man_macro.c

   1 /*      $Id: man_macro.c,v 1.56 2011/03/17 11:56:17 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  void             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 void
  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;
  97         if (MAN_VALID & n->flags)
  98                 return;
  99         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 100                 return;
 101
 102         assert(er < MANDOCERR_FATAL);
 103         man_nmsg(m, n, er);
 104 }
 105
 106
 107 /*
 108  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 109  * will be used if an explicit block scope is being closed out.
 110  */
 111 int
 112 man_unscope(struct man *m, const struct man_node *to,
 113                 enum mandocerr er)
 114 {
 115         struct man_node *n;
 116
 117         assert(to);
 118
 119         /* LINTED */
 120         while (m->last != to) {
 121                 /*
 122                  * Save the parent here, because we may delete the
 123                  * m->last node in the post-validation phase and reset
 124                  * it to m->last->parent, causing a step in the closing
 125                  * out to be lost.
 126                  */
 127                 n = m->last->parent;
 128                 rew_warn(m, m->last, er);
 129                 if ( ! man_valid_post(m))
 130                         return(0);
 131                 m->last = n;
 132                 assert(m->last);
 133         }
 134
 135         rew_warn(m, m->last, er);
 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                 man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE);
 283
 284         if ( ! rew_scope(MAN_BODY, m, ntok))
 285                 return(0);
 286         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 287                 return(0);
 288
 289         return(1);
 290 }
 291
 292
 293 /* ARGSUSED */
 294 int
 295 blk_exp(MACRO_PROT_ARGS)
 296 {
 297         int              w, la;
 298         char            *p;
 299
 300         /*
 301          * Close out prior scopes.  "Regular" explicit macros cannot be
 302          * nested, but we allow roff macros to be placed just about
 303          * anywhere.
 304          */
 305
 306         if ( ! rew_scope(MAN_BODY, m, tok))
 307                 return(0);
 308         if ( ! rew_scope(MAN_BLOCK, m, tok))
 309                 return(0);
 310
 311         if ( ! man_block_alloc(m, line, ppos, tok))
 312                 return(0);
 313         if ( ! man_head_alloc(m, line, ppos, tok))
 314                 return(0);
 315
 316         for (;;) {
 317                 la = *pos;
 318                 w = man_args(m, line, pos, buf, &p);
 319
 320                 if (-1 == w)
 321                         return(0);
 322                 if (0 == w)
 323                         break;
 324
 325                 if ( ! man_word_alloc(m, line, la, p))
 326                         return(0);
 327         }
 328
 329         assert(m);
 330         assert(tok != MAN_MAX);
 331
 332         if ( ! rew_scope(MAN_HEAD, m, tok))
 333                 return(0);
 334         return(man_body_alloc(m, line, ppos, tok));
 335 }
 336
 337
 338
 339 /*
 340  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 341  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 342  * scopes, such as `SH' closing out an `SS', are defined in the rew
 343  * routines.
 344  */
 345 /* ARGSUSED */
 346 int
 347 blk_imp(MACRO_PROT_ARGS)
 348 {
 349         int              w, la;
 350         char            *p;
 351         struct man_node *n;
 352
 353         /* Close out prior scopes. */
 354
 355         if ( ! rew_scope(MAN_BODY, m, tok))
 356                 return(0);
 357         if ( ! rew_scope(MAN_BLOCK, m, tok))
 358                 return(0);
 359
 360         /* Allocate new block & head scope. */
 361
 362         if ( ! man_block_alloc(m, line, ppos, tok))
 363                 return(0);
 364         if ( ! man_head_alloc(m, line, ppos, tok))
 365                 return(0);
 366
 367         n = m->last;
 368
 369         /* Add line arguments. */
 370
 371         for (;;) {
 372                 la = *pos;
 373                 w = man_args(m, line, pos, buf, &p);
 374
 375                 if (-1 == w)
 376                         return(0);
 377                 if (0 == w)
 378                         break;
 379
 380                 if ( ! man_word_alloc(m, line, la, p))
 381                         return(0);
 382         }
 383
 384         /* Close out head and open body (unless MAN_SCOPE). */
 385
 386         if (MAN_SCOPED & man_macros[tok].flags) {
 387                 /* If we're forcing scope (`TP'), keep it open. */
 388                 if (MAN_FSCOPED & man_macros[tok].flags) {
 389                         m->flags |= MAN_BLINE;
 390                         return(1);
 391                 } else if (n == m->last) {
 392                         m->flags |= MAN_BLINE;
 393                         return(1);
 394                 }
 395         }
 396
 397         if ( ! rew_scope(MAN_HEAD, m, tok))
 398                 return(0);
 399         return(man_body_alloc(m, line, ppos, tok));
 400 }
 401
 402
 403 /* ARGSUSED */
 404 int
 405 in_line_eoln(MACRO_PROT_ARGS)
 406 {
 407         int              w, la;
 408         char            *p;
 409         struct man_node *n;
 410
 411         if ( ! man_elem_alloc(m, line, ppos, tok))
 412                 return(0);
 413
 414         n = m->last;
 415
 416         for (;;) {
 417                 la = *pos;
 418                 w = man_args(m, line, pos, buf, &p);
 419
 420                 if (-1 == w)
 421                         return(0);
 422                 if (0 == w)
 423                         break;
 424                 if ( ! man_word_alloc(m, line, la, p))
 425                         return(0);
 426         }
 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 == m->last && MAN_SCOPED & man_macros[tok].flags) {
 435                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 436                 m->flags |= MAN_ELINE;
 437                 return(1);
 438         }
 439
 440         /* Set ignorable context, if applicable. */
 441
 442         if (MAN_NSCOPED & man_macros[tok].flags) {
 443                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 444                 m->flags |= MAN_ILINE;
 445         }
 446
 447         /*
 448          * Rewind our element scope.  Note that when TH is pruned, we'll
 449          * be back at the root, so make sure that we don't clobber as
 450          * its sibling.
 451          */
 452
 453         for ( ; m->last; m->last = m->last->parent) {
 454                 if (m->last == n)
 455                         break;
 456                 if (m->last->type == MAN_ROOT)
 457                         break;
 458                 if ( ! man_valid_post(m))
 459                         return(0);
 460         }
 461
 462         assert(m->last);
 463
 464         /*
 465          * Same here regarding whether we're back at the root.
 466          */
 467
 468         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 469                 return(0);
 470
 471         m->next = MAN_ROOT == m->last->type ?
 472                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 473
 474         return(1);
 475 }
 476
 477
 478 int
 479 man_macroend(struct man *m)
 480 {
 481
 482         return(man_unscope(m, m->first, MANDOCERR_SCOPEEXIT));
 483 }
 484