mdoc.c

   1 /*      $Id: mdoc.c,v 1.250 2015/04/19 14:57:38 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2015 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 <stdarg.h>
  25 #include <stdio.h>
  26 #include <stdlib.h>
  27 #include <string.h>
  28 #include <time.h>
  29
  30 #include "mandoc_aux.h"
  31 #include "mandoc.h"
  32 #include "roff.h"
  33 #include "mdoc.h"
  34 #include "libmandoc.h"
  35 #include "roff_int.h"
  36 #include "libmdoc.h"
  37
  38 const   char *const __mdoc_macronames[MDOC_MAX + 1] = {
  39         "text",
  40         "Ap",           "Dd",           "Dt",           "Os",
  41         "Sh",           "Ss",           "Pp",           "D1",
  42         "Dl",           "Bd",           "Ed",           "Bl",
  43         "El",           "It",           "Ad",           "An",
  44         "Ar",           "Cd",           "Cm",           "Dv",
  45         "Er",           "Ev",           "Ex",           "Fa",
  46         "Fd",           "Fl",           "Fn",           "Ft",
  47         "Ic",           "In",           "Li",           "Nd",
  48         "Nm",           "Op",           "Ot",           "Pa",
  49         "Rv",           "St",           "Va",           "Vt",
  50         "Xr",           "%A",           "%B",           "%D",
  51         "%I",           "%J",           "%N",           "%O",
  52         "%P",           "%R",           "%T",           "%V",
  53         "Ac",           "Ao",           "Aq",           "At",
  54         "Bc",           "Bf",           "Bo",           "Bq",
  55         "Bsx",          "Bx",           "Db",           "Dc",
  56         "Do",           "Dq",           "Ec",           "Ef",
  57         "Em",           "Eo",           "Fx",           "Ms",
  58         "No",           "Ns",           "Nx",           "Ox",
  59         "Pc",           "Pf",           "Po",           "Pq",
  60         "Qc",           "Ql",           "Qo",           "Qq",
  61         "Re",           "Rs",           "Sc",           "So",
  62         "Sq",           "Sm",           "Sx",           "Sy",
  63         "Tn",           "Ux",           "Xc",           "Xo",
  64         "Fo",           "Fc",           "Oo",           "Oc",
  65         "Bk",           "Ek",           "Bt",           "Hf",
  66         "Fr",           "Ud",           "Lb",           "Lp",
  67         "Lk",           "Mt",           "Brq",          "Bro",
  68         "Brc",          "%C",           "Es",           "En",
  69         "Dx",           "%Q",           "br",           "sp",
  70         "%U",           "Ta",           "ll",
  71 };
  72
  73 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  74         "split",                "nosplit",              "ragged",
  75         "unfilled",             "literal",              "file",
  76         "offset",               "bullet",               "dash",
  77         "hyphen",               "item",                 "enum",
  78         "tag",                  "diag",                 "hang",
  79         "ohang",                "inset",                "column",
  80         "width",                "compact",              "std",
  81         "filled",               "words",                "emphasis",
  82         "symbolic",             "nested",               "centered"
  83         };
  84
  85 const   char * const *mdoc_macronames = __mdoc_macronames + 1;
  86 const   char * const *mdoc_argnames = __mdoc_argnames;
  87
  88 static  int               mdoc_ptext(struct roff_man *, int, char *, int);
  89 static  int               mdoc_pmacro(struct roff_man *, int, char *, int);
  90
  91
  92 void
  93 mdoc_endparse(struct roff_man *mdoc)
  94 {
  95
  96         mdoc_macroend(mdoc);
  97 }
  98
  99 /*
 100  * Main parse routine.  Parses a single line -- really just hands off to
 101  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 102  */
 103 int
 104 mdoc_parseln(struct roff_man *mdoc, int ln, char *buf, int offs)
 105 {
 106
 107         if (mdoc->last->type != ROFFT_EQN || ln > mdoc->last->line)
 108                 mdoc->flags |= MDOC_NEWLINE;
 109
 110         /*
 111          * Let the roff nS register switch SYNOPSIS mode early,
 112          * such that the parser knows at all times
 113          * whether this mode is on or off.
 114          * Note that this mode is also switched by the Sh macro.
 115          */
 116         if (roff_getreg(mdoc->roff, "nS"))
 117                 mdoc->flags |= MDOC_SYNOPSIS;
 118         else
 119                 mdoc->flags &= ~MDOC_SYNOPSIS;
 120
 121         return(roff_getcontrol(mdoc->roff, buf, &offs) ?
 122             mdoc_pmacro(mdoc, ln, buf, offs) :
 123             mdoc_ptext(mdoc, ln, buf, offs));
 124 }
 125
 126 void
 127 mdoc_macro(MACRO_PROT_ARGS)
 128 {
 129         assert(tok > TOKEN_NONE && tok < MDOC_MAX);
 130
 131         if (mdoc->flags & MDOC_PBODY) {
 132                 if (tok == MDOC_Dt) {
 133                         mandoc_vmsg(MANDOCERR_DT_LATE,
 134                             mdoc->parse, line, ppos,
 135                             "Dt %s", buf + *pos);
 136                         return;
 137                 }
 138         } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
 139                 if (mdoc->meta.title == NULL) {
 140                         mandoc_vmsg(MANDOCERR_DT_NOTITLE,
 141                             mdoc->parse, line, ppos, "%s %s",
 142                             mdoc_macronames[tok], buf + *pos);
 143                         mdoc->meta.title = mandoc_strdup("UNTITLED");
 144                 }
 145                 if (NULL == mdoc->meta.vol)
 146                         mdoc->meta.vol = mandoc_strdup("LOCAL");
 147                 mdoc->flags |= MDOC_PBODY;
 148         }
 149         (*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf);
 150 }
 151
 152 void
 153 mdoc_tail_alloc(struct roff_man *mdoc, int line, int pos, int tok)
 154 {
 155         struct roff_node *p;
 156
 157         p = roff_node_alloc(mdoc, line, pos, ROFFT_TAIL, tok);
 158         roff_node_append(mdoc, p);
 159         mdoc->next = ROFF_NEXT_CHILD;
 160 }
 161
 162 struct roff_node *
 163 mdoc_endbody_alloc(struct roff_man *mdoc, int line, int pos, int tok,
 164                 struct roff_node *body, enum mdoc_endbody end)
 165 {
 166         struct roff_node *p;
 167
 168         body->flags |= MDOC_ENDED;
 169         body->parent->flags |= MDOC_ENDED;
 170         p = roff_node_alloc(mdoc, line, pos, ROFFT_BODY, tok);
 171         p->body = body;
 172         p->norm = body->norm;
 173         p->end = end;
 174         roff_node_append(mdoc, p);
 175         mdoc->next = ROFF_NEXT_SIBLING;
 176         return(p);
 177 }
 178
 179 struct roff_node *
 180 mdoc_block_alloc(struct roff_man *mdoc, int line, int pos,
 181         int tok, struct mdoc_arg *args)
 182 {
 183         struct roff_node *p;
 184
 185         p = roff_node_alloc(mdoc, line, pos, ROFFT_BLOCK, tok);
 186         p->args = args;
 187         if (p->args)
 188                 (args->refcnt)++;
 189
 190         switch (tok) {
 191         case MDOC_Bd:
 192                 /* FALLTHROUGH */
 193         case MDOC_Bf:
 194                 /* FALLTHROUGH */
 195         case MDOC_Bl:
 196                 /* FALLTHROUGH */
 197         case MDOC_En:
 198                 /* FALLTHROUGH */
 199         case MDOC_Rs:
 200                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 201                 break;
 202         default:
 203                 break;
 204         }
 205         roff_node_append(mdoc, p);
 206         mdoc->next = ROFF_NEXT_CHILD;
 207         return(p);
 208 }
 209
 210 void
 211 mdoc_elem_alloc(struct roff_man *mdoc, int line, int pos,
 212         int tok, struct mdoc_arg *args)
 213 {
 214         struct roff_node *p;
 215
 216         p = roff_node_alloc(mdoc, line, pos, ROFFT_ELEM, tok);
 217         p->args = args;
 218         if (p->args)
 219                 (args->refcnt)++;
 220
 221         switch (tok) {
 222         case MDOC_An:
 223                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 224                 break;
 225         default:
 226                 break;
 227         }
 228         roff_node_append(mdoc, p);
 229         mdoc->next = ROFF_NEXT_CHILD;
 230 }
 231
 232 void
 233 mdoc_node_relink(struct roff_man *mdoc, struct roff_node *p)
 234 {
 235
 236         roff_node_unlink(mdoc, p);
 237         roff_node_append(mdoc, p);
 238 }
 239
 240 /*
 241  * Parse free-form text, that is, a line that does not begin with the
 242  * control character.
 243  */
 244 static int
 245 mdoc_ptext(struct roff_man *mdoc, int line, char *buf, int offs)
 246 {
 247         struct roff_node *n;
 248         char             *c, *ws, *end;
 249
 250         assert(mdoc->last);
 251         n = mdoc->last;
 252
 253         /*
 254          * Divert directly to list processing if we're encountering a
 255          * columnar ROFFT_BLOCK with or without a prior ROFFT_BLOCK entry
 256          * (a ROFFT_BODY means it's already open, in which case we should
 257          * process within its context in the normal way).
 258          */
 259
 260         if (n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
 261             n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
 262                 /* `Bl' is open without any children. */
 263                 mdoc->flags |= MDOC_FREECOL;
 264                 mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
 265                 return(1);
 266         }
 267
 268         if (n->tok == MDOC_It && n->type == ROFFT_BLOCK &&
 269             NULL != n->parent &&
 270             MDOC_Bl == n->parent->tok &&
 271             LIST_column == n->parent->norm->Bl.type) {
 272                 /* `Bl' has block-level `It' children. */
 273                 mdoc->flags |= MDOC_FREECOL;
 274                 mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
 275                 return(1);
 276         }
 277
 278         /*
 279          * Search for the beginning of unescaped trailing whitespace (ws)
 280          * and for the first character not to be output (end).
 281          */
 282
 283         /* FIXME: replace with strcspn(). */
 284         ws = NULL;
 285         for (c = end = buf + offs; *c; c++) {
 286                 switch (*c) {
 287                 case ' ':
 288                         if (NULL == ws)
 289                                 ws = c;
 290                         continue;
 291                 case '\t':
 292                         /*
 293                          * Always warn about trailing tabs,
 294                          * even outside literal context,
 295                          * where they should be put on the next line.
 296                          */
 297                         if (NULL == ws)
 298                                 ws = c;
 299                         /*
 300                          * Strip trailing tabs in literal context only;
 301                          * outside, they affect the next line.
 302                          */
 303                         if (MDOC_LITERAL & mdoc->flags)
 304                                 continue;
 305                         break;
 306                 case '\\':
 307                         /* Skip the escaped character, too, if any. */
 308                         if (c[1])
 309                                 c++;
 310                         /* FALLTHROUGH */
 311                 default:
 312                         ws = NULL;
 313                         break;
 314                 }
 315                 end = c + 1;
 316         }
 317         *end = '\0';
 318
 319         if (ws)
 320                 mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
 321                     line, (int)(ws-buf), NULL);
 322
 323         if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
 324                 mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
 325                     line, (int)(c - buf), NULL);
 326
 327                 /*
 328                  * Insert a `sp' in the case of a blank line.  Technically,
 329                  * blank lines aren't allowed, but enough manuals assume this
 330                  * behaviour that we want to work around it.
 331                  */
 332                 roff_elem_alloc(mdoc, line, offs, MDOC_sp);
 333                 mdoc->next = ROFF_NEXT_SIBLING;
 334                 mdoc_valid_post(mdoc);
 335                 return(1);
 336         }
 337
 338         roff_word_alloc(mdoc, line, offs, buf+offs);
 339
 340         if (mdoc->flags & MDOC_LITERAL)
 341                 return(1);
 342
 343         /*
 344          * End-of-sentence check.  If the last character is an unescaped
 345          * EOS character, then flag the node as being the end of a
 346          * sentence.  The front-end will know how to interpret this.
 347          */
 348
 349         assert(buf < end);
 350
 351         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
 352                 mdoc->last->flags |= MDOC_EOS;
 353         return(1);
 354 }
 355
 356 /*
 357  * Parse a macro line, that is, a line beginning with the control
 358  * character.
 359  */
 360 static int
 361 mdoc_pmacro(struct roff_man *mdoc, int ln, char *buf, int offs)
 362 {
 363         struct roff_node *n;
 364         const char       *cp;
 365         int               tok;
 366         int               i, sv;
 367         char              mac[5];
 368
 369         sv = offs;
 370
 371         /*
 372          * Copy the first word into a nil-terminated buffer.
 373          * Stop when a space, tab, escape, or eoln is encountered.
 374          */
 375
 376         i = 0;
 377         while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
 378                 mac[i++] = buf[offs++];
 379
 380         mac[i] = '\0';
 381
 382         tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : TOKEN_NONE;
 383
 384         if (tok == TOKEN_NONE) {
 385                 mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
 386                     ln, sv, buf + sv - 1);
 387                 return(1);
 388         }
 389
 390         /* Skip a leading escape sequence or tab. */
 391
 392         switch (buf[offs]) {
 393         case '\\':
 394                 cp = buf + offs + 1;
 395                 mandoc_escape(&cp, NULL, NULL);
 396                 offs = cp - buf;
 397                 break;
 398         case '\t':
 399                 offs++;
 400                 break;
 401         default:
 402                 break;
 403         }
 404
 405         /* Jump to the next non-whitespace word. */
 406
 407         while (buf[offs] && ' ' == buf[offs])
 408                 offs++;
 409
 410         /*
 411          * Trailing whitespace.  Note that tabs are allowed to be passed
 412          * into the parser as "text", so we only warn about spaces here.
 413          */
 414
 415         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 416                 mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
 417                     ln, offs - 1, NULL);
 418
 419         /*
 420          * If an initial macro or a list invocation, divert directly
 421          * into macro processing.
 422          */
 423
 424         if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
 425                 mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
 426                 return(1);
 427         }
 428
 429         n = mdoc->last;
 430         assert(mdoc->last);
 431
 432         /*
 433          * If the first macro of a `Bl -column', open an `It' block
 434          * context around the parsed macro.
 435          */
 436
 437         if (n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
 438             n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
 439                 mdoc->flags |= MDOC_FREECOL;
 440                 mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
 441                 return(1);
 442         }
 443
 444         /*
 445          * If we're following a block-level `It' within a `Bl -column'
 446          * context (perhaps opened in the above block or in ptext()),
 447          * then open an `It' block context around the parsed macro.
 448          */
 449
 450         if (n->tok == MDOC_It && n->type == ROFFT_BLOCK &&
 451             NULL != n->parent &&
 452             MDOC_Bl == n->parent->tok &&
 453             LIST_column == n->parent->norm->Bl.type) {
 454                 mdoc->flags |= MDOC_FREECOL;
 455                 mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
 456                 return(1);
 457         }
 458
 459         /* Normal processing of a macro. */
 460
 461         mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
 462
 463         /* In quick mode (for mandocdb), abort after the NAME section. */
 464
 465         if (mdoc->quick && MDOC_Sh == tok &&
 466             SEC_NAME != mdoc->last->sec)
 467                 return(2);
 468
 469         return(1);
 470 }
 471
 472 enum mdelim
 473 mdoc_isdelim(const char *p)
 474 {
 475
 476         if ('\0' == p[0])
 477                 return(DELIM_NONE);
 478
 479         if ('\0' == p[1])
 480                 switch (p[0]) {
 481                 case '(':
 482                         /* FALLTHROUGH */
 483                 case '[':
 484                         return(DELIM_OPEN);
 485                 case '|':
 486                         return(DELIM_MIDDLE);
 487                 case '.':
 488                         /* FALLTHROUGH */
 489                 case ',':
 490                         /* FALLTHROUGH */
 491                 case ';':
 492                         /* FALLTHROUGH */
 493                 case ':':
 494                         /* FALLTHROUGH */
 495                 case '?':
 496                         /* FALLTHROUGH */
 497                 case '!':
 498                         /* FALLTHROUGH */
 499                 case ')':
 500                         /* FALLTHROUGH */
 501                 case ']':
 502                         return(DELIM_CLOSE);
 503                 default:
 504                         return(DELIM_NONE);
 505                 }
 506
 507         if ('\\' != p[0])
 508                 return(DELIM_NONE);
 509
 510         if (0 == strcmp(p + 1, "."))
 511                 return(DELIM_CLOSE);
 512         if (0 == strcmp(p + 1, "fR|\\fP"))
 513                 return(DELIM_MIDDLE);
 514
 515         return(DELIM_NONE);
 516 }
 517
 518 void
 519 mdoc_deroff(char **dest, const struct roff_node *n)
 520 {
 521         char    *cp;
 522         size_t   sz;
 523
 524         if (n->type != ROFFT_TEXT) {
 525                 for (n = n->child; n; n = n->next)
 526                         mdoc_deroff(dest, n);
 527                 return;
 528         }
 529
 530         /* Skip leading whitespace. */
 531
 532         for (cp = n->string; '\0' != *cp; cp++)
 533                 if (0 == isspace((unsigned char)*cp))
 534                         break;
 535
 536         /* Skip trailing whitespace. */
 537
 538         for (sz = strlen(cp); sz; sz--)
 539                 if (0 == isspace((unsigned char)cp[sz-1]))
 540                         break;
 541
 542         /* Skip empty strings. */
 543
 544         if (0 == sz)
 545                 return;
 546
 547         if (NULL == *dest) {
 548                 *dest = mandoc_strndup(cp, sz);
 549                 return;
 550         }
 551
 552         mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
 553         free(*dest);
 554         *dest = cp;
 555 }