mandoc.c

   1 /*      $Id: mandoc.c,v 1.79 2014/04/20 16:46:04 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011, 2012, 2013, 2014 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 <errno.h>
  27 #include <limits.h>
  28 #include <stdlib.h>
  29 #include <stdio.h>
  30 #include <string.h>
  31 #include <time.h>
  32
  33 #include "mandoc.h"
  34 #include "mandoc_aux.h"
  35 #include "libmandoc.h"
  36
  37 #define DATESIZE 32
  38
  39 static  int      a2time(time_t *, const char *, const char *);
  40 static  char    *time2a(time_t);
  41
  42
  43 enum mandoc_esc
  44 mandoc_escape(const char **end, const char **start, int *sz)
  45 {
  46         const char      *local_start;
  47         int              local_sz;
  48         char             term;
  49         enum mandoc_esc  gly;
  50
  51         /*
  52          * When the caller doesn't provide return storage,
  53          * use local storage.
  54          */
  55
  56         if (NULL == start)
  57                 start = &local_start;
  58         if (NULL == sz)
  59                 sz = &local_sz;
  60
  61         /*
  62          * Beyond the backslash, at least one input character
  63          * is part of the escape sequence.  With one exception
  64          * (see below), that character won't be returned.
  65          */
  66
  67         gly = ESCAPE_ERROR;
  68         *start = ++*end;
  69         *sz = 0;
  70         term = '\0';
  71
  72         switch ((*start)[-1]) {
  73         /*
  74          * First the glyphs.  There are several different forms of
  75          * these, but each eventually returns a substring of the glyph
  76          * name.
  77          */
  78         case '(':
  79                 gly = ESCAPE_SPECIAL;
  80                 *sz = 2;
  81                 break;
  82         case '[':
  83                 gly = ESCAPE_SPECIAL;
  84                 /*
  85                  * Unicode escapes are defined in groff as \[uXXXX] to
  86                  * \[u10FFFF], where the contained value must be a valid
  87                  * Unicode codepoint.  Here, however, only check whether
  88                  * it's not a zero-width escape.
  89                  */
  90                 if ('u' == (*start)[0] && ']' != (*start)[1])
  91                         gly = ESCAPE_UNICODE;
  92                 term = ']';
  93                 break;
  94         case 'C':
  95                 if ('\'' != **start)
  96                         return(ESCAPE_ERROR);
  97                 *start = ++*end;
  98                 if ('u' == (*start)[0] && '\'' != (*start)[1])
  99                         gly = ESCAPE_UNICODE;
 100                 else
 101                         gly = ESCAPE_SPECIAL;
 102                 term = '\'';
 103                 break;
 104
 105         /*
 106          * Escapes taking no arguments at all.
 107          */
 108         case 'd':
 109                 /* FALLTHROUGH */
 110         case 'u':
 111                 return(ESCAPE_IGNORE);
 112
 113         /*
 114          * The \z escape is supposed to output the following
 115          * character without advancing the cursor position.
 116          * Since we are mostly dealing with terminal mode,
 117          * let us just skip the next character.
 118          */
 119         case 'z':
 120                 return(ESCAPE_SKIPCHAR);
 121
 122         /*
 123          * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
 124          * 'X' is the trigger.  These have opaque sub-strings.
 125          */
 126         case 'F':
 127                 /* FALLTHROUGH */
 128         case 'g':
 129                 /* FALLTHROUGH */
 130         case 'k':
 131                 /* FALLTHROUGH */
 132         case 'M':
 133                 /* FALLTHROUGH */
 134         case 'm':
 135                 /* FALLTHROUGH */
 136         case 'n':
 137                 /* FALLTHROUGH */
 138         case 'V':
 139                 /* FALLTHROUGH */
 140         case 'Y':
 141                 gly = ESCAPE_IGNORE;
 142                 /* FALLTHROUGH */
 143         case 'f':
 144                 if (ESCAPE_ERROR == gly)
 145                         gly = ESCAPE_FONT;
 146                 switch (**start) {
 147                 case '(':
 148                         *start = ++*end;
 149                         *sz = 2;
 150                         break;
 151                 case '[':
 152                         *start = ++*end;
 153                         term = ']';
 154                         break;
 155                 default:
 156                         *sz = 1;
 157                         break;
 158                 }
 159                 break;
 160
 161         /*
 162          * These escapes are of the form \X'Y', where 'X' is the trigger
 163          * and 'Y' is any string.  These have opaque sub-strings.
 164          * The \B and \w escapes are handled in roff.c, roff_res().
 165          */
 166         case 'A':
 167                 /* FALLTHROUGH */
 168         case 'b':
 169                 /* FALLTHROUGH */
 170         case 'D':
 171                 /* FALLTHROUGH */
 172         case 'o':
 173                 /* FALLTHROUGH */
 174         case 'R':
 175                 /* FALLTHROUGH */
 176         case 'X':
 177                 /* FALLTHROUGH */
 178         case 'Z':
 179                 if ('\0' == **start)
 180                         return(ESCAPE_ERROR);
 181                 gly = ESCAPE_IGNORE;
 182                 term = **start;
 183                 *start = ++*end;
 184                 break;
 185
 186         /*
 187          * These escapes are of the form \X'N', where 'X' is the trigger
 188          * and 'N' resolves to a numerical expression.
 189          */
 190         case 'h':
 191                 /* FALLTHROUGH */
 192         case 'H':
 193                 /* FALLTHROUGH */
 194         case 'L':
 195                 /* FALLTHROUGH */
 196         case 'l':
 197                 /* FALLTHROUGH */
 198         case 'S':
 199                 /* FALLTHROUGH */
 200         case 'v':
 201                 /* FALLTHROUGH */
 202         case 'x':
 203                 if (strchr("\0 %&()*+-./0123456789:<=>", **start))
 204                         return(ESCAPE_ERROR);
 205                 gly = ESCAPE_IGNORE;
 206                 term = **start;
 207                 *start = ++*end;
 208                 break;
 209
 210         /*
 211          * Special handling for the numbered character escape.
 212          * XXX Do any other escapes need similar handling?
 213          */
 214         case 'N':
 215                 if ('\0' == **start)
 216                         return(ESCAPE_ERROR);
 217                 (*end)++;
 218                 if (isdigit((unsigned char)**start)) {
 219                         *sz = 1;
 220                         return(ESCAPE_IGNORE);
 221                 }
 222                 (*start)++;
 223                 while (isdigit((unsigned char)**end))
 224                         (*end)++;
 225                 *sz = *end - *start;
 226                 if ('\0' != **end)
 227                         (*end)++;
 228                 return(ESCAPE_NUMBERED);
 229
 230         /*
 231          * Sizes get a special category of their own.
 232          */
 233         case 's':
 234                 gly = ESCAPE_IGNORE;
 235
 236                 /* See +/- counts as a sign. */
 237                 if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
 238                         (*end)++;
 239
 240                 switch (**end) {
 241                 case '(':
 242                         *start = ++*end;
 243                         *sz = 2;
 244                         break;
 245                 case '[':
 246                         *start = ++*end;
 247                         term = ']';
 248                         break;
 249                 case '\'':
 250                         *start = ++*end;
 251                         term = '\'';
 252                         break;
 253                 default:
 254                         *sz = 1;
 255                         break;
 256                 }
 257
 258                 break;
 259
 260         /*
 261          * Anything else is assumed to be a glyph.
 262          * In this case, pass back the character after the backslash.
 263          */
 264         default:
 265                 gly = ESCAPE_SPECIAL;
 266                 *start = --*end;
 267                 *sz = 1;
 268                 break;
 269         }
 270
 271         assert(ESCAPE_ERROR != gly);
 272
 273         /*
 274          * Read up to the terminating character,
 275          * paying attention to nested escapes.
 276          */
 277
 278         if ('\0' != term) {
 279                 while (**end != term) {
 280                         switch (**end) {
 281                         case '\0':
 282                                 return(ESCAPE_ERROR);
 283                         case '\\':
 284                                 (*end)++;
 285                                 if (ESCAPE_ERROR ==
 286                                     mandoc_escape(end, NULL, NULL))
 287                                         return(ESCAPE_ERROR);
 288                                 break;
 289                         default:
 290                                 (*end)++;
 291                                 break;
 292                         }
 293                 }
 294                 *sz = (*end)++ - *start;
 295         } else {
 296                 assert(*sz > 0);
 297                 if ((size_t)*sz > strlen(*start))
 298                         return(ESCAPE_ERROR);
 299                 *end += *sz;
 300         }
 301
 302         /* Run post-processors. */
 303
 304         switch (gly) {
 305         case ESCAPE_FONT:
 306                 if (2 == *sz) {
 307                         if ('C' == **start) {
 308                                 /*
 309                                  * Treat constant-width font modes
 310                                  * just like regular font modes.
 311                                  */
 312                                 (*start)++;
 313                                 (*sz)--;
 314                         } else {
 315                                 if ('B' == (*start)[0] && 'I' == (*start)[1])
 316                                         gly = ESCAPE_FONTBI;
 317                                 break;
 318                         }
 319                 } else if (1 != *sz)
 320                         break;
 321
 322                 switch (**start) {
 323                 case '3':
 324                         /* FALLTHROUGH */
 325                 case 'B':
 326                         gly = ESCAPE_FONTBOLD;
 327                         break;
 328                 case '2':
 329                         /* FALLTHROUGH */
 330                 case 'I':
 331                         gly = ESCAPE_FONTITALIC;
 332                         break;
 333                 case 'P':
 334                         gly = ESCAPE_FONTPREV;
 335                         break;
 336                 case '1':
 337                         /* FALLTHROUGH */
 338                 case 'R':
 339                         gly = ESCAPE_FONTROMAN;
 340                         break;
 341                 }
 342                 break;
 343         case ESCAPE_SPECIAL:
 344                 if (1 == *sz && 'c' == **start)
 345                         gly = ESCAPE_NOSPACE;
 346                 break;
 347         default:
 348                 break;
 349         }
 350
 351         return(gly);
 352 }
 353
 354 /*
 355  * Parse a quoted or unquoted roff-style request or macro argument.
 356  * Return a pointer to the parsed argument, which is either the original
 357  * pointer or advanced by one byte in case the argument is quoted.
 358  * NUL-terminate the argument in place.
 359  * Collapse pairs of quotes inside quoted arguments.
 360  * Advance the argument pointer to the next argument,
 361  * or to the NUL byte terminating the argument line.
 362  */
 363 char *
 364 mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
 365 {
 366         char     *start, *cp;
 367         int       quoted, pairs, white;
 368
 369         /* Quoting can only start with a new word. */
 370         start = *cpp;
 371         quoted = 0;
 372         if ('"' == *start) {
 373                 quoted = 1;
 374                 start++;
 375         }
 376
 377         pairs = 0;
 378         white = 0;
 379         for (cp = start; '\0' != *cp; cp++) {
 380
 381                 /*
 382                  * Move the following text left
 383                  * after quoted quotes and after "\\" and "\t".
 384                  */
 385                 if (pairs)
 386                         cp[-pairs] = cp[0];
 387
 388                 if ('\\' == cp[0]) {
 389                         /*
 390                          * In copy mode, translate double to single
 391                          * backslashes and backslash-t to literal tabs.
 392                          */
 393                         switch (cp[1]) {
 394                         case 't':
 395                                 cp[0] = '\t';
 396                                 /* FALLTHROUGH */
 397                         case '\\':
 398                                 pairs++;
 399                                 cp++;
 400                                 break;
 401                         case ' ':
 402                                 /* Skip escaped blanks. */
 403                                 if (0 == quoted)
 404                                         cp++;
 405                                 break;
 406                         default:
 407                                 break;
 408                         }
 409                 } else if (0 == quoted) {
 410                         if (' ' == cp[0]) {
 411                                 /* Unescaped blanks end unquoted args. */
 412                                 white = 1;
 413                                 break;
 414                         }
 415                 } else if ('"' == cp[0]) {
 416                         if ('"' == cp[1]) {
 417                                 /* Quoted quotes collapse. */
 418                                 pairs++;
 419                                 cp++;
 420                         } else {
 421                                 /* Unquoted quotes end quoted args. */
 422                                 quoted = 2;
 423                                 break;
 424                         }
 425                 }
 426         }
 427
 428         /* Quoted argument without a closing quote. */
 429         if (1 == quoted)
 430                 mandoc_msg(MANDOCERR_BADQUOTE, parse, ln, *pos, NULL);
 431
 432         /* NUL-terminate this argument and move to the next one. */
 433         if (pairs)
 434                 cp[-pairs] = '\0';
 435         if ('\0' != *cp) {
 436                 *cp++ = '\0';
 437                 while (' ' == *cp)
 438                         cp++;
 439         }
 440         *pos += (int)(cp - start) + (quoted ? 1 : 0);
 441         *cpp = cp;
 442
 443         if ('\0' == *cp && (white || ' ' == cp[-1]))
 444                 mandoc_msg(MANDOCERR_EOLNSPACE, parse, ln, *pos, NULL);
 445
 446         return(start);
 447 }
 448
 449 static int
 450 a2time(time_t *t, const char *fmt, const char *p)
 451 {
 452         struct tm        tm;
 453         char            *pp;
 454
 455         memset(&tm, 0, sizeof(struct tm));
 456
 457         pp = NULL;
 458 #ifdef  HAVE_STRPTIME
 459         pp = strptime(p, fmt, &tm);
 460 #endif
 461         if (NULL != pp && '\0' == *pp) {
 462                 *t = mktime(&tm);
 463                 return(1);
 464         }
 465
 466         return(0);
 467 }
 468
 469 static char *
 470 time2a(time_t t)
 471 {
 472         struct tm       *tm;
 473         char            *buf, *p;
 474         size_t           ssz;
 475         int              isz;
 476
 477         tm = localtime(&t);
 478
 479         /*
 480          * Reserve space:
 481          * up to 9 characters for the month (September) + blank
 482          * up to 2 characters for the day + comma + blank
 483          * 4 characters for the year and a terminating '\0'
 484          */
 485         p = buf = mandoc_malloc(10 + 4 + 4 + 1);
 486
 487         if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
 488                 goto fail;
 489         p += (int)ssz;
 490
 491         if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
 492                 goto fail;
 493         p += isz;
 494
 495         if (0 == strftime(p, 4 + 1, "%Y", tm))
 496                 goto fail;
 497         return(buf);
 498
 499 fail:
 500         free(buf);
 501         return(NULL);
 502 }
 503
 504 char *
 505 mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
 506 {
 507         char            *out;
 508         time_t           t;
 509
 510         if (NULL == in || '\0' == *in ||
 511             0 == strcmp(in, "$" "Mdocdate$")) {
 512                 mandoc_msg(MANDOCERR_NODATE, parse, ln, pos, NULL);
 513                 time(&t);
 514         }
 515         else if (a2time(&t, "%Y-%m-%d", in))
 516                 t = 0;
 517         else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
 518             !a2time(&t, "%b %d, %Y", in)) {
 519                 mandoc_msg(MANDOCERR_BADDATE, parse, ln, pos, NULL);
 520                 t = 0;
 521         }
 522         out = t ? time2a(t) : NULL;
 523         return(out ? out : mandoc_strdup(in));
 524 }
 525
 526 int
 527 mandoc_eos(const char *p, size_t sz)
 528 {
 529         const char      *q;
 530         int              enclosed, found;
 531
 532         if (0 == sz)
 533                 return(0);
 534
 535         /*
 536          * End-of-sentence recognition must include situations where
 537          * some symbols, such as `)', allow prior EOS punctuation to
 538          * propagate outward.
 539          */
 540
 541         enclosed = found = 0;
 542         for (q = p + (int)sz - 1; q >= p; q--) {
 543                 switch (*q) {
 544                 case '\"':
 545                         /* FALLTHROUGH */
 546                 case '\'':
 547                         /* FALLTHROUGH */
 548                 case ']':
 549                         /* FALLTHROUGH */
 550                 case ')':
 551                         if (0 == found)
 552                                 enclosed = 1;
 553                         break;
 554                 case '.':
 555                         /* FALLTHROUGH */
 556                 case '!':
 557                         /* FALLTHROUGH */
 558                 case '?':
 559                         found = 1;
 560                         break;
 561                 default:
 562                         return(found && (!enclosed || isalnum((unsigned char)*q)));
 563                 }
 564         }
 565
 566         return(found && !enclosed);
 567 }
 568
 569 /*
 570  * Convert a string to a long that may not be <0.
 571  * If the string is invalid, or is less than 0, return -1.
 572  */
 573 int
 574 mandoc_strntoi(const char *p, size_t sz, int base)
 575 {
 576         char             buf[32];
 577         char            *ep;
 578         long             v;
 579
 580         if (sz > 31)
 581                 return(-1);
 582
 583         memcpy(buf, p, sz);
 584         buf[(int)sz] = '\0';
 585
 586         errno = 0;
 587         v = strtol(buf, &ep, base);
 588
 589         if (buf[0] == '\0' || *ep != '\0')
 590                 return(-1);
 591
 592         if (v > INT_MAX)
 593                 v = INT_MAX;
 594         if (v < INT_MIN)
 595                 v = INT_MIN;
 596
 597         return((int)v);
 598 }