Deleted Added
sdiff udiff text old ( 201175 ) new ( 203926 )
full compact
strfile.c (201175) strfile.c (203926)
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.
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
16 * 3. 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

--- 12 unchanged lines hidden (view full) ---

41 The Regents of the University of California. All rights reserved.\n";
42#endif /* not lint */
43
44#ifndef lint
45static const char sccsid[] = "@(#)strfile.c 8.1 (Berkeley) 5/31/93";
46#endif /* not lint */
47#endif
48#include <sys/cdefs.h>
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

--- 12 unchanged lines hidden (view full) ---

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>
49__FBSDID("$FreeBSD: head/games/fortune/strfile/strfile.c 201175 2009-12-29 08:42:58Z ed $");
45__FBSDID("$FreeBSD: head/games/fortune/strfile/strfile.c 203926 2010-02-15 15:10:21Z uqs $");
50
46
51# include <sys/param.h>
52# include <sys/endian.h>
53# include <stdio.h>
54# include <stdlib.h>
55# include <ctype.h>
56# include <string.h>
57# include <time.h>
58# include <locale.h>
59# include <unistd.h>
60# include "strfile.h"
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>
61
57
58#include "strfile.h"
59
62/*
63 * This program takes a file composed of strings separated by
64 * lines starting with two consecutive delimiting character (default
65 * character is '%') and creates another file which consists of a table
66 * describing the file (structure from "strfile.h"), a table of seek
67 * pointers to the start of the strings, and the strings, each terminated
68 * by a null byte. Usage:
69 *

--- 8 unchanged lines hidden (view full) ---

78 * r - randomize the order of the strings
79 * x - set rotated bit
80 *
81 * Ken Arnold Sept. 7, 1978 --
82 *
83 * Added ordering options.
84 */
85
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 *

--- 8 unchanged lines hidden (view full) ---

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
86# define TRUE 1
87# define FALSE 0
84#define STORING_PTRS (Oflag || Rflag)
85#define CHUNKSIZE 512
88
86
89# define STORING_PTRS (Oflag || Rflag)
90# define CHUNKSIZE 512
91
92# define ALLOC(ptr,sz) { \
87#define ALLOC(ptr, sz) do { \
93 if (ptr == NULL) \
88 if (ptr == NULL) \
94 ptr = malloc(CHUNKSIZE * sizeof *ptr); \
89 ptr = malloc(CHUNKSIZE * sizeof(*ptr)); \
95 else if (((sz) + 1) % CHUNKSIZE == 0) \
90 else if (((sz) + 1) % CHUNKSIZE == 0) \
96 ptr = realloc(ptr, ((sz) + CHUNKSIZE) * sizeof *ptr); \
91 ptr = realloc(ptr, ((sz) + CHUNKSIZE) * sizeof(*ptr)); \
97 if (ptr == NULL) { \
98 fprintf(stderr, "out of space\n"); \
99 exit(1); \
100 } \
92 if (ptr == NULL) { \
93 fprintf(stderr, "out of space\n"); \
94 exit(1); \
95 } \
101 }
96 } while (0)
102
97
103#ifdef NO_VOID
104# define void char
105#endif
106
107typedef struct {
108 int first;
109 off_t pos;
110} STR;
111
112static char *Infile = NULL, /* input file name */
113 Outfile[MAXPATHLEN] = "", /* output file name */
114 Delimch = '%'; /* delimiting character */
115
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
116static int Cflag = FALSE; /* embedded comments */
117static int Sflag = FALSE; /* silent run flag */
118static int Oflag = FALSE; /* ordering flag */
119static int Iflag = FALSE; /* ignore case flag */
120static int Rflag = FALSE; /* randomize order flag */
121static int Xflag = FALSE; /* set rotated bit */
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 */
122static uint32_t Num_pts = 0; /* number of pointers/strings */
123
124static off_t *Seekpts;
125
126static FILE *Sort_1, *Sort_2; /* pointers for sorting */
127
128static STRFILE Tbl; /* statistics table */
129
130static STR *Firstch; /* first chars of each string */
131
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
132static void add_offset(FILE *, off_t);
133static int cmp_str(const void *, const void *);
134static int stable_collate_range_cmp(int, int);
135static void do_order(void);
136static void getargs(int, char **);
137static void randomize(void);
138static void usage(void);
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);
139
140/*
141 * main:
142 * Drive the sucker. There are two main modes -- either we store
143 * the seek pointers, if the table is to be sorted or randomized,
144 * or we write the pointer directly to the file, if we are to stay
145 * in file order. If the former, we allocate and re-allocate in
146 * CHUNKSIZE blocks; if the latter, we just write each pointer,
147 * and then seek back to the beginning to write in the table.
148 */
149int
150main(int ac, char *av[])
151{
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{
152 char *sp, dc;
153 FILE *inf, *outf;
154 off_t last_off, pos, *p;
155 size_t length;
156 int first;
157 uint32_t cnt;
158 char *nsp;
159 STR *fp;
160 static char string[257];
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];
161
151
162 (void) setlocale(LC_ALL, "");
152 setlocale(LC_ALL, "");
163
164 getargs(ac, av); /* evalute arguments */
165 dc = Delimch;
166 if ((inf = fopen(Infile, "r")) == NULL) {
167 perror(Infile);
168 exit(1);
169 }
170
171 if ((outf = fopen(Outfile, "w")) == NULL) {
172 perror(Outfile);
173 exit(1);
174 }
175 if (!STORING_PTRS)
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)
176 (void) fseek(outf, (long) sizeof Tbl, 0);
166 fseek(outf, (long)sizeof(Tbl), SEEK_SET);
177
178 /*
179 * Write the strings onto the file
180 */
181
182 Tbl.str_longlen = 0;
183 Tbl.str_shortlen = 0xffffffff;
184 Tbl.str_delim = dc;

--- 22 unchanged lines hidden (view full) ---

207 continue;
208 ALLOC(Firstch, Num_pts);
209 fp = &Firstch[Num_pts - 1];
210 if (Iflag && isupper((unsigned char)*nsp))
211 fp->first = tolower((unsigned char)*nsp);
212 else
213 fp->first = *nsp;
214 fp->pos = Seekpts[Num_pts - 1];
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;

--- 22 unchanged lines hidden (view full) ---

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];
215 first = FALSE;
205 first = false;
216 }
217 } while (sp != NULL);
218
219 /*
220 * write the tables in
221 */
222
206 }
207 } while (sp != NULL);
208
209 /*
210 * write the tables in
211 */
212
223 (void) fclose(inf);
213 fclose(inf);
224 Tbl.str_numstr = Num_pts - 1;
225
226 if (Cflag)
227 Tbl.str_flags |= STR_COMMENTS;
228
229 if (Oflag)
230 do_order();
231 else if (Rflag)

--- 15 unchanged lines hidden (view full) ---

247 }
248
249 rewind(outf);
250 Tbl.str_version = htobe32(Tbl.str_version);
251 Tbl.str_numstr = htobe32(Tbl.str_numstr);
252 Tbl.str_longlen = htobe32(Tbl.str_longlen);
253 Tbl.str_shortlen = htobe32(Tbl.str_shortlen);
254 Tbl.str_flags = htobe32(Tbl.str_flags);
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)

--- 15 unchanged lines hidden (view full) ---

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);
255 (void) fwrite((char *) &Tbl, sizeof Tbl, 1, outf);
245 fwrite((char *)&Tbl, sizeof(Tbl), 1, outf);
256 if (STORING_PTRS) {
257 for (p = Seekpts, cnt = Num_pts; cnt--; ++p)
258 *p = htobe64(*p);
246 if (STORING_PTRS) {
247 for (p = Seekpts, cnt = Num_pts; cnt--; ++p)
248 *p = htobe64(*p);
259 (void) fwrite(Seekpts, sizeof *Seekpts, (size_t) Num_pts, outf);
249 fwrite(Seekpts, sizeof(*Seekpts), (size_t)Num_pts, outf);
260 }
250 }
261 (void) fclose(outf);
251 fclose(outf);
262 exit(0);
263}
264
265/*
266 * This routine evaluates arguments from the command line
267 */
268void
269getargs(int argc, char **argv)
270{
252 exit(0);
253}
254
255/*
256 * This routine evaluates arguments from the command line
257 */
258void
259getargs(int argc, char **argv)
260{
271 int ch;
261 int ch;
272
273 while ((ch = getopt(argc, argv, "Cc:iorsx")) != -1)
274 switch(ch) {
275 case 'C': /* embedded comments */
276 Cflag++;
277 break;
278 case 'c': /* new delimiting char */
279 Delimch = *optarg;

--- 21 unchanged lines hidden (view full) ---

301 default:
302 usage();
303 }
304 argv += optind;
305
306 if (*argv) {
307 Infile = *argv;
308 if (*++argv)
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;

--- 21 unchanged lines hidden (view full) ---

291 default:
292 usage();
293 }
294 argv += optind;
295
296 if (*argv) {
297 Infile = *argv;
298 if (*++argv)
309 (void) strcpy(Outfile, *argv);
299 strcpy(Outfile, *argv);
310 }
311 if (!Infile) {
312 puts("No input file name");
313 usage();
314 }
315 if (*Outfile == '\0') {
300 }
301 if (!Infile) {
302 puts("No input file name");
303 usage();
304 }
305 if (*Outfile == '\0') {
316 (void) strcpy(Outfile, Infile);
317 (void) strcat(Outfile, ".dat");
306 strcpy(Outfile, Infile);
307 strcat(Outfile, ".dat");
318 }
319}
320
321void
322usage(void)
323{
308 }
309}
310
311void
312usage(void)
313{
324 (void) fprintf(stderr,
314 fprintf(stderr,
325 "strfile [-Ciorsx] [-c char] source_file [output_file]\n");
326 exit(1);
327}
328
329/*
330 * add_offset:
331 * Add an offset to the list, or write it out, as appropriate.
332 */
333void
334add_offset(FILE *fp, off_t off)
335{
336 off_t beoff;
337
338 if (!STORING_PTRS) {
339 beoff = htobe64(off);
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);
340 fwrite(&beoff, 1, sizeof beoff, fp);
330 fwrite(&beoff, 1, sizeof(beoff), fp);
341 } else {
342 ALLOC(Seekpts, Num_pts + 1);
343 Seekpts[Num_pts] = off;
344 }
345 Num_pts++;
346}
347
348/*
349 * do_order:
350 * Order the strings alphabetically (possibly ignoring case).
351 */
352void
353do_order(void)
354{
355 uint32_t i;
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;
356 off_t *lp;
357 STR *fp;
346 off_t *lp;
347 STR *fp;
358
359 Sort_1 = fopen(Infile, "r");
360 Sort_2 = fopen(Infile, "r");
348
349 Sort_1 = fopen(Infile, "r");
350 Sort_2 = fopen(Infile, "r");
361 qsort(Firstch, (size_t) Tbl.str_numstr, sizeof *Firstch, cmp_str);
351 qsort(Firstch, (size_t)Tbl.str_numstr, sizeof(*Firstch), cmp_str);
362 i = Tbl.str_numstr;
363 lp = Seekpts;
364 fp = Firstch;
365 while (i--)
366 *lp++ = fp++->pos;
352 i = Tbl.str_numstr;
353 lp = Seekpts;
354 fp = Firstch;
355 while (i--)
356 *lp++ = fp++->pos;
367 (void) fclose(Sort_1);
368 (void) fclose(Sort_2);
357 fclose(Sort_1);
358 fclose(Sort_2);
369 Tbl.str_flags |= STR_ORDERED;
370}
371
372static int
373stable_collate_range_cmp(int c1, int c2)
374{
375 static char s1[2], s2[2];
376 int ret;

--- 7 unchanged lines hidden (view full) ---

384
385/*
386 * cmp_str:
387 * Compare two strings in the file
388 */
389int
390cmp_str(const void *s1, const void *s2)
391{
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;

--- 7 unchanged lines hidden (view full) ---

374
375/*
376 * cmp_str:
377 * Compare two strings in the file
378 */
379int
380cmp_str(const void *s1, const void *s2)
381{
392 const STR *p1, *p2;
393 int c1, c2;
394 int n1, n2;
395 int r;
382 const STR *p1, *p2;
383 int c1, c2, n1, n2, r;
396
384
397# define SET_N(nf,ch) (nf = (ch == '\n'))
398# define IS_END(ch,nf) (ch == EOF || (ch == (unsigned char) Delimch && nf))
385#define SET_N(nf,ch) (nf = (ch == '\n'))
386#define IS_END(ch,nf) (ch == EOF || (ch == (unsigned char)Delimch && nf))
399
387
400 p1 = (const STR *) s1;
401 p2 = (const STR *) s2;
402
403 c1 = (unsigned char) p1->first;
404 c2 = (unsigned char) p2->first;
388 p1 = (const STR *)s1;
389 p2 = (const STR *)s2;
390
391 c1 = (unsigned char)p1->first;
392 c2 = (unsigned char)p2->first;
405 if ((r = stable_collate_range_cmp(c1, c2)) != 0)
406 return (r);
407
393 if ((r = stable_collate_range_cmp(c1, c2)) != 0)
394 return (r);
395
408 (void) fseeko(Sort_1, p1->pos, 0);
409 (void) fseeko(Sort_2, p2->pos, 0);
396 fseeko(Sort_1, p1->pos, SEEK_SET);
397 fseeko(Sort_2, p2->pos, SEEK_SET);
410
398
411 n1 = FALSE;
412 n2 = FALSE;
399 n1 = false;
400 n2 = false;
413 while (!isalnum(c1 = getc(Sort_1)) && c1 != '\0' && c1 != EOF)
414 SET_N(n1, c1);
415 while (!isalnum(c2 = getc(Sort_2)) && c2 != '\0' && c2 != EOF)
416 SET_N(n2, c2);
417
418 while (!IS_END(c1, n1) && !IS_END(c2, n2)) {
419 if (Iflag) {
420 if (isupper(c1))

--- 7 unchanged lines hidden (view full) ---

428 SET_N(n2, c2);
429 c1 = getc(Sort_1);
430 c2 = getc(Sort_2);
431 }
432 if (IS_END(c1, n1))
433 c1 = 0;
434 if (IS_END(c2, n2))
435 c2 = 0;
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))

--- 7 unchanged lines hidden (view full) ---

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
436 return (stable_collate_range_cmp(c1, c2));
437}
438
439/*
440 * randomize:
441 * Randomize the order of the string table. We must be careful
442 * not to randomize across delimiter boundaries. All
443 * randomization is done within each block.
444 */
445void
446randomize(void)
447{
448 uint32_t cnt, i;
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;
449 off_t tmp;
450 off_t *sp;
438 off_t tmp;
439 off_t *sp;
451
452#if __FreeBSD_version < 800041
453 srandomdev();
454#endif
455
456 Tbl.str_flags |= STR_RANDOM;
457 cnt = Tbl.str_numstr;
458

--- 15 unchanged lines hidden --- 		440
441#if __FreeBSD_version < 800041
442 srandomdev();
443#endif
444
445 Tbl.str_flags |= STR_RANDOM;
446 cnt = Tbl.str_numstr;
447

--- 15 unchanged lines hidden ---