Home | History | Annotate | Line # | Download | only in mkdict
      1  1.12  rillig /* $NetBSD: mkdict.c,v 1.12 2021/05/02 12:50:44 rillig 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   1.7  simonb static const char copyright[] =
     37   1.7  simonb     "@(#) Copyright (c) 1993\n\
     38   1.7  simonb 	The Regents of the University of California.  All rights reserved.\n";
     39   1.2     cgd #if 0
     40   1.1     jtc static char sccsid[] = "@(#)mkdict.c	8.1 (Berkeley) 6/11/93";
     41   1.2     cgd #else
     42  1.12  rillig static const char rcsid[] =
     43  1.12  rillig     "$NetBSD: mkdict.c,v 1.12 2021/05/02 12:50:44 rillig Exp $";
     44   1.2     cgd #endif
     45   1.1     jtc #endif /* not lint */
     46   1.1     jtc 
     47   1.1     jtc /*
     48   1.1     jtc  * Filter out words that:
     49   1.1     jtc  *	1) Are not completely made up of lower case letters
     50   1.1     jtc  *	2) Contain a 'q' not immediately followed by a 'u'
     51   1.1     jtc  *	3) Are less that 3 characters long
     52   1.1     jtc  *	4) Are greater than MAXWORDLEN characters long
     53   1.1     jtc  */
     54   1.1     jtc 
     55   1.1     jtc #include <ctype.h>
     56   1.1     jtc #include <stdio.h>
     57   1.1     jtc #include <stdlib.h>
     58   1.1     jtc #include <string.h>
     59   1.1     jtc 
     60   1.1     jtc #include "bog.h"
     61   1.1     jtc 
     62   1.1     jtc int
     63  1.11     jmc main(int argc, char *argv[])
     64   1.1     jtc {
     65   1.3   lukem 	char *p, *q;
     66   1.6     jsm 	int ch, common, nwords;
     67   1.1     jtc 	int current, len, prev, qcount;
     68   1.1     jtc 	char buf[2][MAXWORDLEN + 1];
     69   1.1     jtc 
     70   1.1     jtc 	prev = 0;
     71   1.1     jtc 	current = 1;
     72   1.1     jtc 	buf[prev][0] = '\0';
     73   1.1     jtc 
     74   1.1     jtc 	for (nwords = 1;
     75   1.1     jtc 	    fgets(buf[current], MAXWORDLEN + 1, stdin) != NULL; ++nwords) {
     76   1.3   lukem 		if ((p = strchr(buf[current], '\n')) == NULL) {
     77   1.8     jsm 			fprintf(stderr, "word too long: %s\n", buf[current]);
     78   1.1     jtc 			while ((ch = getc(stdin)) != EOF && ch != '\n')
     79   1.1     jtc 				;
     80   1.1     jtc 			if (ch == EOF)
     81   1.1     jtc 				break;
     82   1.1     jtc 			continue;
     83   1.1     jtc 		}
     84   1.1     jtc 		len = 0;
     85   1.1     jtc 		for (p = buf[current]; *p != '\n'; p++) {
     86  1.10  rillig 			if (!islower((unsigned char)*p))
     87   1.1     jtc 				break;
     88   1.1     jtc 			if (*p == 'q') {
     89   1.1     jtc 				q = p + 1;
     90   1.1     jtc 				if (*q != 'u')
     91   1.1     jtc 					break;
     92   1.1     jtc 				else {
     93   1.5     jsm 					while ((*q = *(q + 1)))
     94   1.1     jtc 						q++;
     95   1.1     jtc 				}
     96   1.1     jtc 				len++;
     97   1.1     jtc 			}
     98   1.1     jtc 			len++;
     99   1.1     jtc 		}
    100   1.1     jtc 		if (*p != '\n' || len < 3 || len > MAXWORDLEN)
    101   1.1     jtc 			continue;
    102   1.6     jsm 		if (argc == 2 && nwords % atoi(argv[1]))
    103   1.1     jtc 			continue;
    104   1.1     jtc 
    105   1.1     jtc 		*p = '\0';
    106   1.1     jtc 		p = buf[current];
    107   1.1     jtc 		q = buf[prev];
    108   1.1     jtc 		qcount = 0;
    109   1.1     jtc 		while ((ch = *p++) == *q++ && ch != '\0')
    110   1.1     jtc 			if (ch == 'q')
    111   1.1     jtc 				qcount++;
    112   1.1     jtc 		common = p - buf[current] - 1;
    113   1.1     jtc 		printf("%c%s", common + qcount, p - 1);
    114   1.1     jtc 		prev = !prev;
    115   1.1     jtc 		current = !current;
    116   1.1     jtc 	}
    117   1.7  simonb 	fprintf(stderr, "%d words\n", nwords);
    118   1.4     jsm 	fflush(stdout);
    119   1.7  simonb 	if (ferror(stdout)) {
    120   1.7  simonb 		perror("error writing standard output");
    121   1.7  simonb 		exit(1);
    122   1.7  simonb 	}
    123   1.1     jtc 	exit(0);
    124   1.1     jtc }
    125