man_validate.c

   1 /*      $Id: man_validate.c,v 1.64 2011/03/17 11:56:17 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 "mandoc.h"
  34 #include "libman.h"
  35 #include "libmandoc.h"
  36
  37 #define CHKARGS   struct man *m, struct man_node *n
  38
  39 typedef int     (*v_check)(CHKARGS);
  40
  41 struct  man_valid {
  42         v_check  *pres;
  43         v_check  *posts;
  44 };
  45
  46 static  int       check_bline(CHKARGS);
  47 static  int       check_eq0(CHKARGS);
  48 static  int       check_ft(CHKARGS);
  49 static  int       check_le1(CHKARGS);
  50 static  int       check_ge2(CHKARGS);
  51 static  int       check_le5(CHKARGS);
  52 static  int       check_par(CHKARGS);
  53 static  int       check_part(CHKARGS);
  54 static  int       check_root(CHKARGS);
  55 static  int       check_sec(CHKARGS);
  56 static  int       check_text(CHKARGS);
  57
  58 static  int       post_AT(CHKARGS);
  59 static  int       post_fi(CHKARGS);
  60 static  int       post_nf(CHKARGS);
  61 static  int       post_TH(CHKARGS);
  62 static  int       post_UC(CHKARGS);
  63
  64 static  v_check   posts_at[] = { post_AT, NULL };
  65 static  v_check   posts_eq0[] = { check_eq0, NULL };
  66 static  v_check   posts_fi[] = { check_eq0, post_fi, NULL };
  67 static  v_check   posts_le1[] = { check_le1, NULL };
  68 static  v_check   posts_ft[] = { check_ft, NULL };
  69 static  v_check   posts_nf[] = { check_eq0, post_nf, NULL };
  70 static  v_check   posts_par[] = { check_par, NULL };
  71 static  v_check   posts_part[] = { check_part, NULL };
  72 static  v_check   posts_sec[] = { check_sec, NULL };
  73 static  v_check   posts_th[] = { check_ge2, check_le5, post_TH, NULL };
  74 static  v_check   posts_uc[] = { post_UC, NULL };
  75 static  v_check   pres_bline[] = { check_bline, NULL };
  76
  77
  78 static  const struct man_valid man_valids[MAN_MAX] = {
  79         { NULL, posts_eq0 }, /* br */
  80         { pres_bline, posts_th }, /* TH */
  81         { pres_bline, posts_sec }, /* SH */
  82         { pres_bline, posts_sec }, /* SS */
  83         { pres_bline, NULL }, /* TP */
  84         { pres_bline, posts_par }, /* LP */
  85         { pres_bline, posts_par }, /* PP */
  86         { pres_bline, posts_par }, /* P */
  87         { pres_bline, NULL }, /* IP */
  88         { pres_bline, NULL }, /* HP */
  89         { NULL, NULL }, /* SM */
  90         { NULL, NULL }, /* SB */
  91         { NULL, NULL }, /* BI */
  92         { NULL, NULL }, /* IB */
  93         { NULL, NULL }, /* BR */
  94         { NULL, NULL }, /* RB */
  95         { NULL, NULL }, /* R */
  96         { NULL, NULL }, /* B */
  97         { NULL, NULL }, /* I */
  98         { NULL, NULL }, /* IR */
  99         { NULL, NULL }, /* RI */
 100         { NULL, posts_eq0 }, /* na */ /* FIXME: should warn only. */
 101         { NULL, posts_le1 }, /* sp */ /* FIXME: should warn only. */
 102         { pres_bline, posts_nf }, /* nf */
 103         { pres_bline, posts_fi }, /* fi */
 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         { NULL, posts_ft }, /* ft */
 112 };
 113
 114
 115 int
 116 man_valid_pre(struct man *m, struct man_node *n)
 117 {
 118         v_check         *cp;
 119
 120         switch (n->type) {
 121         case (MAN_TEXT):
 122                 /* FALLTHROUGH */
 123         case (MAN_ROOT):
 124                 /* FALLTHROUGH */
 125         case (MAN_EQN):
 126                 /* FALLTHROUGH */
 127         case (MAN_TBL):
 128                 return(1);
 129         default:
 130                 break;
 131         }
 132
 133         if (NULL == (cp = man_valids[n->tok].pres))
 134                 return(1);
 135         for ( ; *cp; cp++)
 136                 if ( ! (*cp)(m, n))
 137                         return(0);
 138         return(1);
 139 }
 140
 141
 142 int
 143 man_valid_post(struct man *m)
 144 {
 145         v_check         *cp;
 146
 147         if (MAN_VALID & m->last->flags)
 148                 return(1);
 149         m->last->flags |= MAN_VALID;
 150
 151         switch (m->last->type) {
 152         case (MAN_TEXT):
 153                 return(check_text(m, m->last));
 154         case (MAN_ROOT):
 155                 return(check_root(m, m->last));
 156         case (MAN_EQN):
 157                 /* FALLTHROUGH */
 158         case (MAN_TBL):
 159                 return(1);
 160         default:
 161                 break;
 162         }
 163
 164         if (NULL == (cp = man_valids[m->last->tok].posts))
 165                 return(1);
 166         for ( ; *cp; cp++)
 167                 if ( ! (*cp)(m, m->last))
 168                         return(0);
 169
 170         return(1);
 171 }
 172
 173
 174 static int
 175 check_root(CHKARGS)
 176 {
 177
 178         if (MAN_BLINE & m->flags)
 179                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 180         else if (MAN_ELINE & m->flags)
 181                 man_nmsg(m, n, MANDOCERR_SCOPEEXIT);
 182
 183         m->flags &= ~MAN_BLINE;
 184         m->flags &= ~MAN_ELINE;
 185
 186         if (NULL == m->first->child) {
 187                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 188                 return(0);
 189         } else if (NULL == m->meta.title) {
 190                 man_nmsg(m, n, MANDOCERR_NOTITLE);
 191
 192                 /*
 193                  * If a title hasn't been set, do so now (by
 194                  * implication, date and section also aren't set).
 195                  */
 196
 197                 m->meta.title = mandoc_strdup("unknown");
 198                 m->meta.msec = mandoc_strdup("1");
 199                 m->meta.date = mandoc_normdate(NULL,
 200                     m->msg, m->data, n->line, n->pos);
 201         }
 202
 203         return(1);
 204 }
 205
 206
 207 static int
 208 check_text(CHKARGS)
 209 {
 210         char            *p;
 211         int              pos, c;
 212         size_t           sz;
 213
 214         for (p = n->string, pos = n->pos + 1; *p; p++, pos++) {
 215                 sz = strcspn(p, "\t\\");
 216                 p += (int)sz;
 217
 218                 if ('\0' == *p)
 219                         break;
 220
 221                 pos += (int)sz;
 222
 223                 if ('\t' == *p) {
 224                         if (MAN_LITERAL & m->flags)
 225                                 continue;
 226                         man_pmsg(m, n->line, pos, MANDOCERR_BADTAB);
 227                         continue;
 228                 }
 229
 230                 /* Check the special character. */
 231
 232                 c = mandoc_special(p);
 233                 if (c) {
 234                         p += c - 1;
 235                         pos += c - 1;
 236                 } else
 237                         man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
 238         }
 239
 240         return(1);
 241 }
 242
 243
 244 #define INEQ_DEFINE(x, ineq, name) \
 245 static int \
 246 check_##name(CHKARGS) \
 247 { \
 248         if (n->nchild ineq (x)) \
 249                 return(1); \
 250         man_vmsg(m, MANDOCERR_ARGCOUNT, n->line, n->pos, \
 251                         "line arguments %s %d (have %d)", \
 252                         #ineq, (x), n->nchild); \
 253         return(1); \
 254 }
 255
 256 INEQ_DEFINE(0, ==, eq0)
 257 INEQ_DEFINE(1, <=, le1)
 258 INEQ_DEFINE(2, >=, ge2)
 259 INEQ_DEFINE(5, <=, le5)
 260
 261 static int
 262 check_ft(CHKARGS)
 263 {
 264         char    *cp;
 265         int      ok;
 266
 267         if (0 == n->nchild)
 268                 return(1);
 269
 270         ok = 0;
 271         cp = n->child->string;
 272         switch (*cp) {
 273         case ('1'):
 274                 /* FALLTHROUGH */
 275         case ('2'):
 276                 /* FALLTHROUGH */
 277         case ('3'):
 278                 /* FALLTHROUGH */
 279         case ('4'):
 280                 /* FALLTHROUGH */
 281         case ('I'):
 282                 /* FALLTHROUGH */
 283         case ('P'):
 284                 /* FALLTHROUGH */
 285         case ('R'):
 286                 if ('\0' == cp[1])
 287                         ok = 1;
 288                 break;
 289         case ('B'):
 290                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 291                         ok = 1;
 292                 break;
 293         case ('C'):
 294                 if ('W' == cp[1] && '\0' == cp[2])
 295                         ok = 1;
 296                 break;
 297         default:
 298                 break;
 299         }
 300
 301         if (0 == ok) {
 302                 man_vmsg(m, MANDOCERR_BADFONT,
 303                                 n->line, n->pos, "%s", cp);
 304                 *cp = '\0';
 305         }
 306
 307         if (1 < n->nchild)
 308                 man_vmsg(m, MANDOCERR_ARGCOUNT, n->line, n->pos,
 309                                 "want one child (have %d)", n->nchild);
 310
 311         return(1);
 312 }
 313
 314 static int
 315 check_sec(CHKARGS)
 316 {
 317
 318         if (MAN_HEAD == n->type && 0 == n->nchild) {
 319                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 320                 return(0);
 321         } else if (MAN_BODY == n->type && 0 == n->nchild)
 322                 man_nmsg(m, n, MANDOCERR_NOBODY);
 323
 324         return(1);
 325 }
 326
 327
 328 static int
 329 check_part(CHKARGS)
 330 {
 331
 332         if (MAN_BODY == n->type && 0 == n->nchild)
 333                 man_nmsg(m, n, MANDOCERR_NOBODY);
 334
 335         return(1);
 336 }
 337
 338
 339 static int
 340 check_par(CHKARGS)
 341 {
 342
 343         switch (n->type) {
 344         case (MAN_BLOCK):
 345                 if (0 == n->body->nchild)
 346                         man_node_delete(m, n);
 347                 break;
 348         case (MAN_BODY):
 349                 if (0 == n->nchild)
 350                         man_nmsg(m, n, MANDOCERR_IGNPAR);
 351                 break;
 352         case (MAN_HEAD):
 353                 if (n->nchild)
 354                         man_nmsg(m, n, MANDOCERR_ARGSLOST);
 355                 break;
 356         default:
 357                 break;
 358         }
 359
 360         return(1);
 361 }
 362
 363
 364 static int
 365 check_bline(CHKARGS)
 366 {
 367
 368         assert( ! (MAN_ELINE & m->flags));
 369         if (MAN_BLINE & m->flags) {
 370                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 371                 return(0);
 372         }
 373
 374         return(1);
 375 }
 376
 377 static int
 378 post_TH(CHKARGS)
 379 {
 380         const char      *p;
 381         int              line, pos;
 382
 383         if (m->meta.title)
 384                 free(m->meta.title);
 385         if (m->meta.vol)
 386                 free(m->meta.vol);
 387         if (m->meta.source)
 388                 free(m->meta.source);
 389         if (m->meta.msec)
 390                 free(m->meta.msec);
 391         if (m->meta.date)
 392                 free(m->meta.date);
 393
 394         line = n->line;
 395         pos = n->pos;
 396         m->meta.title = m->meta.vol = m->meta.date =
 397                 m->meta.msec = m->meta.source = NULL;
 398
 399         /* ->TITLE<- MSEC DATE SOURCE VOL */
 400
 401         n = n->child;
 402         if (n && n->string) {
 403                 for (p = n->string; '\0' != *p; p++) {
 404                         /* Only warn about this once... */
 405                         if (isalpha((u_char)*p) && ! isupper((u_char)*p)) {
 406                                 man_nmsg(m, n, MANDOCERR_UPPERCASE);
 407                                 break;
 408                         }
 409                 }
 410                 m->meta.title = mandoc_strdup(n->string);
 411         } else
 412                 m->meta.title = mandoc_strdup("");
 413
 414         /* TITLE ->MSEC<- DATE SOURCE VOL */
 415
 416         if (n)
 417                 n = n->next;
 418         if (n && n->string)
 419                 m->meta.msec = mandoc_strdup(n->string);
 420         else
 421                 m->meta.msec = mandoc_strdup("");
 422
 423         /* TITLE MSEC ->DATE<- SOURCE VOL */
 424
 425         if (n)
 426                 n = n->next;
 427         if (n)
 428                 pos = n->pos;
 429         m->meta.date = mandoc_normdate(n ? n->string : NULL,
 430             m->msg, m->data, line, pos);
 431
 432         /* TITLE MSEC DATE ->SOURCE<- VOL */
 433
 434         if (n && (n = n->next))
 435                 m->meta.source = mandoc_strdup(n->string);
 436
 437         /* TITLE MSEC DATE SOURCE ->VOL<- */
 438
 439         if (n && (n = n->next))
 440                 m->meta.vol = mandoc_strdup(n->string);
 441
 442         /*
 443          * Remove the `TH' node after we've processed it for our
 444          * meta-data.
 445          */
 446         man_node_delete(m, m->last);
 447         return(1);
 448 }
 449
 450 static int
 451 post_nf(CHKARGS)
 452 {
 453
 454         if (MAN_LITERAL & m->flags)
 455                 man_nmsg(m, n, MANDOCERR_SCOPEREP);
 456
 457         m->flags |= MAN_LITERAL;
 458         return(1);
 459 }
 460
 461 static int
 462 post_fi(CHKARGS)
 463 {
 464
 465         if ( ! (MAN_LITERAL & m->flags))
 466                 man_nmsg(m, n, MANDOCERR_WNOSCOPE);
 467
 468         m->flags &= ~MAN_LITERAL;
 469         return(1);
 470 }
 471
 472 static int
 473 post_UC(CHKARGS)
 474 {
 475         static const char * const bsd_versions[] = {
 476             "3rd Berkeley Distribution",
 477             "4th Berkeley Distribution",
 478             "4.2 Berkeley Distribution",
 479             "4.3 Berkeley Distribution",
 480             "4.4 Berkeley Distribution",
 481         };
 482
 483         const char      *p, *s;
 484
 485         n = n->child;
 486         n = m->last->child;
 487
 488         if (NULL == n || MAN_TEXT != n->type)
 489                 p = bsd_versions[0];
 490         else {
 491                 s = n->string;
 492                 if (0 == strcmp(s, "3"))
 493                         p = bsd_versions[0];
 494                 else if (0 == strcmp(s, "4"))
 495                         p = bsd_versions[1];
 496                 else if (0 == strcmp(s, "5"))
 497                         p = bsd_versions[2];
 498                 else if (0 == strcmp(s, "6"))
 499                         p = bsd_versions[3];
 500                 else if (0 == strcmp(s, "7"))
 501                         p = bsd_versions[4];
 502                 else
 503                         p = bsd_versions[0];
 504         }
 505
 506         if (m->meta.source)
 507                 free(m->meta.source);
 508
 509         m->meta.source = mandoc_strdup(p);
 510         return(1);
 511 }
 512
 513 static int
 514 post_AT(CHKARGS)
 515 {
 516         static const char * const unix_versions[] = {
 517             "7th Edition",
 518             "System III",
 519             "System V",
 520             "System V Release 2",
 521         };
 522
 523         const char      *p, *s;
 524         struct man_node *nn;
 525
 526         n = n->child;
 527
 528         if (NULL == n || MAN_TEXT != n->type)
 529                 p = unix_versions[0];
 530         else {
 531                 s = n->string;
 532                 if (0 == strcmp(s, "3"))
 533                         p = unix_versions[0];
 534                 else if (0 == strcmp(s, "4"))
 535                         p = unix_versions[1];
 536                 else if (0 == strcmp(s, "5")) {
 537                         nn = n->next;
 538                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 539                                 p = unix_versions[3];
 540                         else
 541                                 p = unix_versions[2];
 542                 } else
 543                         p = unix_versions[0];
 544         }
 545
 546         if (m->meta.source)
 547                 free(m->meta.source);
 548
 549         m->meta.source = mandoc_strdup(p);
 550         return(1);
 551 }