man_macro.c

   1 /*      $Id: man_macro.c,v 1.52 2010/12/05 16:14:16 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, 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, 0 }, /* AT */
  83         { in_line_eoln, 0 }, /* in */
  84 };
  85
  86 const   struct man_macro * const man_macros = __man_macros;
  87
  88
  89 /*
  90  * Warn when "n" is an explicit non-roff macro.
  91  */
  92 static int
  93 rew_warn(struct man *m, struct man_node *n, enum mandocerr er)
  94 {
  95
  96         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
  97                 return(1);
  98         if (MAN_VALID & n->flags)
  99                 return(1);
 100         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 101                 return(1);
 102         return(man_nmsg(m, n, er));
 103 }
 104
 105
 106 /*
 107  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 108  * will be used if an explicit block scope is being closed out.
 109  */
 110 int
 111 man_unscope(struct man *m, const struct man_node *n,
 112                 enum mandocerr er)
 113 {
 114
 115         assert(n);
 116
 117         /* LINTED */
 118         while (m->last != n) {
 119                 if ( ! rew_warn(m, m->last, er))
 120                         return(0);
 121                 if ( ! man_valid_post(m))
 122                         return(0);
 123                 m->last = m->last->parent;
 124                 assert(m->last);
 125         }
 126
 127         if ( ! rew_warn(m, m->last, er))
 128                 return(0);
 129         if ( ! man_valid_post(m))
 130                 return(0);
 131
 132         m->next = MAN_ROOT == m->last->type ?
 133                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 134
 135         return(1);
 136 }
 137
 138
 139 static enum rew
 140 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 141 {
 142
 143         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 144                         MAN_BODY == n->parent->type)
 145                 return(REW_REWIND);
 146         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 147 }
 148
 149
 150 /*
 151  * There are three scope levels: scoped to the root (all), scoped to the
 152  * section (all less sections), and scoped to subsections (all less
 153  * sections and subsections).
 154  */
 155 static enum rew
 156 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 157 {
 158         enum rew         c;
 159
 160         /* We cannot progress beyond the root ever. */
 161         if (MAN_ROOT == n->type)
 162                 return(REW_HALT);
 163
 164         assert(n->parent);
 165
 166         /* Normal nodes shouldn't go to the level of the root. */
 167         if (MAN_ROOT == n->parent->type)
 168                 return(REW_REWIND);
 169
 170         /* Already-validated nodes should be closed out. */
 171         if (MAN_VALID & n->flags)
 172                 return(REW_NOHALT);
 173
 174         /* First: rewind to ourselves. */
 175         if (type == n->type && tok == n->tok)
 176                 return(REW_REWIND);
 177
 178         /*
 179          * Next follow the implicit scope-smashings as defined by man.7:
 180          * section, sub-section, etc.
 181          */
 182
 183         switch (tok) {
 184         case (MAN_SH):
 185                 break;
 186         case (MAN_SS):
 187                 /* Rewind to a section, if a block. */
 188                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 189                         return(c);
 190                 break;
 191         case (MAN_RS):
 192                 /* Rewind to a subsection, if a block. */
 193                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 194                         return(c);
 195                 /* Rewind to a section, if a block. */
 196                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 197                         return(c);
 198                 break;
 199         default:
 200                 /* Rewind to an offsetter, if a block. */
 201                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 202                         return(c);
 203                 /* Rewind to a subsection, if a block. */
 204                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 205                         return(c);
 206                 /* Rewind to a section, if a block. */
 207                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 208                         return(c);
 209                 break;
 210         }
 211
 212         return(REW_NOHALT);
 213 }
 214
 215
 216 /*
 217  * Rewinding entails ascending the parse tree until a coherent point,
 218  * for example, the `SH' macro will close out any intervening `SS'
 219  * scopes.  When a scope is closed, it must be validated and actioned.
 220  */
 221 static int
 222 rew_scope(enum man_type type, struct man *m, enum mant tok)
 223 {
 224         struct man_node *n;
 225         enum rew         c;
 226
 227         /* LINTED */
 228         for (n = m->last; n; n = n->parent) {
 229                 /*
 230                  * Whether we should stop immediately (REW_HALT), stop
 231                  * and rewind until this point (REW_REWIND), or keep
 232                  * rewinding (REW_NOHALT).
 233                  */
 234                 c = rew_dohalt(tok, type, n);
 235                 if (REW_HALT == c)
 236                         return(1);
 237                 if (REW_REWIND == c)
 238                         break;
 239         }
 240
 241         /*
 242          * Rewind until the current point.  Warn if we're a roff
 243          * instruction that's mowing over explicit scopes.
 244          */
 245         assert(n);
 246
 247         return(man_unscope(m, n, MANDOCERR_MAX));
 248 }
 249
 250
 251 /*
 252  * Close out a generic explicit macro.
 253  */
 254 /* ARGSUSED */
 255 int
 256 blk_close(MACRO_PROT_ARGS)
 257 {
 258         enum mant                ntok;
 259         const struct man_node   *nn;
 260
 261         switch (tok) {
 262         case (MAN_RE):
 263                 ntok = MAN_RS;
 264                 break;
 265         default:
 266                 abort();
 267                 /* NOTREACHED */
 268         }
 269
 270         for (nn = m->last->parent; nn; nn = nn->parent)
 271                 if (ntok == nn->tok)
 272                         break;
 273
 274         if (NULL == nn)
 275                 if ( ! man_pmsg(m, line, ppos, MANDOCERR_NOSCOPE))
 276                         return(0);
 277
 278         if ( ! rew_scope(MAN_BODY, m, ntok))
 279                 return(0);
 280         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 281                 return(0);
 282
 283         return(1);
 284 }
 285
 286
 287 /* ARGSUSED */
 288 int
 289 blk_exp(MACRO_PROT_ARGS)
 290 {
 291         int              w, la;
 292         char            *p;
 293
 294         /*
 295          * Close out prior scopes.  "Regular" explicit macros cannot be
 296          * nested, but we allow roff macros to be placed just about
 297          * anywhere.
 298          */
 299
 300         if ( ! rew_scope(MAN_BODY, m, tok))
 301                 return(0);
 302         if ( ! rew_scope(MAN_BLOCK, m, tok))
 303                 return(0);
 304
 305         if ( ! man_block_alloc(m, line, ppos, tok))
 306                 return(0);
 307         if ( ! man_head_alloc(m, line, ppos, tok))
 308                 return(0);
 309
 310         for (;;) {
 311                 la = *pos;
 312                 w = man_args(m, line, pos, buf, &p);
 313
 314                 if (-1 == w)
 315                         return(0);
 316                 if (0 == w)
 317                         break;
 318
 319                 if ( ! man_word_alloc(m, line, la, p))
 320                         return(0);
 321         }
 322
 323         assert(m);
 324         assert(tok != MAN_MAX);
 325
 326         if ( ! rew_scope(MAN_HEAD, m, tok))
 327                 return(0);
 328         return(man_body_alloc(m, line, ppos, tok));
 329 }
 330
 331
 332
 333 /*
 334  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 335  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 336  * scopes, such as `SH' closing out an `SS', are defined in the rew
 337  * routines.
 338  */
 339 /* ARGSUSED */
 340 int
 341 blk_imp(MACRO_PROT_ARGS)
 342 {
 343         int              w, la;
 344         char            *p;
 345         struct man_node *n;
 346
 347         /* Close out prior scopes. */
 348
 349         if ( ! rew_scope(MAN_BODY, m, tok))
 350                 return(0);
 351         if ( ! rew_scope(MAN_BLOCK, m, tok))
 352                 return(0);
 353
 354         /* Allocate new block & head scope. */
 355
 356         if ( ! man_block_alloc(m, line, ppos, tok))
 357                 return(0);
 358         if ( ! man_head_alloc(m, line, ppos, tok))
 359                 return(0);
 360
 361         n = m->last;
 362
 363         /* Add line arguments. */
 364
 365         for (;;) {
 366                 la = *pos;
 367                 w = man_args(m, line, pos, buf, &p);
 368
 369                 if (-1 == w)
 370                         return(0);
 371                 if (0 == w)
 372                         break;
 373
 374                 if ( ! man_word_alloc(m, line, la, p))
 375                         return(0);
 376         }
 377
 378         /* Close out head and open body (unless MAN_SCOPE). */
 379
 380         if (MAN_SCOPED & man_macros[tok].flags) {
 381                 /* If we're forcing scope (`TP'), keep it open. */
 382                 if (MAN_FSCOPED & man_macros[tok].flags) {
 383                         m->flags |= MAN_BLINE;
 384                         return(1);
 385                 } else if (n == m->last) {
 386                         m->flags |= MAN_BLINE;
 387                         return(1);
 388                 }
 389         }
 390
 391         if ( ! rew_scope(MAN_HEAD, m, tok))
 392                 return(0);
 393         return(man_body_alloc(m, line, ppos, tok));
 394 }
 395
 396
 397 /* ARGSUSED */
 398 int
 399 in_line_eoln(MACRO_PROT_ARGS)
 400 {
 401         int              w, la;
 402         char            *p;
 403         struct man_node *n;
 404
 405         if ( ! man_elem_alloc(m, line, ppos, tok))
 406                 return(0);
 407
 408         n = m->last;
 409
 410         for (;;) {
 411                 la = *pos;
 412                 w = man_args(m, line, pos, buf, &p);
 413
 414                 if (-1 == w)
 415                         return(0);
 416                 if (0 == w)
 417                         break;
 418                 if ( ! man_word_alloc(m, line, la, p))
 419                         return(0);
 420         }
 421
 422         /*
 423          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 424          * next-line scoped), then set our mode to indicate that we're
 425          * waiting for terms to load into our context.
 426          */
 427
 428         if (n == m->last && MAN_SCOPED & man_macros[tok].flags) {
 429                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 430                 m->flags |= MAN_ELINE;
 431                 return(1);
 432         }
 433
 434         /* Set ignorable context, if applicable. */
 435
 436         if (MAN_NSCOPED & man_macros[tok].flags) {
 437                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 438                 m->flags |= MAN_ILINE;
 439         }
 440
 441         /*
 442          * Rewind our element scope.  Note that when TH is pruned, we'll
 443          * be back at the root, so make sure that we don't clobber as
 444          * its sibling.
 445          */
 446
 447         for ( ; m->last; m->last = m->last->parent) {
 448                 if (m->last == n)
 449                         break;
 450                 if (m->last->type == MAN_ROOT)
 451                         break;
 452                 if ( ! man_valid_post(m))
 453                         return(0);
 454         }
 455
 456         assert(m->last);
 457
 458         /*
 459          * Same here regarding whether we're back at the root.
 460          */
 461
 462         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 463                 return(0);
 464
 465         m->next = MAN_ROOT == m->last->type ?
 466                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 467
 468         return(1);
 469 }
 470
 471
 472 int
 473 man_macroend(struct man *m)
 474 {
 475
 476         return(man_unscope(m, m->first, MANDOCERR_SCOPEEXIT));
 477 }
 478