man_macro.c

   1 /*      $Id: man_macro.c,v 1.68 2011/11/18 16:43:29 joerg Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 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 "man.h"
  27 #include "mandoc.h"
  28 #include "libmandoc.h"
  29 #include "libman.h"
  30
  31 enum    rew {
  32         REW_REWIND,
  33         REW_NOHALT,
  34         REW_HALT
  35 };
  36
  37 static  int              blk_close(MACRO_PROT_ARGS);
  38 static  int              blk_exp(MACRO_PROT_ARGS);
  39 static  int              blk_imp(MACRO_PROT_ARGS);
  40 static  int              in_line_eoln(MACRO_PROT_ARGS);
  41 static  int              man_args(struct man *, int,
  42                                 int *, char *, char **);
  43
  44 static  int              rew_scope(enum man_type,
  45                                 struct man *, enum mant);
  46 static  enum rew         rew_dohalt(enum mant, enum man_type,
  47                                 const struct man_node *);
  48 static  enum rew         rew_block(enum mant, enum man_type,
  49                                 const struct man_node *);
  50 static  void             rew_warn(struct man *,
  51                                 struct man_node *, enum mandocerr);
  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_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 };
  88
  89 const   struct man_macro * const man_macros = __man_macros;
  90
  91
  92 /*
  93  * Warn when "n" is an explicit non-roff macro.
  94  */
  95 static void
  96 rew_warn(struct man *m, struct man_node *n, enum mandocerr er)
  97 {
  98
  99         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
 100                 return;
 101         if (MAN_VALID & n->flags)
 102                 return;
 103         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 104                 return;
 105
 106         assert(er < MANDOCERR_FATAL);
 107         man_nmsg(m, n, er);
 108 }
 109
 110
 111 /*
 112  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 113  * will be used if an explicit block scope is being closed out.
 114  */
 115 int
 116 man_unscope(struct man *m, const struct man_node *to,
 117                 enum mandocerr er)
 118 {
 119         struct man_node *n;
 120
 121         assert(to);
 122
 123         if (MAN_ROOT != m->last->type)
 124                 return(1);
 125         m->next = MAN_NEXT_SIBLING;
 126
 127         /* LINTED */
 128         while (m->last != to) {
 129                 /*
 130                  * Save the parent here, because we may delete the
 131                  * m->last node in the post-validation phase and reset
 132                  * it to m->last->parent, causing a step in the closing
 133                  * out to be lost.
 134                  */
 135                 n = m->last->parent;
 136                 rew_warn(m, m->last, er);
 137                 if ( ! man_valid_post(m))
 138                         return(0);
 139                 m->last = n;
 140                 assert(m->last);
 141         }
 142
 143         rew_warn(m, m->last, er);
 144         if ( ! man_valid_post(m))
 145                 return(0);
 146
 147         return(1);
 148 }
 149
 150
 151 static enum rew
 152 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 153 {
 154
 155         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 156                         MAN_BODY == n->parent->type)
 157                 return(REW_REWIND);
 158         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 159 }
 160
 161
 162 /*
 163  * There are three scope levels: scoped to the root (all), scoped to the
 164  * section (all less sections), and scoped to subsections (all less
 165  * sections and subsections).
 166  */
 167 static enum rew
 168 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 169 {
 170         enum rew         c;
 171
 172         /* We cannot progress beyond the root ever. */
 173         if (MAN_ROOT == n->type)
 174                 return(REW_HALT);
 175
 176         assert(n->parent);
 177
 178         /* Normal nodes shouldn't go to the level of the root. */
 179         if (MAN_ROOT == n->parent->type)
 180                 return(REW_REWIND);
 181
 182         /* Already-validated nodes should be closed out. */
 183         if (MAN_VALID & n->flags)
 184                 return(REW_NOHALT);
 185
 186         /* First: rewind to ourselves. */
 187         if (type == n->type && tok == n->tok)
 188                 return(REW_REWIND);
 189
 190         /*
 191          * Next follow the implicit scope-smashings as defined by man.7:
 192          * section, sub-section, etc.
 193          */
 194
 195         switch (tok) {
 196         case (MAN_SH):
 197                 break;
 198         case (MAN_SS):
 199                 /* Rewind to a section, if a block. */
 200                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 201                         return(c);
 202                 break;
 203         case (MAN_RS):
 204                 /* Rewind to a subsection, if a block. */
 205                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 206                         return(c);
 207                 /* Rewind to a section, if a block. */
 208                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 209                         return(c);
 210                 break;
 211         default:
 212                 /* Rewind to an offsetter, if a block. */
 213                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 214                         return(c);
 215                 /* Rewind to a subsection, if a block. */
 216                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 217                         return(c);
 218                 /* Rewind to a section, if a block. */
 219                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 220                         return(c);
 221                 break;
 222         }
 223
 224         return(REW_NOHALT);
 225 }
 226
 227
 228 /*
 229  * Rewinding entails ascending the parse tree until a coherent point,
 230  * for example, the `SH' macro will close out any intervening `SS'
 231  * scopes.  When a scope is closed, it must be validated and actioned.
 232  */
 233 static int
 234 rew_scope(enum man_type type, struct man *m, enum mant tok)
 235 {
 236         struct man_node *n;
 237         enum rew         c;
 238
 239         /* LINTED */
 240         for (n = m->last; n; n = n->parent) {
 241                 /*
 242                  * Whether we should stop immediately (REW_HALT), stop
 243                  * and rewind until this point (REW_REWIND), or keep
 244                  * rewinding (REW_NOHALT).
 245                  */
 246                 c = rew_dohalt(tok, type, n);
 247                 if (REW_HALT == c)
 248                         return(1);
 249                 if (REW_REWIND == c)
 250                         break;
 251         }
 252
 253         /*
 254          * Rewind until the current point.  Warn if we're a roff
 255          * instruction that's mowing over explicit scopes.
 256          */
 257         assert(n);
 258
 259         return(man_unscope(m, n, MANDOCERR_MAX));
 260 }
 261
 262
 263 /*
 264  * Close out a generic explicit macro.
 265  */
 266 /* ARGSUSED */
 267 int
 268 blk_close(MACRO_PROT_ARGS)
 269 {
 270         enum mant                ntok;
 271         const struct man_node   *nn;
 272
 273         switch (tok) {
 274         case (MAN_RE):
 275                 ntok = MAN_RS;
 276                 break;
 277         default:
 278                 abort();
 279                 /* NOTREACHED */
 280         }
 281
 282         for (nn = m->last->parent; nn; nn = nn->parent)
 283                 if (ntok == nn->tok)
 284                         break;
 285
 286         if (NULL == nn)
 287                 man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE);
 288
 289         if ( ! rew_scope(MAN_BODY, m, ntok))
 290                 return(0);
 291         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 292                 return(0);
 293
 294         return(1);
 295 }
 296
 297
 298 /* ARGSUSED */
 299 int
 300 blk_exp(MACRO_PROT_ARGS)
 301 {
 302         int              la;
 303         char            *p;
 304
 305         /*
 306          * Close out prior scopes.  "Regular" explicit macros cannot be
 307          * nested, but we allow roff macros to be placed just about
 308          * anywhere.
 309          */
 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                 if ( ! man_args(m, line, pos, buf, &p))
 319                         break;
 320                 if ( ! man_word_alloc(m, line, la, p))
 321                         return(0);
 322         }
 323
 324         assert(m);
 325         assert(tok != MAN_MAX);
 326
 327         if ( ! rew_scope(MAN_HEAD, m, tok))
 328                 return(0);
 329         return(man_body_alloc(m, line, ppos, tok));
 330 }
 331
 332
 333
 334 /*
 335  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 336  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 337  * scopes, such as `SH' closing out an `SS', are defined in the rew
 338  * routines.
 339  */
 340 /* ARGSUSED */
 341 int
 342 blk_imp(MACRO_PROT_ARGS)
 343 {
 344         int              la;
 345         char            *p;
 346         struct man_node *n;
 347
 348         /* Close out prior scopes. */
 349
 350         if ( ! rew_scope(MAN_BODY, m, tok))
 351                 return(0);
 352         if ( ! rew_scope(MAN_BLOCK, m, tok))
 353                 return(0);
 354
 355         /* Allocate new block & head scope. */
 356
 357         if ( ! man_block_alloc(m, line, ppos, tok))
 358                 return(0);
 359         if ( ! man_head_alloc(m, line, ppos, tok))
 360                 return(0);
 361
 362         n = m->last;
 363
 364         /* Add line arguments. */
 365
 366         for (;;) {
 367                 la = *pos;
 368                 if ( ! man_args(m, line, pos, buf, &p))
 369                         break;
 370                 if ( ! man_word_alloc(m, line, la, p))
 371                         return(0);
 372         }
 373
 374         /* Close out head and open body (unless MAN_SCOPE). */
 375
 376         if (MAN_SCOPED & man_macros[tok].flags) {
 377                 /* If we're forcing scope (`TP'), keep it open. */
 378                 if (MAN_FSCOPED & man_macros[tok].flags) {
 379                         m->flags |= MAN_BLINE;
 380                         return(1);
 381                 } else if (n == m->last) {
 382                         m->flags |= MAN_BLINE;
 383                         return(1);
 384                 }
 385         }
 386
 387         if ( ! rew_scope(MAN_HEAD, m, tok))
 388                 return(0);
 389         return(man_body_alloc(m, line, ppos, tok));
 390 }
 391
 392
 393 /* ARGSUSED */
 394 int
 395 in_line_eoln(MACRO_PROT_ARGS)
 396 {
 397         int              la;
 398         char            *p;
 399         struct man_node *n;
 400
 401         if ( ! man_elem_alloc(m, line, ppos, tok))
 402                 return(0);
 403
 404         n = m->last;
 405
 406         for (;;) {
 407                 la = *pos;
 408                 if ( ! man_args(m, line, pos, buf, &p))
 409                         break;
 410                 if ( ! man_word_alloc(m, line, la, p))
 411                         return(0);
 412         }
 413
 414         /*
 415          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 416          * next-line scoped), then set our mode to indicate that we're
 417          * waiting for terms to load into our context.
 418          */
 419
 420         if (n == m->last && MAN_SCOPED & man_macros[tok].flags) {
 421                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 422                 m->flags |= MAN_ELINE;
 423                 return(1);
 424         }
 425
 426         /* Set ignorable context, if applicable. */
 427
 428         if (MAN_NSCOPED & man_macros[tok].flags) {
 429                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 430                 m->flags |= MAN_ILINE;
 431         }
 432
 433         assert(MAN_ROOT != m->last->type);
 434         m->next = MAN_NEXT_SIBLING;
 435
 436         /*
 437          * Rewind our element scope.  Note that when TH is pruned, we'll
 438          * be back at the root, so make sure that we don't clobber as
 439          * its sibling.
 440          */
 441
 442         for ( ; m->last; m->last = m->last->parent) {
 443                 if (m->last == n)
 444                         break;
 445                 if (m->last->type == MAN_ROOT)
 446                         break;
 447                 if ( ! man_valid_post(m))
 448                         return(0);
 449         }
 450
 451         assert(m->last);
 452
 453         /*
 454          * Same here regarding whether we're back at the root.
 455          */
 456
 457         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 458                 return(0);
 459
 460         return(1);
 461 }
 462
 463
 464 int
 465 man_macroend(struct man *m)
 466 {
 467
 468         return(man_unscope(m, m->first, MANDOCERR_SCOPEEXIT));
 469 }
 470
 471 static int
 472 man_args(struct man *m, int line, int *pos, char *buf, char **v)
 473 {
 474         char     *start;
 475
 476         assert(*pos);
 477         *v = start = buf + *pos;
 478         assert(' ' != *start);
 479
 480         if ('\0' == *start)
 481                 return(0);
 482
 483         *v = mandoc_getarg(m->parse, v, line, pos);
 484         return(1);
 485 }