eqn.c

   1 /*      $Id: eqn.c,v 1.57 2015/01/28 21:11:53 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #include "config.h"
  19
  20 #include <sys/types.h>
  21
  22 #include <assert.h>
  23 #include <limits.h>
  24 #include <stdio.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27 #include <time.h>
  28
  29 #include "mandoc.h"
  30 #include "mandoc_aux.h"
  31 #include "libmandoc.h"
  32 #include "libroff.h"
  33
  34 #define EQN_NEST_MAX     128 /* maximum nesting of defines */
  35 #define STRNEQ(p1, sz1, p2, sz2) \
  36         ((sz1) == (sz2) && 0 == strncmp((p1), (p2), (sz1)))
  37
  38 enum    eqn_tok {
  39         EQN_TOK_DYAD = 0,
  40         EQN_TOK_VEC,
  41         EQN_TOK_UNDER,
  42         EQN_TOK_BAR,
  43         EQN_TOK_TILDE,
  44         EQN_TOK_HAT,
  45         EQN_TOK_DOT,
  46         EQN_TOK_DOTDOT,
  47         EQN_TOK_FWD,
  48         EQN_TOK_BACK,
  49         EQN_TOK_DOWN,
  50         EQN_TOK_UP,
  51         EQN_TOK_FAT,
  52         EQN_TOK_ROMAN,
  53         EQN_TOK_ITALIC,
  54         EQN_TOK_BOLD,
  55         EQN_TOK_SIZE,
  56         EQN_TOK_SUB,
  57         EQN_TOK_SUP,
  58         EQN_TOK_SQRT,
  59         EQN_TOK_OVER,
  60         EQN_TOK_FROM,
  61         EQN_TOK_TO,
  62         EQN_TOK_BRACE_OPEN,
  63         EQN_TOK_BRACE_CLOSE,
  64         EQN_TOK_GSIZE,
  65         EQN_TOK_GFONT,
  66         EQN_TOK_MARK,
  67         EQN_TOK_LINEUP,
  68         EQN_TOK_LEFT,
  69         EQN_TOK_RIGHT,
  70         EQN_TOK_PILE,
  71         EQN_TOK_LPILE,
  72         EQN_TOK_RPILE,
  73         EQN_TOK_CPILE,
  74         EQN_TOK_MATRIX,
  75         EQN_TOK_CCOL,
  76         EQN_TOK_LCOL,
  77         EQN_TOK_RCOL,
  78         EQN_TOK_DELIM,
  79         EQN_TOK_DEFINE,
  80         EQN_TOK_TDEFINE,
  81         EQN_TOK_NDEFINE,
  82         EQN_TOK_UNDEF,
  83         EQN_TOK_EOF,
  84         EQN_TOK_ABOVE,
  85         EQN_TOK__MAX
  86 };
  87
  88 static  const char *eqn_toks[EQN_TOK__MAX] = {
  89         "dyad", /* EQN_TOK_DYAD */
  90         "vec", /* EQN_TOK_VEC */
  91         "under", /* EQN_TOK_UNDER */
  92         "bar", /* EQN_TOK_BAR */
  93         "tilde", /* EQN_TOK_TILDE */
  94         "hat", /* EQN_TOK_HAT */
  95         "dot", /* EQN_TOK_DOT */
  96         "dotdot", /* EQN_TOK_DOTDOT */
  97         "fwd", /* EQN_TOK_FWD * */
  98         "back", /* EQN_TOK_BACK */
  99         "down", /* EQN_TOK_DOWN */
 100         "up", /* EQN_TOK_UP */
 101         "fat", /* EQN_TOK_FAT */
 102         "roman", /* EQN_TOK_ROMAN */
 103         "italic", /* EQN_TOK_ITALIC */
 104         "bold", /* EQN_TOK_BOLD */
 105         "size", /* EQN_TOK_SIZE */
 106         "sub", /* EQN_TOK_SUB */
 107         "sup", /* EQN_TOK_SUP */
 108         "sqrt", /* EQN_TOK_SQRT */
 109         "over", /* EQN_TOK_OVER */
 110         "from", /* EQN_TOK_FROM */
 111         "to", /* EQN_TOK_TO */
 112         "{", /* EQN_TOK_BRACE_OPEN */
 113         "}", /* EQN_TOK_BRACE_CLOSE */
 114         "gsize", /* EQN_TOK_GSIZE */
 115         "gfont", /* EQN_TOK_GFONT */
 116         "mark", /* EQN_TOK_MARK */
 117         "lineup", /* EQN_TOK_LINEUP */
 118         "left", /* EQN_TOK_LEFT */
 119         "right", /* EQN_TOK_RIGHT */
 120         "pile", /* EQN_TOK_PILE */
 121         "lpile", /* EQN_TOK_LPILE */
 122         "rpile", /* EQN_TOK_RPILE */
 123         "cpile", /* EQN_TOK_CPILE */
 124         "matrix", /* EQN_TOK_MATRIX */
 125         "ccol", /* EQN_TOK_CCOL */
 126         "lcol", /* EQN_TOK_LCOL */
 127         "rcol", /* EQN_TOK_RCOL */
 128         "delim", /* EQN_TOK_DELIM */
 129         "define", /* EQN_TOK_DEFINE */
 130         "tdefine", /* EQN_TOK_TDEFINE */
 131         "ndefine", /* EQN_TOK_NDEFINE */
 132         "undef", /* EQN_TOK_UNDEF */
 133         NULL, /* EQN_TOK_EOF */
 134         "above", /* EQN_TOK_ABOVE */
 135 };
 136
 137 enum    eqn_symt {
 138         EQNSYM_alpha,
 139         EQNSYM_beta,
 140         EQNSYM_chi,
 141         EQNSYM_delta,
 142         EQNSYM_epsilon,
 143         EQNSYM_eta,
 144         EQNSYM_gamma,
 145         EQNSYM_iota,
 146         EQNSYM_kappa,
 147         EQNSYM_lambda,
 148         EQNSYM_mu,
 149         EQNSYM_nu,
 150         EQNSYM_omega,
 151         EQNSYM_omicron,
 152         EQNSYM_phi,
 153         EQNSYM_pi,
 154         EQNSYM_ps,
 155         EQNSYM_rho,
 156         EQNSYM_sigma,
 157         EQNSYM_tau,
 158         EQNSYM_theta,
 159         EQNSYM_upsilon,
 160         EQNSYM_xi,
 161         EQNSYM_zeta,
 162         EQNSYM_DELTA,
 163         EQNSYM_GAMMA,
 164         EQNSYM_LAMBDA,
 165         EQNSYM_OMEGA,
 166         EQNSYM_PHI,
 167         EQNSYM_PI,
 168         EQNSYM_PSI,
 169         EQNSYM_SIGMA,
 170         EQNSYM_THETA,
 171         EQNSYM_UPSILON,
 172         EQNSYM_XI,
 173         EQNSYM_inter,
 174         EQNSYM_union,
 175         EQNSYM_prod,
 176         EQNSYM_int,
 177         EQNSYM_sum,
 178         EQNSYM_grad,
 179         EQNSYM_del,
 180         EQNSYM_times,
 181         EQNSYM_cdot,
 182         EQNSYM_nothing,
 183         EQNSYM_approx,
 184         EQNSYM_prime,
 185         EQNSYM_half,
 186         EQNSYM_partial,
 187         EQNSYM_inf,
 188         EQNSYM_muchgreat,
 189         EQNSYM_muchless,
 190         EQNSYM_larrow,
 191         EQNSYM_rarrow,
 192         EQNSYM_pm,
 193         EQNSYM_nequal,
 194         EQNSYM_equiv,
 195         EQNSYM_lessequal,
 196         EQNSYM_moreequal,
 197         EQNSYM__MAX
 198 };
 199
 200 struct  eqnsym {
 201         const char      *str;
 202         const char      *sym;
 203 };
 204
 205 static  const struct eqnsym eqnsyms[EQNSYM__MAX] = {
 206         { "alpha", "*a" }, /* EQNSYM_alpha */
 207         { "beta", "*b" }, /* EQNSYM_beta */
 208         { "chi", "*x" }, /* EQNSYM_chi */
 209         { "delta", "*d" }, /* EQNSYM_delta */
 210         { "epsilon", "*e" }, /* EQNSYM_epsilon */
 211         { "eta", "*y" }, /* EQNSYM_eta */
 212         { "gamma", "*g" }, /* EQNSYM_gamma */
 213         { "iota", "*i" }, /* EQNSYM_iota */
 214         { "kappa", "*k" }, /* EQNSYM_kappa */
 215         { "lambda", "*l" }, /* EQNSYM_lambda */
 216         { "mu", "*m" }, /* EQNSYM_mu */
 217         { "nu", "*n" }, /* EQNSYM_nu */
 218         { "omega", "*w" }, /* EQNSYM_omega */
 219         { "omicron", "*o" }, /* EQNSYM_omicron */
 220         { "phi", "*f" }, /* EQNSYM_phi */
 221         { "pi", "*p" }, /* EQNSYM_pi */
 222         { "psi", "*q" }, /* EQNSYM_psi */
 223         { "rho", "*r" }, /* EQNSYM_rho */
 224         { "sigma", "*s" }, /* EQNSYM_sigma */
 225         { "tau", "*t" }, /* EQNSYM_tau */
 226         { "theta", "*h" }, /* EQNSYM_theta */
 227         { "upsilon", "*u" }, /* EQNSYM_upsilon */
 228         { "xi", "*c" }, /* EQNSYM_xi */
 229         { "zeta", "*z" }, /* EQNSYM_zeta */
 230         { "DELTA", "*D" }, /* EQNSYM_DELTA */
 231         { "GAMMA", "*G" }, /* EQNSYM_GAMMA */
 232         { "LAMBDA", "*L" }, /* EQNSYM_LAMBDA */
 233         { "OMEGA", "*W" }, /* EQNSYM_OMEGA */
 234         { "PHI", "*F" }, /* EQNSYM_PHI */
 235         { "PI", "*P" }, /* EQNSYM_PI */
 236         { "PSI", "*Q" }, /* EQNSYM_PSI */
 237         { "SIGMA", "*S" }, /* EQNSYM_SIGMA */
 238         { "THETA", "*H" }, /* EQNSYM_THETA */
 239         { "UPSILON", "*U" }, /* EQNSYM_UPSILON */
 240         { "XI", "*C" }, /* EQNSYM_XI */
 241         { "inter", "ca" }, /* EQNSYM_inter */
 242         { "union", "cu" }, /* EQNSYM_union */
 243         { "prod", "product" }, /* EQNSYM_prod */
 244         { "int", "integral" }, /* EQNSYM_int */
 245         { "sum", "sum" }, /* EQNSYM_sum */
 246         { "grad", "gr" }, /* EQNSYM_grad */
 247         { "del", "gr" }, /* EQNSYM_del */
 248         { "times", "mu" }, /* EQNSYM_times */
 249         { "cdot", "pc" }, /* EQNSYM_cdot */
 250         { "nothing", "&" }, /* EQNSYM_nothing */
 251         { "approx", "~~" }, /* EQNSYM_approx */
 252         { "prime", "aq" }, /* EQNSYM_prime */
 253         { "half", "12" }, /* EQNSYM_half */
 254         { "partial", "pd" }, /* EQNSYM_partial */
 255         { "inf", "if" }, /* EQNSYM_inf */
 256         { ">>", ">>" }, /* EQNSYM_muchgreat */
 257         { "<<", "<<" }, /* EQNSYM_muchless */
 258         { "<-", "<-" }, /* EQNSYM_larrow */
 259         { "->", "->" }, /* EQNSYM_rarrow */
 260         { "+-", "+-" }, /* EQNSYM_pm */
 261         { "!=", "!=" }, /* EQNSYM_nequal */
 262         { "==", "==" }, /* EQNSYM_equiv */
 263         { "<=", "<=" }, /* EQNSYM_lessequal */
 264         { ">=", ">=" }, /* EQNSYM_moreequal */
 265 };
 266
 267 static  struct eqn_box  *eqn_box_alloc(struct eqn_node *, struct eqn_box *);
 268 static  void             eqn_box_free(struct eqn_box *);
 269 static  struct eqn_box  *eqn_box_makebinary(struct eqn_node *,
 270                                 enum eqn_post, struct eqn_box *);
 271 static  void             eqn_def(struct eqn_node *);
 272 static  struct eqn_def  *eqn_def_find(struct eqn_node *, const char *, size_t);
 273 static  void             eqn_delim(struct eqn_node *);
 274 static  const char      *eqn_next(struct eqn_node *, char, size_t *, int);
 275 static  const char      *eqn_nextrawtok(struct eqn_node *, size_t *);
 276 static  const char      *eqn_nexttok(struct eqn_node *, size_t *);
 277 static  enum rofferr     eqn_parse(struct eqn_node *, struct eqn_box *);
 278 static  enum eqn_tok     eqn_tok_parse(struct eqn_node *, char **);
 279 static  void             eqn_undef(struct eqn_node *);
 280
 281
 282 enum rofferr
 283 eqn_read(struct eqn_node **epp, int ln,
 284                 const char *p, int pos, int *offs)
 285 {
 286         size_t           sz;
 287         struct eqn_node *ep;
 288         enum rofferr     er;
 289
 290         ep = *epp;
 291
 292         /*
 293          * If we're the terminating mark, unset our equation status and
 294          * validate the full equation.
 295          */
 296
 297         if (0 == strncmp(p, ".EN", 3)) {
 298                 er = eqn_end(epp);
 299                 p += 3;
 300                 while (' ' == *p || '\t' == *p)
 301                         p++;
 302                 if ('\0' == *p)
 303                         return(er);
 304                 mandoc_vmsg(MANDOCERR_ARG_SKIP, ep->parse,
 305                     ln, pos, "EN %s", p);
 306                 return(er);
 307         }
 308
 309         /*
 310          * Build up the full string, replacing all newlines with regular
 311          * whitespace.
 312          */
 313
 314         sz = strlen(p + pos) + 1;
 315         ep->data = mandoc_realloc(ep->data, ep->sz + sz + 1);
 316
 317         /* First invocation: nil terminate the string. */
 318
 319         if (0 == ep->sz)
 320                 *ep->data = '\0';
 321
 322         ep->sz += sz;
 323         strlcat(ep->data, p + pos, ep->sz + 1);
 324         strlcat(ep->data, " ", ep->sz + 1);
 325         return(ROFF_IGN);
 326 }
 327
 328 struct eqn_node *
 329 eqn_alloc(int pos, int line, struct mparse *parse)
 330 {
 331         struct eqn_node *p;
 332
 333         p = mandoc_calloc(1, sizeof(struct eqn_node));
 334
 335         p->parse = parse;
 336         p->eqn.ln = line;
 337         p->eqn.pos = pos;
 338         p->gsize = EQN_DEFSIZE;
 339
 340         return(p);
 341 }
 342
 343 /*
 344  * Find the key "key" of the give size within our eqn-defined values.
 345  */
 346 static struct eqn_def *
 347 eqn_def_find(struct eqn_node *ep, const char *key, size_t sz)
 348 {
 349         int              i;
 350
 351         for (i = 0; i < (int)ep->defsz; i++)
 352                 if (ep->defs[i].keysz && STRNEQ(ep->defs[i].key,
 353                     ep->defs[i].keysz, key, sz))
 354                         return(&ep->defs[i]);
 355
 356         return(NULL);
 357 }
 358
 359 /*
 360  * Get the next token from the input stream using the given quote
 361  * character.
 362  * Optionally make any replacements.
 363  */
 364 static const char *
 365 eqn_next(struct eqn_node *ep, char quote, size_t *sz, int repl)
 366 {
 367         char            *start, *next;
 368         int              q, diff, lim;
 369         size_t           ssz, dummy;
 370         struct eqn_def  *def;
 371
 372         if (NULL == sz)
 373                 sz = &dummy;
 374
 375         lim = 0;
 376         ep->rew = ep->cur;
 377 again:
 378         /* Prevent self-definitions. */
 379
 380         if (lim >= EQN_NEST_MAX) {
 381                 mandoc_msg(MANDOCERR_ROFFLOOP, ep->parse,
 382                     ep->eqn.ln, ep->eqn.pos, NULL);
 383                 return(NULL);
 384         }
 385
 386         ep->cur = ep->rew;
 387         start = &ep->data[(int)ep->cur];
 388         q = 0;
 389
 390         if ('\0' == *start)
 391                 return(NULL);
 392
 393         if (quote == *start) {
 394                 ep->cur++;
 395                 q = 1;
 396         }
 397
 398         start = &ep->data[(int)ep->cur];
 399
 400         if ( ! q) {
 401                 if ('{' == *start || '}' == *start)
 402                         ssz = 1;
 403                 else
 404                         ssz = strcspn(start + 1, " ^~\"{}\t") + 1;
 405                 next = start + (int)ssz;
 406                 if ('\0' == *next)
 407                         next = NULL;
 408         } else
 409                 next = strchr(start, quote);
 410
 411         if (NULL != next) {
 412                 *sz = (size_t)(next - start);
 413                 ep->cur += *sz;
 414                 if (q)
 415                         ep->cur++;
 416                 while (' ' == ep->data[(int)ep->cur] ||
 417                     '\t' == ep->data[(int)ep->cur] ||
 418                     '^' == ep->data[(int)ep->cur] ||
 419                     '~' == ep->data[(int)ep->cur])
 420                         ep->cur++;
 421         } else {
 422                 if (q)
 423                         mandoc_msg(MANDOCERR_ARG_QUOTE, ep->parse,
 424                             ep->eqn.ln, ep->eqn.pos, NULL);
 425                 next = strchr(start, '\0');
 426                 *sz = (size_t)(next - start);
 427                 ep->cur += *sz;
 428         }
 429
 430         /* Quotes aren't expanded for values. */
 431
 432         if (q || ! repl)
 433                 return(start);
 434
 435         if (NULL != (def = eqn_def_find(ep, start, *sz))) {
 436                 diff = def->valsz - *sz;
 437
 438                 if (def->valsz > *sz) {
 439                         ep->sz += diff;
 440                         ep->data = mandoc_realloc(ep->data, ep->sz + 1);
 441                         ep->data[ep->sz] = '\0';
 442                         start = &ep->data[(int)ep->rew];
 443                 }
 444
 445                 diff = def->valsz - *sz;
 446                 memmove(start + *sz + diff, start + *sz,
 447                     (strlen(start) - *sz) + 1);
 448                 memcpy(start, def->val, def->valsz);
 449                 goto again;
 450         }
 451
 452         return(start);
 453 }
 454
 455 /*
 456  * Get the next delimited token using the default current quote
 457  * character.
 458  */
 459 static const char *
 460 eqn_nexttok(struct eqn_node *ep, size_t *sz)
 461 {
 462
 463         return(eqn_next(ep, '"', sz, 1));
 464 }
 465
 466 /*
 467  * Get next token without replacement.
 468  */
 469 static const char *
 470 eqn_nextrawtok(struct eqn_node *ep, size_t *sz)
 471 {
 472
 473         return(eqn_next(ep, '"', sz, 0));
 474 }
 475
 476 /*
 477  * Parse a token from the stream of text.
 478  * A token consists of one of the recognised eqn(7) strings.
 479  * Strings are separated by delimiting marks.
 480  * This returns EQN_TOK_EOF when there are no more tokens.
 481  * If the token is an unrecognised string literal, then it returns
 482  * EQN_TOK__MAX and sets the "p" pointer to an allocated, nil-terminated
 483  * string.
 484  * This must be later freed with free(3).
 485  */
 486 static enum eqn_tok
 487 eqn_tok_parse(struct eqn_node *ep, char **p)
 488 {
 489         const char      *start;
 490         size_t           i, sz;
 491         int              quoted;
 492
 493         if (NULL != p)
 494                 *p = NULL;
 495
 496         quoted = ep->data[ep->cur] == '"';
 497
 498         if (NULL == (start = eqn_nexttok(ep, &sz)))
 499                 return(EQN_TOK_EOF);
 500
 501         if (quoted) {
 502                 if (p != NULL)
 503                         *p = mandoc_strndup(start, sz);
 504                 return(EQN_TOK__MAX);
 505         }
 506
 507         for (i = 0; i < EQN_TOK__MAX; i++) {
 508                 if (NULL == eqn_toks[i])
 509                         continue;
 510                 if (STRNEQ(start, sz, eqn_toks[i], strlen(eqn_toks[i])))
 511                         break;
 512         }
 513
 514         if (i == EQN_TOK__MAX && NULL != p)
 515                 *p = mandoc_strndup(start, sz);
 516
 517         return(i);
 518 }
 519
 520 static void
 521 eqn_box_free(struct eqn_box *bp)
 522 {
 523
 524         if (bp->first)
 525                 eqn_box_free(bp->first);
 526         if (bp->next)
 527                 eqn_box_free(bp->next);
 528
 529         free(bp->text);
 530         free(bp->left);
 531         free(bp->right);
 532         free(bp->top);
 533         free(bp->bottom);
 534         free(bp);
 535 }
 536
 537 /*
 538  * Allocate a box as the last child of the parent node.
 539  */
 540 static struct eqn_box *
 541 eqn_box_alloc(struct eqn_node *ep, struct eqn_box *parent)
 542 {
 543         struct eqn_box  *bp;
 544
 545         bp = mandoc_calloc(1, sizeof(struct eqn_box));
 546         bp->parent = parent;
 547         bp->parent->args++;
 548         bp->expectargs = UINT_MAX;
 549         bp->size = ep->gsize;
 550
 551         if (NULL != parent->first) {
 552                 parent->last->next = bp;
 553                 bp->prev = parent->last;
 554         } else
 555                 parent->first = bp;
 556
 557         parent->last = bp;
 558         return(bp);
 559 }
 560
 561 /*
 562  * Reparent the current last node (of the current parent) under a new
 563  * EQN_SUBEXPR as the first element.
 564  * Then return the new parent.
 565  * The new EQN_SUBEXPR will have a two-child limit.
 566  */
 567 static struct eqn_box *
 568 eqn_box_makebinary(struct eqn_node *ep,
 569         enum eqn_post pos, struct eqn_box *parent)
 570 {
 571         struct eqn_box  *b, *newb;
 572
 573         assert(NULL != parent->last);
 574         b = parent->last;
 575         if (parent->last == parent->first)
 576                 parent->first = NULL;
 577         parent->args--;
 578         parent->last = b->prev;
 579         b->prev = NULL;
 580         newb = eqn_box_alloc(ep, parent);
 581         newb->pos = pos;
 582         newb->type = EQN_SUBEXPR;
 583         newb->expectargs = 2;
 584         newb->args = 1;
 585         newb->first = newb->last = b;
 586         newb->first->next = NULL;
 587         b->parent = newb;
 588         return(newb);
 589 }
 590
 591 /*
 592  * Parse the "delim" control statement.
 593  */
 594 static void
 595 eqn_delim(struct eqn_node *ep)
 596 {
 597         const char      *start;
 598         size_t           sz;
 599
 600         if ((start = eqn_nextrawtok(ep, &sz)) == NULL)
 601                 mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 602                     ep->eqn.ln, ep->eqn.pos, "delim");
 603         else if (strncmp(start, "off", 3) == 0)
 604                 ep->delim = 0;
 605         else if (strncmp(start, "on", 2) == 0) {
 606                 if (ep->odelim && ep->cdelim)
 607                         ep->delim = 1;
 608         } else if (start[1] != '\0') {
 609                 ep->odelim = start[0];
 610                 ep->cdelim = start[1];
 611                 ep->delim = 1;
 612         }
 613 }
 614
 615 /*
 616  * Undefine a previously-defined string.
 617  */
 618 static void
 619 eqn_undef(struct eqn_node *ep)
 620 {
 621         const char      *start;
 622         struct eqn_def  *def;
 623         size_t           sz;
 624
 625         if ((start = eqn_nextrawtok(ep, &sz)) == NULL) {
 626                 mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 627                     ep->eqn.ln, ep->eqn.pos, "undef");
 628                 return;
 629         }
 630         if ((def = eqn_def_find(ep, start, sz)) == NULL)
 631                 return;
 632         free(def->key);
 633         free(def->val);
 634         def->key = def->val = NULL;
 635         def->keysz = def->valsz = 0;
 636 }
 637
 638 static void
 639 eqn_def(struct eqn_node *ep)
 640 {
 641         const char      *start;
 642         size_t           sz;
 643         struct eqn_def  *def;
 644         int              i;
 645
 646         if ((start = eqn_nextrawtok(ep, &sz)) == NULL) {
 647                 mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 648                     ep->eqn.ln, ep->eqn.pos, "define");
 649                 return;
 650         }
 651
 652         /*
 653          * Search for a key that already exists.
 654          * Create a new key if none is found.
 655          */
 656         if (NULL == (def = eqn_def_find(ep, start, sz))) {
 657                 /* Find holes in string array. */
 658                 for (i = 0; i < (int)ep->defsz; i++)
 659                         if (0 == ep->defs[i].keysz)
 660                                 break;
 661
 662                 if (i == (int)ep->defsz) {
 663                         ep->defsz++;
 664                         ep->defs = mandoc_reallocarray(ep->defs,
 665                             ep->defsz, sizeof(struct eqn_def));
 666                         ep->defs[i].key = ep->defs[i].val = NULL;
 667                 }
 668
 669                 def = ep->defs + i;
 670                 free(def->key);
 671                 def->key = mandoc_strndup(start, sz);
 672                 def->keysz = sz;
 673         }
 674
 675         start = eqn_next(ep, ep->data[(int)ep->cur], &sz, 0);
 676         if (start == NULL) {
 677                 mandoc_vmsg(MANDOCERR_REQ_EMPTY, ep->parse,
 678                     ep->eqn.ln, ep->eqn.pos, "define %s", def->key);
 679                 free(def->key);
 680                 free(def->val);
 681                 def->key = def->val = NULL;
 682                 def->keysz = def->valsz = 0;
 683                 return;
 684         }
 685         free(def->val);
 686         def->val = mandoc_strndup(start, sz);
 687         def->valsz = sz;
 688 }
 689
 690 /*
 691  * Recursively parse an eqn(7) expression.
 692  */
 693 static enum rofferr
 694 eqn_parse(struct eqn_node *ep, struct eqn_box *parent)
 695 {
 696         char             sym[64];
 697         struct eqn_box  *cur;
 698         const char      *start;
 699         char            *p;
 700         size_t           i, sz;
 701         enum eqn_tok     tok, subtok;
 702         enum eqn_post    pos;
 703         int              size;
 704
 705         assert(parent != NULL);
 706
 707         /*
 708          * Empty equation.
 709          * Do not add it to the high-level syntax tree.
 710          */
 711
 712         if (ep->data == NULL)
 713                 return(ROFF_IGN);
 714
 715 next_tok:
 716         tok = eqn_tok_parse(ep, &p);
 717
 718 this_tok:
 719         switch (tok) {
 720         case (EQN_TOK_UNDEF):
 721                 eqn_undef(ep);
 722                 break;
 723         case (EQN_TOK_NDEFINE):
 724         case (EQN_TOK_DEFINE):
 725                 eqn_def(ep);
 726                 break;
 727         case (EQN_TOK_TDEFINE):
 728                 if (eqn_nextrawtok(ep, NULL) == NULL ||
 729                     eqn_next(ep, ep->data[(int)ep->cur], NULL, 0) == NULL)
 730                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 731                             ep->eqn.ln, ep->eqn.pos, "tdefine");
 732                 break;
 733         case (EQN_TOK_DELIM):
 734                 eqn_delim(ep);
 735                 break;
 736         case (EQN_TOK_GFONT):
 737                 if (eqn_nextrawtok(ep, NULL) == NULL)
 738                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 739                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 740                 break;
 741         case (EQN_TOK_MARK):
 742         case (EQN_TOK_LINEUP):
 743                 /* Ignore these. */
 744                 break;
 745         case (EQN_TOK_DYAD):
 746         case (EQN_TOK_VEC):
 747         case (EQN_TOK_UNDER):
 748         case (EQN_TOK_BAR):
 749         case (EQN_TOK_TILDE):
 750         case (EQN_TOK_HAT):
 751         case (EQN_TOK_DOT):
 752         case (EQN_TOK_DOTDOT):
 753                 if (parent->last == NULL) {
 754                         mandoc_msg(MANDOCERR_EQN_NOBOX, ep->parse,
 755                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 756                         cur = eqn_box_alloc(ep, parent);
 757                         cur->type = EQN_TEXT;
 758                         cur->text = mandoc_strdup("");
 759                 }
 760                 parent = eqn_box_makebinary(ep, EQNPOS_NONE, parent);
 761                 parent->type = EQN_LISTONE;
 762                 parent->expectargs = 1;
 763                 switch (tok) {
 764                 case (EQN_TOK_DOTDOT):
 765                         strlcpy(sym, "\\[ad]", sizeof(sym));
 766                         break;
 767                 case (EQN_TOK_VEC):
 768                         strlcpy(sym, "\\[->]", sizeof(sym));
 769                         break;
 770                 case (EQN_TOK_DYAD):
 771                         strlcpy(sym, "\\[<>]", sizeof(sym));
 772                         break;
 773                 case (EQN_TOK_TILDE):
 774                         strlcpy(sym, "\\[a~]", sizeof(sym));
 775                         break;
 776                 case (EQN_TOK_UNDER):
 777                         strlcpy(sym, "\\[ul]", sizeof(sym));
 778                         break;
 779                 case (EQN_TOK_BAR):
 780                         strlcpy(sym, "\\[rl]", sizeof(sym));
 781                         break;
 782                 case (EQN_TOK_DOT):
 783                         strlcpy(sym, "\\[a.]", sizeof(sym));
 784                         break;
 785                 case (EQN_TOK_HAT):
 786                         strlcpy(sym, "\\[ha]", sizeof(sym));
 787                         break;
 788                 default:
 789                         abort();
 790                 }
 791
 792                 switch (tok) {
 793                 case (EQN_TOK_DOTDOT):
 794                 case (EQN_TOK_VEC):
 795                 case (EQN_TOK_DYAD):
 796                 case (EQN_TOK_TILDE):
 797                 case (EQN_TOK_BAR):
 798                 case (EQN_TOK_DOT):
 799                 case (EQN_TOK_HAT):
 800                         parent->top = mandoc_strdup(sym);
 801                         break;
 802                 case (EQN_TOK_UNDER):
 803                         parent->bottom = mandoc_strdup(sym);
 804                         break;
 805                 default:
 806                         abort();
 807                 }
 808                 parent = parent->parent;
 809                 break;
 810         case (EQN_TOK_FWD):
 811         case (EQN_TOK_BACK):
 812         case (EQN_TOK_DOWN):
 813         case (EQN_TOK_UP):
 814                 subtok = eqn_tok_parse(ep, NULL);
 815                 if (subtok != EQN_TOK__MAX) {
 816                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 817                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 818                         tok = subtok;
 819                         goto this_tok;
 820                 }
 821                 break;
 822         case (EQN_TOK_FAT):
 823         case (EQN_TOK_ROMAN):
 824         case (EQN_TOK_ITALIC):
 825         case (EQN_TOK_BOLD):
 826                 while (parent->args == parent->expectargs)
 827                         parent = parent->parent;
 828                 /*
 829                  * These values apply to the next word or sequence of
 830                  * words; thus, we mark that we'll have a child with
 831                  * exactly one of those.
 832                  */
 833                 parent = eqn_box_alloc(ep, parent);
 834                 parent->type = EQN_LISTONE;
 835                 parent->expectargs = 1;
 836                 switch (tok) {
 837                 case (EQN_TOK_FAT):
 838                         parent->font = EQNFONT_FAT;
 839                         break;
 840                 case (EQN_TOK_ROMAN):
 841                         parent->font = EQNFONT_ROMAN;
 842                         break;
 843                 case (EQN_TOK_ITALIC):
 844                         parent->font = EQNFONT_ITALIC;
 845                         break;
 846                 case (EQN_TOK_BOLD):
 847                         parent->font = EQNFONT_BOLD;
 848                         break;
 849                 default:
 850                         abort();
 851                 }
 852                 break;
 853         case (EQN_TOK_SIZE):
 854         case (EQN_TOK_GSIZE):
 855                 /* Accept two values: integral size and a single. */
 856                 if (NULL == (start = eqn_nexttok(ep, &sz))) {
 857                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 858                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 859                         break;
 860                 }
 861                 size = mandoc_strntoi(start, sz, 10);
 862                 if (-1 == size) {
 863                         mandoc_msg(MANDOCERR_IT_NONUM, ep->parse,
 864                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 865                         break;
 866                 }
 867                 if (EQN_TOK_GSIZE == tok) {
 868                         ep->gsize = size;
 869                         break;
 870                 }
 871                 parent = eqn_box_alloc(ep, parent);
 872                 parent->type = EQN_LISTONE;
 873                 parent->expectargs = 1;
 874                 parent->size = size;
 875                 break;
 876         case (EQN_TOK_FROM):
 877         case (EQN_TOK_TO):
 878         case (EQN_TOK_SUB):
 879         case (EQN_TOK_SUP):
 880                 /*
 881                  * We have a left-right-associative expression.
 882                  * Repivot under a positional node, open a child scope
 883                  * and keep on reading.
 884                  */
 885                 if (parent->last == NULL) {
 886                         mandoc_msg(MANDOCERR_EQN_NOBOX, ep->parse,
 887                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 888                         cur = eqn_box_alloc(ep, parent);
 889                         cur->type = EQN_TEXT;
 890                         cur->text = mandoc_strdup("");
 891                 }
 892                 /* Handle the "subsup" and "fromto" positions. */
 893                 if (EQN_TOK_SUP == tok && parent->pos == EQNPOS_SUB) {
 894                         parent->expectargs = 3;
 895                         parent->pos = EQNPOS_SUBSUP;
 896                         break;
 897                 }
 898                 if (EQN_TOK_TO == tok && parent->pos == EQNPOS_FROM) {
 899                         parent->expectargs = 3;
 900                         parent->pos = EQNPOS_FROMTO;
 901                         break;
 902                 }
 903                 switch (tok) {
 904                 case (EQN_TOK_FROM):
 905                         pos = EQNPOS_FROM;
 906                         break;
 907                 case (EQN_TOK_TO):
 908                         pos = EQNPOS_TO;
 909                         break;
 910                 case (EQN_TOK_SUP):
 911                         pos = EQNPOS_SUP;
 912                         break;
 913                 case (EQN_TOK_SUB):
 914                         pos = EQNPOS_SUB;
 915                         break;
 916                 default:
 917                         abort();
 918                 }
 919                 parent = eqn_box_makebinary(ep, pos, parent);
 920                 break;
 921         case (EQN_TOK_SQRT):
 922                 while (parent->args == parent->expectargs)
 923                         parent = parent->parent;
 924                 /*
 925                  * Accept a left-right-associative set of arguments just
 926                  * like sub and sup and friends but without rebalancing
 927                  * under a pivot.
 928                  */
 929                 parent = eqn_box_alloc(ep, parent);
 930                 parent->type = EQN_SUBEXPR;
 931                 parent->pos = EQNPOS_SQRT;
 932                 parent->expectargs = 1;
 933                 break;
 934         case (EQN_TOK_OVER):
 935                 /*
 936                  * We have a right-left-associative fraction.
 937                  * Close out anything that's currently open, then
 938                  * rebalance and continue reading.
 939                  */
 940                 if (parent->last == NULL) {
 941                         mandoc_msg(MANDOCERR_EQN_NOBOX, ep->parse,
 942                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 943                         cur = eqn_box_alloc(ep, parent);
 944                         cur->type = EQN_TEXT;
 945                         cur->text = mandoc_strdup("");
 946                 }
 947                 while (EQN_SUBEXPR == parent->type)
 948                         parent = parent->parent;
 949                 parent = eqn_box_makebinary(ep, EQNPOS_OVER, parent);
 950                 break;
 951         case (EQN_TOK_RIGHT):
 952         case (EQN_TOK_BRACE_CLOSE):
 953                 /*
 954                  * Close out the existing brace.
 955                  * FIXME: this is a shitty sentinel: we should really
 956                  * have a native EQN_BRACE type or whatnot.
 957                  */
 958                 for (cur = parent; cur != NULL; cur = cur->parent)
 959                         if (cur->type == EQN_LIST &&
 960                             (tok == EQN_TOK_BRACE_CLOSE ||
 961                              cur->left != NULL))
 962                                 break;
 963                 if (cur == NULL) {
 964                         mandoc_msg(MANDOCERR_BLK_NOTOPEN, ep->parse,
 965                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 966                         break;
 967                 }
 968                 parent = cur;
 969                 if (EQN_TOK_RIGHT == tok) {
 970                         if (NULL == (start = eqn_nexttok(ep, &sz))) {
 971                                 mandoc_msg(MANDOCERR_REQ_EMPTY,
 972                                     ep->parse, ep->eqn.ln,
 973                                     ep->eqn.pos, eqn_toks[tok]);
 974                                 break;
 975                         }
 976                         /* Handling depends on right/left. */
 977                         if (STRNEQ(start, sz, "ceiling", 7)) {
 978                                 strlcpy(sym, "\\[rc]", sizeof(sym));
 979                                 parent->right = mandoc_strdup(sym);
 980                         } else if (STRNEQ(start, sz, "floor", 5)) {
 981                                 strlcpy(sym, "\\[rf]", sizeof(sym));
 982                                 parent->right = mandoc_strdup(sym);
 983                         } else
 984                                 parent->right = mandoc_strndup(start, sz);
 985                 }
 986                 parent = parent->parent;
 987                 if (EQN_TOK_BRACE_CLOSE == tok && parent &&
 988                     (parent->type == EQN_PILE ||
 989                      parent->type == EQN_MATRIX))
 990                         parent = parent->parent;
 991                 /* Close out any "singleton" lists. */
 992                 while (parent->type == EQN_LISTONE &&
 993                     parent->args == parent->expectargs)
 994                         parent = parent->parent;
 995                 break;
 996         case (EQN_TOK_BRACE_OPEN):
 997         case (EQN_TOK_LEFT):
 998                 /*
 999                  * If we already have something in the stack and we're
1000                  * in an expression, then rewind til we're not any more
1001                  * (just like with the text node).
1002                  */
1003                 while (parent->args == parent->expectargs)
1004                         parent = parent->parent;
1005                 if (EQN_TOK_LEFT == tok &&
1006                     (start = eqn_nexttok(ep, &sz)) == NULL) {
1007                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
1008                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
1009                         break;
1010                 }
1011                 parent = eqn_box_alloc(ep, parent);
1012                 parent->type = EQN_LIST;
1013                 if (EQN_TOK_LEFT == tok) {
1014                         if (STRNEQ(start, sz, "ceiling", 7)) {
1015                                 strlcpy(sym, "\\[lc]", sizeof(sym));
1016                                 parent->left = mandoc_strdup(sym);
1017                         } else if (STRNEQ(start, sz, "floor", 5)) {
1018                                 strlcpy(sym, "\\[lf]", sizeof(sym));
1019                                 parent->left = mandoc_strdup(sym);
1020                         } else
1021                                 parent->left = mandoc_strndup(start, sz);
1022                 }
1023                 break;
1024         case (EQN_TOK_PILE):
1025         case (EQN_TOK_LPILE):
1026         case (EQN_TOK_RPILE):
1027         case (EQN_TOK_CPILE):
1028         case (EQN_TOK_CCOL):
1029         case (EQN_TOK_LCOL):
1030         case (EQN_TOK_RCOL):
1031                 while (parent->args == parent->expectargs)
1032                         parent = parent->parent;
1033                 parent = eqn_box_alloc(ep, parent);
1034                 parent->type = EQN_PILE;
1035                 parent->expectargs = 1;
1036                 break;
1037         case (EQN_TOK_ABOVE):
1038                 for (cur = parent; cur != NULL; cur = cur->parent)
1039                         if (cur->type == EQN_PILE)
1040                                 break;
1041                 if (cur == NULL) {
1042                         mandoc_msg(MANDOCERR_IT_STRAY, ep->parse,
1043                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
1044                         break;
1045                 }
1046                 parent = eqn_box_alloc(ep, cur);
1047                 parent->type = EQN_LIST;
1048                 break;
1049         case (EQN_TOK_MATRIX):
1050                 while (parent->args == parent->expectargs)
1051                         parent = parent->parent;
1052                 parent = eqn_box_alloc(ep, parent);
1053                 parent->type = EQN_MATRIX;
1054                 parent->expectargs = 1;
1055                 break;
1056         case (EQN_TOK_EOF):
1057                 /*
1058                  * End of file!
1059                  * TODO: make sure we're not in an open subexpression.
1060                  */
1061                 return(ROFF_EQN);
1062         default:
1063                 assert(tok == EQN_TOK__MAX);
1064                 assert(NULL != p);
1065                 /*
1066                  * If we already have something in the stack and we're
1067                  * in an expression, then rewind til we're not any more.
1068                  */
1069                 while (parent->args == parent->expectargs)
1070                         parent = parent->parent;
1071                 cur = eqn_box_alloc(ep, parent);
1072                 cur->type = EQN_TEXT;
1073                 for (i = 0; i < EQNSYM__MAX; i++)
1074                         if (0 == strcmp(eqnsyms[i].str, p)) {
1075                                 (void)snprintf(sym, sizeof(sym),
1076                                         "\\[%s]", eqnsyms[i].sym);
1077                                 cur->text = mandoc_strdup(sym);
1078                                 free(p);
1079                                 break;
1080                         }
1081
1082                 if (i == EQNSYM__MAX)
1083                         cur->text = p;
1084                 /*
1085                  * Post-process list status.
1086                  */
1087                 while (parent->type == EQN_LISTONE &&
1088                     parent->args == parent->expectargs)
1089                         parent = parent->parent;
1090                 break;
1091         }
1092         goto next_tok;
1093 }
1094
1095 enum rofferr
1096 eqn_end(struct eqn_node **epp)
1097 {
1098         struct eqn_node *ep;
1099
1100         ep = *epp;
1101         *epp = NULL;
1102
1103         ep->eqn.root = mandoc_calloc(1, sizeof(struct eqn_box));
1104         ep->eqn.root->expectargs = UINT_MAX;
1105         return(eqn_parse(ep, ep->eqn.root));
1106 }
1107
1108 void
1109 eqn_free(struct eqn_node *p)
1110 {
1111         int              i;
1112
1113         eqn_box_free(p->eqn.root);
1114
1115         for (i = 0; i < (int)p->defsz; i++) {
1116                 free(p->defs[i].key);
1117                 free(p->defs[i].val);
1118         }
1119
1120         free(p->data);
1121         free(p->defs);
1122         free(p);
1123 }