strfile.c revision 1.31
1 /* $NetBSD: strfile.c,v 1.31 2011/08/16 19:53:03 dholland Exp $ */ 2 3 /*- 4 * Copyright (c) 1989, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Ken Arnold. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. Neither the name of the University nor the names of its contributors 19 * may be used to endorse or promote products derived from this software 20 * without specific prior written permission. 21 * 22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 32 * SUCH DAMAGE. 33 */ 34 35 #ifdef __NetBSD__ 36 #include <sys/cdefs.h> 37 #ifndef lint 38 __COPYRIGHT("@(#) Copyright (c) 1989, 1993\ 39 The Regents of the University of California. All rights reserved."); 40 #endif /* not lint */ 41 42 #ifndef lint 43 #if 0 44 static char sccsid[] = "@(#)strfile.c 8.1 (Berkeley) 5/31/93"; 45 #else 46 __RCSID("$NetBSD: strfile.c,v 1.31 2011/08/16 19:53:03 dholland Exp $"); 47 #endif 48 #endif /* not lint */ 49 #endif /* __NetBSD__ */ 50 51 /* n.b.: this file is used at build-time - i.e. during build.sh. */ 52 53 #include <sys/types.h> 54 #include <sys/param.h> 55 #include <ctype.h> 56 #include <stdio.h> 57 #include <stdlib.h> 58 #include <string.h> 59 #include <time.h> 60 #include <unistd.h> 61 #include <inttypes.h> 62 #include <err.h> 63 64 #include "strfile.h" 65 66 #ifndef MAXPATHLEN 67 #define MAXPATHLEN 1024 68 #endif /* MAXPATHLEN */ 69 70 static uint32_t h2nl(uint32_t h); 71 static void getargs(int argc, char **argv); 72 static void usage(void) __dead; 73 static void add_offset(FILE *fp, off_t off); 74 static void do_order(void); 75 static int cmp_str(const void *vp1, const void *vp2); 76 static void randomize(void); 77 static void fwrite_be_offt(off_t off, FILE *f); 78 79 static uint32_t 80 h2nl(uint32_t h) 81 { 82 unsigned char c[4]; 83 uint32_t rv; 84 85 c[0] = (h >> 24) & 0xff; 86 c[1] = (h >> 16) & 0xff; 87 c[2] = (h >> 8) & 0xff; 88 c[3] = (h >> 0) & 0xff; 89 memcpy(&rv, c, sizeof rv); 90 91 return (rv); 92 } 93 94 /* 95 * This program takes a file composed of strings separated by 96 * lines starting with two consecutive delimiting character (default 97 * character is '%') and creates another file which consists of a table 98 * describing the file (structure from "strfile.h"), a table of seek 99 * pointers to the start of the strings, and the strings, each terminated 100 * by a null byte. Usage: 101 * 102 * % strfile [-iorsx] [ -cC ] sourcefile [ datafile ] 103 * 104 * c - Change delimiting character from '%' to 'C' 105 * s - Silent. Give no summary of data processed at the end of 106 * the run. 107 * o - order the strings in alphabetic order 108 * i - if ordering, ignore case 109 * r - randomize the order of the strings 110 * x - set rotated bit 111 * 112 * Ken Arnold Sept. 7, 1978 -- 113 * 114 * Added ordering options. 115 */ 116 117 # define TRUE 1 118 # define FALSE 0 119 120 # define STORING_PTRS (Oflag || Rflag) 121 # define CHUNKSIZE 512 122 123 # define ALLOC(ptr,sz) do { \ 124 if (ptr == NULL) \ 125 ptr = malloc(CHUNKSIZE * sizeof *ptr); \ 126 else if (((sz) + 1) % CHUNKSIZE == 0) \ 127 ptr = realloc(ptr, ((sz) + CHUNKSIZE) * sizeof *ptr); \ 128 if (ptr == NULL) \ 129 err(1, "out of space"); \ 130 } while (0) 131 132 typedef struct { 133 char first; 134 off_t pos; 135 } STR; 136 137 static char *Infile = NULL; /* input file name */ 138 static char Outfile[MAXPATHLEN] = ""; /* output file name */ 139 static char Delimch = '%'; /* delimiting character */ 140 141 static int Sflag = FALSE; /* silent run flag */ 142 static int Oflag = FALSE; /* ordering flag */ 143 static int Iflag = FALSE; /* ignore case flag */ 144 static int Rflag = FALSE; /* randomize order flag */ 145 static int Xflag = FALSE; /* set rotated bit */ 146 static long Num_pts = 0; /* number of pointers/strings */ 147 148 static off_t *Seekpts; 149 150 static FILE *Sort_1, *Sort_2; /* pointers for sorting */ 151 152 static STRFILE Tbl; /* statistics table */ 153 154 static STR *Firstch; /* first chars of each string */ 155 156 #ifdef __GNUC__ 157 #define NORETURN __dead 158 #else 159 #define NORETURN 160 #endif 161 162 #ifndef __dead /* not NetBSD, presumably */ 163 #define __dead ; 164 #endif 165 166 void add_offset(FILE *, off_t); 167 int cmp_str(const void *, const void *); 168 void do_order(void); 169 void fwrite_be_offt(off_t, FILE *); 170 void getargs(int, char *[]); 171 int main(int, char *[]); 172 void randomize(void); 173 void usage(void) NORETURN; 174 175 176 /* 177 * main: 178 * Drive the sucker. There are two main modes -- either we store 179 * the seek pointers, if the table is to be sorted or randomized, 180 * or we write the pointer directly to the file, if we are to stay 181 * in file order. If the former, we allocate and re-allocate in 182 * CHUNKSIZE blocks; if the latter, we just write each pointer, 183 * and then seek back to the beginning to write in the table. 184 */ 185 int 186 main(int ac, char **av) 187 { 188 char *sp, dc; 189 FILE *inf, *outf; 190 off_t last_off, length, pos, *p; 191 int first, cnt; 192 char *nsp; 193 STR *fp; 194 static char string[257]; 195 196 /* sanity test */ 197 if (sizeof(uint32_t) != 4) 198 errx(1, "sizeof(uint32_t) != 4"); 199 200 getargs(ac, av); /* evalute arguments */ 201 dc = Delimch; 202 if ((inf = fopen(Infile, "r")) == NULL) 203 err(1, "open `%s'", Infile); 204 205 if ((outf = fopen(Outfile, "w")) == NULL) 206 err(1, "open `%s'", Outfile); 207 if (!STORING_PTRS) 208 (void) fseek(outf, sizeof Tbl, SEEK_SET); 209 210 /* 211 * Write the strings onto the file 212 */ 213 214 Tbl.str_longlen = 0; 215 Tbl.str_shortlen = (unsigned int) 0x7fffffff; 216 Tbl.str_delim = dc; 217 Tbl.str_version = VERSION; 218 first = Oflag; 219 add_offset(outf, ftell(inf)); 220 last_off = 0; 221 do { 222 sp = fgets(string, 256, inf); 223 if (sp == NULL || (sp[0] == dc && sp[1] == '\n')) { 224 pos = ftell(inf); 225 length = pos - last_off - (sp ? strlen(sp) : 0); 226 last_off = pos; 227 if (!length) 228 continue; 229 add_offset(outf, pos); 230 if ((off_t)Tbl.str_longlen < length) 231 Tbl.str_longlen = length; 232 if ((off_t)Tbl.str_shortlen > length) 233 Tbl.str_shortlen = length; 234 first = Oflag; 235 } 236 else if (first) { 237 for (nsp = sp; !isalnum((unsigned char)*nsp); nsp++) 238 continue; 239 ALLOC(Firstch, Num_pts); 240 fp = &Firstch[Num_pts - 1]; 241 if (Iflag && isupper((unsigned char)*nsp)) 242 fp->first = tolower((unsigned char)*nsp); 243 else 244 fp->first = *nsp; 245 fp->pos = Seekpts[Num_pts - 1]; 246 first = FALSE; 247 } 248 } while (sp != NULL); 249 250 /* 251 * write the tables in 252 */ 253 254 (void) fclose(inf); 255 256 if (Oflag) 257 do_order(); 258 else if (Rflag) 259 randomize(); 260 261 if (Xflag) 262 Tbl.str_flags |= STR_ROTATED; 263 264 if (!Sflag) { 265 printf("\"%s\" created\n", Outfile); 266 if (Num_pts == 2) 267 puts("There was 1 string"); 268 else 269 printf("There were %d strings\n", (int)(Num_pts - 1)); 270 printf("Longest string: %lu byte%s\n", (unsigned long)Tbl.str_longlen, 271 Tbl.str_longlen == 1 ? "" : "s"); 272 printf("Shortest string: %lu byte%s\n", (unsigned long)Tbl.str_shortlen, 273 Tbl.str_shortlen == 1 ? "" : "s"); 274 } 275 276 (void) fseek(outf, (off_t) 0, SEEK_SET); 277 Tbl.str_version = h2nl(Tbl.str_version); 278 Tbl.str_numstr = h2nl(Num_pts - 1); 279 Tbl.str_longlen = h2nl(Tbl.str_longlen); 280 Tbl.str_shortlen = h2nl(Tbl.str_shortlen); 281 Tbl.str_flags = h2nl(Tbl.str_flags); 282 (void) fwrite((char *) &Tbl, sizeof Tbl, 1, outf); 283 if (STORING_PTRS) { 284 for (p = Seekpts, cnt = Num_pts; cnt--; ++p) 285 fwrite_be_offt(*p, outf); 286 } 287 fflush(outf); 288 if (ferror(outf)) 289 err(1, "fwrite %s", Outfile); 290 (void) fclose(outf); 291 exit(0); 292 } 293 294 /* 295 * This routine evaluates arguments from the command line 296 */ 297 static void 298 getargs(int argc, char **argv) 299 { 300 int ch; 301 extern int optind; 302 extern char *optarg; 303 304 while ((ch = getopt(argc, argv, "c:iorsx")) != -1) 305 switch(ch) { 306 case 'c': /* new delimiting char */ 307 Delimch = *optarg; 308 if (!isascii(Delimch)) { 309 printf("bad delimiting character: '\\%o\n'", 310 Delimch); 311 } 312 break; 313 case 'i': /* ignore case in ordering */ 314 Iflag++; 315 break; 316 case 'o': /* order strings */ 317 Oflag++; 318 break; 319 case 'r': /* randomize pointers */ 320 Rflag++; 321 break; 322 case 's': /* silent */ 323 Sflag++; 324 break; 325 case 'x': /* set the rotated bit */ 326 Xflag++; 327 break; 328 case '?': 329 default: 330 usage(); 331 } 332 argv += optind; 333 334 if (*argv) { 335 Infile = *argv; 336 if (*++argv) 337 (void) strcpy(Outfile, *argv); 338 } 339 if (!Infile) { 340 puts("No input file name"); 341 usage(); 342 } 343 if (*Outfile == '\0') { 344 (void) strcpy(Outfile, Infile); 345 (void) strcat(Outfile, ".dat"); 346 } 347 } 348 349 static void 350 usage(void) 351 { 352 (void) fprintf(stderr, 353 "Usage: %s [-iorsx] [-c char] sourcefile [datafile]\n", 354 getprogname()); 355 exit(1); 356 } 357 358 /* 359 * add_offset: 360 * Add an offset to the list, or write it out, as appropriate. 361 */ 362 static void 363 add_offset(FILE *fp, off_t off) 364 { 365 366 if (!STORING_PTRS) { 367 fwrite_be_offt(off, fp); 368 } else { 369 ALLOC(Seekpts, Num_pts + 1); 370 Seekpts[Num_pts] = off; 371 } 372 Num_pts++; 373 } 374 375 /* 376 * do_order: 377 * Order the strings alphabetically (possibly ignoring case). 378 */ 379 static void 380 do_order(void) 381 { 382 int i; 383 off_t *lp; 384 STR *fp; 385 386 Sort_1 = fopen(Infile, "r"); 387 Sort_2 = fopen(Infile, "r"); 388 qsort((char *) Firstch, (int) Tbl.str_numstr, sizeof *Firstch, cmp_str); 389 i = Tbl.str_numstr; 390 lp = Seekpts; 391 fp = Firstch; 392 while (i--) 393 *lp++ = fp++->pos; 394 (void) fclose(Sort_1); 395 (void) fclose(Sort_2); 396 Tbl.str_flags |= STR_ORDERED; 397 } 398 399 static int 400 cmp_str(const void *vp1, const void *vp2) 401 { 402 const STR *p1, *p2; 403 int c1, c2; 404 int n1, n2; 405 406 p1 = (const STR *)vp1; 407 p2 = (const STR *)vp2; 408 409 # define SET_N(nf,ch) (nf = (ch == '\n')) 410 # define IS_END(ch,nf) (ch == Delimch && nf) 411 412 c1 = p1->first; 413 c2 = p2->first; 414 if (c1 != c2) 415 return c1 - c2; 416 417 (void) fseek(Sort_1, p1->pos, SEEK_SET); 418 (void) fseek(Sort_2, p2->pos, SEEK_SET); 419 420 n1 = FALSE; 421 n2 = FALSE; 422 while (!isalnum(c1 = getc(Sort_1)) && c1 != '\0') 423 SET_N(n1, c1); 424 while (!isalnum(c2 = getc(Sort_2)) && c2 != '\0') 425 SET_N(n2, c2); 426 427 while (!IS_END(c1, n1) && !IS_END(c2, n2)) { 428 if (Iflag) { 429 if (isupper(c1)) 430 c1 = tolower(c1); 431 if (isupper(c2)) 432 c2 = tolower(c2); 433 } 434 if (c1 != c2) 435 return c1 - c2; 436 SET_N(n1, c1); 437 SET_N(n2, c2); 438 c1 = getc(Sort_1); 439 c2 = getc(Sort_2); 440 } 441 if (IS_END(c1, n1)) 442 c1 = 0; 443 if (IS_END(c2, n2)) 444 c2 = 0; 445 return c1 - c2; 446 } 447 448 /* 449 * randomize: 450 * Randomize the order of the string table. We must be careful 451 * not to randomize across delimiter boundaries. All 452 * randomization is done within each block. 453 */ 454 static void 455 randomize(void) 456 { 457 int cnt, i; 458 off_t tmp; 459 off_t *sp; 460 461 srandom((int)(time((time_t *) NULL) + getpid())); 462 463 Tbl.str_flags |= STR_RANDOM; 464 cnt = Tbl.str_numstr; 465 466 /* 467 * move things around randomly 468 */ 469 470 for (sp = Seekpts; cnt > 0; cnt--, sp++) { 471 i = random() % cnt; 472 tmp = sp[0]; 473 sp[0] = sp[i]; 474 sp[i] = tmp; 475 } 476 } 477 478 /* 479 * fwrite_be_offt: 480 * Write out the off paramater as a 64 bit big endian number 481 */ 482 483 static void 484 fwrite_be_offt(off_t off, FILE *f) 485 { 486 int i; 487 unsigned char c[8]; 488 489 for (i = 7; i >= 0; i--) { 490 c[i] = off & 0xff; 491 off >>= 8; 492 } 493 fwrite(c, sizeof(c), 1, f); 494 } 495
Indexes created Wed Oct 15 07:09:58 GMT 2025