mdoc.c

   1 /*      $Id: mdoc.c,v 1.179 2011/02/06 22:05:20 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 static  int               mdoc_span_alloc(struct mdoc *,
 102                                 const struct tbl_span *);
 103
 104
 105 const struct mdoc_node *
 106 mdoc_node(const struct mdoc *m)
 107 {
 108
 109         assert( ! (MDOC_HALT & m->flags));
 110         return(m->first);
 111 }
 112
 113
 114 const struct mdoc_meta *
 115 mdoc_meta(const struct mdoc *m)
 116 {
 117
 118         assert( ! (MDOC_HALT & m->flags));
 119         return(&m->meta);
 120 }
 121
 122
 123 /*
 124  * Frees volatile resources (parse tree, meta-data, fields).
 125  */
 126 static void
 127 mdoc_free1(struct mdoc *mdoc)
 128 {
 129
 130         if (mdoc->first)
 131                 mdoc_node_delete(mdoc, mdoc->first);
 132         if (mdoc->meta.title)
 133                 free(mdoc->meta.title);
 134         if (mdoc->meta.os)
 135                 free(mdoc->meta.os);
 136         if (mdoc->meta.name)
 137                 free(mdoc->meta.name);
 138         if (mdoc->meta.arch)
 139                 free(mdoc->meta.arch);
 140         if (mdoc->meta.vol)
 141                 free(mdoc->meta.vol);
 142         if (mdoc->meta.msec)
 143                 free(mdoc->meta.msec);
 144 }
 145
 146
 147 /*
 148  * Allocate all volatile resources (parse tree, meta-data, fields).
 149  */
 150 static void
 151 mdoc_alloc1(struct mdoc *mdoc)
 152 {
 153
 154         memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 155         mdoc->flags = 0;
 156         mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
 157         mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 158         mdoc->first = mdoc->last;
 159         mdoc->last->type = MDOC_ROOT;
 160         mdoc->next = MDOC_NEXT_CHILD;
 161 }
 162
 163
 164 /*
 165  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 166  * data with mdoc_alloc1().  After invocation, parse data has been reset
 167  * and the parser is ready for re-invocation on a new tree; however,
 168  * cross-parse non-volatile data is kept intact.
 169  */
 170 void
 171 mdoc_reset(struct mdoc *mdoc)
 172 {
 173
 174         mdoc_free1(mdoc);
 175         mdoc_alloc1(mdoc);
 176 }
 177
 178
 179 /*
 180  * Completely free up all volatile and non-volatile parse resources.
 181  * After invocation, the pointer is no longer usable.
 182  */
 183 void
 184 mdoc_free(struct mdoc *mdoc)
 185 {
 186
 187         mdoc_free1(mdoc);
 188         free(mdoc);
 189 }
 190
 191
 192 /*
 193  * Allocate volatile and non-volatile parse resources.
 194  */
 195 struct mdoc *
 196 mdoc_alloc(struct regset *regs, void *data, mandocmsg msg)
 197 {
 198         struct mdoc     *p;
 199
 200         p = mandoc_calloc(1, sizeof(struct mdoc));
 201
 202         p->msg = msg;
 203         p->data = data;
 204         p->regs = regs;
 205
 206         mdoc_hash_init();
 207         mdoc_alloc1(p);
 208         return(p);
 209 }
 210
 211
 212 /*
 213  * Climb back up the parse tree, validating open scopes.  Mostly calls
 214  * through to macro_end() in macro.c.
 215  */
 216 int
 217 mdoc_endparse(struct mdoc *m)
 218 {
 219
 220         assert( ! (MDOC_HALT & m->flags));
 221         if (mdoc_macroend(m))
 222                 return(1);
 223         m->flags |= MDOC_HALT;
 224         return(0);
 225 }
 226
 227 int
 228 mdoc_addspan(struct mdoc *m, const struct tbl_span *sp)
 229 {
 230
 231         assert( ! (MDOC_HALT & m->flags));
 232
 233         /* No text before an initial macro. */
 234
 235         if (SEC_NONE == m->lastnamed) {
 236                 mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT);
 237                 return(1);
 238         }
 239
 240         return(mdoc_span_alloc(m, sp));
 241 }
 242
 243
 244 /*
 245  * Main parse routine.  Parses a single line -- really just hands off to
 246  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 247  */
 248 int
 249 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 250 {
 251
 252         assert( ! (MDOC_HALT & m->flags));
 253
 254         m->flags |= MDOC_NEWLINE;
 255
 256         /*
 257          * Let the roff nS register switch SYNOPSIS mode early,
 258          * such that the parser knows at all times
 259          * whether this mode is on or off.
 260          * Note that this mode is also switched by the Sh macro.
 261          */
 262         if (m->regs->regs[(int)REG_nS].set) {
 263                 if (m->regs->regs[(int)REG_nS].v.u)
 264                         m->flags |= MDOC_SYNOPSIS;
 265                 else
 266                         m->flags &= ~MDOC_SYNOPSIS;
 267         }
 268
 269         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 270                         mdoc_pmacro(m, ln, buf, offs) :
 271                         mdoc_ptext(m, ln, buf, offs));
 272 }
 273
 274
 275 int
 276 mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
 277                 int ln, int pos, const char *fmt, ...)
 278 {
 279         char             buf[256];
 280         va_list          ap;
 281
 282         va_start(ap, fmt);
 283         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 284         va_end(ap);
 285
 286         return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
 287 }
 288
 289
 290 int
 291 mdoc_macro(MACRO_PROT_ARGS)
 292 {
 293         assert(tok < MDOC_MAX);
 294
 295         /* If we're in the body, deny prologue calls. */
 296
 297         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 298                         MDOC_PBODY & m->flags) {
 299                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY);
 300                 return(1);
 301         }
 302
 303         /* If we're in the prologue, deny "body" macros.  */
 304
 305         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 306                         ! (MDOC_PBODY & m->flags)) {
 307                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG);
 308                 if (NULL == m->meta.msec)
 309                         m->meta.msec = mandoc_strdup("1");
 310                 if (NULL == m->meta.title)
 311                         m->meta.title = mandoc_strdup("UNKNOWN");
 312                 if (NULL == m->meta.vol)
 313                         m->meta.vol = mandoc_strdup("LOCAL");
 314                 if (NULL == m->meta.os)
 315                         m->meta.os = mandoc_strdup("LOCAL");
 316                 if (0 == m->meta.date)
 317                         m->meta.date = time(NULL);
 318                 m->flags |= MDOC_PBODY;
 319         }
 320
 321         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 322 }
 323
 324
 325 static int
 326 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 327 {
 328
 329         assert(mdoc->last);
 330         assert(mdoc->first);
 331         assert(MDOC_ROOT != p->type);
 332
 333         switch (mdoc->next) {
 334         case (MDOC_NEXT_SIBLING):
 335                 mdoc->last->next = p;
 336                 p->prev = mdoc->last;
 337                 p->parent = mdoc->last->parent;
 338                 break;
 339         case (MDOC_NEXT_CHILD):
 340                 mdoc->last->child = p;
 341                 p->parent = mdoc->last;
 342                 break;
 343         default:
 344                 abort();
 345                 /* NOTREACHED */
 346         }
 347
 348         p->parent->nchild++;
 349
 350         /*
 351          * Copy over the normalised-data pointer of our parent.  Not
 352          * everybody has one, but copying a null pointer is fine.
 353          */
 354
 355         switch (p->type) {
 356         case (MDOC_BODY):
 357                 /* FALLTHROUGH */
 358         case (MDOC_TAIL):
 359                 /* FALLTHROUGH */
 360         case (MDOC_HEAD):
 361                 p->norm = p->parent->norm;
 362                 break;
 363         default:
 364                 break;
 365         }
 366
 367         if ( ! mdoc_valid_pre(mdoc, p))
 368                 return(0);
 369
 370         switch (p->type) {
 371         case (MDOC_HEAD):
 372                 assert(MDOC_BLOCK == p->parent->type);
 373                 p->parent->head = p;
 374                 break;
 375         case (MDOC_TAIL):
 376                 assert(MDOC_BLOCK == p->parent->type);
 377                 p->parent->tail = p;
 378                 break;
 379         case (MDOC_BODY):
 380                 if (p->end)
 381                         break;
 382                 assert(MDOC_BLOCK == p->parent->type);
 383                 p->parent->body = p;
 384                 break;
 385         default:
 386                 break;
 387         }
 388
 389         mdoc->last = p;
 390
 391         switch (p->type) {
 392         case (MDOC_TBL):
 393                 /* FALLTHROUGH */
 394         case (MDOC_TEXT):
 395                 if ( ! mdoc_valid_post(mdoc))
 396                         return(0);
 397                 break;
 398         default:
 399                 break;
 400         }
 401
 402         return(1);
 403 }
 404
 405
 406 static struct mdoc_node *
 407 node_alloc(struct mdoc *m, int line, int pos,
 408                 enum mdoct tok, enum mdoc_type type)
 409 {
 410         struct mdoc_node *p;
 411
 412         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 413         p->sec = m->lastsec;
 414         p->line = line;
 415         p->pos = pos;
 416         p->tok = tok;
 417         p->type = type;
 418
 419         /* Flag analysis. */
 420
 421         if (MDOC_SYNOPSIS & m->flags)
 422                 p->flags |= MDOC_SYNPRETTY;
 423         else
 424                 p->flags &= ~MDOC_SYNPRETTY;
 425         if (MDOC_NEWLINE & m->flags)
 426                 p->flags |= MDOC_LINE;
 427         m->flags &= ~MDOC_NEWLINE;
 428
 429         return(p);
 430 }
 431
 432
 433 int
 434 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 435 {
 436         struct mdoc_node *p;
 437
 438         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 439         if ( ! node_append(m, p))
 440                 return(0);
 441         m->next = MDOC_NEXT_CHILD;
 442         return(1);
 443 }
 444
 445
 446 int
 447 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 448 {
 449         struct mdoc_node *p;
 450
 451         assert(m->first);
 452         assert(m->last);
 453
 454         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 455         if ( ! node_append(m, p))
 456                 return(0);
 457         m->next = MDOC_NEXT_CHILD;
 458         return(1);
 459 }
 460
 461
 462 int
 463 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 464 {
 465         struct mdoc_node *p;
 466
 467         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 468         if ( ! node_append(m, p))
 469                 return(0);
 470         m->next = MDOC_NEXT_CHILD;
 471         return(1);
 472 }
 473
 474
 475 int
 476 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 477                 struct mdoc_node *body, enum mdoc_endbody end)
 478 {
 479         struct mdoc_node *p;
 480
 481         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 482         p->pending = body;
 483         p->end = end;
 484         if ( ! node_append(m, p))
 485                 return(0);
 486         m->next = MDOC_NEXT_SIBLING;
 487         return(1);
 488 }
 489
 490
 491 int
 492 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 493                 enum mdoct tok, struct mdoc_arg *args)
 494 {
 495         struct mdoc_node *p;
 496
 497         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 498         p->args = args;
 499         if (p->args)
 500                 (args->refcnt)++;
 501
 502         switch (tok) {
 503         case (MDOC_Bd):
 504                 /* FALLTHROUGH */
 505         case (MDOC_Bf):
 506                 /* FALLTHROUGH */
 507         case (MDOC_Bl):
 508                 /* FALLTHROUGH */
 509         case (MDOC_Rs):
 510                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 511                 break;
 512         default:
 513                 break;
 514         }
 515
 516         if ( ! node_append(m, p))
 517                 return(0);
 518         m->next = MDOC_NEXT_CHILD;
 519         return(1);
 520 }
 521
 522
 523 int
 524 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 525                 enum mdoct tok, struct mdoc_arg *args)
 526 {
 527         struct mdoc_node *p;
 528
 529         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 530         p->args = args;
 531         if (p->args)
 532                 (args->refcnt)++;
 533
 534         switch (tok) {
 535         case (MDOC_An):
 536                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 537                 break;
 538         default:
 539                 break;
 540         }
 541
 542         if ( ! node_append(m, p))
 543                 return(0);
 544         m->next = MDOC_NEXT_CHILD;
 545         return(1);
 546 }
 547
 548 static int
 549 mdoc_span_alloc(struct mdoc *m, const struct tbl_span *sp)
 550 {
 551         struct mdoc_node *n;
 552
 553         n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL);
 554         n->span = sp;
 555
 556         if ( ! node_append(m, n))
 557                 return(0);
 558
 559         m->next = MDOC_NEXT_SIBLING;
 560         return(1);
 561 }
 562
 563
 564 int
 565 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 566 {
 567         struct mdoc_node *n;
 568         size_t            sv, len;
 569
 570         len = strlen(p);
 571
 572         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 573         n->string = mandoc_malloc(len + 1);
 574         sv = strlcpy(n->string, p, len + 1);
 575
 576         /* Prohibit truncation. */
 577         assert(sv < len + 1);
 578
 579         if ( ! node_append(m, n))
 580                 return(0);
 581
 582         m->next = MDOC_NEXT_SIBLING;
 583         return(1);
 584 }
 585
 586
 587 static void
 588 mdoc_node_free(struct mdoc_node *p)
 589 {
 590
 591         if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
 592                 free(p->norm);
 593         if (p->string)
 594                 free(p->string);
 595         if (p->args)
 596                 mdoc_argv_free(p->args);
 597         free(p);
 598 }
 599
 600
 601 static void
 602 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 603 {
 604
 605         /* Adjust siblings. */
 606
 607         if (n->prev)
 608                 n->prev->next = n->next;
 609         if (n->next)
 610                 n->next->prev = n->prev;
 611
 612         /* Adjust parent. */
 613
 614         if (n->parent) {
 615                 n->parent->nchild--;
 616                 if (n->parent->child == n)
 617                         n->parent->child = n->prev ? n->prev : n->next;
 618                 if (n->parent->last == n)
 619                         n->parent->last = n->prev ? n->prev : NULL;
 620         }
 621
 622         /* Adjust parse point, if applicable. */
 623
 624         if (m && m->last == n) {
 625                 if (n->prev) {
 626                         m->last = n->prev;
 627                         m->next = MDOC_NEXT_SIBLING;
 628                 } else {
 629                         m->last = n->parent;
 630                         m->next = MDOC_NEXT_CHILD;
 631                 }
 632         }
 633
 634         if (m && m->first == n)
 635                 m->first = NULL;
 636 }
 637
 638
 639 void
 640 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 641 {
 642
 643         while (p->child) {
 644                 assert(p->nchild);
 645                 mdoc_node_delete(m, p->child);
 646         }
 647         assert(0 == p->nchild);
 648
 649         mdoc_node_unlink(m, p);
 650         mdoc_node_free(p);
 651 }
 652
 653
 654 /*
 655  * Parse free-form text, that is, a line that does not begin with the
 656  * control character.
 657  */
 658 static int
 659 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 660 {
 661         char             *c, *ws, *end;
 662         struct mdoc_node *n;
 663
 664         /* Ignore bogus comments. */
 665
 666         if ('\\' == buf[offs] &&
 667                         '.' == buf[offs + 1] &&
 668                         '"' == buf[offs + 2]) {
 669                 mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT);
 670                 return(1);
 671         }
 672
 673         /* No text before an initial macro. */
 674
 675         if (SEC_NONE == m->lastnamed) {
 676                 mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
 677                 return(1);
 678         }
 679
 680         assert(m->last);
 681         n = m->last;
 682
 683         /*
 684          * Divert directly to list processing if we're encountering a
 685          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 686          * (a MDOC_BODY means it's already open, in which case we should
 687          * process within its context in the normal way).
 688          */
 689
 690         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 691                         LIST_column == n->norm->Bl.type) {
 692                 /* `Bl' is open without any children. */
 693                 m->flags |= MDOC_FREECOL;
 694                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 695         }
 696
 697         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 698                         NULL != n->parent &&
 699                         MDOC_Bl == n->parent->tok &&
 700                         LIST_column == n->parent->norm->Bl.type) {
 701                 /* `Bl' has block-level `It' children. */
 702                 m->flags |= MDOC_FREECOL;
 703                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 704         }
 705
 706         /*
 707          * Search for the beginning of unescaped trailing whitespace (ws)
 708          * and for the first character not to be output (end).
 709          */
 710
 711         /* FIXME: replace with strcspn(). */
 712         ws = NULL;
 713         for (c = end = buf + offs; *c; c++) {
 714                 switch (*c) {
 715                 case '-':
 716                         if (mandoc_hyph(buf + offs, c))
 717                                 *c = ASCII_HYPH;
 718                         ws = NULL;
 719                         break;
 720                 case ' ':
 721                         if (NULL == ws)
 722                                 ws = c;
 723                         continue;
 724                 case '\t':
 725                         /*
 726                          * Always warn about trailing tabs,
 727                          * even outside literal context,
 728                          * where they should be put on the next line.
 729                          */
 730                         if (NULL == ws)
 731                                 ws = c;
 732                         /*
 733                          * Strip trailing tabs in literal context only;
 734                          * outside, they affect the next line.
 735                          */
 736                         if (MDOC_LITERAL & m->flags)
 737                                 continue;
 738                         break;
 739                 case '\\':
 740                         /* Skip the escaped character, too, if any. */
 741                         if (c[1])
 742                                 c++;
 743                         /* FALLTHROUGH */
 744                 default:
 745                         ws = NULL;
 746                         break;
 747                 }
 748                 end = c + 1;
 749         }
 750         *end = '\0';
 751
 752         if (ws)
 753                 mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
 754
 755         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 756                 mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
 757
 758                 /*
 759                  * Insert a `sp' in the case of a blank line.  Technically,
 760                  * blank lines aren't allowed, but enough manuals assume this
 761                  * behaviour that we want to work around it.
 762                  */
 763                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 764                         return(0);
 765
 766                 m->next = MDOC_NEXT_SIBLING;
 767                 return(1);
 768         }
 769
 770         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 771                 return(0);
 772
 773         if (MDOC_LITERAL & m->flags)
 774                 return(1);
 775
 776         /*
 777          * End-of-sentence check.  If the last character is an unescaped
 778          * EOS character, then flag the node as being the end of a
 779          * sentence.  The front-end will know how to interpret this.
 780          */
 781
 782         assert(buf < end);
 783
 784         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 785                 m->last->flags |= MDOC_EOS;
 786
 787         return(1);
 788 }
 789
 790
 791 /*
 792  * Parse a macro line, that is, a line beginning with the control
 793  * character.
 794  */
 795 static int
 796 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 797 {
 798         enum mdoct        tok;
 799         int               i, j, sv;
 800         char              mac[5];
 801         struct mdoc_node *n;
 802
 803         /* Empty lines are ignored. */
 804
 805         offs++;
 806
 807         if ('\0' == buf[offs])
 808                 return(1);
 809
 810         i = offs;
 811
 812         /* Accept tabs/whitespace after the initial control char. */
 813
 814         if (' ' == buf[i] || '\t' == buf[i]) {
 815                 i++;
 816                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 817                         i++;
 818                 if ('\0' == buf[i])
 819                         return(1);
 820         }
 821
 822         sv = i;
 823
 824         /*
 825          * Copy the first word into a nil-terminated buffer.
 826          * Stop copying when a tab, space, or eoln is encountered.
 827          */
 828
 829         j = 0;
 830         while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
 831                 mac[j++] = buf[i++];
 832         mac[j] = '\0';
 833
 834         tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 835         if (MDOC_MAX == tok) {
 836                 mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv, "%s", buf + sv - 1);
 837                 return(1);
 838         }
 839
 840         /* Disregard the first trailing tab, if applicable. */
 841
 842         if ('\t' == buf[i])
 843                 i++;
 844
 845         /* Jump to the next non-whitespace word. */
 846
 847         while (buf[i] && ' ' == buf[i])
 848                 i++;
 849
 850         /*
 851          * Trailing whitespace.  Note that tabs are allowed to be passed
 852          * into the parser as "text", so we only warn about spaces here.
 853          */
 854
 855         if ('\0' == buf[i] && ' ' == buf[i - 1])
 856                 mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE);
 857
 858         /*
 859          * If an initial macro or a list invocation, divert directly
 860          * into macro processing.
 861          */
 862
 863         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 864                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 865                         goto err;
 866                 return(1);
 867         }
 868
 869         n = m->last;
 870         assert(m->last);
 871
 872         /*
 873          * If the first macro of a `Bl -column', open an `It' block
 874          * context around the parsed macro.
 875          */
 876
 877         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 878                         LIST_column == n->norm->Bl.type) {
 879                 m->flags |= MDOC_FREECOL;
 880                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 881                         goto err;
 882                 return(1);
 883         }
 884
 885         /*
 886          * If we're following a block-level `It' within a `Bl -column'
 887          * context (perhaps opened in the above block or in ptext()),
 888          * then open an `It' block context around the parsed macro.
 889          */
 890
 891         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 892                         NULL != n->parent &&
 893                         MDOC_Bl == n->parent->tok &&
 894                         LIST_column == n->parent->norm->Bl.type) {
 895                 m->flags |= MDOC_FREECOL;
 896                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 897                         goto err;
 898                 return(1);
 899         }
 900
 901         /* Normal processing of a macro. */
 902
 903         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 904                 goto err;
 905
 906         return(1);
 907
 908 err:    /* Error out. */
 909
 910         m->flags |= MDOC_HALT;
 911         return(0);
 912 }
 913
 914