mkindex.c revision 1.10 1 1.10 jmc /* $NetBSD: mkindex.c,v 1.10 2005/07/01 16:38:24 jmc Exp $ */
2 1.2 cgd
3 1.1 jtc /*-
4 1.1 jtc * Copyright (c) 1993
5 1.1 jtc * The Regents of the University of California. All rights reserved.
6 1.1 jtc *
7 1.1 jtc * This code is derived from software contributed to Berkeley by
8 1.1 jtc * Barry Brachman.
9 1.1 jtc *
10 1.1 jtc * Redistribution and use in source and binary forms, with or without
11 1.1 jtc * modification, are permitted provided that the following conditions
12 1.1 jtc * are met:
13 1.1 jtc * 1. Redistributions of source code must retain the above copyright
14 1.1 jtc * notice, this list of conditions and the following disclaimer.
15 1.1 jtc * 2. Redistributions in binary form must reproduce the above copyright
16 1.1 jtc * notice, this list of conditions and the following disclaimer in the
17 1.1 jtc * documentation and/or other materials provided with the distribution.
18 1.9 agc * 3. Neither the name of the University nor the names of its contributors
19 1.1 jtc * may be used to endorse or promote products derived from this software
20 1.1 jtc * without specific prior written permission.
21 1.1 jtc *
22 1.1 jtc * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 1.1 jtc * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 1.1 jtc * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 1.1 jtc * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 1.1 jtc * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 1.1 jtc * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 1.1 jtc * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 1.1 jtc * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 1.1 jtc * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 1.1 jtc * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 1.1 jtc * SUCH DAMAGE.
33 1.1 jtc */
34 1.1 jtc
35 1.1 jtc #ifndef lint
36 1.8 simonb static char copyright[] = "@(#) Copyright (c) 1993\n\
37 1.8 simonb The Regents of the University of California. All rights reserved.\n";
38 1.1 jtc
39 1.2 cgd #if 0
40 1.1 jtc static char sccsid[] = "@(#)mkindex.c 8.1 (Berkeley) 6/11/93";
41 1.2 cgd #else
42 1.8 simonb static char rcsid[] =
43 1.10 jmc "$NetBSD: mkindex.c,v 1.10 2005/07/01 16:38:24 jmc Exp $";
44 1.2 cgd #endif
45 1.1 jtc #endif /* not lint */
46 1.1 jtc
47 1.1 jtc #include <stdio.h>
48 1.6 jsm #include <stdlib.h>
49 1.1 jtc
50 1.1 jtc #include "bog.h"
51 1.1 jtc
52 1.8 simonb char *nextword(FILE *, char *, int *, int *);
53 1.1 jtc
54 1.1 jtc int
55 1.4 hubertf main(void)
56 1.1 jtc {
57 1.7 jsm int clen, rlen, prev, i;
58 1.1 jtc long off, start;
59 1.1 jtc char buf[MAXWORDLEN + 1];
60 1.1 jtc
61 1.1 jtc prev = '\0';
62 1.1 jtc off = start = 0L;
63 1.1 jtc while (nextword(stdin, buf, &clen, &rlen) != NULL) {
64 1.1 jtc if (*buf != prev) {
65 1.7 jsm /*
66 1.7 jsm * Boggle expects a full index even if the dictionary
67 1.7 jsm * had no words beginning with some letters.
68 1.7 jsm * So we write out entries for every letter from prev
69 1.7 jsm * to *buf.
70 1.7 jsm */
71 1.1 jtc if (prev != '\0')
72 1.1 jtc printf("%c %6ld %6ld\n", prev, start, off - 1);
73 1.7 jsm for (i = (prev ? prev + 1 : 'a'); i < *buf; i++)
74 1.7 jsm printf("%c %6ld %6ld\n", i, off, off - 1);
75 1.1 jtc prev = *buf;
76 1.1 jtc start = off;
77 1.1 jtc }
78 1.1 jtc off += clen + 1;
79 1.1 jtc }
80 1.1 jtc printf("%c %6ld %6ld\n", prev, start, off - 1);
81 1.7 jsm for (i = prev + 1; i <= 'z'; i++)
82 1.7 jsm printf("%c %6ld %6ld\n", i, off, off - 1);
83 1.5 jsm fflush(stdout);
84 1.8 simonb if (ferror(stdout)) {
85 1.8 simonb perror("error writing standard output");
86 1.8 simonb exit(1);
87 1.8 simonb }
88 1.1 jtc exit(0);
89 1.1 jtc }
90 1.1 jtc
91 1.1 jtc /*
92 1.1 jtc * Return the next word in the compressed dictionary in 'buffer' or
93 1.1 jtc * NULL on end-of-file
94 1.1 jtc * Also set clen to the length of the compressed word (for mkindex) and
95 1.1 jtc * rlen to the strlen() of the real word
96 1.1 jtc */
97 1.1 jtc char *
98 1.10 jmc nextword(FILE *fp, char *buffer, int *clen, int *rlen)
99 1.1 jtc {
100 1.3 lukem int ch, pcount;
101 1.3 lukem char *p, *q;
102 1.1 jtc static char buf[MAXWORDLEN + 1];
103 1.1 jtc static int first = 1;
104 1.1 jtc static int lastch = 0;
105 1.1 jtc
106 1.1 jtc if (first) {
107 1.1 jtc if ((pcount = getc(fp)) == EOF)
108 1.1 jtc return (NULL);
109 1.1 jtc first = 0;
110 1.1 jtc }
111 1.1 jtc else if ((pcount = lastch) == EOF)
112 1.1 jtc return (NULL);
113 1.1 jtc
114 1.1 jtc p = buf + (*clen = pcount);
115 1.1 jtc
116 1.1 jtc while ((ch = getc(fp)) != EOF && ch >= 'a')
117 1.1 jtc *p++ = ch;
118 1.1 jtc lastch = ch;
119 1.1 jtc *p = '\0';
120 1.1 jtc
121 1.1 jtc *rlen = (int) (p - buf);
122 1.1 jtc *clen = *rlen - *clen;
123 1.1 jtc
124 1.1 jtc p = buf;
125 1.1 jtc q = buffer;
126 1.1 jtc while ((*q++ = *p) != '\0') {
127 1.1 jtc if (*p++ == 'q')
128 1.1 jtc *q++ = 'u';
129 1.1 jtc }
130 1.1 jtc return (buffer);
131 1.1 jtc }
132