mdoc.c

   1 /* $Id: mdoc.c,v 1.49 2009/03/01 23:14:15 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         /*
 195          * FIXME: should puke on whitespace in non-literal displays.
 196          */
 197
 198         if ('.' != *buf) {
 199                 if (SEC_PROLOGUE == mdoc->lastnamed)
 200                         return(mdoc_perr(mdoc, line, 0,
 201                                 "no text in document prologue"));
 202                 if ( ! mdoc_word_alloc(mdoc, line, 0, buf))
 203                         return(0);
 204                 mdoc->next = MDOC_NEXT_SIBLING;
 205                 return(1);
 206         }
 207
 208         if (buf[1] && '\\' == buf[1])
 209                 if (buf[2] && '\"' == buf[2])
 210                         return(1);
 211
 212         i = 1;
 213         while (buf[i] && ! isspace((int)buf[i]) && i < (int)sizeof(tmp))
 214                 i++;
 215
 216         if (i == (int)sizeof(tmp)) {
 217                 mdoc->flags |= MDOC_HALT;
 218                 return(mdoc_perr(mdoc, line, 1, "unknown macro"));
 219         } else if (i <= 2) {
 220                 mdoc->flags |= MDOC_HALT;
 221                 return(mdoc_perr(mdoc, line, 1, "unknown macro"));
 222         }
 223
 224         i--;
 225
 226         (void)memcpy(tmp, buf + 1, (size_t)i);
 227         tmp[i++] = 0;
 228
 229         if (MDOC_MAX == (c = mdoc_find(mdoc, tmp))) {
 230                 mdoc->flags |= MDOC_HALT;
 231                 return(mdoc_perr(mdoc, line, 1, "unknown macro"));
 232         }
 233
 234         while (buf[i] && isspace((int)buf[i]))
 235                 i++;
 236
 237         if ( ! mdoc_macro(mdoc, c, line, 1, &i, buf)) {
 238                 mdoc->flags |= MDOC_HALT;
 239                 return(0);
 240         }
 241         return(1);
 242 }
 243
 244
 245 void
 246 mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
 247 {
 248         char              buf[256];
 249         va_list           ap;
 250
 251         if (NULL == mdoc->cb.mdoc_msg)
 252                 return;
 253
 254         va_start(ap, fmt);
 255         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 256         va_end(ap);
 257         (*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf);
 258 }
 259
 260
 261 int
 262 mdoc_verr(struct mdoc *mdoc, int ln, int pos,
 263                 const char *fmt, ...)
 264 {
 265         char             buf[256];
 266         va_list          ap;
 267
 268         if (NULL == mdoc->cb.mdoc_err)
 269                 return(0);
 270
 271         va_start(ap, fmt);
 272         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 273         va_end(ap);
 274         return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
 275 }
 276
 277
 278 int
 279 mdoc_vwarn(struct mdoc *mdoc, int ln, int pos,
 280                 enum mdoc_warn type, const char *fmt, ...)
 281 {
 282         char             buf[256];
 283         va_list          ap;
 284
 285         if (NULL == mdoc->cb.mdoc_warn)
 286                 return(0);
 287
 288         va_start(ap, fmt);
 289         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 290         va_end(ap);
 291         return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf));
 292 }
 293
 294
 295 int
 296 mdoc_macro(struct mdoc *mdoc, int tok,
 297                 int ln, int ppos, int *pos, char *buf)
 298 {
 299
 300         assert(mdoc_macros[tok].fp);
 301
 302         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 303                         SEC_PROLOGUE != mdoc->lastnamed)
 304                 return(mdoc_perr(mdoc, ln, ppos, "macro disallowed in document body"));
 305         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 306                         SEC_PROLOGUE == mdoc->lastnamed)
 307                 return(mdoc_perr(mdoc, ln, ppos, "macro disallowed in document prologue"));
 308         if (1 != ppos && ! (MDOC_CALLABLE & mdoc_macros[tok].flags))
 309                 return(mdoc_perr(mdoc, ln, ppos, "macro not callable"));
 310         return((*mdoc_macros[tok].fp)(mdoc, tok, ln, ppos, pos, buf));
 311 }
 312
 313
 314 static int
 315 mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p)
 316 {
 317         const char       *nn, *nt, *on, *ot, *act;
 318
 319         assert(mdoc->last);
 320         assert(mdoc->first);
 321         assert(MDOC_ROOT != p->type);
 322
 323         /* See if we exceed the suggest line-max. */
 324
 325         switch (p->type) {
 326         case (MDOC_TEXT):
 327                 /* FALLTHROUGH */
 328         case (MDOC_ELEM):
 329                 /* FALLTHROUGH */
 330         case (MDOC_BLOCK):
 331                 mdoc->linetok++;
 332                 break;
 333         default:
 334                 break;
 335         }
 336
 337         /* This sort-of works (re-opening of text macros...). */
 338         if (mdoc->linetok > MDOC_LINEARG_SOFTMAX)
 339                 if ( ! mdoc_nwarn(mdoc, p, WARN_COMPAT,
 340                                         "suggested %d tokens per line exceeded (has %d)",
 341                                         MDOC_LINEARG_SOFTMAX, mdoc->linetok))
 342                         return(0);
 343
 344         if (MDOC_TEXT == mdoc->last->type)
 345                 on = "<text>";
 346         else if (MDOC_ROOT == mdoc->last->type)
 347                 on = "<root>";
 348         else
 349                 on = mdoc_macronames[mdoc->last->tok];
 350
 351         if (MDOC_TEXT == p->type)
 352                 nn = "<text>";
 353         else if (MDOC_ROOT == p->type)
 354                 nn = "<root>";
 355         else
 356                 nn = mdoc_macronames[p->tok];
 357
 358         ot = mdoc_type2a(mdoc->last->type);
 359         nt = mdoc_type2a(p->type);
 360
 361         switch (mdoc->next) {
 362         case (MDOC_NEXT_SIBLING):
 363                 mdoc->last->next = p;
 364                 p->prev = mdoc->last;
 365                 p->parent = mdoc->last->parent;
 366                 act = "sibling";
 367                 break;
 368         case (MDOC_NEXT_CHILD):
 369                 mdoc->last->child = p;
 370                 p->parent = mdoc->last;
 371                 act = "child";
 372                 break;
 373         default:
 374                 abort();
 375                 /* NOTREACHED */
 376         }
 377
 378         if ( ! mdoc_valid_pre(mdoc, p))
 379                 return(0);
 380
 381         switch (p->type) {
 382         case (MDOC_HEAD):
 383                 assert(MDOC_BLOCK == p->parent->type);
 384                 p->parent->data.block.head = p;
 385                 break;
 386         case (MDOC_TAIL):
 387                 assert(MDOC_BLOCK == p->parent->type);
 388                 p->parent->data.block.tail = p;
 389                 break;
 390         case (MDOC_BODY):
 391                 assert(MDOC_BLOCK == p->parent->type);
 392                 p->parent->data.block.body = p;
 393                 break;
 394         default:
 395                 break;
 396         }
 397
 398         mdoc->last = p;
 399         mdoc_msg(mdoc, "parse: %s `%s' %s of %s `%s'",
 400                         nt, nn, act, ot, on);
 401         return(1);
 402 }
 403
 404
 405 static struct mdoc_node *
 406 mdoc_node_alloc(const struct mdoc *mdoc)
 407 {
 408         struct mdoc_node *p;
 409
 410         p = xcalloc(1, sizeof(struct mdoc_node));
 411         p->sec = mdoc->lastsec;
 412
 413         return(p);
 414 }
 415
 416
 417 int
 418 mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 419 {
 420         struct mdoc_node *p;
 421
 422         assert(mdoc->first);
 423         assert(mdoc->last);
 424
 425         p = mdoc_node_alloc(mdoc);
 426
 427         p->line = line;
 428         p->pos = pos;
 429         p->type = MDOC_TAIL;
 430         p->tok = tok;
 431
 432         return(mdoc_node_append(mdoc, p));
 433 }
 434
 435
 436 int
 437 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 438 {
 439         struct mdoc_node *p;
 440
 441         assert(mdoc->first);
 442         assert(mdoc->last);
 443
 444         p = mdoc_node_alloc(mdoc);
 445
 446         p->line = line;
 447         p->pos = pos;
 448         p->type = MDOC_HEAD;
 449         p->tok = tok;
 450
 451         return(mdoc_node_append(mdoc, p));
 452 }
 453
 454
 455 int
 456 mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 457 {
 458         struct mdoc_node *p;
 459
 460         assert(mdoc->first);
 461         assert(mdoc->last);
 462
 463         p = mdoc_node_alloc(mdoc);
 464
 465         p->line = line;
 466         p->pos = pos;
 467         p->type = MDOC_BODY;
 468         p->tok = tok;
 469
 470         return(mdoc_node_append(mdoc, p));
 471 }
 472
 473
 474 int
 475 mdoc_root_alloc(struct mdoc *mdoc)
 476 {
 477         struct mdoc_node *p;
 478
 479         p = mdoc_node_alloc(mdoc);
 480
 481         p->type = MDOC_ROOT;
 482
 483         return(mdoc_node_append(mdoc, p));
 484 }
 485
 486
 487 int
 488 mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
 489                 int tok, size_t argsz, const struct mdoc_arg *args)
 490 {
 491         struct mdoc_node *p;
 492
 493         p = mdoc_node_alloc(mdoc);
 494
 495         p->pos = pos;
 496         p->line = line;
 497         p->type = MDOC_BLOCK;
 498         p->tok = tok;
 499         p->data.block.argc = argsz;
 500         p->data.block.argv = argdup(argsz, args);
 501
 502         return(mdoc_node_append(mdoc, p));
 503 }
 504
 505
 506 int
 507 mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
 508                 int tok, size_t argsz, const struct mdoc_arg *args)
 509 {
 510         struct mdoc_node *p;
 511
 512         p = mdoc_node_alloc(mdoc);
 513
 514         p->line = line;
 515         p->pos = pos;
 516         p->type = MDOC_ELEM;
 517         p->tok = tok;
 518         p->data.elem.argc = argsz;
 519         p->data.elem.argv = argdup(argsz, args);
 520
 521         return(mdoc_node_append(mdoc, p));
 522 }
 523
 524
 525 int
 526 mdoc_word_alloc(struct mdoc *mdoc,
 527                 int line, int pos, const char *word)
 528 {
 529         struct mdoc_node *p;
 530
 531         p = mdoc_node_alloc(mdoc);
 532
 533         p->line = line;
 534         p->pos = pos;
 535         p->type = MDOC_TEXT;
 536         p->data.text.string = xstrdup(word);
 537
 538         return(mdoc_node_append(mdoc, p));
 539 }
 540
 541
 542 static void
 543 argfree(size_t sz, struct mdoc_arg *p)
 544 {
 545         int              i, j;
 546
 547         if (0 == sz)
 548                 return;
 549
 550         assert(p);
 551         /* LINTED */
 552         for (i = 0; i < (int)sz; i++)
 553                 if (p[i].sz > 0) {
 554                         assert(p[i].value);
 555                         /* LINTED */
 556                         for (j = 0; j < (int)p[i].sz; j++)
 557                                 free(p[i].value[j]);
 558                         free(p[i].value);
 559                 }
 560         free(p);
 561 }
 562
 563
 564 static void
 565 mdoc_elem_free(struct mdoc_elem *p)
 566 {
 567
 568         argfree(p->argc, p->argv);
 569 }
 570
 571
 572 static void
 573 mdoc_block_free(struct mdoc_block *p)
 574 {
 575
 576         argfree(p->argc, p->argv);
 577 }
 578
 579
 580 static void
 581 mdoc_text_free(struct mdoc_text *p)
 582 {
 583
 584         if (p->string)
 585                 free(p->string);
 586 }
 587
 588
 589 void
 590 mdoc_node_free(struct mdoc_node *p)
 591 {
 592
 593         switch (p->type) {
 594         case (MDOC_TEXT):
 595                 mdoc_text_free(&p->data.text);
 596                 break;
 597         case (MDOC_ELEM):
 598                 mdoc_elem_free(&p->data.elem);
 599                 break;
 600         case (MDOC_BLOCK):
 601                 mdoc_block_free(&p->data.block);
 602                 break;
 603         default:
 604                 break;
 605         }
 606
 607         free(p);
 608 }
 609
 610
 611 void
 612 mdoc_node_freelist(struct mdoc_node *p)
 613 {
 614
 615         if (p->child)
 616                 mdoc_node_freelist(p->child);
 617         if (p->next)
 618                 mdoc_node_freelist(p->next);
 619
 620         mdoc_node_free(p);
 621 }
 622
 623
 624 int
 625 mdoc_find(const struct mdoc *mdoc, const char *key)
 626 {
 627
 628         return(mdoc_tokhash_find(mdoc->htab, key));
 629 }
 630
 631
 632 static void
 633 argcpy(struct mdoc_arg *dst, const struct mdoc_arg *src)
 634 {
 635         int              i;
 636
 637         dst->line = src->line;
 638         dst->pos = src->pos;
 639         dst->arg = src->arg;
 640         if (0 == (dst->sz = src->sz))
 641                 return;
 642         dst->value = xcalloc(dst->sz, sizeof(char *));
 643         for (i = 0; i < (int)dst->sz; i++)
 644                 dst->value[i] = xstrdup(src->value[i]);
 645 }
 646
 647
 648 static struct mdoc_arg *
 649 argdup(size_t argsz, const struct mdoc_arg *args)
 650 {
 651         struct mdoc_arg *pp;
 652         int              i;
 653
 654         if (0 == argsz)
 655                 return(NULL);
 656
 657         pp = xcalloc((size_t)argsz, sizeof(struct mdoc_arg));
 658         for (i = 0; i < (int)argsz; i++)
 659                 argcpy(&pp[i], &args[i]);
 660
 661         return(pp);
 662 }
 663
 664
 665 /* FIXME: deprecate. */
 666 char *
 667 mdoc_node2a(struct mdoc_node *node)
 668 {
 669         static char      buf[64];
 670
 671         assert(node);
 672
 673         buf[0] = 0;
 674         (void)xstrlcat(buf, mdoc_type2a(node->type), 64);
 675         if (MDOC_ROOT == node->type)
 676                 return(buf);
 677         (void)xstrlcat(buf, " `", 64);
 678         if (MDOC_TEXT == node->type)
 679                 (void)xstrlcat(buf, node->data.text.string, 64);
 680         else
 681                 (void)xstrlcat(buf, mdoc_macronames[node->tok], 64);
 682         (void)xstrlcat(buf, "'", 64);
 683
 684         return(buf);
 685 }
 686
 687