man_macro.c

   1 /*      $Id: man_macro.c,v 1.55 2011/01/12 16:55:22 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <assert.h>
  22 #include <ctype.h>
  23 #include <stdlib.h>
  24 #include <string.h>
  25
  26 #include "mandoc.h"
  27 #include "libman.h"
  28
  29 enum    rew {
  30         REW_REWIND,
  31         REW_NOHALT,
  32         REW_HALT
  33 };
  34
  35 static  int              blk_close(MACRO_PROT_ARGS);
  36 static  int              blk_exp(MACRO_PROT_ARGS);
  37 static  int              blk_imp(MACRO_PROT_ARGS);
  38 static  int              in_line_eoln(MACRO_PROT_ARGS);
  39
  40 static  int              rew_scope(enum man_type,
  41                                 struct man *, enum mant);
  42 static  enum rew         rew_dohalt(enum mant, enum man_type,
  43                                 const struct man_node *);
  44 static  enum rew         rew_block(enum mant, enum man_type,
  45                                 const struct man_node *);
  46 static  int              rew_warn(struct man *,
  47                                 struct man_node *, enum mandocerr);
  48
  49 const   struct man_macro __man_macros[MAN_MAX] = {
  50         { in_line_eoln, MAN_NSCOPED }, /* br */
  51         { in_line_eoln, 0 }, /* TH */
  52         { blk_imp, MAN_SCOPED }, /* SH */
  53         { blk_imp, MAN_SCOPED }, /* SS */
  54         { blk_imp, MAN_SCOPED | MAN_FSCOPED }, /* TP */
  55         { blk_imp, 0 }, /* LP */
  56         { blk_imp, 0 }, /* PP */
  57         { blk_imp, 0 }, /* P */
  58         { blk_imp, 0 }, /* IP */
  59         { blk_imp, 0 }, /* HP */
  60         { in_line_eoln, MAN_SCOPED }, /* SM */
  61         { in_line_eoln, MAN_SCOPED }, /* SB */
  62         { in_line_eoln, 0 }, /* BI */
  63         { in_line_eoln, 0 }, /* IB */
  64         { in_line_eoln, 0 }, /* BR */
  65         { in_line_eoln, 0 }, /* RB */
  66         { in_line_eoln, MAN_SCOPED }, /* R */
  67         { in_line_eoln, MAN_SCOPED }, /* B */
  68         { in_line_eoln, MAN_SCOPED }, /* I */
  69         { in_line_eoln, 0 }, /* IR */
  70         { in_line_eoln, 0 }, /* RI */
  71         { in_line_eoln, MAN_NSCOPED }, /* na */
  72         { in_line_eoln, MAN_NSCOPED }, /* sp */
  73         { in_line_eoln, 0 }, /* nf */
  74         { in_line_eoln, 0 }, /* fi */
  75         { blk_close, 0 }, /* RE */
  76         { blk_exp, MAN_EXPLICIT }, /* RS */
  77         { in_line_eoln, 0 }, /* DT */
  78         { in_line_eoln, 0 }, /* UC */
  79         { in_line_eoln, 0 }, /* PD */
  80         { in_line_eoln, 0 }, /* AT */
  81         { in_line_eoln, 0 }, /* in */
  82         { in_line_eoln, 0 }, /* ft */
  83 };
  84
  85 const   struct man_macro * const man_macros = __man_macros;
  86
  87
  88 /*
  89  * Warn when "n" is an explicit non-roff macro.
  90  */
  91 static int
  92 rew_warn(struct man *m, struct man_node *n, enum mandocerr er)
  93 {
  94
  95         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
  96                 return(1);
  97         if (MAN_VALID & n->flags)
  98                 return(1);
  99         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 100                 return(1);
 101         return(man_nmsg(m, n, er));
 102 }
 103
 104
 105 /*
 106  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 107  * will be used if an explicit block scope is being closed out.
 108  */
 109 int
 110 man_unscope(struct man *m, const struct man_node *to,
 111                 enum mandocerr er)
 112 {
 113         struct man_node *n;
 114
 115         assert(to);
 116
 117         /* LINTED */
 118         while (m->last != to) {
 119                 /*
 120                  * Save the parent here, because we may delete the
 121                  * m->last node in the post-validation phase and reset
 122                  * it to m->last->parent, causing a step in the closing
 123                  * out to be lost.
 124                  */
 125                 n = m->last->parent;
 126                 if ( ! rew_warn(m, m->last, er))
 127                         return(0);
 128                 if ( ! man_valid_post(m))
 129                         return(0);
 130                 m->last = n;
 131                 assert(m->last);
 132         }
 133
 134         if ( ! rew_warn(m, m->last, er))
 135                 return(0);
 136         if ( ! man_valid_post(m))
 137                 return(0);
 138
 139         m->next = MAN_ROOT == m->last->type ?
 140                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 141
 142         return(1);
 143 }
 144
 145
 146 static enum rew
 147 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 148 {
 149
 150         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 151                         MAN_BODY == n->parent->type)
 152                 return(REW_REWIND);
 153         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 154 }
 155
 156
 157 /*
 158  * There are three scope levels: scoped to the root (all), scoped to the
 159  * section (all less sections), and scoped to subsections (all less
 160  * sections and subsections).
 161  */
 162 static enum rew
 163 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 164 {
 165         enum rew         c;
 166
 167         /* We cannot progress beyond the root ever. */
 168         if (MAN_ROOT == n->type)
 169                 return(REW_HALT);
 170
 171         assert(n->parent);
 172
 173         /* Normal nodes shouldn't go to the level of the root. */
 174         if (MAN_ROOT == n->parent->type)
 175                 return(REW_REWIND);
 176
 177         /* Already-validated nodes should be closed out. */
 178         if (MAN_VALID & n->flags)
 179                 return(REW_NOHALT);
 180
 181         /* First: rewind to ourselves. */
 182         if (type == n->type && tok == n->tok)
 183                 return(REW_REWIND);
 184
 185         /*
 186          * Next follow the implicit scope-smashings as defined by man.7:
 187          * section, sub-section, etc.
 188          */
 189
 190         switch (tok) {
 191         case (MAN_SH):
 192                 break;
 193         case (MAN_SS):
 194                 /* Rewind to a section, if a block. */
 195                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 196                         return(c);
 197                 break;
 198         case (MAN_RS):
 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         default:
 207                 /* Rewind to an offsetter, if a block. */
 208                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 209                         return(c);
 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         }
 218
 219         return(REW_NOHALT);
 220 }
 221
 222
 223 /*
 224  * Rewinding entails ascending the parse tree until a coherent point,
 225  * for example, the `SH' macro will close out any intervening `SS'
 226  * scopes.  When a scope is closed, it must be validated and actioned.
 227  */
 228 static int
 229 rew_scope(enum man_type type, struct man *m, enum mant tok)
 230 {
 231         struct man_node *n;
 232         enum rew         c;
 233
 234         /* LINTED */
 235         for (n = m->last; n; n = n->parent) {
 236                 /*
 237                  * Whether we should stop immediately (REW_HALT), stop
 238                  * and rewind until this point (REW_REWIND), or keep
 239                  * rewinding (REW_NOHALT).
 240                  */
 241                 c = rew_dohalt(tok, type, n);
 242                 if (REW_HALT == c)
 243                         return(1);
 244                 if (REW_REWIND == c)
 245                         break;
 246         }
 247
 248         /*
 249          * Rewind until the current point.  Warn if we're a roff
 250          * instruction that's mowing over explicit scopes.
 251          */
 252         assert(n);
 253
 254         return(man_unscope(m, n, MANDOCERR_MAX));
 255 }
 256
 257
 258 /*
 259  * Close out a generic explicit macro.
 260  */
 261 /* ARGSUSED */
 262 int
 263 blk_close(MACRO_PROT_ARGS)
 264 {
 265         enum mant                ntok;
 266         const struct man_node   *nn;
 267
 268         switch (tok) {
 269         case (MAN_RE):
 270                 ntok = MAN_RS;
 271                 break;
 272         default:
 273                 abort();
 274                 /* NOTREACHED */
 275         }
 276
 277         for (nn = m->last->parent; nn; nn = nn->parent)
 278                 if (ntok == nn->tok)
 279                         break;
 280
 281         if (NULL == nn)
 282                 if ( ! man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE))
 283                         return(0);
 284
 285         if ( ! rew_scope(MAN_BODY, m, ntok))
 286                 return(0);
 287         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 288                 return(0);
 289
 290         return(1);
 291 }
 292
 293
 294 /* ARGSUSED */
 295 int
 296 blk_exp(MACRO_PROT_ARGS)
 297 {
 298         int              w, la;
 299         char            *p;
 300
 301         /*
 302          * Close out prior scopes.  "Regular" explicit macros cannot be
 303          * nested, but we allow roff macros to be placed just about
 304          * anywhere.
 305          */
 306
 307         if ( ! rew_scope(MAN_BODY, m, tok))
 308                 return(0);
 309         if ( ! rew_scope(MAN_BLOCK, m, tok))
 310                 return(0);
 311
 312         if ( ! man_block_alloc(m, line, ppos, tok))
 313                 return(0);
 314         if ( ! man_head_alloc(m, line, ppos, tok))
 315                 return(0);
 316
 317         for (;;) {
 318                 la = *pos;
 319                 w = man_args(m, line, pos, buf, &p);
 320
 321                 if (-1 == w)
 322                         return(0);
 323                 if (0 == w)
 324                         break;
 325
 326                 if ( ! man_word_alloc(m, line, la, p))
 327                         return(0);
 328         }
 329
 330         assert(m);
 331         assert(tok != MAN_MAX);
 332
 333         if ( ! rew_scope(MAN_HEAD, m, tok))
 334                 return(0);
 335         return(man_body_alloc(m, line, ppos, tok));
 336 }
 337
 338
 339
 340 /*
 341  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 342  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 343  * scopes, such as `SH' closing out an `SS', are defined in the rew
 344  * routines.
 345  */
 346 /* ARGSUSED */
 347 int
 348 blk_imp(MACRO_PROT_ARGS)
 349 {
 350         int              w, la;
 351         char            *p;
 352         struct man_node *n;
 353
 354         /* Close out prior scopes. */
 355
 356         if ( ! rew_scope(MAN_BODY, m, tok))
 357                 return(0);
 358         if ( ! rew_scope(MAN_BLOCK, m, tok))
 359                 return(0);
 360
 361         /* Allocate new block & head scope. */
 362
 363         if ( ! man_block_alloc(m, line, ppos, tok))
 364                 return(0);
 365         if ( ! man_head_alloc(m, line, ppos, tok))
 366                 return(0);
 367
 368         n = m->last;
 369
 370         /* Add line arguments. */
 371
 372         for (;;) {
 373                 la = *pos;
 374                 w = man_args(m, line, pos, buf, &p);
 375
 376                 if (-1 == w)
 377                         return(0);
 378                 if (0 == w)
 379                         break;
 380
 381                 if ( ! man_word_alloc(m, line, la, p))
 382                         return(0);
 383         }
 384
 385         /* Close out head and open body (unless MAN_SCOPE). */
 386
 387         if (MAN_SCOPED & man_macros[tok].flags) {
 388                 /* If we're forcing scope (`TP'), keep it open. */
 389                 if (MAN_FSCOPED & man_macros[tok].flags) {
 390                         m->flags |= MAN_BLINE;
 391                         return(1);
 392                 } else if (n == m->last) {
 393                         m->flags |= MAN_BLINE;
 394                         return(1);
 395                 }
 396         }
 397
 398         if ( ! rew_scope(MAN_HEAD, m, tok))
 399                 return(0);
 400         return(man_body_alloc(m, line, ppos, tok));
 401 }
 402
 403
 404 /* ARGSUSED */
 405 int
 406 in_line_eoln(MACRO_PROT_ARGS)
 407 {
 408         int              w, la;
 409         char            *p;
 410         struct man_node *n;
 411
 412         if ( ! man_elem_alloc(m, line, ppos, tok))
 413                 return(0);
 414
 415         n = m->last;
 416
 417         for (;;) {
 418                 la = *pos;
 419                 w = man_args(m, line, pos, buf, &p);
 420
 421                 if (-1 == w)
 422                         return(0);
 423                 if (0 == w)
 424                         break;
 425                 if ( ! man_word_alloc(m, line, la, p))
 426                         return(0);
 427         }
 428
 429         /*
 430          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 431          * next-line scoped), then set our mode to indicate that we're
 432          * waiting for terms to load into our context.
 433          */
 434
 435         if (n == m->last && MAN_SCOPED & man_macros[tok].flags) {
 436                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 437                 m->flags |= MAN_ELINE;
 438                 return(1);
 439         }
 440
 441         /* Set ignorable context, if applicable. */
 442
 443         if (MAN_NSCOPED & man_macros[tok].flags) {
 444                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 445                 m->flags |= MAN_ILINE;
 446         }
 447
 448         /*
 449          * Rewind our element scope.  Note that when TH is pruned, we'll
 450          * be back at the root, so make sure that we don't clobber as
 451          * its sibling.
 452          */
 453
 454         for ( ; m->last; m->last = m->last->parent) {
 455                 if (m->last == n)
 456                         break;
 457                 if (m->last->type == MAN_ROOT)
 458                         break;
 459                 if ( ! man_valid_post(m))
 460                         return(0);
 461         }
 462
 463         assert(m->last);
 464
 465         /*
 466          * Same here regarding whether we're back at the root.
 467          */
 468
 469         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 470                 return(0);
 471
 472         m->next = MAN_ROOT == m->last->type ?
 473                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 474
 475         return(1);
 476 }
 477
 478
 479 int
 480 man_macroend(struct man *m)
 481 {
 482
 483         return(man_unscope(m, m->first, MANDOCERR_SCOPEEXIT));
 484 }
 485