man_macro.c

   1 /*      $Id: man_macro.c,v 1.89 2014/08/18 16:36:54 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012, 2013 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  int              blk_close(MACRO_PROT_ARGS);
  40 static  int              blk_exp(MACRO_PROT_ARGS);
  41 static  int              blk_imp(MACRO_PROT_ARGS);
  42 static  int              in_line_eoln(MACRO_PROT_ARGS);
  43 static  int              man_args(struct man *, int,
  44                                 int *, char *, char **);
  45
  46 static  int              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 }, /* SM */
  65         { in_line_eoln, MAN_SCOPED }, /* 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 }, /* R */
  71         { in_line_eoln, MAN_SCOPED }, /* B */
  72         { in_line_eoln, MAN_SCOPED }, /* 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, 0 }, /* 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, 0 }, /* UE */
  92         { in_line_eoln, 0 }, /* ll */
  93 };
  94
  95 const   struct man_macro * const man_macros = __man_macros;
  96
  97
  98 int
  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                 if ( ! man_valid_post(man))
 145                         return(0);
 146         }
 147
 148         /*
 149          * If we ended up at the parent of the node we were
 150          * supposed to rewind to, that means the target node
 151          * got deleted, so add the next node we parse as a child
 152          * of the parent instead of as a sibling of the target.
 153          */
 154
 155         man->next = (man->last == to) ?
 156             MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 157
 158         return(1);
 159 }
 160
 161 static enum rew
 162 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 163 {
 164
 165         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 166             MAN_BODY == n->parent->type)
 167                 return(REW_REWIND);
 168         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 169 }
 170
 171 /*
 172  * There are three scope levels: scoped to the root (all), scoped to the
 173  * section (all less sections), and scoped to subsections (all less
 174  * sections and subsections).
 175  */
 176 static enum rew
 177 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 178 {
 179         enum rew         c;
 180
 181         /* We cannot progress beyond the root ever. */
 182         if (MAN_ROOT == n->type)
 183                 return(REW_HALT);
 184
 185         assert(n->parent);
 186
 187         /* Normal nodes shouldn't go to the level of the root. */
 188         if (MAN_ROOT == n->parent->type)
 189                 return(REW_REWIND);
 190
 191         /* Already-validated nodes should be closed out. */
 192         if (MAN_VALID & n->flags)
 193                 return(REW_NOHALT);
 194
 195         /* First: rewind to ourselves. */
 196         if (type == n->type && tok == n->tok) {
 197                 if (MAN_EXPLICIT & man_macros[n->tok].flags)
 198                         return(REW_HALT);
 199                 else
 200                         return(REW_REWIND);
 201         }
 202
 203         /*
 204          * Next follow the implicit scope-smashings as defined by man.7:
 205          * section, sub-section, etc.
 206          */
 207
 208         switch (tok) {
 209         case MAN_SH:
 210                 break;
 211         case MAN_SS:
 212                 /* Rewind to a section, if a block. */
 213                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 214                         return(c);
 215                 break;
 216         case MAN_RS:
 217                 /* Preserve empty paragraphs before RS. */
 218                 if (0 == n->nchild && (MAN_P == n->tok ||
 219                     MAN_PP == n->tok || MAN_LP == n->tok))
 220                         return(REW_HALT);
 221                 /* Rewind to a subsection, if a block. */
 222                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 223                         return(c);
 224                 /* Rewind to a section, if a block. */
 225                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 226                         return(c);
 227                 break;
 228         default:
 229                 /* Rewind to an offsetter, if a block. */
 230                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 231                         return(c);
 232                 /* Rewind to a subsection, if a block. */
 233                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 234                         return(c);
 235                 /* Rewind to a section, if a block. */
 236                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 237                         return(c);
 238                 break;
 239         }
 240
 241         return(REW_NOHALT);
 242 }
 243
 244 /*
 245  * Rewinding entails ascending the parse tree until a coherent point,
 246  * for example, the `SH' macro will close out any intervening `SS'
 247  * scopes.  When a scope is closed, it must be validated and actioned.
 248  */
 249 static int
 250 rew_scope(enum man_type type, struct man *man, enum mant tok)
 251 {
 252         struct man_node *n;
 253         enum rew         c;
 254
 255         for (n = man->last; n; n = n->parent) {
 256                 /*
 257                  * Whether we should stop immediately (REW_HALT), stop
 258                  * and rewind until this point (REW_REWIND), or keep
 259                  * rewinding (REW_NOHALT).
 260                  */
 261                 c = rew_dohalt(tok, type, n);
 262                 if (REW_HALT == c)
 263                         return(1);
 264                 if (REW_REWIND == c)
 265                         break;
 266         }
 267
 268         /*
 269          * Rewind until the current point.  Warn if we're a roff
 270          * instruction that's mowing over explicit scopes.
 271          */
 272         assert(n);
 273
 274         return(man_unscope(man, n));
 275 }
 276
 277
 278 /*
 279  * Close out a generic explicit macro.
 280  */
 281 int
 282 blk_close(MACRO_PROT_ARGS)
 283 {
 284         enum mant                ntok;
 285         const struct man_node   *nn;
 286
 287         switch (tok) {
 288         case MAN_RE:
 289                 ntok = MAN_RS;
 290                 break;
 291         case MAN_UE:
 292                 ntok = MAN_UR;
 293                 break;
 294         default:
 295                 abort();
 296                 /* NOTREACHED */
 297         }
 298
 299         for (nn = man->last->parent; nn; nn = nn->parent)
 300                 if (ntok == nn->tok && MAN_BLOCK == nn->type)
 301                         break;
 302
 303         if (NULL == nn) {
 304                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 305                     line, ppos, man_macronames[tok]);
 306                 if ( ! rew_scope(MAN_BLOCK, man, MAN_PP))
 307                         return(0);
 308         } else
 309                 man_unscope(man, nn);
 310
 311         return(1);
 312 }
 313
 314 int
 315 blk_exp(MACRO_PROT_ARGS)
 316 {
 317         struct man_node *n;
 318         int              la;
 319         char            *p;
 320
 321         /* Close out prior implicit scopes. */
 322
 323         if ( ! rew_scope(MAN_BLOCK, man, tok))
 324                 return(0);
 325
 326         if ( ! man_block_alloc(man, line, ppos, tok))
 327                 return(0);
 328         if ( ! man_head_alloc(man, line, ppos, tok))
 329                 return(0);
 330
 331         for (;;) {
 332                 la = *pos;
 333                 if ( ! man_args(man, line, pos, buf, &p))
 334                         break;
 335                 if ( ! man_word_alloc(man, line, la, p))
 336                         return(0);
 337         }
 338
 339         assert(man);
 340         assert(tok != MAN_MAX);
 341
 342         for (n = man->last; n; n = n->parent) {
 343                 if (n->tok != tok)
 344                         continue;
 345                 assert(MAN_HEAD == n->type);
 346                 man_unscope(man, n);
 347                 break;
 348         }
 349
 350         return(man_body_alloc(man, line, ppos, tok));
 351 }
 352
 353 /*
 354  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 355  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 356  * scopes, such as `SH' closing out an `SS', are defined in the rew
 357  * routines.
 358  */
 359 int
 360 blk_imp(MACRO_PROT_ARGS)
 361 {
 362         int              la;
 363         char            *p;
 364         struct man_node *n;
 365
 366         /* Close out prior scopes. */
 367
 368         if ( ! rew_scope(MAN_BODY, man, tok))
 369                 return(0);
 370         if ( ! rew_scope(MAN_BLOCK, man, tok))
 371                 return(0);
 372
 373         /* Allocate new block & head scope. */
 374
 375         if ( ! man_block_alloc(man, line, ppos, tok))
 376                 return(0);
 377         if ( ! man_head_alloc(man, line, ppos, tok))
 378                 return(0);
 379
 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                 if ( ! man_word_alloc(man, line, la, p))
 389                         return(0);
 390         }
 391
 392         /* Close out head and open body (unless MAN_SCOPE). */
 393
 394         if (MAN_SCOPED & man_macros[tok].flags) {
 395                 /* If we're forcing scope (`TP'), keep it open. */
 396                 if (MAN_FSCOPED & man_macros[tok].flags) {
 397                         man->flags |= MAN_BLINE;
 398                         return(1);
 399                 } else if (n == man->last) {
 400                         man->flags |= MAN_BLINE;
 401                         return(1);
 402                 }
 403         }
 404
 405         if ( ! rew_scope(MAN_HEAD, man, tok))
 406                 return(0);
 407         return(man_body_alloc(man, line, ppos, tok));
 408 }
 409
 410 int
 411 in_line_eoln(MACRO_PROT_ARGS)
 412 {
 413         int              la;
 414         char            *p;
 415         struct man_node *n;
 416
 417         if ( ! man_elem_alloc(man, line, ppos, tok))
 418                 return(0);
 419
 420         n = man->last;
 421
 422         for (;;) {
 423                 la = *pos;
 424                 if ( ! man_args(man, line, pos, buf, &p))
 425                         break;
 426                 if ( ! man_word_alloc(man, line, la, p))
 427                         return(0);
 428         }
 429
 430         /*
 431          * Append MAN_EOS in case the last snipped argument
 432          * ends with a dot, e.g. `.IR syslog (3).'
 433          */
 434
 435         if (n != man->last &&
 436             mandoc_eos(man->last->string, strlen(man->last->string)))
 437                 man->last->flags |= MAN_EOS;
 438
 439         /*
 440          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 441          * next-line scoped), then set our mode to indicate that we're
 442          * waiting for terms to load into our context.
 443          */
 444
 445         if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
 446                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 447                 man->flags |= MAN_ELINE;
 448                 return(1);
 449         }
 450
 451         assert(MAN_ROOT != man->last->type);
 452         man->next = MAN_NEXT_SIBLING;
 453
 454         /*
 455          * Rewind our element scope.  Note that when TH is pruned, we'll
 456          * be back at the root, so make sure that we don't clobber as
 457          * its sibling.
 458          */
 459
 460         for ( ; man->last; man->last = man->last->parent) {
 461                 if (man->last == n)
 462                         break;
 463                 if (man->last->type == MAN_ROOT)
 464                         break;
 465                 if ( ! man_valid_post(man))
 466                         return(0);
 467         }
 468
 469         assert(man->last);
 470
 471         /*
 472          * Same here regarding whether we're back at the root.
 473          */
 474
 475         if (man->last->type != MAN_ROOT && ! man_valid_post(man))
 476                 return(0);
 477
 478         return(1);
 479 }
 480
 481
 482 int
 483 man_macroend(struct man *man)
 484 {
 485
 486         return(man_unscope(man, man->first));
 487 }
 488
 489 static int
 490 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 491 {
 492         char     *start;
 493
 494         assert(*pos);
 495         *v = start = buf + *pos;
 496         assert(' ' != *start);
 497
 498         if ('\0' == *start)
 499                 return(0);
 500
 501         *v = mandoc_getarg(man->parse, v, line, pos);
 502         return(1);
 503 }