(CHKARGS);
  54 static  int       check_text(CHKARGS);
  55 static  int       check_title(CHKARGS);
  56
  57 static  int       post_AT(CHKARGS);
  58 static  int       post_fi(CHKARGS);
  59 static  int       post_nf(CHKARGS);
  60 static  int       post_TH(CHKARGS);
  61 static  int       post_UC(CHKARGS);
  62
  63 static  v_check   posts_at[] = { post_AT, NULL };
  64 static  v_check   posts_eq0[] = { check_eq0, NULL };
  65 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  66 static  v_check   posts_le1[] = { check_le1, NULL };
  67 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  68 static  v_check   posts_par[] = { check_par, NULL };
  69 static  v_check   posts_part[] = { check_part, NULL };
  70 static  v_check   posts_sec[] = { check_sec, NULL };
  71 static  v_check   posts_th[] = { check_ge2, check_le5, check_title, post_TH, NULL };
  72 static  v_check   posts_uc[] = { post_UC, NULL };
  73 static  v_check   pres_bline[] = { check_bline, NULL };
  74
  75
  76 static  const struct man_valid man_valids[MAN_MAX] = {
  77         { NULL, posts_eq0 }, /* br */
  78         { pres_bline, posts_th }, /* TH */
  79         { pres_bline, posts_sec }, /* SH */
  80         { pres_bline, posts_sec }, /* SS */
  81         { pres_bline, posts_par }, /* TP */
  82         { pres_bline, posts_par }, /* LP */
  83         { pres_bline, posts_par }, /* PP */
  84         { pres_bline, posts_par }, /* P */
  85         { pres_bline, posts_par }, /* IP */
  86         { pres_bline, posts_par }, /* HP */
  87         { NULL, NULL }, /* SM */
  88         { NULL, NULL }, /* SB */
  89         { NULL, NULL }, /* BI */
  90         { NULL, NULL }, /* IB */
  91         { NULL, NULL }, /* BR */
  92         { NULL, NULL }, /* RB */
  93         { NULL, NULL }, /* R */
  94         { NULL, NULL }, /* B */
  95         { NULL, NULL }, /* I */
  96         { NULL, NULL }, /* IR */
  97         { NULL, NULL }, /* RI */
  98         { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
  99         { NULL, NULL }, /* i */
 100         { NULL, posts_le1 }, /* sp */ /* FIXME: should warn only. */
 101         { pres_bline, posts_nf }, /* nf */
 102         { pres_bline, posts_fi }, /* fi */
 103         { NULL, NULL }, /* r */
 104         { NULL, NULL }, /* RE */
 105         { NULL, posts_part }, /* RS */
 106         { NULL, NULL }, /* DT */
 107         { NULL, posts_uc }, /* UC */
 108         { NULL, NULL }, /* PD */
 109         { NULL, posts_at }, /* AT */
 110         { NULL, NULL }, /* in */
 111 };
 112
 113
 114 int
 115 man_valid_pre(struct man *m, struct man_node *n)
 116 {
 117         v_check         *cp;
 118
 119         if (MAN_TEXT == n->type)
 120                 return(1);
 121         if (MAN_ROOT == n->type)
 122                 return(1);
 123
 124         if (NULL == (cp = man_valids[n->tok].pres))
 125                 return(1);
 126         for ( ; *cp; cp++)
 127                 if ( ! (*cp)(m, n))
 128                         return(0);
 129         return(1);
 130 }
 131
 132
 133 int
 134 man_valid_post(struct man *m)
 135 {
 136         v_check         *cp;
 137
 138         if (MAN_VALID & m->last->flags)
 139                 return(1);
 140         m->last->flags |= MAN_VALID;
 141
 142         switch (m->last->type) {
 143         case (MAN_TEXT):
 144                 return(check_text(m, m->last));
 145         case (MAN_ROOT):
 146                 return(check_root(m, m->last));
 147         default:
 148                 break;
 149         }
 150
 151         if (NULL == (cp = man_valids[m->last->tok].posts))
 152                 return(1);
 153         for ( ; *cp; cp++)
 154                 if ( ! (*cp)(m, m->last))
 155                         return(0);
 156
 157         return(1);
 158 }
 159
 160
 161 static int
 162 check_root(CHKARGS)
 163 {
 164
 165         if (MAN_BLINE & m->flags)
 166                 return(man_nmsg(m, n, MANDOCERR_SCOPEEXIT));
 167         if (MAN_ELINE & m->flags)
 168                 return(man_nmsg(m, n, MANDOCERR_SCOPEEXIT));
 169
 170         m->flags &= ~MAN_BLINE;
 171         m->flags &= ~MAN_ELINE;
 172
 173         if (NULL == m->first->child) {
 174                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 175                 return(0);
 176         } else if (NULL == m->meta.title) {
 177                 if ( ! man_nmsg(m, n, MANDOCERR_NOTITLE))
 178                         return(0);
 179                 /*
 180                  * If a title hasn't been set, do so now (by
 181                  * implication, date and section also aren't set).
 182                  *
 183                  * FIXME: this should be in man_action.c.
 184                  */
 185                 m->meta.title = mandoc_strdup("unknown");
 186                 m->meta.date = time(NULL);
 187                 m->meta.msec = mandoc_strdup("1");
 188         }
 189
 190         return(1);
 191 }
 192
 193
 194 static int
 195 check_title(CHKARGS)
 196 {
 197         const char      *p;
 198
 199         assert(n->child);
 200         /* FIXME: is this sufficient? */
 201         if ('\0' == *n->child->string) {
 202                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 203                 return(0);
 204         }
 205
 206         for (p = n->child->string; '\0' != *p; p++)
 207                 if (isalpha((u_char)*p) && ! isupper((u_char)*p))
 208                         if ( ! man_nmsg(m, n, MANDOCERR_UPPERCASE))
 209                                 return(0);
 210
 211         return(1);
 212 }
 213
 214
 215 static int
 216 check_text(CHKARGS)
 217 {
 218         char            *p;
 219         int              pos, c;
 220         size_t           sz;
 221
 222         for (p = n->string, pos = n->pos + 1; *p; p++, pos++) {
 223                 sz = strcspn(p, "\t\\");
 224                 p += (int)sz;
 225
 226                 if ('\0' == *p)
 227                         break;
 228
 229                 pos += (int)sz;
 230
 231                 if ('\t' == *p) {
 232                         if (MAN_LITERAL & m->flags)
 233                                 continue;
 234                         if (man_pmsg(m, n->line, pos, MANDOCERR_BADTAB))
 235                                 continue;
 236                         return(0);
 237                 }
 238
 239                 /* Check the special character. */
 240
 241                 c = mandoc_special(p);
 242                 if (c) {
 243                         p += c - 1;
 244                         pos += c - 1;
 245                 } else
 246                         man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
 247         }
 248
 249         return(1);
 250 }
 251
 252
 253 #define INEQ_DEFINE(x, ineq, name) \
 254 static int \
 255 check_##name(CHKARGS) \
 256 { \
 257         if (n->nchild ineq (x)) \
 258                 return(1); \
 259         man_vmsg(m, MANDOCERR_SYNTARGCOUNT, n->line, n->pos, \
 260                         "line arguments %s %d (have %d)", \
 261                         #ineq, (x), n->nchild); \
 262         return(0); \
 263 }
 264
 265 INEQ_DEFINE(0, ==, eq0)
 266 INEQ_DEFINE(1, <=, le1)
 267 INEQ_DEFINE(2, >=, ge2)
 268 INEQ_DEFINE(5, <=, le5)
 269
 270
 271 static int
 272 check_sec(CHKARGS)
 273 {
 274
 275         if (MAN_HEAD == n->type && 0 == n->nchild) {
 276                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 277                 return(0);
 278         } else if (MAN_BODY == n->type && 0 == n->nchild)
 279                 return(man_nmsg(m, n, MANDOCERR_NOBODY));
 280
 281         return(1);
 282 }
 283
 284
 285 static int
 286 check_part(CHKARGS)
 287 {
 288
 289         if (MAN_BODY == n->type && 0 == n->nchild)
 290                 return(man_nmsg(m, n, MANDOCERR_NOBODY));
 291         return(1);
 292 }
 293
 294
 295 static int
 296 check_par(CHKARGS)
 297 {
 298
 299         if (MAN_BODY == n->type)
 300                 switch (n->tok) {
 301                 case (MAN_IP):
 302                         /* FALLTHROUGH */
 303                 case (MAN_HP):
 304                         /* FALLTHROUGH */
 305                 case (MAN_TP):
 306                         /* Body-less lists are ok. */
 307                         break;
 308                 default:
 309                         if (n->nchild)
 310                                 break;
 311                         return(man_nmsg(m, n, MANDOCERR_NOBODY));
 312                 }
 313         if (MAN_HEAD == n->type)
 314                 switch (n->tok) {
 315                 case (MAN_PP):
 316                         /* FALLTHROUGH */
 317                 case (MAN_P):
 318                         /* FALLTHROUGH */
 319                 case (MAN_LP):
 320                         if (0 == n->nchild)
 321                                 break;
 322                         return(man_nmsg(m, n, MANDOCERR_ARGSLOST));
 323                 default:
 324                         if (n->nchild)
 325                                 break;
 326                         return(man_nmsg(m, n, MANDOCERR_NOARGS));
 327                 }
 328
 329         return(1);
 330 }
 331
 332
 333 static int
 334 check_bline(CHKARGS)
 335 {
 336
 337         assert( ! (MAN_ELINE & m->flags));
 338         if (MAN_BLINE & m->flags) {
 339                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 340                 return(0);
 341         }
 342
 343         return(1);
 344 }
 345
 346 static int
 347 post_TH(CHKARGS)
 348 {
 349
 350         if (m->meta.title)
 351                 free(m->meta.title);
 352         if (m->meta.vol)
 353                 free(m->meta.vol);
 354         if (m->meta.source)
 355                 free(m->meta.source);
 356         if (m->meta.msec)
 357                 free(m->meta.msec);
 358         if (m->meta.rawdate)
 359                 free(m->meta.rawdate);
 360
 361         m->meta.title = m->meta.vol = m->meta.rawdate =
 362                 m->meta.msec = m->meta.source = NULL;
 363         m->meta.date = 0;
 364
 365         /* ->TITLE<- MSEC DATE SOURCE VOL */
 366
 367         n = n->child;
 368         assert(n);
 369         m->meta.title = mandoc_strdup(n->string);
 370
 371         /* TITLE ->MSEC<- DATE SOURCE VOL */
 372
 373         n = n->next;
 374         assert(n);
 375         m->meta.msec = mandoc_strdup(n->string);
 376
 377         /* TITLE MSEC ->DATE<- SOURCE VOL */
 378
 379         /*
 380          * Try to parse the date.  If this works, stash the epoch (this
 381          * is optimal because we can reformat it in the canonical form).
 382          * If it doesn't parse, isn't specified at all, or is an empty
 383          * string, then use the current date.
 384          */
 385
 386         n = n->next;
 387         if (n && n->string && *n->string) {
 388                 m->meta.date = mandoc_a2time
 389                         (MTIME_ISO_8601, n->string);
 390                 if (0 == m->meta.date) {
 391                         man_nmsg(m, n, MANDOCERR_BADDATE);
 392                         m->meta.rawdate = mandoc_strdup(n->string);
 393                 }
 394         } else
 395                 m->meta.date = time(NULL);
 396
 397         /* TITLE MSEC DATE ->SOURCE<- VOL */
 398
 399         if (n && (n = n->next))
 400                 m->meta.source = mandoc_strdup(n->string);
 401
 402         /* TITLE MSEC DATE SOURCE ->VOL<- */
 403
 404         if (n && (n = n->next))
 405                 m->meta.vol = mandoc_strdup(n->string);
 406
 407         /*
 408          * Remove the `TH' node after we've processed it for our
 409          * meta-data.
 410          */
 411         man_node_delete(m, m->last);
 412         return(1);
 413 }
 414
 415 static int
 416 post_nf(CHKARGS)
 417 {
 418
 419         if (MAN_LITERAL & m->flags)
 420                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 421
 422         m->flags |= MAN_LITERAL;
 423         return(1);
 424 }
 425
 426 static int
 427 post_fi(CHKARGS)
 428 {
 429
 430         if ( ! (MAN_LITERAL & m->flags))
 431                 man_nmsg(m, n, MANDOCERR_NOSCOPE);
 432
 433         m->flags &= ~MAN_LITERAL;
 434         return(1);
 435 }
 436
 437 static int
 438 post_UC(CHKARGS)
 439 {
 440         static const char * const bsd_versions[] = {
 441             "3rd Berkeley Distribution",
 442             "4th Berkeley Distribution",
 443             "4.2 Berkeley Distribution",
 444             "4.3 Berkeley Distribution",
 445             "4.4 Berkeley Distribution",
 446         };
 447
 448         const char      *p, *s;
 449
 450         n = n->child;
 451         n = m->last->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         if (m->meta.source)
 472                 free(m->meta.source);
 473
 474         m->meta.source = mandoc_strdup(p);
 475         return(1);
 476 }
 477
 478 static int
 479 post_AT(CHKARGS)
 480 {
 481         static const char * const unix_versions[] = {
 482             "7th Edition",
 483             "System III",
 484             "System V",
 485             "System V Release 2",
 486         };
 487
 488         const char      *p, *s;
 489         struct man_node *nn;
 490
 491         n = n->child;
 492
 493         if (NULL == n || MAN_TEXT != n->type)
 494                 p = unix_versions[0];
 495         else {
 496                 s = n->string;
 497                 if (0 == strcmp(s, "3"))
 498                         p = unix_versions[0];
 499                 else if (0 == strcmp(s, "4"))
 500                         p = unix_versions[1];
 501                 else if (0 == strcmp(s, "5")) {
 502                         nn = n->next;
 503                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 504                                 p = unix_versions[3];
 505                         else
 506                                 p = unix_versions[2];
 507                 } else
 508                         p = unix_versions[0];
 509         }
 510
 511         if (m->meta.source)
 512                 free(m->meta.source);
 513
 514         m->meta.source = mandoc_strdup(p);
 515         return(1);
 516 }