eqn.c

   1 /*      $Id: eqn.c,v 1.63 2017/06/20 17:24:35 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_minus,
 198         EQNSYM__MAX
 199 };
 200
 201 struct  eqnsym {
 202         const char      *str;
 203         const char      *sym;
 204 };
 205
 206 static  const struct eqnsym eqnsyms[EQNSYM__MAX] = {
 207         { "alpha", "*a" }, /* EQNSYM_alpha */
 208         { "beta", "*b" }, /* EQNSYM_beta */
 209         { "chi", "*x" }, /* EQNSYM_chi */
 210         { "delta", "*d" }, /* EQNSYM_delta */
 211         { "epsilon", "*e" }, /* EQNSYM_epsilon */
 212         { "eta", "*y" }, /* EQNSYM_eta */
 213         { "gamma", "*g" }, /* EQNSYM_gamma */
 214         { "iota", "*i" }, /* EQNSYM_iota */
 215         { "kappa", "*k" }, /* EQNSYM_kappa */
 216         { "lambda", "*l" }, /* EQNSYM_lambda */
 217         { "mu", "*m" }, /* EQNSYM_mu */
 218         { "nu", "*n" }, /* EQNSYM_nu */
 219         { "omega", "*w" }, /* EQNSYM_omega */
 220         { "omicron", "*o" }, /* EQNSYM_omicron */
 221         { "phi", "*f" }, /* EQNSYM_phi */
 222         { "pi", "*p" }, /* EQNSYM_pi */
 223         { "psi", "*q" }, /* EQNSYM_psi */
 224         { "rho", "*r" }, /* EQNSYM_rho */
 225         { "sigma", "*s" }, /* EQNSYM_sigma */
 226         { "tau", "*t" }, /* EQNSYM_tau */
 227         { "theta", "*h" }, /* EQNSYM_theta */
 228         { "upsilon", "*u" }, /* EQNSYM_upsilon */
 229         { "xi", "*c" }, /* EQNSYM_xi */
 230         { "zeta", "*z" }, /* EQNSYM_zeta */
 231         { "DELTA", "*D" }, /* EQNSYM_DELTA */
 232         { "GAMMA", "*G" }, /* EQNSYM_GAMMA */
 233         { "LAMBDA", "*L" }, /* EQNSYM_LAMBDA */
 234         { "OMEGA", "*W" }, /* EQNSYM_OMEGA */
 235         { "PHI", "*F" }, /* EQNSYM_PHI */
 236         { "PI", "*P" }, /* EQNSYM_PI */
 237         { "PSI", "*Q" }, /* EQNSYM_PSI */
 238         { "SIGMA", "*S" }, /* EQNSYM_SIGMA */
 239         { "THETA", "*H" }, /* EQNSYM_THETA */
 240         { "UPSILON", "*U" }, /* EQNSYM_UPSILON */
 241         { "XI", "*C" }, /* EQNSYM_XI */
 242         { "inter", "ca" }, /* EQNSYM_inter */
 243         { "union", "cu" }, /* EQNSYM_union */
 244         { "prod", "product" }, /* EQNSYM_prod */
 245         { "int", "integral" }, /* EQNSYM_int */
 246         { "sum", "sum" }, /* EQNSYM_sum */
 247         { "grad", "gr" }, /* EQNSYM_grad */
 248         { "del", "gr" }, /* EQNSYM_del */
 249         { "times", "mu" }, /* EQNSYM_times */
 250         { "cdot", "pc" }, /* EQNSYM_cdot */
 251         { "nothing", "&" }, /* EQNSYM_nothing */
 252         { "approx", "~~" }, /* EQNSYM_approx */
 253         { "prime", "fm" }, /* EQNSYM_prime */
 254         { "half", "12" }, /* EQNSYM_half */
 255         { "partial", "pd" }, /* EQNSYM_partial */
 256         { "inf", "if" }, /* EQNSYM_inf */
 257         { ">>", ">>" }, /* EQNSYM_muchgreat */
 258         { "<<", "<<" }, /* EQNSYM_muchless */
 259         { "<-", "<-" }, /* EQNSYM_larrow */
 260         { "->", "->" }, /* EQNSYM_rarrow */
 261         { "+-", "+-" }, /* EQNSYM_pm */
 262         { "!=", "!=" }, /* EQNSYM_nequal */
 263         { "==", "==" }, /* EQNSYM_equiv */
 264         { "<=", "<=" }, /* EQNSYM_lessequal */
 265         { ">=", ">=" }, /* EQNSYM_moreequal */
 266         { "-", "mi" }, /* EQNSYM_minus */
 267 };
 268
 269 static  struct eqn_box  *eqn_box_alloc(struct eqn_node *, struct eqn_box *);
 270 static  void             eqn_box_free(struct eqn_box *);
 271 static  struct eqn_box  *eqn_box_makebinary(struct eqn_node *,
 272                                 enum eqn_post, struct eqn_box *);
 273 static  void             eqn_def(struct eqn_node *);
 274 static  struct eqn_def  *eqn_def_find(struct eqn_node *, const char *, size_t);
 275 static  void             eqn_delim(struct eqn_node *);
 276 static  const char      *eqn_next(struct eqn_node *, char, size_t *, int);
 277 static  const char      *eqn_nextrawtok(struct eqn_node *, size_t *);
 278 static  const char      *eqn_nexttok(struct eqn_node *, size_t *);
 279 static  enum rofferr     eqn_parse(struct eqn_node *, struct eqn_box *);
 280 static  enum eqn_tok     eqn_tok_parse(struct eqn_node *, char **);
 281 static  void             eqn_undef(struct eqn_node *);
 282
 283
 284 enum rofferr
 285 eqn_read(struct eqn_node **epp, int ln,
 286                 const char *p, int pos, int *offs)
 287 {
 288         size_t           sz;
 289         struct eqn_node *ep;
 290         enum rofferr     er;
 291
 292         ep = *epp;
 293
 294         /*
 295          * If we're the terminating mark, unset our equation status and
 296          * validate the full equation.
 297          */
 298
 299         if (0 == strncmp(p, ".EN", 3)) {
 300                 er = eqn_end(epp);
 301                 p += 3;
 302                 while (' ' == *p || '\t' == *p)
 303                         p++;
 304                 if ('\0' == *p)
 305                         return er;
 306                 mandoc_vmsg(MANDOCERR_ARG_SKIP, ep->parse,
 307                     ln, pos, "EN %s", p);
 308                 return er;
 309         }
 310
 311         /*
 312          * Build up the full string, replacing all newlines with regular
 313          * whitespace.
 314          */
 315
 316         sz = strlen(p + pos) + 1;
 317         ep->data = mandoc_realloc(ep->data, ep->sz + sz + 1);
 318
 319         /* First invocation: nil terminate the string. */
 320
 321         if (0 == ep->sz)
 322                 *ep->data = '\0';
 323
 324         ep->sz += sz;
 325         strlcat(ep->data, p + pos, ep->sz + 1);
 326         strlcat(ep->data, " ", ep->sz + 1);
 327         return ROFF_IGN;
 328 }
 329
 330 struct eqn_node *
 331 eqn_alloc(int pos, int line, struct mparse *parse)
 332 {
 333         struct eqn_node *p;
 334
 335         p = mandoc_calloc(1, sizeof(struct eqn_node));
 336
 337         p->parse = parse;
 338         p->eqn.ln = line;
 339         p->eqn.pos = pos;
 340         p->gsize = EQN_DEFSIZE;
 341
 342         return p;
 343 }
 344
 345 /*
 346  * Find the key "key" of the give size within our eqn-defined values.
 347  */
 348 static struct eqn_def *
 349 eqn_def_find(struct eqn_node *ep, const char *key, size_t sz)
 350 {
 351         int              i;
 352
 353         for (i = 0; i < (int)ep->defsz; i++)
 354                 if (ep->defs[i].keysz && STRNEQ(ep->defs[i].key,
 355                     ep->defs[i].keysz, key, sz))
 356                         return &ep->defs[i];
 357
 358         return NULL;
 359 }
 360
 361 /*
 362  * Get the next token from the input stream using the given quote
 363  * character.
 364  * Optionally make any replacements.
 365  */
 366 static const char *
 367 eqn_next(struct eqn_node *ep, char quote, size_t *sz, int repl)
 368 {
 369         static size_t    last_len;
 370         static int       lim;
 371
 372         char            *start, *next;
 373         int              q, diff;
 374         size_t           ssz, dummy;
 375         struct eqn_def  *def;
 376
 377         if (NULL == sz)
 378                 sz = &dummy;
 379
 380         if (ep->cur >= last_len)
 381                 lim = 0;
 382         ep->rew = ep->cur;
 383 again:
 384         /* Prevent self-definitions. */
 385
 386         if (lim >= EQN_NEST_MAX) {
 387                 mandoc_msg(MANDOCERR_ROFFLOOP, ep->parse,
 388                     ep->eqn.ln, ep->eqn.pos, NULL);
 389                 return NULL;
 390         }
 391
 392         ep->cur = ep->rew;
 393         start = &ep->data[(int)ep->cur];
 394         q = 0;
 395
 396         if ('\0' == *start)
 397                 return NULL;
 398
 399         if (quote == *start) {
 400                 ep->cur++;
 401                 q = 1;
 402         }
 403
 404         start = &ep->data[(int)ep->cur];
 405
 406         if ( ! q) {
 407                 if ('{' == *start || '}' == *start)
 408                         ssz = 1;
 409                 else
 410                         ssz = strcspn(start + 1, " ^~\"{}\t") + 1;
 411                 next = start + (int)ssz;
 412                 if ('\0' == *next)
 413                         next = NULL;
 414         } else
 415                 next = strchr(start, quote);
 416
 417         if (NULL != next) {
 418                 *sz = (size_t)(next - start);
 419                 ep->cur += *sz;
 420                 if (q)
 421                         ep->cur++;
 422                 while (' ' == ep->data[(int)ep->cur] ||
 423                     '\t' == ep->data[(int)ep->cur] ||
 424                     '^' == ep->data[(int)ep->cur] ||
 425                     '~' == ep->data[(int)ep->cur])
 426                         ep->cur++;
 427         } else {
 428                 if (q)
 429                         mandoc_msg(MANDOCERR_ARG_QUOTE, ep->parse,
 430                             ep->eqn.ln, ep->eqn.pos, NULL);
 431                 next = strchr(start, '\0');
 432                 *sz = (size_t)(next - start);
 433                 ep->cur += *sz;
 434         }
 435
 436         /* Quotes aren't expanded for values. */
 437
 438         if (q || ! repl)
 439                 return start;
 440
 441         if (NULL != (def = eqn_def_find(ep, start, *sz))) {
 442                 diff = def->valsz - *sz;
 443
 444                 if (def->valsz > *sz) {
 445                         ep->sz += diff;
 446                         ep->data = mandoc_realloc(ep->data, ep->sz + 1);
 447                         ep->data[ep->sz] = '\0';
 448                         start = &ep->data[(int)ep->rew];
 449                 }
 450
 451                 diff = def->valsz - *sz;
 452                 memmove(start + *sz + diff, start + *sz,
 453                     (strlen(start) - *sz) + 1);
 454                 memcpy(start, def->val, def->valsz);
 455                 last_len = start - ep->data + def->valsz;
 456                 lim++;
 457                 goto again;
 458         }
 459
 460         return start;
 461 }
 462
 463 /*
 464  * Get the next delimited token using the default current quote
 465  * character.
 466  */
 467 static const char *
 468 eqn_nexttok(struct eqn_node *ep, size_t *sz)
 469 {
 470
 471         return eqn_next(ep, '"', sz, 1);
 472 }
 473
 474 /*
 475  * Get next token without replacement.
 476  */
 477 static const char *
 478 eqn_nextrawtok(struct eqn_node *ep, size_t *sz)
 479 {
 480
 481         return eqn_next(ep, '"', sz, 0);
 482 }
 483
 484 /*
 485  * Parse a token from the stream of text.
 486  * A token consists of one of the recognised eqn(7) strings.
 487  * Strings are separated by delimiting marks.
 488  * This returns EQN_TOK_EOF when there are no more tokens.
 489  * If the token is an unrecognised string literal, then it returns
 490  * EQN_TOK__MAX and sets the "p" pointer to an allocated, nil-terminated
 491  * string.
 492  * This must be later freed with free(3).
 493  */
 494 static enum eqn_tok
 495 eqn_tok_parse(struct eqn_node *ep, char **p)
 496 {
 497         const char      *start;
 498         size_t           i, sz;
 499         int              quoted;
 500
 501         if (NULL != p)
 502                 *p = NULL;
 503
 504         quoted = ep->data[ep->cur] == '"';
 505
 506         if (NULL == (start = eqn_nexttok(ep, &sz)))
 507                 return EQN_TOK_EOF;
 508
 509         if (quoted) {
 510                 if (p != NULL)
 511                         *p = mandoc_strndup(start, sz);
 512                 return EQN_TOK__MAX;
 513         }
 514
 515         for (i = 0; i < EQN_TOK__MAX; i++) {
 516                 if (NULL == eqn_toks[i])
 517                         continue;
 518                 if (STRNEQ(start, sz, eqn_toks[i], strlen(eqn_toks[i])))
 519                         break;
 520         }
 521
 522         if (i == EQN_TOK__MAX && NULL != p)
 523                 *p = mandoc_strndup(start, sz);
 524
 525         return i;
 526 }
 527
 528 static void
 529 eqn_box_free(struct eqn_box *bp)
 530 {
 531
 532         if (bp->first)
 533                 eqn_box_free(bp->first);
 534         if (bp->next)
 535                 eqn_box_free(bp->next);
 536
 537         free(bp->text);
 538         free(bp->left);
 539         free(bp->right);
 540         free(bp->top);
 541         free(bp->bottom);
 542         free(bp);
 543 }
 544
 545 /*
 546  * Allocate a box as the last child of the parent node.
 547  */
 548 static struct eqn_box *
 549 eqn_box_alloc(struct eqn_node *ep, struct eqn_box *parent)
 550 {
 551         struct eqn_box  *bp;
 552
 553         bp = mandoc_calloc(1, sizeof(struct eqn_box));
 554         bp->parent = parent;
 555         bp->parent->args++;
 556         bp->expectargs = UINT_MAX;
 557         bp->size = ep->gsize;
 558
 559         if (NULL != parent->first) {
 560                 parent->last->next = bp;
 561                 bp->prev = parent->last;
 562         } else
 563                 parent->first = bp;
 564
 565         parent->last = bp;
 566         return bp;
 567 }
 568
 569 /*
 570  * Reparent the current last node (of the current parent) under a new
 571  * EQN_SUBEXPR as the first element.
 572  * Then return the new parent.
 573  * The new EQN_SUBEXPR will have a two-child limit.
 574  */
 575 static struct eqn_box *
 576 eqn_box_makebinary(struct eqn_node *ep,
 577         enum eqn_post pos, struct eqn_box *parent)
 578 {
 579         struct eqn_box  *b, *newb;
 580
 581         assert(NULL != parent->last);
 582         b = parent->last;
 583         if (parent->last == parent->first)
 584                 parent->first = NULL;
 585         parent->args--;
 586         parent->last = b->prev;
 587         b->prev = NULL;
 588         newb = eqn_box_alloc(ep, parent);
 589         newb->pos = pos;
 590         newb->type = EQN_SUBEXPR;
 591         newb->expectargs = 2;
 592         newb->args = 1;
 593         newb->first = newb->last = b;
 594         newb->first->next = NULL;
 595         b->parent = newb;
 596         return newb;
 597 }
 598
 599 /*
 600  * Parse the "delim" control statement.
 601  */
 602 static void
 603 eqn_delim(struct eqn_node *ep)
 604 {
 605         const char      *start;
 606         size_t           sz;
 607
 608         if ((start = eqn_nextrawtok(ep, &sz)) == NULL)
 609                 mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 610                     ep->eqn.ln, ep->eqn.pos, "delim");
 611         else if (strncmp(start, "off", 3) == 0)
 612                 ep->delim = 0;
 613         else if (strncmp(start, "on", 2) == 0) {
 614                 if (ep->odelim && ep->cdelim)
 615                         ep->delim = 1;
 616         } else if (start[1] != '\0') {
 617                 ep->odelim = start[0];
 618                 ep->cdelim = start[1];
 619                 ep->delim = 1;
 620         }
 621 }
 622
 623 /*
 624  * Undefine a previously-defined string.
 625  */
 626 static void
 627 eqn_undef(struct eqn_node *ep)
 628 {
 629         const char      *start;
 630         struct eqn_def  *def;
 631         size_t           sz;
 632
 633         if ((start = eqn_nextrawtok(ep, &sz)) == NULL) {
 634                 mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 635                     ep->eqn.ln, ep->eqn.pos, "undef");
 636                 return;
 637         }
 638         if ((def = eqn_def_find(ep, start, sz)) == NULL)
 639                 return;
 640         free(def->key);
 641         free(def->val);
 642         def->key = def->val = NULL;
 643         def->keysz = def->valsz = 0;
 644 }
 645
 646 static void
 647 eqn_def(struct eqn_node *ep)
 648 {
 649         const char      *start;
 650         size_t           sz;
 651         struct eqn_def  *def;
 652         int              i;
 653
 654         if ((start = eqn_nextrawtok(ep, &sz)) == NULL) {
 655                 mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 656                     ep->eqn.ln, ep->eqn.pos, "define");
 657                 return;
 658         }
 659
 660         /*
 661          * Search for a key that already exists.
 662          * Create a new key if none is found.
 663          */
 664         if (NULL == (def = eqn_def_find(ep, start, sz))) {
 665                 /* Find holes in string array. */
 666                 for (i = 0; i < (int)ep->defsz; i++)
 667                         if (0 == ep->defs[i].keysz)
 668                                 break;
 669
 670                 if (i == (int)ep->defsz) {
 671                         ep->defsz++;
 672                         ep->defs = mandoc_reallocarray(ep->defs,
 673                             ep->defsz, sizeof(struct eqn_def));
 674                         ep->defs[i].key = ep->defs[i].val = NULL;
 675                 }
 676
 677                 def = ep->defs + i;
 678                 free(def->key);
 679                 def->key = mandoc_strndup(start, sz);
 680                 def->keysz = sz;
 681         }
 682
 683         start = eqn_next(ep, ep->data[(int)ep->cur], &sz, 0);
 684         if (start == NULL) {
 685                 mandoc_vmsg(MANDOCERR_REQ_EMPTY, ep->parse,
 686                     ep->eqn.ln, ep->eqn.pos, "define %s", def->key);
 687                 free(def->key);
 688                 free(def->val);
 689                 def->key = def->val = NULL;
 690                 def->keysz = def->valsz = 0;
 691                 return;
 692         }
 693         free(def->val);
 694         def->val = mandoc_strndup(start, sz);
 695         def->valsz = sz;
 696 }
 697
 698 /*
 699  * Recursively parse an eqn(7) expression.
 700  */
 701 static enum rofferr
 702 eqn_parse(struct eqn_node *ep, struct eqn_box *parent)
 703 {
 704         char             sym[64];
 705         struct eqn_box  *cur;
 706         const char      *start;
 707         char            *p;
 708         size_t           i, sz;
 709         enum eqn_tok     tok, subtok;
 710         enum eqn_post    pos;
 711         int              size;
 712
 713         assert(parent != NULL);
 714
 715         /*
 716          * Empty equation.
 717          * Do not add it to the high-level syntax tree.
 718          */
 719
 720         if (ep->data == NULL)
 721                 return ROFF_IGN;
 722
 723 next_tok:
 724         tok = eqn_tok_parse(ep, &p);
 725
 726 this_tok:
 727         switch (tok) {
 728         case EQN_TOK_UNDEF:
 729                 eqn_undef(ep);
 730                 break;
 731         case EQN_TOK_NDEFINE:
 732         case EQN_TOK_DEFINE:
 733                 eqn_def(ep);
 734                 break;
 735         case EQN_TOK_TDEFINE:
 736                 if (eqn_nextrawtok(ep, NULL) == NULL ||
 737                     eqn_next(ep, ep->data[(int)ep->cur], NULL, 0) == NULL)
 738                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 739                             ep->eqn.ln, ep->eqn.pos, "tdefine");
 740                 break;
 741         case EQN_TOK_DELIM:
 742                 eqn_delim(ep);
 743                 break;
 744         case EQN_TOK_GFONT:
 745                 if (eqn_nextrawtok(ep, NULL) == NULL)
 746                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 747                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 748                 break;
 749         case EQN_TOK_MARK:
 750         case EQN_TOK_LINEUP:
 751                 /* Ignore these. */
 752                 break;
 753         case EQN_TOK_DYAD:
 754         case EQN_TOK_VEC:
 755         case EQN_TOK_UNDER:
 756         case EQN_TOK_BAR:
 757         case EQN_TOK_TILDE:
 758         case EQN_TOK_HAT:
 759         case EQN_TOK_DOT:
 760         case EQN_TOK_DOTDOT:
 761                 if (parent->last == NULL) {
 762                         mandoc_msg(MANDOCERR_EQN_NOBOX, ep->parse,
 763                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 764                         cur = eqn_box_alloc(ep, parent);
 765                         cur->type = EQN_TEXT;
 766                         cur->text = mandoc_strdup("");
 767                 }
 768                 parent = eqn_box_makebinary(ep, EQNPOS_NONE, parent);
 769                 parent->type = EQN_LISTONE;
 770                 parent->expectargs = 1;
 771                 switch (tok) {
 772                 case EQN_TOK_DOTDOT:
 773                         strlcpy(sym, "\\[ad]", sizeof(sym));
 774                         break;
 775                 case EQN_TOK_VEC:
 776                         strlcpy(sym, "\\[->]", sizeof(sym));
 777                         break;
 778                 case EQN_TOK_DYAD:
 779                         strlcpy(sym, "\\[<>]", sizeof(sym));
 780                         break;
 781                 case EQN_TOK_TILDE:
 782                         strlcpy(sym, "\\[a~]", sizeof(sym));
 783                         break;
 784                 case EQN_TOK_UNDER:
 785                         strlcpy(sym, "\\[ul]", sizeof(sym));
 786                         break;
 787                 case EQN_TOK_BAR:
 788                         strlcpy(sym, "\\[rl]", sizeof(sym));
 789                         break;
 790                 case EQN_TOK_DOT:
 791                         strlcpy(sym, "\\[a.]", sizeof(sym));
 792                         break;
 793                 case EQN_TOK_HAT:
 794                         strlcpy(sym, "\\[ha]", sizeof(sym));
 795                         break;
 796                 default:
 797                         abort();
 798                 }
 799
 800                 switch (tok) {
 801                 case EQN_TOK_DOTDOT:
 802                 case EQN_TOK_VEC:
 803                 case EQN_TOK_DYAD:
 804                 case EQN_TOK_TILDE:
 805                 case EQN_TOK_BAR:
 806                 case EQN_TOK_DOT:
 807                 case EQN_TOK_HAT:
 808                         parent->top = mandoc_strdup(sym);
 809                         break;
 810                 case EQN_TOK_UNDER:
 811                         parent->bottom = mandoc_strdup(sym);
 812                         break;
 813                 default:
 814                         abort();
 815                 }
 816                 parent = parent->parent;
 817                 break;
 818         case EQN_TOK_FWD:
 819         case EQN_TOK_BACK:
 820         case EQN_TOK_DOWN:
 821         case EQN_TOK_UP:
 822                 subtok = eqn_tok_parse(ep, NULL);
 823                 if (subtok != EQN_TOK__MAX) {
 824                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 825                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 826                         tok = subtok;
 827                         goto this_tok;
 828                 }
 829                 break;
 830         case EQN_TOK_FAT:
 831         case EQN_TOK_ROMAN:
 832         case EQN_TOK_ITALIC:
 833         case EQN_TOK_BOLD:
 834                 while (parent->args == parent->expectargs)
 835                         parent = parent->parent;
 836                 /*
 837                  * These values apply to the next word or sequence of
 838                  * words; thus, we mark that we'll have a child with
 839                  * exactly one of those.
 840                  */
 841                 parent = eqn_box_alloc(ep, parent);
 842                 parent->type = EQN_LISTONE;
 843                 parent->expectargs = 1;
 844                 switch (tok) {
 845                 case EQN_TOK_FAT:
 846                         parent->font = EQNFONT_FAT;
 847                         break;
 848                 case EQN_TOK_ROMAN:
 849                         parent->font = EQNFONT_ROMAN;
 850                         break;
 851                 case EQN_TOK_ITALIC:
 852                         parent->font = EQNFONT_ITALIC;
 853                         break;
 854                 case EQN_TOK_BOLD:
 855                         parent->font = EQNFONT_BOLD;
 856                         break;
 857                 default:
 858                         abort();
 859                 }
 860                 break;
 861         case EQN_TOK_SIZE:
 862         case EQN_TOK_GSIZE:
 863                 /* Accept two values: integral size and a single. */
 864                 if (NULL == (start = eqn_nexttok(ep, &sz))) {
 865                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
 866                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 867                         break;
 868                 }
 869                 size = mandoc_strntoi(start, sz, 10);
 870                 if (-1 == size) {
 871                         mandoc_msg(MANDOCERR_IT_NONUM, ep->parse,
 872                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 873                         break;
 874                 }
 875                 if (EQN_TOK_GSIZE == tok) {
 876                         ep->gsize = size;
 877                         break;
 878                 }
 879                 parent = eqn_box_alloc(ep, parent);
 880                 parent->type = EQN_LISTONE;
 881                 parent->expectargs = 1;
 882                 parent->size = size;
 883                 break;
 884         case EQN_TOK_FROM:
 885         case EQN_TOK_TO:
 886         case EQN_TOK_SUB:
 887         case EQN_TOK_SUP:
 888                 /*
 889                  * We have a left-right-associative expression.
 890                  * Repivot under a positional node, open a child scope
 891                  * and keep on reading.
 892                  */
 893                 if (parent->last == NULL) {
 894                         mandoc_msg(MANDOCERR_EQN_NOBOX, ep->parse,
 895                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 896                         cur = eqn_box_alloc(ep, parent);
 897                         cur->type = EQN_TEXT;
 898                         cur->text = mandoc_strdup("");
 899                 }
 900                 /* Handle the "subsup" and "fromto" positions. */
 901                 if (EQN_TOK_SUP == tok && parent->pos == EQNPOS_SUB) {
 902                         parent->expectargs = 3;
 903                         parent->pos = EQNPOS_SUBSUP;
 904                         break;
 905                 }
 906                 if (EQN_TOK_TO == tok && parent->pos == EQNPOS_FROM) {
 907                         parent->expectargs = 3;
 908                         parent->pos = EQNPOS_FROMTO;
 909                         break;
 910                 }
 911                 switch (tok) {
 912                 case EQN_TOK_FROM:
 913                         pos = EQNPOS_FROM;
 914                         break;
 915                 case EQN_TOK_TO:
 916                         pos = EQNPOS_TO;
 917                         break;
 918                 case EQN_TOK_SUP:
 919                         pos = EQNPOS_SUP;
 920                         break;
 921                 case EQN_TOK_SUB:
 922                         pos = EQNPOS_SUB;
 923                         break;
 924                 default:
 925                         abort();
 926                 }
 927                 parent = eqn_box_makebinary(ep, pos, parent);
 928                 break;
 929         case EQN_TOK_SQRT:
 930                 while (parent->args == parent->expectargs)
 931                         parent = parent->parent;
 932                 /*
 933                  * Accept a left-right-associative set of arguments just
 934                  * like sub and sup and friends but without rebalancing
 935                  * under a pivot.
 936                  */
 937                 parent = eqn_box_alloc(ep, parent);
 938                 parent->type = EQN_SUBEXPR;
 939                 parent->pos = EQNPOS_SQRT;
 940                 parent->expectargs = 1;
 941                 break;
 942         case EQN_TOK_OVER:
 943                 /*
 944                  * We have a right-left-associative fraction.
 945                  * Close out anything that's currently open, then
 946                  * rebalance and continue reading.
 947                  */
 948                 if (parent->last == NULL) {
 949                         mandoc_msg(MANDOCERR_EQN_NOBOX, ep->parse,
 950                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 951                         cur = eqn_box_alloc(ep, parent);
 952                         cur->type = EQN_TEXT;
 953                         cur->text = mandoc_strdup("");
 954                 }
 955                 while (EQN_SUBEXPR == parent->type)
 956                         parent = parent->parent;
 957                 parent = eqn_box_makebinary(ep, EQNPOS_OVER, parent);
 958                 break;
 959         case EQN_TOK_RIGHT:
 960         case EQN_TOK_BRACE_CLOSE:
 961                 /*
 962                  * Close out the existing brace.
 963                  * FIXME: this is a shitty sentinel: we should really
 964                  * have a native EQN_BRACE type or whatnot.
 965                  */
 966                 for (cur = parent; cur != NULL; cur = cur->parent)
 967                         if (cur->type == EQN_LIST &&
 968                             (tok == EQN_TOK_BRACE_CLOSE ||
 969                              cur->left != NULL))
 970                                 break;
 971                 if (cur == NULL) {
 972                         mandoc_msg(MANDOCERR_BLK_NOTOPEN, ep->parse,
 973                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
 974                         break;
 975                 }
 976                 parent = cur;
 977                 if (EQN_TOK_RIGHT == tok) {
 978                         if (NULL == (start = eqn_nexttok(ep, &sz))) {
 979                                 mandoc_msg(MANDOCERR_REQ_EMPTY,
 980                                     ep->parse, ep->eqn.ln,
 981                                     ep->eqn.pos, eqn_toks[tok]);
 982                                 break;
 983                         }
 984                         /* Handling depends on right/left. */
 985                         if (STRNEQ(start, sz, "ceiling", 7)) {
 986                                 strlcpy(sym, "\\[rc]", sizeof(sym));
 987                                 parent->right = mandoc_strdup(sym);
 988                         } else if (STRNEQ(start, sz, "floor", 5)) {
 989                                 strlcpy(sym, "\\[rf]", sizeof(sym));
 990                                 parent->right = mandoc_strdup(sym);
 991                         } else
 992                                 parent->right = mandoc_strndup(start, sz);
 993                 }
 994                 parent = parent->parent;
 995                 if (tok == EQN_TOK_BRACE_CLOSE &&
 996                     (parent->type == EQN_PILE ||
 997                      parent->type == EQN_MATRIX))
 998                         parent = parent->parent;
 999                 /* Close out any "singleton" lists. */
1000                 while (parent->type == EQN_LISTONE &&
1001                     parent->args == parent->expectargs)
1002                         parent = parent->parent;
1003                 break;
1004         case EQN_TOK_BRACE_OPEN:
1005         case EQN_TOK_LEFT:
1006                 /*
1007                  * If we already have something in the stack and we're
1008                  * in an expression, then rewind til we're not any more
1009                  * (just like with the text node).
1010                  */
1011                 while (parent->args == parent->expectargs)
1012                         parent = parent->parent;
1013                 if (EQN_TOK_LEFT == tok &&
1014                     (start = eqn_nexttok(ep, &sz)) == NULL) {
1015                         mandoc_msg(MANDOCERR_REQ_EMPTY, ep->parse,
1016                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
1017                         break;
1018                 }
1019                 parent = eqn_box_alloc(ep, parent);
1020                 parent->type = EQN_LIST;
1021                 if (EQN_TOK_LEFT == tok) {
1022                         if (STRNEQ(start, sz, "ceiling", 7)) {
1023                                 strlcpy(sym, "\\[lc]", sizeof(sym));
1024                                 parent->left = mandoc_strdup(sym);
1025                         } else if (STRNEQ(start, sz, "floor", 5)) {
1026                                 strlcpy(sym, "\\[lf]", sizeof(sym));
1027                                 parent->left = mandoc_strdup(sym);
1028                         } else
1029                                 parent->left = mandoc_strndup(start, sz);
1030                 }
1031                 break;
1032         case EQN_TOK_PILE:
1033         case EQN_TOK_LPILE:
1034         case EQN_TOK_RPILE:
1035         case EQN_TOK_CPILE:
1036         case EQN_TOK_CCOL:
1037         case EQN_TOK_LCOL:
1038         case EQN_TOK_RCOL:
1039                 while (parent->args == parent->expectargs)
1040                         parent = parent->parent;
1041                 parent = eqn_box_alloc(ep, parent);
1042                 parent->type = EQN_PILE;
1043                 parent->expectargs = 1;
1044                 break;
1045         case EQN_TOK_ABOVE:
1046                 for (cur = parent; cur != NULL; cur = cur->parent)
1047                         if (cur->type == EQN_PILE)
1048                                 break;
1049                 if (cur == NULL) {
1050                         mandoc_msg(MANDOCERR_IT_STRAY, ep->parse,
1051                             ep->eqn.ln, ep->eqn.pos, eqn_toks[tok]);
1052                         break;
1053                 }
1054                 parent = eqn_box_alloc(ep, cur);
1055                 parent->type = EQN_LIST;
1056                 break;
1057         case EQN_TOK_MATRIX:
1058                 while (parent->args == parent->expectargs)
1059                         parent = parent->parent;
1060                 parent = eqn_box_alloc(ep, parent);
1061                 parent->type = EQN_MATRIX;
1062                 parent->expectargs = 1;
1063                 break;
1064         case EQN_TOK_EOF:
1065                 /*
1066                  * End of file!
1067                  * TODO: make sure we're not in an open subexpression.
1068                  */
1069                 return ROFF_EQN;
1070         default:
1071                 assert(tok == EQN_TOK__MAX);
1072                 assert(NULL != p);
1073                 /*
1074                  * If we already have something in the stack and we're
1075                  * in an expression, then rewind til we're not any more.
1076                  */
1077                 while (parent->args == parent->expectargs)
1078                         parent = parent->parent;
1079                 cur = eqn_box_alloc(ep, parent);
1080                 cur->type = EQN_TEXT;
1081                 for (i = 0; i < EQNSYM__MAX; i++)
1082                         if (0 == strcmp(eqnsyms[i].str, p)) {
1083                                 (void)snprintf(sym, sizeof(sym),
1084                                         "\\[%s]", eqnsyms[i].sym);
1085                                 cur->text = mandoc_strdup(sym);
1086                                 free(p);
1087                                 break;
1088                         }
1089
1090                 if (i == EQNSYM__MAX)
1091                         cur->text = p;
1092                 /*
1093                  * Post-process list status.
1094                  */
1095                 while (parent->type == EQN_LISTONE &&
1096                     parent->args == parent->expectargs)
1097                         parent = parent->parent;
1098                 break;
1099         }
1100         goto next_tok;
1101 }
1102
1103 enum rofferr
1104 eqn_end(struct eqn_node **epp)
1105 {
1106         struct eqn_node *ep;
1107
1108         ep = *epp;
1109         *epp = NULL;
1110
1111         ep->eqn.root = mandoc_calloc(1, sizeof(struct eqn_box));
1112         ep->eqn.root->expectargs = UINT_MAX;
1113         return eqn_parse(ep, ep->eqn.root);
1114 }
1115
1116 void
1117 eqn_free(struct eqn_node *p)
1118 {
1119         int              i;
1120
1121         eqn_box_free(p->eqn.root);
1122
1123         for (i = 0; i < (int)p->defsz; i++) {
1124                 free(p->defs[i].key);
1125                 free(p->defs[i].val);
1126         }
1127
1128         free(p->data);
1129         free(p->defs);
1130         free(p);
1131 }