mdoc.c

   1 /* $Id: mdoc.c,v 1.54 2009/03/08 12:40:27 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",
  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"
  86         };
  87
  88 const   char * const *mdoc_macronames = __mdoc_macronames;
  89 const   char * const *mdoc_argnames = __mdoc_argnames;
  90
  91
  92 const struct mdoc_node *
  93 mdoc_node(const struct mdoc *mdoc)
  94 {
  95
  96         return(mdoc->first);
  97 }
  98
  99
 100 const struct mdoc_meta *
 101 mdoc_meta(const struct mdoc *mdoc)
 102 {
 103
 104         return(&mdoc->meta);
 105 }
 106
 107
 108 void
 109 mdoc_free(struct mdoc *mdoc)
 110 {
 111
 112         if (mdoc->first)
 113                 mdoc_node_freelist(mdoc->first);
 114         if (mdoc->htab)
 115                 mdoc_tokhash_free(mdoc->htab);
 116         if (mdoc->meta.title)
 117                 free(mdoc->meta.title);
 118         if (mdoc->meta.os)
 119                 free(mdoc->meta.os);
 120         if (mdoc->meta.name)
 121                 free(mdoc->meta.name);
 122         if (mdoc->meta.arch)
 123                 free(mdoc->meta.arch);
 124         if (mdoc->meta.vol)
 125                 free(mdoc->meta.vol);
 126
 127         free(mdoc);
 128 }
 129
 130
 131 struct mdoc *
 132 mdoc_alloc(void *data, const struct mdoc_cb *cb)
 133 {
 134         struct mdoc     *p;
 135
 136         p = xcalloc(1, sizeof(struct mdoc));
 137
 138         p->data = data;
 139         if (cb)
 140                 (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
 141
 142         p->last = xcalloc(1, sizeof(struct mdoc_node));
 143         p->last->type = MDOC_ROOT;
 144         p->first = p->last;
 145
 146         p->next = MDOC_NEXT_CHILD;
 147         p->htab = mdoc_tokhash_alloc();
 148
 149         return(p);
 150 }
 151
 152
 153 int
 154 mdoc_endparse(struct mdoc *mdoc)
 155 {
 156
 157         if (MDOC_HALT & mdoc->flags)
 158                 return(0);
 159         if (NULL == mdoc->first)
 160                 return(1);
 161
 162         assert(mdoc->last);
 163         if ( ! macro_end(mdoc)) {
 164                 mdoc->flags |= MDOC_HALT;
 165                 return(0);
 166         }
 167         return(1);
 168 }
 169
 170
 171 /*
 172  * Main parse routine.  Parses a single line -- really just hands off to
 173  * the macro or text parser.
 174  */
 175 int
 176 mdoc_parseln(struct mdoc *m, int ln, char *buf)
 177 {
 178
 179         /* If in error-mode, then we parse no more. */
 180
 181         if (MDOC_HALT & m->flags)
 182                 return(0);
 183
 184         return('.' == *buf ? parsemacro(m, ln, buf) :
 185                         parsetext(m, ln, buf));
 186 }
 187
 188
 189 void
 190 mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
 191 {
 192         char              buf[256];
 193         va_list           ap;
 194
 195         if (NULL == mdoc->cb.mdoc_msg)
 196                 return;
 197
 198         va_start(ap, fmt);
 199         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 200         va_end(ap);
 201         (*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf);
 202 }
 203
 204
 205 int
 206 mdoc_verr(struct mdoc *mdoc, int ln, int pos,
 207                 const char *fmt, ...)
 208 {
 209         char             buf[256];
 210         va_list          ap;
 211
 212         if (NULL == mdoc->cb.mdoc_err)
 213                 return(0);
 214
 215         va_start(ap, fmt);
 216         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 217         va_end(ap);
 218         return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
 219 }
 220
 221
 222 int
 223 mdoc_vwarn(struct mdoc *mdoc, int ln, int pos,
 224                 enum mdoc_warn type, const char *fmt, ...)
 225 {
 226         char             buf[256];
 227         va_list          ap;
 228
 229         if (NULL == mdoc->cb.mdoc_warn)
 230                 return(0);
 231
 232         va_start(ap, fmt);
 233         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 234         va_end(ap);
 235         return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf));
 236 }
 237
 238
 239 int
 240 mdoc_macro(struct mdoc *m, int tok,
 241                 int ln, int pp, int *pos, char *buf)
 242 {
 243
 244         /* FIXME - these should happen during validation. */
 245
 246         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 247                         SEC_PROLOGUE != m->lastnamed)
 248                 return(mdoc_perr(m, ln, pp,
 249                                 "disallowed in document body"));
 250
 251         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 252                         SEC_PROLOGUE == m->lastnamed)
 253                 return(mdoc_perr(m, ln, pp,
 254                                 "disallowed in prologue"));
 255
 256         if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags))
 257                 return(mdoc_perr(m, ln, pp, "not callable"));
 258
 259         return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
 260 }
 261
 262
 263 static int
 264 mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p)
 265 {
 266
 267         assert(mdoc->last);
 268         assert(mdoc->first);
 269         assert(MDOC_ROOT != p->type);
 270
 271         switch (mdoc->next) {
 272         case (MDOC_NEXT_SIBLING):
 273                 mdoc->last->next = p;
 274                 p->prev = mdoc->last;
 275                 p->parent = mdoc->last->parent;
 276                 break;
 277         case (MDOC_NEXT_CHILD):
 278                 mdoc->last->child = p;
 279                 p->parent = mdoc->last;
 280                 break;
 281         default:
 282                 abort();
 283                 /* NOTREACHED */
 284         }
 285
 286         if ( ! mdoc_valid_pre(mdoc, p))
 287                 return(0);
 288
 289         switch (p->type) {
 290         case (MDOC_HEAD):
 291                 assert(MDOC_BLOCK == p->parent->type);
 292                 p->parent->head = p;
 293                 break;
 294         case (MDOC_TAIL):
 295                 assert(MDOC_BLOCK == p->parent->type);
 296                 p->parent->tail = p;
 297                 break;
 298         case (MDOC_BODY):
 299                 assert(MDOC_BLOCK == p->parent->type);
 300                 p->parent->body = p;
 301                 break;
 302         default:
 303                 break;
 304         }
 305
 306         mdoc->last = p;
 307         return(1);
 308 }
 309
 310
 311 static struct mdoc_node *
 312 mdoc_node_alloc(const struct mdoc *mdoc)
 313 {
 314         struct mdoc_node *p;
 315
 316         p = xcalloc(1, sizeof(struct mdoc_node));
 317         p->sec = mdoc->lastsec;
 318
 319         return(p);
 320 }
 321
 322
 323 int
 324 mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 325 {
 326         struct mdoc_node *p;
 327
 328         assert(mdoc->first);
 329         assert(mdoc->last);
 330
 331         p = mdoc_node_alloc(mdoc);
 332
 333         p->line = line;
 334         p->pos = pos;
 335         p->type = MDOC_TAIL;
 336         p->tok = tok;
 337
 338         return(mdoc_node_append(mdoc, p));
 339 }
 340
 341
 342 int
 343 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 344 {
 345         struct mdoc_node *p;
 346
 347         assert(mdoc->first);
 348         assert(mdoc->last);
 349
 350         p = mdoc_node_alloc(mdoc);
 351
 352         p->line = line;
 353         p->pos = pos;
 354         p->type = MDOC_HEAD;
 355         p->tok = tok;
 356
 357         return(mdoc_node_append(mdoc, p));
 358 }
 359
 360
 361 int
 362 mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 363 {
 364         struct mdoc_node *p;
 365
 366         assert(mdoc->first);
 367         assert(mdoc->last);
 368
 369         p = mdoc_node_alloc(mdoc);
 370
 371         p->line = line;
 372         p->pos = pos;
 373         p->type = MDOC_BODY;
 374         p->tok = tok;
 375
 376         return(mdoc_node_append(mdoc, p));
 377 }
 378
 379
 380 int
 381 mdoc_root_alloc(struct mdoc *mdoc)
 382 {
 383         struct mdoc_node *p;
 384
 385         p = mdoc_node_alloc(mdoc);
 386
 387         p->type = MDOC_ROOT;
 388
 389         return(mdoc_node_append(mdoc, p));
 390 }
 391
 392
 393 int
 394 mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
 395                 int tok, struct mdoc_arg *args)
 396 {
 397         struct mdoc_node *p;
 398
 399         p = mdoc_node_alloc(mdoc);
 400
 401         p->pos = pos;
 402         p->line = line;
 403         p->type = MDOC_BLOCK;
 404         p->tok = tok;
 405         p->args = args;
 406
 407         if (args)
 408                 (args->refcnt)++;
 409
 410         return(mdoc_node_append(mdoc, p));
 411 }
 412
 413
 414 int
 415 mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
 416                 int tok, struct mdoc_arg *args)
 417 {
 418         struct mdoc_node *p;
 419
 420         p = mdoc_node_alloc(mdoc);
 421
 422         p->line = line;
 423         p->pos = pos;
 424         p->type = MDOC_ELEM;
 425         p->tok = tok;
 426         p->args = args;
 427
 428         if (args)
 429                 (args->refcnt)++;
 430
 431         return(mdoc_node_append(mdoc, p));
 432 }
 433
 434
 435 int
 436 mdoc_word_alloc(struct mdoc *mdoc,
 437                 int line, int pos, const char *word)
 438 {
 439         struct mdoc_node *p;
 440
 441         p = mdoc_node_alloc(mdoc);
 442
 443         p->line = line;
 444         p->pos = pos;
 445         p->type = MDOC_TEXT;
 446         p->string = xstrdup(word);
 447
 448         return(mdoc_node_append(mdoc, p));
 449 }
 450
 451
 452 void
 453 mdoc_node_free(struct mdoc_node *p)
 454 {
 455
 456         if (p->string)
 457                 free(p->string);
 458         if (p->args)
 459                 mdoc_argv_free(p->args);
 460         free(p);
 461 }
 462
 463
 464 void
 465 mdoc_node_freelist(struct mdoc_node *p)
 466 {
 467
 468         if (p->child)
 469                 mdoc_node_freelist(p->child);
 470         if (p->next)
 471                 mdoc_node_freelist(p->next);
 472
 473         mdoc_node_free(p);
 474 }
 475
 476
 477 /*
 478  * Parse free-form text, that is, a line that does not begin with the
 479  * control character.
 480  */
 481 static int
 482 parsetext(struct mdoc *mdoc, int line, char *buf)
 483 {
 484
 485         if (SEC_PROLOGUE == mdoc->lastnamed)
 486                 return(mdoc_perr(mdoc, line, 0,
 487                         "text disallowed in prologue"));
 488
 489         if ( ! mdoc_word_alloc(mdoc, line, 0, buf))
 490                 return(0);
 491
 492         mdoc->next = MDOC_NEXT_SIBLING;
 493         return(1);
 494 }
 495
 496
 497 /*
 498  * Parse a macro line, that is, a line beginning with the control
 499  * character.
 500  */
 501 int
 502 parsemacro(struct mdoc *m, int ln, char *buf)
 503 {
 504         int               i, c;
 505         char              mac[5];
 506
 507         /* Comments are quickly ignored. */
 508
 509         if (buf[1] && '\\' == buf[1])
 510                 if (buf[2] && '\"' == buf[2])
 511                         return(1);
 512
 513         /* Copy the first word into a nil-terminated buffer. */
 514
 515         for (i = 1; i < 5; i++) {
 516                 if (0 == (mac[i - 1] = buf[i]))
 517                         break;
 518                 else if (isspace((unsigned char)buf[i]))
 519                         break;
 520         }
 521
 522         /* FIXME: be able to skip unknown macro lines! */
 523
 524         mac[i - 1] = 0;
 525
 526         if (i == 5 || i <= 2) {
 527                 (void)mdoc_perr(m, ln, 1, "unknown macro: %s%s",
 528                                 mac, i == 5 ? "..." : "");
 529                 goto err;
 530         }
 531
 532         if (MDOC_MAX == (c = mdoc_tokhash_find(m->htab, mac))) {
 533                 (void)mdoc_perr(m, ln, 1, "unknown macro: %s", mac);
 534                 goto err;
 535         }
 536
 537         /* The macro is sane.  Jump to the next word. */
 538
 539         while (buf[i] && isspace((unsigned char)buf[i]))
 540                 i++;
 541
 542         /* Begin recursive parse sequence. */
 543
 544         if ( ! mdoc_macro(m, c, ln, 1, &i, buf))
 545                 goto err;
 546
 547         return(1);
 548
 549 err:    /* Error out. */
 550
 551         m->flags |= MDOC_HALT;
 552         return(0);
 553 }