man.c

   1 /*      $Id: man.c,v 1.173 2017/05/08 20:33:53 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2013, 2014, 2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
   5  * Copyright (c) 2011 Joerg Sonnenberger <joerg@netbsd.org>
   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 AUTHORS DISCLAIM ALL WARRANTIES
  12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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 #include "config.h"
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdarg.h>
  26 #include <stdlib.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29
  30 #include "mandoc_aux.h"
  31 #include "mandoc.h"
  32 #include "roff.h"
  33 #include "man.h"
  34 #include "libmandoc.h"
  35 #include "roff_int.h"
  36 #include "libman.h"
  37
  38 static  void             man_descope(struct roff_man *, int, int);
  39 static  int              man_ptext(struct roff_man *, int, char *, int);
  40 static  int              man_pmacro(struct roff_man *, int, char *, int);
  41
  42
  43 int
  44 man_parseln(struct roff_man *man, int ln, char *buf, int offs)
  45 {
  46
  47         if (man->last->type != ROFFT_EQN || ln > man->last->line)
  48                 man->flags |= MAN_NEWLINE;
  49
  50         return roff_getcontrol(man->roff, buf, &offs) ?
  51             man_pmacro(man, ln, buf, offs) :
  52             man_ptext(man, ln, buf, offs);
  53 }
  54
  55 static void
  56 man_descope(struct roff_man *man, int line, int offs)
  57 {
  58         /*
  59          * Co-ordinate what happens with having a next-line scope open:
  60          * first close out the element scope (if applicable), then close
  61          * out the block scope (also if applicable).
  62          */
  63
  64         if (man->flags & MAN_ELINE) {
  65                 man->flags &= ~MAN_ELINE;
  66                 man_unscope(man, man->last->parent);
  67         }
  68         if ( ! (man->flags & MAN_BLINE))
  69                 return;
  70         man->flags &= ~MAN_BLINE;
  71         man_unscope(man, man->last->parent);
  72         roff_body_alloc(man, line, offs, man->last->tok);
  73 }
  74
  75 static int
  76 man_ptext(struct roff_man *man, int line, char *buf, int offs)
  77 {
  78         int              i;
  79
  80         /* Literal free-form text whitespace is preserved. */
  81
  82         if (man->flags & MAN_LITERAL) {
  83                 roff_word_alloc(man, line, offs, buf + offs);
  84                 man_descope(man, line, offs);
  85                 return 1;
  86         }
  87
  88         for (i = offs; buf[i] == ' '; i++)
  89                 /* Skip leading whitespace. */ ;
  90
  91         /*
  92          * Blank lines are ignored right after headings
  93          * but add a single vertical space elsewhere.
  94          */
  95
  96         if (buf[i] == '\0') {
  97                 man_breakscope(man, ROFF_sp);
  98                 /* Allocate a blank entry. */
  99                 if (man->last->tok != MAN_SH &&
 100                     man->last->tok != MAN_SS) {
 101                         roff_elem_alloc(man, line, offs, ROFF_sp);
 102                         man->next = ROFF_NEXT_SIBLING;
 103                 }
 104                 return 1;
 105         }
 106
 107         /*
 108          * Warn if the last un-escaped character is whitespace. Then
 109          * strip away the remaining spaces (tabs stay!).
 110          */
 111
 112         i = (int)strlen(buf);
 113         assert(i);
 114
 115         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 116                 if (i > 1 && '\\' != buf[i - 2])
 117                         mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 118                             line, i - 1, NULL);
 119
 120                 for (--i; i && ' ' == buf[i]; i--)
 121                         /* Spin back to non-space. */ ;
 122
 123                 /* Jump ahead of escaped whitespace. */
 124                 i += '\\' == buf[i] ? 2 : 1;
 125
 126                 buf[i] = '\0';
 127         }
 128         roff_word_alloc(man, line, offs, buf + offs);
 129
 130         /*
 131          * End-of-sentence check.  If the last character is an unescaped
 132          * EOS character, then flag the node as being the end of a
 133          * sentence.  The front-end will know how to interpret this.
 134          */
 135
 136         assert(i);
 137         if (mandoc_eos(buf, (size_t)i))
 138                 man->last->flags |= NODE_EOS;
 139
 140         man_descope(man, line, offs);
 141         return 1;
 142 }
 143
 144 static int
 145 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 146 {
 147         struct roff_node *n;
 148         const char      *cp;
 149         size_t           sz;
 150         enum roff_tok    tok;
 151         int              ppos;
 152         int              bline;
 153
 154         /* Determine the line macro. */
 155
 156         ppos = offs;
 157         tok = TOKEN_NONE;
 158         for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
 159                 offs++;
 160         if (sz > 0 && sz < 4)
 161                 tok = roffhash_find(man->manmac, buf + ppos, sz);
 162         if (tok == TOKEN_NONE) {
 163                 mandoc_msg(MANDOCERR_MACRO, man->parse,
 164                     ln, ppos, buf + ppos - 1);
 165                 return 1;
 166         }
 167
 168         /* Skip a leading escape sequence or tab. */
 169
 170         switch (buf[offs]) {
 171         case '\\':
 172                 cp = buf + offs + 1;
 173                 mandoc_escape(&cp, NULL, NULL);
 174                 offs = cp - buf;
 175                 break;
 176         case '\t':
 177                 offs++;
 178                 break;
 179         default:
 180                 break;
 181         }
 182
 183         /* Jump to the next non-whitespace word. */
 184
 185         while (buf[offs] == ' ')
 186                 offs++;
 187
 188         /*
 189          * Trailing whitespace.  Note that tabs are allowed to be passed
 190          * into the parser as "text", so we only warn about spaces here.
 191          */
 192
 193         if (buf[offs] == '\0' && buf[offs - 1] == ' ')
 194                 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 195                     ln, offs - 1, NULL);
 196
 197         /*
 198          * Some macros break next-line scopes; otherwise, remember
 199          * whether we are in next-line scope for a block head.
 200          */
 201
 202         man_breakscope(man, tok);
 203         bline = man->flags & MAN_BLINE;
 204
 205         /*
 206          * If the line in next-line scope ends with \c, keep the
 207          * next-line scope open for the subsequent input line.
 208          * That is not at all portable, only groff >= 1.22.4
 209          * does it, but *if* this weird idiom occurs in a manual
 210          * page, that's very likely what the author intended.
 211          */
 212
 213         if (bline) {
 214                 cp = strchr(buf + offs, '\0') - 2;
 215                 if (cp >= buf && cp[0] == '\\' && cp[1] == 'c')
 216                         bline = 0;
 217         }
 218
 219         /* Call to handler... */
 220
 221         assert(man_macros[tok].fp);
 222         (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
 223
 224         /* In quick mode (for mandocdb), abort after the NAME section. */
 225
 226         if (man->quick && tok == MAN_SH) {
 227                 n = man->last;
 228                 if (n->type == ROFFT_BODY &&
 229                     strcmp(n->prev->child->string, "NAME"))
 230                         return 2;
 231         }
 232
 233         /*
 234          * If we are in a next-line scope for a block head,
 235          * close it out now and switch to the body,
 236          * unless the next-line scope is allowed to continue.
 237          */
 238
 239         if ( ! bline || man->flags & MAN_ELINE ||
 240             man_macros[tok].flags & MAN_NSCOPED)
 241                 return 1;
 242
 243         assert(man->flags & MAN_BLINE);
 244         man->flags &= ~MAN_BLINE;
 245
 246         man_unscope(man, man->last->parent);
 247         roff_body_alloc(man, ln, ppos, man->last->tok);
 248         return 1;
 249 }
 250
 251 void
 252 man_breakscope(struct roff_man *man, int tok)
 253 {
 254         struct roff_node *n;
 255
 256         /*
 257          * An element next line scope is open,
 258          * and the new macro is not allowed inside elements.
 259          * Delete the element that is being broken.
 260          */
 261
 262         if (man->flags & MAN_ELINE && (tok < MAN_TH ||
 263             ! (man_macros[tok].flags & MAN_NSCOPED))) {
 264                 n = man->last;
 265                 assert(n->type != ROFFT_TEXT);
 266                 if (man_macros[n->tok].flags & MAN_NSCOPED)
 267                         n = n->parent;
 268
 269                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 270                     n->line, n->pos, "%s breaks %s",
 271                     roff_name[tok], roff_name[n->tok]);
 272
 273                 roff_node_delete(man, n);
 274                 man->flags &= ~MAN_ELINE;
 275         }
 276
 277         /*
 278          * Weird special case:
 279          * Switching fill mode closes section headers.
 280          */
 281
 282         if (man->flags & MAN_BLINE &&
 283             (tok == MAN_nf || tok == MAN_fi) &&
 284             (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
 285                 n = man->last;
 286                 man_unscope(man, n);
 287                 roff_body_alloc(man, n->line, n->pos, n->tok);
 288                 man->flags &= ~MAN_BLINE;
 289         }
 290
 291         /*
 292          * A block header next line scope is open,
 293          * and the new macro is not allowed inside block headers.
 294          * Delete the block that is being broken.
 295          */
 296
 297         if (man->flags & MAN_BLINE && (tok < MAN_TH ||
 298             man_macros[tok].flags & MAN_BSCOPE)) {
 299                 n = man->last;
 300                 if (n->type == ROFFT_TEXT)
 301                         n = n->parent;
 302                 if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
 303                         n = n->parent;
 304
 305                 assert(n->type == ROFFT_HEAD);
 306                 n = n->parent;
 307                 assert(n->type == ROFFT_BLOCK);
 308                 assert(man_macros[n->tok].flags & MAN_SCOPED);
 309
 310                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 311                     n->line, n->pos, "%s breaks %s",
 312                     roff_name[tok], roff_name[n->tok]);
 313
 314                 roff_node_delete(man, n);
 315                 man->flags &= ~MAN_BLINE;
 316         }
 317 }
 318
 319 const struct mparse *
 320 man_mparse(const struct roff_man *man)
 321 {
 322
 323         assert(man && man->parse);
 324         return man->parse;
 325 }
 326
 327 void
 328 man_state(struct roff_man *man, struct roff_node *n)
 329 {
 330
 331         switch(n->tok) {
 332         case MAN_nf:
 333         case MAN_EX:
 334                 if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
 335                         mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 336                             n->line, n->pos, "nf");
 337                 man->flags |= MAN_LITERAL;
 338                 break;
 339         case MAN_fi:
 340         case MAN_EE:
 341                 if ( ! (man->flags & MAN_LITERAL) &&
 342                      ! (n->flags & NODE_VALID))
 343                         mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 344                             n->line, n->pos, "fi");
 345                 man->flags &= ~MAN_LITERAL;
 346                 break;
 347         default:
 348                 break;
 349         }
 350         man->last->flags |= NODE_VALID;
 351 }
 352
 353 void
 354 man_validate(struct roff_man *man)
 355 {
 356
 357         man->last = man->first;
 358         man_node_validate(man);
 359         man->flags &= ~MAN_LITERAL;
 360 }