Home | History | Annotate | Line # | Download | only in mkdict
mkdict.c revision 1.8
      1  1.8     jsm /* $NetBSD: mkdict.c,v 1.8 2001/08/29 18:22:56 jsm 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  1.1     jtc  * 3. All advertising materials mentioning features or use of this software
     19  1.1     jtc  *    must display the following acknowledgement:
     20  1.1     jtc  *	This product includes software developed by the University of
     21  1.1     jtc  *	California, Berkeley and its contributors.
     22  1.1     jtc  * 4. Neither the name of the University nor the names of its contributors
     23  1.1     jtc  *    may be used to endorse or promote products derived from this software
     24  1.1     jtc  *    without specific prior written permission.
     25  1.1     jtc  *
     26  1.1     jtc  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     27  1.1     jtc  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     28  1.1     jtc  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     29  1.1     jtc  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     30  1.1     jtc  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     31  1.1     jtc  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     32  1.1     jtc  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     33  1.1     jtc  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     34  1.1     jtc  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     35  1.1     jtc  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     36  1.1     jtc  * SUCH DAMAGE.
     37  1.1     jtc  */
     38  1.1     jtc 
     39  1.1     jtc #ifndef lint
     40  1.7  simonb static const char copyright[] =
     41  1.7  simonb     "@(#) Copyright (c) 1993\n\
     42  1.7  simonb 	The Regents of the University of California.  All rights reserved.\n";
     43  1.2     cgd #if 0
     44  1.1     jtc static char sccsid[] = "@(#)mkdict.c	8.1 (Berkeley) 6/11/93";
     45  1.2     cgd #else
     46  1.7  simonb static const char rcsid[] =
     47  1.8     jsm     "$NetBSD: mkdict.c,v 1.8 2001/08/29 18:22:56 jsm Exp $";
     48  1.2     cgd #endif
     49  1.1     jtc #endif /* not lint */
     50  1.1     jtc 
     51  1.1     jtc /*
     52  1.1     jtc  * Filter out words that:
     53  1.1     jtc  *	1) Are not completely made up of lower case letters
     54  1.1     jtc  *	2) Contain a 'q' not immediately followed by a 'u'
     55  1.1     jtc  *	3) Are less that 3 characters long
     56  1.1     jtc  *	4) Are greater than MAXWORDLEN characters long
     57  1.1     jtc  */
     58  1.1     jtc 
     59  1.1     jtc #include <ctype.h>
     60  1.1     jtc #include <stdio.h>
     61  1.1     jtc #include <stdlib.h>
     62  1.1     jtc #include <string.h>
     63  1.1     jtc 
     64  1.1     jtc #include "bog.h"
     65  1.1     jtc 
     66  1.7  simonb int main(int, char *[]);
     67  1.5     jsm 
     68  1.1     jtc int
     69  1.1     jtc main(argc, argv)
     70  1.1     jtc 	int argc;
     71  1.1     jtc 	char *argv[];
     72  1.1     jtc {
     73  1.3   lukem 	char *p, *q;
     74  1.6     jsm 	int ch, common, nwords;
     75  1.1     jtc 	int current, len, prev, qcount;
     76  1.1     jtc 	char buf[2][MAXWORDLEN + 1];
     77  1.1     jtc 
     78  1.1     jtc 	prev = 0;
     79  1.1     jtc 	current = 1;
     80  1.1     jtc 	buf[prev][0] = '\0';
     81  1.1     jtc 
     82  1.1     jtc 	for (nwords = 1;
     83  1.1     jtc 	    fgets(buf[current], MAXWORDLEN + 1, stdin) != NULL; ++nwords) {
     84  1.3   lukem 		if ((p = strchr(buf[current], '\n')) == NULL) {
     85  1.8     jsm 			fprintf(stderr, "word too long: %s\n", buf[current]);
     86  1.1     jtc 			while ((ch = getc(stdin)) != EOF && ch != '\n')
     87  1.1     jtc 				;
     88  1.1     jtc 			if (ch == EOF)
     89  1.1     jtc 				break;
     90  1.1     jtc 			continue;
     91  1.1     jtc 		}
     92  1.1     jtc 		len = 0;
     93  1.1     jtc 		for (p = buf[current]; *p != '\n'; p++) {
     94  1.1     jtc 			if (!islower(*p))
     95  1.1     jtc 				break;
     96  1.1     jtc 			if (*p == 'q') {
     97  1.1     jtc 				q = p + 1;
     98  1.1     jtc 				if (*q != 'u')
     99  1.1     jtc 					break;
    100  1.1     jtc 				else {
    101  1.5     jsm 					while ((*q = *(q + 1)))
    102  1.1     jtc 						q++;
    103  1.1     jtc 				}
    104  1.1     jtc 				len++;
    105  1.1     jtc 			}
    106  1.1     jtc 			len++;
    107  1.1     jtc 		}
    108  1.1     jtc 		if (*p != '\n' || len < 3 || len > MAXWORDLEN)
    109  1.1     jtc 			continue;
    110  1.6     jsm 		if (argc == 2 && nwords % atoi(argv[1]))
    111  1.1     jtc 			continue;
    112  1.1     jtc 
    113  1.1     jtc 		*p = '\0';
    114  1.1     jtc 		p = buf[current];
    115  1.1     jtc 		q = buf[prev];
    116  1.1     jtc 		qcount = 0;
    117  1.1     jtc 		while ((ch = *p++) == *q++ && ch != '\0')
    118  1.1     jtc 			if (ch == 'q')
    119  1.1     jtc 				qcount++;
    120  1.1     jtc 		common = p - buf[current] - 1;
    121  1.1     jtc 		printf("%c%s", common + qcount, p - 1);
    122  1.1     jtc 		prev = !prev;
    123  1.1     jtc 		current = !current;
    124  1.1     jtc 	}
    125  1.7  simonb 	fprintf(stderr, "%d words\n", nwords);
    126  1.4     jsm 	fflush(stdout);
    127  1.7  simonb 	if (ferror(stdout)) {
    128  1.7  simonb 		perror("error writing standard output");
    129  1.7  simonb 		exit(1);
    130  1.7  simonb 	}
    131  1.1     jtc 	exit(0);
    132  1.1     jtc }
    133