man_macro.c

   1 /*      $Id: man_macro.c,v 1.101 2015/04/02 21:36:50 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012, 2013, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org>
   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 <stdlib.h>
  26 #include <string.h>
  27
  28 #include "mandoc.h"
  29 #include "roff.h"
  30 #include "man.h"
  31 #include "libmandoc.h"
  32 #include "libman.h"
  33
  34 enum    rew {
  35         REW_REWIND,
  36         REW_NOHALT,
  37         REW_HALT
  38 };
  39
  40 static  void             blk_close(MACRO_PROT_ARGS);
  41 static  void             blk_exp(MACRO_PROT_ARGS);
  42 static  void             blk_imp(MACRO_PROT_ARGS);
  43 static  void             in_line_eoln(MACRO_PROT_ARGS);
  44 static  int              man_args(struct man *, int,
  45                                 int *, char *, char **);
  46
  47 static  void             rew_scope(enum roff_type,
  48                                 struct man *, enum mant);
  49 static  enum rew         rew_dohalt(enum mant, enum roff_type,
  50                                 const struct man_node *);
  51 static  enum rew         rew_block(enum mant, enum roff_type,
  52                                 const struct man_node *);
  53
  54 const   struct man_macro __man_macros[MAN_MAX] = {
  55         { in_line_eoln, MAN_NSCOPED }, /* br */
  56         { in_line_eoln, MAN_BSCOPE }, /* TH */
  57         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
  58         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
  59         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* TP */
  60         { blk_imp, MAN_BSCOPE }, /* LP */
  61         { blk_imp, MAN_BSCOPE }, /* PP */
  62         { blk_imp, MAN_BSCOPE }, /* P */
  63         { blk_imp, MAN_BSCOPE }, /* IP */
  64         { blk_imp, MAN_BSCOPE }, /* HP */
  65         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SM */
  66         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SB */
  67         { in_line_eoln, 0 }, /* BI */
  68         { in_line_eoln, 0 }, /* IB */
  69         { in_line_eoln, 0 }, /* BR */
  70         { in_line_eoln, 0 }, /* RB */
  71         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* R */
  72         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* B */
  73         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* I */
  74         { in_line_eoln, 0 }, /* IR */
  75         { in_line_eoln, 0 }, /* RI */
  76         { in_line_eoln, MAN_NSCOPED }, /* sp */
  77         { in_line_eoln, MAN_BSCOPE }, /* nf */
  78         { in_line_eoln, MAN_BSCOPE }, /* fi */
  79         { blk_close, MAN_BSCOPE }, /* RE */
  80         { blk_exp, MAN_BSCOPE }, /* RS */
  81         { in_line_eoln, 0 }, /* DT */
  82         { in_line_eoln, 0 }, /* UC */
  83         { in_line_eoln, MAN_NSCOPED }, /* PD */
  84         { in_line_eoln, 0 }, /* AT */
  85         { in_line_eoln, 0 }, /* in */
  86         { in_line_eoln, 0 }, /* ft */
  87         { in_line_eoln, 0 }, /* OP */
  88         { in_line_eoln, MAN_BSCOPE }, /* EX */
  89         { in_line_eoln, MAN_BSCOPE }, /* EE */
  90         { blk_exp, MAN_BSCOPE }, /* UR */
  91         { blk_close, MAN_BSCOPE }, /* UE */
  92         { in_line_eoln, 0 }, /* ll */
  93 };
  94
  95 const   struct man_macro * const man_macros = __man_macros;
  96
  97
  98 void
  99 man_unscope(struct man *man, const struct man_node *to)
 100 {
 101         struct man_node *n;
 102
 103         to = to->parent;
 104         n = man->last;
 105         while (n != to) {
 106
 107                 /* Reached the end of the document? */
 108
 109                 if (to == NULL && ! (n->flags & MAN_VALID)) {
 110                         if (man->flags & (MAN_BLINE | MAN_ELINE) &&
 111                             man_macros[n->tok].flags & MAN_SCOPED) {
 112                                 mandoc_vmsg(MANDOCERR_BLK_LINE,
 113                                     man->parse, n->line, n->pos,
 114                                     "EOF breaks %s",
 115                                     man_macronames[n->tok]);
 116                                 if (man->flags & MAN_ELINE)
 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                                 man_node_delete(man, man->last);
 126                                 continue;
 127                         }
 128                         if (n->type == ROFFT_BLOCK &&
 129                             man_macros[n->tok].fp == blk_exp)
 130                                 mandoc_msg(MANDOCERR_BLK_NOEND,
 131                                     man->parse, n->line, n->pos,
 132                                     man_macronames[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_valid_post(man);
 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             MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 156 }
 157
 158 static enum rew
 159 rew_block(enum mant ntok, enum roff_type type, const struct man_node *n)
 160 {
 161
 162         if (type == ROFFT_BLOCK && n->parent->tok == ntok &&
 163             n->parent->type == ROFFT_BODY)
 164                 return(REW_REWIND);
 165         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 166 }
 167
 168 /*
 169  * There are three scope levels: scoped to the root (all), scoped to the
 170  * section (all less sections), and scoped to subsections (all less
 171  * sections and subsections).
 172  */
 173 static enum rew
 174 rew_dohalt(enum mant tok, enum roff_type type, const struct man_node *n)
 175 {
 176         enum rew         c;
 177
 178         /* We cannot progress beyond the root ever. */
 179         if (n->type == ROFFT_ROOT)
 180                 return(REW_HALT);
 181
 182         assert(n->parent);
 183
 184         /* Normal nodes shouldn't go to the level of the root. */
 185         if (n->parent->type == ROFFT_ROOT)
 186                 return(REW_REWIND);
 187
 188         /* Already-validated nodes should be closed out. */
 189         if (MAN_VALID & n->flags)
 190                 return(REW_NOHALT);
 191
 192         /* First: rewind to ourselves. */
 193         if (type == n->type && tok == n->tok) {
 194                 if (man_macros[n->tok].fp == blk_exp)
 195                         return(REW_HALT);
 196                 else
 197                         return(REW_REWIND);
 198         }
 199
 200         /*
 201          * Next follow the implicit scope-smashings as defined by man.7:
 202          * section, sub-section, etc.
 203          */
 204
 205         switch (tok) {
 206         case MAN_SH:
 207                 break;
 208         case MAN_SS:
 209                 /* Rewind to a section, if a block. */
 210                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 211                         return(c);
 212                 break;
 213         case MAN_RS:
 214                 /* Preserve empty paragraphs before RS. */
 215                 if (0 == n->nchild && (MAN_P == n->tok ||
 216                     MAN_PP == n->tok || MAN_LP == n->tok))
 217                         return(REW_HALT);
 218                 /* Rewind to a subsection, if a block. */
 219                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 220                         return(c);
 221                 /* Rewind to a section, if a block. */
 222                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 223                         return(c);
 224                 break;
 225         default:
 226                 /* Rewind to an offsetter, if a block. */
 227                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 228                         return(c);
 229                 /* Rewind to a subsection, if a block. */
 230                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 231                         return(c);
 232                 /* Rewind to a section, if a block. */
 233                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 234                         return(c);
 235                 break;
 236         }
 237
 238         return(REW_NOHALT);
 239 }
 240
 241 /*
 242  * Rewinding entails ascending the parse tree until a coherent point,
 243  * for example, the `SH' macro will close out any intervening `SS'
 244  * scopes.  When a scope is closed, it must be validated and actioned.
 245  */
 246 static void
 247 rew_scope(enum roff_type type, struct man *man, enum mant tok)
 248 {
 249         struct man_node *n;
 250         enum rew         c;
 251
 252         for (n = man->last; n; n = n->parent) {
 253                 /*
 254                  * Whether we should stop immediately (REW_HALT), stop
 255                  * and rewind until this point (REW_REWIND), or keep
 256                  * rewinding (REW_NOHALT).
 257                  */
 258                 c = rew_dohalt(tok, type, n);
 259                 if (REW_HALT == c)
 260                         return;
 261                 if (REW_REWIND == c)
 262                         break;
 263         }
 264
 265         /*
 266          * Rewind until the current point.  Warn if we're a roff
 267          * instruction that's mowing over explicit scopes.
 268          */
 269
 270         man_unscope(man, n);
 271 }
 272
 273
 274 /*
 275  * Close out a generic explicit macro.
 276  */
 277 void
 278 blk_close(MACRO_PROT_ARGS)
 279 {
 280         enum mant                ntok;
 281         const struct man_node   *nn;
 282         char                    *p;
 283         int                      nrew, target;
 284
 285         nrew = 1;
 286         switch (tok) {
 287         case MAN_RE:
 288                 ntok = MAN_RS;
 289                 if ( ! man_args(man, line, pos, buf, &p))
 290                         break;
 291                 for (nn = man->last->parent; nn; nn = nn->parent)
 292                         if (nn->tok == ntok && nn->type == ROFFT_BLOCK)
 293                                 nrew++;
 294                 target = strtol(p, &p, 10);
 295                 if (*p != '\0')
 296                         mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 297                             line, p - buf, "RE ... %s", p);
 298                 if (target == 0)
 299                         target = 1;
 300                 nrew -= target;
 301                 if (nrew < 1) {
 302                         mandoc_vmsg(MANDOCERR_RE_NOTOPEN, man->parse,
 303                             line, ppos, "RE %d", target);
 304                         return;
 305                 }
 306                 break;
 307         case MAN_UE:
 308                 ntok = MAN_UR;
 309                 break;
 310         default:
 311                 abort();
 312                 /* NOTREACHED */
 313         }
 314
 315         for (nn = man->last->parent; nn; nn = nn->parent)
 316                 if (nn->tok == ntok && nn->type == ROFFT_BLOCK && ! --nrew)
 317                         break;
 318
 319         if (nn == NULL) {
 320                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 321                     line, ppos, man_macronames[tok]);
 322                 rew_scope(ROFFT_BLOCK, man, MAN_PP);
 323         } else {
 324                 line = man->last->line;
 325                 ppos = man->last->pos;
 326                 ntok = man->last->tok;
 327                 man_unscope(man, nn);
 328
 329                 /* Move a trailing paragraph behind the block. */
 330
 331                 if (ntok == MAN_LP || ntok == MAN_PP || ntok == MAN_P) {
 332                         *pos = strlen(buf);
 333                         blk_imp(man, ntok, line, ppos, pos, buf);
 334                 }
 335         }
 336 }
 337
 338 void
 339 blk_exp(MACRO_PROT_ARGS)
 340 {
 341         struct man_node *head;
 342         char            *p;
 343         int              la;
 344
 345         rew_scope(ROFFT_BLOCK, man, tok);
 346         man_block_alloc(man, line, ppos, tok);
 347         man_head_alloc(man, line, ppos, tok);
 348         head = man->last;
 349
 350         la = *pos;
 351         if (man_args(man, line, pos, buf, &p))
 352                 man_word_alloc(man, line, la, p);
 353
 354         if (buf[*pos] != '\0')
 355                 mandoc_vmsg(MANDOCERR_ARG_EXCESS,
 356                     man->parse, line, *pos, "%s ... %s",
 357                     man_macronames[tok], buf + *pos);
 358
 359         man_unscope(man, head);
 360         man_body_alloc(man, line, ppos, tok);
 361 }
 362
 363 /*
 364  * Parse an implicit-block macro.  These contain a ROFFT_HEAD and a
 365  * ROFFT_BODY contained within a ROFFT_BLOCK.  Rules for closing out other
 366  * scopes, such as `SH' closing out an `SS', are defined in the rew
 367  * routines.
 368  */
 369 void
 370 blk_imp(MACRO_PROT_ARGS)
 371 {
 372         int              la;
 373         char            *p;
 374         struct man_node *n;
 375
 376         rew_scope(ROFFT_BODY, man, tok);
 377         rew_scope(ROFFT_BLOCK, man, tok);
 378         man_block_alloc(man, line, ppos, tok);
 379         man_head_alloc(man, line, ppos, tok);
 380         n = man->last;
 381
 382         /* Add line arguments. */
 383
 384         for (;;) {
 385                 la = *pos;
 386                 if ( ! man_args(man, line, pos, buf, &p))
 387                         break;
 388                 man_word_alloc(man, line, la, p);
 389         }
 390
 391         /*
 392          * For macros having optional next-line scope,
 393          * keep the head open if there were no arguments.
 394          * For `TP', always keep the head open.
 395          */
 396
 397         if (man_macros[tok].flags & MAN_SCOPED &&
 398             (tok == MAN_TP || n == man->last)) {
 399                 man->flags |= MAN_BLINE;
 400                 return;
 401         }
 402
 403         /* Close out the head and open the body. */
 404
 405         rew_scope(ROFFT_HEAD, man, tok);
 406         man_body_alloc(man, line, ppos, tok);
 407 }
 408
 409 void
 410 in_line_eoln(MACRO_PROT_ARGS)
 411 {
 412         int              la;
 413         char            *p;
 414         struct man_node *n;
 415
 416         man_elem_alloc(man, line, ppos, tok);
 417         n = man->last;
 418
 419         for (;;) {
 420                 if (buf[*pos] != '\0' && (tok == MAN_br ||
 421                     tok == MAN_fi || tok == MAN_nf)) {
 422                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 423                             man->parse, line, *pos, "%s %s",
 424                             man_macronames[tok], buf + *pos);
 425                         break;
 426                 }
 427                 if (buf[*pos] != '\0' && man->last != n &&
 428                     (tok == MAN_PD || tok == MAN_ft || tok == MAN_sp)) {
 429                         mandoc_vmsg(MANDOCERR_ARG_EXCESS,
 430                             man->parse, line, *pos, "%s ... %s",
 431                             man_macronames[tok], buf + *pos);
 432                         break;
 433                 }
 434                 la = *pos;
 435                 if ( ! man_args(man, line, pos, buf, &p))
 436                         break;
 437                 if (man_macros[tok].flags & MAN_JOIN &&
 438                     man->last->type == ROFFT_TEXT)
 439                         man_word_append(man, p);
 440                 else
 441                         man_word_alloc(man, line, la, p);
 442         }
 443
 444         /*
 445          * Append MAN_EOS in case the last snipped argument
 446          * ends with a dot, e.g. `.IR syslog (3).'
 447          */
 448
 449         if (n != man->last &&
 450             mandoc_eos(man->last->string, strlen(man->last->string)))
 451                 man->last->flags |= MAN_EOS;
 452
 453         /*
 454          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 455          * next-line scoped), then set our mode to indicate that we're
 456          * waiting for terms to load into our context.
 457          */
 458
 459         if (n == man->last && man_macros[tok].flags & MAN_SCOPED) {
 460                 assert( ! (man_macros[tok].flags & MAN_NSCOPED));
 461                 man->flags |= MAN_ELINE;
 462                 return;
 463         }
 464
 465         assert(man->last->type != ROFFT_ROOT);
 466         man->next = MAN_NEXT_SIBLING;
 467
 468         /*
 469          * Rewind our element scope.  Note that when TH is pruned, we'll
 470          * be back at the root, so make sure that we don't clobber as
 471          * its sibling.
 472          */
 473
 474         for ( ; man->last; man->last = man->last->parent) {
 475                 if (man->last == n)
 476                         break;
 477                 if (man->last->type == ROFFT_ROOT)
 478                         break;
 479                 man_valid_post(man);
 480         }
 481
 482         assert(man->last);
 483
 484         /*
 485          * Same here regarding whether we're back at the root.
 486          */
 487
 488         if (man->last->type != ROFFT_ROOT)
 489                 man_valid_post(man);
 490 }
 491
 492
 493 void
 494 man_macroend(struct man *man)
 495 {
 496
 497         man_unscope(man, man->first);
 498 }
 499
 500 static int
 501 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 502 {
 503         char     *start;
 504
 505         assert(*pos);
 506         *v = start = buf + *pos;
 507         assert(' ' != *start);
 508
 509         if ('\0' == *start)
 510                 return(0);
 511
 512         *v = mandoc_getarg(man->parse, v, line, pos);
 513         return(1);
 514 }