mdoc.c

   1 /*      $Id: mdoc.c,v 1.173 2010/12/25 13:50:37 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         /*
 332          * Copy over the normalised-data pointer of our parent.  Not
 333          * everybody has one, but copying a null pointer is fine.
 334          */
 335
 336         switch (p->type) {
 337         case (MDOC_BODY):
 338                 /* FALLTHROUGH */
 339         case (MDOC_TAIL):
 340                 /* FALLTHROUGH */
 341         case (MDOC_HEAD):
 342                 p->norm = p->parent->norm;
 343                 break;
 344         default:
 345                 break;
 346         }
 347
 348         if ( ! mdoc_valid_pre(mdoc, p))
 349                 return(0);
 350
 351         switch (p->type) {
 352         case (MDOC_HEAD):
 353                 assert(MDOC_BLOCK == p->parent->type);
 354                 p->parent->head = p;
 355                 break;
 356         case (MDOC_TAIL):
 357                 assert(MDOC_BLOCK == p->parent->type);
 358                 p->parent->tail = p;
 359                 break;
 360         case (MDOC_BODY):
 361                 if (p->end)
 362                         break;
 363                 assert(MDOC_BLOCK == p->parent->type);
 364                 p->parent->body = p;
 365                 break;
 366         default:
 367                 break;
 368         }
 369
 370         mdoc->last = p;
 371
 372         switch (p->type) {
 373         case (MDOC_TEXT):
 374                 if ( ! mdoc_valid_post(mdoc))
 375                         return(0);
 376                 break;
 377         default:
 378                 break;
 379         }
 380
 381         return(1);
 382 }
 383
 384
 385 static struct mdoc_node *
 386 node_alloc(struct mdoc *m, int line, int pos,
 387                 enum mdoct tok, enum mdoc_type type)
 388 {
 389         struct mdoc_node *p;
 390
 391         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 392         p->sec = m->lastsec;
 393         p->line = line;
 394         p->pos = pos;
 395         p->tok = tok;
 396         p->type = type;
 397
 398         /* Flag analysis. */
 399
 400         if (MDOC_SYNOPSIS & m->flags)
 401                 p->flags |= MDOC_SYNPRETTY;
 402         else
 403                 p->flags &= ~MDOC_SYNPRETTY;
 404         if (MDOC_NEWLINE & m->flags)
 405                 p->flags |= MDOC_LINE;
 406         m->flags &= ~MDOC_NEWLINE;
 407
 408         return(p);
 409 }
 410
 411
 412 int
 413 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 414 {
 415         struct mdoc_node *p;
 416
 417         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 418         if ( ! node_append(m, p))
 419                 return(0);
 420         m->next = MDOC_NEXT_CHILD;
 421         return(1);
 422 }
 423
 424
 425 int
 426 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 427 {
 428         struct mdoc_node *p;
 429
 430         assert(m->first);
 431         assert(m->last);
 432
 433         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 434         if ( ! node_append(m, p))
 435                 return(0);
 436         m->next = MDOC_NEXT_CHILD;
 437         return(1);
 438 }
 439
 440
 441 int
 442 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 443 {
 444         struct mdoc_node *p;
 445
 446         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 447         if ( ! node_append(m, p))
 448                 return(0);
 449         m->next = MDOC_NEXT_CHILD;
 450         return(1);
 451 }
 452
 453
 454 int
 455 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 456                 struct mdoc_node *body, enum mdoc_endbody end)
 457 {
 458         struct mdoc_node *p;
 459
 460         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 461         p->pending = body;
 462         p->end = end;
 463         if ( ! node_append(m, p))
 464                 return(0);
 465         m->next = MDOC_NEXT_SIBLING;
 466         return(1);
 467 }
 468
 469
 470 int
 471 mdoc_block_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_BLOCK);
 477         p->args = args;
 478         if (p->args)
 479                 (args->refcnt)++;
 480
 481         switch (tok) {
 482         case (MDOC_Bd):
 483                 /* FALLTHROUGH */
 484         case (MDOC_Bf):
 485                 /* FALLTHROUGH */
 486         case (MDOC_Bl):
 487                 /* FALLTHROUGH */
 488         case (MDOC_Rs):
 489                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 490                 break;
 491         default:
 492                 break;
 493         }
 494
 495         if ( ! node_append(m, p))
 496                 return(0);
 497         m->next = MDOC_NEXT_CHILD;
 498         return(1);
 499 }
 500
 501
 502 int
 503 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 504                 enum mdoct tok, struct mdoc_arg *args)
 505 {
 506         struct mdoc_node *p;
 507
 508         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 509         p->args = args;
 510         if (p->args)
 511                 (args->refcnt)++;
 512
 513         switch (tok) {
 514         case (MDOC_An):
 515                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 516                 break;
 517         default:
 518                 break;
 519         }
 520
 521         if ( ! node_append(m, p))
 522                 return(0);
 523         m->next = MDOC_NEXT_CHILD;
 524         return(1);
 525 }
 526
 527
 528 int
 529 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 530 {
 531         struct mdoc_node *n;
 532         size_t            sv, len;
 533
 534         len = strlen(p);
 535
 536         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 537         n->string = mandoc_malloc(len + 1);
 538         sv = strlcpy(n->string, p, len + 1);
 539
 540         /* Prohibit truncation. */
 541         assert(sv < len + 1);
 542
 543         if ( ! node_append(m, n))
 544                 return(0);
 545
 546         m->next = MDOC_NEXT_SIBLING;
 547         return(1);
 548 }
 549
 550
 551 static void
 552 mdoc_node_free(struct mdoc_node *p)
 553 {
 554
 555         if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
 556                 free(p->norm);
 557         if (p->string)
 558                 free(p->string);
 559         if (p->args)
 560                 mdoc_argv_free(p->args);
 561         free(p);
 562 }
 563
 564
 565 static void
 566 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 567 {
 568
 569         /* Adjust siblings. */
 570
 571         if (n->prev)
 572                 n->prev->next = n->next;
 573         if (n->next)
 574                 n->next->prev = n->prev;
 575
 576         /* Adjust parent. */
 577
 578         if (n->parent) {
 579                 n->parent->nchild--;
 580                 if (n->parent->child == n)
 581                         n->parent->child = n->prev ? n->prev : n->next;
 582                 if (n->parent->last == n)
 583                         n->parent->last = n->prev ? n->prev : NULL;
 584         }
 585
 586         /* Adjust parse point, if applicable. */
 587
 588         if (m && m->last == n) {
 589                 if (n->prev) {
 590                         m->last = n->prev;
 591                         m->next = MDOC_NEXT_SIBLING;
 592                 } else {
 593                         m->last = n->parent;
 594                         m->next = MDOC_NEXT_CHILD;
 595                 }
 596         }
 597
 598         if (m && m->first == n)
 599                 m->first = NULL;
 600 }
 601
 602
 603 void
 604 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 605 {
 606
 607         while (p->child) {
 608                 assert(p->nchild);
 609                 mdoc_node_delete(m, p->child);
 610         }
 611         assert(0 == p->nchild);
 612
 613         mdoc_node_unlink(m, p);
 614         mdoc_node_free(p);
 615 }
 616
 617
 618 /*
 619  * Parse free-form text, that is, a line that does not begin with the
 620  * control character.
 621  */
 622 static int
 623 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 624 {
 625         char             *c, *ws, *end;
 626         struct mdoc_node *n;
 627
 628         /* Ignore bogus comments. */
 629
 630         if ('\\' == buf[offs] &&
 631                         '.' == buf[offs + 1] &&
 632                         '"' == buf[offs + 2])
 633                 return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
 634
 635         /* No text before an initial macro. */
 636
 637         if (SEC_NONE == m->lastnamed)
 638                 return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
 639
 640         assert(m->last);
 641         n = m->last;
 642
 643         /*
 644          * Divert directly to list processing if we're encountering a
 645          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 646          * (a MDOC_BODY means it's already open, in which case we should
 647          * process within its context in the normal way).
 648          */
 649
 650         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 651                         LIST_column == n->norm->Bl.type) {
 652                 /* `Bl' is open without any children. */
 653                 m->flags |= MDOC_FREECOL;
 654                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 655         }
 656
 657         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 658                         NULL != n->parent &&
 659                         MDOC_Bl == n->parent->tok &&
 660                         LIST_column == n->parent->norm->Bl.type) {
 661                 /* `Bl' has block-level `It' children. */
 662                 m->flags |= MDOC_FREECOL;
 663                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 664         }
 665
 666         /*
 667          * Search for the beginning of unescaped trailing whitespace (ws)
 668          * and for the first character not to be output (end).
 669          */
 670
 671         /* FIXME: replace with strcspn(). */
 672         ws = NULL;
 673         for (c = end = buf + offs; *c; c++) {
 674                 switch (*c) {
 675                 case '-':
 676                         if (mandoc_hyph(buf + offs, c))
 677                                 *c = ASCII_HYPH;
 678                         ws = NULL;
 679                         break;
 680                 case ' ':
 681                         if (NULL == ws)
 682                                 ws = c;
 683                         continue;
 684                 case '\t':
 685                         /*
 686                          * Always warn about trailing tabs,
 687                          * even outside literal context,
 688                          * where they should be put on the next line.
 689                          */
 690                         if (NULL == ws)
 691                                 ws = c;
 692                         /*
 693                          * Strip trailing tabs in literal context only;
 694                          * outside, they affect the next line.
 695                          */
 696                         if (MDOC_LITERAL & m->flags)
 697                                 continue;
 698                         break;
 699                 case '\\':
 700                         /* Skip the escaped character, too, if any. */
 701                         if (c[1])
 702                                 c++;
 703                         /* FALLTHROUGH */
 704                 default:
 705                         ws = NULL;
 706                         break;
 707                 }
 708                 end = c + 1;
 709         }
 710         *end = '\0';
 711
 712         if (ws)
 713                 if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
 714                         return(0);
 715
 716         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 717                 if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
 718                         return(0);
 719
 720                 /*
 721                  * Insert a `sp' in the case of a blank line.  Technically,
 722                  * blank lines aren't allowed, but enough manuals assume this
 723                  * behaviour that we want to work around it.
 724                  */
 725                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 726                         return(0);
 727
 728                 m->next = MDOC_NEXT_SIBLING;
 729                 return(1);
 730         }
 731
 732         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 733                 return(0);
 734
 735         if (MDOC_LITERAL & m->flags)
 736                 return(1);
 737
 738         /*
 739          * End-of-sentence check.  If the last character is an unescaped
 740          * EOS character, then flag the node as being the end of a
 741          * sentence.  The front-end will know how to interpret this.
 742          */
 743
 744         assert(buf < end);
 745
 746         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 747                 m->last->flags |= MDOC_EOS;
 748
 749         return(1);
 750 }
 751
 752
 753 /*
 754  * Parse a macro line, that is, a line beginning with the control
 755  * character.
 756  */
 757 static int
 758 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 759 {
 760         enum mdoct        tok;
 761         int               i, j, sv;
 762         char              mac[5];
 763         struct mdoc_node *n;
 764
 765         /* Empty lines are ignored. */
 766
 767         offs++;
 768
 769         if ('\0' == buf[offs])
 770                 return(1);
 771
 772         i = offs;
 773
 774         /* Accept tabs/whitespace after the initial control char. */
 775
 776         if (' ' == buf[i] || '\t' == buf[i]) {
 777                 i++;
 778                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 779                         i++;
 780                 if ('\0' == buf[i])
 781                         return(1);
 782         }
 783
 784         sv = i;
 785
 786         /*
 787          * Copy the first word into a nil-terminated buffer.
 788          * Stop copying when a tab, space, or eoln is encountered.
 789          */
 790
 791         j = 0;
 792         while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
 793                 mac[j++] = buf[i++];
 794         mac[j] = '\0';
 795
 796         tok = (j > 1 || j < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 797         if (MDOC_MAX == tok) {
 798                 mdoc_vmsg(m, MANDOCERR_MACRO, ln, sv, "%s", buf + sv - 1);
 799                 return(1);
 800         }
 801
 802         /* Disregard the first trailing tab, if applicable. */
 803
 804         if ('\t' == buf[i])
 805                 i++;
 806
 807         /* Jump to the next non-whitespace word. */
 808
 809         while (buf[i] && ' ' == buf[i])
 810                 i++;
 811
 812         /*
 813          * Trailing whitespace.  Note that tabs are allowed to be passed
 814          * into the parser as "text", so we only warn about spaces here.
 815          */
 816
 817         if ('\0' == buf[i] && ' ' == buf[i - 1])
 818                 if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
 819                         goto err;
 820
 821         /*
 822          * If an initial macro or a list invocation, divert directly
 823          * into macro processing.
 824          */
 825
 826         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 827                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 828                         goto err;
 829                 return(1);
 830         }
 831
 832         n = m->last;
 833         assert(m->last);
 834
 835         /*
 836          * If the first macro of a `Bl -column', open an `It' block
 837          * context around the parsed macro.
 838          */
 839
 840         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 841                         LIST_column == n->norm->Bl.type) {
 842                 m->flags |= MDOC_FREECOL;
 843                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 844                         goto err;
 845                 return(1);
 846         }
 847
 848         /*
 849          * If we're following a block-level `It' within a `Bl -column'
 850          * context (perhaps opened in the above block or in ptext()),
 851          * then open an `It' block context around the parsed macro.
 852          */
 853
 854         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 855                         NULL != n->parent &&
 856                         MDOC_Bl == n->parent->tok &&
 857                         LIST_column == n->parent->norm->Bl.type) {
 858                 m->flags |= MDOC_FREECOL;
 859                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 860                         goto err;
 861                 return(1);
 862         }
 863
 864         /* Normal processing of a macro. */
 865
 866         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 867                 goto err;
 868
 869         return(1);
 870
 871 err:    /* Error out. */
 872
 873         m->flags |= MDOC_HALT;
 874         return(0);
 875 }
 876
 877