mdoc.c

   1 /*      $Id: mdoc.c,v 1.190 2011/07/21 10:24:35 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 static  int               mdoc_ptext(struct mdoc *, int, char *, int);
 101 static  int               mdoc_pmacro(struct mdoc *, int, char *, int);
 102
 103 const struct mdoc_node *
 104 mdoc_node(const struct mdoc *m)
 105 {
 106
 107         assert( ! (MDOC_HALT & m->flags));
 108         return(m->first);
 109 }
 110
 111
 112 const struct mdoc_meta *
 113 mdoc_meta(const struct mdoc *m)
 114 {
 115
 116         assert( ! (MDOC_HALT & m->flags));
 117         return(&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         if (mdoc->meta.date)
 143                 free(mdoc->meta.date);
 144 }
 145
 146
 147 /*
 148  * Allocate all volatile resources (parse tree, meta-data, fields).
 149  */
 150 static void
 151 mdoc_alloc1(struct mdoc *mdoc)
 152 {
 153
 154         memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 155         mdoc->flags = 0;
 156         mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
 157         mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 158         mdoc->first = mdoc->last;
 159         mdoc->last->type = MDOC_ROOT;
 160         mdoc->next = MDOC_NEXT_CHILD;
 161 }
 162
 163
 164 /*
 165  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 166  * data with mdoc_alloc1().  After invocation, parse data has been reset
 167  * and the parser is ready for re-invocation on a new tree; however,
 168  * cross-parse non-volatile data is kept intact.
 169  */
 170 void
 171 mdoc_reset(struct mdoc *mdoc)
 172 {
 173
 174         mdoc_free1(mdoc);
 175         mdoc_alloc1(mdoc);
 176 }
 177
 178
 179 /*
 180  * Completely free up all volatile and non-volatile parse resources.
 181  * After invocation, the pointer is no longer usable.
 182  */
 183 void
 184 mdoc_free(struct mdoc *mdoc)
 185 {
 186
 187         mdoc_free1(mdoc);
 188         free(mdoc);
 189 }
 190
 191
 192 /*
 193  * Allocate volatile and non-volatile parse resources.
 194  */
 195 struct mdoc *
 196 mdoc_alloc(struct roff *roff, struct mparse *parse)
 197 {
 198         struct mdoc     *p;
 199
 200         p = mandoc_calloc(1, sizeof(struct mdoc));
 201
 202         p->parse = parse;
 203         p->roff = roff;
 204
 205         mdoc_hash_init();
 206         mdoc_alloc1(p);
 207         return(p);
 208 }
 209
 210
 211 /*
 212  * Climb back up the parse tree, validating open scopes.  Mostly calls
 213  * through to macro_end() in macro.c.
 214  */
 215 int
 216 mdoc_endparse(struct mdoc *m)
 217 {
 218
 219         assert( ! (MDOC_HALT & m->flags));
 220         if (mdoc_macroend(m))
 221                 return(1);
 222         m->flags |= MDOC_HALT;
 223         return(0);
 224 }
 225
 226 int
 227 mdoc_addeqn(struct mdoc *m, const struct eqn *ep)
 228 {
 229         struct mdoc_node *n;
 230
 231         assert( ! (MDOC_HALT & m->flags));
 232
 233         /* No text before an initial macro. */
 234
 235         if (SEC_NONE == m->lastnamed) {
 236                 mdoc_pmsg(m, ep->ln, ep->pos, MANDOCERR_NOTEXT);
 237                 return(1);
 238         }
 239
 240         n = node_alloc(m, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
 241         n->eqn = ep;
 242
 243         if ( ! node_append(m, n))
 244                 return(0);
 245
 246         m->next = MDOC_NEXT_SIBLING;
 247         return(1);
 248 }
 249
 250 int
 251 mdoc_addspan(struct mdoc *m, const struct tbl_span *sp)
 252 {
 253         struct mdoc_node *n;
 254
 255         assert( ! (MDOC_HALT & m->flags));
 256
 257         /* No text before an initial macro. */
 258
 259         if (SEC_NONE == m->lastnamed) {
 260                 mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT);
 261                 return(1);
 262         }
 263
 264         n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL);
 265         n->span = sp;
 266
 267         if ( ! node_append(m, n))
 268                 return(0);
 269
 270         m->next = MDOC_NEXT_SIBLING;
 271         return(1);
 272 }
 273
 274
 275 /*
 276  * Main parse routine.  Parses a single line -- really just hands off to
 277  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 278  */
 279 int
 280 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 281 {
 282
 283         assert( ! (MDOC_HALT & m->flags));
 284
 285         m->flags |= MDOC_NEWLINE;
 286
 287         /*
 288          * Let the roff nS register switch SYNOPSIS mode early,
 289          * such that the parser knows at all times
 290          * whether this mode is on or off.
 291          * Note that this mode is also switched by the Sh macro.
 292          */
 293         if (roff_regisset(m->roff, REG_nS)) {
 294                 if (roff_regget(m->roff, REG_nS))
 295                         m->flags |= MDOC_SYNOPSIS;
 296                 else
 297                         m->flags &= ~MDOC_SYNOPSIS;
 298         }
 299
 300         return(mandoc_getcontrol(buf, &offs) ?
 301                         mdoc_pmacro(m, ln, buf, offs) :
 302                         mdoc_ptext(m, ln, buf, offs));
 303 }
 304
 305 int
 306 mdoc_macro(MACRO_PROT_ARGS)
 307 {
 308         assert(tok < MDOC_MAX);
 309
 310         /* If we're in the body, deny prologue calls. */
 311
 312         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 313                         MDOC_PBODY & m->flags) {
 314                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY);
 315                 return(1);
 316         }
 317
 318         /* If we're in the prologue, deny "body" macros.  */
 319
 320         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 321                         ! (MDOC_PBODY & m->flags)) {
 322                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG);
 323                 if (NULL == m->meta.msec)
 324                         m->meta.msec = mandoc_strdup("1");
 325                 if (NULL == m->meta.title)
 326                         m->meta.title = mandoc_strdup("UNKNOWN");
 327                 if (NULL == m->meta.vol)
 328                         m->meta.vol = mandoc_strdup("LOCAL");
 329                 if (NULL == m->meta.os)
 330                         m->meta.os = mandoc_strdup("LOCAL");
 331                 if (NULL == m->meta.date)
 332                         m->meta.date = mandoc_normdate
 333                                 (m->parse, NULL, line, ppos);
 334                 m->flags |= MDOC_PBODY;
 335         }
 336
 337         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 338 }
 339
 340
 341 static int
 342 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 343 {
 344
 345         assert(mdoc->last);
 346         assert(mdoc->first);
 347         assert(MDOC_ROOT != p->type);
 348
 349         switch (mdoc->next) {
 350         case (MDOC_NEXT_SIBLING):
 351                 mdoc->last->next = p;
 352                 p->prev = mdoc->last;
 353                 p->parent = mdoc->last->parent;
 354                 break;
 355         case (MDOC_NEXT_CHILD):
 356                 mdoc->last->child = p;
 357                 p->parent = mdoc->last;
 358                 break;
 359         default:
 360                 abort();
 361                 /* NOTREACHED */
 362         }
 363
 364         p->parent->nchild++;
 365
 366         /*
 367          * Copy over the normalised-data pointer of our parent.  Not
 368          * everybody has one, but copying a null pointer is fine.
 369          */
 370
 371         switch (p->type) {
 372         case (MDOC_BODY):
 373                 /* FALLTHROUGH */
 374         case (MDOC_TAIL):
 375                 /* FALLTHROUGH */
 376         case (MDOC_HEAD):
 377                 p->norm = p->parent->norm;
 378                 break;
 379         default:
 380                 break;
 381         }
 382
 383         if ( ! mdoc_valid_pre(mdoc, p))
 384                 return(0);
 385
 386         switch (p->type) {
 387         case (MDOC_HEAD):
 388                 assert(MDOC_BLOCK == p->parent->type);
 389                 p->parent->head = p;
 390                 break;
 391         case (MDOC_TAIL):
 392                 assert(MDOC_BLOCK == p->parent->type);
 393                 p->parent->tail = p;
 394                 break;
 395         case (MDOC_BODY):
 396                 if (p->end)
 397                         break;
 398                 assert(MDOC_BLOCK == p->parent->type);
 399                 p->parent->body = p;
 400                 break;
 401         default:
 402                 break;
 403         }
 404
 405         mdoc->last = p;
 406
 407         switch (p->type) {
 408         case (MDOC_TBL):
 409                 /* FALLTHROUGH */
 410         case (MDOC_TEXT):
 411                 if ( ! mdoc_valid_post(mdoc))
 412                         return(0);
 413                 break;
 414         default:
 415                 break;
 416         }
 417
 418         return(1);
 419 }
 420
 421
 422 static struct mdoc_node *
 423 node_alloc(struct mdoc *m, int line, int pos,
 424                 enum mdoct tok, enum mdoc_type type)
 425 {
 426         struct mdoc_node *p;
 427
 428         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 429         p->sec = m->lastsec;
 430         p->line = line;
 431         p->pos = pos;
 432         p->tok = tok;
 433         p->type = type;
 434
 435         /* Flag analysis. */
 436
 437         if (MDOC_SYNOPSIS & m->flags)
 438                 p->flags |= MDOC_SYNPRETTY;
 439         else
 440                 p->flags &= ~MDOC_SYNPRETTY;
 441         if (MDOC_NEWLINE & m->flags)
 442                 p->flags |= MDOC_LINE;
 443         m->flags &= ~MDOC_NEWLINE;
 444
 445         return(p);
 446 }
 447
 448
 449 int
 450 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 451 {
 452         struct mdoc_node *p;
 453
 454         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 455         if ( ! node_append(m, p))
 456                 return(0);
 457         m->next = MDOC_NEXT_CHILD;
 458         return(1);
 459 }
 460
 461
 462 int
 463 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 464 {
 465         struct mdoc_node *p;
 466
 467         assert(m->first);
 468         assert(m->last);
 469
 470         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 471         if ( ! node_append(m, p))
 472                 return(0);
 473         m->next = MDOC_NEXT_CHILD;
 474         return(1);
 475 }
 476
 477
 478 int
 479 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 480 {
 481         struct mdoc_node *p;
 482
 483         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 484         if ( ! node_append(m, p))
 485                 return(0);
 486         m->next = MDOC_NEXT_CHILD;
 487         return(1);
 488 }
 489
 490
 491 int
 492 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 493                 struct mdoc_node *body, enum mdoc_endbody end)
 494 {
 495         struct mdoc_node *p;
 496
 497         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 498         p->pending = body;
 499         p->end = end;
 500         if ( ! node_append(m, p))
 501                 return(0);
 502         m->next = MDOC_NEXT_SIBLING;
 503         return(1);
 504 }
 505
 506
 507 int
 508 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 509                 enum mdoct tok, struct mdoc_arg *args)
 510 {
 511         struct mdoc_node *p;
 512
 513         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 514         p->args = args;
 515         if (p->args)
 516                 (args->refcnt)++;
 517
 518         switch (tok) {
 519         case (MDOC_Bd):
 520                 /* FALLTHROUGH */
 521         case (MDOC_Bf):
 522                 /* FALLTHROUGH */
 523         case (MDOC_Bl):
 524                 /* FALLTHROUGH */
 525         case (MDOC_Rs):
 526                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 527                 break;
 528         default:
 529                 break;
 530         }
 531
 532         if ( ! node_append(m, p))
 533                 return(0);
 534         m->next = MDOC_NEXT_CHILD;
 535         return(1);
 536 }
 537
 538
 539 int
 540 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 541                 enum mdoct tok, struct mdoc_arg *args)
 542 {
 543         struct mdoc_node *p;
 544
 545         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 546         p->args = args;
 547         if (p->args)
 548                 (args->refcnt)++;
 549
 550         switch (tok) {
 551         case (MDOC_An):
 552                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 553                 break;
 554         default:
 555                 break;
 556         }
 557
 558         if ( ! node_append(m, p))
 559                 return(0);
 560         m->next = MDOC_NEXT_CHILD;
 561         return(1);
 562 }
 563
 564 int
 565 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 566 {
 567         struct mdoc_node *n;
 568         size_t            sv, len;
 569
 570         len = strlen(p);
 571
 572         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 573         n->string = mandoc_malloc(len + 1);
 574         sv = strlcpy(n->string, p, len + 1);
 575
 576         /* Prohibit truncation. */
 577         assert(sv < len + 1);
 578
 579         if ( ! node_append(m, n))
 580                 return(0);
 581
 582         m->next = MDOC_NEXT_SIBLING;
 583         return(1);
 584 }
 585
 586
 587 static void
 588 mdoc_node_free(struct mdoc_node *p)
 589 {
 590
 591         if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
 592                 free(p->norm);
 593         if (p->string)
 594                 free(p->string);
 595         if (p->args)
 596                 mdoc_argv_free(p->args);
 597         free(p);
 598 }
 599
 600
 601 static void
 602 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 603 {
 604
 605         /* Adjust siblings. */
 606
 607         if (n->prev)
 608                 n->prev->next = n->next;
 609         if (n->next)
 610                 n->next->prev = n->prev;
 611
 612         /* Adjust parent. */
 613
 614         if (n->parent) {
 615                 n->parent->nchild--;
 616                 if (n->parent->child == n)
 617                         n->parent->child = n->prev ? n->prev : n->next;
 618                 if (n->parent->last == n)
 619                         n->parent->last = n->prev ? n->prev : NULL;
 620         }
 621
 622         /* Adjust parse point, if applicable. */
 623
 624         if (m && m->last == n) {
 625                 if (n->prev) {
 626                         m->last = n->prev;
 627                         m->next = MDOC_NEXT_SIBLING;
 628                 } else {
 629                         m->last = n->parent;
 630                         m->next = MDOC_NEXT_CHILD;
 631                 }
 632         }
 633
 634         if (m && m->first == n)
 635                 m->first = NULL;
 636 }
 637
 638
 639 void
 640 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 641 {
 642
 643         while (p->child) {
 644                 assert(p->nchild);
 645                 mdoc_node_delete(m, p->child);
 646         }
 647         assert(0 == p->nchild);
 648
 649         mdoc_node_unlink(m, p);
 650         mdoc_node_free(p);
 651 }
 652
 653
 654 /*
 655  * Parse free-form text, that is, a line that does not begin with the
 656  * control character.
 657  */
 658 static int
 659 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 660 {
 661         char             *c, *ws, *end;
 662         struct mdoc_node *n;
 663
 664         /* No text before an initial macro. */
 665
 666         if (SEC_NONE == m->lastnamed) {
 667                 mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
 668                 return(1);
 669         }
 670
 671         assert(m->last);
 672         n = m->last;
 673
 674         /*
 675          * Divert directly to list processing if we're encountering a
 676          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 677          * (a MDOC_BODY means it's already open, in which case we should
 678          * process within its context in the normal way).
 679          */
 680
 681         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 682                         LIST_column == n->norm->Bl.type) {
 683                 /* `Bl' is open without any children. */
 684                 m->flags |= MDOC_FREECOL;
 685                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 686         }
 687
 688         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 689                         NULL != n->parent &&
 690                         MDOC_Bl == n->parent->tok &&
 691                         LIST_column == n->parent->norm->Bl.type) {
 692                 /* `Bl' has block-level `It' children. */
 693                 m->flags |= MDOC_FREECOL;
 694                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 695         }
 696
 697         /*
 698          * Search for the beginning of unescaped trailing whitespace (ws)
 699          * and for the first character not to be output (end).
 700          */
 701
 702         /* FIXME: replace with strcspn(). */
 703         ws = NULL;
 704         for (c = end = buf + offs; *c; c++) {
 705                 switch (*c) {
 706                 case '-':
 707                         if (mandoc_hyph(buf + offs, c))
 708                                 *c = ASCII_HYPH;
 709                         ws = NULL;
 710                         break;
 711                 case ' ':
 712                         if (NULL == ws)
 713                                 ws = c;
 714                         continue;
 715                 case '\t':
 716                         /*
 717                          * Always warn about trailing tabs,
 718                          * even outside literal context,
 719                          * where they should be put on the next line.
 720                          */
 721                         if (NULL == ws)
 722                                 ws = c;
 723                         /*
 724                          * Strip trailing tabs in literal context only;
 725                          * outside, they affect the next line.
 726                          */
 727                         if (MDOC_LITERAL & m->flags)
 728                                 continue;
 729                         break;
 730                 case '\\':
 731                         /* Skip the escaped character, too, if any. */
 732                         if (c[1])
 733                                 c++;
 734                         /* FALLTHROUGH */
 735                 default:
 736                         ws = NULL;
 737                         break;
 738                 }
 739                 end = c + 1;
 740         }
 741         *end = '\0';
 742
 743         if (ws)
 744                 mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
 745
 746         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 747                 mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
 748
 749                 /*
 750                  * Insert a `sp' in the case of a blank line.  Technically,
 751                  * blank lines aren't allowed, but enough manuals assume this
 752                  * behaviour that we want to work around it.
 753                  */
 754                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 755                         return(0);
 756
 757                 m->next = MDOC_NEXT_SIBLING;
 758                 return(1);
 759         }
 760
 761         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 762                 return(0);
 763
 764         if (MDOC_LITERAL & m->flags)
 765                 return(1);
 766
 767         /*
 768          * End-of-sentence check.  If the last character is an unescaped
 769          * EOS character, then flag the node as being the end of a
 770          * sentence.  The front-end will know how to interpret this.
 771          */
 772
 773         assert(buf < end);
 774
 775         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 776                 m->last->flags |= MDOC_EOS;
 777
 778         return(1);
 779 }
 780
 781
 782 /*
 783  * Parse a macro line, that is, a line beginning with the control
 784  * character.
 785  */
 786 static int
 787 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 788 {
 789         enum mdoct        tok;
 790         int               i, sv;
 791         char              mac[5];
 792         struct mdoc_node *n;
 793
 794         /* Empty post-control lines are ignored. */
 795
 796         if ('"' == buf[offs]) {
 797                 mdoc_pmsg(m, ln, offs, MANDOCERR_BADCOMMENT);
 798                 return(1);
 799         } else if ('\0' == buf[offs])
 800                 return(1);
 801
 802         sv = offs;
 803
 804         /*
 805          * Copy the first word into a nil-terminated buffer.
 806          * Stop copying when a tab, space, or eoln is encountered.
 807          */
 808
 809         i = 0;
 810         while (i < 4 && '\0' != buf[offs] &&
 811                         ' ' != buf[offs] && '\t' != buf[offs])
 812                 mac[i++] = buf[offs++];
 813
 814         mac[i] = '\0';
 815
 816         tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 817
 818         if (MDOC_MAX == tok) {
 819                 mandoc_vmsg(MANDOCERR_MACRO, m->parse,
 820                                 ln, sv, "%s", buf + sv - 1);
 821                 return(1);
 822         }
 823
 824         /* Disregard the first trailing tab, if applicable. */
 825
 826         if ('\t' == buf[offs])
 827                 offs++;
 828
 829         /* Jump to the next non-whitespace word. */
 830
 831         while (buf[offs] && ' ' == buf[offs])
 832                 offs++;
 833
 834         /*
 835          * Trailing whitespace.  Note that tabs are allowed to be passed
 836          * into the parser as "text", so we only warn about spaces here.
 837          */
 838
 839         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 840                 mdoc_pmsg(m, ln, offs - 1, MANDOCERR_EOLNSPACE);
 841
 842         /*
 843          * If an initial macro or a list invocation, divert directly
 844          * into macro processing.
 845          */
 846
 847         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 848                 if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
 849                         goto err;
 850                 return(1);
 851         }
 852
 853         n = m->last;
 854         assert(m->last);
 855
 856         /*
 857          * If the first macro of a `Bl -column', open an `It' block
 858          * context around the parsed macro.
 859          */
 860
 861         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 862                         LIST_column == n->norm->Bl.type) {
 863                 m->flags |= MDOC_FREECOL;
 864                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 865                         goto err;
 866                 return(1);
 867         }
 868
 869         /*
 870          * If we're following a block-level `It' within a `Bl -column'
 871          * context (perhaps opened in the above block or in ptext()),
 872          * then open an `It' block context around the parsed macro.
 873          */
 874
 875         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 876                         NULL != n->parent &&
 877                         MDOC_Bl == n->parent->tok &&
 878                         LIST_column == n->parent->norm->Bl.type) {
 879                 m->flags |= MDOC_FREECOL;
 880                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 881                         goto err;
 882                 return(1);
 883         }
 884
 885         /* Normal processing of a macro. */
 886
 887         if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
 888                 goto err;
 889
 890         return(1);
 891
 892 err:    /* Error out. */
 893
 894         m->flags |= MDOC_HALT;
 895         return(0);
 896 }
 897
 898 enum mdelim
 899 mdoc_isdelim(const char *p)
 900 {
 901
 902         if ('\0' == p[0])
 903                 return(DELIM_NONE);
 904
 905         if ('\0' == p[1])
 906                 switch (p[0]) {
 907                 case('('):
 908                         /* FALLTHROUGH */
 909                 case('['):
 910                         return(DELIM_OPEN);
 911                 case('|'):
 912                         return(DELIM_MIDDLE);
 913                 case('.'):
 914                         /* FALLTHROUGH */
 915                 case(','):
 916                         /* FALLTHROUGH */
 917                 case(';'):
 918                         /* FALLTHROUGH */
 919                 case(':'):
 920                         /* FALLTHROUGH */
 921                 case('?'):
 922                         /* FALLTHROUGH */
 923                 case('!'):
 924                         /* FALLTHROUGH */
 925                 case(')'):
 926                         /* FALLTHROUGH */
 927                 case(']'):
 928                         return(DELIM_CLOSE);
 929                 default:
 930                         return(DELIM_NONE);
 931                 }
 932
 933         if ('\\' != p[0])
 934                 return(DELIM_NONE);
 935
 936         if (0 == strcmp(p + 1, "."))
 937                 return(DELIM_CLOSE);
 938         if (0 == strcmp(p + 1, "*(Ba"))
 939                 return(DELIM_MIDDLE);
 940
 941         return(DELIM_NONE);
 942 }