mdoc.c

   1 /*      $Id: mdoc.c,v 1.160 2010/08/07 20:33:55 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 <ctype.h>
  26 #include <stdarg.h>
  27 #include <stdio.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <time.h>
  31
  32 #include "mandoc.h"
  33 #include "libmdoc.h"
  34 #include "libmandoc.h"
  35
  36 const   char *const __mdoc_macronames[MDOC_MAX] = {
  37         "Ap",           "Dd",           "Dt",           "Os",
  38         "Sh",           "Ss",           "Pp",           "D1",
  39         "Dl",           "Bd",           "Ed",           "Bl",
  40         "El",           "It",           "Ad",           "An",
  41         "Ar",           "Cd",           "Cm",           "Dv",
  42         "Er",           "Ev",           "Ex",           "Fa",
  43         "Fd",           "Fl",           "Fn",           "Ft",
  44         "Ic",           "In",           "Li",           "Nd",
  45         "Nm",           "Op",           "Ot",           "Pa",
  46         "Rv",           "St",           "Va",           "Vt",
  47         /* LINTED */
  48         "Xr",           "%A",           "%B",           "%D",
  49         /* LINTED */
  50         "%I",           "%J",           "%N",           "%O",
  51         /* LINTED */
  52         "%P",           "%R",           "%T",           "%V",
  53         "Ac",           "Ao",           "Aq",           "At",
  54         "Bc",           "Bf",           "Bo",           "Bq",
  55         "Bsx",          "Bx",           "Db",           "Dc",
  56         "Do",           "Dq",           "Ec",           "Ef",
  57         "Em",           "Eo",           "Fx",           "Ms",
  58         "No",           "Ns",           "Nx",           "Ox",
  59         "Pc",           "Pf",           "Po",           "Pq",
  60         "Qc",           "Ql",           "Qo",           "Qq",
  61         "Re",           "Rs",           "Sc",           "So",
  62         "Sq",           "Sm",           "Sx",           "Sy",
  63         "Tn",           "Ux",           "Xc",           "Xo",
  64         "Fo",           "Fc",           "Oo",           "Oc",
  65         "Bk",           "Ek",           "Bt",           "Hf",
  66         "Fr",           "Ud",           "Lb",           "Lp",
  67         "Lk",           "Mt",           "Brq",          "Bro",
  68         /* LINTED */
  69         "Brc",          "%C",           "Es",           "En",
  70         /* LINTED */
  71         "Dx",           "%Q",           "br",           "sp",
  72         /* LINTED */
  73         "%U",           "Ta"
  74         };
  75
  76 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  77         "split",                "nosplit",              "ragged",
  78         "unfilled",             "literal",              "file",
  79         "offset",               "bullet",               "dash",
  80         "hyphen",               "item",                 "enum",
  81         "tag",                  "diag",                 "hang",
  82         "ohang",                "inset",                "column",
  83         "width",                "compact",              "std",
  84         "filled",               "words",                "emphasis",
  85         "symbolic",             "nested",               "centered"
  86         };
  87
  88 const   char * const *mdoc_macronames = __mdoc_macronames;
  89 const   char * const *mdoc_argnames = __mdoc_argnames;
  90
  91 static  void              mdoc_node_free(struct mdoc_node *);
  92 static  void              mdoc_node_unlink(struct mdoc *,
  93                                 struct mdoc_node *);
  94 static  void              mdoc_free1(struct mdoc *);
  95 static  void              mdoc_alloc1(struct mdoc *);
  96 static  struct mdoc_node *node_alloc(struct mdoc *, int, int,
  97                                 enum mdoct, enum mdoc_type);
  98 static  int               node_append(struct mdoc *,
  99                                 struct mdoc_node *);
 100 static  int               mdoc_ptext(struct mdoc *, int, char *, int);
 101 static  int               mdoc_pmacro(struct mdoc *, int, char *, int);
 102 static  int               macrowarn(struct mdoc *, int,
 103                                 const char *, int);
 104
 105
 106 const struct mdoc_node *
 107 mdoc_node(const struct mdoc *m)
 108 {
 109
 110         return(MDOC_HALT & m->flags ? NULL : m->first);
 111 }
 112
 113
 114 const struct mdoc_meta *
 115 mdoc_meta(const struct mdoc *m)
 116 {
 117
 118         return(MDOC_HALT & m->flags ? NULL : &m->meta);
 119 }
 120
 121
 122 /*
 123  * Frees volatile resources (parse tree, meta-data, fields).
 124  */
 125 static void
 126 mdoc_free1(struct mdoc *mdoc)
 127 {
 128
 129         if (mdoc->first)
 130                 mdoc_node_delete(mdoc, mdoc->first);
 131         if (mdoc->meta.title)
 132                 free(mdoc->meta.title);
 133         if (mdoc->meta.os)
 134                 free(mdoc->meta.os);
 135         if (mdoc->meta.name)
 136                 free(mdoc->meta.name);
 137         if (mdoc->meta.arch)
 138                 free(mdoc->meta.arch);
 139         if (mdoc->meta.vol)
 140                 free(mdoc->meta.vol);
 141         if (mdoc->meta.msec)
 142                 free(mdoc->meta.msec);
 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,
 196                 int pflags, 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->pflags = pflags;
 205         p->regs = regs;
 206
 207         mdoc_hash_init();
 208         mdoc_alloc1(p);
 209         return(p);
 210 }
 211
 212
 213 /*
 214  * Climb back up the parse tree, validating open scopes.  Mostly calls
 215  * through to macro_end() in macro.c.
 216  */
 217 int
 218 mdoc_endparse(struct mdoc *m)
 219 {
 220
 221         if (MDOC_HALT & m->flags)
 222                 return(0);
 223         else if (mdoc_macroend(m))
 224                 return(1);
 225         m->flags |= MDOC_HALT;
 226         return(0);
 227 }
 228
 229
 230 /*
 231  * Main parse routine.  Parses a single line -- really just hands off to
 232  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 233  */
 234 int
 235 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 236 {
 237
 238         if (MDOC_HALT & m->flags)
 239                 return(0);
 240
 241         m->flags |= MDOC_NEWLINE;
 242
 243         /*
 244          * Let the roff nS register switch SYNOPSIS mode early,
 245          * such that the parser knows at all times
 246          * whether this mode is on or off.
 247          * Note that this mode is also switched by the Sh macro.
 248          */
 249         if (m->regs->regs[(int)REG_nS].set) {
 250                 if (m->regs->regs[(int)REG_nS].v.u)
 251                         m->flags |= MDOC_SYNOPSIS;
 252                 else
 253                         m->flags &= ~MDOC_SYNOPSIS;
 254         }
 255
 256         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 257                         mdoc_pmacro(m, ln, buf, offs) :
 258                         mdoc_ptext(m, ln, buf, offs));
 259 }
 260
 261
 262 int
 263 mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
 264                 int ln, int pos, const char *fmt, ...)
 265 {
 266         char             buf[256];
 267         va_list          ap;
 268
 269         va_start(ap, fmt);
 270         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 271         va_end(ap);
 272
 273         return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
 274 }
 275
 276
 277 int
 278 mdoc_macro(MACRO_PROT_ARGS)
 279 {
 280         assert(tok < MDOC_MAX);
 281
 282         /* If we're in the body, deny prologue calls. */
 283
 284         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 285                         MDOC_PBODY & m->flags)
 286                 return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
 287
 288         /* If we're in the prologue, deny "body" macros.  */
 289
 290         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 291                         ! (MDOC_PBODY & m->flags)) {
 292                 if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
 293                         return(0);
 294                 if (NULL == m->meta.title)
 295                         m->meta.title = mandoc_strdup("UNKNOWN");
 296                 if (NULL == m->meta.vol)
 297                         m->meta.vol = mandoc_strdup("LOCAL");
 298                 if (NULL == m->meta.os)
 299                         m->meta.os = mandoc_strdup("LOCAL");
 300                 if (0 == m->meta.date)
 301                         m->meta.date = time(NULL);
 302                 m->flags |= MDOC_PBODY;
 303         }
 304
 305         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 306 }
 307
 308
 309 static int
 310 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 311 {
 312
 313         assert(mdoc->last);
 314         assert(mdoc->first);
 315         assert(MDOC_ROOT != p->type);
 316
 317         switch (mdoc->next) {
 318         case (MDOC_NEXT_SIBLING):
 319                 mdoc->last->next = p;
 320                 p->prev = mdoc->last;
 321                 p->parent = mdoc->last->parent;
 322                 break;
 323         case (MDOC_NEXT_CHILD):
 324                 mdoc->last->child = p;
 325                 p->parent = mdoc->last;
 326                 break;
 327         default:
 328                 abort();
 329                 /* NOTREACHED */
 330         }
 331
 332         p->parent->nchild++;
 333
 334         if ( ! mdoc_valid_pre(mdoc, p))
 335                 return(0);
 336         if ( ! mdoc_action_pre(mdoc, p))
 337                 return(0);
 338
 339         switch (p->type) {
 340         case (MDOC_HEAD):
 341                 assert(MDOC_BLOCK == p->parent->type);
 342                 p->parent->head = p;
 343                 break;
 344         case (MDOC_TAIL):
 345                 assert(MDOC_BLOCK == p->parent->type);
 346                 p->parent->tail = p;
 347                 break;
 348         case (MDOC_BODY):
 349                 if (p->end)
 350                         break;
 351                 assert(MDOC_BLOCK == p->parent->type);
 352                 p->parent->body = p;
 353                 break;
 354         default:
 355                 break;
 356         }
 357
 358         mdoc->last = p;
 359
 360         switch (p->type) {
 361         case (MDOC_TEXT):
 362                 if ( ! mdoc_valid_post(mdoc))
 363                         return(0);
 364                 if ( ! mdoc_action_post(mdoc))
 365                         return(0);
 366                 break;
 367         default:
 368                 break;
 369         }
 370
 371         return(1);
 372 }
 373
 374
 375 static struct mdoc_node *
 376 node_alloc(struct mdoc *m, int line, int pos,
 377                 enum mdoct tok, enum mdoc_type type)
 378 {
 379         struct mdoc_node *p;
 380
 381         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 382         p->sec = m->lastsec;
 383         p->line = line;
 384         p->pos = pos;
 385         p->tok = tok;
 386         p->type = type;
 387
 388         /* Flag analysis. */
 389
 390         if (MDOC_SYNOPSIS & m->flags)
 391                 p->flags |= MDOC_SYNPRETTY;
 392         else
 393                 p->flags &= ~MDOC_SYNPRETTY;
 394         if (MDOC_NEWLINE & m->flags)
 395                 p->flags |= MDOC_LINE;
 396         m->flags &= ~MDOC_NEWLINE;
 397
 398         return(p);
 399 }
 400
 401
 402 int
 403 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 404 {
 405         struct mdoc_node *p;
 406
 407         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 408         if ( ! node_append(m, p))
 409                 return(0);
 410         m->next = MDOC_NEXT_CHILD;
 411         return(1);
 412 }
 413
 414
 415 int
 416 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 417 {
 418         struct mdoc_node *p;
 419
 420         assert(m->first);
 421         assert(m->last);
 422
 423         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 424         if ( ! node_append(m, p))
 425                 return(0);
 426         m->next = MDOC_NEXT_CHILD;
 427         return(1);
 428 }
 429
 430
 431 int
 432 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 433 {
 434         struct mdoc_node *p;
 435
 436         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 437         if ( ! node_append(m, p))
 438                 return(0);
 439         m->next = MDOC_NEXT_CHILD;
 440         return(1);
 441 }
 442
 443
 444 int
 445 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 446                 struct mdoc_node *body, enum mdoc_endbody end)
 447 {
 448         struct mdoc_node *p;
 449
 450         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 451         p->pending = body;
 452         p->end = end;
 453         if ( ! node_append(m, p))
 454                 return(0);
 455         m->next = MDOC_NEXT_SIBLING;
 456         return(1);
 457 }
 458
 459
 460 int
 461 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 462                 enum mdoct tok, struct mdoc_arg *args)
 463 {
 464         struct mdoc_node *p;
 465
 466         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 467         p->args = args;
 468         if (p->args)
 469                 (args->refcnt)++;
 470         if ( ! node_append(m, p))
 471                 return(0);
 472         m->next = MDOC_NEXT_CHILD;
 473         return(1);
 474 }
 475
 476
 477 int
 478 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 479                 enum mdoct tok, struct mdoc_arg *args)
 480 {
 481         struct mdoc_node *p;
 482
 483         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 484         p->args = args;
 485         if (p->args)
 486                 (args->refcnt)++;
 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_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 496 {
 497         struct mdoc_node *n;
 498         size_t            sv, len;
 499
 500         len = strlen(p);
 501
 502         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 503         n->string = mandoc_malloc(len + 1);
 504         sv = strlcpy(n->string, p, len + 1);
 505
 506         /* Prohibit truncation. */
 507         assert(sv < len + 1);
 508
 509         if ( ! node_append(m, n))
 510                 return(0);
 511
 512         m->next = MDOC_NEXT_SIBLING;
 513         return(1);
 514 }
 515
 516
 517 static void
 518 mdoc_node_free(struct mdoc_node *p)
 519 {
 520
 521         /*
 522          * XXX: if these end up being problematic in terms of memory
 523          * management and dereferencing freed blocks, then make them
 524          * into reference-counted double-pointers.
 525          */
 526
 527         if (MDOC_Bd == p->tok && MDOC_BLOCK == p->type)
 528                 if (p->data.Bd)
 529                         free(p->data.Bd);
 530         if (MDOC_Bl == p->tok && MDOC_BLOCK == p->type)
 531                 if (p->data.Bl)
 532                         free(p->data.Bl);
 533         if (MDOC_Bf == p->tok && MDOC_HEAD == p->type)
 534                 if (p->data.Bf)
 535                         free(p->data.Bf);
 536
 537         if (p->string)
 538                 free(p->string);
 539         if (p->args)
 540                 mdoc_argv_free(p->args);
 541         free(p);
 542 }
 543
 544
 545 static void
 546 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 547 {
 548
 549         /* Adjust siblings. */
 550
 551         if (n->prev)
 552                 n->prev->next = n->next;
 553         if (n->next)
 554                 n->next->prev = n->prev;
 555
 556         /* Adjust parent. */
 557
 558         if (n->parent) {
 559                 n->parent->nchild--;
 560                 if (n->parent->child == n)
 561                         n->parent->child = n->prev ? n->prev : n->next;
 562         }
 563
 564         /* Adjust parse point, if applicable. */
 565
 566         if (m && m->last == n) {
 567                 if (n->prev) {
 568                         m->last = n->prev;
 569                         m->next = MDOC_NEXT_SIBLING;
 570                 } else {
 571                         m->last = n->parent;
 572                         m->next = MDOC_NEXT_CHILD;
 573                 }
 574         }
 575
 576         if (m && m->first == n)
 577                 m->first = NULL;
 578 }
 579
 580
 581 void
 582 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 583 {
 584
 585         while (p->child) {
 586                 assert(p->nchild);
 587                 mdoc_node_delete(m, p->child);
 588         }
 589         assert(0 == p->nchild);
 590
 591         mdoc_node_unlink(m, p);
 592         mdoc_node_free(p);
 593 }
 594
 595
 596 /*
 597  * Parse free-form text, that is, a line that does not begin with the
 598  * control character.
 599  */
 600 static int
 601 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 602 {
 603         char             *c, *ws, *end;
 604         struct mdoc_node *n;
 605
 606         /* Ignore bogus comments. */
 607
 608         if ('\\' == buf[offs] &&
 609                         '.' == buf[offs + 1] &&
 610                         '"' == buf[offs + 2])
 611                 return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
 612
 613         /* No text before an initial macro. */
 614
 615         if (SEC_NONE == m->lastnamed)
 616                 return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
 617
 618         assert(m->last);
 619         n = m->last;
 620
 621         /*
 622          * Divert directly to list processing if we're encountering a
 623          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 624          * (a MDOC_BODY means it's already open, in which case we should
 625          * process within its context in the normal way).
 626          */
 627
 628         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 629                         LIST_column == n->data.Bl->type) {
 630                 /* `Bl' is open without any children. */
 631                 m->flags |= MDOC_FREECOL;
 632                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 633         }
 634
 635         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 636                         NULL != n->parent &&
 637                         MDOC_Bl == n->parent->tok &&
 638                         LIST_column == n->parent->data.Bl->type) {
 639                 /* `Bl' has block-level `It' children. */
 640                 m->flags |= MDOC_FREECOL;
 641                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 642         }
 643
 644         /*
 645          * Search for the beginning of unescaped trailing whitespace (ws)
 646          * and for the first character not to be output (end).
 647          */
 648
 649         /* FIXME: replace with strcspn(). */
 650         ws = NULL;
 651         for (c = end = buf + offs; *c; c++) {
 652                 switch (*c) {
 653                 case '-':
 654                         if (mandoc_hyph(buf + offs, c))
 655                                 *c = ASCII_HYPH;
 656                         ws = NULL;
 657                         break;
 658                 case ' ':
 659                         if (NULL == ws)
 660                                 ws = c;
 661                         continue;
 662                 case '\t':
 663                         /*
 664                          * Always warn about trailing tabs,
 665                          * even outside literal context,
 666                          * where they should be put on the next line.
 667                          */
 668                         if (NULL == ws)
 669                                 ws = c;
 670                         /*
 671                          * Strip trailing tabs in literal context only;
 672                          * outside, they affect the next line.
 673                          */
 674                         if (MDOC_LITERAL & m->flags)
 675                                 continue;
 676                         break;
 677                 case '\\':
 678                         /* Skip the escaped character, too, if any. */
 679                         if (c[1])
 680                                 c++;
 681                         /* FALLTHROUGH */
 682                 default:
 683                         ws = NULL;
 684                         break;
 685                 }
 686                 end = c + 1;
 687         }
 688         *end = '\0';
 689
 690         if (ws)
 691                 if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
 692                         return(0);
 693
 694         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 695                 if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
 696                         return(0);
 697
 698                 /*
 699                  * Insert a `Pp' in the case of a blank line.  Technically,
 700                  * blank lines aren't allowed, but enough manuals assume this
 701                  * behaviour that we want to work around it.
 702                  */
 703                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL))
 704                         return(0);
 705
 706                 m->next = MDOC_NEXT_SIBLING;
 707                 return(1);
 708         }
 709
 710         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 711                 return(0);
 712
 713         if (MDOC_LITERAL & m->flags)
 714                 return(1);
 715
 716         /*
 717          * End-of-sentence check.  If the last character is an unescaped
 718          * EOS character, then flag the node as being the end of a
 719          * sentence.  The front-end will know how to interpret this.
 720          */
 721
 722         assert(buf < end);
 723
 724         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 725                 m->last->flags |= MDOC_EOS;
 726
 727         return(1);
 728 }
 729
 730
 731 static int
 732 macrowarn(struct mdoc *m, int ln, const char *buf, int offs)
 733 {
 734         int              rc;
 735
 736         rc = mdoc_vmsg(m, MANDOCERR_MACRO, ln, offs,
 737                         "unknown macro: %s%s",
 738                         buf, strlen(buf) > 3 ? "..." : "");
 739
 740         /* FIXME: logic should be in driver. */
 741         /* FIXME: broken, will error out and not omit a message. */
 742         return(MDOC_IGN_MACRO & m->pflags ? rc : 0);
 743 }
 744
 745
 746 /*
 747  * Parse a macro line, that is, a line beginning with the control
 748  * character.
 749  */
 750 static int
 751 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 752 {
 753         enum mdoct        tok;
 754         int               i, j, sv;
 755         char              mac[5];
 756         struct mdoc_node *n;
 757
 758         /* Empty lines are ignored. */
 759
 760         offs++;
 761
 762         if ('\0' == buf[offs])
 763                 return(1);
 764
 765         i = offs;
 766
 767         /* Accept tabs/whitespace after the initial control char. */
 768
 769         if (' ' == buf[i] || '\t' == buf[i]) {
 770                 i++;
 771                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 772                         i++;
 773                 if ('\0' == buf[i])
 774                         return(1);
 775         }
 776
 777         sv = i;
 778
 779         /*
 780          * Copy the first word into a nil-terminated buffer.  Stop
 781          * copying when a tab, space, or eoln is encountered.
 782          */
 783
 784         for (j = 0; j < 4; j++, i++) {
 785                 if ('\0' == (mac[j] = buf[i]))
 786                         break;
 787                 else if (' ' == buf[i] || '\t' == buf[i])
 788                         break;
 789
 790                 /* Check for invalid characters. */
 791                 /* TODO: remove me, already done in main.c. */
 792
 793                 if (isgraph((u_char)buf[i]))
 794                         continue;
 795                 if ( ! mdoc_pmsg(m, ln, i, MANDOCERR_BADCHAR))
 796                         return(0);
 797                 i--;
 798         }
 799
 800         mac[j] = '\0';
 801
 802         if (j == 4 || j < 2) {
 803                 if ( ! macrowarn(m, ln, mac, sv))
 804                         goto err;
 805                 return(1);
 806         }
 807
 808         if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
 809                 if ( ! macrowarn(m, ln, mac, sv))
 810                         goto err;
 811                 return(1);
 812         }
 813
 814         /* Disregard the first trailing tab, if applicable. */
 815
 816         if ('\t' == buf[i])
 817                 i++;
 818
 819         /* Jump to the next non-whitespace word. */
 820
 821         while (buf[i] && ' ' == buf[i])
 822                 i++;
 823
 824         /*
 825          * Trailing whitespace.  Note that tabs are allowed to be passed
 826          * into the parser as "text", so we only warn about spaces here.
 827          */
 828
 829         if ('\0' == buf[i] && ' ' == buf[i - 1])
 830                 if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
 831                         goto err;
 832
 833         /*
 834          * If an initial macro or a list invocation, divert directly
 835          * into macro processing.
 836          */
 837
 838         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 839                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 840                         goto err;
 841                 return(1);
 842         }
 843
 844         n = m->last;
 845         assert(m->last);
 846
 847         /*
 848          * If the first macro of a `Bl -column', open an `It' block
 849          * context around the parsed macro.
 850          */
 851
 852         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 853                         LIST_column == n->data.Bl->type) {
 854                 m->flags |= MDOC_FREECOL;
 855                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 856                         goto err;
 857                 return(1);
 858         }
 859
 860         /*
 861          * If we're following a block-level `It' within a `Bl -column'
 862          * context (perhaps opened in the above block or in ptext()),
 863          * then open an `It' block context around the parsed macro.
 864          */
 865
 866         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 867                         NULL != n->parent &&
 868                         MDOC_Bl == n->parent->tok &&
 869                         LIST_column == n->parent->data.Bl->type) {
 870                 m->flags |= MDOC_FREECOL;
 871                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 872                         goto err;
 873                 return(1);
 874         }
 875
 876         /* Normal processing of a macro. */
 877
 878         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 879                 goto err;
 880
 881         return(1);
 882
 883 err:    /* Error out. */
 884
 885         m->flags |= MDOC_HALT;
 886         return(0);
 887 }
 888
 889