mandoc.c

   1 /*      $Id: mandoc.c,v 1.41 2011/03/17 09:18:12 kristaps 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 #define DATESIZE 32
  35
  36 static  int      a2time(time_t *, const char *, const char *);
  37 static  char    *time2a(time_t);
  38
  39 int
  40 mandoc_special(char *p)
  41 {
  42         int              len, i;
  43         char             term;
  44         char            *sv;
  45
  46         len = 0;
  47         term = '\0';
  48         sv = p;
  49
  50         assert('\\' == *p);
  51         p++;
  52
  53         switch (*p++) {
  54 #if 0
  55         case ('Z'):
  56                 /* FALLTHROUGH */
  57         case ('X'):
  58                 /* FALLTHROUGH */
  59         case ('x'):
  60                 /* FALLTHROUGH */
  61         case ('S'):
  62                 /* FALLTHROUGH */
  63         case ('R'):
  64                 /* FALLTHROUGH */
  65         case ('N'):
  66                 /* FALLTHROUGH */
  67         case ('l'):
  68                 /* FALLTHROUGH */
  69         case ('L'):
  70                 /* FALLTHROUGH */
  71         case ('H'):
  72                 /* FALLTHROUGH */
  73         case ('h'):
  74                 /* FALLTHROUGH */
  75         case ('D'):
  76                 /* FALLTHROUGH */
  77         case ('C'):
  78                 /* FALLTHROUGH */
  79         case ('b'):
  80                 /* FALLTHROUGH */
  81         case ('B'):
  82                 /* FALLTHROUGH */
  83         case ('a'):
  84                 /* FALLTHROUGH */
  85         case ('A'):
  86                 if (*p++ != '\'')
  87                         return(0);
  88                 term = '\'';
  89                 break;
  90 #endif
  91         case ('h'):
  92                 /* FALLTHROUGH */
  93         case ('v'):
  94                 /* FALLTHROUGH */
  95         case ('s'):
  96                 if (ASCII_HYPH == *p)
  97                         *p = '-';
  98
  99                 i = 0;
 100                 if ('+' == *p || '-' == *p) {
 101                         p++;
 102                         i = 1;
 103                 }
 104
 105                 switch (*p++) {
 106                 case ('('):
 107                         len = 2;
 108                         break;
 109                 case ('['):
 110                         term = ']';
 111                         break;
 112                 case ('\''):
 113                         term = '\'';
 114                         break;
 115                 case ('0'):
 116                         i = 1;
 117                         /* FALLTHROUGH */
 118                 default:
 119                         len = 1;
 120                         p--;
 121                         break;
 122                 }
 123
 124                 if (ASCII_HYPH == *p)
 125                         *p = '-';
 126                 if ('+' == *p || '-' == *p) {
 127                         if (i)
 128                                 return(0);
 129                         p++;
 130                 }
 131
 132                 /* Handle embedded numerical subexp or escape. */
 133
 134                 if ('(' == *p) {
 135                         while (*p && ')' != *p)
 136                                 if ('\\' == *p++) {
 137                                         i = mandoc_special(--p);
 138                                         if (0 == i)
 139                                                 return(0);
 140                                         p += i;
 141                                 }
 142
 143                         if (')' == *p++)
 144                                 break;
 145
 146                         return(0);
 147                 } else if ('\\' == *p) {
 148                         if (0 == (i = mandoc_special(p)))
 149                                 return(0);
 150                         p += i;
 151                 }
 152
 153                 break;
 154 #if 0
 155         case ('Y'):
 156                 /* FALLTHROUGH */
 157         case ('V'):
 158                 /* FALLTHROUGH */
 159         case ('$'):
 160                 /* FALLTHROUGH */
 161         case ('n'):
 162                 /* FALLTHROUGH */
 163 #endif
 164         case ('k'):
 165                 /* FALLTHROUGH */
 166         case ('M'):
 167                 /* FALLTHROUGH */
 168         case ('m'):
 169                 /* FALLTHROUGH */
 170         case ('f'):
 171                 /* FALLTHROUGH */
 172         case ('F'):
 173                 /* FALLTHROUGH */
 174         case ('*'):
 175                 switch (*p++) {
 176                 case ('('):
 177                         len = 2;
 178                         break;
 179                 case ('['):
 180                         term = ']';
 181                         break;
 182                 default:
 183                         len = 1;
 184                         p--;
 185                         break;
 186                 }
 187                 break;
 188         case ('('):
 189                 len = 2;
 190                 break;
 191         case ('['):
 192                 term = ']';
 193                 break;
 194         case ('z'):
 195                 len = 1;
 196                 if ('\\' == *p) {
 197                         if (0 == (i = mandoc_special(p)))
 198                                 return(0);
 199                         p += i;
 200                         return(*p ? (int)(p - sv) : 0);
 201                 }
 202                 break;
 203         case ('o'):
 204                 /* FALLTHROUGH */
 205         case ('w'):
 206                 if ('\'' == *p++) {
 207                         term = '\'';
 208                         break;
 209                 }
 210                 /* FALLTHROUGH */
 211         default:
 212                 len = 1;
 213                 p--;
 214                 break;
 215         }
 216
 217         if (term) {
 218                 for ( ; *p && term != *p; p++)
 219                         if (ASCII_HYPH == *p)
 220                                 *p = '-';
 221                 return(*p ? (int)(p - sv) : 0);
 222         }
 223
 224         for (i = 0; *p && i < len; i++, p++)
 225                 if (ASCII_HYPH == *p)
 226                         *p = '-';
 227         return(i == len ? (int)(p - sv) : 0);
 228 }
 229
 230
 231 void *
 232 mandoc_calloc(size_t num, size_t size)
 233 {
 234         void            *ptr;
 235
 236         ptr = calloc(num, size);
 237         if (NULL == ptr) {
 238                 perror(NULL);
 239                 exit((int)MANDOCLEVEL_SYSERR);
 240         }
 241
 242         return(ptr);
 243 }
 244
 245
 246 void *
 247 mandoc_malloc(size_t size)
 248 {
 249         void            *ptr;
 250
 251         ptr = malloc(size);
 252         if (NULL == ptr) {
 253                 perror(NULL);
 254                 exit((int)MANDOCLEVEL_SYSERR);
 255         }
 256
 257         return(ptr);
 258 }
 259
 260
 261 void *
 262 mandoc_realloc(void *ptr, size_t size)
 263 {
 264
 265         ptr = realloc(ptr, size);
 266         if (NULL == ptr) {
 267                 perror(NULL);
 268                 exit((int)MANDOCLEVEL_SYSERR);
 269         }
 270
 271         return(ptr);
 272 }
 273
 274
 275 char *
 276 mandoc_strdup(const char *ptr)
 277 {
 278         char            *p;
 279
 280         p = strdup(ptr);
 281         if (NULL == p) {
 282                 perror(NULL);
 283                 exit((int)MANDOCLEVEL_SYSERR);
 284         }
 285
 286         return(p);
 287 }
 288
 289 /*
 290  * Parse a quoted or unquoted roff-style request or macro argument.
 291  * Return a pointer to the parsed argument, which is either the original
 292  * pointer or advanced by one byte in case the argument is quoted.
 293  * Null-terminate the argument in place.
 294  * Collapse pairs of quotes inside quoted arguments.
 295  * Advance the argument pointer to the next argument,
 296  * or to the null byte terminating the argument line.
 297  */
 298 char *
 299 mandoc_getarg(char **cpp, mandocmsg msg, void *data, int ln, int *pos)
 300 {
 301         char     *start, *cp;
 302         int       quoted, pairs, white;
 303
 304         /* Quoting can only start with a new word. */
 305         start = *cpp;
 306         if ('"' == *start) {
 307                 quoted = 1;
 308                 start++;
 309         } else
 310                 quoted = 0;
 311
 312         pairs = 0;
 313         white = 0;
 314         for (cp = start; '\0' != *cp; cp++) {
 315                 /* Move left after quoted quotes and escaped backslashes. */
 316                 if (pairs)
 317                         cp[-pairs] = cp[0];
 318                 if ('\\' == cp[0]) {
 319                         if ('\\' == cp[1]) {
 320                                 /* Poor man's copy mode. */
 321                                 pairs++;
 322                                 cp++;
 323                         } else if (0 == quoted && ' ' == cp[1])
 324                                 /* Skip escaped blanks. */
 325                                 cp++;
 326                 } else if (0 == quoted) {
 327                         if (' ' == cp[0]) {
 328                                 /* Unescaped blanks end unquoted args. */
 329                                 white = 1;
 330                                 break;
 331                         }
 332                 } else if ('"' == cp[0]) {
 333                         if ('"' == cp[1]) {
 334                                 /* Quoted quotes collapse. */
 335                                 pairs++;
 336                                 cp++;
 337                         } else {
 338                                 /* Unquoted quotes end quoted args. */
 339                                 quoted = 2;
 340                                 break;
 341                         }
 342                 }
 343         }
 344
 345         /* Quoted argument without a closing quote. */
 346         if (1 == quoted && msg)
 347                 (*msg)(MANDOCERR_BADQUOTE, data, ln, *pos, NULL);
 348
 349         /* Null-terminate this argument and move to the next one. */
 350         if (pairs)
 351                 cp[-pairs] = '\0';
 352         if ('\0' != *cp) {
 353                 *cp++ = '\0';
 354                 while (' ' == *cp)
 355                         cp++;
 356         }
 357         *pos += (int)(cp - start) + (quoted ? 1 : 0);
 358         *cpp = cp;
 359
 360         if ('\0' == *cp && msg && (white || ' ' == cp[-1]))
 361                 (*msg)(MANDOCERR_EOLNSPACE, data, ln, *pos, NULL);
 362
 363         return(start);
 364 }
 365
 366 static int
 367 a2time(time_t *t, const char *fmt, const char *p)
 368 {
 369         struct tm        tm;
 370         char            *pp;
 371
 372         memset(&tm, 0, sizeof(struct tm));
 373
 374         pp = strptime(p, fmt, &tm);
 375         if (NULL != pp && '\0' == *pp) {
 376                 *t = mktime(&tm);
 377                 return(1);
 378         }
 379
 380         return(0);
 381 }
 382
 383 static char *
 384 time2a(time_t t)
 385 {
 386         struct tm        tm;
 387         char            *buf, *p;
 388         size_t           ssz;
 389         int              isz;
 390
 391         localtime_r(&t, &tm);
 392
 393         /*
 394          * Reserve space:
 395          * up to 9 characters for the month (September) + blank
 396          * up to 2 characters for the day + comma + blank
 397          * 4 characters for the year and a terminating '\0'
 398          */
 399         p = buf = mandoc_malloc(10 + 4 + 4 + 1);
 400
 401         if (0 == (ssz = strftime(p, 10 + 1, "%B ", &tm)))
 402                 goto fail;
 403         p += (int)ssz;
 404
 405         if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm.tm_mday)))
 406                 goto fail;
 407         p += isz;
 408
 409         if (0 == strftime(p, 4 + 1, "%Y", &tm))
 410                 goto fail;
 411         return(buf);
 412
 413 fail:
 414         free(buf);
 415         return(NULL);
 416 }
 417
 418 char *
 419 mandoc_normdate(char *in, mandocmsg msg, void *data, int ln, int pos)
 420 {
 421         char            *out;
 422         time_t           t;
 423
 424         if (NULL == in || '\0' == *in ||
 425             0 == strcmp(in, "$" "Mdocdate$")) {
 426                 (*msg)(MANDOCERR_NODATE, data, ln, pos, NULL);
 427                 time(&t);
 428         }
 429         else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
 430             !a2time(&t, "%b %d, %Y", in) &&
 431             !a2time(&t, "%Y-%m-%d", in)) {
 432                 (*msg)(MANDOCERR_BADDATE, data, ln, pos, NULL);
 433                 t = 0;
 434         }
 435         out = t ? time2a(t) : NULL;
 436         return(out ? out : mandoc_strdup(in));
 437 }
 438
 439 int
 440 mandoc_eos(const char *p, size_t sz, int enclosed)
 441 {
 442         const char *q;
 443         int found;
 444
 445         if (0 == sz)
 446                 return(0);
 447
 448         /*
 449          * End-of-sentence recognition must include situations where
 450          * some symbols, such as `)', allow prior EOS punctuation to
 451          * propogate outward.
 452          */
 453
 454         found = 0;
 455         for (q = p + (int)sz - 1; q >= p; q--) {
 456                 switch (*q) {
 457                 case ('\"'):
 458                         /* FALLTHROUGH */
 459                 case ('\''):
 460                         /* FALLTHROUGH */
 461                 case (']'):
 462                         /* FALLTHROUGH */
 463                 case (')'):
 464                         if (0 == found)
 465                                 enclosed = 1;
 466                         break;
 467                 case ('.'):
 468                         /* FALLTHROUGH */
 469                 case ('!'):
 470                         /* FALLTHROUGH */
 471                 case ('?'):
 472                         found = 1;
 473                         break;
 474                 default:
 475                         return(found && (!enclosed || isalnum((unsigned char)*q)));
 476                 }
 477         }
 478
 479         return(found && !enclosed);
 480 }
 481
 482 int
 483 mandoc_hyph(const char *start, const char *c)
 484 {
 485
 486         /*
 487          * Choose whether to break at a hyphenated character.  We only
 488          * do this if it's free-standing within a word.
 489          */
 490
 491         /* Skip first/last character of buffer. */
 492         if (c == start || '\0' == *(c + 1))
 493                 return(0);
 494         /* Skip first/last character of word. */
 495         if ('\t' == *(c + 1) || '\t' == *(c - 1))
 496                 return(0);
 497         if (' ' == *(c + 1) || ' ' == *(c - 1))
 498                 return(0);
 499         /* Skip double invocations. */
 500         if ('-' == *(c + 1) || '-' == *(c - 1))
 501                 return(0);
 502         /* Skip escapes. */
 503         if ('\\' == *(c - 1))
 504                 return(0);
 505
 506         return(1);
 507 }
 508
 509 /*
 510  * Check if a string is a punctuation delimiter.  This only applies to
 511  * mdoc(7) documents, but as it's used in both front-ends and back-ends,
 512  * it needs to go here (instead of, say, in libmdoc.h).
 513  */
 514 enum mdelim
 515 mandoc_isdelim(const char *p)
 516 {
 517
 518         if ('\0' == p[0])
 519                 return(DELIM_NONE);
 520
 521         if ('\0' == p[1])
 522                 switch (p[0]) {
 523                 case('('):
 524                         /* FALLTHROUGH */
 525                 case('['):
 526                         return(DELIM_OPEN);
 527                 case('|'):
 528                         return(DELIM_MIDDLE);
 529                 case('.'):
 530                         /* FALLTHROUGH */
 531                 case(','):
 532                         /* FALLTHROUGH */
 533                 case(';'):
 534                         /* FALLTHROUGH */
 535                 case(':'):
 536                         /* FALLTHROUGH */
 537                 case('?'):
 538                         /* FALLTHROUGH */
 539                 case('!'):
 540                         /* FALLTHROUGH */
 541                 case(')'):
 542                         /* FALLTHROUGH */
 543                 case(']'):
 544                         return(DELIM_CLOSE);
 545                 default:
 546                         return(DELIM_NONE);
 547                 }
 548
 549         if ('\\' != p[0])
 550                 return(DELIM_NONE);
 551
 552         if (0 == strcmp(p + 1, "."))
 553                 return(DELIM_CLOSE);
 554         if (0 == strcmp(p + 1, "*(Ba"))
 555                 return(DELIM_MIDDLE);
 556
 557         return(DELIM_NONE);
 558 }