strfile.c revision 33633 
1/*-
2 * Copyright (c) 1989, 1993
3 *	The Regents of the University of California.  All rights reserved.
4 *
5 * This code is derived from software contributed to Berkeley by
6 * Ken Arnold.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 *    notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 *    notice, this list of conditions and the following disclaimer in the
15 *    documentation and/or other materials provided with the distribution.
16 * 3. All advertising materials mentioning features or use of this software
17 *    must display the following acknowledgement:
18 *	This product includes software developed by the University of
19 *	California, Berkeley and its contributors.
20 * 4. Neither the name of the University nor the names of its contributors
21 *    may be used to endorse or promote products derived from this software
22 *    without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 */
36
37#ifndef lint
38static const char copyright[] =
39"@(#) Copyright (c) 1989, 1993\n\
40	The Regents of the University of California.  All rights reserved.\n";
41#endif /* not lint */
42
43#ifndef lint
44static const char sccsid[] = "@(#)strfile.c   8.1 (Berkeley) 5/31/93";
45#endif /* not lint */
46
47# include	<sys/param.h>
48# include	<stdio.h>
49# include       <stdlib.h>
50# include	<ctype.h>
51# include       <string.h>
52# include       <time.h>
53# include       <locale.h>
54# include       <unistd.h>
55# include	"strfile.h"
56
57/*
58 *	This program takes a file composed of strings seperated by
59 * lines starting with two consecutive delimiting character (default
60 * character is '%') and creates another file which consists of a table
61 * describing the file (structure from "strfile.h"), a table of seek
62 * pointers to the start of the strings, and the strings, each terminated
63 * by a null byte.  Usage:
64 *
65 *	% strfile [-iorsx] [ -cC ] sourcefile [ datafile ]
66 *
67 *	c - Change delimiting character from '%' to 'C'
68 *	s - Silent.  Give no summary of data processed at the end of
69 *	    the run.
70 *	o - order the strings in alphabetic order
71 *	i - if ordering, ignore case
72 *	r - randomize the order of the strings
73 *	x - set rotated bit
74 *
75 *		Ken Arnold	Sept. 7, 1978 --
76 *
77 *	Added ordering options.
78 */
79
80# define	TRUE	1
81# define	FALSE	0
82
83# define	STORING_PTRS	(Oflag || Rflag)
84# define	CHUNKSIZE	512
85
86# define        ALLOC(ptr,sz) { \
87			if (ptr == NULL) \
88				ptr = malloc((unsigned int) (CHUNKSIZE * sizeof *ptr)); \
89			else if (((sz) + 1) % CHUNKSIZE == 0) \
90				ptr = realloc((void *) ptr, ((unsigned int) ((sz) + CHUNKSIZE) * sizeof *ptr)); \
91			if (ptr == NULL) { \
92				fprintf(stderr, "out of space\n"); \
93				exit(1); \
94			} \
95		}
96
97#ifdef NO_VOID
98# define	void	char
99#endif
100
101typedef struct {
102	char	first;
103	long    pos;
104} STR;
105
106char	*Infile		= NULL,		/* input file name */
107	Outfile[MAXPATHLEN] = "",	/* output file name */
108	Delimch		= '%';		/* delimiting character */
109
110int	Sflag		= FALSE;	/* silent run flag */
111int	Oflag		= FALSE;	/* ordering flag */
112int	Iflag		= FALSE;	/* ignore case flag */
113int	Rflag		= FALSE;	/* randomize order flag */
114int	Xflag		= FALSE;	/* set rotated bit */
115long	Num_pts		= 0;		/* number of pointers/strings */
116
117long    *Seekpts;
118
119FILE	*Sort_1, *Sort_2;		/* pointers for sorting */
120
121STRFILE	Tbl;				/* statistics table */
122
123STR	*Firstch;			/* first chars of each string */
124
125void getargs(), add_offset(), do_order(), randomize(), usage();
126int cmp_str();
127
128/*
129 * main:
130 *	Drive the sucker.  There are two main modes -- either we store
131 *	the seek pointers, if the table is to be sorted or randomized,
132 *	or we write the pointer directly to the file, if we are to stay
133 *	in file order.  If the former, we allocate and re-allocate in
134 *	CHUNKSIZE blocks; if the latter, we just write each pointer,
135 *	and then seek back to the beginning to write in the table.
136 */
137void main(ac, av)
138int	ac;
139char	**av;
140{
141	register char		*sp, dc;
142	register FILE		*inf, *outf;
143	register long           last_off, length, pos, *p;
144	register int		first, cnt;
145	register char		*nsp;
146	register STR		*fp;
147	static char		string[257];
148
149	(void) setlocale(LC_ALL, "");
150
151	getargs(ac, av);		/* evalute arguments */
152	dc = Delimch;
153	if ((inf = fopen(Infile, "r")) == NULL) {
154		perror(Infile);
155		exit(1);
156	}
157
158	if ((outf = fopen(Outfile, "w")) == NULL) {
159		perror(Outfile);
160		exit(1);
161	}
162	if (!STORING_PTRS)
163		(void) fseek(outf, (long) sizeof Tbl, 0);
164
165	/*
166	 * Write the strings onto the file
167	 */
168
169	Tbl.str_longlen = 0;
170	Tbl.str_shortlen = ~((unsigned long) 0);
171	Tbl.str_delim = dc;
172	Tbl.str_version = VERSION;
173	first = Oflag;
174	add_offset(outf, ftell(inf));
175	last_off = 0;
176	do {
177		sp = fgets(string, 256, inf);
178		if (sp == NULL || (sp[0] == dc && sp[1] == '\n')) {
179			pos = ftell(inf);
180			length = pos - last_off - (sp ? strlen(sp) : 0);
181			last_off = pos;
182			if (!length)
183				continue;
184			add_offset(outf, pos);
185			if (Tbl.str_longlen < length)
186				Tbl.str_longlen = length;
187			if (Tbl.str_shortlen > length)
188				Tbl.str_shortlen = length;
189			first = Oflag;
190		}
191		else if (first) {
192			for (nsp = sp; !isalnum((unsigned char)*nsp); nsp++)
193				continue;
194			ALLOC(Firstch, Num_pts);
195			fp = &Firstch[Num_pts - 1];
196			if (Iflag && isupper((unsigned char)*nsp))
197				fp->first = tolower((unsigned char)*nsp);
198			else
199				fp->first = *nsp;
200			fp->pos = Seekpts[Num_pts - 1];
201			first = FALSE;
202		}
203	} while (sp != NULL);
204
205	/*
206	 * write the tables in
207	 */
208
209	(void) fclose(inf);
210	Tbl.str_numstr = Num_pts - 1;
211
212	if (Oflag)
213		do_order();
214	else if (Rflag)
215		randomize();
216
217	if (Xflag)
218		Tbl.str_flags |= STR_ROTATED;
219
220	if (!Sflag) {
221		printf("\"%s\" created\n", Outfile);
222		if (Num_pts == 2)
223			puts("There was 1 string");
224		else
225			printf("There were %ld strings\n", Num_pts - 1);
226		printf("Longest string: %lu byte%s\n", Tbl.str_longlen,
227		       Tbl.str_longlen == 1 ? "" : "s");
228		printf("Shortest string: %lu byte%s\n", Tbl.str_shortlen,
229		       Tbl.str_shortlen == 1 ? "" : "s");
230	}
231
232	rewind(outf);
233	Tbl.str_version = htonl(Tbl.str_version);
234	Tbl.str_numstr = htonl(Tbl.str_numstr);
235	Tbl.str_longlen = htonl(Tbl.str_longlen);
236	Tbl.str_shortlen = htonl(Tbl.str_shortlen);
237	Tbl.str_flags = htonl(Tbl.str_flags);
238	(void) fwrite((char *) &Tbl, sizeof Tbl, 1, outf);
239	if (STORING_PTRS) {
240		for (p = Seekpts, cnt = Num_pts; cnt--; ++p)
241			*p = htonl(*p);
242		(void) fwrite((char *) Seekpts, sizeof *Seekpts, (int) Num_pts, outf);
243	}
244	(void) fclose(outf);
245	exit(0);
246}
247
248/*
249 *	This routine evaluates arguments from the command line
250 */
251void getargs(argc, argv)
252int	argc;
253char	**argv;
254{
255	extern char	*optarg;
256	extern int	optind;
257	int	ch;
258
259	while ((ch = getopt(argc, argv, "c:iorsx")) != EOF)
260		switch(ch) {
261		case 'c':			/* new delimiting char */
262			Delimch = *optarg;
263			if (!isascii(Delimch)) {
264				printf("bad delimiting character: '\\%o\n'",
265				       (unsigned char)Delimch);
266			}
267			break;
268		case 'i':			/* ignore case in ordering */
269			Iflag++;
270			break;
271		case 'o':			/* order strings */
272			Oflag++;
273			break;
274		case 'r':			/* randomize pointers */
275			Rflag++;
276			break;
277		case 's':			/* silent */
278			Sflag++;
279			break;
280		case 'x':			/* set the rotated bit */
281			Xflag++;
282			break;
283		case '?':
284		default:
285			usage();
286		}
287	argv += optind;
288
289	if (*argv) {
290		Infile = *argv;
291		if (*++argv)
292			(void) strcpy(Outfile, *argv);
293	}
294	if (!Infile) {
295		puts("No input file name");
296		usage();
297	}
298	if (*Outfile == '\0') {
299		(void) strcpy(Outfile, Infile);
300		(void) strcat(Outfile, ".dat");
301	}
302}
303
304void usage()
305{
306	(void) fprintf(stderr,
307	    "strfile [-iorsx] [-c char] sourcefile [datafile]\n");
308	exit(1);
309}
310
311/*
312 * add_offset:
313 *	Add an offset to the list, or write it out, as appropriate.
314 */
315void add_offset(fp, off)
316FILE	*fp;
317long    off;
318{
319	long net;
320
321	if (!STORING_PTRS) {
322		net = htonl(off);
323		fwrite(&net, 1, sizeof net, fp);
324	} else {
325		ALLOC(Seekpts, Num_pts + 1);
326		Seekpts[Num_pts] = off;
327	}
328	Num_pts++;
329}
330
331/*
332 * do_order:
333 *	Order the strings alphabetically (possibly ignoring case).
334 */
335void do_order()
336{
337	register int	i;
338	register long   *lp;
339	register STR	*fp;
340
341	Sort_1 = fopen(Infile, "r");
342	Sort_2 = fopen(Infile, "r");
343	qsort((char *) Firstch, (int) Tbl.str_numstr, sizeof *Firstch, cmp_str);
344	i = Tbl.str_numstr;
345	lp = Seekpts;
346	fp = Firstch;
347	while (i--)
348		*lp++ = fp++->pos;
349	(void) fclose(Sort_1);
350	(void) fclose(Sort_2);
351	Tbl.str_flags |= STR_ORDERED;
352}
353
354static int collate_range_cmp (c1, c2)
355	int c1, c2;
356{
357	static char s1[2], s2[2];
358	int ret;
359
360	c1 &= UCHAR_MAX;
361	c2 &= UCHAR_MAX;
362	if (c1 == c2)
363		return (0);
364	s1[0] = c1;
365	s2[0] = c2;
366	if ((ret = strcoll(s1, s2)) != 0)
367		return (ret);
368	return (c1 - c2);
369}
370
371/*
372 * cmp_str:
373 *	Compare two strings in the file
374 */
375int cmp_str(p1, p2)
376STR	*p1, *p2;
377{
378	register int	c1, c2;
379	register int	n1, n2;
380	int r;
381
382# define	SET_N(nf,ch)	(nf = (ch == '\n'))
383# define        IS_END(ch,nf)   (ch == EOF || (ch == (unsigned char) Delimch && nf))
384
385	c1 = (unsigned char) p1->first;
386	c2 = (unsigned char) p2->first;
387	if ((r = collate_range_cmp(c1, c2)) != 0)
388		return r;
389
390	(void) fseek(Sort_1, p1->pos, 0);
391	(void) fseek(Sort_2, p2->pos, 0);
392
393	n1 = FALSE;
394	n2 = FALSE;
395	while (!isalnum(c1 = getc(Sort_1)) && c1 != '\0' && c1 != EOF)
396		SET_N(n1, c1);
397	while (!isalnum(c2 = getc(Sort_2)) && c2 != '\0' && c2 != EOF)
398		SET_N(n2, c2);
399
400	while (!IS_END(c1, n1) && !IS_END(c2, n2)) {
401		if (Iflag) {
402			if (isupper(c1))
403				c1 = tolower(c1);
404			if (isupper(c2))
405				c2 = tolower(c2);
406		}
407		if ((r = collate_range_cmp(c1, c2)) != 0)
408			return r;
409		SET_N(n1, c1);
410		SET_N(n2, c2);
411		c1 = getc(Sort_1);
412		c2 = getc(Sort_2);
413	}
414	if (IS_END(c1, n1))
415		c1 = 0;
416	if (IS_END(c2, n2))
417		c2 = 0;
418	return collate_range_cmp(c1, c2);
419}
420
421/*
422 * randomize:
423 *	Randomize the order of the string table.  We must be careful
424 *	not to randomize across delimiter boundaries.  All
425 *	randomization is done within each block.
426 */
427void randomize()
428{
429	register int	cnt, i;
430	register long   tmp;
431	register long   *sp;
432
433	srandomdev();
434
435	Tbl.str_flags |= STR_RANDOM;
436	cnt = Tbl.str_numstr;
437
438	/*
439	 * move things around randomly
440	 */
441
442	for (sp = Seekpts; cnt > 0; cnt--, sp++) {
443		i = random() % cnt;
444		tmp = sp[0];
445		sp[0] = sp[i];
446		sp[i] = tmp;
447	}
448}
449