mandoc.c

   1 /*      $Id: mandoc.c,v 1.91 2015/01/21 20:33:25 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                 /* FALLTHROUGH */
  97         case 'u':
  98                 return(ESCAPE_IGNORE);
  99
 100         /*
 101          * The \z escape is supposed to output the following
 102          * character without advancing the cursor position.
 103          * Since we are mostly dealing with terminal mode,
 104          * let us just skip the next character.
 105          */
 106         case 'z':
 107                 return(ESCAPE_SKIPCHAR);
 108
 109         /*
 110          * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
 111          * 'X' is the trigger.  These have opaque sub-strings.
 112          */
 113         case 'F':
 114                 /* FALLTHROUGH */
 115         case 'g':
 116                 /* FALLTHROUGH */
 117         case 'k':
 118                 /* FALLTHROUGH */
 119         case 'M':
 120                 /* FALLTHROUGH */
 121         case 'm':
 122                 /* FALLTHROUGH */
 123         case 'n':
 124                 /* FALLTHROUGH */
 125         case 'V':
 126                 /* FALLTHROUGH */
 127         case 'Y':
 128                 gly = ESCAPE_IGNORE;
 129                 /* FALLTHROUGH */
 130         case 'f':
 131                 if (ESCAPE_ERROR == gly)
 132                         gly = ESCAPE_FONT;
 133                 switch (**start) {
 134                 case '(':
 135                         *start = ++*end;
 136                         *sz = 2;
 137                         break;
 138                 case '[':
 139                         *start = ++*end;
 140                         term = ']';
 141                         break;
 142                 default:
 143                         *sz = 1;
 144                         break;
 145                 }
 146                 break;
 147
 148         /*
 149          * These escapes are of the form \X'Y', where 'X' is the trigger
 150          * and 'Y' is any string.  These have opaque sub-strings.
 151          * The \B and \w escapes are handled in roff.c, roff_res().
 152          */
 153         case 'A':
 154                 /* FALLTHROUGH */
 155         case 'b':
 156                 /* FALLTHROUGH */
 157         case 'D':
 158                 /* FALLTHROUGH */
 159         case 'R':
 160                 /* FALLTHROUGH */
 161         case 'X':
 162                 /* FALLTHROUGH */
 163         case 'Z':
 164                 gly = ESCAPE_IGNORE;
 165                 /* FALLTHROUGH */
 166         case 'o':
 167                 if (**start == '\0')
 168                         return(ESCAPE_ERROR);
 169                 if (gly == ESCAPE_ERROR)
 170                         gly = ESCAPE_OVERSTRIKE;
 171                 term = **start;
 172                 *start = ++*end;
 173                 break;
 174
 175         /*
 176          * These escapes are of the form \X'N', where 'X' is the trigger
 177          * and 'N' resolves to a numerical expression.
 178          */
 179         case 'h':
 180                 /* FALLTHROUGH */
 181         case 'H':
 182                 /* FALLTHROUGH */
 183         case 'L':
 184                 /* FALLTHROUGH */
 185         case 'l':
 186                 /* FALLTHROUGH */
 187         case 'S':
 188                 /* FALLTHROUGH */
 189         case 'v':
 190                 /* FALLTHROUGH */
 191         case 'x':
 192                 if (strchr(" %&()*+-./0123456789:<=>", **start)) {
 193                         if ('\0' != **start)
 194                                 ++*end;
 195                         return(ESCAPE_ERROR);
 196                 }
 197                 gly = ESCAPE_IGNORE;
 198                 term = **start;
 199                 *start = ++*end;
 200                 break;
 201
 202         /*
 203          * Special handling for the numbered character escape.
 204          * XXX Do any other escapes need similar handling?
 205          */
 206         case 'N':
 207                 if ('\0' == **start)
 208                         return(ESCAPE_ERROR);
 209                 (*end)++;
 210                 if (isdigit((unsigned char)**start)) {
 211                         *sz = 1;
 212                         return(ESCAPE_IGNORE);
 213                 }
 214                 (*start)++;
 215                 while (isdigit((unsigned char)**end))
 216                         (*end)++;
 217                 *sz = *end - *start;
 218                 if ('\0' != **end)
 219                         (*end)++;
 220                 return(ESCAPE_NUMBERED);
 221
 222         /*
 223          * Sizes get a special category of their own.
 224          */
 225         case 's':
 226                 gly = ESCAPE_IGNORE;
 227
 228                 /* See +/- counts as a sign. */
 229                 if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
 230                         *start = ++*end;
 231
 232                 switch (**end) {
 233                 case '(':
 234                         *start = ++*end;
 235                         *sz = 2;
 236                         break;
 237                 case '[':
 238                         *start = ++*end;
 239                         term = ']';
 240                         break;
 241                 case '\'':
 242                         *start = ++*end;
 243                         term = '\'';
 244                         break;
 245                 default:
 246                         *sz = 1;
 247                         break;
 248                 }
 249
 250                 break;
 251
 252         /*
 253          * Anything else is assumed to be a glyph.
 254          * In this case, pass back the character after the backslash.
 255          */
 256         default:
 257                 gly = ESCAPE_SPECIAL;
 258                 *start = --*end;
 259                 *sz = 1;
 260                 break;
 261         }
 262
 263         assert(ESCAPE_ERROR != gly);
 264
 265         /*
 266          * Read up to the terminating character,
 267          * paying attention to nested escapes.
 268          */
 269
 270         if ('\0' != term) {
 271                 while (**end != term) {
 272                         switch (**end) {
 273                         case '\0':
 274                                 return(ESCAPE_ERROR);
 275                         case '\\':
 276                                 (*end)++;
 277                                 if (ESCAPE_ERROR ==
 278                                     mandoc_escape(end, NULL, NULL))
 279                                         return(ESCAPE_ERROR);
 280                                 break;
 281                         default:
 282                                 (*end)++;
 283                                 break;
 284                         }
 285                 }
 286                 *sz = (*end)++ - *start;
 287         } else {
 288                 assert(*sz > 0);
 289                 if ((size_t)*sz > strlen(*start))
 290                         return(ESCAPE_ERROR);
 291                 *end += *sz;
 292         }
 293
 294         /* Run post-processors. */
 295
 296         switch (gly) {
 297         case ESCAPE_FONT:
 298                 if (2 == *sz) {
 299                         if ('C' == **start) {
 300                                 /*
 301                                  * Treat constant-width font modes
 302                                  * just like regular font modes.
 303                                  */
 304                                 (*start)++;
 305                                 (*sz)--;
 306                         } else {
 307                                 if ('B' == (*start)[0] && 'I' == (*start)[1])
 308                                         gly = ESCAPE_FONTBI;
 309                                 break;
 310                         }
 311                 } else if (1 != *sz)
 312                         break;
 313
 314                 switch (**start) {
 315                 case '3':
 316                         /* FALLTHROUGH */
 317                 case 'B':
 318                         gly = ESCAPE_FONTBOLD;
 319                         break;
 320                 case '2':
 321                         /* FALLTHROUGH */
 322                 case 'I':
 323                         gly = ESCAPE_FONTITALIC;
 324                         break;
 325                 case 'P':
 326                         gly = ESCAPE_FONTPREV;
 327                         break;
 328                 case '1':
 329                         /* FALLTHROUGH */
 330                 case 'R':
 331                         gly = ESCAPE_FONTROMAN;
 332                         break;
 333                 }
 334                 break;
 335         case ESCAPE_SPECIAL:
 336                 if (1 == *sz && 'c' == **start)
 337                         gly = ESCAPE_NOSPACE;
 338                 /*
 339                  * Unicode escapes are defined in groff as \[u0000]
 340                  * to \[u10FFFF], where the contained value must be
 341                  * a valid Unicode codepoint.  Here, however, only
 342                  * check the length and range.
 343                  */
 344                 if (**start != 'u' || *sz < 5 || *sz > 7)
 345                         break;
 346                 if (*sz == 7 && ((*start)[1] != '1' || (*start)[2] != '0'))
 347                         break;
 348                 if (*sz == 6 && (*start)[1] == '0')
 349                         break;
 350                 if ((int)strspn(*start + 1, "0123456789ABCDEFabcdef")
 351                     + 1 == *sz)
 352                         gly = ESCAPE_UNICODE;
 353                 break;
 354         default:
 355                 break;
 356         }
 357
 358         return(gly);
 359 }
 360
 361 /*
 362  * Parse a quoted or unquoted roff-style request or macro argument.
 363  * Return a pointer to the parsed argument, which is either the original
 364  * pointer or advanced by one byte in case the argument is quoted.
 365  * NUL-terminate the argument in place.
 366  * Collapse pairs of quotes inside quoted arguments.
 367  * Advance the argument pointer to the next argument,
 368  * or to the NUL byte terminating the argument line.
 369  */
 370 char *
 371 mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
 372 {
 373         char     *start, *cp;
 374         int       quoted, pairs, white;
 375
 376         /* Quoting can only start with a new word. */
 377         start = *cpp;
 378         quoted = 0;
 379         if ('"' == *start) {
 380                 quoted = 1;
 381                 start++;
 382         }
 383
 384         pairs = 0;
 385         white = 0;
 386         for (cp = start; '\0' != *cp; cp++) {
 387
 388                 /*
 389                  * Move the following text left
 390                  * after quoted quotes and after "\\" and "\t".
 391                  */
 392                 if (pairs)
 393                         cp[-pairs] = cp[0];
 394
 395                 if ('\\' == cp[0]) {
 396                         /*
 397                          * In copy mode, translate double to single
 398                          * backslashes and backslash-t to literal tabs.
 399                          */
 400                         switch (cp[1]) {
 401                         case 't':
 402                                 cp[0] = '\t';
 403                                 /* FALLTHROUGH */
 404                         case '\\':
 405                                 pairs++;
 406                                 cp++;
 407                                 break;
 408                         case ' ':
 409                                 /* Skip escaped blanks. */
 410                                 if (0 == quoted)
 411                                         cp++;
 412                                 break;
 413                         default:
 414                                 break;
 415                         }
 416                 } else if (0 == quoted) {
 417                         if (' ' == cp[0]) {
 418                                 /* Unescaped blanks end unquoted args. */
 419                                 white = 1;
 420                                 break;
 421                         }
 422                 } else if ('"' == cp[0]) {
 423                         if ('"' == cp[1]) {
 424                                 /* Quoted quotes collapse. */
 425                                 pairs++;
 426                                 cp++;
 427                         } else {
 428                                 /* Unquoted quotes end quoted args. */
 429                                 quoted = 2;
 430                                 break;
 431                         }
 432                 }
 433         }
 434
 435         /* Quoted argument without a closing quote. */
 436         if (1 == quoted)
 437                 mandoc_msg(MANDOCERR_ARG_QUOTE, parse, ln, *pos, NULL);
 438
 439         /* NUL-terminate this argument and move to the next one. */
 440         if (pairs)
 441                 cp[-pairs] = '\0';
 442         if ('\0' != *cp) {
 443                 *cp++ = '\0';
 444                 while (' ' == *cp)
 445                         cp++;
 446         }
 447         *pos += (int)(cp - start) + (quoted ? 1 : 0);
 448         *cpp = cp;
 449
 450         if ('\0' == *cp && (white || ' ' == cp[-1]))
 451                 mandoc_msg(MANDOCERR_SPACE_EOL, parse, ln, *pos, NULL);
 452
 453         return(start);
 454 }
 455
 456 static int
 457 a2time(time_t *t, const char *fmt, const char *p)
 458 {
 459         struct tm        tm;
 460         char            *pp;
 461
 462         memset(&tm, 0, sizeof(struct tm));
 463
 464         pp = NULL;
 465 #if HAVE_STRPTIME
 466         pp = strptime(p, fmt, &tm);
 467 #endif
 468         if (NULL != pp && '\0' == *pp) {
 469                 *t = mktime(&tm);
 470                 return(1);
 471         }
 472
 473         return(0);
 474 }
 475
 476 static char *
 477 time2a(time_t t)
 478 {
 479         struct tm       *tm;
 480         char            *buf, *p;
 481         size_t           ssz;
 482         int              isz;
 483
 484         tm = localtime(&t);
 485         if (tm == NULL)
 486                 return(NULL);
 487
 488         /*
 489          * Reserve space:
 490          * up to 9 characters for the month (September) + blank
 491          * up to 2 characters for the day + comma + blank
 492          * 4 characters for the year and a terminating '\0'
 493          */
 494         p = buf = mandoc_malloc(10 + 4 + 4 + 1);
 495
 496         if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
 497                 goto fail;
 498         p += (int)ssz;
 499
 500         if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
 501                 goto fail;
 502         p += isz;
 503
 504         if (0 == strftime(p, 4 + 1, "%Y", tm))
 505                 goto fail;
 506         return(buf);
 507
 508 fail:
 509         free(buf);
 510         return(NULL);
 511 }
 512
 513 char *
 514 mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
 515 {
 516         char            *out;
 517         time_t           t;
 518
 519         if (NULL == in || '\0' == *in ||
 520             0 == strcmp(in, "$" "Mdocdate$")) {
 521                 mandoc_msg(MANDOCERR_DATE_MISSING, parse, ln, pos, NULL);
 522                 time(&t);
 523         }
 524         else if (a2time(&t, "%Y-%m-%d", in))
 525                 t = 0;
 526         else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
 527             !a2time(&t, "%b %d, %Y", in)) {
 528                 mandoc_msg(MANDOCERR_DATE_BAD, parse, ln, pos, in);
 529                 t = 0;
 530         }
 531         out = t ? time2a(t) : NULL;
 532         return(out ? out : mandoc_strdup(in));
 533 }
 534
 535 int
 536 mandoc_eos(const char *p, size_t sz)
 537 {
 538         const char      *q;
 539         int              enclosed, found;
 540
 541         if (0 == sz)
 542                 return(0);
 543
 544         /*
 545          * End-of-sentence recognition must include situations where
 546          * some symbols, such as `)', allow prior EOS punctuation to
 547          * propagate outward.
 548          */
 549
 550         enclosed = found = 0;
 551         for (q = p + (int)sz - 1; q >= p; q--) {
 552                 switch (*q) {
 553                 case '\"':
 554                         /* FALLTHROUGH */
 555                 case '\'':
 556                         /* FALLTHROUGH */
 557                 case ']':
 558                         /* FALLTHROUGH */
 559                 case ')':
 560                         if (0 == found)
 561                                 enclosed = 1;
 562                         break;
 563                 case '.':
 564                         /* FALLTHROUGH */
 565                 case '!':
 566                         /* FALLTHROUGH */
 567                 case '?':
 568                         found = 1;
 569                         break;
 570                 default:
 571                         return(found && (!enclosed || isalnum((unsigned char)*q)));
 572                 }
 573         }
 574
 575         return(found && !enclosed);
 576 }
 577
 578 /*
 579  * Convert a string to a long that may not be <0.
 580  * If the string is invalid, or is less than 0, return -1.
 581  */
 582 int
 583 mandoc_strntoi(const char *p, size_t sz, int base)
 584 {
 585         char             buf[32];
 586         char            *ep;
 587         long             v;
 588
 589         if (sz > 31)
 590                 return(-1);
 591
 592         memcpy(buf, p, sz);
 593         buf[(int)sz] = '\0';
 594
 595         errno = 0;
 596         v = strtol(buf, &ep, base);
 597
 598         if (buf[0] == '\0' || *ep != '\0')
 599                 return(-1);
 600
 601         if (v > INT_MAX)
 602                 v = INT_MAX;
 603         if (v < INT_MIN)
 604                 v = INT_MIN;
 605
 606         return((int)v);
 607 }