mdoc.c

   1 /*      $Id: mdoc.c,v 1.167 2010/12/01 16:38:57 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010 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 AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR 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 <stdarg.h>
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mandoc.h"
  32 #include "libmdoc.h"
  33 #include "libmandoc.h"
  34
  35 const   char *const __mdoc_macronames[MDOC_MAX] = {
  36         "Ap",           "Dd",           "Dt",           "Os",
  37         "Sh",           "Ss",           "Pp",           "D1",
  38         "Dl",           "Bd",           "Ed",           "Bl",
  39         "El",           "It",           "Ad",           "An",
  40         "Ar",           "Cd",           "Cm",           "Dv",
  41         "Er",           "Ev",           "Ex",           "Fa",
  42         "Fd",           "Fl",           "Fn",           "Ft",
  43         "Ic",           "In",           "Li",           "Nd",
  44         "Nm",           "Op",           "Ot",           "Pa",
  45         "Rv",           "St",           "Va",           "Vt",
  46         /* LINTED */
  47         "Xr",           "%A",           "%B",           "%D",
  48         /* LINTED */
  49         "%I",           "%J",           "%N",           "%O",
  50         /* LINTED */
  51         "%P",           "%R",           "%T",           "%V",
  52         "Ac",           "Ao",           "Aq",           "At",
  53         "Bc",           "Bf",           "Bo",           "Bq",
  54         "Bsx",          "Bx",           "Db",           "Dc",
  55         "Do",           "Dq",           "Ec",           "Ef",
  56         "Em",           "Eo",           "Fx",           "Ms",
  57         "No",           "Ns",           "Nx",           "Ox",
  58         "Pc",           "Pf",           "Po",           "Pq",
  59         "Qc",           "Ql",           "Qo",           "Qq",
  60         "Re",           "Rs",           "Sc",           "So",
  61         "Sq",           "Sm",           "Sx",           "Sy",
  62         "Tn",           "Ux",           "Xc",           "Xo",
  63         "Fo",           "Fc",           "Oo",           "Oc",
  64         "Bk",           "Ek",           "Bt",           "Hf",
  65         "Fr",           "Ud",           "Lb",           "Lp",
  66         "Lk",           "Mt",           "Brq",          "Bro",
  67         /* LINTED */
  68         "Brc",          "%C",           "Es",           "En",
  69         /* LINTED */
  70         "Dx",           "%Q",           "br",           "sp",
  71         /* LINTED */
  72         "%U",           "Ta"
  73         };
  74
  75 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  76         "split",                "nosplit",              "ragged",
  77         "unfilled",             "literal",              "file",
  78         "offset",               "bullet",               "dash",
  79         "hyphen",               "item",                 "enum",
  80         "tag",                  "diag",                 "hang",
  81         "ohang",                "inset",                "column",
  82         "width",                "compact",              "std",
  83         "filled",               "words",                "emphasis",
  84         "symbolic",             "nested",               "centered"
  85         };
  86
  87 const   char * const *mdoc_macronames = __mdoc_macronames;
  88 const   char * const *mdoc_argnames = __mdoc_argnames;
  89
  90 static  void              mdoc_node_free(struct mdoc_node *);
  91 static  void              mdoc_node_unlink(struct mdoc *,
  92                                 struct mdoc_node *);
  93 static  void              mdoc_free1(struct mdoc *);
  94 static  void              mdoc_alloc1(struct mdoc *);
  95 static  struct mdoc_node *node_alloc(struct mdoc *, int, int,
  96                                 enum mdoct, enum mdoc_type);
  97 static  int               node_append(struct mdoc *,
  98                                 struct mdoc_node *);
  99 static  int               mdoc_ptext(struct mdoc *, int, char *, int);
 100 static  int               mdoc_pmacro(struct mdoc *, int, char *, int);
 101
 102
 103 const struct mdoc_node *
 104 mdoc_node(const struct mdoc *m)
 105 {
 106
 107         return(MDOC_HALT & m->flags ? NULL : m->first);
 108 }
 109
 110
 111 const struct mdoc_meta *
 112 mdoc_meta(const struct mdoc *m)
 113 {
 114
 115         return(MDOC_HALT & m->flags ? NULL : &m->meta);
 116 }
 117
 118
 119 /*
 120  * Frees volatile resources (parse tree, meta-data, fields).
 121  */
 122 static void
 123 mdoc_free1(struct mdoc *mdoc)
 124 {
 125
 126         if (mdoc->first)
 127                 mdoc_node_delete(mdoc, mdoc->first);
 128         if (mdoc->meta.title)
 129                 free(mdoc->meta.title);
 130         if (mdoc->meta.os)
 131                 free(mdoc->meta.os);
 132         if (mdoc->meta.name)
 133                 free(mdoc->meta.name);
 134         if (mdoc->meta.arch)
 135                 free(mdoc->meta.arch);
 136         if (mdoc->meta.vol)
 137                 free(mdoc->meta.vol);
 138         if (mdoc->meta.msec)
 139                 free(mdoc->meta.msec);
 140 }
 141
 142
 143 /*
 144  * Allocate all volatile resources (parse tree, meta-data, fields).
 145  */
 146 static void
 147 mdoc_alloc1(struct mdoc *mdoc)
 148 {
 149
 150         memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 151         mdoc->flags = 0;
 152         mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
 153         mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 154         mdoc->first = mdoc->last;
 155         mdoc->last->type = MDOC_ROOT;
 156         mdoc->next = MDOC_NEXT_CHILD;
 157 }
 158
 159
 160 /*
 161  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 162  * data with mdoc_alloc1().  After invocation, parse data has been reset
 163  * and the parser is ready for re-invocation on a new tree; however,
 164  * cross-parse non-volatile data is kept intact.
 165  */
 166 void
 167 mdoc_reset(struct mdoc *mdoc)
 168 {
 169
 170         mdoc_free1(mdoc);
 171         mdoc_alloc1(mdoc);
 172 }
 173
 174
 175 /*
 176  * Completely free up all volatile and non-volatile parse resources.
 177  * After invocation, the pointer is no longer usable.
 178  */
 179 void
 180 mdoc_free(struct mdoc *mdoc)
 181 {
 182
 183         mdoc_free1(mdoc);
 184         free(mdoc);
 185 }
 186
 187
 188 /*
 189  * Allocate volatile and non-volatile parse resources.
 190  */
 191 struct mdoc *
 192 mdoc_alloc(struct regset *regs, void *data, mandocmsg msg)
 193 {
 194         struct mdoc     *p;
 195
 196         p = mandoc_calloc(1, sizeof(struct mdoc));
 197
 198         p->msg = msg;
 199         p->data = data;
 200         p->regs = regs;
 201
 202         mdoc_hash_init();
 203         mdoc_alloc1(p);
 204         return(p);
 205 }
 206
 207
 208 /*
 209  * Climb back up the parse tree, validating open scopes.  Mostly calls
 210  * through to macro_end() in macro.c.
 211  */
 212 int
 213 mdoc_endparse(struct mdoc *m)
 214 {
 215
 216         if (MDOC_HALT & m->flags)
 217                 return(0);
 218         else if (mdoc_macroend(m))
 219                 return(1);
 220         m->flags |= MDOC_HALT;
 221         return(0);
 222 }
 223
 224
 225 /*
 226  * Main parse routine.  Parses a single line -- really just hands off to
 227  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 228  */
 229 int
 230 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 231 {
 232
 233         if (MDOC_HALT & m->flags)
 234                 return(0);
 235
 236         m->flags |= MDOC_NEWLINE;
 237
 238         /*
 239          * Let the roff nS register switch SYNOPSIS mode early,
 240          * such that the parser knows at all times
 241          * whether this mode is on or off.
 242          * Note that this mode is also switched by the Sh macro.
 243          */
 244         if (m->regs->regs[(int)REG_nS].set) {
 245                 if (m->regs->regs[(int)REG_nS].v.u)
 246                         m->flags |= MDOC_SYNOPSIS;
 247                 else
 248                         m->flags &= ~MDOC_SYNOPSIS;
 249         }
 250
 251         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 252                         mdoc_pmacro(m, ln, buf, offs) :
 253                         mdoc_ptext(m, ln, buf, offs));
 254 }
 255
 256
 257 int
 258 mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
 259                 int ln, int pos, const char *fmt, ...)
 260 {
 261         char             buf[256];
 262         va_list          ap;
 263
 264         va_start(ap, fmt);
 265         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 266         va_end(ap);
 267
 268         return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
 269 }
 270
 271
 272 int
 273 mdoc_macro(MACRO_PROT_ARGS)
 274 {
 275         assert(tok < MDOC_MAX);
 276
 277         /* If we're in the body, deny prologue calls. */
 278
 279         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 280                         MDOC_PBODY & m->flags)
 281                 return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
 282
 283         /* If we're in the prologue, deny "body" macros.  */
 284
 285         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 286                         ! (MDOC_PBODY & m->flags)) {
 287                 if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
 288                         return(0);
 289                 if (NULL == m->meta.msec)
 290                         m->meta.msec = mandoc_strdup("1");
 291                 if (NULL == m->meta.title)
 292                         m->meta.title = mandoc_strdup("UNKNOWN");
 293                 if (NULL == m->meta.vol)
 294                         m->meta.vol = mandoc_strdup("LOCAL");
 295                 if (NULL == m->meta.os)
 296                         m->meta.os = mandoc_strdup("LOCAL");
 297                 if (0 == m->meta.date)
 298                         m->meta.date = time(NULL);
 299                 m->flags |= MDOC_PBODY;
 300         }
 301
 302         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 303 }
 304
 305
 306 static int
 307 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 308 {
 309
 310         assert(mdoc->last);
 311         assert(mdoc->first);
 312         assert(MDOC_ROOT != p->type);
 313
 314         switch (mdoc->next) {
 315         case (MDOC_NEXT_SIBLING):
 316                 mdoc->last->next = p;
 317                 p->prev = mdoc->last;
 318                 p->parent = mdoc->last->parent;
 319                 break;
 320         case (MDOC_NEXT_CHILD):
 321                 mdoc->last->child = p;
 322                 p->parent = mdoc->last;
 323                 break;
 324         default:
 325                 abort();
 326                 /* NOTREACHED */
 327         }
 328
 329         p->parent->nchild++;
 330
 331         if ( ! mdoc_valid_pre(mdoc, p))
 332                 return(0);
 333
 334         switch (p->type) {
 335         case (MDOC_HEAD):
 336                 assert(MDOC_BLOCK == p->parent->type);
 337                 p->parent->head = p;
 338                 break;
 339         case (MDOC_TAIL):
 340                 assert(MDOC_BLOCK == p->parent->type);
 341                 p->parent->tail = p;
 342                 break;
 343         case (MDOC_BODY):
 344                 if (p->end)
 345                         break;
 346                 assert(MDOC_BLOCK == p->parent->type);
 347                 p->parent->body = p;
 348                 break;
 349         default:
 350                 break;
 351         }
 352
 353         mdoc->last = p;
 354
 355         switch (p->type) {
 356         case (MDOC_TEXT):
 357                 if ( ! mdoc_valid_post(mdoc))
 358                         return(0);
 359                 break;
 360         default:
 361                 break;
 362         }
 363
 364         return(1);
 365 }
 366
 367
 368 static struct mdoc_node *
 369 node_alloc(struct mdoc *m, int line, int pos,
 370                 enum mdoct tok, enum mdoc_type type)
 371 {
 372         struct mdoc_node *p;
 373
 374         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 375         p->sec = m->lastsec;
 376         p->line = line;
 377         p->pos = pos;
 378         p->tok = tok;
 379         p->type = type;
 380
 381         /* Flag analysis. */
 382
 383         if (MDOC_SYNOPSIS & m->flags)
 384                 p->flags |= MDOC_SYNPRETTY;
 385         else
 386                 p->flags &= ~MDOC_SYNPRETTY;
 387         if (MDOC_NEWLINE & m->flags)
 388                 p->flags |= MDOC_LINE;
 389         m->flags &= ~MDOC_NEWLINE;
 390
 391         return(p);
 392 }
 393
 394
 395 int
 396 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 397 {
 398         struct mdoc_node *p;
 399
 400         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 401         if ( ! node_append(m, p))
 402                 return(0);
 403         m->next = MDOC_NEXT_CHILD;
 404         return(1);
 405 }
 406
 407
 408 int
 409 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 410 {
 411         struct mdoc_node *p;
 412
 413         assert(m->first);
 414         assert(m->last);
 415
 416         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 417         if ( ! node_append(m, p))
 418                 return(0);
 419         m->next = MDOC_NEXT_CHILD;
 420         return(1);
 421 }
 422
 423
 424 int
 425 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 426 {
 427         struct mdoc_node *p;
 428
 429         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 430         if ( ! node_append(m, p))
 431                 return(0);
 432         m->next = MDOC_NEXT_CHILD;
 433         return(1);
 434 }
 435
 436
 437 int
 438 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 439                 struct mdoc_node *body, enum mdoc_endbody end)
 440 {
 441         struct mdoc_node *p;
 442
 443         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 444         p->pending = body;
 445         p->end = end;
 446         if ( ! node_append(m, p))
 447                 return(0);
 448         m->next = MDOC_NEXT_SIBLING;
 449         return(1);
 450 }
 451
 452
 453 int
 454 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 455                 enum mdoct tok, struct mdoc_arg *args)
 456 {
 457         struct mdoc_node *p;
 458
 459         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 460         p->args = args;
 461         if (p->args)
 462                 (args->refcnt)++;
 463         if ( ! node_append(m, p))
 464                 return(0);
 465         m->next = MDOC_NEXT_CHILD;
 466         return(1);
 467 }
 468
 469
 470 int
 471 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 472                 enum mdoct tok, struct mdoc_arg *args)
 473 {
 474         struct mdoc_node *p;
 475
 476         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 477         p->args = args;
 478         if (p->args)
 479                 (args->refcnt)++;
 480         if ( ! node_append(m, p))
 481                 return(0);
 482         m->next = MDOC_NEXT_CHILD;
 483         return(1);
 484 }
 485
 486
 487 int
 488 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 489 {
 490         struct mdoc_node *n;
 491         size_t            sv, len;
 492
 493         len = strlen(p);
 494
 495         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 496         n->string = mandoc_malloc(len + 1);
 497         sv = strlcpy(n->string, p, len + 1);
 498
 499         /* Prohibit truncation. */
 500         assert(sv < len + 1);
 501
 502         if ( ! node_append(m, n))
 503                 return(0);
 504
 505         m->next = MDOC_NEXT_SIBLING;
 506         return(1);
 507 }
 508
 509
 510 static void
 511 mdoc_node_free(struct mdoc_node *p)
 512 {
 513
 514         /*
 515          * XXX: if these end up being problematic in terms of memory
 516          * management and dereferencing freed blocks, then make them
 517          * into reference-counted double-pointers.
 518          */
 519
 520         if (MDOC_Bd == p->tok && MDOC_BLOCK == p->type)
 521                 if (p->data.Bd)
 522                         free(p->data.Bd);
 523         if (MDOC_Bl == p->tok && MDOC_BLOCK == p->type)
 524                 if (p->data.Bl)
 525                         free(p->data.Bl);
 526         if (MDOC_Bf == p->tok && MDOC_HEAD == p->type)
 527                 if (p->data.Bf)
 528                         free(p->data.Bf);
 529
 530         if (p->string)
 531                 free(p->string);
 532         if (p->args)
 533                 mdoc_argv_free(p->args);
 534         free(p);
 535 }
 536
 537
 538 static void
 539 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 540 {
 541
 542         /* Adjust siblings. */
 543
 544         if (n->prev)
 545                 n->prev->next = n->next;
 546         if (n->next)
 547                 n->next->prev = n->prev;
 548
 549         /* Adjust parent. */
 550
 551         if (n->parent) {
 552                 n->parent->nchild--;
 553                 if (n->parent->child == n)
 554                         n->parent->child = n->prev ? n->prev : n->next;
 555         }
 556
 557         /* Adjust parse point, if applicable. */
 558
 559         if (m && m->last == n) {
 560                 if (n->prev) {
 561                         m->last = n->prev;
 562                         m->next = MDOC_NEXT_SIBLING;
 563                 } else {
 564                         m->last = n->parent;
 565                         m->next = MDOC_NEXT_CHILD;
 566                 }
 567         }
 568
 569         if (m && m->first == n)
 570                 m->first = NULL;
 571 }
 572
 573
 574 void
 575 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 576 {
 577
 578         while (p->child) {
 579                 assert(p->nchild);
 580                 mdoc_node_delete(m, p->child);
 581         }
 582         assert(0 == p->nchild);
 583
 584         mdoc_node_unlink(m, p);
 585         mdoc_node_free(p);
 586 }
 587
 588
 589 /*
 590  * Parse free-form text, that is, a line that does not begin with the
 591  * control character.
 592  */
 593 static int
 594 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 595 {
 596         char             *c, *ws, *end;
 597         struct mdoc_node *n;
 598
 599         /* Ignore bogus comments. */
 600
 601         if ('\\' == buf[offs] &&
 602                         '.' == buf[offs + 1] &&
 603                         '"' == buf[offs + 2])
 604                 return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
 605
 606         /* No text before an initial macro. */
 607
 608         if (SEC_NONE == m->lastnamed)
 609                 return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
 610
 611         assert(m->last);
 612         n = m->last;
 613
 614         /*
 615          * Divert directly to list processing if we're encountering a
 616          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 617          * (a MDOC_BODY means it's already open, in which case we should
 618          * process within its context in the normal way).
 619          */
 620
 621         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 622                         LIST_column == n->data.Bl->type) {
 623                 /* `Bl' is open without any children. */
 624                 m->flags |= MDOC_FREECOL;
 625                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 626         }
 627
 628         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 629                         NULL != n->parent &&
 630                         MDOC_Bl == n->parent->tok &&
 631                         LIST_column == n->parent->data.Bl->type) {
 632                 /* `Bl' has block-level `It' children. */
 633                 m->flags |= MDOC_FREECOL;
 634                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 635         }
 636
 637         /*
 638          * Search for the beginning of unescaped trailing whitespace (ws)
 639          * and for the first character not to be output (end).
 640          */
 641
 642         /* FIXME: replace with strcspn(). */
 643         ws = NULL;
 644         for (c = end = buf + offs; *c; c++) {
 645                 switch (*c) {
 646                 case '-':
 647                         if (mandoc_hyph(buf + offs, c))
 648                                 *c = ASCII_HYPH;
 649                         ws = NULL;
 650                         break;
 651                 case ' ':
 652                         if (NULL == ws)
 653                                 ws = c;
 654                         continue;
 655                 case '\t':
 656                         /*
 657                          * Always warn about trailing tabs,
 658                          * even outside literal context,
 659                          * where they should be put on the next line.
 660                          */
 661                         if (NULL == ws)
 662                                 ws = c;
 663                         /*
 664                          * Strip trailing tabs in literal context only;
 665                          * outside, they affect the next line.
 666                          */
 667                         if (MDOC_LITERAL & m->flags)
 668                                 continue;
 669                         break;
 670                 case '\\':
 671                         /* Skip the escaped character, too, if any. */
 672                         if (c[1])
 673                                 c++;
 674                         /* FALLTHROUGH */
 675                 default:
 676                         ws = NULL;
 677                         break;
 678                 }
 679                 end = c + 1;
 680         }
 681         *end = '\0';
 682
 683         if (ws)
 684                 if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
 685                         return(0);
 686
 687         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 688                 if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
 689                         return(0);
 690
 691                 /*
 692                  * Insert a `sp' in the case of a blank line.  Technically,
 693                  * blank lines aren't allowed, but enough manuals assume this
 694                  * behaviour that we want to work around it.
 695                  */
 696                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 697                         return(0);
 698
 699                 m->next = MDOC_NEXT_SIBLING;
 700                 return(1);
 701         }
 702
 703         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 704                 return(0);
 705
 706         if (MDOC_LITERAL & m->flags)
 707                 return(1);
 708
 709         /*
 710          * End-of-sentence check.  If the last character is an unescaped
 711          * EOS character, then flag the node as being the end of a
 712          * sentence.  The front-end will know how to interpret this.
 713          */
 714
 715         assert(buf < end);
 716
 717         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 718                 m->last->flags |= MDOC_EOS;
 719
 720         return(1);
 721 }
 722
 723
 724 /*
 725  * Parse a macro line, that is, a line beginning with the control
 726  * character.
 727  */
 728 static int
 729 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 730 {
 731         enum mdoct        tok;
 732         int               i, j, sv;
 733         char              mac[5];
 734         struct mdoc_node *n;
 735
 736         /* Empty lines are ignored. */
 737
 738         offs++;
 739
 740         if ('\0' == buf[offs])
 741                 return(1);
 742
 743         i = offs;
 744
 745         /* Accept tabs/whitespace after the initial control char. */
 746
 747         if (' ' == buf[i] || '\t' == buf[i]) {
 748                 i++;
 749                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 750                         i++;
 751                 if ('\0' == buf[i])
 752                         return(1);
 753         }
 754
 755         sv = i;
 756
 757         /*
 758          * Copy the first word into a nil-terminated buffer.
 759          * Stop copying when a tab, space, or eoln is encountered.
 760          */
 761
 762         j = 0;
 763         while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
 764                 mac[j++] = buf[i++];
 765         mac[j] = '\0';
 766
 767         tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 768         if (MDOC_MAX == tok) {
 769                 mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv,
 770                     "unknown macro: %s%s",
 771                     buf, strlen(buf) > 3 ? "..." : "");
 772                 return(1);
 773         }
 774
 775         /* Disregard the first trailing tab, if applicable. */
 776
 777         if ('\t' == buf[i])
 778                 i++;
 779
 780         /* Jump to the next non-whitespace word. */
 781
 782         while (buf[i] && ' ' == buf[i])
 783                 i++;
 784
 785         /*
 786          * Trailing whitespace.  Note that tabs are allowed to be passed
 787          * into the parser as "text", so we only warn about spaces here.
 788          */
 789
 790         if ('\0' == buf[i] && ' ' == buf[i - 1])
 791                 if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
 792                         goto err;
 793
 794         /*
 795          * If an initial macro or a list invocation, divert directly
 796          * into macro processing.
 797          */
 798
 799         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 800                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 801                         goto err;
 802                 return(1);
 803         }
 804
 805         n = m->last;
 806         assert(m->last);
 807
 808         /*
 809          * If the first macro of a `Bl -column', open an `It' block
 810          * context around the parsed macro.
 811          */
 812
 813         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 814                         LIST_column == n->data.Bl->type) {
 815                 m->flags |= MDOC_FREECOL;
 816                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 817                         goto err;
 818                 return(1);
 819         }
 820
 821         /*
 822          * If we're following a block-level `It' within a `Bl -column'
 823          * context (perhaps opened in the above block or in ptext()),
 824          * then open an `It' block context around the parsed macro.
 825          */
 826
 827         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 828                         NULL != n->parent &&
 829                         MDOC_Bl == n->parent->tok &&
 830                         LIST_column == n->parent->data.Bl->type) {
 831                 m->flags |= MDOC_FREECOL;
 832                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 833                         goto err;
 834                 return(1);
 835         }
 836
 837         /* Normal processing of a macro. */
 838
 839         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 840                 goto err;
 841
 842         return(1);
 843
 844 err:    /* Error out. */
 845
 846         m->flags |= MDOC_HALT;
 847         return(0);
 848 }
 849
 850