mdoc.c

   1 /* $Id: mdoc.c,v 1.61 2009/03/11 00:39:58 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 static  int               macrowarn(struct mdoc *, int, const char *);
  42
  43
  44 const   char *const __mdoc_macronames[MDOC_MAX] = {
  45         "\\\"",         "Dd",           "Dt",           "Os",
  46         "Sh",           "Ss",           "Pp",           "D1",
  47         "Dl",           "Bd",           "Ed",           "Bl",
  48         "El",           "It",           "Ad",           "An",
  49         "Ar",           "Cd",           "Cm",           "Dv",
  50         "Er",           "Ev",           "Ex",           "Fa",
  51         "Fd",           "Fl",           "Fn",           "Ft",
  52         "Ic",           "In",           "Li",           "Nd",
  53         "Nm",           "Op",           "Ot",           "Pa",
  54         "Rv",           "St",           "Va",           "Vt",
  55         /* LINTED */
  56         "Xr",           "\%A",          "\%B",          "\%D",
  57         /* LINTED */
  58         "\%I",          "\%J",          "\%N",          "\%O",
  59         /* LINTED */
  60         "\%P",          "\%R",          "\%T",          "\%V",
  61         "Ac",           "Ao",           "Aq",           "At",
  62         "Bc",           "Bf",           "Bo",           "Bq",
  63         "Bsx",          "Bx",           "Db",           "Dc",
  64         "Do",           "Dq",           "Ec",           "Ef",
  65         "Em",           "Eo",           "Fx",           "Ms",
  66         "No",           "Ns",           "Nx",           "Ox",
  67         "Pc",           "Pf",           "Po",           "Pq",
  68         "Qc",           "Ql",           "Qo",           "Qq",
  69         "Re",           "Rs",           "Sc",           "So",
  70         "Sq",           "Sm",           "Sx",           "Sy",
  71         "Tn",           "Ux",           "Xc",           "Xo",
  72         "Fo",           "Fc",           "Oo",           "Oc",
  73         "Bk",           "Ek",           "Bt",           "Hf",
  74         "Fr",           "Ud",           "Lb",           "Ap",
  75         "Lp",           "Lk",           "Mt",           "Brq",
  76         "Bro",          "Brc"
  77         };
  78
  79 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  80         "split",                "nosplit",              "ragged",
  81         "unfilled",             "literal",              "file",
  82         "offset",               "bullet",               "dash",
  83         "hyphen",               "item",                 "enum",
  84         "tag",                  "diag",                 "hang",
  85         "ohang",                "inset",                "column",
  86         "width",                "compact",              "std",
  87         "filled",               "words",                "emphasis",
  88         "symbolic"
  89         };
  90
  91 const   char * const *mdoc_macronames = __mdoc_macronames;
  92 const   char * const *mdoc_argnames = __mdoc_argnames;
  93
  94
  95 const struct mdoc_node *
  96 mdoc_node(const struct mdoc *mdoc)
  97 {
  98
  99         return(mdoc->first);
 100 }
 101
 102
 103 const struct mdoc_meta *
 104 mdoc_meta(const struct mdoc *mdoc)
 105 {
 106
 107         return(&mdoc->meta);
 108 }
 109
 110
 111 void
 112 mdoc_free(struct mdoc *mdoc)
 113 {
 114
 115         if (mdoc->first)
 116                 mdoc_node_freelist(mdoc->first);
 117         if (mdoc->htab)
 118                 mdoc_tokhash_free(mdoc->htab);
 119         if (mdoc->meta.title)
 120                 free(mdoc->meta.title);
 121         if (mdoc->meta.os)
 122                 free(mdoc->meta.os);
 123         if (mdoc->meta.name)
 124                 free(mdoc->meta.name);
 125         if (mdoc->meta.arch)
 126                 free(mdoc->meta.arch);
 127         if (mdoc->meta.vol)
 128                 free(mdoc->meta.vol);
 129
 130         free(mdoc);
 131 }
 132
 133
 134 struct mdoc *
 135 mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb)
 136 {
 137         struct mdoc     *p;
 138
 139         p = xcalloc(1, sizeof(struct mdoc));
 140
 141         p->data = data;
 142         if (cb)
 143                 (void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
 144
 145         p->last = xcalloc(1, sizeof(struct mdoc_node));
 146         p->last->type = MDOC_ROOT;
 147         p->first = p->last;
 148         p->pflags = pflags;
 149         p->next = MDOC_NEXT_CHILD;
 150         p->htab = mdoc_tokhash_alloc();
 151
 152         return(p);
 153 }
 154
 155
 156 int
 157 mdoc_endparse(struct mdoc *mdoc)
 158 {
 159
 160         if (MDOC_HALT & mdoc->flags)
 161                 return(0);
 162         if (NULL == mdoc->first)
 163                 return(1);
 164
 165         assert(mdoc->last);
 166         if ( ! macro_end(mdoc)) {
 167                 mdoc->flags |= MDOC_HALT;
 168                 return(0);
 169         }
 170         return(1);
 171 }
 172
 173
 174 /*
 175  * Main parse routine.  Parses a single line -- really just hands off to
 176  * the macro or text parser.
 177  */
 178 int
 179 mdoc_parseln(struct mdoc *m, int ln, char *buf)
 180 {
 181
 182         /* If in error-mode, then we parse no more. */
 183
 184         if (MDOC_HALT & m->flags)
 185                 return(0);
 186
 187         return('.' == *buf ? parsemacro(m, ln, buf) :
 188                         parsetext(m, ln, buf));
 189 }
 190
 191
 192 void
 193 mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...)
 194 {
 195         char              buf[256];
 196         va_list           ap;
 197
 198         if (NULL == mdoc->cb.mdoc_msg)
 199                 return;
 200
 201         va_start(ap, fmt);
 202         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 203         va_end(ap);
 204         (*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf);
 205 }
 206
 207
 208 int
 209 mdoc_verr(struct mdoc *mdoc, int ln, int pos,
 210                 const char *fmt, ...)
 211 {
 212         char             buf[256];
 213         va_list          ap;
 214
 215         if (NULL == mdoc->cb.mdoc_err)
 216                 return(0);
 217
 218         va_start(ap, fmt);
 219         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 220         va_end(ap);
 221         return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf));
 222 }
 223
 224
 225 int
 226 mdoc_vwarn(struct mdoc *mdoc, int ln, int pos,
 227                 enum mdoc_warn type, const char *fmt, ...)
 228 {
 229         char             buf[256];
 230         va_list          ap;
 231
 232         if (NULL == mdoc->cb.mdoc_warn)
 233                 return(0);
 234
 235         va_start(ap, fmt);
 236         (void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 237         va_end(ap);
 238         return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf));
 239 }
 240
 241
 242 int
 243 mdoc_macro(struct mdoc *m, int tok,
 244                 int ln, int pp, int *pos, char *buf)
 245 {
 246
 247         /* FIXME - these should happen during validation. */
 248
 249         if (MDOC_PROLOGUE & mdoc_macros[tok].flags &&
 250                         SEC_PROLOGUE != m->lastnamed)
 251                 return(mdoc_perr(m, ln, pp,
 252                                 "disallowed in document body"));
 253
 254         if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) &&
 255                         SEC_PROLOGUE == m->lastnamed)
 256                 return(mdoc_perr(m, ln, pp,
 257                                 "disallowed in prologue"));
 258
 259         if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags))
 260                 return(mdoc_perr(m, ln, pp, "not callable"));
 261
 262         return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
 263 }
 264
 265
 266 static int
 267 mdoc_node_append(struct mdoc *mdoc, struct mdoc_node *p)
 268 {
 269
 270         assert(mdoc->last);
 271         assert(mdoc->first);
 272         assert(MDOC_ROOT != p->type);
 273
 274         switch (mdoc->next) {
 275         case (MDOC_NEXT_SIBLING):
 276                 mdoc->last->next = p;
 277                 p->prev = mdoc->last;
 278                 p->parent = mdoc->last->parent;
 279                 break;
 280         case (MDOC_NEXT_CHILD):
 281                 mdoc->last->child = p;
 282                 p->parent = mdoc->last;
 283                 break;
 284         default:
 285                 abort();
 286                 /* NOTREACHED */
 287         }
 288
 289         if ( ! mdoc_valid_pre(mdoc, p))
 290                 return(0);
 291
 292         switch (p->type) {
 293         case (MDOC_HEAD):
 294                 assert(MDOC_BLOCK == p->parent->type);
 295                 p->parent->head = p;
 296                 break;
 297         case (MDOC_TAIL):
 298                 assert(MDOC_BLOCK == p->parent->type);
 299                 p->parent->tail = p;
 300                 break;
 301         case (MDOC_BODY):
 302                 assert(MDOC_BLOCK == p->parent->type);
 303                 p->parent->body = p;
 304                 break;
 305         default:
 306                 break;
 307         }
 308
 309         mdoc->last = p;
 310         return(1);
 311 }
 312
 313
 314 static struct mdoc_node *
 315 mdoc_node_alloc(const struct mdoc *mdoc)
 316 {
 317         struct mdoc_node *p;
 318
 319         p = xcalloc(1, sizeof(struct mdoc_node));
 320         p->sec = mdoc->lastsec;
 321
 322         return(p);
 323 }
 324
 325
 326 int
 327 mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 328 {
 329         struct mdoc_node *p;
 330
 331         assert(mdoc->first);
 332         assert(mdoc->last);
 333
 334         p = mdoc_node_alloc(mdoc);
 335
 336         p->line = line;
 337         p->pos = pos;
 338         p->type = MDOC_TAIL;
 339         p->tok = tok;
 340
 341         return(mdoc_node_append(mdoc, p));
 342 }
 343
 344
 345 int
 346 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 347 {
 348         struct mdoc_node *p;
 349
 350         assert(mdoc->first);
 351         assert(mdoc->last);
 352
 353         p = mdoc_node_alloc(mdoc);
 354
 355         p->line = line;
 356         p->pos = pos;
 357         p->type = MDOC_HEAD;
 358         p->tok = tok;
 359
 360         return(mdoc_node_append(mdoc, p));
 361 }
 362
 363
 364 int
 365 mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok)
 366 {
 367         struct mdoc_node *p;
 368
 369         assert(mdoc->first);
 370         assert(mdoc->last);
 371
 372         p = mdoc_node_alloc(mdoc);
 373
 374         p->line = line;
 375         p->pos = pos;
 376         p->type = MDOC_BODY;
 377         p->tok = tok;
 378
 379         return(mdoc_node_append(mdoc, p));
 380 }
 381
 382
 383 int
 384 mdoc_root_alloc(struct mdoc *mdoc)
 385 {
 386         struct mdoc_node *p;
 387
 388         p = mdoc_node_alloc(mdoc);
 389
 390         p->type = MDOC_ROOT;
 391
 392         return(mdoc_node_append(mdoc, p));
 393 }
 394
 395
 396 int
 397 mdoc_block_alloc(struct mdoc *mdoc, int line, int pos,
 398                 int tok, struct mdoc_arg *args)
 399 {
 400         struct mdoc_node *p;
 401
 402         p = mdoc_node_alloc(mdoc);
 403
 404         p->pos = pos;
 405         p->line = line;
 406         p->type = MDOC_BLOCK;
 407         p->tok = tok;
 408         p->args = args;
 409
 410         if (args)
 411                 (args->refcnt)++;
 412
 413         return(mdoc_node_append(mdoc, p));
 414 }
 415
 416
 417 int
 418 mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos,
 419                 int tok, struct mdoc_arg *args)
 420 {
 421         struct mdoc_node *p;
 422
 423         p = mdoc_node_alloc(mdoc);
 424
 425         p->line = line;
 426         p->pos = pos;
 427         p->type = MDOC_ELEM;
 428         p->tok = tok;
 429         p->args = args;
 430
 431         if (args)
 432                 (args->refcnt)++;
 433
 434         return(mdoc_node_append(mdoc, p));
 435 }
 436
 437
 438 int
 439 mdoc_word_alloc(struct mdoc *mdoc,
 440                 int line, int pos, const char *word)
 441 {
 442         struct mdoc_node *p;
 443
 444         p = mdoc_node_alloc(mdoc);
 445
 446         p->line = line;
 447         p->pos = pos;
 448         p->type = MDOC_TEXT;
 449         p->string = xstrdup(word);
 450
 451         return(mdoc_node_append(mdoc, p));
 452 }
 453
 454
 455 void
 456 mdoc_node_free(struct mdoc_node *p)
 457 {
 458
 459         if (p->string)
 460                 free(p->string);
 461         if (p->args)
 462                 mdoc_argv_free(p->args);
 463         free(p);
 464 }
 465
 466
 467 void
 468 mdoc_node_freelist(struct mdoc_node *p)
 469 {
 470
 471         if (p->child)
 472                 mdoc_node_freelist(p->child);
 473         if (p->next)
 474                 mdoc_node_freelist(p->next);
 475
 476         mdoc_node_free(p);
 477 }
 478
 479
 480 /*
 481  * Parse free-form text, that is, a line that does not begin with the
 482  * control character.
 483  */
 484 static int
 485 parsetext(struct mdoc *mdoc, int line, char *buf)
 486 {
 487
 488         if (SEC_PROLOGUE == mdoc->lastnamed)
 489                 return(mdoc_perr(mdoc, line, 0,
 490                         "text disallowed in prologue"));
 491
 492         if ( ! mdoc_word_alloc(mdoc, line, 0, buf))
 493                 return(0);
 494
 495         mdoc->next = MDOC_NEXT_SIBLING;
 496         return(1);
 497 }
 498
 499
 500 static int
 501 macrowarn(struct mdoc *m, int ln, const char *buf)
 502 {
 503         if ( ! (MDOC_IGN_MACRO & m->pflags))
 504                 return(mdoc_perr(m, ln, 1, "unknown macro: %s%s",
 505                                 buf, strlen(buf) > 3 ? "..." : ""));
 506         return(mdoc_pwarn(m, ln, 1, WARN_SYNTAX,
 507                                 "unknown macro: %s%s",
 508                                 buf, strlen(buf) > 3 ? "..." : ""));
 509 }
 510
 511
 512
 513 /*
 514  * Parse a macro line, that is, a line beginning with the control
 515  * character.
 516  */
 517 int
 518 parsemacro(struct mdoc *m, int ln, char *buf)
 519 {
 520         int               i, c;
 521         char              mac[5];
 522
 523         /* Comments are quickly ignored. */
 524
 525         if (buf[1] && '\\' == buf[1])
 526                 if (buf[2] && '\"' == buf[2])
 527                         return(1);
 528
 529         /* Copy the first word into a nil-terminated buffer. */
 530
 531         for (i = 1; i < 5; i++) {
 532                 if (0 == (mac[i - 1] = buf[i]))
 533                         break;
 534                 else if (isspace((unsigned char)buf[i]))
 535                         break;
 536         }
 537
 538         /* FIXME: be able to skip unknown macro lines! */
 539
 540         mac[i - 1] = 0;
 541
 542         if (i == 5 || i <= 2) {
 543                 if ( ! macrowarn(m, ln, mac))
 544                         goto err;
 545                 return(1);
 546         }
 547
 548         if (MDOC_MAX == (c = mdoc_tokhash_find(m->htab, mac))) {
 549                 if ( ! macrowarn(m, ln, mac))
 550                         goto err;
 551                 return(1);
 552         }
 553
 554         /* The macro is sane.  Jump to the next word. */
 555
 556         while (buf[i] && isspace((unsigned char)buf[i]))
 557                 i++;
 558
 559         /* Begin recursive parse sequence. */
 560
 561         if ( ! mdoc_macro(m, c, ln, 1, &i, buf))
 562                 goto err;
 563
 564         return(1);
 565
 566 err:    /* Error out. */
 567
 568         m->flags |= MDOC_HALT;
 569         return(0);
 570 }