Version bump.
[fwd] / conf.c
1 /* -*-c-*-
2 *
3 * $Id: conf.c,v 1.10 2002/02/22 23:42:56 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 /*----- Revision history --------------------------------------------------*
30 *
31 * $Log: conf.c,v $
32 * Revision 1.10 2002/02/22 23:42:56 mdw
33 * `fw'-specific configuration code moved out. This file might become part
34 * of a library some day.
35 *
36 * Revision 1.9 2002/01/13 14:48:16 mdw
37 * Make delimiters be a property of a scanner. Change the delimiter-
38 * changing functions' names.
39 *
40 * Revision 1.8 2001/02/03 20:33:26 mdw
41 * Fix flags to be unsigned.
42 *
43 * Revision 1.7 2000/08/01 17:58:10 mdw
44 * Fix subtleties with <ctype.h> functions.
45 *
46 * Revision 1.6 2000/02/12 18:13:20 mdw
47 * Terminate tables of sources and targets.
48 *
49 * Revision 1.5 1999/10/22 22:46:44 mdw
50 * Improve documentation for conf_enum.
51 *
52 * Revision 1.4 1999/10/15 21:12:36 mdw
53 * Remove redundant debugging code.
54 *
55 * Revision 1.3 1999/08/19 18:32:48 mdw
56 * Improve lexical analysis. In particular, `chmod' patterns don't have to
57 * be quoted any more.
58 *
59 * Revision 1.2 1999/07/26 23:28:39 mdw
60 * Major reconstruction work for new design.
61 *
62 * Revision 1.1.1.1 1999/07/01 08:56:23 mdw
63 * Initial revision.
64 *
65 */
66
67 /*----- Header files ------------------------------------------------------*/
68
69 #include <ctype.h>
70 #include <errno.h>
71 #include <stdarg.h>
72 #include <stdio.h>
73 #include <stdlib.h>
74 #include <string.h>
75
76 #include <mLib/dstr.h>
77 #include <mLib/quis.h>
78 #include <mLib/report.h>
79
80 #include "conf.h"
81 #include "scan.h"
82
83 /*----- Main code ---------------------------------------------------------*/
84
85 /* --- @conf_undelim@ --- *
86 *
87 * Arguments: @scanner *sc@ = pointer to scanner definition
88 * @const char *d, *dd@ = pointer to characters to escape
89 *
90 * Returns: ---
91 *
92 * Use: Modifies the tokenizer. Characters in the first list will
93 * always be considered to begin a word. Characters in the
94 * second list will always be allowed to continue a word.
95 */
96
97 void conf_undelim(scanner *sc, const char *d, const char *dd)
98 {
99 sc->wbegin = d;
100 sc->wcont = dd;
101 }
102
103 /* --- @token@ --- *
104 *
105 * Arguments: @scanner *sc@ = pointer to scanner definition
106 *
107 * Returns: Type of token scanned.
108 *
109 * Use: Reads the next token from the character scanner.
110 */
111
112 int token(scanner *sc)
113 {
114 #define SELFDELIM \
115 '{': case '}': case '/': case ',': \
116 case '=': case ':': case ';': \
117 case '.': case '[': case ']'
118
119 int ch;
120
121 DRESET(&sc->d);
122
123 /* --- Main tokenization --- */
124
125 for (;;) {
126 ch = scan(sc);
127
128 /* --- Deal with pushed-back tokens --- */
129
130 if (sc->head->tok) {
131 dstr_puts(&sc->d, sc->head->tok);
132 xfree(sc->head->tok);
133 sc->head->tok = 0;
134 sc->t = sc->head->t;
135 goto done;
136 }
137
138 else if (isspace(ch))
139 ;
140 else switch (ch) {
141
142 /* --- End of file --- */
143
144 case EOF:
145 sc->t = CTOK_EOF;
146 goto done;
147
148 /* --- Comment character --- */
149
150 case '#':
151 do ch = scan(sc); while (ch != EOF && ch != '\n');
152 break;
153
154 /* --- Various self-delimiting characters --- */
155
156 case SELFDELIM:
157 if (!sc->wbegin || strchr(sc->wbegin, ch) == 0) {
158 dstr_putc(&sc->d, ch);
159 dstr_putz(&sc->d);
160 sc->t = ch;
161 goto done;
162 }
163
164 /* --- Bare words --- *
165 *
166 * These aren't as bare any more. You can now backslash-escape
167 * individual characters, and enclose sections in double-quotes.
168 */
169
170 default: {
171 int q = 0;
172
173 for (;;) {
174 switch (ch) {
175 case EOF:
176 goto word;
177 case '\\':
178 ch = scan(sc);
179 if (ch == EOF)
180 goto word;
181 DPUTC(&sc->d, ch);
182 break;
183 case '\"':
184 q = !q;
185 break;
186 case SELFDELIM:
187 if (q || (sc->wcont && strchr(sc->wcont, ch)))
188 goto insert;
189 goto word;
190 default:
191 if (!q && isspace(ch))
192 goto word;
193 insert:
194 DPUTC(&sc->d, ch);
195 break;
196 }
197 ch = scan(sc);
198 }
199 word:
200 unscan(sc, ch);
201 DPUTZ(&sc->d);
202 sc->t = CTOK_WORD;
203 goto done;
204 }
205 }
206 }
207
208 done:
209 return (sc->t);
210 }
211
212 /* --- @pushback@ --- *
213 *
214 * Arguments: @scanner *sc@ = pointer to scanner definition
215 *
216 * Returns: ---
217 *
218 * Use: Pushes the current token back. This is normally a precursor
219 * to pushing a new scanner source.
220 */
221
222 void pushback(scanner *sc)
223 {
224 sc->head->tok = xstrdup(sc->d.buf);
225 sc->head->t = sc->t;
226 }
227
228 /* --- @error@ --- *
229 *
230 * Arguments: @scanner *sc@ = pointer to scanner definition
231 * @const char *msg@ = message skeleton string
232 * @...@ = extra arguments for the skeleton
233 *
234 * Returns: Doesn't
235 *
236 * Use: Reports an error at the current scanner location.
237 */
238
239 void error(scanner *sc, const char *msg, ...)
240 {
241 va_list ap;
242 va_start(ap, msg);
243 fprintf(stderr, "%s: %s:%i: ", QUIS, sc->head->src, sc->head->line);
244 vfprintf(stderr, msg, ap);
245 fputc('\n', stderr);
246 exit(1);
247 }
248
249 /* --- @conf_enum@ --- *
250 *
251 * Arguments: @scanner *sc@ = pointer to a scanner object
252 * @const char *list@ = comma-separated things to allow
253 * @unsigned f@ = flags for the search
254 * @const char *err@ = error message if not found
255 *
256 * Returns: Index into list, zero-based, or @-1@.
257 *
258 * Use: Checks whether the current token is a string which matches
259 * one of the comma-separated items given. The return value is
260 * the index (zero-based) of the matched string in the list.
261 *
262 * The flags control the behaviour if no exact match is found.
263 * If @ENUM_ABBREV@ is set, and the current token is a left
264 * substring of exactly one of the possibilities, then that one
265 * is chosen. If @ENUM_NONE@ is set, the value @-1@ is
266 * returned; otherwise an error is reported and the program is
267 * terminated.
268 */
269
270 int conf_enum(scanner *sc, const char *list, unsigned f, const char *err)
271 {
272 const char *p, *q;
273 int chosen = -1;
274 int ok;
275 int index;
276
277 /* --- Make sure it's a string --- */
278
279 if (sc->t != CTOK_WORD)
280 error(sc, "parse error, expected %s", err);
281
282 /* --- Grind through the list --- */
283
284 q = sc->d.buf;
285 ok = 1;
286 index = 0;
287 p = list;
288 for (;;) {
289 switch (*p) {
290 case 0:
291 if (ok && !*q) {
292 token(sc);
293 return (index);
294 } else if (chosen != -1) {
295 token(sc);
296 return (chosen);
297 }
298 else if (f & ENUM_NONE)
299 return (-1);
300 else
301 error(sc, "unknown %s `%s'", err, sc->d.buf);
302 break;
303 case ',':
304 if (ok && !*q) {
305 token(sc);
306 return (index);
307 }
308 ok = 1;
309 q = sc->d.buf;
310 index++;
311 break;
312 default:
313 if (!ok)
314 break;
315 if ((f & ENUM_ABBREV) && !*q) {
316 if (chosen != -1)
317 error(sc, "ambiguous %s `%s'", err, sc->d.buf);
318 chosen = index;
319 ok = 0;
320 }
321 if (*p == *q)
322 q++;
323 else
324 ok = 0;
325 break;
326 }
327 p++;
328 }
329 }
330
331 /* --- @conf_prefix@ --- *
332 *
333 * Arguments: @scanner *sc@ = pointer to a scanner object
334 * @const char *p@ = pointer to prefix string to check
335 *
336 * Returns: Nonzero if the prefix matches.
337 *
338 * Use: If the current token is a word matching the given prefix
339 * string, then it and an optional `.' character are removed and
340 * a nonzero result is returned. Otherwise the current token is
341 * left as it is, and zero is returned.
342 *
343 * Typical options parsing code would remove an expected prefix,
344 * scan an option anyway (since qualifying prefixes are
345 * optional) and if a match is found, claim the option. If no
346 * match is found, and a prefix was stripped, then an error
347 * should be reported.
348 */
349
350 int conf_prefix(scanner *sc, const char *p)
351 {
352 if (sc->t == CTOK_WORD && strcmp(p, sc->d.buf) == 0) {
353 token(sc);
354 if (sc->t == '.')
355 token(sc);
356 return (1);
357 }
358 return (0);
359 }
360
361 /* --- @conf_name@ --- *
362 *
363 * Arguments: @scanner *sc@ = pointer to scanner
364 * @char delim@ = delimiter character to look for
365 * @dstr *d@ = pointer to dynamic string for output
366 *
367 * Returns: ---
368 *
369 * Use: Reads in a compound name consisting of words separated by
370 * delimiters. Leading and trailing delimiters are permitted,
371 * although they'll probably cause confusion if used. The name
372 * may be enclosed in square brackets if that helps at all.
373 *
374 * Examples of compound names are filenames (delimited by `/')
375 * and IP addresses (delimited by `.').
376 */
377
378 void conf_name(scanner *sc, char delim, dstr *d)
379 {
380 unsigned f = 0;
381
382 #define f_ok 1u
383 #define f_bra 2u
384
385 /* --- Read an optional opening bracket --- */
386
387 if (sc->t == '[') {
388 token(sc);
389 f |= f_bra | f_ok;
390 }
391
392 /* --- Do the main reading sequence --- */
393
394 do {
395 if (sc->t == delim) {
396 DPUTC(d, delim);
397 f |= f_ok;
398 token(sc);
399 }
400 if (sc->t == CTOK_WORD) {
401 DPUTD(d, &sc->d);
402 f |= f_ok;
403 token(sc);
404 }
405 } while (sc->t == delim);
406
407 /* --- Check that the string was OK --- */
408
409 if (!(f & f_ok))
410 error(sc, "parse error, name expected");
411
412 /* --- Read a closing bracket --- */
413
414 if (f & f_bra) {
415 if (sc->t == ']')
416 token(sc);
417 else
418 error(sc, "parse error, missing `]'");
419 }
420 DPUTZ(d);
421
422 #undef f_ok
423 #undef f_bra
424 }
425
426 /*----- That's all, folks -------------------------------------------------*/