mdoc.c

   1 /*      $Id: mdoc.c,v 1.183 2011/03/15 13:23:33 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 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 const struct mdoc_node *
 103 mdoc_node(const struct mdoc *m)
 104 {
 105
 106         assert( ! (MDOC_HALT & m->flags));
 107         return(m->first);
 108 }
 109
 110
 111 const struct mdoc_meta *
 112 mdoc_meta(const struct mdoc *m)
 113 {
 114
 115         assert( ! (MDOC_HALT & m->flags));
 116         return(&m->meta);
 117 }
 118
 119
 120 /*
 121  * Frees volatile resources (parse tree, meta-data, fields).
 122  */
 123 static void
 124 mdoc_free1(struct mdoc *mdoc)
 125 {
 126
 127         if (mdoc->first)
 128                 mdoc_node_delete(mdoc, mdoc->first);
 129         if (mdoc->meta.title)
 130                 free(mdoc->meta.title);
 131         if (mdoc->meta.os)
 132                 free(mdoc->meta.os);
 133         if (mdoc->meta.name)
 134                 free(mdoc->meta.name);
 135         if (mdoc->meta.arch)
 136                 free(mdoc->meta.arch);
 137         if (mdoc->meta.vol)
 138                 free(mdoc->meta.vol);
 139         if (mdoc->meta.msec)
 140                 free(mdoc->meta.msec);
 141         if (mdoc->meta.date)
 142                 free(mdoc->meta.date);
 143 }
 144
 145
 146 /*
 147  * Allocate all volatile resources (parse tree, meta-data, fields).
 148  */
 149 static void
 150 mdoc_alloc1(struct mdoc *mdoc)
 151 {
 152
 153         memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 154         mdoc->flags = 0;
 155         mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
 156         mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 157         mdoc->first = mdoc->last;
 158         mdoc->last->type = MDOC_ROOT;
 159         mdoc->next = MDOC_NEXT_CHILD;
 160 }
 161
 162
 163 /*
 164  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 165  * data with mdoc_alloc1().  After invocation, parse data has been reset
 166  * and the parser is ready for re-invocation on a new tree; however,
 167  * cross-parse non-volatile data is kept intact.
 168  */
 169 void
 170 mdoc_reset(struct mdoc *mdoc)
 171 {
 172
 173         mdoc_free1(mdoc);
 174         mdoc_alloc1(mdoc);
 175 }
 176
 177
 178 /*
 179  * Completely free up all volatile and non-volatile parse resources.
 180  * After invocation, the pointer is no longer usable.
 181  */
 182 void
 183 mdoc_free(struct mdoc *mdoc)
 184 {
 185
 186         mdoc_free1(mdoc);
 187         free(mdoc);
 188 }
 189
 190
 191 /*
 192  * Allocate volatile and non-volatile parse resources.
 193  */
 194 struct mdoc *
 195 mdoc_alloc(struct regset *regs, void *data, mandocmsg msg)
 196 {
 197         struct mdoc     *p;
 198
 199         p = mandoc_calloc(1, sizeof(struct mdoc));
 200
 201         p->msg = msg;
 202         p->data = data;
 203         p->regs = regs;
 204
 205         mdoc_hash_init();
 206         mdoc_alloc1(p);
 207         return(p);
 208 }
 209
 210
 211 /*
 212  * Climb back up the parse tree, validating open scopes.  Mostly calls
 213  * through to macro_end() in macro.c.
 214  */
 215 int
 216 mdoc_endparse(struct mdoc *m)
 217 {
 218
 219         assert( ! (MDOC_HALT & m->flags));
 220         if (mdoc_macroend(m))
 221                 return(1);
 222         m->flags |= MDOC_HALT;
 223         return(0);
 224 }
 225
 226 int
 227 mdoc_addeqn(struct mdoc *m, const struct eqn *ep)
 228 {
 229         struct mdoc_node *n;
 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, ep->line, ep->pos, MANDOCERR_NOTEXT);
 237                 return(1);
 238         }
 239
 240         n = node_alloc(m, ep->line, ep->pos, MDOC_MAX, MDOC_EQN);
 241         n->eqn = ep;
 242
 243         if ( ! node_append(m, n))
 244                 return(0);
 245
 246         m->next = MDOC_NEXT_SIBLING;
 247         return(1);
 248 }
 249
 250 int
 251 mdoc_addspan(struct mdoc *m, const struct tbl_span *sp)
 252 {
 253         struct mdoc_node *n;
 254
 255         assert( ! (MDOC_HALT & m->flags));
 256
 257         /* No text before an initial macro. */
 258
 259         if (SEC_NONE == m->lastnamed) {
 260                 mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT);
 261                 return(1);
 262         }
 263
 264         n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL);
 265         n->span = sp;
 266
 267         if ( ! node_append(m, n))
 268                 return(0);
 269
 270         m->next = MDOC_NEXT_SIBLING;
 271         return(1);
 272 }
 273
 274
 275 /*
 276  * Main parse routine.  Parses a single line -- really just hands off to
 277  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 278  */
 279 int
 280 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 281 {
 282
 283         assert( ! (MDOC_HALT & m->flags));
 284
 285         m->flags |= MDOC_NEWLINE;
 286
 287         /*
 288          * Let the roff nS register switch SYNOPSIS mode early,
 289          * such that the parser knows at all times
 290          * whether this mode is on or off.
 291          * Note that this mode is also switched by the Sh macro.
 292          */
 293         if (m->regs->regs[(int)REG_nS].set) {
 294                 if (m->regs->regs[(int)REG_nS].v.u)
 295                         m->flags |= MDOC_SYNOPSIS;
 296                 else
 297                         m->flags &= ~MDOC_SYNOPSIS;
 298         }
 299
 300         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 301                         mdoc_pmacro(m, ln, buf, offs) :
 302                         mdoc_ptext(m, ln, buf, offs));
 303 }
 304
 305
 306 int
 307 mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
 308                 int ln, int pos, const char *fmt, ...)
 309 {
 310         char             buf[256];
 311         va_list          ap;
 312
 313         va_start(ap, fmt);
 314         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 315         va_end(ap);
 316
 317         return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
 318 }
 319
 320
 321 int
 322 mdoc_macro(MACRO_PROT_ARGS)
 323 {
 324         assert(tok < MDOC_MAX);
 325
 326         /* If we're in the body, deny prologue calls. */
 327
 328         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 329                         MDOC_PBODY & m->flags) {
 330                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY);
 331                 return(1);
 332         }
 333
 334         /* If we're in the prologue, deny "body" macros.  */
 335
 336         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 337                         ! (MDOC_PBODY & m->flags)) {
 338                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG);
 339                 if (NULL == m->meta.msec)
 340                         m->meta.msec = mandoc_strdup("1");
 341                 if (NULL == m->meta.title)
 342                         m->meta.title = mandoc_strdup("UNKNOWN");
 343                 if (NULL == m->meta.vol)
 344                         m->meta.vol = mandoc_strdup("LOCAL");
 345                 if (NULL == m->meta.os)
 346                         m->meta.os = mandoc_strdup("LOCAL");
 347                 if (NULL == m->meta.date)
 348                         m->meta.date = mandoc_normdate(NULL,
 349                             m->msg, m->data, line, ppos);
 350                 m->flags |= MDOC_PBODY;
 351         }
 352
 353         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 354 }
 355
 356
 357 static int
 358 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 359 {
 360
 361         assert(mdoc->last);
 362         assert(mdoc->first);
 363         assert(MDOC_ROOT != p->type);
 364
 365         switch (mdoc->next) {
 366         case (MDOC_NEXT_SIBLING):
 367                 mdoc->last->next = p;
 368                 p->prev = mdoc->last;
 369                 p->parent = mdoc->last->parent;
 370                 break;
 371         case (MDOC_NEXT_CHILD):
 372                 mdoc->last->child = p;
 373                 p->parent = mdoc->last;
 374                 break;
 375         default:
 376                 abort();
 377                 /* NOTREACHED */
 378         }
 379
 380         p->parent->nchild++;
 381
 382         /*
 383          * Copy over the normalised-data pointer of our parent.  Not
 384          * everybody has one, but copying a null pointer is fine.
 385          */
 386
 387         switch (p->type) {
 388         case (MDOC_BODY):
 389                 /* FALLTHROUGH */
 390         case (MDOC_TAIL):
 391                 /* FALLTHROUGH */
 392         case (MDOC_HEAD):
 393                 p->norm = p->parent->norm;
 394                 break;
 395         default:
 396                 break;
 397         }
 398
 399         if ( ! mdoc_valid_pre(mdoc, p))
 400                 return(0);
 401
 402         switch (p->type) {
 403         case (MDOC_HEAD):
 404                 assert(MDOC_BLOCK == p->parent->type);
 405                 p->parent->head = p;
 406                 break;
 407         case (MDOC_TAIL):
 408                 assert(MDOC_BLOCK == p->parent->type);
 409                 p->parent->tail = p;
 410                 break;
 411         case (MDOC_BODY):
 412                 if (p->end)
 413                         break;
 414                 assert(MDOC_BLOCK == p->parent->type);
 415                 p->parent->body = p;
 416                 break;
 417         default:
 418                 break;
 419         }
 420
 421         mdoc->last = p;
 422
 423         switch (p->type) {
 424         case (MDOC_TBL):
 425                 /* FALLTHROUGH */
 426         case (MDOC_TEXT):
 427                 if ( ! mdoc_valid_post(mdoc))
 428                         return(0);
 429                 break;
 430         default:
 431                 break;
 432         }
 433
 434         return(1);
 435 }
 436
 437
 438 static struct mdoc_node *
 439 node_alloc(struct mdoc *m, int line, int pos,
 440                 enum mdoct tok, enum mdoc_type type)
 441 {
 442         struct mdoc_node *p;
 443
 444         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 445         p->sec = m->lastsec;
 446         p->line = line;
 447         p->pos = pos;
 448         p->tok = tok;
 449         p->type = type;
 450
 451         /* Flag analysis. */
 452
 453         if (MDOC_SYNOPSIS & m->flags)
 454                 p->flags |= MDOC_SYNPRETTY;
 455         else
 456                 p->flags &= ~MDOC_SYNPRETTY;
 457         if (MDOC_NEWLINE & m->flags)
 458                 p->flags |= MDOC_LINE;
 459         m->flags &= ~MDOC_NEWLINE;
 460
 461         return(p);
 462 }
 463
 464
 465 int
 466 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 467 {
 468         struct mdoc_node *p;
 469
 470         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 471         if ( ! node_append(m, p))
 472                 return(0);
 473         m->next = MDOC_NEXT_CHILD;
 474         return(1);
 475 }
 476
 477
 478 int
 479 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 480 {
 481         struct mdoc_node *p;
 482
 483         assert(m->first);
 484         assert(m->last);
 485
 486         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 487         if ( ! node_append(m, p))
 488                 return(0);
 489         m->next = MDOC_NEXT_CHILD;
 490         return(1);
 491 }
 492
 493
 494 int
 495 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 496 {
 497         struct mdoc_node *p;
 498
 499         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 500         if ( ! node_append(m, p))
 501                 return(0);
 502         m->next = MDOC_NEXT_CHILD;
 503         return(1);
 504 }
 505
 506
 507 int
 508 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 509                 struct mdoc_node *body, enum mdoc_endbody end)
 510 {
 511         struct mdoc_node *p;
 512
 513         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 514         p->pending = body;
 515         p->end = end;
 516         if ( ! node_append(m, p))
 517                 return(0);
 518         m->next = MDOC_NEXT_SIBLING;
 519         return(1);
 520 }
 521
 522
 523 int
 524 mdoc_block_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_BLOCK);
 530         p->args = args;
 531         if (p->args)
 532                 (args->refcnt)++;
 533
 534         switch (tok) {
 535         case (MDOC_Bd):
 536                 /* FALLTHROUGH */
 537         case (MDOC_Bf):
 538                 /* FALLTHROUGH */
 539         case (MDOC_Bl):
 540                 /* FALLTHROUGH */
 541         case (MDOC_Rs):
 542                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 543                 break;
 544         default:
 545                 break;
 546         }
 547
 548         if ( ! node_append(m, p))
 549                 return(0);
 550         m->next = MDOC_NEXT_CHILD;
 551         return(1);
 552 }
 553
 554
 555 int
 556 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 557                 enum mdoct tok, struct mdoc_arg *args)
 558 {
 559         struct mdoc_node *p;
 560
 561         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 562         p->args = args;
 563         if (p->args)
 564                 (args->refcnt)++;
 565
 566         switch (tok) {
 567         case (MDOC_An):
 568                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 569                 break;
 570         default:
 571                 break;
 572         }
 573
 574         if ( ! node_append(m, p))
 575                 return(0);
 576         m->next = MDOC_NEXT_CHILD;
 577         return(1);
 578 }
 579
 580 int
 581 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 582 {
 583         struct mdoc_node *n;
 584         size_t            sv, len;
 585
 586         len = strlen(p);
 587
 588         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 589         n->string = mandoc_malloc(len + 1);
 590         sv = strlcpy(n->string, p, len + 1);
 591
 592         /* Prohibit truncation. */
 593         assert(sv < len + 1);
 594
 595         if ( ! node_append(m, n))
 596                 return(0);
 597
 598         m->next = MDOC_NEXT_SIBLING;
 599         return(1);
 600 }
 601
 602
 603 static void
 604 mdoc_node_free(struct mdoc_node *p)
 605 {
 606
 607         if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
 608                 free(p->norm);
 609         if (p->string)
 610                 free(p->string);
 611         if (p->args)
 612                 mdoc_argv_free(p->args);
 613         free(p);
 614 }
 615
 616
 617 static void
 618 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 619 {
 620
 621         /* Adjust siblings. */
 622
 623         if (n->prev)
 624                 n->prev->next = n->next;
 625         if (n->next)
 626                 n->next->prev = n->prev;
 627
 628         /* Adjust parent. */
 629
 630         if (n->parent) {
 631                 n->parent->nchild--;
 632                 if (n->parent->child == n)
 633                         n->parent->child = n->prev ? n->prev : n->next;
 634                 if (n->parent->last == n)
 635                         n->parent->last = n->prev ? n->prev : NULL;
 636         }
 637
 638         /* Adjust parse point, if applicable. */
 639
 640         if (m && m->last == n) {
 641                 if (n->prev) {
 642                         m->last = n->prev;
 643                         m->next = MDOC_NEXT_SIBLING;
 644                 } else {
 645                         m->last = n->parent;
 646                         m->next = MDOC_NEXT_CHILD;
 647                 }
 648         }
 649
 650         if (m && m->first == n)
 651                 m->first = NULL;
 652 }
 653
 654
 655 void
 656 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 657 {
 658
 659         while (p->child) {
 660                 assert(p->nchild);
 661                 mdoc_node_delete(m, p->child);
 662         }
 663         assert(0 == p->nchild);
 664
 665         mdoc_node_unlink(m, p);
 666         mdoc_node_free(p);
 667 }
 668
 669
 670 /*
 671  * Parse free-form text, that is, a line that does not begin with the
 672  * control character.
 673  */
 674 static int
 675 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 676 {
 677         char             *c, *ws, *end;
 678         struct mdoc_node *n;
 679
 680         /* Ignore bogus comments. */
 681
 682         if ('\\' == buf[offs] &&
 683                         '.' == buf[offs + 1] &&
 684                         '"' == buf[offs + 2]) {
 685                 mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT);
 686                 return(1);
 687         }
 688
 689         /* No text before an initial macro. */
 690
 691         if (SEC_NONE == m->lastnamed) {
 692                 mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
 693                 return(1);
 694         }
 695
 696         assert(m->last);
 697         n = m->last;
 698
 699         /*
 700          * Divert directly to list processing if we're encountering a
 701          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 702          * (a MDOC_BODY means it's already open, in which case we should
 703          * process within its context in the normal way).
 704          */
 705
 706         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 707                         LIST_column == n->norm->Bl.type) {
 708                 /* `Bl' is open without any children. */
 709                 m->flags |= MDOC_FREECOL;
 710                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 711         }
 712
 713         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 714                         NULL != n->parent &&
 715                         MDOC_Bl == n->parent->tok &&
 716                         LIST_column == n->parent->norm->Bl.type) {
 717                 /* `Bl' has block-level `It' children. */
 718                 m->flags |= MDOC_FREECOL;
 719                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 720         }
 721
 722         /*
 723          * Search for the beginning of unescaped trailing whitespace (ws)
 724          * and for the first character not to be output (end).
 725          */
 726
 727         /* FIXME: replace with strcspn(). */
 728         ws = NULL;
 729         for (c = end = buf + offs; *c; c++) {
 730                 switch (*c) {
 731                 case '-':
 732                         if (mandoc_hyph(buf + offs, c))
 733                                 *c = ASCII_HYPH;
 734                         ws = NULL;
 735                         break;
 736                 case ' ':
 737                         if (NULL == ws)
 738                                 ws = c;
 739                         continue;
 740                 case '\t':
 741                         /*
 742                          * Always warn about trailing tabs,
 743                          * even outside literal context,
 744                          * where they should be put on the next line.
 745                          */
 746                         if (NULL == ws)
 747                                 ws = c;
 748                         /*
 749                          * Strip trailing tabs in literal context only;
 750                          * outside, they affect the next line.
 751                          */
 752                         if (MDOC_LITERAL & m->flags)
 753                                 continue;
 754                         break;
 755                 case '\\':
 756                         /* Skip the escaped character, too, if any. */
 757                         if (c[1])
 758                                 c++;
 759                         /* FALLTHROUGH */
 760                 default:
 761                         ws = NULL;
 762                         break;
 763                 }
 764                 end = c + 1;
 765         }
 766         *end = '\0';
 767
 768         if (ws)
 769                 mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
 770
 771         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 772                 mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
 773
 774                 /*
 775                  * Insert a `sp' in the case of a blank line.  Technically,
 776                  * blank lines aren't allowed, but enough manuals assume this
 777                  * behaviour that we want to work around it.
 778                  */
 779                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 780                         return(0);
 781
 782                 m->next = MDOC_NEXT_SIBLING;
 783                 return(1);
 784         }
 785
 786         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 787                 return(0);
 788
 789         if (MDOC_LITERAL & m->flags)
 790                 return(1);
 791
 792         /*
 793          * End-of-sentence check.  If the last character is an unescaped
 794          * EOS character, then flag the node as being the end of a
 795          * sentence.  The front-end will know how to interpret this.
 796          */
 797
 798         assert(buf < end);
 799
 800         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 801                 m->last->flags |= MDOC_EOS;
 802
 803         return(1);
 804 }
 805
 806
 807 /*
 808  * Parse a macro line, that is, a line beginning with the control
 809  * character.
 810  */
 811 static int
 812 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 813 {
 814         enum mdoct        tok;
 815         int               i, j, sv;
 816         char              mac[5];
 817         struct mdoc_node *n;
 818
 819         /* Empty lines are ignored. */
 820
 821         offs++;
 822
 823         if ('\0' == buf[offs])
 824                 return(1);
 825
 826         i = offs;
 827
 828         /* Accept tabs/whitespace after the initial control char. */
 829
 830         if (' ' == buf[i] || '\t' == buf[i]) {
 831                 i++;
 832                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 833                         i++;
 834                 if ('\0' == buf[i])
 835                         return(1);
 836         }
 837
 838         sv = i;
 839
 840         /*
 841          * Copy the first word into a nil-terminated buffer.
 842          * Stop copying when a tab, space, or eoln is encountered.
 843          */
 844
 845         j = 0;
 846         while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
 847                 mac[j++] = buf[i++];
 848         mac[j] = '\0';
 849
 850         tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 851         if (MDOC_MAX == tok) {
 852                 mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv, "%s", buf + sv - 1);
 853                 return(1);
 854         }
 855
 856         /* Disregard the first trailing tab, if applicable. */
 857
 858         if ('\t' == buf[i])
 859                 i++;
 860
 861         /* Jump to the next non-whitespace word. */
 862
 863         while (buf[i] && ' ' == buf[i])
 864                 i++;
 865
 866         /*
 867          * Trailing whitespace.  Note that tabs are allowed to be passed
 868          * into the parser as "text", so we only warn about spaces here.
 869          */
 870
 871         if ('\0' == buf[i] && ' ' == buf[i - 1])
 872                 mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE);
 873
 874         /*
 875          * If an initial macro or a list invocation, divert directly
 876          * into macro processing.
 877          */
 878
 879         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 880                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 881                         goto err;
 882                 return(1);
 883         }
 884
 885         n = m->last;
 886         assert(m->last);
 887
 888         /*
 889          * If the first macro of a `Bl -column', open an `It' block
 890          * context around the parsed macro.
 891          */
 892
 893         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 894                         LIST_column == n->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         /*
 902          * If we're following a block-level `It' within a `Bl -column'
 903          * context (perhaps opened in the above block or in ptext()),
 904          * then open an `It' block context around the parsed macro.
 905          */
 906
 907         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 908                         NULL != n->parent &&
 909                         MDOC_Bl == n->parent->tok &&
 910                         LIST_column == n->parent->norm->Bl.type) {
 911                 m->flags |= MDOC_FREECOL;
 912                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 913                         goto err;
 914                 return(1);
 915         }
 916
 917         /* Normal processing of a macro. */
 918
 919         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 920                 goto err;
 921
 922         return(1);
 923
 924 err:    /* Error out. */
 925
 926         m->flags |= MDOC_HALT;
 927         return(0);
 928 }
 929
 930