man.c

   1 /*      $Id: man.c,v 1.160 2015/04/19 14:25:41 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2013, 2014, 2015 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 const   char *const __man_macronames[MAN_MAX] = {
  39         "br",           "TH",           "SH",           "SS",
  40         "TP",           "LP",           "PP",           "P",
  41         "IP",           "HP",           "SM",           "SB",
  42         "BI",           "IB",           "BR",           "RB",
  43         "R",            "B",            "I",            "IR",
  44         "RI",           "sp",           "nf",
  45         "fi",           "RE",           "RS",           "DT",
  46         "UC",           "PD",           "AT",           "in",
  47         "ft",           "OP",           "EX",           "EE",
  48         "UR",           "UE",           "ll"
  49         };
  50
  51 const   char * const *man_macronames = __man_macronames;
  52
  53 static  void             man_descope(struct roff_man *, int, int);
  54 static  int              man_ptext(struct roff_man *, int, char *, int);
  55 static  int              man_pmacro(struct roff_man *, int, char *, int);
  56
  57
  58 void
  59 man_endparse(struct roff_man *man)
  60 {
  61
  62         man_macroend(man);
  63 }
  64
  65 int
  66 man_parseln(struct roff_man *man, int ln, char *buf, int offs)
  67 {
  68
  69         if (man->last->type != ROFFT_EQN || ln > man->last->line)
  70                 man->flags |= MAN_NEWLINE;
  71
  72         return (roff_getcontrol(man->roff, buf, &offs) ?
  73             man_pmacro(man, ln, buf, offs) :
  74             man_ptext(man, ln, buf, offs));
  75 }
  76
  77 void
  78 man_elem_alloc(struct roff_man *man, int line, int pos, int tok)
  79 {
  80         struct roff_node *p;
  81
  82         p = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
  83         roff_node_append(man, p);
  84         man->next = ROFF_NEXT_CHILD;
  85 }
  86
  87 void
  88 man_block_alloc(struct roff_man *man, int line, int pos, int tok)
  89 {
  90         struct roff_node *p;
  91
  92         p = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
  93         roff_node_append(man, p);
  94         man->next = ROFF_NEXT_CHILD;
  95 }
  96
  97 static void
  98 man_descope(struct roff_man *man, int line, int offs)
  99 {
 100         /*
 101          * Co-ordinate what happens with having a next-line scope open:
 102          * first close out the element scope (if applicable), then close
 103          * out the block scope (also if applicable).
 104          */
 105
 106         if (man->flags & MAN_ELINE) {
 107                 man->flags &= ~MAN_ELINE;
 108                 man_unscope(man, man->last->parent);
 109         }
 110         if ( ! (man->flags & MAN_BLINE))
 111                 return;
 112         man->flags &= ~MAN_BLINE;
 113         man_unscope(man, man->last->parent);
 114         roff_body_alloc(man, line, offs, man->last->tok);
 115 }
 116
 117 static int
 118 man_ptext(struct roff_man *man, int line, char *buf, int offs)
 119 {
 120         int              i;
 121
 122         /* Literal free-form text whitespace is preserved. */
 123
 124         if (man->flags & MAN_LITERAL) {
 125                 roff_word_alloc(man, line, offs, buf + offs);
 126                 man_descope(man, line, offs);
 127                 return(1);
 128         }
 129
 130         for (i = offs; buf[i] == ' '; i++)
 131                 /* Skip leading whitespace. */ ;
 132
 133         /*
 134          * Blank lines are ignored right after headings
 135          * but add a single vertical space elsewhere.
 136          */
 137
 138         if (buf[i] == '\0') {
 139                 /* Allocate a blank entry. */
 140                 if (man->last->tok != MAN_SH &&
 141                     man->last->tok != MAN_SS) {
 142                         man_elem_alloc(man, line, offs, MAN_sp);
 143                         man->next = ROFF_NEXT_SIBLING;
 144                 }
 145                 return(1);
 146         }
 147
 148         /*
 149          * Warn if the last un-escaped character is whitespace. Then
 150          * strip away the remaining spaces (tabs stay!).
 151          */
 152
 153         i = (int)strlen(buf);
 154         assert(i);
 155
 156         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 157                 if (i > 1 && '\\' != buf[i - 2])
 158                         mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 159                             line, i - 1, NULL);
 160
 161                 for (--i; i && ' ' == buf[i]; i--)
 162                         /* Spin back to non-space. */ ;
 163
 164                 /* Jump ahead of escaped whitespace. */
 165                 i += '\\' == buf[i] ? 2 : 1;
 166
 167                 buf[i] = '\0';
 168         }
 169         roff_word_alloc(man, line, offs, buf + offs);
 170
 171         /*
 172          * End-of-sentence check.  If the last character is an unescaped
 173          * EOS character, then flag the node as being the end of a
 174          * sentence.  The front-end will know how to interpret this.
 175          */
 176
 177         assert(i);
 178         if (mandoc_eos(buf, (size_t)i))
 179                 man->last->flags |= MAN_EOS;
 180
 181         man_descope(man, line, offs);
 182         return(1);
 183 }
 184
 185 static int
 186 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 187 {
 188         struct roff_node *n;
 189         const char      *cp;
 190         int              tok;
 191         int              i, ppos;
 192         int              bline;
 193         char             mac[5];
 194
 195         ppos = offs;
 196
 197         /*
 198          * Copy the first word into a nil-terminated buffer.
 199          * Stop when a space, tab, escape, or eoln is encountered.
 200          */
 201
 202         i = 0;
 203         while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
 204                 mac[i++] = buf[offs++];
 205
 206         mac[i] = '\0';
 207
 208         tok = (i > 0 && i < 4) ? man_hash_find(mac) : TOKEN_NONE;
 209
 210         if (tok == TOKEN_NONE) {
 211                 mandoc_msg(MANDOCERR_MACRO, man->parse,
 212                     ln, ppos, buf + ppos - 1);
 213                 return(1);
 214         }
 215
 216         /* Skip a leading escape sequence or tab. */
 217
 218         switch (buf[offs]) {
 219         case '\\':
 220                 cp = buf + offs + 1;
 221                 mandoc_escape(&cp, NULL, NULL);
 222                 offs = cp - buf;
 223                 break;
 224         case '\t':
 225                 offs++;
 226                 break;
 227         default:
 228                 break;
 229         }
 230
 231         /* Jump to the next non-whitespace word. */
 232
 233         while (buf[offs] && buf[offs] == ' ')
 234                 offs++;
 235
 236         /*
 237          * Trailing whitespace.  Note that tabs are allowed to be passed
 238          * into the parser as "text", so we only warn about spaces here.
 239          */
 240
 241         if (buf[offs] == '\0' && buf[offs - 1] == ' ')
 242                 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 243                     ln, offs - 1, NULL);
 244
 245         /*
 246          * Some macros break next-line scopes; otherwise, remember
 247          * whether we are in next-line scope for a block head.
 248          */
 249
 250         man_breakscope(man, tok);
 251         bline = man->flags & MAN_BLINE;
 252
 253         /* Call to handler... */
 254
 255         assert(man_macros[tok].fp);
 256         (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
 257
 258         /* In quick mode (for mandocdb), abort after the NAME section. */
 259
 260         if (man->quick && tok == MAN_SH) {
 261                 n = man->last;
 262                 if (n->type == ROFFT_BODY &&
 263                     strcmp(n->prev->child->string, "NAME"))
 264                         return(2);
 265         }
 266
 267         /*
 268          * If we are in a next-line scope for a block head,
 269          * close it out now and switch to the body,
 270          * unless the next-line scope is allowed to continue.
 271          */
 272
 273         if ( ! bline || man->flags & MAN_ELINE ||
 274             man_macros[tok].flags & MAN_NSCOPED)
 275                 return(1);
 276
 277         assert(man->flags & MAN_BLINE);
 278         man->flags &= ~MAN_BLINE;
 279
 280         man_unscope(man, man->last->parent);
 281         roff_body_alloc(man, ln, ppos, man->last->tok);
 282         return(1);
 283 }
 284
 285 void
 286 man_breakscope(struct roff_man *man, int tok)
 287 {
 288         struct roff_node *n;
 289
 290         /*
 291          * An element next line scope is open,
 292          * and the new macro is not allowed inside elements.
 293          * Delete the element that is being broken.
 294          */
 295
 296         if (man->flags & MAN_ELINE && (tok == TOKEN_NONE ||
 297             ! (man_macros[tok].flags & MAN_NSCOPED))) {
 298                 n = man->last;
 299                 assert(n->type != ROFFT_TEXT);
 300                 if (man_macros[n->tok].flags & MAN_NSCOPED)
 301                         n = n->parent;
 302
 303                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 304                     n->line, n->pos, "%s breaks %s",
 305                     tok == TOKEN_NONE ? "TS" : man_macronames[tok],
 306                     man_macronames[n->tok]);
 307
 308                 roff_node_delete(man, n);
 309                 man->flags &= ~MAN_ELINE;
 310         }
 311
 312         /*
 313          * A block header next line scope is open,
 314          * and the new macro is not allowed inside block headers.
 315          * Delete the block that is being broken.
 316          */
 317
 318         if (man->flags & MAN_BLINE && (tok == TOKEN_NONE ||
 319             man_macros[tok].flags & MAN_BSCOPE)) {
 320                 n = man->last;
 321                 if (n->type == ROFFT_TEXT)
 322                         n = n->parent;
 323                 if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
 324                         n = n->parent;
 325
 326                 assert(n->type == ROFFT_HEAD);
 327                 n = n->parent;
 328                 assert(n->type == ROFFT_BLOCK);
 329                 assert(man_macros[n->tok].flags & MAN_SCOPED);
 330
 331                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 332                     n->line, n->pos, "%s breaks %s",
 333                     tok == TOKEN_NONE ? "TS" : man_macronames[tok],
 334                     man_macronames[n->tok]);
 335
 336                 roff_node_delete(man, n);
 337                 man->flags &= ~MAN_BLINE;
 338         }
 339 }
 340
 341 const struct mparse *
 342 man_mparse(const struct roff_man *man)
 343 {
 344
 345         assert(man && man->parse);
 346         return(man->parse);
 347 }
 348
 349 void
 350 man_deroff(char **dest, const struct roff_node *n)
 351 {
 352         char    *cp;
 353         size_t   sz;
 354
 355         if (n->type != ROFFT_TEXT) {
 356                 for (n = n->child; n; n = n->next)
 357                         man_deroff(dest, n);
 358                 return;
 359         }
 360
 361         /* Skip leading whitespace and escape sequences. */
 362
 363         cp = n->string;
 364         while ('\0' != *cp) {
 365                 if ('\\' == *cp) {
 366                         cp++;
 367                         mandoc_escape((const char **)&cp, NULL, NULL);
 368                 } else if (isspace((unsigned char)*cp))
 369                         cp++;
 370                 else
 371                         break;
 372         }
 373
 374         /* Skip trailing whitespace. */
 375
 376         for (sz = strlen(cp); sz; sz--)
 377                 if (0 == isspace((unsigned char)cp[sz-1]))
 378                         break;
 379
 380         /* Skip empty strings. */
 381
 382         if (0 == sz)
 383                 return;
 384
 385         if (NULL == *dest) {
 386                 *dest = mandoc_strndup(cp, sz);
 387                 return;
 388         }
 389
 390         mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
 391         free(*dest);
 392         *dest = cp;
 393 }