man_macro.c

   1 /*      $Id: man_macro.c,v 1.92 2014/12/16 17:26:00 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012, 2013, 2014 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 AUTHOR DISCLAIMS ALL WARRANTIES
  12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR 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 "man.h"
  29 #include "mandoc.h"
  30 #include "libmandoc.h"
  31 #include "libman.h"
  32
  33 enum    rew {
  34         REW_REWIND,
  35         REW_NOHALT,
  36         REW_HALT
  37 };
  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 man *, int,
  44                                 int *, char *, char **);
  45
  46 static  void             rew_scope(enum man_type,
  47                                 struct man *, enum mant);
  48 static  enum rew         rew_dohalt(enum mant, enum man_type,
  49                                 const struct man_node *);
  50 static  enum rew         rew_block(enum mant, enum man_type,
  51                                 const struct man_node *);
  52
  53 const   struct man_macro __man_macros[MAN_MAX] = {
  54         { in_line_eoln, MAN_NSCOPED }, /* br */
  55         { in_line_eoln, MAN_BSCOPE }, /* TH */
  56         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
  57         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
  58         { blk_imp, MAN_BSCOPE | MAN_SCOPED | MAN_FSCOPED }, /* TP */
  59         { blk_imp, MAN_BSCOPE }, /* LP */
  60         { blk_imp, MAN_BSCOPE }, /* PP */
  61         { blk_imp, MAN_BSCOPE }, /* P */
  62         { blk_imp, MAN_BSCOPE }, /* IP */
  63         { blk_imp, MAN_BSCOPE }, /* HP */
  64         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SM */
  65         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SB */
  66         { in_line_eoln, 0 }, /* BI */
  67         { in_line_eoln, 0 }, /* IB */
  68         { in_line_eoln, 0 }, /* BR */
  69         { in_line_eoln, 0 }, /* RB */
  70         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* R */
  71         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* B */
  72         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* I */
  73         { in_line_eoln, 0 }, /* IR */
  74         { in_line_eoln, 0 }, /* RI */
  75         { in_line_eoln, MAN_NSCOPED }, /* na */
  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 | MAN_EXPLICIT }, /* RS */
  81         { in_line_eoln, 0 }, /* DT */
  82         { in_line_eoln, 0 }, /* UC */
  83         { in_line_eoln, 0 }, /* 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 | MAN_EXPLICIT }, /* 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 == MAN_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 == MAN_BLOCK &&
 129                             man_macros[n->tok].flags & MAN_EXPLICIT)
 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 man_type type, const struct man_node *n)
 160 {
 161
 162         if (type == MAN_BLOCK && ntok == n->parent->tok &&
 163             n->parent->type == MAN_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 man_type type, const struct man_node *n)
 175 {
 176         enum rew         c;
 177
 178         /* We cannot progress beyond the root ever. */
 179         if (MAN_ROOT == n->type)
 180                 return(REW_HALT);
 181
 182         assert(n->parent);
 183
 184         /* Normal nodes shouldn't go to the level of the root. */
 185         if (MAN_ROOT == n->parent->type)
 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_EXPLICIT & man_macros[n->tok].flags)
 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 man_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
 283         switch (tok) {
 284         case MAN_RE:
 285                 ntok = MAN_RS;
 286                 break;
 287         case MAN_UE:
 288                 ntok = MAN_UR;
 289                 break;
 290         default:
 291                 abort();
 292                 /* NOTREACHED */
 293         }
 294
 295         for (nn = man->last->parent; nn; nn = nn->parent)
 296                 if (nn->tok == ntok && nn->type == MAN_BLOCK)
 297                         break;
 298
 299         if (nn == NULL) {
 300                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 301                     line, ppos, man_macronames[tok]);
 302                 rew_scope(MAN_BLOCK, man, MAN_PP);
 303         } else
 304                 man_unscope(man, nn);
 305 }
 306
 307 void
 308 blk_exp(MACRO_PROT_ARGS)
 309 {
 310         struct man_node *n;
 311         int              la;
 312         char            *p;
 313
 314         rew_scope(MAN_BLOCK, man, tok);
 315         man_block_alloc(man, line, ppos, tok);
 316         man_head_alloc(man, line, ppos, tok);
 317
 318         for (;;) {
 319                 la = *pos;
 320                 if ( ! man_args(man, line, pos, buf, &p))
 321                         break;
 322                 man_word_alloc(man, line, la, p);
 323         }
 324
 325         assert(man);
 326         assert(tok != MAN_MAX);
 327
 328         for (n = man->last; n; n = n->parent)
 329                 if (n->tok == tok) {
 330                         assert(n->type == MAN_HEAD);
 331                         man_unscope(man, n);
 332                         break;
 333                 }
 334
 335         man_body_alloc(man, line, ppos, tok);
 336 }
 337
 338 /*
 339  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 340  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 341  * scopes, such as `SH' closing out an `SS', are defined in the rew
 342  * routines.
 343  */
 344 void
 345 blk_imp(MACRO_PROT_ARGS)
 346 {
 347         int              la;
 348         char            *p;
 349         struct man_node *n;
 350
 351         rew_scope(MAN_BODY, man, tok);
 352         rew_scope(MAN_BLOCK, man, tok);
 353         man_block_alloc(man, line, ppos, tok);
 354         man_head_alloc(man, line, ppos, tok);
 355         n = man->last;
 356
 357         /* Add line arguments. */
 358
 359         for (;;) {
 360                 la = *pos;
 361                 if ( ! man_args(man, line, pos, buf, &p))
 362                         break;
 363                 man_word_alloc(man, line, la, p);
 364         }
 365
 366         /* Close out head and open body (unless MAN_SCOPE). */
 367
 368         if (man_macros[tok].flags & MAN_SCOPED) {
 369                 /* If we're forcing scope (`TP'), keep it open. */
 370                 if (man_macros[tok].flags & MAN_FSCOPED) {
 371                         man->flags |= MAN_BLINE;
 372                         return;
 373                 } else if (n == man->last) {
 374                         man->flags |= MAN_BLINE;
 375                         return;
 376                 }
 377         }
 378         rew_scope(MAN_HEAD, man, tok);
 379         man_body_alloc(man, line, ppos, tok);
 380 }
 381
 382 void
 383 in_line_eoln(MACRO_PROT_ARGS)
 384 {
 385         int              la;
 386         char            *p;
 387         struct man_node *n;
 388
 389         man_elem_alloc(man, line, ppos, tok);
 390         n = man->last;
 391
 392         for (;;) {
 393                 la = *pos;
 394                 if ( ! man_args(man, line, pos, buf, &p))
 395                         break;
 396                 if (man_macros[tok].flags & MAN_JOIN &&
 397                     man->last->type == MAN_TEXT)
 398                         man_word_append(man, p);
 399                 else
 400                         man_word_alloc(man, line, la, p);
 401         }
 402
 403         /*
 404          * Append MAN_EOS in case the last snipped argument
 405          * ends with a dot, e.g. `.IR syslog (3).'
 406          */
 407
 408         if (n != man->last &&
 409             mandoc_eos(man->last->string, strlen(man->last->string)))
 410                 man->last->flags |= MAN_EOS;
 411
 412         /*
 413          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 414          * next-line scoped), then set our mode to indicate that we're
 415          * waiting for terms to load into our context.
 416          */
 417
 418         if (n == man->last && man_macros[tok].flags & MAN_SCOPED) {
 419                 assert( ! (man_macros[tok].flags & MAN_NSCOPED));
 420                 man->flags |= MAN_ELINE;
 421                 return;
 422         }
 423
 424         assert(man->last->type != MAN_ROOT);
 425         man->next = MAN_NEXT_SIBLING;
 426
 427         /*
 428          * Rewind our element scope.  Note that when TH is pruned, we'll
 429          * be back at the root, so make sure that we don't clobber as
 430          * its sibling.
 431          */
 432
 433         for ( ; man->last; man->last = man->last->parent) {
 434                 if (man->last == n)
 435                         break;
 436                 if (man->last->type == MAN_ROOT)
 437                         break;
 438                 man_valid_post(man);
 439         }
 440
 441         assert(man->last);
 442
 443         /*
 444          * Same here regarding whether we're back at the root.
 445          */
 446
 447         if (man->last->type != MAN_ROOT)
 448                 man_valid_post(man);
 449 }
 450
 451
 452 void
 453 man_macroend(struct man *man)
 454 {
 455
 456         man_unscope(man, man->first);
 457 }
 458
 459 static int
 460 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 461 {
 462         char     *start;
 463
 464         assert(*pos);
 465         *v = start = buf + *pos;
 466         assert(' ' != *start);
 467
 468         if ('\0' == *start)
 469                 return(0);
 470
 471         *v = mandoc_getarg(man->parse, v, line, pos);
 472         return(1);
 473 }