man_macro.c

   1 /* $Id: man_macro.c,v 1.148 2022/04/27 17:11:24 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2012-2015,2017-2020,2022 Ingo Schwarze <schwarze@openbsd.org>
   4  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   5  * Copyright (c) 2013 Franco Fichtner <franco@lastsummer.de>
   6  *
   7  * Permission to use, copy, modify, and distribute this software for any
   8  * purpose with or without fee is hereby granted, provided that the above
   9  * copyright notice and this permission notice appear in all copies.
  10  *
  11  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
  12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
  14  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  15  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  16  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  17  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #include "config.h"
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdio.h>
  26 #include <stdlib.h>
  27 #include <string.h>
  28
  29 #if DEBUG_MEMORY
  30 #include "mandoc_dbg.h"
  31 #endif
  32 #include "mandoc.h"
  33 #include "roff.h"
  34 #include "man.h"
  35 #include "libmandoc.h"
  36 #include "roff_int.h"
  37 #include "libman.h"
  38
  39 static  void             blk_close(MACRO_PROT_ARGS);
  40 static  void             blk_exp(MACRO_PROT_ARGS);
  41 static  void             blk_imp(MACRO_PROT_ARGS);
  42 static  void             in_line_eoln(MACRO_PROT_ARGS);
  43 static  int              man_args(struct roff_man *, int,
  44                                 int *, char *, char **);
  45 static  void             rew_scope(struct roff_man *, enum roff_tok);
  46
  47 static const struct man_macro man_macros[MAN_MAX - MAN_TH] = {
  48         { in_line_eoln, MAN_XSCOPE }, /* TH */
  49         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* SH */
  50         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* SS */
  51         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* TP */
  52         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* TQ */
  53         { blk_imp, MAN_XSCOPE }, /* LP */
  54         { blk_imp, MAN_XSCOPE }, /* PP */
  55         { blk_imp, MAN_XSCOPE }, /* P */
  56         { blk_imp, MAN_XSCOPE }, /* IP */
  57         { blk_imp, MAN_XSCOPE }, /* HP */
  58         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* SM */
  59         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* SB */
  60         { in_line_eoln, 0 }, /* BI */
  61         { in_line_eoln, 0 }, /* IB */
  62         { in_line_eoln, 0 }, /* BR */
  63         { in_line_eoln, 0 }, /* RB */
  64         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* R */
  65         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* B */
  66         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* I */
  67         { in_line_eoln, 0 }, /* IR */
  68         { in_line_eoln, 0 }, /* RI */
  69         { blk_close, MAN_XSCOPE }, /* RE */
  70         { blk_exp, MAN_XSCOPE }, /* RS */
  71         { in_line_eoln, MAN_NSCOPED }, /* DT */
  72         { in_line_eoln, MAN_NSCOPED }, /* UC */
  73         { in_line_eoln, MAN_NSCOPED }, /* PD */
  74         { in_line_eoln, MAN_NSCOPED }, /* AT */
  75         { in_line_eoln, MAN_NSCOPED }, /* in */
  76         { blk_imp, MAN_XSCOPE }, /* SY */
  77         { blk_close, MAN_XSCOPE }, /* YS */
  78         { in_line_eoln, 0 }, /* OP */
  79         { in_line_eoln, MAN_XSCOPE }, /* EX */
  80         { in_line_eoln, MAN_XSCOPE }, /* EE */
  81         { blk_exp, MAN_XSCOPE }, /* UR */
  82         { blk_close, MAN_XSCOPE }, /* UE */
  83         { blk_exp, MAN_XSCOPE }, /* MT */
  84         { blk_close, MAN_XSCOPE }, /* ME */
  85 };
  86
  87
  88 const struct man_macro *
  89 man_macro(enum roff_tok tok)
  90 {
  91         assert(tok >= MAN_TH && tok <= MAN_MAX);
  92         return man_macros + (tok - MAN_TH);
  93 }
  94
  95 void
  96 man_unscope(struct roff_man *man, const struct roff_node *to)
  97 {
  98         struct roff_node *n;
  99
 100         to = to->parent;
 101         n = man->last;
 102         while (n != to) {
 103
 104                 /* Reached the end of the document? */
 105
 106                 if (to == NULL && ! (n->flags & NODE_VALID)) {
 107                         if (man->flags & (MAN_BLINE | MAN_ELINE) &&
 108                             man_macro(n->tok)->flags &
 109                              (MAN_BSCOPED | MAN_NSCOPED)) {
 110                                 mandoc_msg(MANDOCERR_BLK_LINE,
 111                                     n->line, n->pos,
 112                                     "EOF breaks %s", roff_name[n->tok]);
 113                                 if (man->flags & MAN_ELINE) {
 114                                         if (n->parent->type == ROFFT_ROOT ||
 115                                             (man_macro(n->parent->tok)->flags &
 116                                             MAN_ESCOPED) == 0)
 117                                                 man->flags &= ~MAN_ELINE;
 118                                 } else {
 119                                         assert(n->type == ROFFT_HEAD);
 120                                         n = n->parent;
 121                                         man->flags &= ~MAN_BLINE;
 122                                 }
 123                                 man->last = n;
 124                                 n = n->parent;
 125                                 roff_node_delete(man, man->last);
 126                                 continue;
 127                         }
 128                         if (n->type == ROFFT_BLOCK &&
 129                             man_macro(n->tok)->fp == blk_exp)
 130                                 mandoc_msg(MANDOCERR_BLK_NOEND,
 131                                     n->line, n->pos, "%s",
 132                                     roff_name[n->tok]);
 133                 }
 134
 135                 /*
 136                  * We might delete the man->last node
 137                  * in the post-validation phase.
 138                  * Save a pointer to the parent such that
 139                  * we know where to continue the iteration.
 140                  */
 141
 142                 man->last = n;
 143                 n = n->parent;
 144                 man->last->flags |= NODE_VALID;
 145         }
 146
 147         /*
 148          * If we ended up at the parent of the node we were
 149          * supposed to rewind to, that means the target node
 150          * got deleted, so add the next node we parse as a child
 151          * of the parent instead of as a sibling of the target.
 152          */
 153
 154         man->next = (man->last == to) ?
 155             ROFF_NEXT_CHILD : ROFF_NEXT_SIBLING;
 156 }
 157
 158 /*
 159  * Rewinding entails ascending the parse tree until a coherent point,
 160  * for example, the `SH' macro will close out any intervening `SS'
 161  * scopes.  When a scope is closed, it must be validated and actioned.
 162  */
 163 static void
 164 rew_scope(struct roff_man *man, enum roff_tok tok)
 165 {
 166         struct roff_node *n;
 167
 168         /* Preserve empty paragraphs before RS. */
 169
 170         n = man->last;
 171         if (tok == MAN_RS && n->child == NULL &&
 172             (n->tok == MAN_P || n->tok == MAN_PP || n->tok == MAN_LP))
 173                 return;
 174
 175         for (;;) {
 176                 if (n->type == ROFFT_ROOT)
 177                         return;
 178                 if (n->flags & NODE_VALID) {
 179                         n = n->parent;
 180                         continue;
 181                 }
 182                 if (n->type != ROFFT_BLOCK) {
 183                         if (n->parent->type == ROFFT_ROOT) {
 184                                 man_unscope(man, n);
 185                                 return;
 186                         } else {
 187                                 n = n->parent;
 188                                 continue;
 189                         }
 190                 }
 191                 if (tok != MAN_SH && (n->tok == MAN_SH ||
 192                     (tok != MAN_SS && (n->tok == MAN_SS ||
 193                      man_macro(n->tok)->fp == blk_exp))))
 194                         return;
 195                 man_unscope(man, n);
 196                 n = man->last;
 197         }
 198 }
 199
 200
 201 /*
 202  * Close out a generic explicit macro.
 203  */
 204 void
 205 blk_close(MACRO_PROT_ARGS)
 206 {
 207         enum roff_tok            ctok, ntok;
 208         const struct roff_node  *nn;
 209         char                    *p, *ep;
 210         int                      cline, cpos, la, nrew, target;
 211
 212         nrew = 1;
 213         switch (tok) {
 214         case MAN_RE:
 215                 ntok = MAN_RS;
 216                 la = *pos;
 217                 if ( ! man_args(man, line, pos, buf, &p))
 218                         break;
 219                 for (nn = man->last->parent; nn; nn = nn->parent)
 220                         if (nn->tok == ntok && nn->type == ROFFT_BLOCK)
 221                                 nrew++;
 222                 target = strtol(p, &ep, 10);
 223                 if (*ep != '\0')
 224                         mandoc_msg(MANDOCERR_ARG_EXCESS, line,
 225                             la + (buf[la] == '"') + (int)(ep - p),
 226                             "RE ... %s", ep);
 227                 free(p);
 228                 if (target == 0)
 229                         target = 1;
 230                 nrew -= target;
 231                 if (nrew < 1) {
 232                         mandoc_msg(MANDOCERR_RE_NOTOPEN,
 233                             line, ppos, "RE %d", target);
 234                         return;
 235                 }
 236                 break;
 237         case MAN_YS:
 238                 ntok = MAN_SY;
 239                 break;
 240         case MAN_UE:
 241                 ntok = MAN_UR;
 242                 break;
 243         case MAN_ME:
 244                 ntok = MAN_MT;
 245                 break;
 246         default:
 247                 abort();
 248         }
 249
 250         for (nn = man->last->parent; nn; nn = nn->parent)
 251                 if (nn->tok == ntok && nn->type == ROFFT_BLOCK && ! --nrew)
 252                         break;
 253
 254         if (nn == NULL) {
 255                 mandoc_msg(MANDOCERR_BLK_NOTOPEN,
 256                     line, ppos, "%s", roff_name[tok]);
 257                 rew_scope(man, MAN_PP);
 258                 if (tok == MAN_RE) {
 259                         roff_elem_alloc(man, line, ppos, ROFF_br);
 260                         man->last->flags |= NODE_LINE |
 261                             NODE_VALID | NODE_ENDED;
 262                         man->next = ROFF_NEXT_SIBLING;
 263                 }
 264                 return;
 265         }
 266
 267         cline = man->last->line;
 268         cpos = man->last->pos;
 269         ctok = man->last->tok;
 270         man_unscope(man, nn);
 271
 272         if (tok == MAN_RE && nn->head->aux > 0)
 273                 roff_setreg(man->roff, "an-margin", nn->head->aux, '-');
 274
 275         /* Trailing text. */
 276
 277         if (buf[*pos] != '\0') {
 278                 roff_word_alloc(man, line, ppos, buf + *pos);
 279                 man->last->flags |= NODE_DELIMC;
 280                 if (mandoc_eos(man->last->string, strlen(man->last->string)))
 281                         man->last->flags |= NODE_EOS;
 282         }
 283
 284         /* Move a trailing paragraph behind the block. */
 285
 286         if (ctok == MAN_LP || ctok == MAN_PP || ctok == MAN_P) {
 287                 *pos = strlen(buf);
 288                 blk_imp(man, ctok, cline, cpos, pos, buf);
 289         }
 290
 291         /* Synopsis blocks need an explicit end marker for spacing. */
 292
 293         if (tok == MAN_YS && man->last == nn) {
 294                 roff_elem_alloc(man, line, ppos, tok);
 295                 man_unscope(man, man->last);
 296         }
 297 }
 298
 299 void
 300 blk_exp(MACRO_PROT_ARGS)
 301 {
 302         struct roff_node *head;
 303         char            *p;
 304         int              la;
 305
 306         if (tok == MAN_RS) {
 307                 rew_scope(man, tok);
 308                 man->flags |= ROFF_NONOFILL;
 309         }
 310         roff_block_alloc(man, line, ppos, tok);
 311         head = roff_head_alloc(man, line, ppos, tok);
 312
 313         la = *pos;
 314         if (man_args(man, line, pos, buf, &p)) {
 315                 roff_word_alloc(man, line, la, p);
 316                 if (tok == MAN_RS) {
 317                         if (roff_getreg(man->roff, "an-margin") == 0)
 318                                 roff_setreg(man->roff, "an-margin",
 319                                     7 * 24, '=');
 320                         if ((head->aux = strtod(p, NULL) * 24.0) > 0)
 321                                 roff_setreg(man->roff, "an-margin",
 322                                     head->aux, '+');
 323                 }
 324                 free(p);
 325         }
 326
 327         if (buf[*pos] != '\0')
 328                 mandoc_msg(MANDOCERR_ARG_EXCESS, line, *pos,
 329                     "%s ... %s", roff_name[tok], buf + *pos);
 330
 331         man_unscope(man, head);
 332         roff_body_alloc(man, line, ppos, tok);
 333         man->flags &= ~ROFF_NONOFILL;
 334 }
 335
 336 /*
 337  * Parse an implicit-block macro.  These contain a ROFFT_HEAD and a
 338  * ROFFT_BODY contained within a ROFFT_BLOCK.  Rules for closing out other
 339  * scopes, such as `SH' closing out an `SS', are defined in the rew
 340  * routines.
 341  */
 342 void
 343 blk_imp(MACRO_PROT_ARGS)
 344 {
 345         int              la;
 346         char            *p;
 347         struct roff_node *n;
 348
 349         rew_scope(man, tok);
 350         man->flags |= ROFF_NONOFILL;
 351         if (tok == MAN_SH || tok == MAN_SS)
 352                 man->flags &= ~ROFF_NOFILL;
 353         roff_block_alloc(man, line, ppos, tok);
 354         n = roff_head_alloc(man, line, ppos, tok);
 355
 356         /* Add line arguments. */
 357
 358         for (;;) {
 359                 la = *pos;
 360                 if ( ! man_args(man, line, pos, buf, &p))
 361                         break;
 362                 roff_word_alloc(man, line, la, p);
 363                 free(p);
 364         }
 365
 366         /*
 367          * For macros having optional next-line scope,
 368          * keep the head open if there were no arguments.
 369          * For `TP' and `TQ', always keep the head open.
 370          */
 371
 372         if (man_macro(tok)->flags & MAN_BSCOPED &&
 373             (tok == MAN_TP || tok == MAN_TQ || n == man->last)) {
 374                 man->flags |= MAN_BLINE;
 375                 return;
 376         }
 377
 378         /* Close out the head and open the body. */
 379
 380         man_unscope(man, n);
 381         roff_body_alloc(man, line, ppos, tok);
 382         man->flags &= ~ROFF_NONOFILL;
 383 }
 384
 385 void
 386 in_line_eoln(MACRO_PROT_ARGS)
 387 {
 388         int              la;
 389         char            *p;
 390         struct roff_node *n;
 391
 392         roff_elem_alloc(man, line, ppos, tok);
 393         n = man->last;
 394
 395         if (tok == MAN_EX)
 396                 man->flags |= ROFF_NOFILL;
 397         else if (tok == MAN_EE)
 398                 man->flags &= ~ROFF_NOFILL;
 399
 400 #if DEBUG_MEMORY
 401         if (tok == MAN_TH)
 402                 mandoc_dbg_name(buf);
 403 #endif
 404
 405         for (;;) {
 406                 if (buf[*pos] != '\0' && man->last != n && tok == MAN_PD) {
 407                         mandoc_msg(MANDOCERR_ARG_EXCESS, line, *pos,
 408                             "%s ... %s", roff_name[tok], buf + *pos);
 409                         break;
 410                 }
 411                 la = *pos;
 412                 if ( ! man_args(man, line, pos, buf, &p))
 413                         break;
 414                 if (man_macro(tok)->flags & MAN_JOIN &&
 415                     man->last->type == ROFFT_TEXT)
 416                         roff_word_append(man, p);
 417                 else
 418                         roff_word_alloc(man, line, la, p);
 419                 free(p);
 420         }
 421
 422         /*
 423          * Append NODE_EOS in case the last snipped argument
 424          * ends with a dot, e.g. `.IR syslog (3).'
 425          */
 426
 427         if (n != man->last &&
 428             mandoc_eos(man->last->string, strlen(man->last->string)))
 429                 man->last->flags |= NODE_EOS;
 430
 431         /*
 432          * If no arguments are specified and this is MAN_ESCOPED (i.e.,
 433          * next-line scoped), then set our mode to indicate that we're
 434          * waiting for terms to load into our context.
 435          */
 436
 437         if (n == man->last && man_macro(tok)->flags & MAN_ESCOPED) {
 438                 man->flags |= MAN_ELINE;
 439                 return;
 440         }
 441
 442         assert(man->last->type != ROFFT_ROOT);
 443         man->next = ROFF_NEXT_SIBLING;
 444
 445         /* Rewind our element scope. */
 446
 447         for ( ; man->last; man->last = man->last->parent) {
 448                 man->last->flags |= NODE_VALID;
 449                 if (man->last == n)
 450                         break;
 451         }
 452
 453         /* Rewind next-line scoped ancestors, if any. */
 454
 455         if (man_macro(tok)->flags & MAN_ESCOPED)
 456                 man_descope(man, line, ppos, NULL);
 457 }
 458
 459 void
 460 man_endparse(struct roff_man *man)
 461 {
 462         man_unscope(man, man->meta.first);
 463 }
 464
 465 static int
 466 man_args(struct roff_man *man, int line, int *pos, char *buf, char **v)
 467 {
 468         char     *start;
 469
 470         assert(*pos);
 471         *v = start = buf + *pos;
 472         assert(' ' != *start);
 473
 474         if ('\0' == *start)
 475                 return 0;
 476
 477         *v = roff_getarg(man->roff, v, line, pos);
 478         return 1;
 479 }