man_macro.c

   1 /*      $Id: man_macro.c,v 1.54 2010/12/08 10:58:22 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 *n,
 111                 enum mandocerr er)
 112 {
 113
 114         assert(n);
 115
 116         /* LINTED */
 117         while (m->last != n) {
 118                 if ( ! rew_warn(m, m->last, er))
 119                         return(0);
 120                 if ( ! man_valid_post(m))
 121                         return(0);
 122                 m->last = m->last->parent;
 123                 assert(m->last);
 124         }
 125
 126         if ( ! rew_warn(m, m->last, er))
 127                 return(0);
 128         if ( ! man_valid_post(m))
 129                 return(0);
 130
 131         m->next = MAN_ROOT == m->last->type ?
 132                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 133
 134         return(1);
 135 }
 136
 137
 138 static enum rew
 139 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 140 {
 141
 142         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 143                         MAN_BODY == n->parent->type)
 144                 return(REW_REWIND);
 145         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 146 }
 147
 148
 149 /*
 150  * There are three scope levels: scoped to the root (all), scoped to the
 151  * section (all less sections), and scoped to subsections (all less
 152  * sections and subsections).
 153  */
 154 static enum rew
 155 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 156 {
 157         enum rew         c;
 158
 159         /* We cannot progress beyond the root ever. */
 160         if (MAN_ROOT == n->type)
 161                 return(REW_HALT);
 162
 163         assert(n->parent);
 164
 165         /* Normal nodes shouldn't go to the level of the root. */
 166         if (MAN_ROOT == n->parent->type)
 167                 return(REW_REWIND);
 168
 169         /* Already-validated nodes should be closed out. */
 170         if (MAN_VALID & n->flags)
 171                 return(REW_NOHALT);
 172
 173         /* First: rewind to ourselves. */
 174         if (type == n->type && tok == n->tok)
 175                 return(REW_REWIND);
 176
 177         /*
 178          * Next follow the implicit scope-smashings as defined by man.7:
 179          * section, sub-section, etc.
 180          */
 181
 182         switch (tok) {
 183         case (MAN_SH):
 184                 break;
 185         case (MAN_SS):
 186                 /* Rewind to a section, if a block. */
 187                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 188                         return(c);
 189                 break;
 190         case (MAN_RS):
 191                 /* Rewind to a subsection, if a block. */
 192                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 193                         return(c);
 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         default:
 199                 /* Rewind to an offsetter, if a block. */
 200                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 201                         return(c);
 202                 /* Rewind to a subsection, if a block. */
 203                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 204                         return(c);
 205                 /* Rewind to a section, if a block. */
 206                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 207                         return(c);
 208                 break;
 209         }
 210
 211         return(REW_NOHALT);
 212 }
 213
 214
 215 /*
 216  * Rewinding entails ascending the parse tree until a coherent point,
 217  * for example, the `SH' macro will close out any intervening `SS'
 218  * scopes.  When a scope is closed, it must be validated and actioned.
 219  */
 220 static int
 221 rew_scope(enum man_type type, struct man *m, enum mant tok)
 222 {
 223         struct man_node *n;
 224         enum rew         c;
 225
 226         /* LINTED */
 227         for (n = m->last; n; n = n->parent) {
 228                 /*
 229                  * Whether we should stop immediately (REW_HALT), stop
 230                  * and rewind until this point (REW_REWIND), or keep
 231                  * rewinding (REW_NOHALT).
 232                  */
 233                 c = rew_dohalt(tok, type, n);
 234                 if (REW_HALT == c)
 235                         return(1);
 236                 if (REW_REWIND == c)
 237                         break;
 238         }
 239
 240         /*
 241          * Rewind until the current point.  Warn if we're a roff
 242          * instruction that's mowing over explicit scopes.
 243          */
 244         assert(n);
 245
 246         return(man_unscope(m, n, MANDOCERR_MAX));
 247 }
 248
 249
 250 /*
 251  * Close out a generic explicit macro.
 252  */
 253 /* ARGSUSED */
 254 int
 255 blk_close(MACRO_PROT_ARGS)
 256 {
 257         enum mant                ntok;
 258         const struct man_node   *nn;
 259
 260         switch (tok) {
 261         case (MAN_RE):
 262                 ntok = MAN_RS;
 263                 break;
 264         default:
 265                 abort();
 266                 /* NOTREACHED */
 267         }
 268
 269         for (nn = m->last->parent; nn; nn = nn->parent)
 270                 if (ntok == nn->tok)
 271                         break;
 272
 273         if (NULL == nn)
 274                 if ( ! man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE))
 275                         return(0);
 276
 277         if ( ! rew_scope(MAN_BODY, m, ntok))
 278                 return(0);
 279         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 280                 return(0);
 281
 282         return(1);
 283 }
 284
 285
 286 /* ARGSUSED */
 287 int
 288 blk_exp(MACRO_PROT_ARGS)
 289 {
 290         int              w, la;
 291         char            *p;
 292
 293         /*
 294          * Close out prior scopes.  "Regular" explicit macros cannot be
 295          * nested, but we allow roff macros to be placed just about
 296          * anywhere.
 297          */
 298
 299         if ( ! rew_scope(MAN_BODY, m, tok))
 300                 return(0);
 301         if ( ! rew_scope(MAN_BLOCK, m, tok))
 302                 return(0);
 303
 304         if ( ! man_block_alloc(m, line, ppos, tok))
 305                 return(0);
 306         if ( ! man_head_alloc(m, line, ppos, tok))
 307                 return(0);
 308
 309         for (;;) {
 310                 la = *pos;
 311                 w = man_args(m, line, pos, buf, &p);
 312
 313                 if (-1 == w)
 314                         return(0);
 315                 if (0 == w)
 316                         break;
 317
 318                 if ( ! man_word_alloc(m, line, la, p))
 319                         return(0);
 320         }
 321
 322         assert(m);
 323         assert(tok != MAN_MAX);
 324
 325         if ( ! rew_scope(MAN_HEAD, m, tok))
 326                 return(0);
 327         return(man_body_alloc(m, line, ppos, tok));
 328 }
 329
 330
 331
 332 /*
 333  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 334  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 335  * scopes, such as `SH' closing out an `SS', are defined in the rew
 336  * routines.
 337  */
 338 /* ARGSUSED */
 339 int
 340 blk_imp(MACRO_PROT_ARGS)
 341 {
 342         int              w, la;
 343         char            *p;
 344         struct man_node *n;
 345
 346         /* Close out prior scopes. */
 347
 348         if ( ! rew_scope(MAN_BODY, m, tok))
 349                 return(0);
 350         if ( ! rew_scope(MAN_BLOCK, m, tok))
 351                 return(0);
 352
 353         /* Allocate new block & head scope. */
 354
 355         if ( ! man_block_alloc(m, line, ppos, tok))
 356                 return(0);
 357         if ( ! man_head_alloc(m, line, ppos, tok))
 358                 return(0);
 359
 360         n = m->last;
 361
 362         /* Add line arguments. */
 363
 364         for (;;) {
 365                 la = *pos;
 366                 w = man_args(m, line, pos, buf, &p);
 367
 368                 if (-1 == w)
 369                         return(0);
 370                 if (0 == w)
 371                         break;
 372
 373                 if ( ! man_word_alloc(m, line, la, p))
 374                         return(0);
 375         }
 376
 377         /* Close out head and open body (unless MAN_SCOPE). */
 378
 379         if (MAN_SCOPED & man_macros[tok].flags) {
 380                 /* If we're forcing scope (`TP'), keep it open. */
 381                 if (MAN_FSCOPED & man_macros[tok].flags) {
 382                         m->flags |= MAN_BLINE;
 383                         return(1);
 384                 } else if (n == m->last) {
 385                         m->flags |= MAN_BLINE;
 386                         return(1);
 387                 }
 388         }
 389
 390         if ( ! rew_scope(MAN_HEAD, m, tok))
 391                 return(0);
 392         return(man_body_alloc(m, line, ppos, tok));
 393 }
 394
 395
 396 /* ARGSUSED */
 397 int
 398 in_line_eoln(MACRO_PROT_ARGS)
 399 {
 400         int              w, la;
 401         char            *p;
 402         struct man_node *n;
 403
 404         if ( ! man_elem_alloc(m, line, ppos, tok))
 405                 return(0);
 406
 407         n = m->last;
 408
 409         for (;;) {
 410                 la = *pos;
 411                 w = man_args(m, line, pos, buf, &p);
 412
 413                 if (-1 == w)
 414                         return(0);
 415                 if (0 == w)
 416                         break;
 417                 if ( ! man_word_alloc(m, line, la, p))
 418                         return(0);
 419         }
 420
 421         /*
 422          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 423          * next-line scoped), then set our mode to indicate that we're
 424          * waiting for terms to load into our context.
 425          */
 426
 427         if (n == m->last && MAN_SCOPED & man_macros[tok].flags) {
 428                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 429                 m->flags |= MAN_ELINE;
 430                 return(1);
 431         }
 432
 433         /* Set ignorable context, if applicable. */
 434
 435         if (MAN_NSCOPED & man_macros[tok].flags) {
 436                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 437                 m->flags |= MAN_ILINE;
 438         }
 439
 440         /*
 441          * Rewind our element scope.  Note that when TH is pruned, we'll
 442          * be back at the root, so make sure that we don't clobber as
 443          * its sibling.
 444          */
 445
 446         for ( ; m->last; m->last = m->last->parent) {
 447                 if (m->last == n)
 448                         break;
 449                 if (m->last->type == MAN_ROOT)
 450                         break;
 451                 if ( ! man_valid_post(m))
 452                         return(0);
 453         }
 454
 455         assert(m->last);
 456
 457         /*
 458          * Same here regarding whether we're back at the root.
 459          */
 460
 461         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 462                 return(0);
 463
 464         m->next = MAN_ROOT == m->last->type ?
 465                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 466
 467         return(1);
 468 }
 469
 470
 471 int
 472 man_macroend(struct man *m)
 473 {
 474
 475         return(man_unscope(m, m->first, MANDOCERR_SCOPEEXIT));
 476 }
 477