man_validate.c

   1 /*      $Id: man_validate.c,v 1.45 2010/06/28 14:39:17 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009 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
  30 #include "mandoc.h"
  31 #include "libman.h"
  32 #include "libmandoc.h"
  33
  34 #define CHKARGS   struct man *m, struct man_node *n
  35
  36 typedef int     (*v_check)(CHKARGS);
  37
  38 struct  man_valid {
  39         v_check  *pres;
  40         v_check  *posts;
  41 };
  42
  43 static  int       check_bline(CHKARGS);
  44 static  int       check_eq0(CHKARGS);
  45 static  int       check_le1(CHKARGS);
  46 static  int       check_ge2(CHKARGS);
  47 static  int       check_le5(CHKARGS);
  48 static  int       check_par(CHKARGS);
  49 static  int       check_part(CHKARGS);
  50 static  int       check_root(CHKARGS);
  51 static  int       check_sec(CHKARGS);
  52 static  int       check_text(CHKARGS);
  53 static  int       check_title(CHKARGS);
  54
  55 static  v_check   posts_eq0[] = { check_eq0, NULL };
  56 static  v_check   posts_th[] = { check_ge2, check_le5, check_title, NULL };
  57 static  v_check   posts_par[] = { check_par, NULL };
  58 static  v_check   posts_part[] = { check_part, NULL };
  59 static  v_check   posts_sec[] = { check_sec, NULL };
  60 static  v_check   posts_le1[] = { check_le1, NULL };
  61 static  v_check   pres_bline[] = { check_bline, NULL };
  62
  63 static  const struct man_valid man_valids[MAN_MAX] = {
  64         { NULL, posts_eq0 }, /* br */
  65         { pres_bline, posts_th }, /* TH */
  66         { pres_bline, posts_sec }, /* SH */
  67         { pres_bline, posts_sec }, /* SS */
  68         { pres_bline, posts_par }, /* TP */
  69         { pres_bline, posts_par }, /* LP */
  70         { pres_bline, posts_par }, /* PP */
  71         { pres_bline, posts_par }, /* P */
  72         { pres_bline, posts_par }, /* IP */
  73         { pres_bline, posts_par }, /* HP */
  74         { NULL, NULL }, /* SM */
  75         { NULL, NULL }, /* SB */
  76         { NULL, NULL }, /* BI */
  77         { NULL, NULL }, /* IB */
  78         { NULL, NULL }, /* BR */
  79         { NULL, NULL }, /* RB */
  80         { NULL, NULL }, /* R */
  81         { NULL, NULL }, /* B */
  82         { NULL, NULL }, /* I */
  83         { NULL, NULL }, /* IR */
  84         { NULL, NULL }, /* RI */
  85         { NULL, posts_eq0 }, /* na */
  86         { NULL, NULL }, /* i */
  87         { NULL, posts_le1 }, /* sp */
  88         { pres_bline, posts_eq0 }, /* nf */
  89         { pres_bline, posts_eq0 }, /* fi */
  90         { NULL, NULL }, /* r */
  91         { NULL, NULL }, /* RE */
  92         { NULL, posts_part }, /* RS */
  93         { NULL, NULL }, /* DT */
  94         { NULL, NULL }, /* UC */
  95         { NULL, NULL }, /* PD */
  96         { NULL, posts_le1 }, /* Sp */
  97         { pres_bline, posts_le1 }, /* Vb */
  98         { pres_bline, posts_eq0 }, /* Ve */
  99         { NULL, NULL }, /* AT */
 100 };
 101
 102
 103 int
 104 man_valid_pre(struct man *m, struct man_node *n)
 105 {
 106         v_check         *cp;
 107
 108         if (MAN_TEXT == n->type)
 109                 return(1);
 110         if (MAN_ROOT == n->type)
 111                 return(1);
 112
 113         if (NULL == (cp = man_valids[n->tok].pres))
 114                 return(1);
 115         for ( ; *cp; cp++)
 116                 if ( ! (*cp)(m, n))
 117                         return(0);
 118         return(1);
 119 }
 120
 121
 122 int
 123 man_valid_post(struct man *m)
 124 {
 125         v_check         *cp;
 126
 127         if (MAN_VALID & m->last->flags)
 128                 return(1);
 129         m->last->flags |= MAN_VALID;
 130
 131         switch (m->last->type) {
 132         case (MAN_TEXT):
 133                 return(check_text(m, m->last));
 134         case (MAN_ROOT):
 135                 return(check_root(m, m->last));
 136         default:
 137                 break;
 138         }
 139
 140         if (NULL == (cp = man_valids[m->last->tok].posts))
 141                 return(1);
 142         for ( ; *cp; cp++)
 143                 if ( ! (*cp)(m, m->last))
 144                         return(0);
 145
 146         return(1);
 147 }
 148
 149
 150 static int
 151 check_root(CHKARGS)
 152 {
 153
 154         if (MAN_BLINE & m->flags)
 155                 return(man_nmsg(m, n, MANDOCERR_SCOPEEXIT));
 156         if (MAN_ELINE & m->flags)
 157                 return(man_nmsg(m, n, MANDOCERR_SCOPEEXIT));
 158
 159         m->flags &= ~MAN_BLINE;
 160         m->flags &= ~MAN_ELINE;
 161
 162         if (NULL == m->first->child) {
 163                 man_nmsg(m, n, MANDOCERR_NODOCBODY);
 164                 return(0);
 165         } else if (NULL == m->meta.title) {
 166                 if ( ! man_nmsg(m, n, MANDOCERR_NOTITLE))
 167                         return(0);
 168                 /*
 169                  * If a title hasn't been set, do so now (by
 170                  * implication, date and section also aren't set).
 171                  *
 172                  * FIXME: this should be in man_action.c.
 173                  */
 174                 m->meta.title = mandoc_strdup("unknown");
 175                 m->meta.date = time(NULL);
 176                 m->meta.msec = mandoc_strdup("1");
 177         }
 178
 179         return(1);
 180 }
 181
 182
 183 static int
 184 check_title(CHKARGS)
 185 {
 186         const char      *p;
 187
 188         assert(n->child);
 189         /* FIXME: is this sufficient? */
 190         if ('\0' == *n->child->string) {
 191                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 192                 return(0);
 193         }
 194
 195         for (p = n->child->string; '\0' != *p; p++)
 196                 if (isalpha((u_char)*p) && ! isupper((u_char)*p))
 197                         if ( ! man_nmsg(m, n, MANDOCERR_UPPERCASE))
 198                                 return(0);
 199
 200         return(1);
 201 }
 202
 203
 204 static int
 205 check_text(CHKARGS)
 206 {
 207         char            *p;
 208         int              pos, c;
 209
 210         assert(n->string);
 211
 212         for (p = n->string, pos = n->pos + 1; *p; p++, pos++) {
 213                 if ('\\' == *p) {
 214                         c = mandoc_special(p);
 215                         if (c) {
 216                                 p += c - 1;
 217                                 pos += c - 1;
 218                                 continue;
 219                         }
 220
 221                         c = man_pmsg(m, n->line, pos, MANDOCERR_BADESCAPE);
 222                         if ( ! (MAN_IGN_ESCAPE & m->pflags) && ! c)
 223                                 return(c);
 224                 }
 225
 226                 /*
 227                  * FIXME: we absolutely cannot let \b get through or it
 228                  * will destroy some assumptions in terms of format.
 229                  */
 230
 231                 if ('\t' == *p || isprint((u_char)*p) || ASCII_HYPH == *p)
 232                         continue;
 233                 if ( ! man_pmsg(m, n->line, pos, MANDOCERR_BADCHAR))
 234                         return(0);
 235         }
 236
 237         return(1);
 238 }
 239
 240
 241 #define INEQ_DEFINE(x, ineq, name) \
 242 static int \
 243 check_##name(CHKARGS) \
 244 { \
 245         if (n->nchild ineq (x)) \
 246                 return(1); \
 247         man_vmsg(m, MANDOCERR_SYNTARGCOUNT, n->line, n->pos, \
 248                         "line arguments %s %d (have %d)", \
 249                         #ineq, (x), n->nchild); \
 250         return(0); \
 251 }
 252
 253 INEQ_DEFINE(0, ==, eq0)
 254 INEQ_DEFINE(1, <=, le1)
 255 INEQ_DEFINE(2, >=, ge2)
 256 INEQ_DEFINE(5, <=, le5)
 257
 258
 259 static int
 260 check_sec(CHKARGS)
 261 {
 262
 263         if (MAN_HEAD == n->type && 0 == n->nchild) {
 264                 man_nmsg(m, n, MANDOCERR_SYNTARGCOUNT);
 265                 return(0);
 266         } else if (MAN_BODY == n->type && 0 == n->nchild)
 267                 return(man_nmsg(m, n, MANDOCERR_NOBODY));
 268
 269         return(1);
 270 }
 271
 272
 273 static int
 274 check_part(CHKARGS)
 275 {
 276
 277         if (MAN_BODY == n->type && 0 == n->nchild)
 278                 return(man_nmsg(m, n, MANDOCERR_NOBODY));
 279         return(1);
 280 }
 281
 282
 283 static int
 284 check_par(CHKARGS)
 285 {
 286
 287         if (MAN_BODY == n->type)
 288                 switch (n->tok) {
 289                 case (MAN_IP):
 290                         /* FALLTHROUGH */
 291                 case (MAN_HP):
 292                         /* FALLTHROUGH */
 293                 case (MAN_TP):
 294                         /* Body-less lists are ok. */
 295                         break;
 296                 default:
 297                         if (n->nchild)
 298                                 break;
 299                         return(man_nmsg(m, n, MANDOCERR_NOBODY));
 300                 }
 301         if (MAN_HEAD == n->type)
 302                 switch (n->tok) {
 303                 case (MAN_PP):
 304                         /* FALLTHROUGH */
 305                 case (MAN_P):
 306                         /* FALLTHROUGH */
 307                 case (MAN_LP):
 308                         if (0 == n->nchild)
 309                                 break;
 310                         return(man_nmsg(m, n, MANDOCERR_ARGSLOST));
 311                 default:
 312                         if (n->nchild)
 313                                 break;
 314                         return(man_nmsg(m, n, MANDOCERR_NOARGS));
 315                 }
 316
 317         return(1);
 318 }
 319
 320
 321 static int
 322 check_bline(CHKARGS)
 323 {
 324
 325         assert( ! (MAN_ELINE & m->flags));
 326         if (MAN_BLINE & m->flags) {
 327                 man_nmsg(m, n, MANDOCERR_SYNTLINESCOPE);
 328                 return(0);
 329         }
 330
 331         return(1);
 332 }
 333