man.c

   1 /*      $Id: man.c,v 1.176 2017/06/28 12:52:45 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         const char      *cp, *sp;
  80         char            *ep;
  81
  82         /* Literal free-form text whitespace is preserved. */
  83
  84         if (man->flags & MAN_LITERAL) {
  85                 roff_word_alloc(man, line, offs, buf + offs);
  86                 man_descope(man, line, offs);
  87                 return 1;
  88         }
  89
  90         for (i = offs; buf[i] == ' '; i++)
  91                 /* Skip leading whitespace. */ ;
  92
  93         /*
  94          * Blank lines are ignored in next line scope
  95          * and right after headings and cancel preceding \c,
  96          * but add a single vertical space elsewhere.
  97          */
  98
  99         if (buf[i] == '\0') {
 100                 if (man->flags & (MAN_ELINE | MAN_BLINE)) {
 101                         mandoc_msg(MANDOCERR_BLK_BLANK, man->parse,
 102                             line, 0, NULL);
 103                         return 1;
 104                 }
 105                 if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
 106                         return 1;
 107                 switch (man->last->type) {
 108                 case ROFFT_TEXT:
 109                         sp = man->last->string;
 110                         cp = ep = strchr(sp, '\0') - 2;
 111                         if (cp < sp || cp[0] != '\\' || cp[1] != 'c')
 112                                 break;
 113                         while (cp > sp && cp[-1] == '\\')
 114                                 cp--;
 115                         if ((ep - cp) % 2)
 116                                 break;
 117                         *ep = '\0';
 118                         return 1;
 119                 default:
 120                         break;
 121                 }
 122                 roff_elem_alloc(man, line, offs, ROFF_sp);
 123                 man->next = ROFF_NEXT_SIBLING;
 124                 return 1;
 125         }
 126
 127         /*
 128          * Warn if the last un-escaped character is whitespace. Then
 129          * strip away the remaining spaces (tabs stay!).
 130          */
 131
 132         i = (int)strlen(buf);
 133         assert(i);
 134
 135         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 136                 if (i > 1 && '\\' != buf[i - 2])
 137                         mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 138                             line, i - 1, NULL);
 139
 140                 for (--i; i && ' ' == buf[i]; i--)
 141                         /* Spin back to non-space. */ ;
 142
 143                 /* Jump ahead of escaped whitespace. */
 144                 i += '\\' == buf[i] ? 2 : 1;
 145
 146                 buf[i] = '\0';
 147         }
 148         roff_word_alloc(man, line, offs, buf + offs);
 149
 150         /*
 151          * End-of-sentence check.  If the last character is an unescaped
 152          * EOS character, then flag the node as being the end of a
 153          * sentence.  The front-end will know how to interpret this.
 154          */
 155
 156         assert(i);
 157         if (mandoc_eos(buf, (size_t)i))
 158                 man->last->flags |= NODE_EOS;
 159
 160         man_descope(man, line, offs);
 161         return 1;
 162 }
 163
 164 static int
 165 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 166 {
 167         struct roff_node *n;
 168         const char      *cp;
 169         size_t           sz;
 170         enum roff_tok    tok;
 171         int              ppos;
 172         int              bline;
 173
 174         /* Determine the line macro. */
 175
 176         ppos = offs;
 177         tok = TOKEN_NONE;
 178         for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
 179                 offs++;
 180         if (sz > 0 && sz < 4)
 181                 tok = roffhash_find(man->manmac, buf + ppos, sz);
 182         if (tok == TOKEN_NONE) {
 183                 mandoc_msg(MANDOCERR_MACRO, man->parse,
 184                     ln, ppos, buf + ppos - 1);
 185                 return 1;
 186         }
 187
 188         /* Skip a leading escape sequence or tab. */
 189
 190         switch (buf[offs]) {
 191         case '\\':
 192                 cp = buf + offs + 1;
 193                 mandoc_escape(&cp, NULL, NULL);
 194                 offs = cp - buf;
 195                 break;
 196         case '\t':
 197                 offs++;
 198                 break;
 199         default:
 200                 break;
 201         }
 202
 203         /* Jump to the next non-whitespace word. */
 204
 205         while (buf[offs] == ' ')
 206                 offs++;
 207
 208         /*
 209          * Trailing whitespace.  Note that tabs are allowed to be passed
 210          * into the parser as "text", so we only warn about spaces here.
 211          */
 212
 213         if (buf[offs] == '\0' && buf[offs - 1] == ' ')
 214                 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 215                     ln, offs - 1, NULL);
 216
 217         /*
 218          * Some macros break next-line scopes; otherwise, remember
 219          * whether we are in next-line scope for a block head.
 220          */
 221
 222         man_breakscope(man, tok);
 223         bline = man->flags & MAN_BLINE;
 224
 225         /*
 226          * If the line in next-line scope ends with \c, keep the
 227          * next-line scope open for the subsequent input line.
 228          * That is not at all portable, only groff >= 1.22.4
 229          * does it, but *if* this weird idiom occurs in a manual
 230          * page, that's very likely what the author intended.
 231          */
 232
 233         if (bline) {
 234                 cp = strchr(buf + offs, '\0') - 2;
 235                 if (cp >= buf && cp[0] == '\\' && cp[1] == 'c')
 236                         bline = 0;
 237         }
 238
 239         /* Call to handler... */
 240
 241         assert(man_macros[tok].fp);
 242         (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
 243
 244         /* In quick mode (for mandocdb), abort after the NAME section. */
 245
 246         if (man->quick && tok == MAN_SH) {
 247                 n = man->last;
 248                 if (n->type == ROFFT_BODY &&
 249                     strcmp(n->prev->child->string, "NAME"))
 250                         return 2;
 251         }
 252
 253         /*
 254          * If we are in a next-line scope for a block head,
 255          * close it out now and switch to the body,
 256          * unless the next-line scope is allowed to continue.
 257          */
 258
 259         if ( ! bline || man->flags & MAN_ELINE ||
 260             man_macros[tok].flags & MAN_NSCOPED)
 261                 return 1;
 262
 263         assert(man->flags & MAN_BLINE);
 264         man->flags &= ~MAN_BLINE;
 265
 266         man_unscope(man, man->last->parent);
 267         roff_body_alloc(man, ln, ppos, man->last->tok);
 268         return 1;
 269 }
 270
 271 void
 272 man_breakscope(struct roff_man *man, int tok)
 273 {
 274         struct roff_node *n;
 275
 276         /*
 277          * An element next line scope is open,
 278          * and the new macro is not allowed inside elements.
 279          * Delete the element that is being broken.
 280          */
 281
 282         if (man->flags & MAN_ELINE && (tok < MAN_TH ||
 283             ! (man_macros[tok].flags & MAN_NSCOPED))) {
 284                 n = man->last;
 285                 if (n->type == ROFFT_TEXT)
 286                         n = n->parent;
 287                 if (n->tok < MAN_TH ||
 288                     man_macros[n->tok].flags & MAN_NSCOPED)
 289                         n = n->parent;
 290
 291                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 292                     n->line, n->pos, "%s breaks %s",
 293                     roff_name[tok], roff_name[n->tok]);
 294
 295                 roff_node_delete(man, n);
 296                 man->flags &= ~MAN_ELINE;
 297         }
 298
 299         /*
 300          * Weird special case:
 301          * Switching fill mode closes section headers.
 302          */
 303
 304         if (man->flags & MAN_BLINE &&
 305             (tok == MAN_nf || tok == MAN_fi) &&
 306             (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
 307                 n = man->last;
 308                 man_unscope(man, n);
 309                 roff_body_alloc(man, n->line, n->pos, n->tok);
 310                 man->flags &= ~MAN_BLINE;
 311         }
 312
 313         /*
 314          * A block header next line scope is open,
 315          * and the new macro is not allowed inside block headers.
 316          * Delete the block that is being broken.
 317          */
 318
 319         if (man->flags & MAN_BLINE && (tok < MAN_TH ||
 320             man_macros[tok].flags & MAN_BSCOPE)) {
 321                 n = man->last;
 322                 if (n->type == ROFFT_TEXT)
 323                         n = n->parent;
 324                 if (n->tok < MAN_TH ||
 325                     (man_macros[n->tok].flags & MAN_BSCOPE) == 0)
 326                         n = n->parent;
 327
 328                 assert(n->type == ROFFT_HEAD);
 329                 n = n->parent;
 330                 assert(n->type == ROFFT_BLOCK);
 331                 assert(man_macros[n->tok].flags & MAN_SCOPED);
 332
 333                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 334                     n->line, n->pos, "%s breaks %s",
 335                     roff_name[tok], roff_name[n->tok]);
 336
 337                 roff_node_delete(man, n);
 338                 man->flags &= ~MAN_BLINE;
 339         }
 340 }
 341
 342 const struct mparse *
 343 man_mparse(const struct roff_man *man)
 344 {
 345
 346         assert(man && man->parse);
 347         return man->parse;
 348 }
 349
 350 void
 351 man_state(struct roff_man *man, struct roff_node *n)
 352 {
 353
 354         switch(n->tok) {
 355         case MAN_nf:
 356         case MAN_EX:
 357                 if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
 358                         mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 359                             n->line, n->pos, "nf");
 360                 man->flags |= MAN_LITERAL;
 361                 break;
 362         case MAN_fi:
 363         case MAN_EE:
 364                 if ( ! (man->flags & MAN_LITERAL) &&
 365                      ! (n->flags & NODE_VALID))
 366                         mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 367                             n->line, n->pos, "fi");
 368                 man->flags &= ~MAN_LITERAL;
 369                 break;
 370         default:
 371                 break;
 372         }
 373         man->last->flags |= NODE_VALID;
 374 }
 375
 376 void
 377 man_validate(struct roff_man *man)
 378 {
 379
 380         man->last = man->first;
 381         man_node_validate(man);
 382         man->flags &= ~MAN_LITERAL;
 383 }