man_macro.c

   1 /*      $Id: man_macro.c,v 1.81 2014/03/30 19:47:48 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 static  void             rew_warn(struct man *,
  53                                 struct man_node *, enum mandocerr);
  54
  55 const   struct man_macro __man_macros[MAN_MAX] = {
  56         { in_line_eoln, MAN_NSCOPED }, /* br */
  57         { in_line_eoln, MAN_BSCOPE }, /* TH */
  58         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
  59         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
  60         { blk_imp, MAN_BSCOPE | MAN_SCOPED | MAN_FSCOPED }, /* TP */
  61         { blk_imp, MAN_BSCOPE }, /* LP */
  62         { blk_imp, MAN_BSCOPE }, /* PP */
  63         { blk_imp, MAN_BSCOPE }, /* P */
  64         { blk_imp, MAN_BSCOPE }, /* IP */
  65         { blk_imp, MAN_BSCOPE }, /* HP */
  66         { in_line_eoln, MAN_SCOPED }, /* SM */
  67         { in_line_eoln, MAN_SCOPED }, /* SB */
  68         { in_line_eoln, 0 }, /* BI */
  69         { in_line_eoln, 0 }, /* IB */
  70         { in_line_eoln, 0 }, /* BR */
  71         { in_line_eoln, 0 }, /* RB */
  72         { in_line_eoln, MAN_SCOPED }, /* R */
  73         { in_line_eoln, MAN_SCOPED }, /* B */
  74         { in_line_eoln, MAN_SCOPED }, /* I */
  75         { in_line_eoln, 0 }, /* IR */
  76         { in_line_eoln, 0 }, /* RI */
  77         { in_line_eoln, MAN_NSCOPED }, /* na */
  78         { in_line_eoln, MAN_NSCOPED }, /* sp */
  79         { in_line_eoln, MAN_BSCOPE }, /* nf */
  80         { in_line_eoln, MAN_BSCOPE }, /* fi */
  81         { blk_close, 0 }, /* RE */
  82         { blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* RS */
  83         { in_line_eoln, 0 }, /* DT */
  84         { in_line_eoln, 0 }, /* UC */
  85         { in_line_eoln, 0 }, /* PD */
  86         { in_line_eoln, 0 }, /* AT */
  87         { in_line_eoln, 0 }, /* in */
  88         { in_line_eoln, 0 }, /* ft */
  89         { in_line_eoln, 0 }, /* OP */
  90         { in_line_eoln, MAN_BSCOPE }, /* EX */
  91         { in_line_eoln, MAN_BSCOPE }, /* EE */
  92         { blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* UR */
  93         { blk_close, 0 }, /* UE */
  94         { in_line_eoln, 0 }, /* ll */
  95 };
  96
  97 const   struct man_macro * const man_macros = __man_macros;
  98
  99
 100 /*
 101  * Warn when "n" is an explicit non-roff macro.
 102  */
 103 static void
 104 rew_warn(struct man *man, struct man_node *n, enum mandocerr er)
 105 {
 106
 107         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
 108                 return;
 109         if (MAN_VALID & n->flags)
 110                 return;
 111         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 112                 return;
 113
 114         assert(er < MANDOCERR_FATAL);
 115         man_nmsg(man, n, er);
 116 }
 117
 118
 119 /*
 120  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 121  * will be used if an explicit block scope is being closed out.
 122  */
 123 int
 124 man_unscope(struct man *man, const struct man_node *to,
 125                 enum mandocerr er)
 126 {
 127         struct man_node *n;
 128
 129         assert(to);
 130
 131         man->next = MAN_NEXT_SIBLING;
 132
 133         /* LINTED */
 134         while (man->last != to) {
 135                 /*
 136                  * Save the parent here, because we may delete the
 137                  * man->last node in the post-validation phase and reset
 138                  * it to man->last->parent, causing a step in the closing
 139                  * out to be lost.
 140                  */
 141                 n = man->last->parent;
 142                 rew_warn(man, man->last, er);
 143                 if ( ! man_valid_post(man))
 144                         return(0);
 145                 man->last = n;
 146                 assert(man->last);
 147         }
 148
 149         rew_warn(man, man->last, er);
 150         if ( ! man_valid_post(man))
 151                 return(0);
 152
 153         return(1);
 154 }
 155
 156
 157 static enum rew
 158 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 159 {
 160
 161         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 162                         MAN_BODY == n->parent->type)
 163                 return(REW_REWIND);
 164         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 165 }
 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 /*
 243  * Rewinding entails ascending the parse tree until a coherent point,
 244  * for example, the `SH' macro will close out any intervening `SS'
 245  * scopes.  When a scope is closed, it must be validated and actioned.
 246  */
 247 static int
 248 rew_scope(enum man_type type, struct man *man, enum mant tok)
 249 {
 250         struct man_node *n;
 251         enum rew         c;
 252
 253         /* LINTED */
 254         for (n = man->last; n; n = n->parent) {
 255                 /*
 256                  * Whether we should stop immediately (REW_HALT), stop
 257                  * and rewind until this point (REW_REWIND), or keep
 258                  * rewinding (REW_NOHALT).
 259                  */
 260                 c = rew_dohalt(tok, type, n);
 261                 if (REW_HALT == c)
 262                         return(1);
 263                 if (REW_REWIND == c)
 264                         break;
 265         }
 266
 267         /*
 268          * Rewind until the current point.  Warn if we're a roff
 269          * instruction that's mowing over explicit scopes.
 270          */
 271         assert(n);
 272
 273         return(man_unscope(man, n, MANDOCERR_MAX));
 274 }
 275
 276
 277 /*
 278  * Close out a generic explicit macro.
 279  */
 280 /* ARGSUSED */
 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                 man_pmsg(man, line, ppos, MANDOCERR_NOSCOPE);
 305                 if ( ! rew_scope(MAN_BLOCK, man, MAN_PP))
 306                         return(0);
 307         } else
 308                 man_unscope(man, nn, MANDOCERR_MAX);
 309
 310         return(1);
 311 }
 312
 313
 314 /* ARGSUSED */
 315 int
 316 blk_exp(MACRO_PROT_ARGS)
 317 {
 318         struct man_node *n;
 319         int              la;
 320         char            *p;
 321
 322         /* Close out prior implicit scopes. */
 323
 324         if ( ! rew_scope(MAN_BLOCK, man, tok))
 325                 return(0);
 326
 327         if ( ! man_block_alloc(man, line, ppos, tok))
 328                 return(0);
 329         if ( ! man_head_alloc(man, line, ppos, tok))
 330                 return(0);
 331
 332         for (;;) {
 333                 la = *pos;
 334                 if ( ! man_args(man, line, pos, buf, &p))
 335                         break;
 336                 if ( ! man_word_alloc(man, line, la, p))
 337                         return(0);
 338         }
 339
 340         assert(man);
 341         assert(tok != MAN_MAX);
 342
 343         for (n = man->last; n; n = n->parent) {
 344                 if (n->tok != tok)
 345                         continue;
 346                 assert(MAN_HEAD == n->type);
 347                 man_unscope(man, n, MANDOCERR_MAX);
 348                 break;
 349         }
 350
 351         return(man_body_alloc(man, line, ppos, tok));
 352 }
 353
 354
 355
 356 /*
 357  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 358  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 359  * scopes, such as `SH' closing out an `SS', are defined in the rew
 360  * routines.
 361  */
 362 /* ARGSUSED */
 363 int
 364 blk_imp(MACRO_PROT_ARGS)
 365 {
 366         int              la;
 367         char            *p;
 368         struct man_node *n;
 369
 370         /* Close out prior scopes. */
 371
 372         if ( ! rew_scope(MAN_BODY, man, tok))
 373                 return(0);
 374         if ( ! rew_scope(MAN_BLOCK, man, tok))
 375                 return(0);
 376
 377         /* Allocate new block & head scope. */
 378
 379         if ( ! man_block_alloc(man, line, ppos, tok))
 380                 return(0);
 381         if ( ! man_head_alloc(man, line, ppos, tok))
 382                 return(0);
 383
 384         n = man->last;
 385
 386         /* Add line arguments. */
 387
 388         for (;;) {
 389                 la = *pos;
 390                 if ( ! man_args(man, line, pos, buf, &p))
 391                         break;
 392                 if ( ! man_word_alloc(man, line, la, p))
 393                         return(0);
 394         }
 395
 396         /* Close out head and open body (unless MAN_SCOPE). */
 397
 398         if (MAN_SCOPED & man_macros[tok].flags) {
 399                 /* If we're forcing scope (`TP'), keep it open. */
 400                 if (MAN_FSCOPED & man_macros[tok].flags) {
 401                         man->flags |= MAN_BLINE;
 402                         return(1);
 403                 } else if (n == man->last) {
 404                         man->flags |= MAN_BLINE;
 405                         return(1);
 406                 }
 407         }
 408
 409         if ( ! rew_scope(MAN_HEAD, man, tok))
 410                 return(0);
 411         return(man_body_alloc(man, line, ppos, tok));
 412 }
 413
 414
 415 /* ARGSUSED */
 416 int
 417 in_line_eoln(MACRO_PROT_ARGS)
 418 {
 419         int              la;
 420         char            *p;
 421         struct man_node *n;
 422
 423         if ( ! man_elem_alloc(man, line, ppos, tok))
 424                 return(0);
 425
 426         n = man->last;
 427
 428         for (;;) {
 429                 la = *pos;
 430                 if ( ! man_args(man, line, pos, buf, &p))
 431                         break;
 432                 if ( ! man_word_alloc(man, line, la, p))
 433                         return(0);
 434         }
 435
 436         /*
 437          * Append MAN_EOS in case the last snipped argument
 438          * ends with a dot, e.g. `.IR syslog (3).'
 439          */
 440
 441         if (n != man->last &&
 442             mandoc_eos(man->last->string, strlen(man->last->string)))
 443                 man->last->flags |= MAN_EOS;
 444
 445         /*
 446          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 447          * next-line scoped), then set our mode to indicate that we're
 448          * waiting for terms to load into our context.
 449          */
 450
 451         if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
 452                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 453                 man->flags |= MAN_ELINE;
 454                 return(1);
 455         }
 456
 457         /* Set ignorable context, if applicable. */
 458
 459         if (MAN_NSCOPED & man_macros[tok].flags) {
 460                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 461                 man->flags |= MAN_ILINE;
 462         }
 463
 464         assert(MAN_ROOT != man->last->type);
 465         man->next = MAN_NEXT_SIBLING;
 466
 467         /*
 468          * Rewind our element scope.  Note that when TH is pruned, we'll
 469          * be back at the root, so make sure that we don't clobber as
 470          * its sibling.
 471          */
 472
 473         for ( ; man->last; man->last = man->last->parent) {
 474                 if (man->last == n)
 475                         break;
 476                 if (man->last->type == MAN_ROOT)
 477                         break;
 478                 if ( ! man_valid_post(man))
 479                         return(0);
 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 != MAN_ROOT && ! man_valid_post(man))
 489                 return(0);
 490
 491         return(1);
 492 }
 493
 494
 495 int
 496 man_macroend(struct man *man)
 497 {
 498
 499         return(man_unscope(man, man->first, MANDOCERR_SCOPEEXIT));
 500 }
 501
 502 static int
 503 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 504 {
 505         char     *start;
 506
 507         assert(*pos);
 508         *v = start = buf + *pos;
 509         assert(' ' != *start);
 510
 511         if ('\0' == *start)
 512                 return(0);
 513
 514         *v = mandoc_getarg(man->parse, v, line, pos);
 515         return(1);
 516 }