man_macro.c

   1 /*      $Id: man_macro.c,v 1.78 2013/12/22 13:25:17 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  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <assert.h>
  23 #include <ctype.h>
  24 #include <stdlib.h>
  25 #include <string.h>
  26
  27 #include "man.h"
  28 #include "mandoc.h"
  29 #include "libmandoc.h"
  30 #include "libman.h"
  31
  32 enum    rew {
  33         REW_REWIND,
  34         REW_NOHALT,
  35         REW_HALT
  36 };
  37
  38 static  int              blk_close(MACRO_PROT_ARGS);
  39 static  int              blk_exp(MACRO_PROT_ARGS);
  40 static  int              blk_imp(MACRO_PROT_ARGS);
  41 static  int              in_line_eoln(MACRO_PROT_ARGS);
  42 static  int              man_args(struct man *, int,
  43                                 int *, char *, char **);
  44
  45 static  int              rew_scope(enum man_type,
  46                                 struct man *, enum mant);
  47 static  enum rew         rew_dohalt(enum mant, enum man_type,
  48                                 const struct man_node *);
  49 static  enum rew         rew_block(enum mant, enum man_type,
  50                                 const struct man_node *);
  51 static  void             rew_warn(struct man *,
  52                                 struct man_node *, enum mandocerr);
  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 | MAN_FSCOPED }, /* 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 }, /* SM */
  66         { in_line_eoln, MAN_SCOPED }, /* 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 }, /* R */
  72         { in_line_eoln, MAN_SCOPED }, /* B */
  73         { in_line_eoln, MAN_SCOPED }, /* I */
  74         { in_line_eoln, 0 }, /* IR */
  75         { in_line_eoln, 0 }, /* RI */
  76         { in_line_eoln, MAN_NSCOPED }, /* na */
  77         { in_line_eoln, MAN_NSCOPED }, /* sp */
  78         { in_line_eoln, MAN_BSCOPE }, /* nf */
  79         { in_line_eoln, MAN_BSCOPE }, /* fi */
  80         { blk_close, 0 }, /* RE */
  81         { blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* RS */
  82         { in_line_eoln, 0 }, /* DT */
  83         { in_line_eoln, 0 }, /* UC */
  84         { in_line_eoln, 0 }, /* PD */
  85         { in_line_eoln, 0 }, /* AT */
  86         { in_line_eoln, 0 }, /* in */
  87         { in_line_eoln, 0 }, /* ft */
  88         { in_line_eoln, 0 }, /* OP */
  89         { in_line_eoln, MAN_BSCOPE }, /* EX */
  90         { in_line_eoln, MAN_BSCOPE }, /* EE */
  91         { blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* UR */
  92         { blk_close, 0 }, /* UE */
  93 };
  94
  95 const   struct man_macro * const man_macros = __man_macros;
  96
  97
  98 /*
  99  * Warn when "n" is an explicit non-roff macro.
 100  */
 101 static void
 102 rew_warn(struct man *man, struct man_node *n, enum mandocerr er)
 103 {
 104
 105         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
 106                 return;
 107         if (MAN_VALID & n->flags)
 108                 return;
 109         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 110                 return;
 111
 112         assert(er < MANDOCERR_FATAL);
 113         man_nmsg(man, n, er);
 114 }
 115
 116
 117 /*
 118  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 119  * will be used if an explicit block scope is being closed out.
 120  */
 121 int
 122 man_unscope(struct man *man, const struct man_node *to,
 123                 enum mandocerr er)
 124 {
 125         struct man_node *n;
 126
 127         assert(to);
 128
 129         man->next = MAN_NEXT_SIBLING;
 130
 131         /* LINTED */
 132         while (man->last != to) {
 133                 /*
 134                  * Save the parent here, because we may delete the
 135                  * man->last node in the post-validation phase and reset
 136                  * it to man->last->parent, causing a step in the closing
 137                  * out to be lost.
 138                  */
 139                 n = man->last->parent;
 140                 rew_warn(man, man->last, er);
 141                 if ( ! man_valid_post(man))
 142                         return(0);
 143                 man->last = n;
 144                 assert(man->last);
 145         }
 146
 147         rew_warn(man, man->last, er);
 148         if ( ! man_valid_post(man))
 149                 return(0);
 150
 151         return(1);
 152 }
 153
 154
 155 static enum rew
 156 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 157 {
 158
 159         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 160                         MAN_BODY == n->parent->type)
 161                 return(REW_REWIND);
 162         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 163 }
 164
 165
 166 /*
 167  * There are three scope levels: scoped to the root (all), scoped to the
 168  * section (all less sections), and scoped to subsections (all less
 169  * sections and subsections).
 170  */
 171 static enum rew
 172 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 173 {
 174         enum rew         c;
 175
 176         /* We cannot progress beyond the root ever. */
 177         if (MAN_ROOT == n->type)
 178                 return(REW_HALT);
 179
 180         assert(n->parent);
 181
 182         /* Normal nodes shouldn't go to the level of the root. */
 183         if (MAN_ROOT == n->parent->type)
 184                 return(REW_REWIND);
 185
 186         /* Already-validated nodes should be closed out. */
 187         if (MAN_VALID & n->flags)
 188                 return(REW_NOHALT);
 189
 190         /* First: rewind to ourselves. */
 191         if (type == n->type && tok == n->tok) {
 192                 if (MAN_EXPLICIT & man_macros[n->tok].flags)
 193                         return(REW_HALT);
 194                 else
 195                         return(REW_REWIND);
 196         }
 197
 198         /*
 199          * Next follow the implicit scope-smashings as defined by man.7:
 200          * section, sub-section, etc.
 201          */
 202
 203         switch (tok) {
 204         case (MAN_SH):
 205                 break;
 206         case (MAN_SS):
 207                 /* Rewind to a section, if a block. */
 208                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 209                         return(c);
 210                 break;
 211         case (MAN_RS):
 212                 /* Preserve empty paragraphs before RS. */
 213                 if (0 == n->nchild && (MAN_P == n->tok ||
 214                     MAN_PP == n->tok || MAN_LP == n->tok))
 215                         return(REW_HALT);
 216                 /* Rewind to a subsection, if a block. */
 217                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 218                         return(c);
 219                 /* Rewind to a section, if a block. */
 220                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 221                         return(c);
 222                 break;
 223         default:
 224                 /* Rewind to an offsetter, if a block. */
 225                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 226                         return(c);
 227                 /* Rewind to a subsection, if a block. */
 228                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 229                         return(c);
 230                 /* Rewind to a section, if a block. */
 231                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 232                         return(c);
 233                 break;
 234         }
 235
 236         return(REW_NOHALT);
 237 }
 238
 239
 240 /*
 241  * Rewinding entails ascending the parse tree until a coherent point,
 242  * for example, the `SH' macro will close out any intervening `SS'
 243  * scopes.  When a scope is closed, it must be validated and actioned.
 244  */
 245 static int
 246 rew_scope(enum man_type type, struct man *man, enum mant tok)
 247 {
 248         struct man_node *n;
 249         enum rew         c;
 250
 251         /* LINTED */
 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(1);
 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         assert(n);
 270
 271         return(man_unscope(man, n, MANDOCERR_MAX));
 272 }
 273
 274
 275 /*
 276  * Close out a generic explicit macro.
 277  */
 278 /* ARGSUSED */
 279 int
 280 blk_close(MACRO_PROT_ARGS)
 281 {
 282         enum mant                ntok;
 283         const struct man_node   *nn;
 284
 285         switch (tok) {
 286         case (MAN_RE):
 287                 ntok = MAN_RS;
 288                 break;
 289         case (MAN_UE):
 290                 ntok = MAN_UR;
 291                 break;
 292         default:
 293                 abort();
 294                 /* NOTREACHED */
 295         }
 296
 297         for (nn = man->last->parent; nn; nn = nn->parent)
 298                 if (ntok == nn->tok && MAN_BLOCK == nn->type)
 299                         break;
 300
 301         if (NULL == nn) {
 302                 man_pmsg(man, line, ppos, MANDOCERR_NOSCOPE);
 303                 if ( ! rew_scope(MAN_BLOCK, man, MAN_PP))
 304                         return(0);
 305         } else
 306                 man_unscope(man, nn, MANDOCERR_MAX);
 307
 308         return(1);
 309 }
 310
 311
 312 /* ARGSUSED */
 313 int
 314 blk_exp(MACRO_PROT_ARGS)
 315 {
 316         struct man_node *n;
 317         int              la;
 318         char            *p;
 319
 320         /* Close out prior implicit scopes. */
 321
 322         if ( ! rew_scope(MAN_BLOCK, man, tok))
 323                 return(0);
 324
 325         if ( ! man_block_alloc(man, line, ppos, tok))
 326                 return(0);
 327         if ( ! man_head_alloc(man, line, ppos, tok))
 328                 return(0);
 329
 330         for (;;) {
 331                 la = *pos;
 332                 if ( ! man_args(man, line, pos, buf, &p))
 333                         break;
 334                 if ( ! man_word_alloc(man, line, la, p))
 335                         return(0);
 336         }
 337
 338         assert(man);
 339         assert(tok != MAN_MAX);
 340
 341         for (n = man->last; n; n = n->parent) {
 342                 if (n->tok != tok)
 343                         continue;
 344                 assert(MAN_HEAD == n->type);
 345                 man_unscope(man, n, MANDOCERR_MAX);
 346                 break;
 347         }
 348
 349         return(man_body_alloc(man, line, ppos, tok));
 350 }
 351
 352
 353
 354 /*
 355  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 356  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 357  * scopes, such as `SH' closing out an `SS', are defined in the rew
 358  * routines.
 359  */
 360 /* ARGSUSED */
 361 int
 362 blk_imp(MACRO_PROT_ARGS)
 363 {
 364         int              la;
 365         char            *p;
 366         struct man_node *n;
 367
 368         /* Close out prior scopes. */
 369
 370         if ( ! rew_scope(MAN_BODY, man, tok))
 371                 return(0);
 372         if ( ! rew_scope(MAN_BLOCK, man, tok))
 373                 return(0);
 374
 375         /* Allocate new block & head scope. */
 376
 377         if ( ! man_block_alloc(man, line, ppos, tok))
 378                 return(0);
 379         if ( ! man_head_alloc(man, line, ppos, tok))
 380                 return(0);
 381
 382         n = man->last;
 383
 384         /* Add line arguments. */
 385
 386         for (;;) {
 387                 la = *pos;
 388                 if ( ! man_args(man, line, pos, buf, &p))
 389                         break;
 390                 if ( ! man_word_alloc(man, line, la, p))
 391                         return(0);
 392         }
 393
 394         /* Close out head and open body (unless MAN_SCOPE). */
 395
 396         if (MAN_SCOPED & man_macros[tok].flags) {
 397                 /* If we're forcing scope (`TP'), keep it open. */
 398                 if (MAN_FSCOPED & man_macros[tok].flags) {
 399                         man->flags |= MAN_BLINE;
 400                         return(1);
 401                 } else if (n == man->last) {
 402                         man->flags |= MAN_BLINE;
 403                         return(1);
 404                 }
 405         }
 406
 407         if ( ! rew_scope(MAN_HEAD, man, tok))
 408                 return(0);
 409         return(man_body_alloc(man, line, ppos, tok));
 410 }
 411
 412
 413 /* ARGSUSED */
 414 int
 415 in_line_eoln(MACRO_PROT_ARGS)
 416 {
 417         int              la;
 418         char            *p;
 419         struct man_node *n;
 420
 421         if ( ! man_elem_alloc(man, line, ppos, tok))
 422                 return(0);
 423
 424         n = man->last;
 425
 426         for (;;) {
 427                 la = *pos;
 428                 if ( ! man_args(man, line, pos, buf, &p))
 429                         break;
 430                 if ( ! man_word_alloc(man, line, la, p))
 431                         return(0);
 432         }
 433
 434         /*
 435          * Append MAN_EOS in case the last snipped argument
 436          * ends with a dot, e.g. `.IR syslog (3).'
 437          */
 438
 439         if (n != man->last &&
 440             mandoc_eos(man->last->string, strlen(man->last->string), 0))
 441                 man->last->flags |= MAN_EOS;
 442
 443         /*
 444          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 445          * next-line scoped), then set our mode to indicate that we're
 446          * waiting for terms to load into our context.
 447          */
 448
 449         if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
 450                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 451                 man->flags |= MAN_ELINE;
 452                 return(1);
 453         }
 454
 455         /* Set ignorable context, if applicable. */
 456
 457         if (MAN_NSCOPED & man_macros[tok].flags) {
 458                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 459                 man->flags |= MAN_ILINE;
 460         }
 461
 462         assert(MAN_ROOT != man->last->type);
 463         man->next = MAN_NEXT_SIBLING;
 464
 465         /*
 466          * Rewind our element scope.  Note that when TH is pruned, we'll
 467          * be back at the root, so make sure that we don't clobber as
 468          * its sibling.
 469          */
 470
 471         for ( ; man->last; man->last = man->last->parent) {
 472                 if (man->last == n)
 473                         break;
 474                 if (man->last->type == MAN_ROOT)
 475                         break;
 476                 if ( ! man_valid_post(man))
 477                         return(0);
 478         }
 479
 480         assert(man->last);
 481
 482         /*
 483          * Same here regarding whether we're back at the root.
 484          */
 485
 486         if (man->last->type != MAN_ROOT && ! man_valid_post(man))
 487                 return(0);
 488
 489         return(1);
 490 }
 491
 492
 493 int
 494 man_macroend(struct man *man)
 495 {
 496
 497         return(man_unscope(man, man->first, MANDOCERR_SCOPEEXIT));
 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 }