mdoc.c

   1 /*      $Id: mdoc.c,v 1.149 2010/06/27 15:52:41 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.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 "regs.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(const struct regset *regs,
 196                 void *data, 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         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 243                         mdoc_pmacro(m, ln, buf, offs) :
 244                         mdoc_ptext(m, ln, buf, offs));
 245 }
 246
 247
 248 int
 249 mdoc_vmsg(struct mdoc *mdoc, enum mandocerr t,
 250                 int ln, int pos, const char *fmt, ...)
 251 {
 252         char             buf[256];
 253         va_list          ap;
 254
 255         va_start(ap, fmt);
 256         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 257         va_end(ap);
 258
 259         return((*mdoc->msg)(t, mdoc->data, ln, pos, buf));
 260 }
 261
 262
 263 int
 264 mdoc_macro(MACRO_PROT_ARGS)
 265 {
 266         assert(tok < MDOC_MAX);
 267
 268         /* If we're in the body, deny prologue calls. */
 269
 270         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 271                         MDOC_PBODY & m->flags)
 272                 return(mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY));
 273
 274         /* If we're in the prologue, deny "body" macros.  */
 275
 276         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 277                         ! (MDOC_PBODY & m->flags)) {
 278                 if ( ! mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG))
 279                         return(0);
 280                 if (NULL == m->meta.title)
 281                         m->meta.title = mandoc_strdup("UNKNOWN");
 282                 if (NULL == m->meta.vol)
 283                         m->meta.vol = mandoc_strdup("LOCAL");
 284                 if (NULL == m->meta.os)
 285                         m->meta.os = mandoc_strdup("LOCAL");
 286                 if (0 == m->meta.date)
 287                         m->meta.date = time(NULL);
 288                 m->flags |= MDOC_PBODY;
 289         }
 290
 291         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 292 }
 293
 294
 295 static int
 296 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 297 {
 298
 299         assert(mdoc->last);
 300         assert(mdoc->first);
 301         assert(MDOC_ROOT != p->type);
 302
 303         switch (mdoc->next) {
 304         case (MDOC_NEXT_SIBLING):
 305                 mdoc->last->next = p;
 306                 p->prev = mdoc->last;
 307                 p->parent = mdoc->last->parent;
 308                 break;
 309         case (MDOC_NEXT_CHILD):
 310                 mdoc->last->child = p;
 311                 p->parent = mdoc->last;
 312                 break;
 313         default:
 314                 abort();
 315                 /* NOTREACHED */
 316         }
 317
 318         p->parent->nchild++;
 319
 320         if ( ! mdoc_valid_pre(mdoc, p))
 321                 return(0);
 322         if ( ! mdoc_action_pre(mdoc, p))
 323                 return(0);
 324
 325         switch (p->type) {
 326         case (MDOC_HEAD):
 327                 assert(MDOC_BLOCK == p->parent->type);
 328                 p->parent->head = p;
 329                 break;
 330         case (MDOC_TAIL):
 331                 assert(MDOC_BLOCK == p->parent->type);
 332                 p->parent->tail = p;
 333                 break;
 334         case (MDOC_BODY):
 335                 assert(MDOC_BLOCK == p->parent->type);
 336                 p->parent->body = p;
 337                 break;
 338         default:
 339                 break;
 340         }
 341
 342         mdoc->last = p;
 343
 344         switch (p->type) {
 345         case (MDOC_TEXT):
 346                 if ( ! mdoc_valid_post(mdoc))
 347                         return(0);
 348                 if ( ! mdoc_action_post(mdoc))
 349                         return(0);
 350                 break;
 351         default:
 352                 break;
 353         }
 354
 355         return(1);
 356 }
 357
 358
 359 static struct mdoc_node *
 360 node_alloc(struct mdoc *m, int line, int pos,
 361                 enum mdoct tok, enum mdoc_type type)
 362 {
 363         struct mdoc_node *p;
 364
 365         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 366         p->sec = m->lastsec;
 367         p->line = line;
 368         p->pos = pos;
 369         p->tok = tok;
 370         p->type = type;
 371         if (MDOC_NEWLINE & m->flags)
 372                 p->flags |= MDOC_LINE;
 373         m->flags &= ~MDOC_NEWLINE;
 374         return(p);
 375 }
 376
 377
 378 int
 379 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 380 {
 381         struct mdoc_node *p;
 382
 383         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 384         if ( ! node_append(m, p))
 385                 return(0);
 386         m->next = MDOC_NEXT_CHILD;
 387         return(1);
 388 }
 389
 390
 391 int
 392 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 393 {
 394         struct mdoc_node *p;
 395
 396         assert(m->first);
 397         assert(m->last);
 398
 399         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 400         if ( ! node_append(m, p))
 401                 return(0);
 402         m->next = MDOC_NEXT_CHILD;
 403         return(1);
 404 }
 405
 406
 407 int
 408 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 409 {
 410         struct mdoc_node *p;
 411
 412         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 413         if ( ! node_append(m, p))
 414                 return(0);
 415         m->next = MDOC_NEXT_CHILD;
 416         return(1);
 417 }
 418
 419
 420 int
 421 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 422                 enum mdoct tok, struct mdoc_arg *args)
 423 {
 424         struct mdoc_node *p;
 425
 426         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 427         p->args = args;
 428         if (p->args)
 429                 (args->refcnt)++;
 430         if ( ! node_append(m, p))
 431                 return(0);
 432         m->next = MDOC_NEXT_CHILD;
 433         return(1);
 434 }
 435
 436
 437 int
 438 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 439                 enum mdoct tok, struct mdoc_arg *args)
 440 {
 441         struct mdoc_node *p;
 442
 443         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 444         p->args = args;
 445         if (p->args)
 446                 (args->refcnt)++;
 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_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 456 {
 457         struct mdoc_node *n;
 458         size_t            sv, len;
 459
 460         len = strlen(p);
 461
 462         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 463         n->string = mandoc_malloc(len + 1);
 464         sv = strlcpy(n->string, p, len + 1);
 465
 466         /* Prohibit truncation. */
 467         assert(sv < len + 1);
 468
 469         if ( ! node_append(m, n))
 470                 return(0);
 471
 472         m->next = MDOC_NEXT_SIBLING;
 473         return(1);
 474 }
 475
 476
 477 void
 478 mdoc_node_free(struct mdoc_node *p)
 479 {
 480
 481         if (p->string)
 482                 free(p->string);
 483         if (p->args)
 484                 mdoc_argv_free(p->args);
 485         free(p);
 486 }
 487
 488
 489 static void
 490 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 491 {
 492
 493         /* Adjust siblings. */
 494
 495         if (n->prev)
 496                 n->prev->next = n->next;
 497         if (n->next)
 498                 n->next->prev = n->prev;
 499
 500         /* Adjust parent. */
 501
 502         if (n->parent) {
 503                 n->parent->nchild--;
 504                 if (n->parent->child == n)
 505                         n->parent->child = n->prev ? n->prev : n->next;
 506         }
 507
 508         /* Adjust parse point, if applicable. */
 509
 510         if (m && m->last == n) {
 511                 if (n->prev) {
 512                         m->last = n->prev;
 513                         m->next = MDOC_NEXT_SIBLING;
 514                 } else {
 515                         m->last = n->parent;
 516                         m->next = MDOC_NEXT_CHILD;
 517                 }
 518         }
 519
 520         if (m && m->first == n)
 521                 m->first = NULL;
 522 }
 523
 524
 525 void
 526 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 527 {
 528
 529         while (p->child) {
 530                 assert(p->nchild);
 531                 mdoc_node_delete(m, p->child);
 532         }
 533         assert(0 == p->nchild);
 534
 535         mdoc_node_unlink(m, p);
 536         mdoc_node_free(p);
 537 }
 538
 539
 540 /*
 541  * Parse free-form text, that is, a line that does not begin with the
 542  * control character.
 543  */
 544 static int
 545 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 546 {
 547         char             *c, *ws, *end;
 548         struct mdoc_node *n;
 549
 550         /* Ignore bogus comments. */
 551
 552         if ('\\' == buf[offs] &&
 553                         '.' == buf[offs + 1] &&
 554                         '"' == buf[offs + 2])
 555                 return(mdoc_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
 556
 557         /* No text before an initial macro. */
 558
 559         if (SEC_NONE == m->lastnamed)
 560                 return(mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT));
 561
 562         assert(m->last);
 563         n = m->last;
 564
 565         /*
 566          * Divert directly to list processing if we're encountering a
 567          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 568          * (a MDOC_BODY means it's already open, in which case we should
 569          * process within its context in the normal way).
 570          */
 571
 572         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 573                         LIST_column == n->data.Bl.type) {
 574                 /* `Bl' is open without any children. */
 575                 m->flags |= MDOC_FREECOL;
 576                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 577         }
 578
 579         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 580                         NULL != n->parent &&
 581                         MDOC_Bl == n->parent->tok &&
 582                         LIST_column == n->parent->data.Bl.type) {
 583                 /* `Bl' has block-level `It' children. */
 584                 m->flags |= MDOC_FREECOL;
 585                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 586         }
 587
 588         /*
 589          * Search for the beginning of unescaped trailing whitespace (ws)
 590          * and for the first character not to be output (end).
 591          */
 592
 593         /* FIXME: replace with strcspn(). */
 594         ws = NULL;
 595         for (c = end = buf + offs; *c; c++) {
 596                 switch (*c) {
 597                 case '-':
 598                         if (mandoc_hyph(buf + offs, c))
 599                                 *c = ASCII_HYPH;
 600                         ws = NULL;
 601                         break;
 602                 case ' ':
 603                         if (NULL == ws)
 604                                 ws = c;
 605                         continue;
 606                 case '\t':
 607                         /*
 608                          * Always warn about trailing tabs,
 609                          * even outside literal context,
 610                          * where they should be put on the next line.
 611                          */
 612                         if (NULL == ws)
 613                                 ws = c;
 614                         /*
 615                          * Strip trailing tabs in literal context only;
 616                          * outside, they affect the next line.
 617                          */
 618                         if (MDOC_LITERAL & m->flags)
 619                                 continue;
 620                         break;
 621                 case '\\':
 622                         /* Skip the escaped character, too, if any. */
 623                         if (c[1])
 624                                 c++;
 625                         /* FALLTHROUGH */
 626                 default:
 627                         ws = NULL;
 628                         break;
 629                 }
 630                 end = c + 1;
 631         }
 632         *end = '\0';
 633
 634         if (ws)
 635                 if ( ! mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE))
 636                         return(0);
 637
 638         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 639                 if ( ! mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN))
 640                         return(0);
 641
 642                 /*
 643                  * Insert a `Pp' in the case of a blank line.  Technically,
 644                  * blank lines aren't allowed, but enough manuals assume this
 645                  * behaviour that we want to work around it.
 646                  */
 647                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_Pp, NULL))
 648                         return(0);
 649
 650                 m->next = MDOC_NEXT_SIBLING;
 651                 return(1);
 652         }
 653
 654         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 655                 return(0);
 656
 657         if (MDOC_LITERAL & m->flags)
 658                 return(1);
 659
 660         /*
 661          * End-of-sentence check.  If the last character is an unescaped
 662          * EOS character, then flag the node as being the end of a
 663          * sentence.  The front-end will know how to interpret this.
 664          */
 665
 666         assert(buf < end);
 667
 668         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
 669                 m->last->flags |= MDOC_EOS;
 670
 671         return(1);
 672 }
 673
 674
 675 static int
 676 macrowarn(struct mdoc *m, int ln, const char *buf, int offs)
 677 {
 678         int              rc;
 679
 680         rc = mdoc_vmsg(m, MANDOCERR_MACRO, ln, offs,
 681                         "unknown macro: %s%s",
 682                         buf, strlen(buf) > 3 ? "..." : "");
 683
 684         /* FIXME: logic should be in driver. */
 685         /* FIXME: broken, will error out and not omit a message. */
 686         return(MDOC_IGN_MACRO & m->pflags ? rc : 0);
 687 }
 688
 689
 690 /*
 691  * Parse a macro line, that is, a line beginning with the control
 692  * character.
 693  */
 694 int
 695 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 696 {
 697         enum mdoct        tok;
 698         int               i, j, sv;
 699         char              mac[5];
 700         struct mdoc_node *n;
 701
 702         /* Empty lines are ignored. */
 703
 704         offs++;
 705
 706         if ('\0' == buf[offs])
 707                 return(1);
 708
 709         i = offs;
 710
 711         /* Accept whitespace after the initial control char. */
 712
 713         if (' ' == buf[i]) {
 714                 i++;
 715                 while (buf[i] && ' ' == buf[i])
 716                         i++;
 717                 if ('\0' == buf[i])
 718                         return(1);
 719         }
 720
 721         sv = i;
 722
 723         /* Copy the first word into a nil-terminated buffer. */
 724
 725         for (j = 0; j < 4; j++, i++) {
 726                 if ('\0' == (mac[j] = buf[i]))
 727                         break;
 728                 else if (' ' == buf[i])
 729                         break;
 730
 731                 /* Check for invalid characters. */
 732
 733                 if (isgraph((u_char)buf[i]))
 734                         continue;
 735                 if ( ! mdoc_pmsg(m, ln, i, MANDOCERR_BADCHAR))
 736                         return(0);
 737                 i--;
 738         }
 739
 740         mac[j] = '\0';
 741
 742         if (j == 4 || j < 2) {
 743                 if ( ! macrowarn(m, ln, mac, sv))
 744                         goto err;
 745                 return(1);
 746         }
 747
 748         if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
 749                 if ( ! macrowarn(m, ln, mac, sv))
 750                         goto err;
 751                 return(1);
 752         }
 753
 754         /* The macro is sane.  Jump to the next word. */
 755
 756         while (buf[i] && ' ' == buf[i])
 757                 i++;
 758
 759         /*
 760          * Trailing whitespace.  Note that tabs are allowed to be passed
 761          * into the parser as "text", so we only warn about spaces here.
 762          */
 763
 764         if ('\0' == buf[i] && ' ' == buf[i - 1])
 765                 if ( ! mdoc_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
 766                         goto err;
 767
 768         /*
 769          * If an initial macro or a list invocation, divert directly
 770          * into macro processing.
 771          */
 772
 773         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 774                 if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 775                         goto err;
 776                 return(1);
 777         }
 778
 779         n = m->last;
 780         assert(m->last);
 781
 782         /*
 783          * If the first macro of a `Bl -column', open an `It' block
 784          * context around the parsed macro.
 785          */
 786
 787         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 788                         LIST_column == n->data.Bl.type) {
 789                 m->flags |= MDOC_FREECOL;
 790                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 791                         goto err;
 792                 return(1);
 793         }
 794
 795         /*
 796          * If we're following a block-level `It' within a `Bl -column'
 797          * context (perhaps opened in the above block or in ptext()),
 798          * then open an `It' block context around the parsed macro.
 799          */
 800
 801         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 802                         NULL != n->parent &&
 803                         MDOC_Bl == n->parent->tok &&
 804                         LIST_column == n->parent->data.Bl.type) {
 805                 m->flags |= MDOC_FREECOL;
 806                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 807                         goto err;
 808                 return(1);
 809         }
 810
 811         /* Normal processing of a macro. */
 812
 813         if ( ! mdoc_macro(m, tok, ln, sv, &i, buf))
 814                 goto err;
 815
 816         return(1);
 817
 818 err:    /* Error out. */
 819
 820         m->flags |= MDOC_HALT;
 821         return(0);
 822 }
 823
 824