mdw@git.distorted.org.uk Git - fwd/blob - conf.c

   1 /* -*-c-*-
   2  *
   3  * $Id: conf.c,v 1.11 2004/04/08 01:36:25 mdw Exp $
   4  *
   5  * Configuration parsing
   6  *
   7  * (c) 1999 Straylight/Edgeware
   8  */
   9
  10 /*----- Licensing notice --------------------------------------------------*
  11  *
  12  * This file is part of the `fw' port forwarder.
  13  *
  14  * `fw' is free software; you can redistribute it and/or modify
  15  * it under the terms of the GNU General Public License as published by
  16  * the Free Software Foundation; either version 2 of the License, or
  17  * (at your option) any later version.
  18  *
  19  * `fw' is distributed in the hope that it will be useful,
  20  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  21  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22  * GNU General Public License for more details.
  23  *
  24  * You should have received a copy of the GNU General Public License
  25  * along with `fw'; if not, write to the Free Software Foundation,
  26  * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  27  */
  28
  29 /*----- Header files ------------------------------------------------------*/
  30
  31 #include <ctype.h>
  32 #include <errno.h>
  33 #include <stdarg.h>
  34 #include <stdio.h>
  35 #include <stdlib.h>
  36 #include <string.h>
  37
  38 #include <mLib/dstr.h>
  39 #include <mLib/quis.h>
  40 #include <mLib/report.h>
  41
  42 #include "conf.h"
  43 #include "scan.h"
  44
  45 /*----- Main code ---------------------------------------------------------*/
  46
  47 /* --- @conf_undelim@ --- *
  48  *
  49  * Arguments:   @scanner *sc@ = pointer to scanner definition
  50  *              @const char *d, *dd@ = pointer to characters to escape
  51  *
  52  * Returns:     ---
  53  *
  54  * Use:         Modifies the tokenizer.  Characters in the first list will
  55  *              always be considered to begin a word.  Characters in the
  56  *              second list will always be allowed to continue a word.
  57  */
  58
  59 void conf_undelim(scanner *sc, const char *d, const char *dd)
  60 {
  61   sc->wbegin = d;
  62   sc->wcont = dd;
  63 }
  64
  65 /* --- @token@ --- *
  66  *
  67  * Arguments:   @scanner *sc@ = pointer to scanner definition
  68  *
  69  * Returns:     Type of token scanned.
  70  *
  71  * Use:         Reads the next token from the character scanner.
  72  */
  73
  74 int token(scanner *sc)
  75 {
  76 #define SELFDELIM \
  77   '{': case '}': case '/': case ',': \
  78   case '=': case ':': case ';': \
  79   case '.': case '[': case ']'
  80
  81   int ch;
  82
  83   DRESET(&sc->d);
  84
  85   /* --- Main tokenization --- */
  86
  87   for (;;) {
  88     ch = scan(sc);
  89
  90     /* --- Deal with pushed-back tokens --- */
  91
  92     if (sc->head->tok) {
  93       dstr_puts(&sc->d, sc->head->tok);
  94       xfree(sc->head->tok);
  95       sc->head->tok = 0;
  96       sc->t = sc->head->t;
  97       goto done;
  98     }
  99
 100     else if (isspace(ch))
 101       ;
 102     else switch (ch) {
 103
 104       /* --- End of file --- */
 105
 106       case EOF:
 107         sc->t = CTOK_EOF;
 108         goto done;
 109
 110       /* --- Comment character --- */
 111
 112       case '#':
 113         do ch = scan(sc); while (ch != EOF && ch != '\n');
 114         break;
 115
 116       /* --- Various self-delimiting characters --- */
 117
 118       case SELFDELIM:
 119         if (!sc->wbegin || strchr(sc->wbegin, ch) == 0) {
 120           dstr_putc(&sc->d, ch);
 121           dstr_putz(&sc->d);
 122           sc->t = ch;
 123           goto done;
 124         }
 125
 126       /* --- Bare words --- *
 127        *
 128        * These aren't as bare any more.  You can now backslash-escape
 129        * individual characters, and enclose sections in double-quotes.
 130        */
 131
 132       default: {
 133         int q = 0;
 134
 135         for (;;) {
 136           switch (ch) {
 137             case EOF:
 138               goto word;
 139             case '\\':
 140               ch = scan(sc);
 141               if (ch == EOF)
 142                 goto word;
 143               DPUTC(&sc->d, ch);
 144               break;
 145             case '\"':
 146               q = !q;
 147               break;
 148             case SELFDELIM:
 149               if (q || (sc->wcont && strchr(sc->wcont, ch)))
 150                 goto insert;
 151               goto word;
 152             default:
 153               if (!q && isspace(ch))
 154                 goto word;
 155             insert:
 156               DPUTC(&sc->d, ch);
 157               break;
 158           }
 159           ch = scan(sc);
 160         }
 161       word:
 162         unscan(sc, ch);
 163         DPUTZ(&sc->d);
 164         sc->t = CTOK_WORD;
 165         goto done;
 166       }
 167     }
 168   }
 169
 170 done:
 171   return (sc->t);
 172 }
 173
 174 /* --- @pushback@ --- *
 175  *
 176  * Arguments:   @scanner *sc@ = pointer to scanner definition
 177  *
 178  * Returns:     ---
 179  *
 180  * Use:         Pushes the current token back.  This is normally a precursor
 181  *              to pushing a new scanner source.
 182  */
 183
 184 void pushback(scanner *sc)
 185 {
 186   sc->head->tok = xstrdup(sc->d.buf);
 187   sc->head->t = sc->t;
 188 }
 189
 190 /* --- @error@ --- *
 191  *
 192  * Arguments:   @scanner *sc@ = pointer to scanner definition
 193  *              @const char *msg@ = message skeleton string
 194  *              @...@ = extra arguments for the skeleton
 195  *
 196  * Returns:     Doesn't
 197  *
 198  * Use:         Reports an error at the current scanner location.
 199  */
 200
 201 void error(scanner *sc, const char *msg, ...)
 202 {
 203   va_list ap;
 204   va_start(ap, msg);
 205   fprintf(stderr, "%s: %s:%i: ", QUIS, sc->head->src, sc->head->line);
 206   vfprintf(stderr, msg, ap);
 207   fputc('\n', stderr);
 208   exit(1);
 209 }
 210
 211 /* --- @conf_enum@ --- *
 212  *
 213  * Arguments:   @scanner *sc@ = pointer to a scanner object
 214  *              @const char *list@ = comma-separated things to allow
 215  *              @unsigned f@ = flags for the search
 216  *              @const char *err@ = error message if not found
 217  *
 218  * Returns:     Index into list, zero-based, or @-1@.
 219  *
 220  * Use:         Checks whether the current token is a string which matches
 221  *              one of the comma-separated items given.  The return value is
 222  *              the index (zero-based) of the matched string in the list.
 223  *
 224  *              The flags control the behaviour if no exact match is found.
 225  *              If @ENUM_ABBREV@ is set, and the current token is a left
 226  *              substring of exactly one of the possibilities, then that one
 227  *              is chosen.  If @ENUM_NONE@ is set, the value @-1@ is
 228  *              returned; otherwise an error is reported and the program is
 229  *              terminated.
 230  */
 231
 232 int conf_enum(scanner *sc, const char *list, unsigned f, const char *err)
 233 {
 234   const char *p, *q;
 235   int chosen = -1;
 236   int ok;
 237   int index;
 238
 239   /* --- Make sure it's a string --- */
 240
 241   if (sc->t != CTOK_WORD)
 242     error(sc, "parse error, expected %s", err);
 243
 244   /* --- Grind through the list --- */
 245
 246   q = sc->d.buf;
 247   ok = 1;
 248   index = 0;
 249   p = list;
 250   for (;;) {
 251     switch (*p) {
 252       case 0:
 253         if (ok && !*q) {
 254           token(sc);
 255           return (index);
 256         } else if (chosen != -1) {
 257           token(sc);
 258           return (chosen);
 259         }
 260         else if (f & ENUM_NONE)
 261           return (-1);
 262         else
 263           error(sc, "unknown %s `%s'", err, sc->d.buf);
 264         break;
 265       case ',':
 266         if (ok && !*q) {
 267           token(sc);
 268           return (index);
 269         }
 270         ok = 1;
 271         q = sc->d.buf;
 272         index++;
 273         break;
 274       default:
 275         if (!ok)
 276           break;
 277         if ((f & ENUM_ABBREV) && !*q) {
 278           if (chosen != -1)
 279             error(sc, "ambiguous %s `%s'", err, sc->d.buf);
 280           chosen = index;
 281           ok = 0;
 282         }
 283         if (*p == *q)
 284           q++;
 285         else
 286           ok = 0;
 287         break;
 288     }
 289     p++;
 290   }
 291 }
 292
 293 /* --- @conf_prefix@ --- *
 294  *
 295  * Arguments:   @scanner *sc@ = pointer to a scanner object
 296  *              @const char *p@ = pointer to prefix string to check
 297  *
 298  * Returns:     Nonzero if the prefix matches.
 299  *
 300  * Use:         If the current token is a word matching the given prefix
 301  *              string, then it and an optional `.' character are removed and
 302  *              a nonzero result is returned.  Otherwise the current token is
 303  *              left as it is, and zero is returned.
 304  *
 305  *              Typical options parsing code would remove an expected prefix,
 306  *              scan an option anyway (since qualifying prefixes are
 307  *              optional) and if a match is found, claim the option.  If no
 308  *              match is found, and a prefix was stripped, then an error
 309  *              should be reported.
 310  */
 311
 312 int conf_prefix(scanner *sc, const char *p)
 313 {
 314   if (sc->t == CTOK_WORD && strcmp(p, sc->d.buf) == 0) {
 315     token(sc);
 316     if (sc->t == '.')
 317       token(sc);
 318     return (1);
 319   }
 320   return (0);
 321 }
 322
 323 /* --- @conf_name@ --- *
 324  *
 325  * Arguments:   @scanner *sc@ = pointer to scanner
 326  *              @char delim@ = delimiter character to look for
 327  *              @dstr *d@ = pointer to dynamic string for output
 328  *
 329  * Returns:     ---
 330  *
 331  * Use:         Reads in a compound name consisting of words separated by
 332  *              delimiters.  Leading and trailing delimiters are permitted,
 333  *              although they'll probably cause confusion if used.  The name
 334  *              may be enclosed in square brackets if that helps at all.
 335  *
 336  *              Examples of compound names are filenames (delimited by `/')
 337  *              and IP addresses (delimited by `.').
 338  */
 339
 340 void conf_name(scanner *sc, char delim, dstr *d)
 341 {
 342   unsigned f = 0;
 343
 344 #define f_ok 1u
 345 #define f_bra 2u
 346
 347   /* --- Read an optional opening bracket --- */
 348
 349   if (sc->t == '[') {
 350     token(sc);
 351     f |= f_bra | f_ok;
 352   }
 353
 354   /* --- Do the main reading sequence --- */
 355
 356   do {
 357     if (sc->t == delim) {
 358       DPUTC(d, delim);
 359       f |= f_ok;
 360       token(sc);
 361     }
 362     if (sc->t == CTOK_WORD) {
 363       DPUTD(d, &sc->d);
 364       f |= f_ok;
 365       token(sc);
 366     }
 367   } while (sc->t == delim);
 368
 369   /* --- Check that the string was OK --- */
 370
 371   if (!(f & f_ok))
 372     error(sc, "parse error, name expected");
 373
 374   /* --- Read a closing bracket --- */
 375
 376   if (f & f_bra) {
 377     if (sc->t == ']')
 378       token(sc);
 379     else
 380       error(sc, "parse error, missing `]'");
 381   }
 382   DPUTZ(d);
 383
 384 #undef f_ok
 385 #undef f_bra
 386 }
 387
 388 /*----- That's all, folks -------------------------------------------------*/