mdoc.c

   1 /*      $Id: mdoc.c,v 1.201 2012/07/18 11:11:12 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012 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->last->tok = MDOC_MAX;
 164         mdoc->next = MDOC_NEXT_CHILD;
 165 }
 166
 167
 168 /*
 169  * Free up volatile resources (see mdoc_free1()) then re-initialises the
 170  * data with mdoc_alloc1().  After invocation, parse data has been reset
 171  * and the parser is ready for re-invocation on a new tree; however,
 172  * cross-parse non-volatile data is kept intact.
 173  */
 174 void
 175 mdoc_reset(struct mdoc *mdoc)
 176 {
 177
 178         mdoc_free1(mdoc);
 179         mdoc_alloc1(mdoc);
 180 }
 181
 182
 183 /*
 184  * Completely free up all volatile and non-volatile parse resources.
 185  * After invocation, the pointer is no longer usable.
 186  */
 187 void
 188 mdoc_free(struct mdoc *mdoc)
 189 {
 190
 191         mdoc_free1(mdoc);
 192         free(mdoc);
 193 }
 194
 195
 196 /*
 197  * Allocate volatile and non-volatile parse resources.
 198  */
 199 struct mdoc *
 200 mdoc_alloc(struct roff *roff, struct mparse *parse, char *defos)
 201 {
 202         struct mdoc     *p;
 203
 204         p = mandoc_calloc(1, sizeof(struct mdoc));
 205
 206         p->parse = parse;
 207         p->defos = defos;
 208         p->roff = roff;
 209
 210         mdoc_hash_init();
 211         mdoc_alloc1(p);
 212         return(p);
 213 }
 214
 215
 216 /*
 217  * Climb back up the parse tree, validating open scopes.  Mostly calls
 218  * through to macro_end() in macro.c.
 219  */
 220 int
 221 mdoc_endparse(struct mdoc *m)
 222 {
 223
 224         assert( ! (MDOC_HALT & m->flags));
 225         if (mdoc_macroend(m))
 226                 return(1);
 227         m->flags |= MDOC_HALT;
 228         return(0);
 229 }
 230
 231 int
 232 mdoc_addeqn(struct mdoc *m, const struct eqn *ep)
 233 {
 234         struct mdoc_node *n;
 235
 236         assert( ! (MDOC_HALT & m->flags));
 237
 238         /* No text before an initial macro. */
 239
 240         if (SEC_NONE == m->lastnamed) {
 241                 mdoc_pmsg(m, ep->ln, ep->pos, MANDOCERR_NOTEXT);
 242                 return(1);
 243         }
 244
 245         n = node_alloc(m, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN);
 246         n->eqn = ep;
 247
 248         if ( ! node_append(m, n))
 249                 return(0);
 250
 251         m->next = MDOC_NEXT_SIBLING;
 252         return(1);
 253 }
 254
 255 int
 256 mdoc_addspan(struct mdoc *m, const struct tbl_span *sp)
 257 {
 258         struct mdoc_node *n;
 259
 260         assert( ! (MDOC_HALT & m->flags));
 261
 262         /* No text before an initial macro. */
 263
 264         if (SEC_NONE == m->lastnamed) {
 265                 mdoc_pmsg(m, sp->line, 0, MANDOCERR_NOTEXT);
 266                 return(1);
 267         }
 268
 269         n = node_alloc(m, sp->line, 0, MDOC_MAX, MDOC_TBL);
 270         n->span = sp;
 271
 272         if ( ! node_append(m, n))
 273                 return(0);
 274
 275         m->next = MDOC_NEXT_SIBLING;
 276         return(1);
 277 }
 278
 279
 280 /*
 281  * Main parse routine.  Parses a single line -- really just hands off to
 282  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
 283  */
 284 int
 285 mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 286 {
 287
 288         assert( ! (MDOC_HALT & m->flags));
 289
 290         m->flags |= MDOC_NEWLINE;
 291
 292         /*
 293          * Let the roff nS register switch SYNOPSIS mode early,
 294          * such that the parser knows at all times
 295          * whether this mode is on or off.
 296          * Note that this mode is also switched by the Sh macro.
 297          */
 298         if (roff_regisset(m->roff, REG_nS)) {
 299                 if (roff_regget(m->roff, REG_nS))
 300                         m->flags |= MDOC_SYNOPSIS;
 301                 else
 302                         m->flags &= ~MDOC_SYNOPSIS;
 303         }
 304
 305         return(roff_getcontrol(m->roff, buf, &offs) ?
 306                         mdoc_pmacro(m, ln, buf, offs) :
 307                         mdoc_ptext(m, ln, buf, offs));
 308 }
 309
 310 int
 311 mdoc_macro(MACRO_PROT_ARGS)
 312 {
 313         assert(tok < MDOC_MAX);
 314
 315         /* If we're in the body, deny prologue calls. */
 316
 317         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 318                         MDOC_PBODY & m->flags) {
 319                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADBODY);
 320                 return(1);
 321         }
 322
 323         /* If we're in the prologue, deny "body" macros.  */
 324
 325         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 326                         ! (MDOC_PBODY & m->flags)) {
 327                 mdoc_pmsg(m, line, ppos, MANDOCERR_BADPROLOG);
 328                 if (NULL == m->meta.msec)
 329                         m->meta.msec = mandoc_strdup("1");
 330                 if (NULL == m->meta.title)
 331                         m->meta.title = mandoc_strdup("UNKNOWN");
 332                 if (NULL == m->meta.vol)
 333                         m->meta.vol = mandoc_strdup("LOCAL");
 334                 if (NULL == m->meta.os)
 335                         m->meta.os = mandoc_strdup("LOCAL");
 336                 if (NULL == m->meta.date)
 337                         m->meta.date = mandoc_normdate
 338                                 (m->parse, NULL, line, ppos);
 339                 m->flags |= MDOC_PBODY;
 340         }
 341
 342         return((*mdoc_macros[tok].fp)(m, tok, line, ppos, pos, buf));
 343 }
 344
 345
 346 static int
 347 node_append(struct mdoc *mdoc, struct mdoc_node *p)
 348 {
 349
 350         assert(mdoc->last);
 351         assert(mdoc->first);
 352         assert(MDOC_ROOT != p->type);
 353
 354         switch (mdoc->next) {
 355         case (MDOC_NEXT_SIBLING):
 356                 mdoc->last->next = p;
 357                 p->prev = mdoc->last;
 358                 p->parent = mdoc->last->parent;
 359                 break;
 360         case (MDOC_NEXT_CHILD):
 361                 mdoc->last->child = p;
 362                 p->parent = mdoc->last;
 363                 break;
 364         default:
 365                 abort();
 366                 /* NOTREACHED */
 367         }
 368
 369         p->parent->nchild++;
 370
 371         /*
 372          * Copy over the normalised-data pointer of our parent.  Not
 373          * everybody has one, but copying a null pointer is fine.
 374          */
 375
 376         switch (p->type) {
 377         case (MDOC_BODY):
 378                 /* FALLTHROUGH */
 379         case (MDOC_TAIL):
 380                 /* FALLTHROUGH */
 381         case (MDOC_HEAD):
 382                 p->norm = p->parent->norm;
 383                 break;
 384         default:
 385                 break;
 386         }
 387
 388         if ( ! mdoc_valid_pre(mdoc, p))
 389                 return(0);
 390
 391         switch (p->type) {
 392         case (MDOC_HEAD):
 393                 assert(MDOC_BLOCK == p->parent->type);
 394                 p->parent->head = p;
 395                 break;
 396         case (MDOC_TAIL):
 397                 assert(MDOC_BLOCK == p->parent->type);
 398                 p->parent->tail = p;
 399                 break;
 400         case (MDOC_BODY):
 401                 if (p->end)
 402                         break;
 403                 assert(MDOC_BLOCK == p->parent->type);
 404                 p->parent->body = p;
 405                 break;
 406         default:
 407                 break;
 408         }
 409
 410         mdoc->last = p;
 411
 412         switch (p->type) {
 413         case (MDOC_TBL):
 414                 /* FALLTHROUGH */
 415         case (MDOC_TEXT):
 416                 if ( ! mdoc_valid_post(mdoc))
 417                         return(0);
 418                 break;
 419         default:
 420                 break;
 421         }
 422
 423         return(1);
 424 }
 425
 426
 427 static struct mdoc_node *
 428 node_alloc(struct mdoc *m, int line, int pos,
 429                 enum mdoct tok, enum mdoc_type type)
 430 {
 431         struct mdoc_node *p;
 432
 433         p = mandoc_calloc(1, sizeof(struct mdoc_node));
 434         p->sec = m->lastsec;
 435         p->line = line;
 436         p->pos = pos;
 437         p->tok = tok;
 438         p->type = type;
 439
 440         /* Flag analysis. */
 441
 442         if (MDOC_SYNOPSIS & m->flags)
 443                 p->flags |= MDOC_SYNPRETTY;
 444         else
 445                 p->flags &= ~MDOC_SYNPRETTY;
 446         if (MDOC_NEWLINE & m->flags)
 447                 p->flags |= MDOC_LINE;
 448         m->flags &= ~MDOC_NEWLINE;
 449
 450         return(p);
 451 }
 452
 453
 454 int
 455 mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 456 {
 457         struct mdoc_node *p;
 458
 459         p = node_alloc(m, line, pos, tok, MDOC_TAIL);
 460         if ( ! node_append(m, p))
 461                 return(0);
 462         m->next = MDOC_NEXT_CHILD;
 463         return(1);
 464 }
 465
 466
 467 int
 468 mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 469 {
 470         struct mdoc_node *p;
 471
 472         assert(m->first);
 473         assert(m->last);
 474
 475         p = node_alloc(m, line, pos, tok, MDOC_HEAD);
 476         if ( ! node_append(m, p))
 477                 return(0);
 478         m->next = MDOC_NEXT_CHILD;
 479         return(1);
 480 }
 481
 482
 483 int
 484 mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 485 {
 486         struct mdoc_node *p;
 487
 488         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 489         if ( ! node_append(m, p))
 490                 return(0);
 491         m->next = MDOC_NEXT_CHILD;
 492         return(1);
 493 }
 494
 495
 496 int
 497 mdoc_endbody_alloc(struct mdoc *m, int line, int pos, enum mdoct tok,
 498                 struct mdoc_node *body, enum mdoc_endbody end)
 499 {
 500         struct mdoc_node *p;
 501
 502         p = node_alloc(m, line, pos, tok, MDOC_BODY);
 503         p->pending = body;
 504         p->end = end;
 505         if ( ! node_append(m, p))
 506                 return(0);
 507         m->next = MDOC_NEXT_SIBLING;
 508         return(1);
 509 }
 510
 511
 512 int
 513 mdoc_block_alloc(struct mdoc *m, int line, int pos,
 514                 enum mdoct tok, struct mdoc_arg *args)
 515 {
 516         struct mdoc_node *p;
 517
 518         p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
 519         p->args = args;
 520         if (p->args)
 521                 (args->refcnt)++;
 522
 523         switch (tok) {
 524         case (MDOC_Bd):
 525                 /* FALLTHROUGH */
 526         case (MDOC_Bf):
 527                 /* FALLTHROUGH */
 528         case (MDOC_Bl):
 529                 /* FALLTHROUGH */
 530         case (MDOC_Rs):
 531                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 532                 break;
 533         default:
 534                 break;
 535         }
 536
 537         if ( ! node_append(m, p))
 538                 return(0);
 539         m->next = MDOC_NEXT_CHILD;
 540         return(1);
 541 }
 542
 543
 544 int
 545 mdoc_elem_alloc(struct mdoc *m, int line, int pos,
 546                 enum mdoct tok, struct mdoc_arg *args)
 547 {
 548         struct mdoc_node *p;
 549
 550         p = node_alloc(m, line, pos, tok, MDOC_ELEM);
 551         p->args = args;
 552         if (p->args)
 553                 (args->refcnt)++;
 554
 555         switch (tok) {
 556         case (MDOC_An):
 557                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 558                 break;
 559         default:
 560                 break;
 561         }
 562
 563         if ( ! node_append(m, p))
 564                 return(0);
 565         m->next = MDOC_NEXT_CHILD;
 566         return(1);
 567 }
 568
 569 int
 570 mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 571 {
 572         struct mdoc_node *n;
 573
 574         n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
 575         n->string = roff_strdup(m->roff, p);
 576
 577         if ( ! node_append(m, n))
 578                 return(0);
 579
 580         m->next = MDOC_NEXT_SIBLING;
 581         return(1);
 582 }
 583
 584
 585 static void
 586 mdoc_node_free(struct mdoc_node *p)
 587 {
 588
 589         if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type)
 590                 free(p->norm);
 591         if (p->string)
 592                 free(p->string);
 593         if (p->args)
 594                 mdoc_argv_free(p->args);
 595         free(p);
 596 }
 597
 598
 599 static void
 600 mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 601 {
 602
 603         /* Adjust siblings. */
 604
 605         if (n->prev)
 606                 n->prev->next = n->next;
 607         if (n->next)
 608                 n->next->prev = n->prev;
 609
 610         /* Adjust parent. */
 611
 612         if (n->parent) {
 613                 n->parent->nchild--;
 614                 if (n->parent->child == n)
 615                         n->parent->child = n->prev ? n->prev : n->next;
 616                 if (n->parent->last == n)
 617                         n->parent->last = n->prev ? n->prev : NULL;
 618         }
 619
 620         /* Adjust parse point, if applicable. */
 621
 622         if (m && m->last == n) {
 623                 if (n->prev) {
 624                         m->last = n->prev;
 625                         m->next = MDOC_NEXT_SIBLING;
 626                 } else {
 627                         m->last = n->parent;
 628                         m->next = MDOC_NEXT_CHILD;
 629                 }
 630         }
 631
 632         if (m && m->first == n)
 633                 m->first = NULL;
 634 }
 635
 636
 637 void
 638 mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 639 {
 640
 641         while (p->child) {
 642                 assert(p->nchild);
 643                 mdoc_node_delete(m, p->child);
 644         }
 645         assert(0 == p->nchild);
 646
 647         mdoc_node_unlink(m, p);
 648         mdoc_node_free(p);
 649 }
 650
 651 int
 652 mdoc_node_relink(struct mdoc *m, struct mdoc_node *p)
 653 {
 654
 655         mdoc_node_unlink(m, p);
 656         return(node_append(m, p));
 657 }
 658
 659 #if 0
 660 /*
 661  * Pre-treat a text line.
 662  * Text lines can consist of equations, which must be handled apart from
 663  * the regular text.
 664  * Thus, use this function to step through a line checking if it has any
 665  * equations embedded in it.
 666  * This must handle multiple equations AND equations that do not end at
 667  * the end-of-line, i.e., will re-enter in the next roff parse.
 668  */
 669 static int
 670 mdoc_preptext(struct mdoc *m, int line, char *buf, int offs)
 671 {
 672         char            *start, *end;
 673         char             delim;
 674
 675         while ('\0' != buf[offs]) {
 676                 /* Mark starting position if eqn is set. */
 677                 start = NULL;
 678                 if ('\0' != (delim = roff_eqndelim(m->roff)))
 679                         if (NULL != (start = strchr(buf + offs, delim)))
 680                                 *start++ = '\0';
 681
 682                 /* Parse text as normal. */
 683                 if ( ! mdoc_ptext(m, line, buf, offs))
 684                         return(0);
 685
 686                 /* Continue only if an equation exists. */
 687                 if (NULL == start)
 688                         break;
 689
 690                 /* Read past the end of the equation. */
 691                 offs += start - (buf + offs);
 692                 assert(start == &buf[offs]);
 693                 if (NULL != (end = strchr(buf + offs, delim))) {
 694                         *end++ = '\0';
 695                         while (' ' == *end)
 696                                 end++;
 697                 }
 698
 699                 /* Parse the equation itself. */
 700                 roff_openeqn(m->roff, NULL, line, offs, buf);
 701
 702                 /* Process a finished equation? */
 703                 if (roff_closeeqn(m->roff))
 704                         if ( ! mdoc_addeqn(m, roff_eqn(m->roff)))
 705                                 return(0);
 706                 offs += (end - (buf + offs));
 707         }
 708
 709         return(1);
 710 }
 711 #endif
 712
 713 /*
 714  * Parse free-form text, that is, a line that does not begin with the
 715  * control character.
 716  */
 717 static int
 718 mdoc_ptext(struct mdoc *m, int line, char *buf, int offs)
 719 {
 720         char             *c, *ws, *end;
 721         struct mdoc_node *n;
 722
 723         /* No text before an initial macro. */
 724
 725         if (SEC_NONE == m->lastnamed) {
 726                 mdoc_pmsg(m, line, offs, MANDOCERR_NOTEXT);
 727                 return(1);
 728         }
 729
 730         assert(m->last);
 731         n = m->last;
 732
 733         /*
 734          * Divert directly to list processing if we're encountering a
 735          * columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry
 736          * (a MDOC_BODY means it's already open, in which case we should
 737          * process within its context in the normal way).
 738          */
 739
 740         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 741                         LIST_column == n->norm->Bl.type) {
 742                 /* `Bl' is open without any children. */
 743                 m->flags |= MDOC_FREECOL;
 744                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 745         }
 746
 747         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 748                         NULL != n->parent &&
 749                         MDOC_Bl == n->parent->tok &&
 750                         LIST_column == n->parent->norm->Bl.type) {
 751                 /* `Bl' has block-level `It' children. */
 752                 m->flags |= MDOC_FREECOL;
 753                 return(mdoc_macro(m, MDOC_It, line, offs, &offs, buf));
 754         }
 755
 756         /*
 757          * Search for the beginning of unescaped trailing whitespace (ws)
 758          * and for the first character not to be output (end).
 759          */
 760
 761         /* FIXME: replace with strcspn(). */
 762         ws = NULL;
 763         for (c = end = buf + offs; *c; c++) {
 764                 switch (*c) {
 765                 case ' ':
 766                         if (NULL == ws)
 767                                 ws = c;
 768                         continue;
 769                 case '\t':
 770                         /*
 771                          * Always warn about trailing tabs,
 772                          * even outside literal context,
 773                          * where they should be put on the next line.
 774                          */
 775                         if (NULL == ws)
 776                                 ws = c;
 777                         /*
 778                          * Strip trailing tabs in literal context only;
 779                          * outside, they affect the next line.
 780                          */
 781                         if (MDOC_LITERAL & m->flags)
 782                                 continue;
 783                         break;
 784                 case '\\':
 785                         /* Skip the escaped character, too, if any. */
 786                         if (c[1])
 787                                 c++;
 788                         /* FALLTHROUGH */
 789                 default:
 790                         ws = NULL;
 791                         break;
 792                 }
 793                 end = c + 1;
 794         }
 795         *end = '\0';
 796
 797         if (ws)
 798                 mdoc_pmsg(m, line, (int)(ws-buf), MANDOCERR_EOLNSPACE);
 799
 800         if ('\0' == buf[offs] && ! (MDOC_LITERAL & m->flags)) {
 801                 mdoc_pmsg(m, line, (int)(c-buf), MANDOCERR_NOBLANKLN);
 802
 803                 /*
 804                  * Insert a `sp' in the case of a blank line.  Technically,
 805                  * blank lines aren't allowed, but enough manuals assume this
 806                  * behaviour that we want to work around it.
 807                  */
 808                 if ( ! mdoc_elem_alloc(m, line, offs, MDOC_sp, NULL))
 809                         return(0);
 810
 811                 m->next = MDOC_NEXT_SIBLING;
 812
 813                 return(mdoc_valid_post(m));
 814         }
 815
 816         if ( ! mdoc_word_alloc(m, line, offs, buf+offs))
 817                 return(0);
 818
 819         if (MDOC_LITERAL & m->flags)
 820                 return(1);
 821
 822         /*
 823          * End-of-sentence check.  If the last character is an unescaped
 824          * EOS character, then flag the node as being the end of a
 825          * sentence.  The front-end will know how to interpret this.
 826          */
 827
 828         assert(buf < end);
 829
 830         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs), 0))
 831                 m->last->flags |= MDOC_EOS;
 832
 833         return(1);
 834 }
 835
 836
 837 /*
 838  * Parse a macro line, that is, a line beginning with the control
 839  * character.
 840  */
 841 static int
 842 mdoc_pmacro(struct mdoc *m, int ln, char *buf, int offs)
 843 {
 844         enum mdoct        tok;
 845         int               i, sv;
 846         char              mac[5];
 847         struct mdoc_node *n;
 848
 849         /* Empty post-control lines are ignored. */
 850
 851         if ('"' == buf[offs]) {
 852                 mdoc_pmsg(m, ln, offs, MANDOCERR_BADCOMMENT);
 853                 return(1);
 854         } else if ('\0' == buf[offs])
 855                 return(1);
 856
 857         sv = offs;
 858
 859         /*
 860          * Copy the first word into a nil-terminated buffer.
 861          * Stop copying when a tab, space, or eoln is encountered.
 862          */
 863
 864         i = 0;
 865         while (i < 4 && '\0' != buf[offs] &&
 866                         ' ' != buf[offs] && '\t' != buf[offs])
 867                 mac[i++] = buf[offs++];
 868
 869         mac[i] = '\0';
 870
 871         tok = (i > 1 || i < 4) ? mdoc_hash_find(mac) : MDOC_MAX;
 872
 873         if (MDOC_MAX == tok) {
 874                 mandoc_vmsg(MANDOCERR_MACRO, m->parse,
 875                                 ln, sv, "%s", buf + sv - 1);
 876                 return(1);
 877         }
 878
 879         /* Disregard the first trailing tab, if applicable. */
 880
 881         if ('\t' == buf[offs])
 882                 offs++;
 883
 884         /* Jump to the next non-whitespace word. */
 885
 886         while (buf[offs] && ' ' == buf[offs])
 887                 offs++;
 888
 889         /*
 890          * Trailing whitespace.  Note that tabs are allowed to be passed
 891          * into the parser as "text", so we only warn about spaces here.
 892          */
 893
 894         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 895                 mdoc_pmsg(m, ln, offs - 1, MANDOCERR_EOLNSPACE);
 896
 897         /*
 898          * If an initial macro or a list invocation, divert directly
 899          * into macro processing.
 900          */
 901
 902         if (NULL == m->last || MDOC_It == tok || MDOC_El == tok) {
 903                 if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
 904                         goto err;
 905                 return(1);
 906         }
 907
 908         n = m->last;
 909         assert(m->last);
 910
 911         /*
 912          * If the first macro of a `Bl -column', open an `It' block
 913          * context around the parsed macro.
 914          */
 915
 916         if (MDOC_Bl == n->tok && MDOC_BODY == n->type &&
 917                         LIST_column == n->norm->Bl.type) {
 918                 m->flags |= MDOC_FREECOL;
 919                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 920                         goto err;
 921                 return(1);
 922         }
 923
 924         /*
 925          * If we're following a block-level `It' within a `Bl -column'
 926          * context (perhaps opened in the above block or in ptext()),
 927          * then open an `It' block context around the parsed macro.
 928          */
 929
 930         if (MDOC_It == n->tok && MDOC_BLOCK == n->type &&
 931                         NULL != n->parent &&
 932                         MDOC_Bl == n->parent->tok &&
 933                         LIST_column == n->parent->norm->Bl.type) {
 934                 m->flags |= MDOC_FREECOL;
 935                 if ( ! mdoc_macro(m, MDOC_It, ln, sv, &sv, buf))
 936                         goto err;
 937                 return(1);
 938         }
 939
 940         /* Normal processing of a macro. */
 941
 942         if ( ! mdoc_macro(m, tok, ln, sv, &offs, buf))
 943                 goto err;
 944
 945         return(1);
 946
 947 err:    /* Error out. */
 948
 949         m->flags |= MDOC_HALT;
 950         return(0);
 951 }
 952
 953 enum mdelim
 954 mdoc_isdelim(const char *p)
 955 {
 956
 957         if ('\0' == p[0])
 958                 return(DELIM_NONE);
 959
 960         if ('\0' == p[1])
 961                 switch (p[0]) {
 962                 case('('):
 963                         /* FALLTHROUGH */
 964                 case('['):
 965                         return(DELIM_OPEN);
 966                 case('|'):
 967                         return(DELIM_MIDDLE);
 968                 case('.'):
 969                         /* FALLTHROUGH */
 970                 case(','):
 971                         /* FALLTHROUGH */
 972                 case(';'):
 973                         /* FALLTHROUGH */
 974                 case(':'):
 975                         /* FALLTHROUGH */
 976                 case('?'):
 977                         /* FALLTHROUGH */
 978                 case('!'):
 979                         /* FALLTHROUGH */
 980                 case(')'):
 981                         /* FALLTHROUGH */
 982                 case(']'):
 983                         return(DELIM_CLOSE);
 984                 default:
 985                         return(DELIM_NONE);
 986                 }
 987
 988         if ('\\' != p[0])
 989                 return(DELIM_NONE);
 990
 991         if (0 == strcmp(p + 1, "."))
 992                 return(DELIM_CLOSE);
 993         if (0 == strcmp(p + 1, "fR|\\fP"))
 994                 return(DELIM_MIDDLE);
 995
 996         return(DELIM_NONE);
 997 }