mandoc.c

   1 /*      $Id: mandoc.c,v 1.36 2011/01/03 22:42:37 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011 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 #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 <stdlib.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mandoc.h"
  32 #include "libmandoc.h"
  33
  34 static  int      a2time(time_t *, const char *, const char *);
  35
  36
  37 int
  38 mandoc_special(char *p)
  39 {
  40         int              len, i;
  41         char             term;
  42         char            *sv;
  43
  44         len = 0;
  45         term = '\0';
  46         sv = p;
  47
  48         assert('\\' == *p);
  49         p++;
  50
  51         switch (*p++) {
  52 #if 0
  53         case ('Z'):
  54                 /* FALLTHROUGH */
  55         case ('X'):
  56                 /* FALLTHROUGH */
  57         case ('x'):
  58                 /* FALLTHROUGH */
  59         case ('S'):
  60                 /* FALLTHROUGH */
  61         case ('R'):
  62                 /* FALLTHROUGH */
  63         case ('N'):
  64                 /* FALLTHROUGH */
  65         case ('l'):
  66                 /* FALLTHROUGH */
  67         case ('L'):
  68                 /* FALLTHROUGH */
  69         case ('H'):
  70                 /* FALLTHROUGH */
  71         case ('h'):
  72                 /* FALLTHROUGH */
  73         case ('D'):
  74                 /* FALLTHROUGH */
  75         case ('C'):
  76                 /* FALLTHROUGH */
  77         case ('b'):
  78                 /* FALLTHROUGH */
  79         case ('B'):
  80                 /* FALLTHROUGH */
  81         case ('a'):
  82                 /* FALLTHROUGH */
  83         case ('A'):
  84                 if (*p++ != '\'')
  85                         return(0);
  86                 term = '\'';
  87                 break;
  88 #endif
  89         case ('h'):
  90                 /* FALLTHROUGH */
  91         case ('v'):
  92                 /* FALLTHROUGH */
  93         case ('s'):
  94                 if (ASCII_HYPH == *p)
  95                         *p = '-';
  96
  97                 i = 0;
  98                 if ('+' == *p || '-' == *p) {
  99                         p++;
 100                         i = 1;
 101                 }
 102
 103                 switch (*p++) {
 104                 case ('('):
 105                         len = 2;
 106                         break;
 107                 case ('['):
 108                         term = ']';
 109                         break;
 110                 case ('\''):
 111                         term = '\'';
 112                         break;
 113                 case ('0'):
 114                         i = 1;
 115                         /* FALLTHROUGH */
 116                 default:
 117                         len = 1;
 118                         p--;
 119                         break;
 120                 }
 121
 122                 if (ASCII_HYPH == *p)
 123                         *p = '-';
 124                 if ('+' == *p || '-' == *p) {
 125                         if (i)
 126                                 return(0);
 127                         p++;
 128                 }
 129
 130                 /* Handle embedded numerical subexp or escape. */
 131
 132                 if ('(' == *p) {
 133                         while (*p && ')' != *p)
 134                                 if ('\\' == *p++) {
 135                                         i = mandoc_special(--p);
 136                                         if (0 == i)
 137                                                 return(0);
 138                                         p += i;
 139                                 }
 140
 141                         if (')' == *p++)
 142                                 break;
 143
 144                         return(0);
 145                 } else if ('\\' == *p) {
 146                         if (0 == (i = mandoc_special(p)))
 147                                 return(0);
 148                         p += i;
 149                 }
 150
 151                 break;
 152 #if 0
 153         case ('Y'):
 154                 /* FALLTHROUGH */
 155         case ('V'):
 156                 /* FALLTHROUGH */
 157         case ('$'):
 158                 /* FALLTHROUGH */
 159         case ('n'):
 160                 /* FALLTHROUGH */
 161 #endif
 162         case ('k'):
 163                 /* FALLTHROUGH */
 164         case ('M'):
 165                 /* FALLTHROUGH */
 166         case ('m'):
 167                 /* FALLTHROUGH */
 168         case ('f'):
 169                 /* FALLTHROUGH */
 170         case ('F'):
 171                 /* FALLTHROUGH */
 172         case ('*'):
 173                 switch (*p++) {
 174                 case ('('):
 175                         len = 2;
 176                         break;
 177                 case ('['):
 178                         term = ']';
 179                         break;
 180                 default:
 181                         len = 1;
 182                         p--;
 183                         break;
 184                 }
 185                 break;
 186         case ('('):
 187                 len = 2;
 188                 break;
 189         case ('['):
 190                 term = ']';
 191                 break;
 192         case ('z'):
 193                 len = 1;
 194                 if ('\\' == *p) {
 195                         if (0 == (i = mandoc_special(p)))
 196                                 return(0);
 197                         p += i;
 198                         return(*p ? (int)(p - sv) : 0);
 199                 }
 200                 break;
 201         case ('o'):
 202                 /* FALLTHROUGH */
 203         case ('w'):
 204                 if ('\'' == *p++) {
 205                         term = '\'';
 206                         break;
 207                 }
 208                 /* FALLTHROUGH */
 209         default:
 210                 len = 1;
 211                 p--;
 212                 break;
 213         }
 214
 215         if (term) {
 216                 for ( ; *p && term != *p; p++)
 217                         if (ASCII_HYPH == *p)
 218                                 *p = '-';
 219                 return(*p ? (int)(p - sv) : 0);
 220         }
 221
 222         for (i = 0; *p && i < len; i++, p++)
 223                 if (ASCII_HYPH == *p)
 224                         *p = '-';
 225         return(i == len ? (int)(p - sv) : 0);
 226 }
 227
 228
 229 void *
 230 mandoc_calloc(size_t num, size_t size)
 231 {
 232         void            *ptr;
 233
 234         ptr = calloc(num, size);
 235         if (NULL == ptr) {
 236                 perror(NULL);
 237                 exit((int)MANDOCLEVEL_SYSERR);
 238         }
 239
 240         return(ptr);
 241 }
 242
 243
 244 void *
 245 mandoc_malloc(size_t size)
 246 {
 247         void            *ptr;
 248
 249         ptr = malloc(size);
 250         if (NULL == ptr) {
 251                 perror(NULL);
 252                 exit((int)MANDOCLEVEL_SYSERR);
 253         }
 254
 255         return(ptr);
 256 }
 257
 258
 259 void *
 260 mandoc_realloc(void *ptr, size_t size)
 261 {
 262
 263         ptr = realloc(ptr, size);
 264         if (NULL == ptr) {
 265                 perror(NULL);
 266                 exit((int)MANDOCLEVEL_SYSERR);
 267         }
 268
 269         return(ptr);
 270 }
 271
 272
 273 char *
 274 mandoc_strdup(const char *ptr)
 275 {
 276         char            *p;
 277
 278         p = strdup(ptr);
 279         if (NULL == p) {
 280                 perror(NULL);
 281                 exit((int)MANDOCLEVEL_SYSERR);
 282         }
 283
 284         return(p);
 285 }
 286
 287 /*
 288  * Parse a quoted or unquoted roff-style request or macro argument.
 289  * Return a pointer to the parsed argument, which is either the original
 290  * pointer or advanced by one byte in case the argument is quoted.
 291  * Null-terminate the argument in place.
 292  * Collapse pairs of quotes inside quoted arguments.
 293  * Advance the argument pointer to the next argument,
 294  * or to the null byte terminating the argument line.
 295  */
 296 char *
 297 mandoc_getarg(char **cpp, mandocmsg msg, void *data, int ln, int *pos)
 298 {
 299         char     *start, *cp;
 300         int       quoted, pairs, white;
 301
 302         /* Quoting can only start with a new word. */
 303         start = *cpp;
 304         if ('"' == *start) {
 305                 quoted = 1;
 306                 start++;
 307         } else
 308                 quoted = 0;
 309
 310         pairs = 0;
 311         white = 0;
 312         for (cp = start; '\0' != *cp; cp++) {
 313                 /* Move left after quoted quotes and escaped backslashes. */
 314                 if (pairs)
 315                         cp[-pairs] = cp[0];
 316                 if ('\\' == cp[0]) {
 317                         if ('\\' == cp[1]) {
 318                                 /* Poor man's copy mode. */
 319                                 pairs++;
 320                                 cp++;
 321                         } else if (0 == quoted && ' ' == cp[1])
 322                                 /* Skip escaped blanks. */
 323                                 cp++;
 324                 } else if (0 == quoted) {
 325                         if (' ' == cp[0]) {
 326                                 /* Unescaped blanks end unquoted args. */
 327                                 white = 1;
 328                                 break;
 329                         }
 330                 } else if ('"' == cp[0]) {
 331                         if ('"' == cp[1]) {
 332                                 /* Quoted quotes collapse. */
 333                                 pairs++;
 334                                 cp++;
 335                         } else {
 336                                 /* Unquoted quotes end quoted args. */
 337                                 quoted = 2;
 338                                 break;
 339                         }
 340                 }
 341         }
 342
 343         /* Quoted argument without a closing quote. */
 344         if (1 == quoted && msg)
 345                 (*msg)(MANDOCERR_BADQUOTE, data, ln, *pos, NULL);
 346
 347         /* Null-terminate this argument and move to the next one. */
 348         if (pairs)
 349                 cp[-pairs] = '\0';
 350         if ('\0' != *cp) {
 351                 *cp++ = '\0';
 352                 while (' ' == *cp)
 353                         cp++;
 354         }
 355         *pos += (cp - start) + (quoted ? 1 : 0);
 356         *cpp = cp;
 357
 358         if ('\0' == *cp && msg && (white || ' ' == cp[-1]))
 359                 (*msg)(MANDOCERR_EOLNSPACE, data, ln, *pos, NULL);
 360
 361         return(start);
 362 }
 363
 364
 365 static int
 366 a2time(time_t *t, const char *fmt, const char *p)
 367 {
 368         struct tm        tm;
 369         char            *pp;
 370
 371         memset(&tm, 0, sizeof(struct tm));
 372
 373         pp = strptime(p, fmt, &tm);
 374         if (NULL != pp && '\0' == *pp) {
 375                 *t = mktime(&tm);
 376                 return(1);
 377         }
 378
 379         return(0);
 380 }
 381
 382
 383 /*
 384  * Convert from a manual date string (see mdoc(7) and man(7)) into a
 385  * date according to the stipulated date type.
 386  */
 387 time_t
 388 mandoc_a2time(int flags, const char *p)
 389 {
 390         time_t           t;
 391
 392         if (MTIME_MDOCDATE & flags) {
 393                 if (0 == strcmp(p, "$" "Mdocdate$"))
 394                         return(time(NULL));
 395                 if (a2time(&t, "$" "Mdocdate: %b %d %Y $", p))
 396                         return(t);
 397         }
 398
 399         if (MTIME_CANONICAL & flags || MTIME_REDUCED & flags)
 400                 if (a2time(&t, "%b %d, %Y", p))
 401                         return(t);
 402
 403         if (MTIME_ISO_8601 & flags)
 404                 if (a2time(&t, "%Y-%m-%d", p))
 405                         return(t);
 406
 407         if (MTIME_REDUCED & flags) {
 408                 if (a2time(&t, "%d, %Y", p))
 409                         return(t);
 410                 if (a2time(&t, "%Y", p))
 411                         return(t);
 412         }
 413
 414         return(0);
 415 }
 416
 417
 418 int
 419 mandoc_eos(const char *p, size_t sz, int enclosed)
 420 {
 421         const char *q;
 422         int found;
 423
 424         if (0 == sz)
 425                 return(0);
 426
 427         /*
 428          * End-of-sentence recognition must include situations where
 429          * some symbols, such as `)', allow prior EOS punctuation to
 430          * propogate outward.
 431          */
 432
 433         found = 0;
 434         for (q = p + (int)sz - 1; q >= p; q--) {
 435                 switch (*q) {
 436                 case ('\"'):
 437                         /* FALLTHROUGH */
 438                 case ('\''):
 439                         /* FALLTHROUGH */
 440                 case (']'):
 441                         /* FALLTHROUGH */
 442                 case (')'):
 443                         if (0 == found)
 444                                 enclosed = 1;
 445                         break;
 446                 case ('.'):
 447                         /* FALLTHROUGH */
 448                 case ('!'):
 449                         /* FALLTHROUGH */
 450                 case ('?'):
 451                         found = 1;
 452                         break;
 453                 default:
 454                         return(found && (!enclosed || isalnum((unsigned char)*q)));
 455                 }
 456         }
 457
 458         return(found && !enclosed);
 459 }
 460
 461
 462 int
 463 mandoc_hyph(const char *start, const char *c)
 464 {
 465
 466         /*
 467          * Choose whether to break at a hyphenated character.  We only
 468          * do this if it's free-standing within a word.
 469          */
 470
 471         /* Skip first/last character of buffer. */
 472         if (c == start || '\0' == *(c + 1))
 473                 return(0);
 474         /* Skip first/last character of word. */
 475         if ('\t' == *(c + 1) || '\t' == *(c - 1))
 476                 return(0);
 477         if (' ' == *(c + 1) || ' ' == *(c - 1))
 478                 return(0);
 479         /* Skip double invocations. */
 480         if ('-' == *(c + 1) || '-' == *(c - 1))
 481                 return(0);
 482         /* Skip escapes. */
 483         if ('\\' == *(c - 1))
 484                 return(0);
 485
 486         return(1);
 487 }