mdoc.c

   1 /*      $Id: mdoc.c,v 1.162 2010/08/08 14:51:32 schwarze 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               macrowarn(struct mdoc *, int,
 102                                 const char *, int);
 103
 104
 105 const struct mdoc_node *
 106 mdoc_node(const struct mdoc *m)
 107 {
 108
 109         return(MDOC_HALT & m->flags ? NULL : m->first);
 110 }
 111
 112
 113 const struct mdoc_meta *
 114 mdoc_meta(const struct mdoc *m)
 115 {
 116
 117         return(MDOC_HALT & m->flags ? NULL : &m->meta);
 118 }
 119
 120
 121 /*
 122  * Frees volatile resources (parse tree, meta-data, fields).
 123  */
 124 static void
 125 mdoc_free1(struct mdoc *mdoc)
 126 {
 127
 128         if (mdoc->first)
 129                 mdoc_node_delete(mdoc, mdoc->first);
 130         if (mdoc->meta.title)
 131                 free(mdoc->meta.title);
 132         if (mdoc->meta.os)
 133                 free(mdoc->meta.os);
 134         if (mdoc->meta.name)
 135                 free(mdoc->meta.name);
 136         if (mdoc->meta.arch)
 137                 free(mdoc->meta.arch);
 138         if (mdoc->meta.vol)
 139                 free(mdoc->meta.vol);
 140         if (mdoc->meta.msec)
 141                 free(mdoc->meta.msec);
 142 }
 143
 144
 145 /*
 146  * Allocate all volatile resources (parse tree, meta-data, fields).
 147  */
 148 static void
 149 mdoc_alloc1(struct mdoc *mdoc)
 150 {
 151
 152         memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 153         mdoc->flags = 0;
 154         mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
 155         mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 156         mdoc->first = mdoc->last;
 157         mdoc->last->type = MDOC_ROOT;
 158         mdoc->next = MDOC_NEXT_CHILD;
 159 }
 160
 161
 162 /*
 163  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 164  * data with mdoc_alloc1().  After invocation, parse data has been reset
 165  * and the parser is ready for re-invocation on a new tree; however,
 166  * cross-parse non-volatile data is kept intact.
 167  */
 168 void
 169 mdoc_reset(struct mdoc *mdoc)
 170 {
 171
 172         mdoc_free1(mdoc);
 173         mdoc_alloc1(mdoc);
 174 }
 175
 176
 177 /*
 178  * Completely free up all volatile and non-volatile parse resources.
 179  * After invocation, the pointer is no longer usable.
 180  */
 181 void
 182 mdoc_free(struct mdoc *mdoc)
 183 {
 184
 185         mdoc_free1(mdoc);
 186         free(mdoc);
 187 }
 188
 189
 190 /*
 191  * Allocate volatile and non-volatile parse resources.
 192  */
 193 struct mdoc *
 194 mdoc_alloc(struct regset *regs, void *data,
 195                 int pflags, 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->pflags = pflags;
 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         if (MDOC_HALT & m->flags)
 221                 return(0);
 222         else if (mdoc_macroend(m))
 223                 return(1);
 224         m->flags |= MDOC_HALT;
 225         return(0);
 226 }
 227
 228
 229 /*
 230  * Main parse routine.  Parses a single line -- really just hands off to
 231  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 232  */
 233 int
 234 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 235 {
 236
 237         if (MDOC_HALT & m->flags)
 238                 return(0);
 239
 240         m->flags |= MDOC_NEWLINE;
 241
 242         /*
 243          * Let the roff nS register switch SYNOPSIS mode early,
 244          * such that the parser knows at all times
 245          * whether this mode is on or off.
 246          * Note that this mode is also switched by the Sh macro.
 247          */
 248         if (m->regs->regs[(int)REG_nS].set) {
 249                 if (m->regs->regs[(int)REG_nS].v.u)
 250                         m->flags |= MDOC_SYNOPSIS;
 251                 else
 252                         m->flags &= ~MDOC_SYNOPSIS;
 253         }
 254
 255         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 256                         mdoc_pmacro(m, ln, buf, offs) :
 257                         mdoc_ptext(m, ln, buf, offs));
 258 }
 259
 260
 261 int
 262 mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
 263                 int ln, int pos, const char *fmt, ...)
 264 {
 265         char             buf[256];
 266         va_list          ap;
 267
 268         va_start(ap, fmt);
 269         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 270         va_end(ap);
 271
 272         return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
 273 }
 274
 275
 276 int
 277 mdoc_macro(MACRO_PROT_ARGS)
 278 {
 279         assert(tok < MDOC_MAX);
 280
 281         /* If we're in the body, deny prologue calls. */
 282
 283         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 284                         MDOC_PBODY & m->flags)
 285                 return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
 286
 287         /* If we're in the prologue, deny "body" macros.  */
 288
 289         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 290                         ! (MDOC_PBODY & m->flags)) {
 291                 if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
 292                         return(0);
 293                 if (NULL == m->meta.title)
 294                         m->meta.title = mandoc_strdup("UNKNOWN");
 295                 if (NULL == m->meta.vol)
 296                         m->meta.vol = mandoc_strdup("LOCAL");
 297                 if (NULL == m->meta.os)
 298                         m->meta.os = mandoc_strdup("LOCAL");
 299                 if (0 == m->meta.date)
 300                         m->meta.date = time(NULL);
 301                 m->flags |= MDOC_PBODY;
 302         }
 303
 304         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 305 }
 306
 307
 308 static int
 309 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 310 {
 311
 312         assert(mdoc->last);
 313         assert(mdoc->first);
 314         assert(MDOC_ROOT != p->type);
 315
 316         switch (mdoc->next) {
 317         case (MDOC_NEXT_SIBLING):
 318                 mdoc->last->next = p;
 319                 p->prev = mdoc->last;
 320                 p->parent = mdoc->last->parent;
 321                 break;
 322         case (MDOC_NEXT_CHILD):
 323                 mdoc->last->child = p;
 324                 p->parent = mdoc->last;
 325                 break;
 326         default:
 327                 abort();
 328                 /* NOTREACHED */
 329         }
 330
 331         p->parent->nchild++;
 332
 333         if ( ! mdoc_valid_pre(mdoc, p))
 334                 return(0);
 335         if ( ! mdoc_action_pre(mdoc, p))
 336                 return(0);
 337
 338         switch (p->type) {
 339         case (MDOC_HEAD):
 340                 assert(MDOC_BLOCK == p->parent->type);
 341                 p->parent->head = p;
 342                 break;
 343         case (MDOC_TAIL):
 344                 assert(MDOC_BLOCK == p->parent->type);
 345                 p->parent->tail = p;
 346                 break;
 347         case (MDOC_BODY):
 348                 if (p->end)
 349                         break;
 350                 assert(MDOC_BLOCK == p->parent->type);
 351                 p->parent->body = p;
 352                 break;
 353         default:
 354                 break;
 355         }
 356
 357         mdoc->last = p;
 358
 359         switch (p->type) {
 360         case (MDOC_TEXT):
 361                 if ( ! mdoc_valid_post(mdoc))
 362                         return(0);
 363                 if ( ! mdoc_action_post(mdoc))
 364                         return(0);
 365                 break;
 366         default:
 367                 break;
 368         }
 369
 370         return(1);
 371 }
 372
 373
 374 static struct mdoc_node *
 375 node_alloc(struct mdoc *m, int line, int pos,
 376                 enum mdoct tok, enum mdoc_type type)
 377 {
 378         struct mdoc_node *p;
 379
 380         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 381         p->sec = m->lastsec;
 382         p->line = line;
 383         p->pos = pos;
 384         p->tok = tok;
 385         p->type = type;
 386
 387         /* Flag analysis. */
 388
 389         if (MDOC_SYNOPSIS & m->flags)
 390                 p->flags |= MDOC_SYNPRETTY;
 391         else
 392                 p->flags &= ~MDOC_SYNPRETTY;
 393         if (MDOC_NEWLINE & m->flags)
 394                 p->flags |= MDOC_LINE;
 395         m->flags &= ~MDOC_NEWLINE;
 396
 397         return(p);
 398 }
 399
 400
 401 int
 402 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 403 {
 404         struct mdoc_node *p;
 405
 406         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 407         if ( ! node_append(m, p))
 408                 return(0);
 409         m->next = MDOC_NEXT_CHILD;
 410         return(1);
 411 }
 412
 413
 414 int
 415 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 416 {
 417         struct mdoc_node *p;
 418
 419         assert(m->first);
 420         assert(m->last);
 421
 422         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 423         if ( ! node_append(m, p))
 424                 return(0);
 425         m->next = MDOC_NEXT_CHILD;
 426         return(1);
 427 }
 428
 429
 430 int
 431 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 432 {
 433         struct mdoc_node *p;
 434
 435         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 436         if ( ! node_append(m, p))
 437                 return(0);
 438         m->next = MDOC_NEXT_CHILD;
 439         return(1);
 440 }
 441
 442
 443 int
 444 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 445                 struct mdoc_node *body, enum mdoc_endbody end)
 446 {
 447         struct mdoc_node *p;
 448
 449         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 450         p->pending = body;
 451         p->end = end;
 452         if ( ! node_append(m, p))
 453                 return(0);
 454         m->next = MDOC_NEXT_SIBLING;
 455         return(1);
 456 }
 457
 458
 459 int
 460 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 461                 enum mdoct tok, struct mdoc_arg *args)
 462 {
 463         struct mdoc_node *p;
 464
 465         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 466         p->args = args;
 467         if (p->args)
 468                 (args->refcnt)++;
 469         if ( ! node_append(m, p))
 470                 return(0);
 471         m->next = MDOC_NEXT_CHILD;
 472         return(1);
 473 }
 474
 475
 476 int
 477 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 478                 enum mdoct tok, struct mdoc_arg *args)
 479 {
 480         struct mdoc_node *p;
 481
 482         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 483         p->args = args;
 484         if (p->args)
 485                 (args->refcnt)++;
 486         if ( ! node_append(m, p))
 487                 return(0);
 488         m->next = MDOC_NEXT_CHILD;
 489         return(1);
 490 }
 491
 492
 493 int
 494 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 495 {
 496         struct mdoc_node *n;
 497         size_t            sv, len;
 498
 499         len = strlen(p);
 500
 501         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 502         n->string = mandoc_malloc(len + 1);
 503         sv = strlcpy(n->string, p, len + 1);
 504
 505         /* Prohibit truncation. */
 506         assert(sv < len + 1);
 507
 508         if ( ! node_append(m, n))
 509                 return(0);
 510
 511         m->next = MDOC_NEXT_SIBLING;
 512         return(1);
 513 }
 514
 515
 516 static void
 517 mdoc_node_free(struct mdoc_node *p)
 518 {
 519
 520         /*
 521          * XXX: if these end up being problematic in terms of memory
 522          * management and dereferencing freed blocks, then make them
 523          * into reference-counted double-pointers.
 524          */
 525
 526         if (MDOC_Bd == p->tok && MDOC_BLOCK == p->type)
 527                 if (p->data.Bd)
 528                         free(p->data.Bd);
 529         if (MDOC_Bl == p->tok && MDOC_BLOCK == p->type)
 530                 if (p->data.Bl)
 531                         free(p->data.Bl);
 532         if (MDOC_Bf == p->tok && MDOC_HEAD == p->type)
 533                 if (p->data.Bf)
 534                         free(p->data.Bf);
 535
 536         if (p->string)
 537                 free(p->string);
 538         if (p->args)
 539                 mdoc_argv_free(p->args);
 540         free(p);
 541 }
 542
 543
 544 static void
 545 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 546 {
 547
 548         /* Adjust siblings. */
 549
 550         if (n->prev)
 551                 n->prev->next = n->next;
 552         if (n->next)
 553                 n->next->prev = n->prev;
 554
 555         /* Adjust parent. */
 556
 557         if (n->parent) {
 558                 n->parent->nchild--;
 559                 if (n->parent->child == n)
 560                         n->parent->child = n->prev ? n->prev : n->next;
 561         }
 562
 563         /* Adjust parse point, if applicable. */
 564
 565         if (m && m->last == n) {
 566                 if (n->prev) {
 567                         m->last = n->prev;
 568                         m->next = MDOC_NEXT_SIBLING;
 569                 } else {
 570                         m->last = n->parent;
 571                         m->next = MDOC_NEXT_CHILD;
 572                 }
 573         }
 574
 575         if (m && m->first == n)
 576                 m->first = NULL;
 577 }
 578
 579
 580 void
 581 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 582 {
 583
 584         while (p->child) {
 585                 assert(p->nchild);
 586                 mdoc_node_delete(m, p->child);
 587         }
 588         assert(0 == p->nchild);
 589
 590         mdoc_node_unlink(m, p);
 591         mdoc_node_free(p);
 592 }
 593
 594
 595 /*
 596  * Parse free-form text, that is, a line that does not begin with the
 597  * control character.
 598  */
 599 static int
 600 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 601 {
 602         char             *c, *ws, *end;
 603         struct mdoc_node *n;
 604
 605         /* Ignore bogus comments. */
 606
 607         if ('\\' == buf[offs] &&
 608                         '.' == buf[offs + 1] &&
 609                         '"' == buf[offs + 2])
 610                 return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
 611
 612         /* No text before an initial macro. */
 613
 614         if (SEC_NONE == m->lastnamed)
 615                 return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
 616
 617         assert(m->last);
 618         n = m->last;
 619
 620         /*
 621          * Divert directly to list processing if we're encountering a
 622          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 623          * (a MDOC_BODY means it's already open, in which case we should
 624          * process within its context in the normal way).
 625          */
 626
 627         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 628                         LIST_column == n->data.Bl->type) {
 629                 /* `Bl' is open without any children. */
 630                 m->flags |= MDOC_FREECOL;
 631                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 632         }
 633
 634         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 635                         NULL != n->parent &&
 636                         MDOC_Bl == n->parent->tok &&
 637                         LIST_column == n->parent->data.Bl->type) {
 638                 /* `Bl' has block-level `It' children. */
 639                 m->flags |= MDOC_FREECOL;
 640                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 641         }
 642
 643         /*
 644          * Search for the beginning of unescaped trailing whitespace (ws)
 645          * and for the first character not to be output (end).
 646          */
 647
 648         /* FIXME: replace with strcspn(). */
 649         ws = NULL;
 650         for (c = end = buf + offs; *c; c++) {
 651                 switch (*c) {
 652                 case '-':
 653                         if (mandoc_hyph(buf + offs, c))
 654                                 *c = ASCII_HYPH;
 655                         ws = NULL;
 656                         break;
 657                 case ' ':
 658                         if (NULL == ws)
 659                                 ws = c;
 660                         continue;
 661                 case '\t':
 662                         /*
 663                          * Always warn about trailing tabs,
 664                          * even outside literal context,
 665                          * where they should be put on the next line.
 666                          */
 667                         if (NULL == ws)
 668                                 ws = c;
 669                         /*
 670                          * Strip trailing tabs in literal context only;
 671                          * outside, they affect the next line.
 672                          */
 673                         if (MDOC_LITERAL & m->flags)
 674                                 continue;
 675                         break;
 676                 case '\\':
 677                         /* Skip the escaped character, too, if any. */
 678                         if (c[1])
 679                                 c++;
 680                         /* FALLTHROUGH */
 681                 default:
 682                         ws = NULL;
 683                         break;
 684                 }
 685                 end = c + 1;
 686         }
 687         *end = '\0';
 688
 689         if (ws)
 690                 if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
 691                         return(0);
 692
 693         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 694                 if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
 695                         return(0);
 696
 697                 /*
 698                  * Insert a `Pp' in the case of a blank line.  Technically,
 699                  * blank lines aren't allowed, but enough manuals assume this
 700                  * behaviour that we want to work around it.
 701                  */
 702                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL))
 703                         return(0);
 704
 705                 m->next = MDOC_NEXT_SIBLING;
 706                 return(1);
 707         }
 708
 709         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 710                 return(0);
 711
 712         if (MDOC_LITERAL & m->flags)
 713                 return(1);
 714
 715         /*
 716          * End-of-sentence check.  If the last character is an unescaped
 717          * EOS character, then flag the node as being the end of a
 718          * sentence.  The front-end will know how to interpret this.
 719          */
 720
 721         assert(buf < end);
 722
 723         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 724                 m->last->flags |= MDOC_EOS;
 725
 726         return(1);
 727 }
 728
 729
 730 static int
 731 macrowarn(struct mdoc *m, int ln, const char *buf, int offs)
 732 {
 733         int              rc;
 734
 735         rc = mdoc_vmsg(m, MANDOCERR_MACRO, ln, offs,
 736                         "unknown macro: %s%s",
 737                         buf, strlen(buf) > 3 ? "..." : "");
 738
 739         /* FIXME: logic should be in driver. */
 740         /* FIXME: broken, will error out and not omit a message. */
 741         return(MDOC_IGN_MACRO & m->pflags ? rc : 0);
 742 }
 743
 744
 745 /*
 746  * Parse a macro line, that is, a line beginning with the control
 747  * character.
 748  */
 749 static int
 750 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 751 {
 752         enum mdoct        tok;
 753         int               i, j, sv;
 754         char              mac[5];
 755         struct mdoc_node *n;
 756
 757         /* Empty lines are ignored. */
 758
 759         offs++;
 760
 761         if ('\0' == buf[offs])
 762                 return(1);
 763
 764         i = offs;
 765
 766         /* Accept tabs/whitespace after the initial control char. */
 767
 768         if (' ' == buf[i] || '\t' == buf[i]) {
 769                 i++;
 770                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 771                         i++;
 772                 if ('\0' == buf[i])
 773                         return(1);
 774         }
 775
 776         sv = i;
 777
 778         /*
 779          * Copy the first word into a nil-terminated buffer.
 780          * Stop copying when a tab, space, or eoln is encountered.
 781          */
 782
 783         j = 0;
 784         while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
 785                 mac[j++] = buf[i++];
 786         mac[j] = '\0';
 787
 788         if (j == 4 || j < 2) {
 789                 if ( ! macrowarn(m, ln, mac, sv))
 790                         goto err;
 791                 return(1);
 792         }
 793
 794         if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
 795                 if ( ! macrowarn(m, ln, mac, sv))
 796                         goto err;
 797                 return(1);
 798         }
 799
 800         /* Disregard the first trailing tab, if applicable. */
 801
 802         if ('\t' == buf[i])
 803                 i++;
 804
 805         /* Jump to the next non-whitespace word. */
 806
 807         while (buf[i] && ' ' == buf[i])
 808                 i++;
 809
 810         /*
 811          * Trailing whitespace.  Note that tabs are allowed to be passed
 812          * into the parser as "text", so we only warn about spaces here.
 813          */
 814
 815         if ('\0' == buf[i] && ' ' == buf[i - 1])
 816                 if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
 817                         goto err;
 818
 819         /*
 820          * If an initial macro or a list invocation, divert directly
 821          * into macro processing.
 822          */
 823
 824         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 825                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 826                         goto err;
 827                 return(1);
 828         }
 829
 830         n = m->last;
 831         assert(m->last);
 832
 833         /*
 834          * If the first macro of a `Bl -column', open an `It' block
 835          * context around the parsed macro.
 836          */
 837
 838         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 839                         LIST_column == n->data.Bl->type) {
 840                 m->flags |= MDOC_FREECOL;
 841                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 842                         goto err;
 843                 return(1);
 844         }
 845
 846         /*
 847          * If we're following a block-level `It' within a `Bl -column'
 848          * context (perhaps opened in the above block or in ptext()),
 849          * then open an `It' block context around the parsed macro.
 850          */
 851
 852         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 853                         NULL != n->parent &&
 854                         MDOC_Bl == n->parent->tok &&
 855                         LIST_column == n->parent->data.Bl->type) {
 856                 m->flags |= MDOC_FREECOL;
 857                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 858                         goto err;
 859                 return(1);
 860         }
 861
 862         /* Normal processing of a macro. */
 863
 864         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 865                 goto err;
 866
 867         return(1);
 868
 869 err:    /* Error out. */
 870
 871         m->flags |= MDOC_HALT;
 872         return(0);
 873 }
 874
 875