mandoc.c

   1 /*      $Id: mandoc.c,v 1.95 2015/10/12 00:08:15 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011-2015 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 #include "config.h"
  19
  20 #include <sys/types.h>
  21
  22 #include <assert.h>
  23 #include <ctype.h>
  24 #include <errno.h>
  25 #include <limits.h>
  26 #include <stdlib.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mandoc.h"
  32 #include "mandoc_aux.h"
  33 #include "libmandoc.h"
  34
  35 #define DATESIZE 32
  36
  37 static  int      a2time(time_t *, const char *, const char *);
  38 static  char    *time2a(time_t);
  39
  40
  41 enum mandoc_esc
  42 mandoc_escape(const char **end, const char **start, int *sz)
  43 {
  44         const char      *local_start;
  45         int              local_sz;
  46         char             term;
  47         enum mandoc_esc  gly;
  48
  49         /*
  50          * When the caller doesn't provide return storage,
  51          * use local storage.
  52          */
  53
  54         if (NULL == start)
  55                 start = &local_start;
  56         if (NULL == sz)
  57                 sz = &local_sz;
  58
  59         /*
  60          * Beyond the backslash, at least one input character
  61          * is part of the escape sequence.  With one exception
  62          * (see below), that character won't be returned.
  63          */
  64
  65         gly = ESCAPE_ERROR;
  66         *start = ++*end;
  67         *sz = 0;
  68         term = '\0';
  69
  70         switch ((*start)[-1]) {
  71         /*
  72          * First the glyphs.  There are several different forms of
  73          * these, but each eventually returns a substring of the glyph
  74          * name.
  75          */
  76         case '(':
  77                 gly = ESCAPE_SPECIAL;
  78                 *sz = 2;
  79                 break;
  80         case '[':
  81                 gly = ESCAPE_SPECIAL;
  82                 term = ']';
  83                 break;
  84         case 'C':
  85                 if ('\'' != **start)
  86                         return ESCAPE_ERROR;
  87                 *start = ++*end;
  88                 gly = ESCAPE_SPECIAL;
  89                 term = '\'';
  90                 break;
  91
  92         /*
  93          * Escapes taking no arguments at all.
  94          */
  95         case 'd':
  96         case 'u':
  97         case ',':
  98         case '/':
  99                 return ESCAPE_IGNORE;
 100
 101         /*
 102          * The \z escape is supposed to output the following
 103          * character without advancing the cursor position.
 104          * Since we are mostly dealing with terminal mode,
 105          * let us just skip the next character.
 106          */
 107         case 'z':
 108                 return ESCAPE_SKIPCHAR;
 109
 110         /*
 111          * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
 112          * 'X' is the trigger.  These have opaque sub-strings.
 113          */
 114         case 'F':
 115         case 'g':
 116         case 'k':
 117         case 'M':
 118         case 'm':
 119         case 'n':
 120         case 'V':
 121         case 'Y':
 122                 gly = ESCAPE_IGNORE;
 123                 /* FALLTHROUGH */
 124         case 'f':
 125                 if (ESCAPE_ERROR == gly)
 126                         gly = ESCAPE_FONT;
 127                 switch (**start) {
 128                 case '(':
 129                         *start = ++*end;
 130                         *sz = 2;
 131                         break;
 132                 case '[':
 133                         *start = ++*end;
 134                         term = ']';
 135                         break;
 136                 default:
 137                         *sz = 1;
 138                         break;
 139                 }
 140                 break;
 141
 142         /*
 143          * These escapes are of the form \X'Y', where 'X' is the trigger
 144          * and 'Y' is any string.  These have opaque sub-strings.
 145          * The \B and \w escapes are handled in roff.c, roff_res().
 146          */
 147         case 'A':
 148         case 'b':
 149         case 'D':
 150         case 'R':
 151         case 'X':
 152         case 'Z':
 153                 gly = ESCAPE_IGNORE;
 154                 /* FALLTHROUGH */
 155         case 'o':
 156                 if (**start == '\0')
 157                         return ESCAPE_ERROR;
 158                 if (gly == ESCAPE_ERROR)
 159                         gly = ESCAPE_OVERSTRIKE;
 160                 term = **start;
 161                 *start = ++*end;
 162                 break;
 163
 164         /*
 165          * These escapes are of the form \X'N', where 'X' is the trigger
 166          * and 'N' resolves to a numerical expression.
 167          */
 168         case 'h':
 169         case 'H':
 170         case 'L':
 171         case 'l':
 172         case 'S':
 173         case 'v':
 174         case 'x':
 175                 if (strchr(" %&()*+-./0123456789:<=>", **start)) {
 176                         if ('\0' != **start)
 177                                 ++*end;
 178                         return ESCAPE_ERROR;
 179                 }
 180                 gly = ESCAPE_IGNORE;
 181                 term = **start;
 182                 *start = ++*end;
 183                 break;
 184
 185         /*
 186          * Special handling for the numbered character escape.
 187          * XXX Do any other escapes need similar handling?
 188          */
 189         case 'N':
 190                 if ('\0' == **start)
 191                         return ESCAPE_ERROR;
 192                 (*end)++;
 193                 if (isdigit((unsigned char)**start)) {
 194                         *sz = 1;
 195                         return ESCAPE_IGNORE;
 196                 }
 197                 (*start)++;
 198                 while (isdigit((unsigned char)**end))
 199                         (*end)++;
 200                 *sz = *end - *start;
 201                 if ('\0' != **end)
 202                         (*end)++;
 203                 return ESCAPE_NUMBERED;
 204
 205         /*
 206          * Sizes get a special category of their own.
 207          */
 208         case 's':
 209                 gly = ESCAPE_IGNORE;
 210
 211                 /* See +/- counts as a sign. */
 212                 if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
 213                         *start = ++*end;
 214
 215                 switch (**end) {
 216                 case '(':
 217                         *start = ++*end;
 218                         *sz = 2;
 219                         break;
 220                 case '[':
 221                         *start = ++*end;
 222                         term = ']';
 223                         break;
 224                 case '\'':
 225                         *start = ++*end;
 226                         term = '\'';
 227                         break;
 228                 case '3':
 229                 case '2':
 230                 case '1':
 231                         *sz = (*end)[-1] == 's' &&
 232                             isdigit((unsigned char)(*end)[1]) ? 2 : 1;
 233                         break;
 234                 default:
 235                         *sz = 1;
 236                         break;
 237                 }
 238
 239                 break;
 240
 241         /*
 242          * Anything else is assumed to be a glyph.
 243          * In this case, pass back the character after the backslash.
 244          */
 245         default:
 246                 gly = ESCAPE_SPECIAL;
 247                 *start = --*end;
 248                 *sz = 1;
 249                 break;
 250         }
 251
 252         assert(ESCAPE_ERROR != gly);
 253
 254         /*
 255          * Read up to the terminating character,
 256          * paying attention to nested escapes.
 257          */
 258
 259         if ('\0' != term) {
 260                 while (**end != term) {
 261                         switch (**end) {
 262                         case '\0':
 263                                 return ESCAPE_ERROR;
 264                         case '\\':
 265                                 (*end)++;
 266                                 if (ESCAPE_ERROR ==
 267                                     mandoc_escape(end, NULL, NULL))
 268                                         return ESCAPE_ERROR;
 269                                 break;
 270                         default:
 271                                 (*end)++;
 272                                 break;
 273                         }
 274                 }
 275                 *sz = (*end)++ - *start;
 276         } else {
 277                 assert(*sz > 0);
 278                 if ((size_t)*sz > strlen(*start))
 279                         return ESCAPE_ERROR;
 280                 *end += *sz;
 281         }
 282
 283         /* Run post-processors. */
 284
 285         switch (gly) {
 286         case ESCAPE_FONT:
 287                 if (2 == *sz) {
 288                         if ('C' == **start) {
 289                                 /*
 290                                  * Treat constant-width font modes
 291                                  * just like regular font modes.
 292                                  */
 293                                 (*start)++;
 294                                 (*sz)--;
 295                         } else {
 296                                 if ('B' == (*start)[0] && 'I' == (*start)[1])
 297                                         gly = ESCAPE_FONTBI;
 298                                 break;
 299                         }
 300                 } else if (1 != *sz)
 301                         break;
 302
 303                 switch (**start) {
 304                 case '3':
 305                 case 'B':
 306                         gly = ESCAPE_FONTBOLD;
 307                         break;
 308                 case '2':
 309                 case 'I':
 310                         gly = ESCAPE_FONTITALIC;
 311                         break;
 312                 case 'P':
 313                         gly = ESCAPE_FONTPREV;
 314                         break;
 315                 case '1':
 316                 case 'R':
 317                         gly = ESCAPE_FONTROMAN;
 318                         break;
 319                 }
 320                 break;
 321         case ESCAPE_SPECIAL:
 322                 if (1 == *sz && 'c' == **start)
 323                         gly = ESCAPE_NOSPACE;
 324                 /*
 325                  * Unicode escapes are defined in groff as \[u0000]
 326                  * to \[u10FFFF], where the contained value must be
 327                  * a valid Unicode codepoint.  Here, however, only
 328                  * check the length and range.
 329                  */
 330                 if (**start != 'u' || *sz < 5 || *sz > 7)
 331                         break;
 332                 if (*sz == 7 && ((*start)[1] != '1' || (*start)[2] != '0'))
 333                         break;
 334                 if (*sz == 6 && (*start)[1] == '0')
 335                         break;
 336                 if ((int)strspn(*start + 1, "0123456789ABCDEFabcdef")
 337                     + 1 == *sz)
 338                         gly = ESCAPE_UNICODE;
 339                 break;
 340         default:
 341                 break;
 342         }
 343
 344         return gly;
 345 }
 346
 347 /*
 348  * Parse a quoted or unquoted roff-style request or macro argument.
 349  * Return a pointer to the parsed argument, which is either the original
 350  * pointer or advanced by one byte in case the argument is quoted.
 351  * NUL-terminate the argument in place.
 352  * Collapse pairs of quotes inside quoted arguments.
 353  * Advance the argument pointer to the next argument,
 354  * or to the NUL byte terminating the argument line.
 355  */
 356 char *
 357 mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
 358 {
 359         char     *start, *cp;
 360         int       quoted, pairs, white;
 361
 362         /* Quoting can only start with a new word. */
 363         start = *cpp;
 364         quoted = 0;
 365         if ('"' == *start) {
 366                 quoted = 1;
 367                 start++;
 368         }
 369
 370         pairs = 0;
 371         white = 0;
 372         for (cp = start; '\0' != *cp; cp++) {
 373
 374                 /*
 375                  * Move the following text left
 376                  * after quoted quotes and after "\\" and "\t".
 377                  */
 378                 if (pairs)
 379                         cp[-pairs] = cp[0];
 380
 381                 if ('\\' == cp[0]) {
 382                         /*
 383                          * In copy mode, translate double to single
 384                          * backslashes and backslash-t to literal tabs.
 385                          */
 386                         switch (cp[1]) {
 387                         case 't':
 388                                 cp[0] = '\t';
 389                                 /* FALLTHROUGH */
 390                         case '\\':
 391                                 pairs++;
 392                                 cp++;
 393                                 break;
 394                         case ' ':
 395                                 /* Skip escaped blanks. */
 396                                 if (0 == quoted)
 397                                         cp++;
 398                                 break;
 399                         default:
 400                                 break;
 401                         }
 402                 } else if (0 == quoted) {
 403                         if (' ' == cp[0]) {
 404                                 /* Unescaped blanks end unquoted args. */
 405                                 white = 1;
 406                                 break;
 407                         }
 408                 } else if ('"' == cp[0]) {
 409                         if ('"' == cp[1]) {
 410                                 /* Quoted quotes collapse. */
 411                                 pairs++;
 412                                 cp++;
 413                         } else {
 414                                 /* Unquoted quotes end quoted args. */
 415                                 quoted = 2;
 416                                 break;
 417                         }
 418                 }
 419         }
 420
 421         /* Quoted argument without a closing quote. */
 422         if (1 == quoted)
 423                 mandoc_msg(MANDOCERR_ARG_QUOTE, parse, ln, *pos, NULL);
 424
 425         /* NUL-terminate this argument and move to the next one. */
 426         if (pairs)
 427                 cp[-pairs] = '\0';
 428         if ('\0' != *cp) {
 429                 *cp++ = '\0';
 430                 while (' ' == *cp)
 431                         cp++;
 432         }
 433         *pos += (int)(cp - start) + (quoted ? 1 : 0);
 434         *cpp = cp;
 435
 436         if ('\0' == *cp && (white || ' ' == cp[-1]))
 437                 mandoc_msg(MANDOCERR_SPACE_EOL, parse, ln, *pos, NULL);
 438
 439         return start;
 440 }
 441
 442 static int
 443 a2time(time_t *t, const char *fmt, const char *p)
 444 {
 445         struct tm        tm;
 446         char            *pp;
 447
 448         memset(&tm, 0, sizeof(struct tm));
 449
 450         pp = NULL;
 451 #if HAVE_STRPTIME
 452         pp = strptime(p, fmt, &tm);
 453 #endif
 454         if (NULL != pp && '\0' == *pp) {
 455                 *t = mktime(&tm);
 456                 return 1;
 457         }
 458
 459         return 0;
 460 }
 461
 462 static char *
 463 time2a(time_t t)
 464 {
 465         struct tm       *tm;
 466         char            *buf, *p;
 467         size_t           ssz;
 468         int              isz;
 469
 470         tm = localtime(&t);
 471         if (tm == NULL)
 472                 return NULL;
 473
 474         /*
 475          * Reserve space:
 476          * up to 9 characters for the month (September) + blank
 477          * up to 2 characters for the day + comma + blank
 478          * 4 characters for the year and a terminating '\0'
 479          */
 480         p = buf = mandoc_malloc(10 + 4 + 4 + 1);
 481
 482         if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
 483                 goto fail;
 484         p += (int)ssz;
 485
 486         if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
 487                 goto fail;
 488         p += isz;
 489
 490         if (0 == strftime(p, 4 + 1, "%Y", tm))
 491                 goto fail;
 492         return buf;
 493
 494 fail:
 495         free(buf);
 496         return NULL;
 497 }
 498
 499 char *
 500 mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
 501 {
 502         char            *out;
 503         time_t           t;
 504
 505         if (NULL == in || '\0' == *in ||
 506             0 == strcmp(in, "$" "Mdocdate$")) {
 507                 mandoc_msg(MANDOCERR_DATE_MISSING, parse, ln, pos, NULL);
 508                 time(&t);
 509         }
 510         else if (a2time(&t, "%Y-%m-%d", in))
 511                 t = 0;
 512         else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
 513             !a2time(&t, "%b %d, %Y", in)) {
 514                 mandoc_msg(MANDOCERR_DATE_BAD, parse, ln, pos, in);
 515                 t = 0;
 516         }
 517         out = t ? time2a(t) : NULL;
 518         return out ? out : mandoc_strdup(in);
 519 }
 520
 521 int
 522 mandoc_eos(const char *p, size_t sz)
 523 {
 524         const char      *q;
 525         int              enclosed, found;
 526
 527         if (0 == sz)
 528                 return 0;
 529
 530         /*
 531          * End-of-sentence recognition must include situations where
 532          * some symbols, such as `)', allow prior EOS punctuation to
 533          * propagate outward.
 534          */
 535
 536         enclosed = found = 0;
 537         for (q = p + (int)sz - 1; q >= p; q--) {
 538                 switch (*q) {
 539                 case '\"':
 540                 case '\'':
 541                 case ']':
 542                 case ')':
 543                         if (0 == found)
 544                                 enclosed = 1;
 545                         break;
 546                 case '.':
 547                 case '!':
 548                 case '?':
 549                         found = 1;
 550                         break;
 551                 default:
 552                         return found &&
 553                             (!enclosed || isalnum((unsigned char)*q));
 554                 }
 555         }
 556
 557         return found && !enclosed;
 558 }
 559
 560 /*
 561  * Convert a string to a long that may not be <0.
 562  * If the string is invalid, or is less than 0, return -1.
 563  */
 564 int
 565 mandoc_strntoi(const char *p, size_t sz, int base)
 566 {
 567         char             buf[32];
 568         char            *ep;
 569         long             v;
 570
 571         if (sz > 31)
 572                 return -1;
 573
 574         memcpy(buf, p, sz);
 575         buf[(int)sz] = '\0';
 576
 577         errno = 0;
 578         v = strtol(buf, &ep, base);
 579
 580         if (buf[0] == '\0' || *ep != '\0')
 581                 return -1;
 582
 583         if (v > INT_MAX)
 584                 v = INT_MAX;
 585         if (v < INT_MIN)
 586                 v = INT_MIN;
 587
 588         return (int)v;
 589 }