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 #include <sys/param.h>
34 #include <sys/endian.h>
35 #include <ctype.h>
36 #include <locale.h>
37 #include <stdbool.h>
38 #include <stdio.h>
39 #include <stdlib.h>
40 #include <string.h>
41 #include <time.h>
42 #include <unistd.h>
43
44 #include "strfile.h"
45
46 /*
47 * This program takes a file composed of strings separated by
48 * lines starting with two consecutive delimiting character (default
49 * character is '%') and creates another file which consists of a table
50 * describing the file (structure from "strfile.h"), a table of seek
51 * pointers to the start of the strings, and the strings, each terminated
52 * by a null byte. Usage:
53 *
54 * % strfile [-iorsx] [ -cC ] sourcefile [ datafile ]
55 *
56 * C - Allow comments marked by a double delimiter at line's beginning
57 * c - Change delimiting character from '%' to 'C'
58 * s - Silent. Give no summary of data processed at the end of
59 * the run.
60 * o - order the strings in alphabetic order
61 * i - if ordering, ignore case
62 * r - randomize the order of the strings
63 * x - set rotated bit
64 *
65 * Ken Arnold Sept. 7, 1978 --
66 *
67 * Added ordering options.
68 */
69
70 #define STORING_PTRS (Oflag || Rflag)
71 #define CHUNKSIZE 512
72
73 #define ALLOC(ptr, sz) do { \
74 if (ptr == NULL) \
75 ptr = malloc(CHUNKSIZE * sizeof(*ptr)); \
76 else if (((sz) + 1) % CHUNKSIZE == 0) \
77 ptr = realloc(ptr, ((sz) + CHUNKSIZE) * sizeof(*ptr)); \
78 if (ptr == NULL) { \
79 fprintf(stderr, "out of space\n"); \
80 exit(1); \
81 } \
82 } while (0)
83
84 typedef struct {
85 int first;
86 off_t pos;
87 } STR;
88
89 static char *Infile = NULL, /* input file name */
90 Outfile[MAXPATHLEN] = "", /* output file name */
91 Delimch = '%'; /* delimiting character */
92
93 static int Cflag = false; /* embedded comments */
94 static int Sflag = false; /* silent run flag */
95 static int Oflag = false; /* ordering flag */
96 static int Iflag = false; /* ignore case flag */
97 static int Rflag = false; /* randomize order flag */
98 static int Xflag = false; /* set rotated bit */
99 static uint32_t Num_pts = 0; /* number of pointers/strings */
100
101 static off_t *Seekpts;
102
103 static FILE *Sort_1, *Sort_2; /* pointers for sorting */
104
105 static STRFILE Tbl; /* statistics table */
106
107 static STR *Firstch; /* first chars of each string */
108
109 static void add_offset(FILE *, off_t);
110 static int cmp_str(const void *, const void *);
111 static int stable_collate_range_cmp(int, int);
112 static void do_order(void);
113 static void getargs(int, char **);
114 static void randomize(void);
115 static void usage(void) __dead2;
116
117 /*
118 * main:
119 * Drive the sucker. There are two main modes -- either we store
120 * the seek pointers, if the table is to be sorted or randomized,
121 * or we write the pointer directly to the file, if we are to stay
122 * in file order. If the former, we allocate and re-allocate in
123 * CHUNKSIZE blocks; if the latter, we just write each pointer,
124 * and then seek back to the beginning to write in the table.
125 */
126 int
main(int ac,char * av[])127 main(int ac, char *av[])
128 {
129 char *sp, *nsp, dc;
130 FILE *inf, *outf;
131 off_t last_off, pos, *p;
132 size_t length;
133 int first;
134 uint32_t cnt;
135 STR *fp;
136 static char string[257];
137
138 setlocale(LC_ALL, "");
139
140 getargs(ac, av); /* evalute arguments */
141 dc = Delimch;
142 if ((inf = fopen(Infile, "r")) == NULL) {
143 perror(Infile);
144 exit(1);
145 }
146
147 if ((outf = fopen(Outfile, "w")) == NULL) {
148 perror(Outfile);
149 exit(1);
150 }
151 if (!STORING_PTRS)
152 fseek(outf, (long)sizeof(Tbl), SEEK_SET);
153
154 /*
155 * Write the strings onto the file
156 */
157
158 Tbl.str_longlen = 0;
159 Tbl.str_shortlen = 0xffffffff;
160 Tbl.str_delim = dc;
161 Tbl.str_version = VERSION;
162 first = Oflag;
163 add_offset(outf, ftello(inf));
164 last_off = 0;
165 do {
166 sp = fgets(string, 256, inf);
167 if (sp == NULL || (sp[0] == dc && sp[1] == '\n')) {
168 pos = ftello(inf);
169 length = (size_t)(pos - last_off) -
170 (sp != NULL ? strlen(sp) : 0);
171 last_off = pos;
172 if (length == 0)
173 continue;
174 add_offset(outf, pos);
175 if ((size_t)Tbl.str_longlen < length)
176 Tbl.str_longlen = length;
177 if ((size_t)Tbl.str_shortlen > length)
178 Tbl.str_shortlen = length;
179 first = Oflag;
180 }
181 else if (first) {
182 for (nsp = sp; !isalnum((unsigned char)*nsp); nsp++)
183 continue;
184 ALLOC(Firstch, Num_pts);
185 fp = &Firstch[Num_pts - 1];
186 if (Iflag && isupper((unsigned char)*nsp))
187 fp->first = tolower((unsigned char)*nsp);
188 else
189 fp->first = *nsp;
190 fp->pos = Seekpts[Num_pts - 1];
191 first = false;
192 }
193 } while (sp != NULL);
194
195 /*
196 * write the tables in
197 */
198
199 fclose(inf);
200 Tbl.str_numstr = Num_pts - 1;
201
202 if (Cflag)
203 Tbl.str_flags |= STR_COMMENTS;
204
205 if (Oflag)
206 do_order();
207 else if (Rflag)
208 randomize();
209
210 if (Xflag)
211 Tbl.str_flags |= STR_ROTATED;
212
213 if (!Sflag) {
214 printf("\"%s\" created\n", Outfile);
215 if (Num_pts == 2)
216 puts("There was 1 string");
217 else
218 printf("There were %u strings\n", Num_pts - 1);
219 printf("Longest string: %u byte%s\n", Tbl.str_longlen,
220 Tbl.str_longlen == 1 ? "" : "s");
221 printf("Shortest string: %u byte%s\n", Tbl.str_shortlen,
222 Tbl.str_shortlen == 1 ? "" : "s");
223 }
224
225 rewind(outf);
226 Tbl.str_version = htobe32(Tbl.str_version);
227 Tbl.str_numstr = htobe32(Tbl.str_numstr);
228 Tbl.str_longlen = htobe32(Tbl.str_longlen);
229 Tbl.str_shortlen = htobe32(Tbl.str_shortlen);
230 Tbl.str_flags = htobe32(Tbl.str_flags);
231 fwrite((char *)&Tbl, sizeof(Tbl), 1, outf);
232 if (STORING_PTRS) {
233 for (p = Seekpts, cnt = Num_pts; cnt--; ++p)
234 *p = htobe64(*p);
235 fwrite(Seekpts, sizeof(*Seekpts), (size_t)Num_pts, outf);
236 }
237 fclose(outf);
238 exit(0);
239 }
240
241 /*
242 * This routine evaluates arguments from the command line
243 */
244 void
getargs(int argc,char ** argv)245 getargs(int argc, char **argv)
246 {
247 int ch;
248
249 while ((ch = getopt(argc, argv, "Cc:iorsx")) != -1)
250 switch(ch) {
251 case 'C': /* embedded comments */
252 Cflag++;
253 break;
254 case 'c': /* new delimiting char */
255 Delimch = *optarg;
256 if (!isascii(Delimch)) {
257 printf("bad delimiting character: '\\%o\n'",
258 (unsigned char)Delimch);
259 }
260 break;
261 case 'i': /* ignore case in ordering */
262 Iflag++;
263 break;
264 case 'o': /* order strings */
265 Oflag++;
266 break;
267 case 'r': /* randomize pointers */
268 Rflag++;
269 break;
270 case 's': /* silent */
271 Sflag++;
272 break;
273 case 'x': /* set the rotated bit */
274 Xflag++;
275 break;
276 case '?':
277 default:
278 usage();
279 }
280 argv += optind;
281
282 if (*argv) {
283 Infile = *argv;
284 if (*++argv) {
285 if (strlcpy(Outfile, *argv, sizeof(Outfile)) >=
286 sizeof(Outfile)) {
287 fprintf(stderr,
288 "output_file path is too long\n");
289 exit(1);
290 }
291 }
292 }
293 if (!Infile) {
294 puts("No input file name");
295 usage();
296 }
297 if (*Outfile == '\0') {
298 if ((size_t)snprintf(Outfile, sizeof(Outfile), "%s.dat",
299 Infile) >= sizeof(Outfile)) {
300 fprintf(stderr,
301 "generated output_file path is too long\n");
302 exit(1);
303 }
304 }
305 }
306
307 void
usage(void)308 usage(void)
309 {
310 fprintf(stderr,
311 "strfile [-Ciorsx] [-c char] source_file [output_file]\n");
312 exit(1);
313 }
314
315 /*
316 * add_offset:
317 * Add an offset to the list, or write it out, as appropriate.
318 */
319 void
add_offset(FILE * fp,off_t off)320 add_offset(FILE *fp, off_t off)
321 {
322 off_t beoff;
323
324 if (!STORING_PTRS) {
325 beoff = htobe64(off);
326 fwrite(&beoff, 1, sizeof(beoff), fp);
327 } else {
328 ALLOC(Seekpts, Num_pts + 1);
329 Seekpts[Num_pts] = off;
330 }
331 Num_pts++;
332 }
333
334 /*
335 * do_order:
336 * Order the strings alphabetically (possibly ignoring case).
337 */
338 void
do_order(void)339 do_order(void)
340 {
341 uint32_t i;
342 off_t *lp;
343 STR *fp;
344
345 Sort_1 = fopen(Infile, "r");
346 Sort_2 = fopen(Infile, "r");
347 qsort(Firstch, (size_t)Tbl.str_numstr, sizeof(*Firstch), cmp_str);
348 i = Tbl.str_numstr;
349 lp = Seekpts;
350 fp = Firstch;
351 while (i--)
352 *lp++ = fp++->pos;
353 fclose(Sort_1);
354 fclose(Sort_2);
355 Tbl.str_flags |= STR_ORDERED;
356 }
357
358 static int
stable_collate_range_cmp(int c1,int c2)359 stable_collate_range_cmp(int c1, int c2)
360 {
361 static char s1[2], s2[2];
362 int ret;
363
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 */
375 int
cmp_str(const void * s1,const void * s2)376 cmp_str(const void *s1, const void *s2)
377 {
378 const STR *p1, *p2;
379 int c1, c2, n1, n2, r;
380
381 #define SET_N(nf,ch) (nf = (ch == '\n'))
382 #define IS_END(ch,nf) (ch == EOF || (ch == (unsigned char)Delimch && nf))
383
384 p1 = (const STR *)s1;
385 p2 = (const STR *)s2;
386
387 c1 = (unsigned char)p1->first;
388 c2 = (unsigned char)p2->first;
389 if ((r = stable_collate_range_cmp(c1, c2)) != 0)
390 return (r);
391
392 fseeko(Sort_1, p1->pos, SEEK_SET);
393 fseeko(Sort_2, p2->pos, SEEK_SET);
394
395 n1 = false;
396 n2 = false;
397 while (!isalnum(c1 = getc(Sort_1)) && c1 != '\0' && c1 != EOF)
398 SET_N(n1, c1);
399 while (!isalnum(c2 = getc(Sort_2)) && c2 != '\0' && c2 != EOF)
400 SET_N(n2, c2);
401
402 while (!IS_END(c1, n1) && !IS_END(c2, n2)) {
403 if (Iflag) {
404 if (isupper(c1))
405 c1 = tolower(c1);
406 if (isupper(c2))
407 c2 = tolower(c2);
408 }
409 if ((r = stable_collate_range_cmp(c1, c2)) != 0)
410 return (r);
411 SET_N(n1, c1);
412 SET_N(n2, c2);
413 c1 = getc(Sort_1);
414 c2 = getc(Sort_2);
415 }
416 if (IS_END(c1, n1))
417 c1 = 0;
418 if (IS_END(c2, n2))
419 c2 = 0;
420
421 return (stable_collate_range_cmp(c1, c2));
422 }
423
424 /*
425 * randomize:
426 * Randomize the order of the string table. We must be careful
427 * not to randomize across delimiter boundaries. All
428 * randomization is done within each block.
429 */
430 void
randomize(void)431 randomize(void)
432 {
433 uint32_t cnt, i;
434 off_t tmp;
435 off_t *sp;
436
437 Tbl.str_flags |= STR_RANDOM;
438 cnt = Tbl.str_numstr;
439
440 /*
441 * move things around randomly
442 */
443
444 for (sp = Seekpts; cnt > 0; cnt--, sp++) {
445 i = arc4random_uniform(cnt);
446 tmp = sp[0];
447 sp[0] = sp[i];
448 sp[i] = tmp;
449 }
450 }
451