contrib/mdocml/man_macro.c

   1 /*      $Id: man_macro.c,v 1.87 2014/07/30 23:01:39 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
 108                 /* Reached the end of the document? */
 109
 110                 if (to == NULL && ! (n->flags & MAN_VALID)) {
 111                         if (man->flags & (MAN_BLINE | MAN_ELINE) &&
 112                             man_macros[n->tok].flags & MAN_SCOPED) {
 113                                 mandoc_vmsg(MANDOCERR_BLK_LINE,
 114                                     man->parse, n->line, n->pos,
 115                                     "EOF breaks %s",
 116                                     man_macronames[n->tok]);
 117                                 if (man->flags & MAN_ELINE)
 118                                         man->flags &= ~MAN_ELINE;
 119                                 else {
 120                                         assert(n->type == MAN_HEAD);
 121                                         n = n->parent;
 122                                         man->flags &= ~MAN_BLINE;
 123                                 }
 124                                 man->last = n;
 125                                 n = n->parent;
 126                                 man_node_delete(man, man->last);
 127                                 continue;
 128                         }
 129                         if (n->type == MAN_BLOCK &&
 130                             man_macros[n->tok].flags & MAN_EXPLICIT)
 131                                 mandoc_msg(MANDOCERR_BLK_NOEND,
 132                                     man->parse, n->line, n->pos,
 133                                     man_macronames[n->tok]);
 134                 }
 135
 136                 /*
 137                  * We might delete the man->last node
 138                  * in the post-validation phase.
 139                  * Save a pointer to the parent such that
 140                  * we know where to continue the iteration.
 141                  */
 142                 man->last = n;
 143                 n = n->parent;
 144                 if ( ! man_valid_post(man))
 145                         return(0);
 146         }
 147         return(1);
 148 }
 149
 150 static enum rew
 151 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 152 {
 153
 154         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 155             MAN_BODY == n->parent->type)
 156                 return(REW_REWIND);
 157         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 158 }
 159
 160 /*
 161  * There are three scope levels: scoped to the root (all), scoped to the
 162  * section (all less sections), and scoped to subsections (all less
 163  * sections and subsections).
 164  */
 165 static enum rew
 166 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 167 {
 168         enum rew         c;
 169
 170         /* We cannot progress beyond the root ever. */
 171         if (MAN_ROOT == n->type)
 172                 return(REW_HALT);
 173
 174         assert(n->parent);
 175
 176         /* Normal nodes shouldn't go to the level of the root. */
 177         if (MAN_ROOT == n->parent->type)
 178                 return(REW_REWIND);
 179
 180         /* Already-validated nodes should be closed out. */
 181         if (MAN_VALID & n->flags)
 182                 return(REW_NOHALT);
 183
 184         /* First: rewind to ourselves. */
 185         if (type == n->type && tok == n->tok) {
 186                 if (MAN_EXPLICIT & man_macros[n->tok].flags)
 187                         return(REW_HALT);
 188                 else
 189                         return(REW_REWIND);
 190         }
 191
 192         /*
 193          * Next follow the implicit scope-smashings as defined by man.7:
 194          * section, sub-section, etc.
 195          */
 196
 197         switch (tok) {
 198         case MAN_SH:
 199                 break;
 200         case MAN_SS:
 201                 /* Rewind to a section, if a block. */
 202                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 203                         return(c);
 204                 break;
 205         case MAN_RS:
 206                 /* Preserve empty paragraphs before RS. */
 207                 if (0 == n->nchild && (MAN_P == n->tok ||
 208                     MAN_PP == n->tok || MAN_LP == n->tok))
 209                         return(REW_HALT);
 210                 /* Rewind to a subsection, if a block. */
 211                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 212                         return(c);
 213                 /* Rewind to a section, if a block. */
 214                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 215                         return(c);
 216                 break;
 217         default:
 218                 /* Rewind to an offsetter, if a block. */
 219                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 220                         return(c);
 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         }
 229
 230         return(REW_NOHALT);
 231 }
 232
 233 /*
 234  * Rewinding entails ascending the parse tree until a coherent point,
 235  * for example, the `SH' macro will close out any intervening `SS'
 236  * scopes.  When a scope is closed, it must be validated and actioned.
 237  */
 238 static int
 239 rew_scope(enum man_type type, struct man *man, enum mant tok)
 240 {
 241         struct man_node *n;
 242         enum rew         c;
 243
 244         for (n = man->last; n; n = n->parent) {
 245                 /*
 246                  * Whether we should stop immediately (REW_HALT), stop
 247                  * and rewind until this point (REW_REWIND), or keep
 248                  * rewinding (REW_NOHALT).
 249                  */
 250                 c = rew_dohalt(tok, type, n);
 251                 if (REW_HALT == c)
 252                         return(1);
 253                 if (REW_REWIND == c)
 254                         break;
 255         }
 256
 257         /*
 258          * Rewind until the current point.  Warn if we're a roff
 259          * instruction that's mowing over explicit scopes.
 260          */
 261         assert(n);
 262
 263         return(man_unscope(man, n));
 264 }
 265
 266
 267 /*
 268  * Close out a generic explicit macro.
 269  */
 270 int
 271 blk_close(MACRO_PROT_ARGS)
 272 {
 273         enum mant                ntok;
 274         const struct man_node   *nn;
 275
 276         switch (tok) {
 277         case MAN_RE:
 278                 ntok = MAN_RS;
 279                 break;
 280         case MAN_UE:
 281                 ntok = MAN_UR;
 282                 break;
 283         default:
 284                 abort();
 285                 /* NOTREACHED */
 286         }
 287
 288         for (nn = man->last->parent; nn; nn = nn->parent)
 289                 if (ntok == nn->tok && MAN_BLOCK == nn->type)
 290                         break;
 291
 292         if (NULL == nn) {
 293                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 294                     line, ppos, man_macronames[tok]);
 295                 if ( ! rew_scope(MAN_BLOCK, man, MAN_PP))
 296                         return(0);
 297         } else
 298                 man_unscope(man, nn);
 299
 300         return(1);
 301 }
 302
 303 int
 304 blk_exp(MACRO_PROT_ARGS)
 305 {
 306         struct man_node *n;
 307         int              la;
 308         char            *p;
 309
 310         /* Close out prior implicit scopes. */
 311
 312         if ( ! rew_scope(MAN_BLOCK, man, tok))
 313                 return(0);
 314
 315         if ( ! man_block_alloc(man, line, ppos, tok))
 316                 return(0);
 317         if ( ! man_head_alloc(man, line, ppos, tok))
 318                 return(0);
 319
 320         for (;;) {
 321                 la = *pos;
 322                 if ( ! man_args(man, line, pos, buf, &p))
 323                         break;
 324                 if ( ! man_word_alloc(man, line, la, p))
 325                         return(0);
 326         }
 327
 328         assert(man);
 329         assert(tok != MAN_MAX);
 330
 331         for (n = man->last; n; n = n->parent) {
 332                 if (n->tok != tok)
 333                         continue;
 334                 assert(MAN_HEAD == n->type);
 335                 man_unscope(man, n);
 336                 break;
 337         }
 338
 339         return(man_body_alloc(man, line, ppos, tok));
 340 }
 341
 342 /*
 343  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 344  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 345  * scopes, such as `SH' closing out an `SS', are defined in the rew
 346  * routines.
 347  */
 348 int
 349 blk_imp(MACRO_PROT_ARGS)
 350 {
 351         int              la;
 352         char            *p;
 353         struct man_node *n;
 354
 355         /* Close out prior scopes. */
 356
 357         if ( ! rew_scope(MAN_BODY, man, tok))
 358                 return(0);
 359         if ( ! rew_scope(MAN_BLOCK, man, tok))
 360                 return(0);
 361
 362         /* Allocate new block & head scope. */
 363
 364         if ( ! man_block_alloc(man, line, ppos, tok))
 365                 return(0);
 366         if ( ! man_head_alloc(man, line, ppos, tok))
 367                 return(0);
 368
 369         n = man->last;
 370
 371         /* Add line arguments. */
 372
 373         for (;;) {
 374                 la = *pos;
 375                 if ( ! man_args(man, line, pos, buf, &p))
 376                         break;
 377                 if ( ! man_word_alloc(man, line, la, p))
 378                         return(0);
 379         }
 380
 381         /* Close out head and open body (unless MAN_SCOPE). */
 382
 383         if (MAN_SCOPED & man_macros[tok].flags) {
 384                 /* If we're forcing scope (`TP'), keep it open. */
 385                 if (MAN_FSCOPED & man_macros[tok].flags) {
 386                         man->flags |= MAN_BLINE;
 387                         return(1);
 388                 } else if (n == man->last) {
 389                         man->flags |= MAN_BLINE;
 390                         return(1);
 391                 }
 392         }
 393
 394         if ( ! rew_scope(MAN_HEAD, man, tok))
 395                 return(0);
 396         return(man_body_alloc(man, line, ppos, tok));
 397 }
 398
 399 int
 400 in_line_eoln(MACRO_PROT_ARGS)
 401 {
 402         int              la;
 403         char            *p;
 404         struct man_node *n;
 405
 406         if ( ! man_elem_alloc(man, line, ppos, tok))
 407                 return(0);
 408
 409         n = man->last;
 410
 411         for (;;) {
 412                 la = *pos;
 413                 if ( ! man_args(man, line, pos, buf, &p))
 414                         break;
 415                 if ( ! man_word_alloc(man, line, la, p))
 416                         return(0);
 417         }
 418
 419         /*
 420          * Append MAN_EOS in case the last snipped argument
 421          * ends with a dot, e.g. `.IR syslog (3).'
 422          */
 423
 424         if (n != man->last &&
 425             mandoc_eos(man->last->string, strlen(man->last->string)))
 426                 man->last->flags |= MAN_EOS;
 427
 428         /*
 429          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 430          * next-line scoped), then set our mode to indicate that we're
 431          * waiting for terms to load into our context.
 432          */
 433
 434         if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
 435                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 436                 man->flags |= MAN_ELINE;
 437                 return(1);
 438         }
 439
 440         assert(MAN_ROOT != man->last->type);
 441         man->next = MAN_NEXT_SIBLING;
 442
 443         /*
 444          * Rewind our element scope.  Note that when TH is pruned, we'll
 445          * be back at the root, so make sure that we don't clobber as
 446          * its sibling.
 447          */
 448
 449         for ( ; man->last; man->last = man->last->parent) {
 450                 if (man->last == n)
 451                         break;
 452                 if (man->last->type == MAN_ROOT)
 453                         break;
 454                 if ( ! man_valid_post(man))
 455                         return(0);
 456         }
 457
 458         assert(man->last);
 459
 460         /*
 461          * Same here regarding whether we're back at the root.
 462          */
 463
 464         if (man->last->type != MAN_ROOT && ! man_valid_post(man))
 465                 return(0);
 466
 467         return(1);
 468 }
 469
 470
 471 int
 472 man_macroend(struct man *man)
 473 {
 474
 475         return(man_unscope(man, man->first));
 476 }
 477
 478 static int
 479 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 480 {
 481         char     *start;
 482
 483         assert(*pos);
 484         *v = start = buf + *pos;
 485         assert(' ' != *start);
 486
 487         if ('\0' == *start)
 488                 return(0);
 489
 490         *v = mandoc_getarg(man->parse, v, line, pos);
 491         return(1);
 492 }