man_macro.c

   1 /*      $Id: man_macro.c,v 1.85 2014/07/09 11:28:26 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 #ifdef HAVE_CONFIG_H
  20 #include "config.h"
  21 #endif
  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         man->next = MAN_NEXT_SIBLING;
 104         to = to->parent;
 105         n = man->last;
 106         while (n != to) {
 107                 if (NULL == to &&
 108                     MAN_BLOCK == n->type &&
 109                     0 == (MAN_VALID & n->flags) &&
 110                     MAN_EXPLICIT & man_macros[n->tok].flags)
 111                         mandoc_msg(MANDOCERR_BLK_NOEND,
 112                             man->parse, n->line, n->pos,
 113                             man_macronames[n->tok]);
 114                 /*
 115                  * We might delete the man->last node
 116                  * in the post-validation phase.
 117                  * Save a pointer to the parent such that
 118                  * we know where to continue the iteration.
 119                  */
 120                 man->last = n;
 121                 n = n->parent;
 122                 if ( ! man_valid_post(man))
 123                         return(0);
 124         }
 125         return(1);
 126 }
 127
 128 static enum rew
 129 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 130 {
 131
 132         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 133             MAN_BODY == n->parent->type)
 134                 return(REW_REWIND);
 135         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 136 }
 137
 138 /*
 139  * There are three scope levels: scoped to the root (all), scoped to the
 140  * section (all less sections), and scoped to subsections (all less
 141  * sections and subsections).
 142  */
 143 static enum rew
 144 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 145 {
 146         enum rew         c;
 147
 148         /* We cannot progress beyond the root ever. */
 149         if (MAN_ROOT == n->type)
 150                 return(REW_HALT);
 151
 152         assert(n->parent);
 153
 154         /* Normal nodes shouldn't go to the level of the root. */
 155         if (MAN_ROOT == n->parent->type)
 156                 return(REW_REWIND);
 157
 158         /* Already-validated nodes should be closed out. */
 159         if (MAN_VALID & n->flags)
 160                 return(REW_NOHALT);
 161
 162         /* First: rewind to ourselves. */
 163         if (type == n->type && tok == n->tok) {
 164                 if (MAN_EXPLICIT & man_macros[n->tok].flags)
 165                         return(REW_HALT);
 166                 else
 167                         return(REW_REWIND);
 168         }
 169
 170         /*
 171          * Next follow the implicit scope-smashings as defined by man.7:
 172          * section, sub-section, etc.
 173          */
 174
 175         switch (tok) {
 176         case MAN_SH:
 177                 break;
 178         case MAN_SS:
 179                 /* Rewind to a section, if a block. */
 180                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 181                         return(c);
 182                 break;
 183         case MAN_RS:
 184                 /* Preserve empty paragraphs before RS. */
 185                 if (0 == n->nchild && (MAN_P == n->tok ||
 186                     MAN_PP == n->tok || MAN_LP == n->tok))
 187                         return(REW_HALT);
 188                 /* Rewind to a subsection, if a block. */
 189                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 190                         return(c);
 191                 /* Rewind to a section, if a block. */
 192                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 193                         return(c);
 194                 break;
 195         default:
 196                 /* Rewind to an offsetter, if a block. */
 197                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 198                         return(c);
 199                 /* Rewind to a subsection, if a block. */
 200                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 201                         return(c);
 202                 /* Rewind to a section, if a block. */
 203                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 204                         return(c);
 205                 break;
 206         }
 207
 208         return(REW_NOHALT);
 209 }
 210
 211 /*
 212  * Rewinding entails ascending the parse tree until a coherent point,
 213  * for example, the `SH' macro will close out any intervening `SS'
 214  * scopes.  When a scope is closed, it must be validated and actioned.
 215  */
 216 static int
 217 rew_scope(enum man_type type, struct man *man, enum mant tok)
 218 {
 219         struct man_node *n;
 220         enum rew         c;
 221
 222         for (n = man->last; n; n = n->parent) {
 223                 /*
 224                  * Whether we should stop immediately (REW_HALT), stop
 225                  * and rewind until this point (REW_REWIND), or keep
 226                  * rewinding (REW_NOHALT).
 227                  */
 228                 c = rew_dohalt(tok, type, n);
 229                 if (REW_HALT == c)
 230                         return(1);
 231                 if (REW_REWIND == c)
 232                         break;
 233         }
 234
 235         /*
 236          * Rewind until the current point.  Warn if we're a roff
 237          * instruction that's mowing over explicit scopes.
 238          */
 239         assert(n);
 240
 241         return(man_unscope(man, n));
 242 }
 243
 244
 245 /*
 246  * Close out a generic explicit macro.
 247  */
 248 int
 249 blk_close(MACRO_PROT_ARGS)
 250 {
 251         enum mant                ntok;
 252         const struct man_node   *nn;
 253
 254         switch (tok) {
 255         case MAN_RE:
 256                 ntok = MAN_RS;
 257                 break;
 258         case MAN_UE:
 259                 ntok = MAN_UR;
 260                 break;
 261         default:
 262                 abort();
 263                 /* NOTREACHED */
 264         }
 265
 266         for (nn = man->last->parent; nn; nn = nn->parent)
 267                 if (ntok == nn->tok && MAN_BLOCK == nn->type)
 268                         break;
 269
 270         if (NULL == nn) {
 271                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 272                     line, ppos, man_macronames[tok]);
 273                 if ( ! rew_scope(MAN_BLOCK, man, MAN_PP))
 274                         return(0);
 275         } else
 276                 man_unscope(man, nn);
 277
 278         return(1);
 279 }
 280
 281 int
 282 blk_exp(MACRO_PROT_ARGS)
 283 {
 284         struct man_node *n;
 285         int              la;
 286         char            *p;
 287
 288         /* Close out prior implicit scopes. */
 289
 290         if ( ! rew_scope(MAN_BLOCK, man, tok))
 291                 return(0);
 292
 293         if ( ! man_block_alloc(man, line, ppos, tok))
 294                 return(0);
 295         if ( ! man_head_alloc(man, line, ppos, tok))
 296                 return(0);
 297
 298         for (;;) {
 299                 la = *pos;
 300                 if ( ! man_args(man, line, pos, buf, &p))
 301                         break;
 302                 if ( ! man_word_alloc(man, line, la, p))
 303                         return(0);
 304         }
 305
 306         assert(man);
 307         assert(tok != MAN_MAX);
 308
 309         for (n = man->last; n; n = n->parent) {
 310                 if (n->tok != tok)
 311                         continue;
 312                 assert(MAN_HEAD == n->type);
 313                 man_unscope(man, n);
 314                 break;
 315         }
 316
 317         return(man_body_alloc(man, line, ppos, tok));
 318 }
 319
 320 /*
 321  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 322  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 323  * scopes, such as `SH' closing out an `SS', are defined in the rew
 324  * routines.
 325  */
 326 int
 327 blk_imp(MACRO_PROT_ARGS)
 328 {
 329         int              la;
 330         char            *p;
 331         struct man_node *n;
 332
 333         /* Close out prior scopes. */
 334
 335         if ( ! rew_scope(MAN_BODY, man, tok))
 336                 return(0);
 337         if ( ! rew_scope(MAN_BLOCK, man, tok))
 338                 return(0);
 339
 340         /* Allocate new block & head scope. */
 341
 342         if ( ! man_block_alloc(man, line, ppos, tok))
 343                 return(0);
 344         if ( ! man_head_alloc(man, line, ppos, tok))
 345                 return(0);
 346
 347         n = man->last;
 348
 349         /* Add line arguments. */
 350
 351         for (;;) {
 352                 la = *pos;
 353                 if ( ! man_args(man, line, pos, buf, &p))
 354                         break;
 355                 if ( ! man_word_alloc(man, line, la, p))
 356                         return(0);
 357         }
 358
 359         /* Close out head and open body (unless MAN_SCOPE). */
 360
 361         if (MAN_SCOPED & man_macros[tok].flags) {
 362                 /* If we're forcing scope (`TP'), keep it open. */
 363                 if (MAN_FSCOPED & man_macros[tok].flags) {
 364                         man->flags |= MAN_BLINE;
 365                         return(1);
 366                 } else if (n == man->last) {
 367                         man->flags |= MAN_BLINE;
 368                         return(1);
 369                 }
 370         }
 371
 372         if ( ! rew_scope(MAN_HEAD, man, tok))
 373                 return(0);
 374         return(man_body_alloc(man, line, ppos, tok));
 375 }
 376
 377 int
 378 in_line_eoln(MACRO_PROT_ARGS)
 379 {
 380         int              la;
 381         char            *p;
 382         struct man_node *n;
 383
 384         if ( ! man_elem_alloc(man, line, ppos, tok))
 385                 return(0);
 386
 387         n = man->last;
 388
 389         for (;;) {
 390                 la = *pos;
 391                 if ( ! man_args(man, line, pos, buf, &p))
 392                         break;
 393                 if ( ! man_word_alloc(man, line, la, p))
 394                         return(0);
 395         }
 396
 397         /*
 398          * Append MAN_EOS in case the last snipped argument
 399          * ends with a dot, e.g. `.IR syslog (3).'
 400          */
 401
 402         if (n != man->last &&
 403             mandoc_eos(man->last->string, strlen(man->last->string)))
 404                 man->last->flags |= MAN_EOS;
 405
 406         /*
 407          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 408          * next-line scoped), then set our mode to indicate that we're
 409          * waiting for terms to load into our context.
 410          */
 411
 412         if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
 413                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 414                 man->flags |= MAN_ELINE;
 415                 return(1);
 416         }
 417
 418         /* Set ignorable context, if applicable. */
 419
 420         if (MAN_NSCOPED & man_macros[tok].flags) {
 421                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 422                 man->flags |= MAN_ILINE;
 423         }
 424
 425         assert(MAN_ROOT != man->last->type);
 426         man->next = MAN_NEXT_SIBLING;
 427
 428         /*
 429          * Rewind our element scope.  Note that when TH is pruned, we'll
 430          * be back at the root, so make sure that we don't clobber as
 431          * its sibling.
 432          */
 433
 434         for ( ; man->last; man->last = man->last->parent) {
 435                 if (man->last == n)
 436                         break;
 437                 if (man->last->type == MAN_ROOT)
 438                         break;
 439                 if ( ! man_valid_post(man))
 440                         return(0);
 441         }
 442
 443         assert(man->last);
 444
 445         /*
 446          * Same here regarding whether we're back at the root.
 447          */
 448
 449         if (man->last->type != MAN_ROOT && ! man_valid_post(man))
 450                 return(0);
 451
 452         return(1);
 453 }
 454
 455
 456 int
 457 man_macroend(struct man *man)
 458 {
 459
 460         return(man_unscope(man, man->first));
 461 }
 462
 463 static int
 464 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 465 {
 466         char     *start;
 467
 468         assert(*pos);
 469         *v = start = buf + *pos;
 470         assert(' ' != *start);
 471
 472         if ('\0' == *start)
 473                 return(0);
 474
 475         *v = mandoc_getarg(man->parse, v, line, pos);
 476         return(1);
 477 }