mdoc.c

   1 /*      $Id: mdoc.c,v 1.253 2015/10/06 18:32:19 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 /*
  93  * Main parse routine.  Parses a single line -- really just hands off to
  94  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
  95  */
  96 int
  97 mdoc_parseln(struct roff_man *mdoc, int ln, char *buf, int offs)
  98 {
  99
 100         if (mdoc->last->type != ROFFT_EQN || ln > mdoc->last->line)
 101                 mdoc->flags |= MDOC_NEWLINE;
 102
 103         /*
 104          * Let the roff nS register switch SYNOPSIS mode early,
 105          * such that the parser knows at all times
 106          * whether this mode is on or off.
 107          * Note that this mode is also switched by the Sh macro.
 108          */
 109         if (roff_getreg(mdoc->roff, "nS"))
 110                 mdoc->flags |= MDOC_SYNOPSIS;
 111         else
 112                 mdoc->flags &= ~MDOC_SYNOPSIS;
 113
 114         return roff_getcontrol(mdoc->roff, buf, &offs) ?
 115             mdoc_pmacro(mdoc, ln, buf, offs) :
 116             mdoc_ptext(mdoc, ln, buf, offs);
 117 }
 118
 119 void
 120 mdoc_macro(MACRO_PROT_ARGS)
 121 {
 122         assert(tok > TOKEN_NONE && tok < MDOC_MAX);
 123
 124         if (mdoc->flags & MDOC_PBODY) {
 125                 if (tok == MDOC_Dt) {
 126                         mandoc_vmsg(MANDOCERR_DT_LATE,
 127                             mdoc->parse, line, ppos,
 128                             "Dt %s", buf + *pos);
 129                         return;
 130                 }
 131         } else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) {
 132                 if (mdoc->meta.title == NULL) {
 133                         mandoc_vmsg(MANDOCERR_DT_NOTITLE,
 134                             mdoc->parse, line, ppos, "%s %s",
 135                             mdoc_macronames[tok], buf + *pos);
 136                         mdoc->meta.title = mandoc_strdup("UNTITLED");
 137                 }
 138                 if (NULL == mdoc->meta.vol)
 139                         mdoc->meta.vol = mandoc_strdup("LOCAL");
 140                 mdoc->flags |= MDOC_PBODY;
 141         }
 142         (*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf);
 143 }
 144
 145 void
 146 mdoc_tail_alloc(struct roff_man *mdoc, int line, int pos, int tok)
 147 {
 148         struct roff_node *p;
 149
 150         p = roff_node_alloc(mdoc, line, pos, ROFFT_TAIL, tok);
 151         roff_node_append(mdoc, p);
 152         mdoc->next = ROFF_NEXT_CHILD;
 153 }
 154
 155 struct roff_node *
 156 mdoc_endbody_alloc(struct roff_man *mdoc, int line, int pos, int tok,
 157                 struct roff_node *body, enum mdoc_endbody end)
 158 {
 159         struct roff_node *p;
 160
 161         body->flags |= MDOC_ENDED;
 162         body->parent->flags |= MDOC_ENDED;
 163         p = roff_node_alloc(mdoc, line, pos, ROFFT_BODY, tok);
 164         p->body = body;
 165         p->norm = body->norm;
 166         p->end = end;
 167         roff_node_append(mdoc, p);
 168         mdoc->next = ROFF_NEXT_SIBLING;
 169         return p;
 170 }
 171
 172 struct roff_node *
 173 mdoc_block_alloc(struct roff_man *mdoc, int line, int pos,
 174         int tok, struct mdoc_arg *args)
 175 {
 176         struct roff_node *p;
 177
 178         p = roff_node_alloc(mdoc, line, pos, ROFFT_BLOCK, tok);
 179         p->args = args;
 180         if (p->args)
 181                 (args->refcnt)++;
 182
 183         switch (tok) {
 184         case MDOC_Bd:
 185                 /* FALLTHROUGH */
 186         case MDOC_Bf:
 187                 /* FALLTHROUGH */
 188         case MDOC_Bl:
 189                 /* FALLTHROUGH */
 190         case MDOC_En:
 191                 /* FALLTHROUGH */
 192         case MDOC_Rs:
 193                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 194                 break;
 195         default:
 196                 break;
 197         }
 198         roff_node_append(mdoc, p);
 199         mdoc->next = ROFF_NEXT_CHILD;
 200         return p;
 201 }
 202
 203 void
 204 mdoc_elem_alloc(struct roff_man *mdoc, int line, int pos,
 205         int tok, struct mdoc_arg *args)
 206 {
 207         struct roff_node *p;
 208
 209         p = roff_node_alloc(mdoc, line, pos, ROFFT_ELEM, tok);
 210         p->args = args;
 211         if (p->args)
 212                 (args->refcnt)++;
 213
 214         switch (tok) {
 215         case MDOC_An:
 216                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 217                 break;
 218         default:
 219                 break;
 220         }
 221         roff_node_append(mdoc, p);
 222         mdoc->next = ROFF_NEXT_CHILD;
 223 }
 224
 225 void
 226 mdoc_node_relink(struct roff_man *mdoc, struct roff_node *p)
 227 {
 228
 229         roff_node_unlink(mdoc, p);
 230         roff_node_append(mdoc, p);
 231 }
 232
 233 /*
 234  * Parse free-form text, that is, a line that does not begin with the
 235  * control character.
 236  */
 237 static int
 238 mdoc_ptext(struct roff_man *mdoc, int line, char *buf, int offs)
 239 {
 240         struct roff_node *n;
 241         char             *c, *ws, *end;
 242
 243         assert(mdoc->last);
 244         n = mdoc->last;
 245
 246         /*
 247          * Divert directly to list processing if we're encountering a
 248          * columnar ROFFT_BLOCK with or without a prior ROFFT_BLOCK entry
 249          * (a ROFFT_BODY means it's already open, in which case we should
 250          * process within its context in the normal way).
 251          */
 252
 253         if (n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
 254             n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
 255                 /* `Bl' is open without any children. */
 256                 mdoc->flags |= MDOC_FREECOL;
 257                 mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
 258                 return 1;
 259         }
 260
 261         if (n->tok == MDOC_It && n->type == ROFFT_BLOCK &&
 262             NULL != n->parent &&
 263             MDOC_Bl == n->parent->tok &&
 264             LIST_column == n->parent->norm->Bl.type) {
 265                 /* `Bl' has block-level `It' children. */
 266                 mdoc->flags |= MDOC_FREECOL;
 267                 mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
 268                 return 1;
 269         }
 270
 271         /*
 272          * Search for the beginning of unescaped trailing whitespace (ws)
 273          * and for the first character not to be output (end).
 274          */
 275
 276         /* FIXME: replace with strcspn(). */
 277         ws = NULL;
 278         for (c = end = buf + offs; *c; c++) {
 279                 switch (*c) {
 280                 case ' ':
 281                         if (NULL == ws)
 282                                 ws = c;
 283                         continue;
 284                 case '\t':
 285                         /*
 286                          * Always warn about trailing tabs,
 287                          * even outside literal context,
 288                          * where they should be put on the next line.
 289                          */
 290                         if (NULL == ws)
 291                                 ws = c;
 292                         /*
 293                          * Strip trailing tabs in literal context only;
 294                          * outside, they affect the next line.
 295                          */
 296                         if (MDOC_LITERAL & mdoc->flags)
 297                                 continue;
 298                         break;
 299                 case '\\':
 300                         /* Skip the escaped character, too, if any. */
 301                         if (c[1])
 302                                 c++;
 303                         /* FALLTHROUGH */
 304                 default:
 305                         ws = NULL;
 306                         break;
 307                 }
 308                 end = c + 1;
 309         }
 310         *end = '\0';
 311
 312         if (ws)
 313                 mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
 314                     line, (int)(ws-buf), NULL);
 315
 316         if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
 317                 mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
 318                     line, (int)(c - buf), NULL);
 319
 320                 /*
 321                  * Insert a `sp' in the case of a blank line.  Technically,
 322                  * blank lines aren't allowed, but enough manuals assume this
 323                  * behaviour that we want to work around it.
 324                  */
 325                 roff_elem_alloc(mdoc, line, offs, MDOC_sp);
 326                 mdoc->next = ROFF_NEXT_SIBLING;
 327                 mdoc_valid_post(mdoc);
 328                 return 1;
 329         }
 330
 331         roff_word_alloc(mdoc, line, offs, buf+offs);
 332
 333         if (mdoc->flags & MDOC_LITERAL)
 334                 return 1;
 335
 336         /*
 337          * End-of-sentence check.  If the last character is an unescaped
 338          * EOS character, then flag the node as being the end of a
 339          * sentence.  The front-end will know how to interpret this.
 340          */
 341
 342         assert(buf < end);
 343
 344         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
 345                 mdoc->last->flags |= MDOC_EOS;
 346         return 1;
 347 }
 348
 349 /*
 350  * Parse a macro line, that is, a line beginning with the control
 351  * character.
 352  */
 353 static int
 354 mdoc_pmacro(struct roff_man *mdoc, int ln, char *buf, int offs)
 355 {
 356         struct roff_node *n;
 357         const char       *cp;
 358         int               tok;
 359         int               i, sv;
 360         char              mac[5];
 361
 362         sv = offs;
 363
 364         /*
 365          * Copy the first word into a nil-terminated buffer.
 366          * Stop when a space, tab, escape, or eoln is encountered.
 367          */
 368
 369         i = 0;
 370         while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
 371                 mac[i++] = buf[offs++];
 372
 373         mac[i] = '\0';
 374
 375         tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : TOKEN_NONE;
 376
 377         if (tok == TOKEN_NONE) {
 378                 mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
 379                     ln, sv, buf + sv - 1);
 380                 return 1;
 381         }
 382
 383         /* Skip a leading escape sequence or tab. */
 384
 385         switch (buf[offs]) {
 386         case '\\':
 387                 cp = buf + offs + 1;
 388                 mandoc_escape(&cp, NULL, NULL);
 389                 offs = cp - buf;
 390                 break;
 391         case '\t':
 392                 offs++;
 393                 break;
 394         default:
 395                 break;
 396         }
 397
 398         /* Jump to the next non-whitespace word. */
 399
 400         while (buf[offs] && ' ' == buf[offs])
 401                 offs++;
 402
 403         /*
 404          * Trailing whitespace.  Note that tabs are allowed to be passed
 405          * into the parser as "text", so we only warn about spaces here.
 406          */
 407
 408         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 409                 mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
 410                     ln, offs - 1, NULL);
 411
 412         /*
 413          * If an initial macro or a list invocation, divert directly
 414          * into macro processing.
 415          */
 416
 417         if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) {
 418                 mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
 419                 return 1;
 420         }
 421
 422         n = mdoc->last;
 423         assert(mdoc->last);
 424
 425         /*
 426          * If the first macro of a `Bl -column', open an `It' block
 427          * context around the parsed macro.
 428          */
 429
 430         if (n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
 431             n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) {
 432                 mdoc->flags |= MDOC_FREECOL;
 433                 mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
 434                 return 1;
 435         }
 436
 437         /*
 438          * If we're following a block-level `It' within a `Bl -column'
 439          * context (perhaps opened in the above block or in ptext()),
 440          * then open an `It' block context around the parsed macro.
 441          */
 442
 443         if (n->tok == MDOC_It && n->type == ROFFT_BLOCK &&
 444             NULL != n->parent &&
 445             MDOC_Bl == n->parent->tok &&
 446             LIST_column == n->parent->norm->Bl.type) {
 447                 mdoc->flags |= MDOC_FREECOL;
 448                 mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
 449                 return 1;
 450         }
 451
 452         /* Normal processing of a macro. */
 453
 454         mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
 455
 456         /* In quick mode (for mandocdb), abort after the NAME section. */
 457
 458         if (mdoc->quick && MDOC_Sh == tok &&
 459             SEC_NAME != mdoc->last->sec)
 460                 return 2;
 461
 462         return 1;
 463 }
 464
 465 enum mdelim
 466 mdoc_isdelim(const char *p)
 467 {
 468
 469         if ('\0' == p[0])
 470                 return DELIM_NONE;
 471
 472         if ('\0' == p[1])
 473                 switch (p[0]) {
 474                 case '(':
 475                         /* FALLTHROUGH */
 476                 case '[':
 477                         return DELIM_OPEN;
 478                 case '|':
 479                         return DELIM_MIDDLE;
 480                 case '.':
 481                         /* FALLTHROUGH */
 482                 case ',':
 483                         /* FALLTHROUGH */
 484                 case ';':
 485                         /* FALLTHROUGH */
 486                 case ':':
 487                         /* FALLTHROUGH */
 488                 case '?':
 489                         /* FALLTHROUGH */
 490                 case '!':
 491                         /* FALLTHROUGH */
 492                 case ')':
 493                         /* FALLTHROUGH */
 494                 case ']':
 495                         return DELIM_CLOSE;
 496                 default:
 497                         return DELIM_NONE;
 498                 }
 499
 500         if ('\\' != p[0])
 501                 return DELIM_NONE;
 502
 503         if (0 == strcmp(p + 1, "."))
 504                 return DELIM_CLOSE;
 505         if (0 == strcmp(p + 1, "fR|\\fP"))
 506                 return DELIM_MIDDLE;
 507
 508         return DELIM_NONE;
 509 }