mdoc.c

   1 /*      $Id: mdoc.c,v 1.195 2011/07/28 14:17:11 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <sys/types.h>
  23
  24 #include <assert.h>
  25 #include <stdarg.h>
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mdoc.h"
  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 #if 0
 101 static  int               mdoc_preptext(struct mdoc *, int, char *, int);
 102 #endif
 103 static  int               mdoc_ptext(struct mdoc *, int, char *, int);
 104 static  int               mdoc_pmacro(struct mdoc *, int, char *, int);
 105
 106 const struct mdoc_node *
 107 mdoc_node(const struct mdoc *m)
 108 {
 109
 110         assert( ! (MDOC_HALT & m->flags));
 111         return(m->first);
 112 }
 113
 114
 115 const struct mdoc_meta *
 116 mdoc_meta(const struct mdoc *m)
 117 {
 118
 119         assert( ! (MDOC_HALT & m->flags));
 120         return(&m->meta);
 121 }
 122
 123
 124 /*
 125  * Frees volatile resources (parse tree, meta-data, fields).
 126  */
 127 static void
 128 mdoc_free1(struct mdoc *mdoc)
 129 {
 130
 131         if (mdoc->first)
 132                 mdoc_node_delete(mdoc, mdoc->first);
 133         if (mdoc->meta.title)
 134                 free(mdoc->meta.title);
 135         if (mdoc->meta.os)
 136                 free(mdoc->meta.os);
 137         if (mdoc->meta.name)
 138                 free(mdoc->meta.name);
 139         if (mdoc->meta.arch)
 140                 free(mdoc->meta.arch);
 141         if (mdoc->meta.vol)
 142                 free(mdoc->meta.vol);
 143         if (mdoc->meta.msec)
 144                 free(mdoc->meta.msec);
 145         if (mdoc->meta.date)
 146                 free(mdoc->meta.date);
 147 }
 148
 149
 150 /*
 151  * Allocate all volatile resources (parse tree, meta-data, fields).
 152  */
 153 static void
 154 mdoc_alloc1(struct mdoc *mdoc)
 155 {
 156
 157         memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 158         mdoc->flags = 0;
 159         mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
 160         mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 161         mdoc->first = mdoc->last;
 162         mdoc->last->type = MDOC_ROOT;
 163         mdoc->next = MDOC_NEXT_CHILD;
 164 }
 165
 166
 167 /*
 168  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 169  * data with mdoc_alloc1().  After invocation, parse data has been reset
 170  * and the parser is ready for re-invocation on a new tree; however,
 171  * cross-parse non-volatile data is kept intact.
 172  */
 173 void
 174 mdoc_reset(struct mdoc *mdoc)
 175 {
 176
 177         mdoc_free1(mdoc);
 178         mdoc_alloc1(mdoc);
 179 }
 180
 181
 182 /*
 183  * Completely free up all volatile and non-volatile parse resources.
 184  * After invocation, the pointer is no longer usable.
 185  */
 186 void
 187 mdoc_free(struct mdoc *mdoc)
 188 {
 189
 190         mdoc_free1(mdoc);
 191         free(mdoc);
 192 }
 193
 194
 195 /*
 196  * Allocate volatile and non-volatile parse resources.
 197  */
 198 struct mdoc *
 199 mdoc_alloc(struct roff *roff, struct mparse *parse)
 200 {
 201         struct mdoc     *p;
 202
 203         p = mandoc_calloc(1, sizeof(struct mdoc));
 204
 205         p->parse = parse;
 206         p->roff = roff;
 207
 208         mdoc_hash_init();
 209         mdoc_alloc1(p);
 210         return(p);
 211 }
 212
 213
 214 /*
 215  * Climb back up the parse tree, validating open scopes.  Mostly calls
 216  * through to macro_end() in macro.c.
 217  */
 218 int
 219 mdoc_endparse(struct mdoc *m)
 220 {
 221
 222         assert( ! (MDOC_HALT & m->flags));
 223         if (mdoc_macroend(m))
 224                 return(1);
 225         m->flags |= MDOC_HALT;
 226         return(0);
 227 }
 228
 229 int
 230 mdoc_addeqn(struct mdoc *m, const struct eqn *ep)
 231 {
 232         struct mdoc_node *n;
 233
 234         assert( ! (MDOC_HALT & m->flags));
 235
 236         /* No text before an initial macro. */
 237
 238         if (SEC_NONE == m->lastnamed) {
 239                 mdoc_pmsg(m, ep->ln, ep->pos, MANDOCERR_NOTEXT);
 240                 return(1);
 241         }
 242
 243         n = node_alloc(m, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
 244         n->eqn = ep;
 245
 246         if ( ! node_append(m, n))
 247                 return(0);
 248
 249         m->next = MDOC_NEXT_SIBLING;
 250         return(1);
 251 }
 252
 253 int
 254 mdoc_addspan(struct mdoc *m, const struct tbl_span *sp)
 255 {
 256         struct mdoc_node *n;
 257
 258         assert( ! (MDOC_HALT & m->flags));
 259
 260         /* No text before an initial macro. */
 261
 262         if (SEC_NONE == m->lastnamed) {
 263                 mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT);
 264                 return(1);
 265         }
 266
 267         n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL);
 268         n->span = sp;
 269
 270         if ( ! node_append(m, n))
 271                 return(0);
 272
 273         m->next = MDOC_NEXT_SIBLING;
 274         return(1);
 275 }
 276
 277
 278 /*
 279  * Main parse routine.  Parses a single line -- really just hands off to
 280  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 281  */
 282 int
 283 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 284 {
 285
 286         assert( ! (MDOC_HALT & m->flags));
 287
 288         m->flags |= MDOC_NEWLINE;
 289
 290         /*
 291          * Let the roff nS register switch SYNOPSIS mode early,
 292          * such that the parser knows at all times
 293          * whether this mode is on or off.
 294          * Note that this mode is also switched by the Sh macro.
 295          */
 296         if (roff_regisset(m->roff, REG_nS)) {
 297                 if (roff_regget(m->roff, REG_nS))
 298                         m->flags |= MDOC_SYNOPSIS;
 299                 else
 300                         m->flags &= ~MDOC_SYNOPSIS;
 301         }
 302
 303         return(mandoc_getcontrol(buf, &offs) ?
 304                         mdoc_pmacro(m, ln, buf, offs) :
 305                         mdoc_ptext(m, ln, buf, offs));
 306 }
 307
 308 int
 309 mdoc_macro(MACRO_PROT_ARGS)
 310 {
 311         assert(tok < MDOC_MAX);
 312
 313         /* If we're in the body, deny prologue calls. */
 314
 315         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 316                         MDOC_PBODY & m->flags) {
 317                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY);
 318                 return(1);
 319         }
 320
 321         /* If we're in the prologue, deny "body" macros.  */
 322
 323         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 324                         ! (MDOC_PBODY & m->flags)) {
 325                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG);
 326                 if (NULL == m->meta.msec)
 327                         m->meta.msec = mandoc_strdup("1");
 328                 if (NULL == m->meta.title)
 329                         m->meta.title = mandoc_strdup("UNKNOWN");
 330                 if (NULL == m->meta.vol)
 331                         m->meta.vol = mandoc_strdup("LOCAL");
 332                 if (NULL == m->meta.os)
 333                         m->meta.os = mandoc_strdup("LOCAL");
 334                 if (NULL == m->meta.date)
 335                         m->meta.date = mandoc_normdate
 336                                 (m->parse, NULL, line, ppos);
 337                 m->flags |= MDOC_PBODY;
 338         }
 339
 340         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 341 }
 342
 343
 344 static int
 345 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 346 {
 347
 348         assert(mdoc->last);
 349         assert(mdoc->first);
 350         assert(MDOC_ROOT != p->type);
 351
 352         switch (mdoc->next) {
 353         case (MDOC_NEXT_SIBLING):
 354                 mdoc->last->next = p;
 355                 p->prev = mdoc->last;
 356                 p->parent = mdoc->last->parent;
 357                 break;
 358         case (MDOC_NEXT_CHILD):
 359                 mdoc->last->child = p;
 360                 p->parent = mdoc->last;
 361                 break;
 362         default:
 363                 abort();
 364                 /* NOTREACHED */
 365         }
 366
 367         p->parent->nchild++;
 368
 369         /*
 370          * Copy over the normalised-data pointer of our parent.  Not
 371          * everybody has one, but copying a null pointer is fine.
 372          */
 373
 374         switch (p->type) {
 375         case (MDOC_BODY):
 376                 /* FALLTHROUGH */
 377         case (MDOC_TAIL):
 378                 /* FALLTHROUGH */
 379         case (MDOC_HEAD):
 380                 p->norm = p->parent->norm;
 381                 break;
 382         default:
 383                 break;
 384         }
 385
 386         if ( ! mdoc_valid_pre(mdoc, p))
 387                 return(0);
 388
 389         switch (p->type) {
 390         case (MDOC_HEAD):
 391                 assert(MDOC_BLOCK == p->parent->type);
 392                 p->parent->head = p;
 393                 break;
 394         case (MDOC_TAIL):
 395                 assert(MDOC_BLOCK == p->parent->type);
 396                 p->parent->tail = p;
 397                 break;
 398         case (MDOC_BODY):
 399                 if (p->end)
 400                         break;
 401                 assert(MDOC_BLOCK == p->parent->type);
 402                 p->parent->body = p;
 403                 break;
 404         default:
 405                 break;
 406         }
 407
 408         mdoc->last = p;
 409
 410         switch (p->type) {
 411         case (MDOC_TBL):
 412                 /* FALLTHROUGH */
 413         case (MDOC_TEXT):
 414                 if ( ! mdoc_valid_post(mdoc))
 415                         return(0);
 416                 break;
 417         default:
 418                 break;
 419         }
 420
 421         return(1);
 422 }
 423
 424
 425 static struct mdoc_node *
 426 node_alloc(struct mdoc *m, int line, int pos,
 427                 enum mdoct tok, enum mdoc_type type)
 428 {
 429         struct mdoc_node *p;
 430
 431         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 432         p->sec = m->lastsec;
 433         p->line = line;
 434         p->pos = pos;
 435         p->tok = tok;
 436         p->type = type;
 437
 438         /* Flag analysis. */
 439
 440         if (MDOC_SYNOPSIS & m->flags)
 441                 p->flags |= MDOC_SYNPRETTY;
 442         else
 443                 p->flags &= ~MDOC_SYNPRETTY;
 444         if (MDOC_NEWLINE & m->flags)
 445                 p->flags |= MDOC_LINE;
 446         m->flags &= ~MDOC_NEWLINE;
 447
 448         return(p);
 449 }
 450
 451
 452 int
 453 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 454 {
 455         struct mdoc_node *p;
 456
 457         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 458         if ( ! node_append(m, p))
 459                 return(0);
 460         m->next = MDOC_NEXT_CHILD;
 461         return(1);
 462 }
 463
 464
 465 int
 466 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 467 {
 468         struct mdoc_node *p;
 469
 470         assert(m->first);
 471         assert(m->last);
 472
 473         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 474         if ( ! node_append(m, p))
 475                 return(0);
 476         m->next = MDOC_NEXT_CHILD;
 477         return(1);
 478 }
 479
 480
 481 int
 482 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 483 {
 484         struct mdoc_node *p;
 485
 486         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 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_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 496                 struct mdoc_node *body, enum mdoc_endbody end)
 497 {
 498         struct mdoc_node *p;
 499
 500         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 501         p->pending = body;
 502         p->end = end;
 503         if ( ! node_append(m, p))
 504                 return(0);
 505         m->next = MDOC_NEXT_SIBLING;
 506         return(1);
 507 }
 508
 509
 510 int
 511 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 512                 enum mdoct tok, struct mdoc_arg *args)
 513 {
 514         struct mdoc_node *p;
 515
 516         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 517         p->args = args;
 518         if (p->args)
 519                 (args->refcnt)++;
 520
 521         switch (tok) {
 522         case (MDOC_Bd):
 523                 /* FALLTHROUGH */
 524         case (MDOC_Bf):
 525                 /* FALLTHROUGH */
 526         case (MDOC_Bl):
 527                 /* FALLTHROUGH */
 528         case (MDOC_Rs):
 529                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 530                 break;
 531         default:
 532                 break;
 533         }
 534
 535         if ( ! node_append(m, p))
 536                 return(0);
 537         m->next = MDOC_NEXT_CHILD;
 538         return(1);
 539 }
 540
 541
 542 int
 543 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 544                 enum mdoct tok, struct mdoc_arg *args)
 545 {
 546         struct mdoc_node *p;
 547
 548         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 549         p->args = args;
 550         if (p->args)
 551                 (args->refcnt)++;
 552
 553         switch (tok) {
 554         case (MDOC_An):
 555                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 556                 break;
 557         default:
 558                 break;
 559         }
 560
 561         if ( ! node_append(m, p))
 562                 return(0);
 563         m->next = MDOC_NEXT_CHILD;
 564         return(1);
 565 }
 566
 567 int
 568 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 569 {
 570         struct mdoc_node *n;
 571
 572         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 573         n->string = roff_strdup(m->roff, p);
 574
 575         if ( ! node_append(m, n))
 576                 return(0);
 577
 578         m->next = MDOC_NEXT_SIBLING;
 579         return(1);
 580 }
 581
 582
 583 static void
 584 mdoc_node_free(struct mdoc_node *p)
 585 {
 586
 587         if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
 588                 free(p->norm);
 589         if (p->string)
 590                 free(p->string);
 591         if (p->args)
 592                 mdoc_argv_free(p->args);
 593         free(p);
 594 }
 595
 596
 597 static void
 598 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 599 {
 600
 601         /* Adjust siblings. */
 602
 603         if (n->prev)
 604                 n->prev->next = n->next;
 605         if (n->next)
 606                 n->next->prev = n->prev;
 607
 608         /* Adjust parent. */
 609
 610         if (n->parent) {
 611                 n->parent->nchild--;
 612                 if (n->parent->child == n)
 613                         n->parent->child = n->prev ? n->prev : n->next;
 614                 if (n->parent->last == n)
 615                         n->parent->last = n->prev ? n->prev : NULL;
 616         }
 617
 618         /* Adjust parse point, if applicable. */
 619
 620         if (m && m->last == n) {
 621                 if (n->prev) {
 622                         m->last = n->prev;
 623                         m->next = MDOC_NEXT_SIBLING;
 624                 } else {
 625                         m->last = n->parent;
 626                         m->next = MDOC_NEXT_CHILD;
 627                 }
 628         }
 629
 630         if (m && m->first == n)
 631                 m->first = NULL;
 632 }
 633
 634
 635 void
 636 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 637 {
 638
 639         while (p->child) {
 640                 assert(p->nchild);
 641                 mdoc_node_delete(m, p->child);
 642         }
 643         assert(0 == p->nchild);
 644
 645         mdoc_node_unlink(m, p);
 646         mdoc_node_free(p);
 647 }
 648
 649 #if 0
 650 /*
 651  * Pre-treat a text line.
 652  * Text lines can consist of equations, which must be handled apart from
 653  * the regular text.
 654  * Thus, use this function to step through a line checking if it has any
 655  * equations embedded in it.
 656  * This must handle multiple equations AND equations that do not end at
 657  * the end-of-line, i.e., will re-enter in the next roff parse.
 658  */
 659 static int
 660 mdoc_preptext(struct mdoc *m, int line, char *buf, int offs)
 661 {
 662         char            *start, *end;
 663         char             delim;
 664
 665         while ('\0' != buf[offs]) {
 666                 /* Mark starting position if eqn is set. */
 667                 start = NULL;
 668                 if ('\0' != (delim = roff_eqndelim(m->roff)))
 669                         if (NULL != (start = strchr(buf + offs, delim)))
 670                                 *start++ = '\0';
 671
 672                 /* Parse text as normal. */
 673                 if ( ! mdoc_ptext(m, line, buf, offs))
 674                         return(0);
 675
 676                 /* Continue only if an equation exists. */
 677                 if (NULL == start)
 678                         break;
 679
 680                 /* Read past the end of the equation. */
 681                 offs += start - (buf + offs);
 682                 assert(start == &buf[offs]);
 683                 if (NULL != (end = strchr(buf + offs, delim))) {
 684                         *end++ = '\0';
 685                         while (' ' == *end)
 686                                 end++;
 687                 }
 688
 689                 /* Parse the equation itself. */
 690                 roff_openeqn(m->roff, NULL, line, offs, buf);
 691
 692                 /* Process a finished equation? */
 693                 if (roff_closeeqn(m->roff))
 694                         if ( ! mdoc_addeqn(m, roff_eqn(m->roff)))
 695                                 return(0);
 696                 offs += (end - (buf + offs));
 697         }
 698
 699         return(1);
 700 }
 701 #endif
 702
 703 /*
 704  * Parse free-form text, that is, a line that does not begin with the
 705  * control character.
 706  */
 707 static int
 708 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 709 {
 710         char             *c, *ws, *end;
 711         struct mdoc_node *n;
 712
 713         /* No text before an initial macro. */
 714
 715         if (SEC_NONE == m->lastnamed) {
 716                 mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
 717                 return(1);
 718         }
 719
 720         assert(m->last);
 721         n = m->last;
 722
 723         /*
 724          * Divert directly to list processing if we're encountering a
 725          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 726          * (a MDOC_BODY means it's already open, in which case we should
 727          * process within its context in the normal way).
 728          */
 729
 730         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 731                         LIST_column == n->norm->Bl.type) {
 732                 /* `Bl' is open without any children. */
 733                 m->flags |= MDOC_FREECOL;
 734                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 735         }
 736
 737         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 738                         NULL != n->parent &&
 739                         MDOC_Bl == n->parent->tok &&
 740                         LIST_column == n->parent->norm->Bl.type) {
 741                 /* `Bl' has block-level `It' children. */
 742                 m->flags |= MDOC_FREECOL;
 743                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 744         }
 745
 746         /*
 747          * Search for the beginning of unescaped trailing whitespace (ws)
 748          * and for the first character not to be output (end).
 749          */
 750
 751         /* FIXME: replace with strcspn(). */
 752         ws = NULL;
 753         for (c = end = buf + offs; *c; c++) {
 754                 switch (*c) {
 755                 case ' ':
 756                         if (NULL == ws)
 757                                 ws = c;
 758                         continue;
 759                 case '\t':
 760                         /*
 761                          * Always warn about trailing tabs,
 762                          * even outside literal context,
 763                          * where they should be put on the next line.
 764                          */
 765                         if (NULL == ws)
 766                                 ws = c;
 767                         /*
 768                          * Strip trailing tabs in literal context only;
 769                          * outside, they affect the next line.
 770                          */
 771                         if (MDOC_LITERAL & m->flags)
 772                                 continue;
 773                         break;
 774                 case '\\':
 775                         /* Skip the escaped character, too, if any. */
 776                         if (c[1])
 777                                 c++;
 778                         /* FALLTHROUGH */
 779                 default:
 780                         ws = NULL;
 781                         break;
 782                 }
 783                 end = c + 1;
 784         }
 785         *end = '\0';
 786
 787         if (ws)
 788                 mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
 789
 790         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 791                 mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
 792
 793                 /*
 794                  * Insert a `sp' in the case of a blank line.  Technically,
 795                  * blank lines aren't allowed, but enough manuals assume this
 796                  * behaviour that we want to work around it.
 797                  */
 798                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 799                         return(0);
 800
 801                 m->next = MDOC_NEXT_SIBLING;
 802                 return(1);
 803         }
 804
 805         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 806                 return(0);
 807
 808         if (MDOC_LITERAL & m->flags)
 809                 return(1);
 810
 811         /*
 812          * End-of-sentence check.  If the last character is an unescaped
 813          * EOS character, then flag the node as being the end of a
 814          * sentence.  The front-end will know how to interpret this.
 815          */
 816
 817         assert(buf < end);
 818
 819         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 820                 m->last->flags |= MDOC_EOS;
 821
 822         return(1);
 823 }
 824
 825
 826 /*
 827  * Parse a macro line, that is, a line beginning with the control
 828  * character.
 829  */
 830 static int
 831 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 832 {
 833         enum mdoct        tok;
 834         int               i, sv;
 835         char              mac[5];
 836         struct mdoc_node *n;
 837
 838         /* Empty post-control lines are ignored. */
 839
 840         if ('"' == buf[offs]) {
 841                 mdoc_pmsg(m, ln, offs, MANDOCERR_BADCOMMENT);
 842                 return(1);
 843         } else if ('\0' == buf[offs])
 844                 return(1);
 845
 846         sv = offs;
 847
 848         /*
 849          * Copy the first word into a nil-terminated buffer.
 850          * Stop copying when a tab, space, or eoln is encountered.
 851          */
 852
 853         i = 0;
 854         while (i < 4 && '\0' != buf[offs] &&
 855                         ' ' != buf[offs] && '\t' != buf[offs])
 856                 mac[i++] = buf[offs++];
 857
 858         mac[i] = '\0';
 859
 860         tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 861
 862         if (MDOC_MAX == tok) {
 863                 mandoc_vmsg(MANDOCERR_MACRO, m->parse,
 864                                 ln, sv, "%s", buf + sv - 1);
 865                 return(1);
 866         }
 867
 868         /* Disregard the first trailing tab, if applicable. */
 869
 870         if ('\t' == buf[offs])
 871                 offs++;
 872
 873         /* Jump to the next non-whitespace word. */
 874
 875         while (buf[offs] && ' ' == buf[offs])
 876                 offs++;
 877
 878         /*
 879          * Trailing whitespace.  Note that tabs are allowed to be passed
 880          * into the parser as "text", so we only warn about spaces here.
 881          */
 882
 883         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 884                 mdoc_pmsg(m, ln, offs - 1, MANDOCERR_EOLNSPACE);
 885
 886         /*
 887          * If an initial macro or a list invocation, divert directly
 888          * into macro processing.
 889          */
 890
 891         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 892                 if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
 893                         goto err;
 894                 return(1);
 895         }
 896
 897         n = m->last;
 898         assert(m->last);
 899
 900         /*
 901          * If the first macro of a `Bl -column', open an `It' block
 902          * context around the parsed macro.
 903          */
 904
 905         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 906                         LIST_column == n->norm->Bl.type) {
 907                 m->flags |= MDOC_FREECOL;
 908                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 909                         goto err;
 910                 return(1);
 911         }
 912
 913         /*
 914          * If we're following a block-level `It' within a `Bl -column'
 915          * context (perhaps opened in the above block or in ptext()),
 916          * then open an `It' block context around the parsed macro.
 917          */
 918
 919         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 920                         NULL != n->parent &&
 921                         MDOC_Bl == n->parent->tok &&
 922                         LIST_column == n->parent->norm->Bl.type) {
 923                 m->flags |= MDOC_FREECOL;
 924                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 925                         goto err;
 926                 return(1);
 927         }
 928
 929         /* Normal processing of a macro. */
 930
 931         if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
 932                 goto err;
 933
 934         return(1);
 935
 936 err:    /* Error out. */
 937
 938         m->flags |= MDOC_HALT;
 939         return(0);
 940 }
 941
 942 enum mdelim
 943 mdoc_isdelim(const char *p)
 944 {
 945
 946         if ('\0' == p[0])
 947                 return(DELIM_NONE);
 948
 949         if ('\0' == p[1])
 950                 switch (p[0]) {
 951                 case('('):
 952                         /* FALLTHROUGH */
 953                 case('['):
 954                         return(DELIM_OPEN);
 955                 case('|'):
 956                         return(DELIM_MIDDLE);
 957                 case('.'):
 958                         /* FALLTHROUGH */
 959                 case(','):
 960                         /* FALLTHROUGH */
 961                 case(';'):
 962                         /* FALLTHROUGH */
 963                 case(':'):
 964                         /* FALLTHROUGH */
 965                 case('?'):
 966                         /* FALLTHROUGH */
 967                 case('!'):
 968                         /* FALLTHROUGH */
 969                 case(')'):
 970                         /* FALLTHROUGH */
 971                 case(']'):
 972                         return(DELIM_CLOSE);
 973                 default:
 974                         return(DELIM_NONE);
 975                 }
 976
 977         if ('\\' != p[0])
 978                 return(DELIM_NONE);
 979
 980         if (0 == strcmp(p + 1, "."))
 981                 return(DELIM_CLOSE);
 982         if (0 == strcmp(p + 1, "*(Ba"))
 983                 return(DELIM_MIDDLE);
 984
 985         return(DELIM_NONE);
 986 }