mdoc.c

   1 /* $Id: mdoc.c,v 1.57 2009/03/08 20:57:35 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_node *mdoc_node_alloc(const struct mdoc *);
  36 static  int               mdoc_node_append(struct mdoc *,
  37                                 struct mdoc_node *);
  38
  39 static  int               parsetext(struct mdoc *, int, char *);
  40 static  int               parsemacro(struct mdoc *, int, char *);
  41
  42
  43 const   char *const __mdoc_macronames[MDOC_MAX] = {
  44         "\\\"",         "Dd",           "Dt",           "Os",
  45         "Sh",           "Ss",           "Pp",           "D1",
  46         "Dl",           "Bd",           "Ed",           "Bl",
  47         "El",           "It",           "Ad",           "An",
  48         "Ar",           "Cd",           "Cm",           "Dv",
  49         "Er",           "Ev",           "Ex",           "Fa",
  50         "Fd",           "Fl",           "Fn",           "Ft",
  51         "Ic",           "In",           "Li",           "Nd",
  52         "Nm",           "Op",           "Ot",           "Pa",
  53         "Rv",           "St",           "Va",           "Vt",
  54         /* LINTED */
  55         "Xr",           "\%A",          "\%B",          "\%D",
  56         /* LINTED */
  57         "\%I",          "\%J",          "\%N",          "\%O",
  58         /* LINTED */
  59         "\%P",          "\%R",          "\%T",          "\%V",
  60         "Ac",           "Ao",           "Aq",           "At",
  61         "Bc",           "Bf",           "Bo",           "Bq",
  62         "Bsx",          "Bx",           "Db",           "Dc",
  63         "Do",           "Dq",           "Ec",           "Ef",
  64         "Em",           "Eo",           "Fx",           "Ms",
  65         "No",           "Ns",           "Nx",           "Ox",
  66         "Pc",           "Pf",           "Po",           "Pq",
  67         "Qc",           "Ql",           "Qo",           "Qq",
  68         "Re",           "Rs",           "Sc",           "So",
  69         "Sq",           "Sm",           "Sx",           "Sy",
  70         "Tn",           "Ux",           "Xc",           "Xo",
  71         "Fo",           "Fc",           "Oo",           "Oc",
  72         "Bk",           "Ek",           "Bt",           "Hf",
  73         "Fr",           "Ud",           "Lb",           "Ap",
  74         "Lp"
  75         };
  76
  77 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  78         "split",                "nosplit",              "ragged",
  79         "unfilled",             "literal",              "file",
  80         "offset",               "bullet",               "dash",
  81         "hyphen",               "item",                 "enum",
  82         "tag",                  "diag",                 "hang",
  83         "ohang",                "inset",                "column",
  84         "width",                "compact",              "std",
  85         "filled",               "words",                "emphasis",
  86         "symbolic"
  87         };
  88
  89 const   char * const *mdoc_macronames = __mdoc_macronames;
  90 const   char * const *mdoc_argnames = __mdoc_argnames;
  91
  92
  93 const struct mdoc_node *
  94 mdoc_node(const struct mdoc *mdoc)
  95 {
  96
  97         return(mdoc->first);
  98 }
  99
 100
 101 const struct mdoc_meta *
 102 mdoc_meta(const struct mdoc *mdoc)
 103 {
 104
 105         return(&mdoc->meta);
 106 }
 107
 108
 109 void
 110 mdoc_free(struct mdoc *mdoc)
 111 {
 112
 113         if (mdoc->first)
 114                 mdoc_node_freelist(mdoc->first);
 115         if (mdoc->htab)
 116                 mdoc_tokhash_free(mdoc->htab);
 117         if (mdoc->meta.title)
 118                 free(mdoc->meta.title);
 119         if (mdoc->meta.os)
 120                 free(mdoc->meta.os);
 121         if (mdoc->meta.name)
 122                 free(mdoc->meta.name);
 123         if (mdoc->meta.arch)
 124                 free(mdoc->meta.arch);
 125         if (mdoc->meta.vol)
 126                 free(mdoc->meta.vol);
 127
 128         free(mdoc);
 129 }
 130
 131
 132 struct mdoc *
 133 mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb)
 134 {
 135         struct mdoc     *p;
 136
 137         p = xcalloc(1, sizeof(struct mdoc));
 138
 139         p->data = data;
 140         if (cb)
 141                 (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
 142
 143         p->last = xcalloc(1, sizeof(struct mdoc_node));
 144         p->last->type = MDOC_ROOT;
 145         p->first = p->last;
 146         p->pflags = pflags;
 147         p->next = MDOC_NEXT_CHILD;
 148         p->htab = mdoc_tokhash_alloc();
 149
 150         return(p);
 151 }
 152
 153
 154 int
 155 mdoc_endparse(struct mdoc *mdoc)
 156 {
 157
 158         if (MDOC_HALT & mdoc->flags)
 159                 return(0);
 160         if (NULL == mdoc->first)
 161                 return(1);
 162
 163         assert(mdoc->last);
 164         if ( ! macro_end(mdoc)) {
 165                 mdoc->flags |= MDOC_HALT;
 166                 return(0);
 167         }
 168         return(1);
 169 }
 170
 171
 172 /*
 173  * Main parse routine.  Parses a single line -- really just hands off to
 174  * the macro or text parser.
 175  */
 176 int
 177 mdoc_parseln(struct mdoc *m, int ln, char *buf)
 178 {
 179
 180         /* If in error-mode, then we parse no more. */
 181
 182         if (MDOC_HALT & m->flags)
 183                 return(0);
 184
 185         return('.' == *buf ? parsemacro(m, ln, buf) :
 186                         parsetext(m, ln, buf));
 187 }
 188
 189
 190 void
 191 mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
 192 {
 193         char              buf[256];
 194         va_list           ap;
 195
 196         if (NULL == mdoc->cb.mdoc_msg)
 197                 return;
 198
 199         va_start(ap, fmt);
 200         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 201         va_end(ap);
 202         (*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf);
 203 }
 204
 205
 206 int
 207 mdoc_verr(struct mdoc *mdoc, int ln, int pos,
 208                 const char *fmt, ...)
 209 {
 210         char             buf[256];
 211         va_list          ap;
 212
 213         if (NULL == mdoc->cb.mdoc_err)
 214                 return(0);
 215
 216         va_start(ap, fmt);
 217         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 218         va_end(ap);
 219         return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
 220 }
 221
 222
 223 int
 224 mdoc_vwarn(struct mdoc *mdoc, int ln, int pos,
 225                 enum mdoc_warn type, const char *fmt, ...)
 226 {
 227         char             buf[256];
 228         va_list          ap;
 229
 230         if (NULL == mdoc->cb.mdoc_warn)
 231                 return(0);
 232
 233         va_start(ap, fmt);
 234         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 235         va_end(ap);
 236         return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf));
 237 }
 238
 239
 240 int
 241 mdoc_macro(struct mdoc *m, int tok,
 242                 int ln, int pp, int *pos, char *buf)
 243 {
 244
 245         /* FIXME - these should happen during validation. */
 246
 247         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 248                         SEC_PROLOGUE != m->lastnamed)
 249                 return(mdoc_perr(m, ln, pp,
 250                                 "disallowed in document body"));
 251
 252         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 253                         SEC_PROLOGUE == m->lastnamed)
 254                 return(mdoc_perr(m, ln, pp,
 255                                 "disallowed in prologue"));
 256
 257         if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags))
 258                 return(mdoc_perr(m, ln, pp, "not callable"));
 259
 260         return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
 261 }
 262
 263
 264 static int
 265 mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p)
 266 {
 267
 268         assert(mdoc->last);
 269         assert(mdoc->first);
 270         assert(MDOC_ROOT != p->type);
 271
 272         switch (mdoc->next) {
 273         case (MDOC_NEXT_SIBLING):
 274                 mdoc->last->next = p;
 275                 p->prev = mdoc->last;
 276                 p->parent = mdoc->last->parent;
 277                 break;
 278         case (MDOC_NEXT_CHILD):
 279                 mdoc->last->child = p;
 280                 p->parent = mdoc->last;
 281                 break;
 282         default:
 283                 abort();
 284                 /* NOTREACHED */
 285         }
 286
 287         if ( ! mdoc_valid_pre(mdoc, p))
 288                 return(0);
 289
 290         switch (p->type) {
 291         case (MDOC_HEAD):
 292                 assert(MDOC_BLOCK == p->parent->type);
 293                 p->parent->head = p;
 294                 break;
 295         case (MDOC_TAIL):
 296                 assert(MDOC_BLOCK == p->parent->type);
 297                 p->parent->tail = p;
 298                 break;
 299         case (MDOC_BODY):
 300                 assert(MDOC_BLOCK == p->parent->type);
 301                 p->parent->body = p;
 302                 break;
 303         default:
 304                 break;
 305         }
 306
 307         mdoc->last = p;
 308         return(1);
 309 }
 310
 311
 312 static struct mdoc_node *
 313 mdoc_node_alloc(const struct mdoc *mdoc)
 314 {
 315         struct mdoc_node *p;
 316
 317         p = xcalloc(1, sizeof(struct mdoc_node));
 318         p->sec = mdoc->lastsec;
 319
 320         return(p);
 321 }
 322
 323
 324 int
 325 mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 326 {
 327         struct mdoc_node *p;
 328
 329         assert(mdoc->first);
 330         assert(mdoc->last);
 331
 332         p = mdoc_node_alloc(mdoc);
 333
 334         p->line = line;
 335         p->pos = pos;
 336         p->type = MDOC_TAIL;
 337         p->tok = tok;
 338
 339         return(mdoc_node_append(mdoc, p));
 340 }
 341
 342
 343 int
 344 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 345 {
 346         struct mdoc_node *p;
 347
 348         assert(mdoc->first);
 349         assert(mdoc->last);
 350
 351         p = mdoc_node_alloc(mdoc);
 352
 353         p->line = line;
 354         p->pos = pos;
 355         p->type = MDOC_HEAD;
 356         p->tok = tok;
 357
 358         return(mdoc_node_append(mdoc, p));
 359 }
 360
 361
 362 int
 363 mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 364 {
 365         struct mdoc_node *p;
 366
 367         assert(mdoc->first);
 368         assert(mdoc->last);
 369
 370         p = mdoc_node_alloc(mdoc);
 371
 372         p->line = line;
 373         p->pos = pos;
 374         p->type = MDOC_BODY;
 375         p->tok = tok;
 376
 377         return(mdoc_node_append(mdoc, p));
 378 }
 379
 380
 381 int
 382 mdoc_root_alloc(struct mdoc *mdoc)
 383 {
 384         struct mdoc_node *p;
 385
 386         p = mdoc_node_alloc(mdoc);
 387
 388         p->type = MDOC_ROOT;
 389
 390         return(mdoc_node_append(mdoc, p));
 391 }
 392
 393
 394 int
 395 mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
 396                 int tok, struct mdoc_arg *args)
 397 {
 398         struct mdoc_node *p;
 399
 400         p = mdoc_node_alloc(mdoc);
 401
 402         p->pos = pos;
 403         p->line = line;
 404         p->type = MDOC_BLOCK;
 405         p->tok = tok;
 406         p->args = args;
 407
 408         if (args)
 409                 (args->refcnt)++;
 410
 411         return(mdoc_node_append(mdoc, p));
 412 }
 413
 414
 415 int
 416 mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
 417                 int tok, struct mdoc_arg *args)
 418 {
 419         struct mdoc_node *p;
 420
 421         p = mdoc_node_alloc(mdoc);
 422
 423         p->line = line;
 424         p->pos = pos;
 425         p->type = MDOC_ELEM;
 426         p->tok = tok;
 427         p->args = args;
 428
 429         if (args)
 430                 (args->refcnt)++;
 431
 432         return(mdoc_node_append(mdoc, p));
 433 }
 434
 435
 436 int
 437 mdoc_word_alloc(struct mdoc *mdoc,
 438                 int line, int pos, const char *word)
 439 {
 440         struct mdoc_node *p;
 441
 442         p = mdoc_node_alloc(mdoc);
 443
 444         p->line = line;
 445         p->pos = pos;
 446         p->type = MDOC_TEXT;
 447         p->string = xstrdup(word);
 448
 449         return(mdoc_node_append(mdoc, p));
 450 }
 451
 452
 453 void
 454 mdoc_node_free(struct mdoc_node *p)
 455 {
 456
 457         if (p->string)
 458                 free(p->string);
 459         if (p->args)
 460                 mdoc_argv_free(p->args);
 461         free(p);
 462 }
 463
 464
 465 void
 466 mdoc_node_freelist(struct mdoc_node *p)
 467 {
 468
 469         if (p->child)
 470                 mdoc_node_freelist(p->child);
 471         if (p->next)
 472                 mdoc_node_freelist(p->next);
 473
 474         mdoc_node_free(p);
 475 }
 476
 477
 478 /*
 479  * Parse free-form text, that is, a line that does not begin with the
 480  * control character.
 481  */
 482 static int
 483 parsetext(struct mdoc *mdoc, int line, char *buf)
 484 {
 485
 486         if (SEC_PROLOGUE == mdoc->lastnamed)
 487                 return(mdoc_perr(mdoc, line, 0,
 488                         "text disallowed in prologue"));
 489
 490         if ( ! mdoc_word_alloc(mdoc, line, 0, buf))
 491                 return(0);
 492
 493         mdoc->next = MDOC_NEXT_SIBLING;
 494         return(1);
 495 }
 496
 497
 498 /*
 499  * Parse a macro line, that is, a line beginning with the control
 500  * character.
 501  */
 502 int
 503 parsemacro(struct mdoc *m, int ln, char *buf)
 504 {
 505         int               i, c;
 506         char              mac[5];
 507
 508         /* Comments are quickly ignored. */
 509
 510         if (buf[1] && '\\' == buf[1])
 511                 if (buf[2] && '\"' == buf[2])
 512                         return(1);
 513
 514         /* Copy the first word into a nil-terminated buffer. */
 515
 516         for (i = 1; i < 5; i++) {
 517                 if (0 == (mac[i - 1] = buf[i]))
 518                         break;
 519                 else if (isspace((unsigned char)buf[i]))
 520                         break;
 521         }
 522
 523         /* FIXME: be able to skip unknown macro lines! */
 524
 525         mac[i - 1] = 0;
 526
 527         if (i == 5 || i <= 2) {
 528                 (void)mdoc_perr(m, ln, 1, "unknown macro: %s%s",
 529                                 mac, i == 5 ? "..." : "");
 530                 goto err;
 531         }
 532
 533         if (MDOC_MAX == (c = mdoc_tokhash_find(m->htab, mac))) {
 534                 (void)mdoc_perr(m, ln, 1, "unknown macro: %s", mac);
 535                 goto err;
 536         }
 537
 538         /* The macro is sane.  Jump to the next word. */
 539
 540         while (buf[i] && isspace((unsigned char)buf[i]))
 541                 i++;
 542
 543         /* Begin recursive parse sequence. */
 544
 545         if ( ! mdoc_macro(m, c, ln, 1, &i, buf))
 546                 goto err;
 547
 548         return(1);
 549
 550 err:    /* Error out. */
 551
 552         m->flags |= MDOC_HALT;
 553         return(0);
 554 }