man_macro.c

   1 /*      $Id: man_macro.c,v 1.44 2010/05/15 22:44:04 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 "libman.h"
  27
  28 enum    rew {
  29         REW_REWIND,
  30         REW_NOHALT,
  31         REW_HALT
  32 };
  33
  34 static  int              blk_close(MACRO_PROT_ARGS);
  35 static  int              blk_exp(MACRO_PROT_ARGS);
  36 static  int              blk_imp(MACRO_PROT_ARGS);
  37 static  int              in_line_eoln(MACRO_PROT_ARGS);
  38
  39 static  int              rew_scope(enum man_type,
  40                                 struct man *, enum mant);
  41 static  enum rew         rew_dohalt(enum mant, enum man_type,
  42                                 const struct man_node *);
  43 static  enum rew         rew_block(enum mant, enum man_type,
  44                                 const struct man_node *);
  45 static  int              rew_warn(struct man *,
  46                                 struct man_node *, enum merr);
  47
  48 const   struct man_macro __man_macros[MAN_MAX] = {
  49         { in_line_eoln, MAN_NSCOPED }, /* br */
  50         { in_line_eoln, 0 }, /* TH */
  51         { blk_imp, MAN_SCOPED }, /* SH */
  52         { blk_imp, MAN_SCOPED }, /* SS */
  53         { blk_imp, MAN_SCOPED | MAN_FSCOPED }, /* TP */
  54         { blk_imp, 0 }, /* LP */
  55         { blk_imp, 0 }, /* PP */
  56         { blk_imp, 0 }, /* P */
  57         { blk_imp, 0 }, /* IP */
  58         { blk_imp, 0 }, /* HP */
  59         { in_line_eoln, MAN_SCOPED }, /* SM */
  60         { in_line_eoln, MAN_SCOPED }, /* SB */
  61         { in_line_eoln, 0 }, /* BI */
  62         { in_line_eoln, 0 }, /* IB */
  63         { in_line_eoln, 0 }, /* BR */
  64         { in_line_eoln, 0 }, /* RB */
  65         { in_line_eoln, MAN_SCOPED }, /* R */
  66         { in_line_eoln, MAN_SCOPED }, /* B */
  67         { in_line_eoln, MAN_SCOPED }, /* I */
  68         { in_line_eoln, 0 }, /* IR */
  69         { in_line_eoln, 0 }, /* RI */
  70         { in_line_eoln, MAN_NSCOPED }, /* na */
  71         { in_line_eoln, 0 }, /* i */
  72         { in_line_eoln, MAN_NSCOPED }, /* sp */
  73         { in_line_eoln, 0 }, /* nf */
  74         { in_line_eoln, 0 }, /* fi */
  75         { in_line_eoln, 0 }, /* r */
  76         { blk_close, 0 }, /* RE */
  77         { blk_exp, MAN_EXPLICIT }, /* RS */
  78         { in_line_eoln, 0 }, /* DT */
  79         { in_line_eoln, 0 }, /* UC */
  80         { in_line_eoln, 0 }, /* PD */
  81         { in_line_eoln, MAN_NSCOPED }, /* Sp */
  82         { in_line_eoln, 0 }, /* Vb */
  83         { in_line_eoln, 0 }, /* Ve */
  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 merr er)
  94 {
  95
  96         if (er == WERRMAX || 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_nwarn(m, n, er));
 103 }
 104
 105
 106 /*
 107  * Rewind scope.  If a code "er" != WERRMAX has been provided, it will
 108  * 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, enum merr er)
 112 {
 113
 114         assert(n);
 115
 116         /* LINTED */
 117         while (m->last != n) {
 118                 if ( ! rew_warn(m, m->last, er))
 119                         return(0);
 120                 if ( ! man_valid_post(m))
 121                         return(0);
 122                 if ( ! man_action_post(m))
 123                         return(0);
 124                 m->last = m->last->parent;
 125                 assert(m->last);
 126         }
 127
 128         if ( ! rew_warn(m, m->last, er))
 129                 return(0);
 130         if ( ! man_valid_post(m))
 131                 return(0);
 132         if ( ! man_action_post(m))
 133                 return(0);
 134
 135         m->next = MAN_ROOT == m->last->type ?
 136                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 137
 138         return(1);
 139 }
 140
 141
 142 static enum rew
 143 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 144 {
 145
 146         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 147                         MAN_BODY == n->parent->type)
 148                 return(REW_REWIND);
 149         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 150 }
 151
 152
 153 /*
 154  * There are three scope levels: scoped to the root (all), scoped to the
 155  * section (all less sections), and scoped to subsections (all less
 156  * sections and subsections).
 157  */
 158 static enum rew
 159 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 160 {
 161         enum rew         c;
 162
 163         /* We cannot progress beyond the root ever. */
 164         if (MAN_ROOT == n->type)
 165                 return(REW_HALT);
 166
 167         assert(n->parent);
 168
 169         /* Normal nodes shouldn't go to the level of the root. */
 170         if (MAN_ROOT == n->parent->type)
 171                 return(REW_REWIND);
 172
 173         /* Already-validated nodes should be closed out. */
 174         if (MAN_VALID & n->flags)
 175                 return(REW_NOHALT);
 176
 177         /* First: rewind to ourselves. */
 178         if (type == n->type && tok == n->tok)
 179                 return(REW_REWIND);
 180
 181         /*
 182          * Next follow the implicit scope-smashings as defined by man.7:
 183          * section, sub-section, etc.
 184          */
 185
 186         switch (tok) {
 187         case (MAN_SH):
 188                 break;
 189         case (MAN_SS):
 190                 /* Rewind to a section, if a block. */
 191                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 192                         return(c);
 193                 break;
 194         case (MAN_RS):
 195                 /* Rewind to a subsection, if a block. */
 196                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 197                         return(c);
 198                 /* Rewind to a section, if a block. */
 199                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 200                         return(c);
 201                 break;
 202         default:
 203                 /* Rewind to an offsetter, if a block. */
 204                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 205                         return(c);
 206                 /* Rewind to a subsection, if a block. */
 207                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 208                         return(c);
 209                 /* Rewind to a section, if a block. */
 210                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 211                         return(c);
 212                 break;
 213         }
 214
 215         return(REW_NOHALT);
 216 }
 217
 218
 219 /*
 220  * Rewinding entails ascending the parse tree until a coherent point,
 221  * for example, the `SH' macro will close out any intervening `SS'
 222  * scopes.  When a scope is closed, it must be validated and actioned.
 223  */
 224 static int
 225 rew_scope(enum man_type type, struct man *m, enum mant tok)
 226 {
 227         struct man_node *n;
 228         enum rew         c;
 229
 230         /* LINTED */
 231         for (n = m->last; n; n = n->parent) {
 232                 /*
 233                  * Whether we should stop immediately (REW_HALT), stop
 234                  * and rewind until this point (REW_REWIND), or keep
 235                  * rewinding (REW_NOHALT).
 236                  */
 237                 c = rew_dohalt(tok, type, n);
 238                 if (REW_HALT == c)
 239                         return(1);
 240                 if (REW_REWIND == c)
 241                         break;
 242         }
 243
 244         /*
 245          * Rewind until the current point.  Warn if we're a roff
 246          * instruction that's mowing over explicit scopes.
 247          */
 248         assert(n);
 249
 250         return(man_unscope(m, n, WERRMAX));
 251 }
 252
 253
 254 /*
 255  * Close out a generic explicit macro.
 256  */
 257 /* ARGSUSED */
 258 int
 259 blk_close(MACRO_PROT_ARGS)
 260 {
 261         enum mant                ntok;
 262         const struct man_node   *nn;
 263
 264         switch (tok) {
 265         case (MAN_RE):
 266                 ntok = MAN_RS;
 267                 break;
 268         default:
 269                 abort();
 270                 /* NOTREACHED */
 271         }
 272
 273         for (nn = m->last->parent; nn; nn = nn->parent)
 274                 if (ntok == nn->tok)
 275                         break;
 276
 277         if (NULL == nn)
 278                 if ( ! man_pwarn(m, line, ppos, WNOSCOPE))
 279                         return(0);
 280
 281         if ( ! rew_scope(MAN_BODY, m, ntok))
 282                 return(0);
 283         if ( ! rew_scope(MAN_BLOCK, m, ntok))
 284                 return(0);
 285
 286         return(1);
 287 }
 288
 289
 290 int
 291 blk_exp(MACRO_PROT_ARGS)
 292 {
 293         int              w, la;
 294         char            *p;
 295
 296         /*
 297          * Close out prior scopes.  "Regular" explicit macros cannot be
 298          * nested, but we allow roff macros to be placed just about
 299          * anywhere.
 300          */
 301
 302         if ( ! rew_scope(MAN_BODY, m, tok))
 303                 return(0);
 304         if ( ! rew_scope(MAN_BLOCK, m, tok))
 305                 return(0);
 306
 307         if ( ! man_block_alloc(m, line, ppos, tok))
 308                 return(0);
 309         if ( ! man_head_alloc(m, line, ppos, tok))
 310                 return(0);
 311
 312         for (;;) {
 313                 la = *pos;
 314                 w = man_args(m, line, pos, buf, &p);
 315
 316                 if (-1 == w)
 317                         return(0);
 318                 if (0 == w)
 319                         break;
 320
 321                 if ( ! man_word_alloc(m, line, la, p))
 322                         return(0);
 323         }
 324
 325         assert(m);
 326         assert(tok != MAN_MAX);
 327
 328         if ( ! rew_scope(MAN_HEAD, m, tok))
 329                 return(0);
 330         return(man_body_alloc(m, line, ppos, tok));
 331 }
 332
 333
 334
 335 /*
 336  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 337  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 338  * scopes, such as `SH' closing out an `SS', are defined in the rew
 339  * routines.
 340  */
 341 int
 342 blk_imp(MACRO_PROT_ARGS)
 343 {
 344         int              w, 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                 w = man_args(m, line, pos, buf, &p);
 369
 370                 if (-1 == w)
 371                         return(0);
 372                 if (0 == w)
 373                         break;
 374
 375                 if ( ! man_word_alloc(m, line, la, p))
 376                         return(0);
 377         }
 378
 379         /* Close out head and open body (unless MAN_SCOPE). */
 380
 381         if (MAN_SCOPED & man_macros[tok].flags) {
 382                 /* If we're forcing scope (`TP'), keep it open. */
 383                 if (MAN_FSCOPED & man_macros[tok].flags) {
 384                         m->flags |= MAN_BLINE;
 385                         return(1);
 386                 } else if (n == m->last) {
 387                         m->flags |= MAN_BLINE;
 388                         return(1);
 389                 }
 390         }
 391
 392         if ( ! rew_scope(MAN_HEAD, m, tok))
 393                 return(0);
 394         return(man_body_alloc(m, line, ppos, tok));
 395 }
 396
 397
 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                 if ( ! man_action_post(m))
 455                         return(0);
 456         }
 457
 458         assert(m->last);
 459
 460         /*
 461          * Same here regarding whether we're back at the root.
 462          */
 463
 464         if (m->last->type != MAN_ROOT && ! man_valid_post(m))
 465                 return(0);
 466         if (m->last->type != MAN_ROOT && ! man_action_post(m))
 467                 return(0);
 468
 469         m->next = MAN_ROOT == m->last->type ?
 470                 MAN_NEXT_CHILD : MAN_NEXT_SIBLING;
 471
 472         return(1);
 473 }
 474
 475
 476 int
 477 man_macroend(struct man *m)
 478 {
 479
 480         return(man_unscope(m, m->first, WEXITSCOPE));
 481 }
 482