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. Neither the name of the University nor the names of its contributors
17 *    may be used to endorse or promote products derived from this software
18 *    without specific prior written permission.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 * SUCH DAMAGE.
31 */
32
33#if 0
34#ifndef lint
35static const char copyright[] =
36"@(#) Copyright (c) 1989, 1993\n\
37	The Regents of the University of California.  All rights reserved.\n";
38#endif /* not lint */
39
40#ifndef lint
41static const char sccsid[] = "@(#)strfile.c   8.1 (Berkeley) 5/31/93";
42#endif /* not lint */
43#endif
44#include <sys/cdefs.h>
45__FBSDID("$FreeBSD$");
46
47#include <sys/param.h>
48#include <sys/endian.h>
49#include <ctype.h>
50#include <locale.h>
51#include <stdbool.h>
52#include <stdio.h>
53#include <stdlib.h>
54#include <string.h>
55#include <time.h>
56#include <unistd.h>
57
58#include "strfile.h"
59
60/*
61 *	This program takes a file composed of strings separated by
62 * lines starting with two consecutive delimiting character (default
63 * character is '%') and creates another file which consists of a table
64 * describing the file (structure from "strfile.h"), a table of seek
65 * pointers to the start of the strings, and the strings, each terminated
66 * by a null byte.  Usage:
67 *
68 *	% strfile [-iorsx] [ -cC ] sourcefile [ datafile ]
69 *
70 *	C - Allow comments marked by a double delimiter at line's beginning
71 *	c - Change delimiting character from '%' to 'C'
72 *	s - Silent.  Give no summary of data processed at the end of
73 *	    the run.
74 *	o - order the strings in alphabetic order
75 *	i - if ordering, ignore case
76 *	r - randomize the order of the strings
77 *	x - set rotated bit
78 *
79 *		Ken Arnold	Sept. 7, 1978 --
80 *
81 *	Added ordering options.
82 */
83
84#define	STORING_PTRS	(Oflag || Rflag)
85#define	CHUNKSIZE	512
86
87#define		ALLOC(ptr, sz)	do { \
88			if (ptr == NULL) \
89				ptr = malloc(CHUNKSIZE * sizeof(*ptr)); \
90			else if (((sz) + 1) % CHUNKSIZE == 0) \
91				ptr = realloc(ptr, ((sz) + CHUNKSIZE) * sizeof(*ptr)); \
92			if (ptr == NULL) { \
93				fprintf(stderr, "out of space\n"); \
94				exit(1); \
95			} \
96		} while (0)
97
98typedef struct {
99	int	first;
100	off_t	pos;
101} STR;
102
103static char	*Infile		= NULL,		/* input file name */
104		Outfile[MAXPATHLEN] = "",	/* output file name */
105		Delimch		= '%';		/* delimiting character */
106
107static int	Cflag		= false;	/* embedded comments */
108static int	Sflag		= false;	/* silent run flag */
109static int	Oflag		= false;	/* ordering flag */
110static int	Iflag		= false;	/* ignore case flag */
111static int	Rflag		= false;	/* randomize order flag */
112static int	Xflag		= false;	/* set rotated bit */
113static uint32_t	Num_pts		= 0;		/* number of pointers/strings */
114
115static off_t	*Seekpts;
116
117static FILE	*Sort_1, *Sort_2;		/* pointers for sorting */
118
119static STRFILE	Tbl;				/* statistics table */
120
121static STR	*Firstch;			/* first chars of each string */
122
123static void add_offset(FILE *, off_t);
124static int cmp_str(const void *, const void *);
125static int stable_collate_range_cmp(int, int);
126static void do_order(void);
127static void getargs(int, char **);
128static void randomize(void);
129static void usage(void);
130
131/*
132 * main:
133 *	Drive the sucker.  There are two main modes -- either we store
134 *	the seek pointers, if the table is to be sorted or randomized,
135 *	or we write the pointer directly to the file, if we are to stay
136 *	in file order.  If the former, we allocate and re-allocate in
137 *	CHUNKSIZE blocks; if the latter, we just write each pointer,
138 *	and then seek back to the beginning to write in the table.
139 */
140int
141main(int ac, char *av[])
142{
143	char *sp, *nsp, dc;
144	FILE *inf, *outf;
145	off_t last_off, pos, *p;
146	size_t length;
147	int first;
148	uint32_t cnt;
149	STR *fp;
150	static char string[257];
151
152	setlocale(LC_ALL, "");
153
154	getargs(ac, av);		/* evalute arguments */
155	dc = Delimch;
156	if ((inf = fopen(Infile, "r")) == NULL) {
157		perror(Infile);
158		exit(1);
159	}
160
161	if ((outf = fopen(Outfile, "w")) == NULL) {
162		perror(Outfile);
163		exit(1);
164	}
165	if (!STORING_PTRS)
166		fseek(outf, (long)sizeof(Tbl), SEEK_SET);
167
168	/*
169	 * Write the strings onto the file
170	 */
171
172	Tbl.str_longlen = 0;
173	Tbl.str_shortlen = 0xffffffff;
174	Tbl.str_delim = dc;
175	Tbl.str_version = VERSION;
176	first = Oflag;
177	add_offset(outf, ftello(inf));
178	last_off = 0;
179	do {
180		sp = fgets(string, 256, inf);
181		if (sp == NULL || (sp[0] == dc && sp[1] == '\n')) {
182			pos = ftello(inf);
183			length = (size_t)(pos - last_off) -
184			    (sp != NULL ? strlen(sp) : 0);
185			last_off = pos;
186			if (length == 0)
187				continue;
188			add_offset(outf, pos);
189			if ((size_t)Tbl.str_longlen < length)
190				Tbl.str_longlen = length;
191			if ((size_t)Tbl.str_shortlen > length)
192				Tbl.str_shortlen = length;
193			first = Oflag;
194		}
195		else if (first) {
196			for (nsp = sp; !isalnum((unsigned char)*nsp); nsp++)
197				continue;
198			ALLOC(Firstch, Num_pts);
199			fp = &Firstch[Num_pts - 1];
200			if (Iflag && isupper((unsigned char)*nsp))
201				fp->first = tolower((unsigned char)*nsp);
202			else
203				fp->first = *nsp;
204			fp->pos = Seekpts[Num_pts - 1];
205			first = false;
206		}
207	} while (sp != NULL);
208
209	/*
210	 * write the tables in
211	 */
212
213	fclose(inf);
214	Tbl.str_numstr = Num_pts - 1;
215
216	if (Cflag)
217		Tbl.str_flags |= STR_COMMENTS;
218
219	if (Oflag)
220		do_order();
221	else if (Rflag)
222		randomize();
223
224	if (Xflag)
225		Tbl.str_flags |= STR_ROTATED;
226
227	if (!Sflag) {
228		printf("\"%s\" created\n", Outfile);
229		if (Num_pts == 2)
230			puts("There was 1 string");
231		else
232			printf("There were %u strings\n", Num_pts - 1);
233		printf("Longest string: %u byte%s\n", Tbl.str_longlen,
234		       Tbl.str_longlen == 1 ? "" : "s");
235		printf("Shortest string: %u byte%s\n", Tbl.str_shortlen,
236		       Tbl.str_shortlen == 1 ? "" : "s");
237	}
238
239	rewind(outf);
240	Tbl.str_version = htobe32(Tbl.str_version);
241	Tbl.str_numstr = htobe32(Tbl.str_numstr);
242	Tbl.str_longlen = htobe32(Tbl.str_longlen);
243	Tbl.str_shortlen = htobe32(Tbl.str_shortlen);
244	Tbl.str_flags = htobe32(Tbl.str_flags);
245	fwrite((char *)&Tbl, sizeof(Tbl), 1, outf);
246	if (STORING_PTRS) {
247		for (p = Seekpts, cnt = Num_pts; cnt--; ++p)
248			*p = htobe64(*p);
249		fwrite(Seekpts, sizeof(*Seekpts), (size_t)Num_pts, outf);
250	}
251	fclose(outf);
252	exit(0);
253}
254
255/*
256 *	This routine evaluates arguments from the command line
257 */
258void
259getargs(int argc, char **argv)
260{
261	int ch;
262
263	while ((ch = getopt(argc, argv, "Cc:iorsx")) != -1)
264		switch(ch) {
265		case 'C':			/* embedded comments */
266			Cflag++;
267			break;
268		case 'c':			/* new delimiting char */
269			Delimch = *optarg;
270			if (!isascii(Delimch)) {
271				printf("bad delimiting character: '\\%o\n'",
272				       (unsigned char)Delimch);
273			}
274			break;
275		case 'i':			/* ignore case in ordering */
276			Iflag++;
277			break;
278		case 'o':			/* order strings */
279			Oflag++;
280			break;
281		case 'r':			/* randomize pointers */
282			Rflag++;
283			break;
284		case 's':			/* silent */
285			Sflag++;
286			break;
287		case 'x':			/* set the rotated bit */
288			Xflag++;
289			break;
290		case '?':
291		default:
292			usage();
293		}
294	argv += optind;
295
296	if (*argv) {
297		Infile = *argv;
298		if (*++argv)
299			strcpy(Outfile, *argv);
300	}
301	if (!Infile) {
302		puts("No input file name");
303		usage();
304	}
305	if (*Outfile == '\0') {
306		strlcpy(Outfile, Infile, sizeof(Outfile));
307		strlcat(Outfile, ".dat", sizeof(Outfile));
308	}
309}
310
311void
312usage(void)
313{
314	fprintf(stderr,
315	    "strfile [-Ciorsx] [-c char] source_file [output_file]\n");
316	exit(1);
317}
318
319/*
320 * add_offset:
321 *	Add an offset to the list, or write it out, as appropriate.
322 */
323void
324add_offset(FILE *fp, off_t off)
325{
326	off_t beoff;
327
328	if (!STORING_PTRS) {
329		beoff = htobe64(off);
330		fwrite(&beoff, 1, sizeof(beoff), fp);
331	} else {
332		ALLOC(Seekpts, Num_pts + 1);
333		Seekpts[Num_pts] = off;
334	}
335	Num_pts++;
336}
337
338/*
339 * do_order:
340 *	Order the strings alphabetically (possibly ignoring case).
341 */
342void
343do_order(void)
344{
345	uint32_t i;
346	off_t *lp;
347	STR *fp;
348
349	Sort_1 = fopen(Infile, "r");
350	Sort_2 = fopen(Infile, "r");
351	qsort(Firstch, (size_t)Tbl.str_numstr, sizeof(*Firstch), cmp_str);
352	i = Tbl.str_numstr;
353	lp = Seekpts;
354	fp = Firstch;
355	while (i--)
356		*lp++ = fp++->pos;
357	fclose(Sort_1);
358	fclose(Sort_2);
359	Tbl.str_flags |= STR_ORDERED;
360}
361
362static int
363stable_collate_range_cmp(int c1, int c2)
364{
365	static char s1[2], s2[2];
366	int ret;
367
368	s1[0] = c1;
369	s2[0] = c2;
370	if ((ret = strcoll(s1, s2)) != 0)
371		return (ret);
372	return (c1 - c2);
373}
374
375/*
376 * cmp_str:
377 *	Compare two strings in the file
378 */
379int
380cmp_str(const void *s1, const void *s2)
381{
382	const STR *p1, *p2;
383	int c1, c2, n1, n2, r;
384
385#define	SET_N(nf,ch)	(nf = (ch == '\n'))
386#define	IS_END(ch,nf)	(ch == EOF || (ch == (unsigned char)Delimch && nf))
387
388	p1 = (const STR *)s1;
389	p2 = (const STR *)s2;
390
391	c1 = (unsigned char)p1->first;
392	c2 = (unsigned char)p2->first;
393	if ((r = stable_collate_range_cmp(c1, c2)) != 0)
394		return (r);
395
396	fseeko(Sort_1, p1->pos, SEEK_SET);
397	fseeko(Sort_2, p2->pos, SEEK_SET);
398
399	n1 = false;
400	n2 = false;
401	while (!isalnum(c1 = getc(Sort_1)) && c1 != '\0' && c1 != EOF)
402		SET_N(n1, c1);
403	while (!isalnum(c2 = getc(Sort_2)) && c2 != '\0' && c2 != EOF)
404		SET_N(n2, c2);
405
406	while (!IS_END(c1, n1) && !IS_END(c2, n2)) {
407		if (Iflag) {
408			if (isupper(c1))
409				c1 = tolower(c1);
410			if (isupper(c2))
411				c2 = tolower(c2);
412		}
413		if ((r = stable_collate_range_cmp(c1, c2)) != 0)
414			return (r);
415		SET_N(n1, c1);
416		SET_N(n2, c2);
417		c1 = getc(Sort_1);
418		c2 = getc(Sort_2);
419	}
420	if (IS_END(c1, n1))
421		c1 = 0;
422	if (IS_END(c2, n2))
423		c2 = 0;
424
425	return (stable_collate_range_cmp(c1, c2));
426}
427
428/*
429 * randomize:
430 *	Randomize the order of the string table.  We must be careful
431 *	not to randomize across delimiter boundaries.  All
432 *	randomization is done within each block.
433 */
434void
435randomize(void)
436{
437	uint32_t cnt, i;
438	off_t tmp;
439	off_t *sp;
440
441	Tbl.str_flags |= STR_RANDOM;
442	cnt = Tbl.str_numstr;
443
444	/*
445	 * move things around randomly
446	 */
447
448	for (sp = Seekpts; cnt > 0; cnt--, sp++) {
449		i = arc4random_uniform(cnt);
450		tmp = sp[0];
451		sp[0] = sp[i];
452		sp[i] = tmp;
453	}
454}
455