mdoc.c

   1 /* $Id: mdoc.c,v 1.48 2009/02/23 12:45:19 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the
   7  * above copyright notice and this permission notice appear in all
   8  * copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL
  11  * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED
  12  * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE
  13  * AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
  14  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
  15  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
  16  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  17  * PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #include <assert.h>
  20 #include <ctype.h>
  21 #include <err.h>
  22 #include <stdarg.h>
  23 #include <stdlib.h>
  24 #include <stdio.h>
  25 #include <string.h>
  26
  27 #include "private.h"
  28
  29 /*
  30  * Main caller in the libmdoc library.  This begins the parsing routine,
  31  * handles allocation of data, and so forth.  Most of the "work" is done
  32  * in macro.c and validate.c.
  33  */
  34
  35 static  struct mdoc_arg  *argdup(size_t, const struct mdoc_arg *);
  36 static  void              argfree(size_t, struct mdoc_arg *);
  37 static  void              argcpy(struct mdoc_arg *,
  38                                 const struct mdoc_arg *);
  39
  40 static  struct mdoc_node *mdoc_node_alloc(const struct mdoc *);
  41 static  int               mdoc_node_append(struct mdoc *,
  42                                 struct mdoc_node *);
  43 static  void              mdoc_elem_free(struct mdoc_elem *);
  44 static  void              mdoc_text_free(struct mdoc_text *);
  45
  46
  47 const   char *const __mdoc_macronames[MDOC_MAX] = {
  48         "\\\"",         "Dd",           "Dt",           "Os",
  49         "Sh",           "Ss",           "Pp",           "D1",
  50         "Dl",           "Bd",           "Ed",           "Bl",
  51         "El",           "It",           "Ad",           "An",
  52         "Ar",           "Cd",           "Cm",           "Dv",
  53         "Er",           "Ev",           "Ex",           "Fa",
  54         "Fd",           "Fl",           "Fn",           "Ft",
  55         "Ic",           "In",           "Li",           "Nd",
  56         "Nm",           "Op",           "Ot",           "Pa",
  57         "Rv",           "St",           "Va",           "Vt",
  58         /* LINTED */
  59         "Xr",           "\%A",          "\%B",          "\%D",
  60         /* LINTED */
  61         "\%I",          "\%J",          "\%N",          "\%O",
  62         /* LINTED */
  63         "\%P",          "\%R",          "\%T",          "\%V",
  64         "Ac",           "Ao",           "Aq",           "At",
  65         "Bc",           "Bf",           "Bo",           "Bq",
  66         "Bsx",          "Bx",           "Db",           "Dc",
  67         "Do",           "Dq",           "Ec",           "Ef",
  68         "Em",           "Eo",           "Fx",           "Ms",
  69         "No",           "Ns",           "Nx",           "Ox",
  70         "Pc",           "Pf",           "Po",           "Pq",
  71         "Qc",           "Ql",           "Qo",           "Qq",
  72         "Re",           "Rs",           "Sc",           "So",
  73         "Sq",           "Sm",           "Sx",           "Sy",
  74         "Tn",           "Ux",           "Xc",           "Xo",
  75         "Fo",           "Fc",           "Oo",           "Oc",
  76         "Bk",           "Ek",           "Bt",           "Hf",
  77         "Fr",           "Ud",
  78         };
  79
  80 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  81         "split",                "nosplit",              "ragged",
  82         "unfilled",             "literal",              "file",
  83         "offset",               "bullet",               "dash",
  84         "hyphen",               "item",                 "enum",
  85         "tag",                  "diag",                 "hang",
  86         "ohang",                "inset",                "column",
  87         "width",                "compact",              "std",
  88         "p1003.1-88",           "p1003.1-90",           "p1003.1-96",
  89         "p1003.1-2001",         "p1003.1-2004",         "p1003.1",
  90         "p1003.1b",             "p1003.1b-93",          "p1003.1c-95",
  91         "p1003.1g-2000",        "p1003.2-92",           "p1387.2-95",
  92         "p1003.2",              "p1387.2",              "isoC-90",
  93         "isoC-amd1",            "isoC-tcor1",           "isoC-tcor2",
  94         "isoC-99",              "ansiC",                "ansiC-89",
  95         "ansiC-99",             "ieee754",              "iso8802-3",
  96         "xpg3",                 "xpg4",                 "xpg4.2",
  97         "xpg4.3",               "xbd5",                 "xcu5",
  98         "xsh5",                 "xns5",                 "xns5.2d2.0",
  99         "xcurses4.2",           "susv2",                "susv3",
 100         "svid4",                "filled",               "words",
 101         "emphasis",             "symbolic",
 102         };
 103
 104 const   char * const *mdoc_macronames = __mdoc_macronames;
 105 const   char * const *mdoc_argnames = __mdoc_argnames;
 106
 107
 108 const struct mdoc_node *
 109 mdoc_node(const struct mdoc *mdoc)
 110 {
 111
 112         return(mdoc->first);
 113 }
 114
 115
 116 const struct mdoc_meta *
 117 mdoc_meta(const struct mdoc *mdoc)
 118 {
 119
 120         return(&mdoc->meta);
 121 }
 122
 123
 124 void
 125 mdoc_free(struct mdoc *mdoc)
 126 {
 127
 128         if (mdoc->first)
 129                 mdoc_node_freelist(mdoc->first);
 130         if (mdoc->htab)
 131                 mdoc_tokhash_free(mdoc->htab);
 132         if (mdoc->meta.title)
 133                 free(mdoc->meta.title);
 134         if (mdoc->meta.os)
 135                 free(mdoc->meta.os);
 136         if (mdoc->meta.name)
 137                 free(mdoc->meta.name);
 138
 139         free(mdoc);
 140 }
 141
 142
 143 struct mdoc *
 144 mdoc_alloc(void *data, const struct mdoc_cb *cb)
 145 {
 146         struct mdoc     *p;
 147
 148         p = xcalloc(1, sizeof(struct mdoc));
 149
 150         p->data = data;
 151         if (cb)
 152                 (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
 153
 154         p->last = xcalloc(1, sizeof(struct mdoc_node));
 155         p->last->type = MDOC_ROOT;
 156         p->first = p->last;
 157
 158         p->next = MDOC_NEXT_CHILD;
 159         p->htab = mdoc_tokhash_alloc();
 160
 161         return(p);
 162 }
 163
 164
 165 int
 166 mdoc_endparse(struct mdoc *mdoc)
 167 {
 168
 169         if (MDOC_HALT & mdoc->flags)
 170                 return(0);
 171         if (NULL == mdoc->first)
 172                 return(1);
 173
 174         assert(mdoc->last);
 175         if ( ! macro_end(mdoc)) {
 176                 mdoc->flags |= MDOC_HALT;
 177                 return(0);
 178         }
 179         return(1);
 180 }
 181
 182
 183 int
 184 mdoc_parseln(struct mdoc *mdoc, int line, char *buf)
 185 {
 186         int               c, i;
 187         char              tmp[5];
 188
 189         if (MDOC_HALT & mdoc->flags)
 190                 return(0);
 191
 192         mdoc->linetok = 0;
 193
 194         if ('.' != *buf) {
 195                 if (SEC_PROLOGUE == mdoc->lastnamed)
 196                         return(mdoc_perr(mdoc, line, 0, "text disallowed in document prologue"));
 197                 if ( ! mdoc_word_alloc(mdoc, line, 0, buf))
 198                         return(0);
 199                 mdoc->next = MDOC_NEXT_SIBLING;
 200                 return(1);
 201         }
 202
 203         if (buf[1] && '\\' == buf[1])
 204                 if (buf[2] && '\"' == buf[2])
 205                         return(1);
 206
 207         i = 1;
 208         while (buf[i] && ! isspace((int)buf[i]) && i < (int)sizeof(tmp))
 209                 i++;
 210
 211         if (i == (int)sizeof(tmp)) {
 212                 mdoc->flags |= MDOC_HALT;
 213                 return(mdoc_perr(mdoc, line, 1, "unknown macro"));
 214         } else if (i <= 2) {
 215                 mdoc->flags |= MDOC_HALT;
 216                 return(mdoc_perr(mdoc, line, 1, "unknown macro"));
 217         }
 218
 219         i--;
 220
 221         (void)memcpy(tmp, buf + 1, (size_t)i);
 222         tmp[i++] = 0;
 223
 224         if (MDOC_MAX == (c = mdoc_find(mdoc, tmp))) {
 225                 mdoc->flags |= MDOC_HALT;
 226                 return(mdoc_perr(mdoc, line, 1, "unknown macro"));
 227         }
 228
 229         while (buf[i] && isspace((int)buf[i]))
 230                 i++;
 231
 232         if ( ! mdoc_macro(mdoc, c, line, 1, &i, buf)) {
 233                 mdoc->flags |= MDOC_HALT;
 234                 return(0);
 235         }
 236         return(1);
 237 }
 238
 239
 240 void
 241 mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
 242 {
 243         char              buf[256];
 244         va_list           ap;
 245
 246         if (NULL == mdoc->cb.mdoc_msg)
 247                 return;
 248
 249         va_start(ap, fmt);
 250         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 251         va_end(ap);
 252         (*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf);
 253 }
 254
 255
 256 int
 257 mdoc_verr(struct mdoc *mdoc, int ln, int pos,
 258                 const char *fmt, ...)
 259 {
 260         char             buf[256];
 261         va_list          ap;
 262
 263         if (NULL == mdoc->cb.mdoc_err)
 264                 return(0);
 265
 266         va_start(ap, fmt);
 267         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 268         va_end(ap);
 269         return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
 270 }
 271
 272
 273 int
 274 mdoc_vwarn(struct mdoc *mdoc, int ln, int pos,
 275                 enum mdoc_warn type, const char *fmt, ...)
 276 {
 277         char             buf[256];
 278         va_list          ap;
 279
 280         if (NULL == mdoc->cb.mdoc_warn)
 281                 return(0);
 282
 283         va_start(ap, fmt);
 284         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 285         va_end(ap);
 286         return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf));
 287 }
 288
 289
 290 int
 291 mdoc_macro(struct mdoc *mdoc, int tok,
 292                 int ln, int ppos, int *pos, char *buf)
 293 {
 294
 295         assert(mdoc_macros[tok].fp);
 296
 297         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 298                         SEC_PROLOGUE != mdoc->lastnamed)
 299                 return(mdoc_perr(mdoc, ln, ppos, "macro disallowed in document body"));
 300         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 301                         SEC_PROLOGUE == mdoc->lastnamed)
 302                 return(mdoc_perr(mdoc, ln, ppos, "macro disallowed in document prologue"));
 303         if (1 != ppos && ! (MDOC_CALLABLE & mdoc_macros[tok].flags))
 304                 return(mdoc_perr(mdoc, ln, ppos, "macro not callable"));
 305         return((*mdoc_macros[tok].fp)(mdoc, tok, ln, ppos, pos, buf));
 306 }
 307
 308
 309 static int
 310 mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p)
 311 {
 312         const char       *nn, *nt, *on, *ot, *act;
 313
 314         assert(mdoc->last);
 315         assert(mdoc->first);
 316         assert(MDOC_ROOT != p->type);
 317
 318         /* See if we exceed the suggest line-max. */
 319
 320         switch (p->type) {
 321         case (MDOC_TEXT):
 322                 /* FALLTHROUGH */
 323         case (MDOC_ELEM):
 324                 /* FALLTHROUGH */
 325         case (MDOC_BLOCK):
 326                 mdoc->linetok++;
 327                 break;
 328         default:
 329                 break;
 330         }
 331
 332         /* This sort-of works (re-opening of text macros...). */
 333         if (mdoc->linetok > MDOC_LINEARG_SOFTMAX)
 334                 if ( ! mdoc_nwarn(mdoc, p, WARN_COMPAT,
 335                                         "suggested %d tokens per line exceeded (has %d)",
 336                                         MDOC_LINEARG_SOFTMAX, mdoc->linetok))
 337                         return(0);
 338
 339         if (MDOC_TEXT == mdoc->last->type)
 340                 on = "<text>";
 341         else if (MDOC_ROOT == mdoc->last->type)
 342                 on = "<root>";
 343         else
 344                 on = mdoc_macronames[mdoc->last->tok];
 345
 346         if (MDOC_TEXT == p->type)
 347                 nn = "<text>";
 348         else if (MDOC_ROOT == p->type)
 349                 nn = "<root>";
 350         else
 351                 nn = mdoc_macronames[p->tok];
 352
 353         ot = mdoc_type2a(mdoc->last->type);
 354         nt = mdoc_type2a(p->type);
 355
 356         switch (mdoc->next) {
 357         case (MDOC_NEXT_SIBLING):
 358                 mdoc->last->next = p;
 359                 p->prev = mdoc->last;
 360                 p->parent = mdoc->last->parent;
 361                 act = "sibling";
 362                 break;
 363         case (MDOC_NEXT_CHILD):
 364                 mdoc->last->child = p;
 365                 p->parent = mdoc->last;
 366                 act = "child";
 367                 break;
 368         default:
 369                 abort();
 370                 /* NOTREACHED */
 371         }
 372
 373         if ( ! mdoc_valid_pre(mdoc, p))
 374                 return(0);
 375
 376         switch (p->type) {
 377         case (MDOC_HEAD):
 378                 assert(MDOC_BLOCK == p->parent->type);
 379                 p->parent->data.block.head = p;
 380                 break;
 381         case (MDOC_TAIL):
 382                 assert(MDOC_BLOCK == p->parent->type);
 383                 p->parent->data.block.tail = p;
 384                 break;
 385         case (MDOC_BODY):
 386                 assert(MDOC_BLOCK == p->parent->type);
 387                 p->parent->data.block.body = p;
 388                 break;
 389         default:
 390                 break;
 391         }
 392
 393         mdoc->last = p;
 394         mdoc_msg(mdoc, "parse: %s `%s' %s of %s `%s'",
 395                         nt, nn, act, ot, on);
 396         return(1);
 397 }
 398
 399
 400 static struct mdoc_node *
 401 mdoc_node_alloc(const struct mdoc *mdoc)
 402 {
 403         struct mdoc_node *p;
 404
 405         p = xcalloc(1, sizeof(struct mdoc_node));
 406         p->sec = mdoc->lastsec;
 407
 408         return(p);
 409 }
 410
 411
 412 int
 413 mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 414 {
 415         struct mdoc_node *p;
 416
 417         assert(mdoc->first);
 418         assert(mdoc->last);
 419
 420         p = mdoc_node_alloc(mdoc);
 421
 422         p->line = line;
 423         p->pos = pos;
 424         p->type = MDOC_TAIL;
 425         p->tok = tok;
 426
 427         return(mdoc_node_append(mdoc, p));
 428 }
 429
 430
 431 int
 432 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 433 {
 434         struct mdoc_node *p;
 435
 436         assert(mdoc->first);
 437         assert(mdoc->last);
 438
 439         p = mdoc_node_alloc(mdoc);
 440
 441         p->line = line;
 442         p->pos = pos;
 443         p->type = MDOC_HEAD;
 444         p->tok = tok;
 445
 446         return(mdoc_node_append(mdoc, p));
 447 }
 448
 449
 450 int
 451 mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 452 {
 453         struct mdoc_node *p;
 454
 455         assert(mdoc->first);
 456         assert(mdoc->last);
 457
 458         p = mdoc_node_alloc(mdoc);
 459
 460         p->line = line;
 461         p->pos = pos;
 462         p->type = MDOC_BODY;
 463         p->tok = tok;
 464
 465         return(mdoc_node_append(mdoc, p));
 466 }
 467
 468
 469 int
 470 mdoc_root_alloc(struct mdoc *mdoc)
 471 {
 472         struct mdoc_node *p;
 473
 474         p = mdoc_node_alloc(mdoc);
 475
 476         p->type = MDOC_ROOT;
 477
 478         return(mdoc_node_append(mdoc, p));
 479 }
 480
 481
 482 int
 483 mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
 484                 int tok, size_t argsz, const struct mdoc_arg *args)
 485 {
 486         struct mdoc_node *p;
 487
 488         p = mdoc_node_alloc(mdoc);
 489
 490         p->pos = pos;
 491         p->line = line;
 492         p->type = MDOC_BLOCK;
 493         p->tok = tok;
 494         p->data.block.argc = argsz;
 495         p->data.block.argv = argdup(argsz, args);
 496
 497         return(mdoc_node_append(mdoc, p));
 498 }
 499
 500
 501 int
 502 mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
 503                 int tok, size_t argsz, const struct mdoc_arg *args)
 504 {
 505         struct mdoc_node *p;
 506
 507         p = mdoc_node_alloc(mdoc);
 508
 509         p->line = line;
 510         p->pos = pos;
 511         p->type = MDOC_ELEM;
 512         p->tok = tok;
 513         p->data.elem.argc = argsz;
 514         p->data.elem.argv = argdup(argsz, args);
 515
 516         return(mdoc_node_append(mdoc, p));
 517 }
 518
 519
 520 int
 521 mdoc_word_alloc(struct mdoc *mdoc,
 522                 int line, int pos, const char *word)
 523 {
 524         struct mdoc_node *p;
 525
 526         p = mdoc_node_alloc(mdoc);
 527
 528         p->line = line;
 529         p->pos = pos;
 530         p->type = MDOC_TEXT;
 531         p->data.text.string = xstrdup(word);
 532
 533         return(mdoc_node_append(mdoc, p));
 534 }
 535
 536
 537 static void
 538 argfree(size_t sz, struct mdoc_arg *p)
 539 {
 540         int              i, j;
 541
 542         if (0 == sz)
 543                 return;
 544
 545         assert(p);
 546         /* LINTED */
 547         for (i = 0; i < (int)sz; i++)
 548                 if (p[i].sz > 0) {
 549                         assert(p[i].value);
 550                         /* LINTED */
 551                         for (j = 0; j < (int)p[i].sz; j++)
 552                                 free(p[i].value[j]);
 553                         free(p[i].value);
 554                 }
 555         free(p);
 556 }
 557
 558
 559 static void
 560 mdoc_elem_free(struct mdoc_elem *p)
 561 {
 562
 563         argfree(p->argc, p->argv);
 564 }
 565
 566
 567 static void
 568 mdoc_block_free(struct mdoc_block *p)
 569 {
 570
 571         argfree(p->argc, p->argv);
 572 }
 573
 574
 575 static void
 576 mdoc_text_free(struct mdoc_text *p)
 577 {
 578
 579         if (p->string)
 580                 free(p->string);
 581 }
 582
 583
 584 void
 585 mdoc_node_free(struct mdoc_node *p)
 586 {
 587
 588         switch (p->type) {
 589         case (MDOC_TEXT):
 590                 mdoc_text_free(&p->data.text);
 591                 break;
 592         case (MDOC_ELEM):
 593                 mdoc_elem_free(&p->data.elem);
 594                 break;
 595         case (MDOC_BLOCK):
 596                 mdoc_block_free(&p->data.block);
 597                 break;
 598         default:
 599                 break;
 600         }
 601
 602         free(p);
 603 }
 604
 605
 606 void
 607 mdoc_node_freelist(struct mdoc_node *p)
 608 {
 609
 610         if (p->child)
 611                 mdoc_node_freelist(p->child);
 612         if (p->next)
 613                 mdoc_node_freelist(p->next);
 614
 615         mdoc_node_free(p);
 616 }
 617
 618
 619 int
 620 mdoc_find(const struct mdoc *mdoc, const char *key)
 621 {
 622
 623         return(mdoc_tokhash_find(mdoc->htab, key));
 624 }
 625
 626
 627 static void
 628 argcpy(struct mdoc_arg *dst, const struct mdoc_arg *src)
 629 {
 630         int              i;
 631
 632         dst->line = src->line;
 633         dst->pos = src->pos;
 634         dst->arg = src->arg;
 635         if (0 == (dst->sz = src->sz))
 636                 return;
 637         dst->value = xcalloc(dst->sz, sizeof(char *));
 638         for (i = 0; i < (int)dst->sz; i++)
 639                 dst->value[i] = xstrdup(src->value[i]);
 640 }
 641
 642
 643 static struct mdoc_arg *
 644 argdup(size_t argsz, const struct mdoc_arg *args)
 645 {
 646         struct mdoc_arg *pp;
 647         int              i;
 648
 649         if (0 == argsz)
 650                 return(NULL);
 651
 652         pp = xcalloc((size_t)argsz, sizeof(struct mdoc_arg));
 653         for (i = 0; i < (int)argsz; i++)
 654                 argcpy(&pp[i], &args[i]);
 655
 656         return(pp);
 657 }
 658
 659
 660 /* FIXME: deprecate. */
 661 char *
 662 mdoc_node2a(struct mdoc_node *node)
 663 {
 664         static char      buf[64];
 665
 666         assert(node);
 667
 668         buf[0] = 0;
 669         (void)xstrlcat(buf, mdoc_type2a(node->type), 64);
 670         if (MDOC_ROOT == node->type)
 671                 return(buf);
 672         (void)xstrlcat(buf, " `", 64);
 673         if (MDOC_TEXT == node->type)
 674                 (void)xstrlcat(buf, node->data.text.string, 64);
 675         else
 676                 (void)xstrlcat(buf, mdoc_macronames[node->tok], 64);
 677         (void)xstrlcat(buf, "'", 64);
 678
 679         return(buf);
 680 }
 681
 682