man_macro.c

   1 /*      $Id: man_macro.c,v 1.82 2014/04/20 16:46:04 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012, 2013 Ingo Schwarze <schwarze@openbsd.org>
   5  * Copyright (c) 2013 Franco Fichtner <franco@lastsummer.de>
   6  *
   7  * Permission to use, copy, modify, and distribute this software for any
   8  * purpose with or without fee is hereby granted, provided that the above
   9  * copyright notice and this permission notice appear in all copies.
  10  *
  11  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  14  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  15  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  16  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  17  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #ifdef HAVE_CONFIG_H
  20 #include "config.h"
  21 #endif
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27
  28 #include "man.h"
  29 #include "mandoc.h"
  30 #include "libmandoc.h"
  31 #include "libman.h"
  32
  33 enum    rew {
  34         REW_REWIND,
  35         REW_NOHALT,
  36         REW_HALT
  37 };
  38
  39 static  int              blk_close(MACRO_PROT_ARGS);
  40 static  int              blk_exp(MACRO_PROT_ARGS);
  41 static  int              blk_imp(MACRO_PROT_ARGS);
  42 static  int              in_line_eoln(MACRO_PROT_ARGS);
  43 static  int              man_args(struct man *, int,
  44                                 int *, char *, char **);
  45
  46 static  int              rew_scope(enum man_type,
  47                                 struct man *, enum mant);
  48 static  enum rew         rew_dohalt(enum mant, enum man_type,
  49                                 const struct man_node *);
  50 static  enum rew         rew_block(enum mant, enum man_type,
  51                                 const struct man_node *);
  52 static  void             rew_warn(struct man *,
  53                                 struct man_node *, enum mandocerr);
  54
  55 const   struct man_macro __man_macros[MAN_MAX] = {
  56         { in_line_eoln, MAN_NSCOPED }, /* br */
  57         { in_line_eoln, MAN_BSCOPE }, /* TH */
  58         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
  59         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
  60         { blk_imp, MAN_BSCOPE | MAN_SCOPED | MAN_FSCOPED }, /* TP */
  61         { blk_imp, MAN_BSCOPE }, /* LP */
  62         { blk_imp, MAN_BSCOPE }, /* PP */
  63         { blk_imp, MAN_BSCOPE }, /* P */
  64         { blk_imp, MAN_BSCOPE }, /* IP */
  65         { blk_imp, MAN_BSCOPE }, /* HP */
  66         { in_line_eoln, MAN_SCOPED }, /* SM */
  67         { in_line_eoln, MAN_SCOPED }, /* SB */
  68         { in_line_eoln, 0 }, /* BI */
  69         { in_line_eoln, 0 }, /* IB */
  70         { in_line_eoln, 0 }, /* BR */
  71         { in_line_eoln, 0 }, /* RB */
  72         { in_line_eoln, MAN_SCOPED }, /* R */
  73         { in_line_eoln, MAN_SCOPED }, /* B */
  74         { in_line_eoln, MAN_SCOPED }, /* I */
  75         { in_line_eoln, 0 }, /* IR */
  76         { in_line_eoln, 0 }, /* RI */
  77         { in_line_eoln, MAN_NSCOPED }, /* na */
  78         { in_line_eoln, MAN_NSCOPED }, /* sp */
  79         { in_line_eoln, MAN_BSCOPE }, /* nf */
  80         { in_line_eoln, MAN_BSCOPE }, /* fi */
  81         { blk_close, 0 }, /* RE */
  82         { blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* RS */
  83         { in_line_eoln, 0 }, /* DT */
  84         { in_line_eoln, 0 }, /* UC */
  85         { in_line_eoln, 0 }, /* PD */
  86         { in_line_eoln, 0 }, /* AT */
  87         { in_line_eoln, 0 }, /* in */
  88         { in_line_eoln, 0 }, /* ft */
  89         { in_line_eoln, 0 }, /* OP */
  90         { in_line_eoln, MAN_BSCOPE }, /* EX */
  91         { in_line_eoln, MAN_BSCOPE }, /* EE */
  92         { blk_exp, MAN_BSCOPE | MAN_EXPLICIT }, /* UR */
  93         { blk_close, 0 }, /* UE */
  94         { in_line_eoln, 0 }, /* ll */
  95 };
  96
  97 const   struct man_macro * const man_macros = __man_macros;
  98
  99
 100 /*
 101  * Warn when "n" is an explicit non-roff macro.
 102  */
 103 static void
 104 rew_warn(struct man *man, struct man_node *n, enum mandocerr er)
 105 {
 106
 107         if (er == MANDOCERR_MAX || MAN_BLOCK != n->type)
 108                 return;
 109         if (MAN_VALID & n->flags)
 110                 return;
 111         if ( ! (MAN_EXPLICIT & man_macros[n->tok].flags))
 112                 return;
 113
 114         assert(er < MANDOCERR_FATAL);
 115         man_nmsg(man, n, er);
 116 }
 117
 118 /*
 119  * Rewind scope.  If a code "er" != MANDOCERR_MAX has been provided, it
 120  * will be used if an explicit block scope is being closed out.
 121  */
 122 int
 123 man_unscope(struct man *man, const struct man_node *to,
 124                 enum mandocerr er)
 125 {
 126         struct man_node *n;
 127
 128         assert(to);
 129
 130         man->next = MAN_NEXT_SIBLING;
 131
 132         while (man->last != to) {
 133                 /*
 134                  * Save the parent here, because we may delete the
 135                  * man->last node in the post-validation phase and reset
 136                  * it to man->last->parent, causing a step in the closing
 137                  * out to be lost.
 138                  */
 139                 n = man->last->parent;
 140                 rew_warn(man, man->last, er);
 141                 if ( ! man_valid_post(man))
 142                         return(0);
 143                 man->last = n;
 144                 assert(man->last);
 145         }
 146
 147         rew_warn(man, man->last, er);
 148         if ( ! man_valid_post(man))
 149                 return(0);
 150
 151         return(1);
 152 }
 153
 154 static enum rew
 155 rew_block(enum mant ntok, enum man_type type, const struct man_node *n)
 156 {
 157
 158         if (MAN_BLOCK == type && ntok == n->parent->tok &&
 159             MAN_BODY == n->parent->type)
 160                 return(REW_REWIND);
 161         return(ntok == n->tok ? REW_HALT : REW_NOHALT);
 162 }
 163
 164 /*
 165  * There are three scope levels: scoped to the root (all), scoped to the
 166  * section (all less sections), and scoped to subsections (all less
 167  * sections and subsections).
 168  */
 169 static enum rew
 170 rew_dohalt(enum mant tok, enum man_type type, const struct man_node *n)
 171 {
 172         enum rew         c;
 173
 174         /* We cannot progress beyond the root ever. */
 175         if (MAN_ROOT == n->type)
 176                 return(REW_HALT);
 177
 178         assert(n->parent);
 179
 180         /* Normal nodes shouldn't go to the level of the root. */
 181         if (MAN_ROOT == n->parent->type)
 182                 return(REW_REWIND);
 183
 184         /* Already-validated nodes should be closed out. */
 185         if (MAN_VALID & n->flags)
 186                 return(REW_NOHALT);
 187
 188         /* First: rewind to ourselves. */
 189         if (type == n->type && tok == n->tok) {
 190                 if (MAN_EXPLICIT & man_macros[n->tok].flags)
 191                         return(REW_HALT);
 192                 else
 193                         return(REW_REWIND);
 194         }
 195
 196         /*
 197          * Next follow the implicit scope-smashings as defined by man.7:
 198          * section, sub-section, etc.
 199          */
 200
 201         switch (tok) {
 202         case MAN_SH:
 203                 break;
 204         case MAN_SS:
 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         case MAN_RS:
 210                 /* Preserve empty paragraphs before RS. */
 211                 if (0 == n->nchild && (MAN_P == n->tok ||
 212                     MAN_PP == n->tok || MAN_LP == n->tok))
 213                         return(REW_HALT);
 214                 /* Rewind to a subsection, if a block. */
 215                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 216                         return(c);
 217                 /* Rewind to a section, if a block. */
 218                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 219                         return(c);
 220                 break;
 221         default:
 222                 /* Rewind to an offsetter, if a block. */
 223                 if (REW_NOHALT != (c = rew_block(MAN_RS, type, n)))
 224                         return(c);
 225                 /* Rewind to a subsection, if a block. */
 226                 if (REW_NOHALT != (c = rew_block(MAN_SS, type, n)))
 227                         return(c);
 228                 /* Rewind to a section, if a block. */
 229                 if (REW_NOHALT != (c = rew_block(MAN_SH, type, n)))
 230                         return(c);
 231                 break;
 232         }
 233
 234         return(REW_NOHALT);
 235 }
 236
 237 /*
 238  * Rewinding entails ascending the parse tree until a coherent point,
 239  * for example, the `SH' macro will close out any intervening `SS'
 240  * scopes.  When a scope is closed, it must be validated and actioned.
 241  */
 242 static int
 243 rew_scope(enum man_type type, struct man *man, enum mant tok)
 244 {
 245         struct man_node *n;
 246         enum rew         c;
 247
 248         for (n = man->last; n; n = n->parent) {
 249                 /*
 250                  * Whether we should stop immediately (REW_HALT), stop
 251                  * and rewind until this point (REW_REWIND), or keep
 252                  * rewinding (REW_NOHALT).
 253                  */
 254                 c = rew_dohalt(tok, type, n);
 255                 if (REW_HALT == c)
 256                         return(1);
 257                 if (REW_REWIND == c)
 258                         break;
 259         }
 260
 261         /*
 262          * Rewind until the current point.  Warn if we're a roff
 263          * instruction that's mowing over explicit scopes.
 264          */
 265         assert(n);
 266
 267         return(man_unscope(man, n, MANDOCERR_MAX));
 268 }
 269
 270
 271 /*
 272  * Close out a generic explicit macro.
 273  */
 274 int
 275 blk_close(MACRO_PROT_ARGS)
 276 {
 277         enum mant                ntok;
 278         const struct man_node   *nn;
 279
 280         switch (tok) {
 281         case MAN_RE:
 282                 ntok = MAN_RS;
 283                 break;
 284         case MAN_UE:
 285                 ntok = MAN_UR;
 286                 break;
 287         default:
 288                 abort();
 289                 /* NOTREACHED */
 290         }
 291
 292         for (nn = man->last->parent; nn; nn = nn->parent)
 293                 if (ntok == nn->tok && MAN_BLOCK == nn->type)
 294                         break;
 295
 296         if (NULL == nn) {
 297                 man_pmsg(man, line, ppos, MANDOCERR_NOSCOPE);
 298                 if ( ! rew_scope(MAN_BLOCK, man, MAN_PP))
 299                         return(0);
 300         } else
 301                 man_unscope(man, nn, MANDOCERR_MAX);
 302
 303         return(1);
 304 }
 305
 306 int
 307 blk_exp(MACRO_PROT_ARGS)
 308 {
 309         struct man_node *n;
 310         int              la;
 311         char            *p;
 312
 313         /* Close out prior implicit scopes. */
 314
 315         if ( ! rew_scope(MAN_BLOCK, man, tok))
 316                 return(0);
 317
 318         if ( ! man_block_alloc(man, line, ppos, tok))
 319                 return(0);
 320         if ( ! man_head_alloc(man, line, ppos, tok))
 321                 return(0);
 322
 323         for (;;) {
 324                 la = *pos;
 325                 if ( ! man_args(man, line, pos, buf, &p))
 326                         break;
 327                 if ( ! man_word_alloc(man, line, la, p))
 328                         return(0);
 329         }
 330
 331         assert(man);
 332         assert(tok != MAN_MAX);
 333
 334         for (n = man->last; n; n = n->parent) {
 335                 if (n->tok != tok)
 336                         continue;
 337                 assert(MAN_HEAD == n->type);
 338                 man_unscope(man, n, MANDOCERR_MAX);
 339                 break;
 340         }
 341
 342         return(man_body_alloc(man, line, ppos, tok));
 343 }
 344
 345 /*
 346  * Parse an implicit-block macro.  These contain a MAN_HEAD and a
 347  * MAN_BODY contained within a MAN_BLOCK.  Rules for closing out other
 348  * scopes, such as `SH' closing out an `SS', are defined in the rew
 349  * routines.
 350  */
 351 int
 352 blk_imp(MACRO_PROT_ARGS)
 353 {
 354         int              la;
 355         char            *p;
 356         struct man_node *n;
 357
 358         /* Close out prior scopes. */
 359
 360         if ( ! rew_scope(MAN_BODY, man, tok))
 361                 return(0);
 362         if ( ! rew_scope(MAN_BLOCK, man, tok))
 363                 return(0);
 364
 365         /* Allocate new block & head scope. */
 366
 367         if ( ! man_block_alloc(man, line, ppos, tok))
 368                 return(0);
 369         if ( ! man_head_alloc(man, line, ppos, tok))
 370                 return(0);
 371
 372         n = man->last;
 373
 374         /* Add line arguments. */
 375
 376         for (;;) {
 377                 la = *pos;
 378                 if ( ! man_args(man, line, pos, buf, &p))
 379                         break;
 380                 if ( ! man_word_alloc(man, line, la, p))
 381                         return(0);
 382         }
 383
 384         /* Close out head and open body (unless MAN_SCOPE). */
 385
 386         if (MAN_SCOPED & man_macros[tok].flags) {
 387                 /* If we're forcing scope (`TP'), keep it open. */
 388                 if (MAN_FSCOPED & man_macros[tok].flags) {
 389                         man->flags |= MAN_BLINE;
 390                         return(1);
 391                 } else if (n == man->last) {
 392                         man->flags |= MAN_BLINE;
 393                         return(1);
 394                 }
 395         }
 396
 397         if ( ! rew_scope(MAN_HEAD, man, tok))
 398                 return(0);
 399         return(man_body_alloc(man, line, ppos, tok));
 400 }
 401
 402 int
 403 in_line_eoln(MACRO_PROT_ARGS)
 404 {
 405         int              la;
 406         char            *p;
 407         struct man_node *n;
 408
 409         if ( ! man_elem_alloc(man, line, ppos, tok))
 410                 return(0);
 411
 412         n = man->last;
 413
 414         for (;;) {
 415                 la = *pos;
 416                 if ( ! man_args(man, line, pos, buf, &p))
 417                         break;
 418                 if ( ! man_word_alloc(man, line, la, p))
 419                         return(0);
 420         }
 421
 422         /*
 423          * Append MAN_EOS in case the last snipped argument
 424          * ends with a dot, e.g. `.IR syslog (3).'
 425          */
 426
 427         if (n != man->last &&
 428             mandoc_eos(man->last->string, strlen(man->last->string)))
 429                 man->last->flags |= MAN_EOS;
 430
 431         /*
 432          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 433          * next-line scoped), then set our mode to indicate that we're
 434          * waiting for terms to load into our context.
 435          */
 436
 437         if (n == man->last && MAN_SCOPED & man_macros[tok].flags) {
 438                 assert( ! (MAN_NSCOPED & man_macros[tok].flags));
 439                 man->flags |= MAN_ELINE;
 440                 return(1);
 441         }
 442
 443         /* Set ignorable context, if applicable. */
 444
 445         if (MAN_NSCOPED & man_macros[tok].flags) {
 446                 assert( ! (MAN_SCOPED & man_macros[tok].flags));
 447                 man->flags |= MAN_ILINE;
 448         }
 449
 450         assert(MAN_ROOT != man->last->type);
 451         man->next = MAN_NEXT_SIBLING;
 452
 453         /*
 454          * Rewind our element scope.  Note that when TH is pruned, we'll
 455          * be back at the root, so make sure that we don't clobber as
 456          * its sibling.
 457          */
 458
 459         for ( ; man->last; man->last = man->last->parent) {
 460                 if (man->last == n)
 461                         break;
 462                 if (man->last->type == MAN_ROOT)
 463                         break;
 464                 if ( ! man_valid_post(man))
 465                         return(0);
 466         }
 467
 468         assert(man->last);
 469
 470         /*
 471          * Same here regarding whether we're back at the root.
 472          */
 473
 474         if (man->last->type != MAN_ROOT && ! man_valid_post(man))
 475                 return(0);
 476
 477         return(1);
 478 }
 479
 480
 481 int
 482 man_macroend(struct man *man)
 483 {
 484
 485         return(man_unscope(man, man->first, MANDOCERR_SCOPEEXIT));
 486 }
 487
 488 static int
 489 man_args(struct man *man, int line, int *pos, char *buf, char **v)
 490 {
 491         char     *start;
 492
 493         assert(*pos);
 494         *v = start = buf + *pos;
 495         assert(' ' != *start);
 496
 497         if ('\0' == *start)
 498                 return(0);
 499
 500         *v = mandoc_getarg(man->parse, v, line, pos);
 501         return(1);
 502 }