man_macro.c

   1 /*      $Id: man_macro.c,v 1.46 2010/05/17 22:11:42 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
   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, 0 }, /* i */
  73         { in_line_eoln, MAN_NSCOPED }, /* sp */
  74         { in_line_eoln, 0 }, /* nf */
  75         { in_line_eoln, 0 }, /* fi */
  76         { in_line_eoln, 0 }, /* r */
  77         { blk_close, 0 }, /* RE */
  78         { blk_exp, MAN_EXPLICIT }, /* RS */
  79         { in_line_eoln, 0 }, /* DT */
  80         { in_line_eoln, 0 }, /* UC */
  81         { in_line_eoln, 0 }, /* PD */
  82         { in_line_eoln, MAN_NSCOPED }, /* Sp */
  83         { in_line_eoln, 0 }, /* Vb */
  84         { in_line_eoln, 0 }, /* Ve */
  85         { in_line_eoln, 0 }, /* AT */
  86 };
  87
  88 const   struct man_macro * const man_macros = __man_macros;
  89
  90
  91 /*
  92  * Warn when "n" is an explicit non-roff macro.
  93  */
  94 static int
  95 rew_warn(struct man *m, struct man_node *n, enum mandocerr er)
  96 {
  97
  98         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
  99                 return(1);
 100         if (MAN_VALID & n->flags)
 101                 return(1);
 102         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 103                 return(1);
 104         return(man_nmsg(m, n, er));
 105 }
 106
 107
 108 /*
 109  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 110  * will be used if an explicit block scope is being closed out.
 111  */
 112 int
 113 man_unscope(struct man *m, const struct man_node *n,
 114                 enum mandocerr er)
 115 {
 116
 117         assert(n);
 118
 119         /* LINTED */
 120         while (m->last != n) {
 121                 if ( ! rew_warn(m, m->last, er))
 122                         return(0);
 123                 if ( ! man_valid_post(m))
 124                         return(0);
 125                 if ( ! man_action_post(m))
 126                         return(0);
 127                 m->last = m->last->parent;
 128                 assert(m->last);
 129         }
 130
 131         if ( ! rew_warn(m, m->last, er))
 132                 return(0);
 133         if ( ! man_valid_post(m))
 134                 return(0);
 135         if ( ! man_action_post(m))
 136                 return(0);
 137
 138         m->next = MAN_ROOT == m->last->type ?
 139                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 140
 141         return(1);
 142 }
 143
 144
 145 static enum rew
 146 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 147 {
 148
 149         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 150                         MAN_BODY == n->parent->type)
 151                 return(REW_REWIND);
 152         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 153 }
 154
 155
 156 /*
 157  * There are three scope levels: scoped to the root (all), scoped to the
 158  * section (all less sections), and scoped to subsections (all less
 159  * sections and subsections).
 160  */
 161 static enum rew
 162 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 163 {
 164         enum rew         c;
 165
 166         /* We cannot progress beyond the root ever. */
 167         if (MAN_ROOT == n->type)
 168                 return(REW_HALT);
 169
 170         assert(n->parent);
 171
 172         /* Normal nodes shouldn't go to the level of the root. */
 173         if (MAN_ROOT == n->parent->type)
 174                 return(REW_REWIND);
 175
 176         /* Already-validated nodes should be closed out. */
 177         if (MAN_VALID & n->flags)
 178                 return(REW_NOHALT);
 179
 180         /* First: rewind to ourselves. */
 181         if (type == n->type && tok == n->tok)
 182                 return(REW_REWIND);
 183
 184         /*
 185          * Next follow the implicit scope-smashings as defined by man.7:
 186          * section, sub-section, etc.
 187          */
 188
 189         switch (tok) {
 190         case (MAN_SH):
 191                 break;
 192         case (MAN_SS):
 193                 /* Rewind to a section, if a block. */
 194                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 195                         return(c);
 196                 break;
 197         case (MAN_RS):
 198                 /* Rewind to a subsection, if a block. */
 199                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 200                         return(c);
 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         default:
 206                 /* Rewind to an offsetter, if a block. */
 207                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 208                         return(c);
 209                 /* Rewind to a subsection, if a block. */
 210                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 211                         return(c);
 212                 /* Rewind to a section, if a block. */
 213                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 214                         return(c);
 215                 break;
 216         }
 217
 218         return(REW_NOHALT);
 219 }
 220
 221
 222 /*
 223  * Rewinding entails ascending the parse tree until a coherent point,
 224  * for example, the `SH' macro will close out any intervening `SS'
 225  * scopes.  When a scope is closed, it must be validated and actioned.
 226  */
 227 static int
 228 rew_scope(enum man_type type, struct man *m, enum mant tok)
 229 {
 230         struct man_node *n;
 231         enum rew         c;
 232
 233         /* LINTED */
 234         for (n = m->last; n; n = n->parent) {
 235                 /*
 236                  * Whether we should stop immediately (REW_HALT), stop
 237                  * and rewind until this point (REW_REWIND), or keep
 238                  * rewinding (REW_NOHALT).
 239                  */
 240                 c = rew_dohalt(tok, type, n);
 241                 if (REW_HALT == c)
 242                         return(1);
 243                 if (REW_REWIND == c)
 244                         break;
 245         }
 246
 247         /*
 248          * Rewind until the current point.  Warn if we're a roff
 249          * instruction that's mowing over explicit scopes.
 250          */
 251         assert(n);
 252
 253         return(man_unscope(m, n, MANDOCERR_MAX));
 254 }
 255
 256
 257 /*
 258  * Close out a generic explicit macro.
 259  */
 260 /* ARGSUSED */
 261 int
 262 blk_close(MACRO_PROT_ARGS)
 263 {
 264         enum mant                ntok;
 265         const struct man_node   *nn;
 266
 267         switch (tok) {
 268         case (MAN_RE):
 269                 ntok = MAN_RS;
 270                 break;
 271         default:
 272                 abort();
 273                 /* NOTREACHED */
 274         }
 275
 276         for (nn = m->last->parent; nn; nn = nn->parent)
 277                 if (ntok == nn->tok)
 278                         break;
 279
 280         if (NULL == nn)
 281                 if ( ! man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE))
 282                         return(0);
 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 int
 294 blk_exp(MACRO_PROT_ARGS)
 295 {
 296         int              w, la;
 297         char            *p;
 298
 299         /*
 300          * Close out prior scopes.  "Regular" explicit macros cannot be
 301          * nested, but we allow roff macros to be placed just about
 302          * anywhere.
 303          */
 304
 305         if ( ! rew_scope(MAN_BODY, m, tok))
 306                 return(0);
 307         if ( ! rew_scope(MAN_BLOCK, m, tok))
 308                 return(0);
 309
 310         if ( ! man_block_alloc(m, line, ppos, tok))
 311                 return(0);
 312         if ( ! man_head_alloc(m, line, ppos, tok))
 313                 return(0);
 314
 315         for (;;) {
 316                 la = *pos;
 317                 w = man_args(m, line, pos, buf, &p);
 318
 319                 if (-1 == w)
 320                         return(0);
 321                 if (0 == w)
 322                         break;
 323
 324                 if ( ! man_word_alloc(m, line, la, p))
 325                         return(0);
 326         }
 327
 328         assert(m);
 329         assert(tok != MAN_MAX);
 330
 331         if ( ! rew_scope(MAN_HEAD, m, tok))
 332                 return(0);
 333         return(man_body_alloc(m, line, ppos, tok));
 334 }
 335
 336
 337
 338 /*
 339  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 340  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 341  * scopes, such as `SH' closing out an `SS', are defined in the rew
 342  * routines.
 343  */
 344 int
 345 blk_imp(MACRO_PROT_ARGS)
 346 {
 347         int              w, la;
 348         char            *p;
 349         struct man_node *n;
 350
 351         /* Close out prior scopes. */
 352
 353         if ( ! rew_scope(MAN_BODY, m, tok))
 354                 return(0);
 355         if ( ! rew_scope(MAN_BLOCK, m, tok))
 356                 return(0);
 357
 358         /* Allocate new block & head scope. */
 359
 360         if ( ! man_block_alloc(m, line, ppos, tok))
 361                 return(0);
 362         if ( ! man_head_alloc(m, line, ppos, tok))
 363                 return(0);
 364
 365         n = m->last;
 366
 367         /* Add line arguments. */
 368
 369         for (;;) {
 370                 la = *pos;
 371                 w = man_args(m, line, pos, buf, &p);
 372
 373                 if (-1 == w)
 374                         return(0);
 375                 if (0 == w)
 376                         break;
 377
 378                 if ( ! man_word_alloc(m, line, la, p))
 379                         return(0);
 380         }
 381
 382         /* Close out head and open body (unless MAN_SCOPE). */
 383
 384         if (MAN_SCOPED & man_macros[tok].flags) {
 385                 /* If we're forcing scope (`TP'), keep it open. */
 386                 if (MAN_FSCOPED & man_macros[tok].flags) {
 387                         m->flags |= MAN_BLINE;
 388                         return(1);
 389                 } else if (n == m->last) {
 390                         m->flags |= MAN_BLINE;
 391                         return(1);
 392                 }
 393         }
 394
 395         if ( ! rew_scope(MAN_HEAD, m, tok))
 396                 return(0);
 397         return(man_body_alloc(m, line, ppos, tok));
 398 }
 399
 400
 401 int
 402 in_line_eoln(MACRO_PROT_ARGS)
 403 {
 404         int              w, la;
 405         char            *p;
 406         struct man_node *n;
 407
 408         if ( ! man_elem_alloc(m, line, ppos, tok))
 409                 return(0);
 410
 411         n = m->last;
 412
 413         for (;;) {
 414                 la = *pos;
 415                 w = man_args(m, line, pos, buf, &p);
 416
 417                 if (-1 == w)
 418                         return(0);
 419                 if (0 == w)
 420                         break;
 421                 if ( ! man_word_alloc(m, line, la, p))
 422                         return(0);
 423         }
 424
 425         /*
 426          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 427          * next-line scoped), then set our mode to indicate that we're
 428          * waiting for terms to load into our context.
 429          */
 430
 431         if (n == m->last && MAN_SCOPED & man_macros[tok].flags) {
 432                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 433                 m->flags |= MAN_ELINE;
 434                 return(1);
 435         }
 436
 437         /* Set ignorable context, if applicable. */
 438
 439         if (MAN_NSCOPED & man_macros[tok].flags) {
 440                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 441                 m->flags |= MAN_ILINE;
 442         }
 443
 444         /*
 445          * Rewind our element scope.  Note that when TH is pruned, we'll
 446          * be back at the root, so make sure that we don't clobber as
 447          * its sibling.
 448          */
 449
 450         for ( ; m->last; m->last = m->last->parent) {
 451                 if (m->last == n)
 452                         break;
 453                 if (m->last->type == MAN_ROOT)
 454                         break;
 455                 if ( ! man_valid_post(m))
 456                         return(0);
 457                 if ( ! man_action_post(m))
 458                         return(0);
 459         }
 460
 461         assert(m->last);
 462
 463         /*
 464          * Same here regarding whether we're back at the root.
 465          */
 466
 467         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 468                 return(0);
 469         if (m->last->type != MAN_ROOT && ! man_action_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