man_macro.c

   1 /*      $Id: man_macro.c,v 1.70 2011/12/04 00:44:12 schwarze 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         m->next = MAN_NEXT_SIBLING;
 124
 125         /* LINTED */
 126         while (m->last != to) {
 127                 /*
 128                  * Save the parent here, because we may delete the
 129                  * m->last node in the post-validation phase and reset
 130                  * it to m->last->parent, causing a step in the closing
 131                  * out to be lost.
 132                  */
 133                 n = m->last->parent;
 134                 rew_warn(m, m->last, er);
 135                 if ( ! man_valid_post(m))
 136                         return(0);
 137                 m->last = n;
 138                 assert(m->last);
 139         }
 140
 141         rew_warn(m, m->last, er);
 142         if ( ! man_valid_post(m))
 143                 return(0);
 144
 145         return(1);
 146 }
 147
 148
 149 static enum rew
 150 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 151 {
 152
 153         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 154                         MAN_BODY == n->parent->type)
 155                 return(REW_REWIND);
 156         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 157 }
 158
 159
 160 /*
 161  * There are three scope levels: scoped to the root (all), scoped to the
 162  * section (all less sections), and scoped to subsections (all less
 163  * sections and subsections).
 164  */
 165 static enum rew
 166 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 167 {
 168         enum rew         c;
 169
 170         /* We cannot progress beyond the root ever. */
 171         if (MAN_ROOT == n->type)
 172                 return(REW_HALT);
 173
 174         assert(n->parent);
 175
 176         /* Normal nodes shouldn't go to the level of the root. */
 177         if (MAN_ROOT == n->parent->type)
 178                 return(REW_REWIND);
 179
 180         /* Already-validated nodes should be closed out. */
 181         if (MAN_VALID & n->flags)
 182                 return(REW_NOHALT);
 183
 184         /* First: rewind to ourselves. */
 185         if (type == n->type && tok == n->tok)
 186                 return(REW_REWIND);
 187
 188         /*
 189          * Next follow the implicit scope-smashings as defined by man.7:
 190          * section, sub-section, etc.
 191          */
 192
 193         switch (tok) {
 194         case (MAN_SH):
 195                 break;
 196         case (MAN_SS):
 197                 /* Rewind to a section, if a block. */
 198                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 199                         return(c);
 200                 break;
 201         case (MAN_RS):
 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         default:
 210                 /* Rewind to an offsetter, if a block. */
 211                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 212                         return(c);
 213                 /* Rewind to a subsection, if a block. */
 214                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 215                         return(c);
 216                 /* Rewind to a section, if a block. */
 217                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 218                         return(c);
 219                 break;
 220         }
 221
 222         return(REW_NOHALT);
 223 }
 224
 225
 226 /*
 227  * Rewinding entails ascending the parse tree until a coherent point,
 228  * for example, the `SH' macro will close out any intervening `SS'
 229  * scopes.  When a scope is closed, it must be validated and actioned.
 230  */
 231 static int
 232 rew_scope(enum man_type type, struct man *m, enum mant tok)
 233 {
 234         struct man_node *n;
 235         enum rew         c;
 236
 237         /* LINTED */
 238         for (n = m->last; n; n = n->parent) {
 239                 /*
 240                  * Whether we should stop immediately (REW_HALT), stop
 241                  * and rewind until this point (REW_REWIND), or keep
 242                  * rewinding (REW_NOHALT).
 243                  */
 244                 c = rew_dohalt(tok, type, n);
 245                 if (REW_HALT == c)
 246                         return(1);
 247                 if (REW_REWIND == c)
 248                         break;
 249         }
 250
 251         /*
 252          * Rewind until the current point.  Warn if we're a roff
 253          * instruction that's mowing over explicit scopes.
 254          */
 255         assert(n);
 256
 257         return(man_unscope(m, n, MANDOCERR_MAX));
 258 }
 259
 260
 261 /*
 262  * Close out a generic explicit macro.
 263  */
 264 /* ARGSUSED */
 265 int
 266 blk_close(MACRO_PROT_ARGS)
 267 {
 268         enum mant                ntok;
 269         const struct man_node   *nn;
 270
 271         switch (tok) {
 272         case (MAN_RE):
 273                 ntok = MAN_RS;
 274                 break;
 275         default:
 276                 abort();
 277                 /* NOTREACHED */
 278         }
 279
 280         for (nn = m->last->parent; nn; nn = nn->parent)
 281                 if (ntok == nn->tok)
 282                         break;
 283
 284         if (NULL == nn)
 285                 man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE);
 286
 287         if ( ! rew_scope(MAN_BODY, m, ntok))
 288                 return(0);
 289         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 290                 return(0);
 291
 292         return(1);
 293 }
 294
 295
 296 /* ARGSUSED */
 297 int
 298 blk_exp(MACRO_PROT_ARGS)
 299 {
 300         int              la;
 301         char            *p;
 302
 303         /*
 304          * Close out prior scopes.  "Regular" explicit macros cannot be
 305          * nested, but we allow roff macros to be placed just about
 306          * anywhere.
 307          */
 308
 309         if ( ! man_block_alloc(m, line, ppos, tok))
 310                 return(0);
 311         if ( ! man_head_alloc(m, line, ppos, tok))
 312                 return(0);
 313
 314         for (;;) {
 315                 la = *pos;
 316                 if ( ! man_args(m, line, pos, buf, &p))
 317                         break;
 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              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                 if ( ! man_args(m, line, pos, buf, &p))
 367                         break;
 368                 if ( ! man_word_alloc(m, line, la, p))
 369                         return(0);
 370         }
 371
 372         /* Close out head and open body (unless MAN_SCOPE). */
 373
 374         if (MAN_SCOPED & man_macros[tok].flags) {
 375                 /* If we're forcing scope (`TP'), keep it open. */
 376                 if (MAN_FSCOPED & man_macros[tok].flags) {
 377                         m->flags |= MAN_BLINE;
 378                         return(1);
 379                 } else if (n == m->last) {
 380                         m->flags |= MAN_BLINE;
 381                         return(1);
 382                 }
 383         }
 384
 385         if ( ! rew_scope(MAN_HEAD, m, tok))
 386                 return(0);
 387         return(man_body_alloc(m, line, ppos, tok));
 388 }
 389
 390
 391 /* ARGSUSED */
 392 int
 393 in_line_eoln(MACRO_PROT_ARGS)
 394 {
 395         int              la;
 396         char            *p;
 397         struct man_node *n;
 398
 399         if ( ! man_elem_alloc(m, line, ppos, tok))
 400                 return(0);
 401
 402         n = m->last;
 403
 404         for (;;) {
 405                 la = *pos;
 406                 if ( ! man_args(m, line, pos, buf, &p))
 407                         break;
 408                 if ( ! man_word_alloc(m, line, la, p))
 409                         return(0);
 410         }
 411
 412         /*
 413          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 414          * next-line scoped), then set our mode to indicate that we're
 415          * waiting for terms to load into our context.
 416          */
 417
 418         if (n == m->last && MAN_SCOPED & man_macros[tok].flags) {
 419                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 420                 m->flags |= MAN_ELINE;
 421                 return(1);
 422         }
 423
 424         /* Set ignorable context, if applicable. */
 425
 426         if (MAN_NSCOPED & man_macros[tok].flags) {
 427                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 428                 m->flags |= MAN_ILINE;
 429         }
 430
 431         assert(MAN_ROOT != m->last->type);
 432         m->next = MAN_NEXT_SIBLING;
 433
 434         /*
 435          * Rewind our element scope.  Note that when TH is pruned, we'll
 436          * be back at the root, so make sure that we don't clobber as
 437          * its sibling.
 438          */
 439
 440         for ( ; m->last; m->last = m->last->parent) {
 441                 if (m->last == n)
 442                         break;
 443                 if (m->last->type == MAN_ROOT)
 444                         break;
 445                 if ( ! man_valid_post(m))
 446                         return(0);
 447         }
 448
 449         assert(m->last);
 450
 451         /*
 452          * Same here regarding whether we're back at the root.
 453          */
 454
 455         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 456                 return(0);
 457
 458         return(1);
 459 }
 460
 461
 462 int
 463 man_macroend(struct man *m)
 464 {
 465
 466         return(man_unscope(m, m->first, MANDOCERR_SCOPEEXIT));
 467 }
 468
 469 static int
 470 man_args(struct man *m, int line, int *pos, char *buf, char **v)
 471 {
 472         char     *start;
 473
 474         assert(*pos);
 475         *v = start = buf + *pos;
 476         assert(' ' != *start);
 477
 478         if ('\0' == *start)
 479                 return(0);
 480
 481         *v = mandoc_getarg(m->parse, v, line, pos);
 482         return(1);
 483 }