1/* 2 * Compress - data compression program 3 */ 4 5/* 6 * machine variants which require cc -Dmachine: pdp11, z8000, pcxt 7 */ 8 9/* 10 * Set USERMEM to the maximum amount of physical user memory available 11 * in bytes. USERMEM is used to determine the maximum BITS that can be used 12 * for compression. 13 * 14 * SACREDMEM is the amount of physical memory saved for others; compress 15 * will hog the rest. 16 */ 17#ifndef SACREDMEM 18#define SACREDMEM 0 19#endif 20 21#ifndef USERMEM 22# define USERMEM 450000 /* default user memory */ 23#endif 24 25#ifdef interdata /* (Perkin-Elmer) */ 26#define SIGNED_COMPARE_SLOW /* signed compare is slower than unsigned */ 27#endif 28 29#ifdef pdp11 30# define BITS 12 /* max bits/code for 16-bit machine */ 31# define NO_UCHAR /* also if "unsigned char" functions as signed char */ 32# undef USERMEM 33#endif /* pdp11 */ /* don't forget to compile with -i */ 34 35#ifdef z8000 36# define BITS 12 37# undef vax /* weird preprocessor */ 38# undef USERMEM 39#endif /* z8000 */ 40 41#ifdef MSDOS /* Microsoft C 3.0 for MS-DOS */ 42# undef USERMEM 43# ifdef BIG /* then this is a large data compilation */ 44# undef DEBUG /* DEBUG makes the executible too big */ 45# define BITS 16 46# define XENIX_16 47# else /* this is a small model compilation */ 48# define BITS 12 49# endif 50#else 51#undef BIG 52#endif /* MSDOS */ 53 54#ifdef pcxt 55# define BITS 12 56# undef USERMEM 57#endif /* pcxt */ 58 59#ifdef USERMEM 60# if USERMEM >= (433484+SACREDMEM) 61# define PBITS 16 62# else 63# if USERMEM >= (229600+SACREDMEM) 64# define PBITS 15 65# else 66# if USERMEM >= (127536+SACREDMEM) 67# define PBITS 14 68# else 69# if USERMEM >= (73464+SACREDMEM) 70# define PBITS 13 71# else 72# define PBITS 12 73# endif 74# endif 75# endif 76# endif 77# undef USERMEM 78#endif /* USERMEM */ 79 80#ifdef PBITS /* Preferred BITS for this memory size */ 81# ifndef BITS 82# define BITS PBITS 83# endif BITS 84#endif /* PBITS */ 85 86#if BITS == 16 87# define HSIZE 69001 /* 95% occupancy */ 88#endif 89#if BITS == 15 90# define HSIZE 35023 /* 94% occupancy */ 91#endif 92#if BITS == 14 93# define HSIZE 18013 /* 91% occupancy */ 94#endif 95#if BITS == 13 96# define HSIZE 9001 /* 91% occupancy */ 97#endif 98#if BITS <= 12 99# define HSIZE 5003 /* 80% occupancy */ 100#endif 101 102#ifdef M_XENIX /* Stupid compiler can't handle arrays with */ 103# if BITS == 16 /* more than 65535 bytes - so we fake it */ 104# define XENIX_16 105# else 106# if BITS > 13 /* Code only handles BITS = 12, 13, or 16 */ 107# define BITS 13 108# endif 109# endif 110#endif 111 112/* 113 * a code_int must be able to hold 2**BITS values of type int, and also -1 114 */ 115#if BITS > 15 116typedef long int code_int; 117#else 118typedef int code_int; 119#endif 120 121#ifdef SIGNED_COMPARE_SLOW 122typedef unsigned long int count_int; 123typedef unsigned short int count_short; 124#else 125typedef long int count_int; 126#endif 127 128#ifdef NO_UCHAR 129 typedef char char_type; 130#else 131 typedef unsigned char char_type; 132#endif /* UCHAR */ 133char_type magic_header[] = { "\037\235" }; /* 1F 9D */ 134 135/* Defines for third byte of header */ 136#define BIT_MASK 0x1f 137#define BLOCK_MASK 0x80 138/* Masks 0x40 and 0x20 are free. I think 0x20 should mean that there is 139 a fourth header byte (for expansion). 140*/ 141#define INIT_BITS 9 /* initial number of bits/code */ 142 143#define min(a, b) ((a) > (b) ? (b) : (a)) 144 145/* 146 * compress.c - File compression ala IEEE Computer, June 1984. 147 * 148 * Authors: Spencer W. Thomas (decvax!harpo!utah-cs!utah-gr!thomas) 149 * Jim McKie (decvax!mcvax!jim) 150 * Steve Davies (decvax!vax135!petsd!peora!srd) 151 * Ken Turkowski (decvax!decwrl!turtlevax!ken) 152 * James A. Woods (decvax!ihnp4!ames!jaw) 153 * Joe Orost (decvax!vax135!petsd!joe) 154 * 155 * $Header: /tmp/bonefish/open-beos/current/src/apps/bin/compress/compress.c,v 1.1 2004/06/07 21:23:09 korli Exp $ 156 * $Log: compress.c,v $ 157 * Revision 1.1 2004/06/07 21:23:09 korli 158 * Added compress-4.0 159 * 160 * Revision 1.12 1996/02/24 02:45:03 cyril 161 * got rid of Unix.h 162 * 163 * Revision 1.11 1996/02/18 23:50:39 cyril 164 * final clean up for new resource policy. 165 * 166 * Revision 1.10 1996/01/25 09:53:14 dbg 167 * deal with the now correct posix headers 168 * 169 * Revision 1.9 1996/01/23 19:17:39 btaylor 170 * prep for CW8 171 * 172 * Revision 1.8 1996/01/11 18:28:19 dbg 173 * deal with the new posix environment (i.e. we can now 174 * include <sys/stat.h> and <sys/types.h> because they 175 * exist and we don't need as many workarounds for missing 176 * things in the posix environment). 177 * 178 * Revision 1.7 1995/12/21 18:58:15 erich 179 * Add macro for isascii, which doesn't really exist. 180 * 181 * Revision 1.1 1995/12/13 22:51:52 ming 182 * DR6 FROZEN ON 12/13/95 14:00:00 183 * 184 * Revision 1.6 1995/11/08 22:49:32 robert 185 * FILE_NAME_LENGTH -> B_FILE_NAME_LENGTH 186 * 187 * Revision 1.5 1995/08/30 02:42:55 herold 188 * ### API ### remove all traces of int type from OS.h. Change 189 * FILENAME_LENGTH to FILE_NAME_LENGTH. Lotsa buck for not much 190 * bang, but there it is... 191 * 192 * Revision 1.4 1995/06/05 22:34:07 erich 193 * Attempt to make stuff in ./gnu compile...sort of. 194 * 195 * Revision 1.3 1995/02/22 19:38:43 peter 196 * Update 'tar' and 'compress' tools to work with new resource/data files. 197 * 'tar' as some glitches (when using the 'u' flag), but we'll worry about 198 * that later. 199 * 200 * Revision 1.2 1994/10/04 17:38:00 herold 201 * don't depend on /usr/include being around - use Unix.h instead 202 * 203 * Revision 1.1 1993/10/14 22:46:10 moofie 204 * Initial revision 205 * 206 * Revision 4.0.1 87/11/27 21:45:00 rms (Richard Stallman) 207 * Once copystat has run, don't delete output file on interrupt. 208 * Mention more flags in the Usage string. 209 * 210 * Revision 4.0 85/07/30 12:50:00 joe 211 * Removed ferror() calls in output routine on every output except first. 212 * Prepared for release to the world. 213 * 214 * Revision 3.6 85/07/04 01:22:21 joe 215 * Remove much wasted storage by overlaying hash table with the tables 216 * used by decompress: tab_suffix[1<<BITS], stack[8000]. Updated USERMEM 217 * computations. Fixed dump_tab() DEBUG routine. 218 * 219 * Revision 3.5 85/06/30 20:47:21 jaw 220 * Change hash function to use exclusive-or. Rip out hash cache. These 221 * speedups render the megamemory version defunct, for now. Make decoder 222 * stack global. Parts of the RCS trunks 2.7, 2.6, and 2.1 no longer apply. 223 * 224 * Revision 3.4 85/06/27 12:00:00 ken 225 * Get rid of all floating-point calculations by doing all compression ratio 226 * calculations in fixed point. 227 * 228 * Revision 3.3 85/06/24 21:53:24 joe 229 * Incorporate portability suggestion for M_XENIX. Got rid of text on #else 230 * and #endif lines. Cleaned up #ifdefs for vax and interdata. 231 * 232 * Revision 3.2 85/06/06 21:53:24 jaw 233 * Incorporate portability suggestions for Z8000, IBM PC/XT from mailing list. 234 * Default to "quiet" output (no compression statistics). 235 * 236 * Revision 3.1 85/05/12 18:56:13 jaw 237 * Integrate decompress() stack speedups (from early pointer mods by McKie). 238 * Repair multi-file USERMEM gaffe. Unify 'force' flags to mimic semantics 239 * of SVR2 'pack'. Streamline block-compress table clear logic. Increase 240 * output byte count by magic number size. 241 * 242 * Revision 3.0 84/11/27 11:50:00 petsd!joe 243 * Set HSIZE depending on BITS. Set BITS depending on USERMEM. Unrolled 244 * loops in clear routines. Added "-C" flag for 2.0 compatibility. Used 245 * unsigned compares on Perkin-Elmer. Fixed foreground check. 246 * 247 * Revision 2.7 84/11/16 19:35:39 ames!jaw 248 * Cache common hash codes based on input statistics; this improves 249 * performance for low-density raster images. Pass on #ifdef bundle 250 * from Turkowski. 251 * 252 * Revision 2.6 84/11/05 19:18:21 ames!jaw 253 * Vary size of hash tables to reduce time for small files. 254 * Tune PDP-11 hash function. 255 * 256 * Revision 2.5 84/10/30 20:15:14 ames!jaw 257 * Junk chaining; replace with the simpler (and, on the VAX, faster) 258 * double hashing, discussed within. Make block compression standard. 259 * 260 * Revision 2.4 84/10/16 11:11:11 ames!jaw 261 * Introduce adaptive reset for block compression, to boost the rate 262 * another several percent. (See mailing list notes.) 263 * 264 * Revision 2.3 84/09/22 22:00:00 petsd!joe 265 * Implemented "-B" block compress. Implemented REVERSE sorting of tab_next. 266 * Bug fix for last bits. Changed fwrite to putchar loop everywhere. 267 * 268 * Revision 2.2 84/09/18 14:12:21 ames!jaw 269 * Fold in news changes, small machine typedef from thomas, 270 * #ifdef interdata from joe. 271 * 272 * Revision 2.1 84/09/10 12:34:56 ames!jaw 273 * Configured fast table lookup for 32-bit machines. 274 * This cuts user time in half for b <= FBITS, and is useful for news batching 275 * from VAX to PDP sites. Also sped up decompress() [fwrite->putc] and 276 * added signal catcher [plus beef in writeerr()] to delete effluvia. 277 * 278 * Revision 2.0 84/08/28 22:00:00 petsd!joe 279 * Add check for foreground before prompting user. Insert maxbits into 280 * compressed file. Force file being uncompressed to end with ".Z". 281 * Added "-c" flag and "zcat". Prepared for release. 282 * 283 * Revision 1.10 84/08/24 18:28:00 turtlevax!ken 284 * Will only compress regular files (no directories), added a magic number 285 * header (plus an undocumented -n flag to handle old files without headers), 286 * added -f flag to force overwriting of possibly existing destination file, 287 * otherwise the user is prompted for a response. Will tack on a .Z to a 288 * filename if it doesn't have one when decompressing. Will only replace 289 * file if it was compressed. 290 * 291 * Revision 1.9 84/08/16 17:28:00 turtlevax!ken 292 * Removed scanargs(), getopt(), added .Z extension and unlimited number of 293 * filenames to compress. Flags may be clustered (-Ddvb12) or separated 294 * (-D -d -v -b 12), or combination thereof. Modes and other status is 295 * copied with copystat(). -O bug for 4.2 seems to have disappeared with 296 * 1.8. 297 * 298 * Revision 1.8 84/08/09 23:15:00 joe 299 * Made it compatible with vax version, installed jim's fixes/enhancements 300 * 301 * Revision 1.6 84/08/01 22:08:00 joe 302 * Sped up algorithm significantly by sorting the compress chain. 303 * 304 * Revision 1.5 84/07/13 13:11:00 srd 305 * Added C version of vax asm routines. Changed structure to arrays to 306 * save much memory. Do unsigned compares where possible (faster on 307 * Perkin-Elmer) 308 * 309 * Revision 1.4 84/07/05 03:11:11 thomas 310 * Clean up the code a little and lint it. (Lint complains about all 311 * the regs used in the asm, but I'm not going to "fix" this.) 312 * 313 * Revision 1.3 84/07/05 02:06:54 thomas 314 * Minor fixes. 315 * 316 * Revision 1.2 84/07/05 00:27:27 thomas 317 * Add variable bit length output. 318 * 319 */ 320static char rcs_ident[] = "$Header: /tmp/bonefish/open-beos/current/src/apps/bin/compress/compress.c,v 1.1 2004/06/07 21:23:09 korli Exp $"; 321 322#include <stdio.h> 323#include <ctype.h> 324#include <signal.h> 325#include <sys/types.h> 326#include <sys/stat.h> 327#include <utime.h> 328 329#ifdef MSDOS 330#include <stdlib.h> 331#endif 332 333#ifdef __HOBBIT__ 334# undef putc 335#endif 336 337#define ARGVAL() (*++(*argv) || (--argc && *++argv)) 338 339int n_bits; /* number of bits/code */ 340int maxbits = BITS; /* user settable max # bits/code */ 341code_int maxcode; /* maximum code, given n_bits */ 342code_int maxmaxcode = (code_int)1 << BITS; /* should NEVER generate this code */ 343#ifdef COMPATIBLE /* But wrong! */ 344# define MAXCODE(n_bits) ((code_int) 1 << (n_bits) - 1) 345#else 346# define MAXCODE(n_bits) (((code_int) 1 << (n_bits)) - 1) 347#endif /* COMPATIBLE */ 348 349#ifdef XENIX_16 350# ifdef MSDOS 351 352count_int far htab0[8192]; 353count_int far htab1[8192]; 354count_int far htab2[8192]; 355count_int far htab3[8192]; 356count_int far htab4[8192]; 357count_int far htab5[8192]; 358count_int far htab6[8192]; 359count_int far htab7[8192]; 360count_int far htab8[HSIZE-65536]; 361count_int far * htab[9] = { 362 htab0, htab1, htab2, htab3, htab4, htab5, htab6, htab7, htab8 }; 363 364unsigned short far code0tab[16384]; 365unsigned short far code1tab[16384]; 366unsigned short far code2tab[16384]; 367unsigned short far code3tab[16384]; 368unsigned short far code4tab[16384]; 369unsigned short far * codetab[5] = { 370 code0tab, code1tab, code2tab, code3tab, code4tab }; 371 372# else 373 374count_int htab0[8192]; 375count_int htab1[8192]; 376count_int htab2[8192]; 377count_int htab3[8192]; 378count_int htab4[8192]; 379count_int htab5[8192]; 380count_int htab6[8192]; 381count_int htab7[8192]; 382count_int htab8[HSIZE-65536]; 383count_int * htab[9] = { 384 htab0, htab1, htab2, htab3, htab4, htab5, htab6, htab7, htab8 }; 385 386unsigned short code0tab[16384]; 387unsigned short code1tab[16384]; 388unsigned short code2tab[16384]; 389unsigned short code3tab[16384]; 390unsigned short code4tab[16384]; 391unsigned short * codetab[5] = { 392 code0tab, code1tab, code2tab, code3tab, code4tab }; 393 394# endif /* MSDOS */ 395 396#define htabof(i) (htab[(i) >> 13][(i) & 0x1fff]) 397#define codetabof(i) (codetab[(i) >> 14][(i) & 0x3fff]) 398 399#else /* Normal machine */ 400count_int htab [HSIZE]; 401unsigned short codetab [HSIZE]; 402#define htabof(i) htab[i] 403#define codetabof(i) codetab[i] 404 405#endif /* XENIX_16 */ 406code_int hsize = HSIZE; /* for dynamic table sizing */ 407count_int fsize; 408 409/* 410 * To save much memory, we overlay the table used by compress() with those 411 * used by decompress(). The tab_prefix table is the same size and type 412 * as the codetab. The tab_suffix table needs 2**BITS characters. We 413 * get this from the beginning of htab. The output stack uses the rest 414 * of htab, and contains characters. There is plenty of room for any 415 * possible stack (stack used to be 8000 characters). 416 */ 417 418#define tab_prefixof(i) codetabof(i) 419 420#ifdef XENIX_16 421# ifdef MSDOS 422# define tab_suffixof(i) ((char_type far *)htab[(i)>>15])[(i) & 0x7fff] 423# define de_stack ((char_type far *)(htab2)) 424# else 425# define tab_suffixof(i) ((char_type *)htab[(i)>>15])[(i) & 0x7fff] 426# define de_stack ((char_type *)(htab2)) 427# endif /* MSDOS */ 428#else /* Normal machine */ 429# define tab_suffixof(i) ((char_type *)(htab))[i] 430# define de_stack ((char_type *)&tab_suffixof((code_int)1<<BITS)) 431#endif /* XENIX_16 */ 432 433code_int free_ent = 0; /* first unused entry */ 434int exit_stat = 0; 435 436code_int getcode(); 437 438Usage() { 439#ifdef DEBUG 440 441# ifdef MSDOS 442 fprintf(stderr,"Usage: compress [-cdDfivV] [-b maxbits] [file ...]\n"); 443# else 444 fprintf(stderr,"Usage: compress [-cdDfvV] [-b maxbits] [file ...]\n"); 445# endif /* MSDOS */ 446 447} 448int debug = 0; 449#else 450 451# ifdef MSDOS 452 fprintf(stderr,"Usage: compress [-cdfivV] [-b maxbits] [file ...]\n"); 453# else 454 fprintf(stderr,"Usage: compress [-cdfvV] [-b maxbits] [file ...]\n"); 455# endif /* MSDOS */ 456 457} 458#endif /* DEBUG */ 459int nomagic = 0; /* Use a 3-byte magic number header, unless old file */ 460int zcat_flg = 0; /* Write output on stdout, suppress messages */ 461int precious = 1; /* Don't unlink output file on interrupt */ 462int quiet = 1; /* don't tell me about compression */ 463 464/* 465 * block compression parameters -- after all codes are used up, 466 * and compression rate changes, start over. 467 */ 468int block_compress = BLOCK_MASK; 469int clear_flg = 0; 470long int ratio = 0; 471#define CHECK_GAP 10000 /* ratio check interval */ 472count_int checkpoint = CHECK_GAP; 473/* 474 * the next two codes should not be changed lightly, as they must not 475 * lie within the contiguous general code space. 476 */ 477#define FIRST 257 /* first free entry */ 478#define CLEAR 256 /* table clear output code */ 479 480int force = 0; 481char ofname [100]; 482 483#ifdef MSDOS 484# define PATH_SEP '\\' 485int image = 2; /* 1 <=> image (binary) mode; 2 <=> text mode */ 486#else 487# define PATH_SEP '/' 488#endif 489 490#ifdef DEBUG 491int verbose = 0; 492#endif /* DEBUG */ 493void (*bgnd_flag)(); 494 495int do_decomp = 0; 496 497/***************************************************************** 498 * TAG( main ) 499 * 500 * Algorithm from "A Technique for High Performance Data Compression", 501 * Terry A. Welch, IEEE Computer Vol 17, No 6 (June 1984), pp 8-19. 502 * 503 * Usage: compress [-cdfivV] [-b bits] [file ...] 504 * Inputs: 505 * 506 * -c: Write output on stdout, don't remove original. 507 * 508 * -d: If given, decompression is done instead. 509 * 510 * -f: Forces output file to be generated, even if one already 511 * exists, and even if no space is saved by compressing. 512 * If -f is not used, the user will be prompted if stdin is 513 * a tty, otherwise, the output file will not be overwritten. 514 * 515 * -i: Image mode (defined only under MS-DOS). Prevents 516 * conversion between UNIX text representation (LF line 517 * termination) in compressed form and MS-DOS text 518 * representation (CR-LF line termination) in uncompressed 519 * form. Useful with non-text files. 520 * 521 * -v: Write compression statistics 522 * 523 * -V: Write version and compilation options. 524 * 525 * -b: Parameter limits the max number of bits/code. 526 * 527 * file ...: Files to be compressed. If none specified, stdin 528 * is used. 529 * Outputs: 530 * file.Z: Compressed form of file with same mode, owner, and utimes 531 * or stdout (if stdin used as input) 532 * 533 * Assumptions: 534 * When filenames are given, replaces with the compressed version 535 * (.Z suffix) only if the file decreases in size. 536 * Algorithm: 537 * Modified Lempel-Ziv method (LZW). Basically finds common 538 * substrings and replaces them with a variable size code. This is 539 * deterministic, and can be done on the fly. Thus, the decompression 540 * procedure needs no input table, but tracks the way the table was built. 541 */ 542 543main( argc, argv ) 544register int argc; char **argv; 545{ 546 int overwrite = 0; /* Do not overwrite unless given -f flag */ 547 char tempname[100]; 548 char **filelist, **fileptr; 549 char *cp, *rindex(), *malloc(); 550 struct stat statbuf; 551 extern void onintr(); 552 553#ifdef MSDOS 554 char *sufp; 555#else 556 extern void oops(); 557#endif 558 559#ifndef MSDOS 560 if ( (bgnd_flag = signal ( SIGINT, SIG_IGN )) != SIG_IGN ) { 561#endif 562 563 signal ( SIGINT, onintr ); 564 565#ifndef MSDOS 566 signal ( SIGSEGV, oops ); 567 } 568#endif 569 570#ifdef COMPATIBLE 571 nomagic = 1; /* Original didn't have a magic number */ 572#endif /* COMPATIBLE */ 573 574 filelist = fileptr = (char **)(malloc(argc * sizeof(*argv))); 575 *filelist = NULL; 576 577 if((cp = rindex(argv[0], PATH_SEP)) != 0) { 578 cp++; 579 } else { 580 cp = argv[0]; 581 } 582 583#ifdef MSDOS 584 if(strcmp(cp, "UNCOMPRE.EXE") == 0) { 585#else 586 if(strcmp(cp, "uncompress") == 0) { 587#endif 588 589 do_decomp = 1; 590 591#ifdef MSDOS 592 } else if(strcmp(cp, "ZCAT.EXE") == 0) { 593#else 594 } else if(strcmp(cp, "zcat") == 0) { 595#endif 596 597 do_decomp = 1; 598 zcat_flg = 1; 599 } 600 601#ifdef BSD4_2 602 /* 4.2BSD dependent - take it out if not */ 603 setlinebuf( stderr ); 604#endif /* BSD4_2 */ 605 606 /* Argument Processing 607 * All flags are optional. 608 * -D => debug 609 * -V => print Version; debug verbose 610 * -d => do_decomp 611 * -v => unquiet 612 * -f => force overwrite of output file 613 * -n => no header: useful to uncompress old files 614 * -b maxbits => maxbits. If -b is specified, then maxbits MUST be 615 * given also. 616 * -c => cat all output to stdout 617 * -C => generate output compatible with compress 2.0. 618 * if a string is left, must be an input filename. 619 */ 620 for (argc--, argv++; argc > 0; argc--, argv++) { 621 if (**argv == '-') { /* A flag argument */ 622 while (*++(*argv)) { /* Process all flags in this arg */ 623 switch (**argv) { 624#ifdef DEBUG 625 case 'D': 626 debug = 1; 627 break; 628 case 'V': 629 verbose = 1; 630 version(); 631 break; 632#else 633 case 'V': 634 version(); 635 break; 636#endif /* DEBUG */ 637 638#ifdef MSDOS 639 case 'i': 640 image = 1; 641 break; 642#endif 643 644 case 'v': 645 quiet = 0; 646 break; 647 case 'd': 648 do_decomp = 1; 649 break; 650 case 'f': 651 case 'F': 652 overwrite = 1; 653 force = 1; 654 break; 655 case 'n': 656 nomagic = 1; 657 break; 658 case 'C': 659 block_compress = 0; 660 break; 661 case 'b': 662 if (!ARGVAL()) { 663 fprintf(stderr, "Missing maxbits\n"); 664 Usage(); 665 exit(1); 666 } 667 maxbits = atoi(*argv); 668 goto nextarg; 669 case 'c': 670 zcat_flg = 1; 671 break; 672 case 'q': 673 quiet = 1; 674 break; 675 default: 676 fprintf(stderr, "Unknown flag: '%c'; ", **argv); 677 Usage(); 678 exit(1); 679 } 680 } 681 } 682 else { /* Input file name */ 683 *fileptr++ = *argv; /* Build input file list */ 684 *fileptr = NULL; 685 /* process nextarg; */ 686 } 687 nextarg: continue; 688 } 689 690 if(maxbits < INIT_BITS) maxbits = INIT_BITS; 691 if (maxbits > BITS) maxbits = BITS; 692 maxmaxcode = (code_int) 1 << maxbits; 693 694 if (*filelist != NULL) { 695 for (fileptr = filelist; *fileptr; fileptr++) { 696 exit_stat = 0; 697 if (do_decomp != 0) { /* DECOMPRESSION */ 698 699#ifdef MSDOS 700 /* Check for .Z or XZ suffix; add one if necessary */ 701 cp = *fileptr + strlen(*fileptr) - 2; 702 if ((*cp != '.' && *cp != 'X' && *cp != 'x') || 703 (*(++cp) != 'Z' && *cp != 'z')) { 704 strcpy(tempname, *fileptr); 705 if ((cp=rindex(tempname,'.')) == NULL) 706 strcat(tempname, ".Z"); 707 else if(*(++cp) == '\0') strcat(tempname, "Z"); 708 else { 709 *(++cp) = '\0'; 710 strcat(tempname, "XZ"); 711 } 712 *fileptr = tempname; 713 } 714#else 715 /* Check for .Z suffix */ 716 if (strcmp(*fileptr + strlen(*fileptr) - 2, ".Z") != 0) { 717 /* No .Z: tack one on */ 718 strcpy(tempname, *fileptr); 719 strcat(tempname, ".Z"); 720 *fileptr = tempname; 721 } 722#endif /*MSDOS */ 723 724 /* Open input file for decompression */ 725 726#ifdef MSDOS 727 if ((freopen(*fileptr, "rb", stdin)) == NULL) { 728#else 729 if ((freopen(*fileptr, "r", stdin)) == NULL) { 730#endif 731 732 perror(*fileptr); continue; 733 } 734 /* Check the magic number */ 735 if (nomagic == 0) { 736 if ((getchar() != (magic_header[0] & 0xFF)) 737 || (getchar() != (magic_header[1] & 0xFF))) { 738 fprintf(stderr, "%s: not in compressed format\n", 739 *fileptr); 740 continue; 741 } 742 maxbits = getchar(); /* set -b from file */ 743 block_compress = maxbits & BLOCK_MASK; 744 maxbits &= BIT_MASK; 745 maxmaxcode = (code_int) 1 << maxbits; 746 if(maxbits > BITS) { 747 fprintf(stderr, 748 "%s: compressed with %d bits, can only handle %d bits\n", 749 *fileptr, maxbits, BITS); 750 continue; 751 } 752 } 753 /* Generate output filename */ 754 strcpy(ofname, *fileptr); 755 ofname[strlen(*fileptr) - 2] = '\0'; /* Strip off .Z */ 756 } else { /* COMPRESSION */ 757 758#ifdef MSDOS 759 cp = *fileptr + strlen(*fileptr) - 2; 760 if ((*cp == '.' || *cp == 'X' || *cp == 'x') && 761 (*(++cp) == 'Z' || *cp == 'z')) { 762 fprintf(stderr,"%s: already has %s suffix -- no change\n", 763 *fileptr,--cp); 764#else 765 if (strcmp(*fileptr + strlen(*fileptr) - 2, ".Z") == 0) { 766 fprintf(stderr, "%s: already has .Z suffix -- no change\n", 767 *fileptr); 768#endif /* MSDOS */ 769 770 continue; 771 } 772 /* Open input file for compression */ 773 774#ifdef MSDOS 775 if ((freopen(*fileptr, image == 2 ? "rt" : "rb", stdin)) 776 == NULL) { 777#else 778 if ((freopen(*fileptr, "r", stdin)) == NULL) { 779#endif 780 781 perror(*fileptr); continue; 782 } 783 stat ( *fileptr, &statbuf ); 784 fsize = (long) statbuf.st_size; 785 /* 786 * tune hash table size for small files -- ad hoc, 787 * but the sizes match earlier #defines, which 788 * serve as upper bounds on the number of output codes. 789 */ 790 hsize = HSIZE; 791 if ( fsize < (1 << 12) ) 792 hsize = min ( 5003, HSIZE ); 793 else if ( fsize < (1 << 13) ) 794 hsize = min ( 9001, HSIZE ); 795 else if ( fsize < (1 << 14) ) 796 hsize = min ( 18013, HSIZE ); 797 else if ( fsize < (1 << 15) ) 798 hsize = min ( 35023, HSIZE ); 799 else if ( fsize < 47000 ) 800 hsize = min ( 50021, HSIZE ); 801 802 /* Generate output filename */ 803 strcpy(ofname, *fileptr); 804#if !defined BSD4_2 && !defined BEOS /* Short filenames */ 805 if ((cp = rindex(ofname, PATH_SEP)) != NULL) cp++; 806 else cp = ofname; 807# ifdef MSDOS 808 if (zcat_flg == 0 && (sufp = rindex(cp, '.')) != NULL && 809 strlen(sufp) > 2) fprintf(stderr, 810 "%s: part of filename extension will be replaced by XZ\n", 811 cp); 812# else 813 if (strlen(cp) > 12) { 814 fprintf(stderr,"%s: filename too long to tack on .Z\n",cp); 815 continue; 816 } 817# endif 818#endif /* BSD4_2 Long filenames allowed */ 819 820#ifdef MSDOS 821 if ((cp = rindex(ofname, '.')) == NULL) strcat(ofname, ".Z"); 822 else { 823 if(*(++cp) != '\0') *(++cp) = '\0'; 824 strcat(ofname, "XZ"); 825 } 826#else 827 strcat(ofname, ".Z"); 828#endif /* MSDOS */ 829 830 } 831 precious = 0; 832 /* Check for overwrite of existing file */ 833 if (overwrite == 0 && zcat_flg == 0) { 834 if (stat(ofname, &statbuf) == 0) { 835 char response[2]; 836 response[0] = 'n'; 837 fprintf(stderr, "%s already exists;", ofname); 838#ifndef MSDOS 839 if (foreground()) { 840#endif 841 fprintf(stderr, 842 " do you wish to overwrite %s (y or n)? ", ofname); 843 fflush(stderr); 844 read(2, response, 2); 845 while (response[1] != '\n') { 846 if (read(2, response+1, 1) < 0) { /* Ack! */ 847 perror("stderr"); break; 848 } 849 } 850#ifndef MSDOS 851 } 852#endif 853 if (response[0] != 'y') { 854 fprintf(stderr, "\tnot overwritten\n"); 855 continue; 856 } 857 } 858 } 859 if(zcat_flg == 0) { /* Open output file */ 860 861#ifdef MSDOS 862 if (freopen(ofname, do_decomp && image == 2 ? "wt" : "wb", 863 stdout) == NULL) { 864#else 865 if (freopen(ofname, "w", stdout) == NULL) { 866#endif 867 868 perror(ofname); continue; 869 } 870 if(!quiet) 871 fprintf(stderr, "%s: ", *fileptr); 872 } 873 874 /* Actually do the compression/decompression */ 875 if (do_decomp == 0) compress(); 876#ifndef DEBUG 877 else decompress(); 878#else 879 else if (debug == 0) decompress(); 880 else printcodes(); 881 if (verbose) dump_tab(); 882#endif /* DEBUG */ 883 if(zcat_flg == 0) { 884 copystat(*fileptr, ofname); /* Copy stats */ 885 if((exit_stat == 1) || (!quiet)) 886 putc('\n', stderr); 887 } 888 } 889 } else { /* Standard input */ 890 if (do_decomp == 0) { 891 compress(); 892#ifdef DEBUG 893 if(verbose) dump_tab(); 894#endif /* DEBUG */ 895 if(!quiet) 896 putc('\n', stderr); 897 } else { 898 /* Check the magic number */ 899 if (nomagic == 0) { 900 if ((getchar()!=(magic_header[0] & 0xFF)) 901 || (getchar()!=(magic_header[1] & 0xFF))) { 902 fprintf(stderr, "stdin: not in compressed format\n"); 903 exit(1); 904 } 905 maxbits = getchar(); /* set -b from file */ 906 block_compress = maxbits & BLOCK_MASK; 907 maxbits &= BIT_MASK; 908 maxmaxcode = (code_int) 1 << maxbits; 909 fsize = 100000; /* assume stdin large for USERMEM */ 910 if(maxbits > BITS) { 911 fprintf(stderr, 912 "stdin: compressed with %d bits, can only handle %d bits\n", 913 maxbits, BITS); 914 exit(1); 915 } 916 } 917#ifndef DEBUG 918 decompress(); 919#else 920 if (debug == 0) decompress(); 921 else printcodes(); 922 if (verbose) dump_tab(); 923#endif /* DEBUG */ 924 } 925 } 926 exit(exit_stat); 927} 928 929static int offset; 930long int in_count = 1; /* length of input */ 931long int bytes_out; /* length of compressed output */ 932long int out_count = 0; /* # of codes output (for debugging) */ 933 934/* 935 * compress stdin to stdout 936 * 937 * Algorithm: use open addressing double hashing (no chaining) on the 938 * prefix code / next character combination. We do a variant of Knuth's 939 * algorithm D (vol. 3, sec. 6.4) along with G. Knott's relatively-prime 940 * secondary probe. Here, the modular division first probe is gives way 941 * to a faster exclusive-or manipulation. Also do block compression with 942 * an adaptive reset, whereby the code table is cleared when the compression 943 * ratio decreases, but after the table fills. The variable-length output 944 * codes are re-sized at this point, and a special CLEAR code is generated 945 * for the decompressor. Late addition: construct the table according to 946 * file size for noticeable speed improvement on small files. Please direct 947 * questions about this implementation to ames!jaw. 948 */ 949 950compress() { 951 register long fcode; 952 register code_int i = 0; 953 register int c; 954 register code_int ent; 955 register code_int disp; 956 register code_int hsize_reg; 957 register int hshift; 958 959#ifndef COMPATIBLE 960 if (nomagic == 0) { 961 putchar(magic_header[0]); putchar(magic_header[1]); 962 putchar((char)(maxbits | block_compress)); 963 if(ferror(stdout)) 964 writeerr(); 965 } 966#endif /* COMPATIBLE */ 967 968 offset = 0; 969 bytes_out = 3; /* includes 3-byte header mojo */ 970 out_count = 0; 971 clear_flg = 0; 972 ratio = 0; 973 in_count = 1; 974 checkpoint = CHECK_GAP; 975 maxcode = MAXCODE(n_bits = INIT_BITS); 976 free_ent = ((block_compress) ? FIRST : 256 ); 977 978 ent = getchar (); 979 980 hshift = 0; 981 for ( fcode = (long) hsize; fcode < 65536L; fcode *= 2L ) 982 hshift++; 983 hshift = 8 - hshift; /* set hash code range bound */ 984 985 hsize_reg = hsize; 986 cl_hash( (count_int) hsize_reg); /* clear hash table */ 987 988#ifdef SIGNED_COMPARE_SLOW 989 while ( (c = getchar()) != (unsigned) EOF ) { 990#else 991 while ( (c = getchar()) != EOF ) { 992#endif 993 994#ifdef MSDOS 995 if (c == '\n') in_count += image; else /* include CR if text mode */ 996#endif 997 998 in_count++; 999 1000 fcode = (long) (((long) c << maxbits) + ent); 1001 i = (((code_int)c << hshift) ^ ent); /* xor hashing */ 1002 1003 if ( htabof (i) == fcode ) { 1004 ent = codetabof (i); 1005 continue; 1006 } else if ( (long)htabof (i) < 0 ) /* empty slot */ 1007 goto nomatch; 1008 disp = hsize_reg - i; /* secondary hash (after G. Knott) */ 1009 if ( i == 0 ) 1010 disp = 1; 1011probe: 1012 if ( (i -= disp) < 0 ) 1013 i += hsize_reg; 1014 1015 if ( htabof (i) == fcode ) { 1016 ent = codetabof (i); 1017 continue; 1018 } 1019 if ( (long)htabof (i) > 0 ) 1020 goto probe; 1021nomatch: 1022 output ( (code_int) ent ); 1023 out_count++; 1024 ent = c; 1025#ifdef SIGNED_COMPARE_SLOW 1026 if ( (unsigned) free_ent < (unsigned) maxmaxcode) { 1027#else 1028 if ( free_ent < maxmaxcode ) { 1029#endif 1030 codetabof (i) = free_ent++; /* code -> hashtable */ 1031 htabof (i) = fcode; 1032 } 1033 else if ( (count_int)in_count >= checkpoint && block_compress ) 1034 cl_block (); 1035 } 1036 /* 1037 * Put out the final code. 1038 */ 1039 output( (code_int)ent ); 1040 out_count++; 1041 output( (code_int)-1 ); 1042 1043 /* 1044 * Print out stats on stderr 1045 */ 1046 if(zcat_flg == 0 && !quiet) { 1047#ifdef DEBUG 1048 fprintf( stderr, 1049 "%ld chars in, %ld codes (%ld bytes) out, compression factor: ", 1050 in_count, out_count, bytes_out ); 1051 prratio( stderr, in_count, bytes_out ); 1052 fprintf( stderr, "\n"); 1053 fprintf( stderr, "\tCompression as in compact: " ); 1054 prratio( stderr, in_count-bytes_out, in_count ); 1055 fprintf( stderr, "\n"); 1056 fprintf( stderr, "\tLargest code (of last block) was %d (%d bits)\n", 1057 free_ent - 1, n_bits ); 1058#else /* !DEBUG */ 1059 fprintf( stderr, "Compression: " ); 1060 prratio( stderr, in_count-bytes_out, in_count ); 1061#endif /* DEBUG */ 1062 } 1063 if(bytes_out > in_count) /* exit(2) if no savings */ 1064 exit_stat = 2; 1065 return; 1066} 1067 1068/***************************************************************** 1069 * TAG( output ) 1070 * 1071 * Output the given code. 1072 * Inputs: 1073 * code: A n_bits-bit integer. If == -1, then EOF. This assumes 1074 * that n_bits =< (long)wordsize - 1. 1075 * Outputs: 1076 * Outputs code to the file. 1077 * Assumptions: 1078 * Chars are 8 bits long. 1079 * Algorithm: 1080 * Maintain a BITS character long buffer (so that 8 codes will 1081 * fit in it exactly). Use the VAX insv instruction to insert each 1082 * code in turn. When the buffer fills up empty it and start over. 1083 */ 1084 1085static char buf[BITS]; 1086 1087#ifndef vax 1088char_type lmask[9] = {0xff, 0xfe, 0xfc, 0xf8, 0xf0, 0xe0, 0xc0, 0x80, 0x00}; 1089char_type rmask[9] = {0x00, 0x01, 0x03, 0x07, 0x0f, 0x1f, 0x3f, 0x7f, 0xff}; 1090#endif /* vax */ 1091 1092output( code ) 1093code_int code; 1094{ 1095#ifdef DEBUG 1096 static int col = 0; 1097#endif /* DEBUG */ 1098 1099 /* 1100 * On the VAX, it is important to have the register declarations 1101 * in exactly the order given, or the asm will break. 1102 */ 1103 register int r_off = offset, bits= n_bits; 1104 register char * bp = buf; 1105 1106#ifdef DEBUG 1107 if ( verbose ) 1108 fprintf( stderr, "%5d%c", code, 1109 (col+=6) >= 74 ? (col = 0, '\n') : ' ' ); 1110#endif /* DEBUG */ 1111 if ( code >= 0 ) { 1112#ifdef vax 1113 /* VAX DEPENDENT!! Implementation on other machines is below. 1114 * 1115 * Translation: Insert BITS bits from the argument starting at 1116 * offset bits from the beginning of buf. 1117 */ 1118 0; /* Work around for pcc -O bug with asm and if stmt */ 1119 asm( "insv 4(ap),r11,r10,(r9)" ); 1120#else /* not a vax */ 1121/* 1122 * byte/bit numbering on the VAX is simulated by the following code 1123 */ 1124 /* 1125 * Get to the first byte. 1126 */ 1127 bp += (r_off >> 3); 1128 r_off &= 7; 1129 /* 1130 * Since code is always >= 8 bits, only need to mask the first 1131 * hunk on the left. 1132 */ 1133 *bp = (*bp & rmask[r_off]) | (code << r_off) & lmask[r_off]; 1134 bp++; 1135 bits -= (8 - r_off); 1136 code >>= 8 - r_off; 1137 /* Get any 8 bit parts in the middle (<=1 for up to 16 bits). */ 1138 if ( bits >= 8 ) { 1139 *bp++ = code; 1140 code >>= 8; 1141 bits -= 8; 1142 } 1143 /* Last bits. */ 1144 if(bits) 1145 *bp = code; 1146#endif /* vax */ 1147 offset += n_bits; 1148 if ( offset == (n_bits << 3) ) { 1149 bp = buf; 1150 bits = n_bits; 1151 bytes_out += bits; 1152 do 1153 putchar(*bp++); 1154 while(--bits); 1155 offset = 0; 1156 } 1157 1158 /* 1159 * If the next entry is going to be too big for the code size, 1160 * then increase it, if possible. 1161 */ 1162 if ( free_ent > maxcode || (clear_flg > 0)) 1163 { 1164 /* 1165 * Write the whole buffer, because the input side won't 1166 * discover the size increase until after it has read it. 1167 */ 1168 if ( offset > 0 ) { 1169 if( fwrite( buf, 1, n_bits, stdout ) != n_bits) 1170 writeerr(); 1171 bytes_out += n_bits; 1172 } 1173 offset = 0; 1174 1175 if ( clear_flg ) { 1176 maxcode = MAXCODE (n_bits = INIT_BITS); 1177 clear_flg = 0; 1178 } 1179 else { 1180 n_bits++; 1181 if ( n_bits == maxbits ) 1182 maxcode = maxmaxcode; 1183 else 1184 maxcode = MAXCODE(n_bits); 1185 } 1186#ifdef DEBUG 1187 if ( debug ) { 1188 fprintf( stderr, "\nChange to %d bits\n", n_bits ); 1189 col = 0; 1190 } 1191#endif /* DEBUG */ 1192 } 1193 } else { 1194 /* 1195 * At EOF, write the rest of the buffer. 1196 */ 1197 if ( offset > 0 ) 1198 fwrite( buf, 1, (offset + 7) / 8, stdout ); 1199 bytes_out += (offset + 7) / 8; 1200 offset = 0; 1201 fflush( stdout ); 1202#ifdef DEBUG 1203 if ( verbose ) 1204 fprintf( stderr, "\n" ); 1205#endif /* DEBUG */ 1206 if( ferror( stdout ) ) 1207 writeerr(); 1208 } 1209} 1210 1211/* 1212 * Decompress stdin to stdout. This routine adapts to the codes in the 1213 * file building the "string" table on-the-fly; requiring no table to 1214 * be stored in the compressed file. The tables used herein are shared 1215 * with those of the compress() routine. See the definitions above. 1216 */ 1217 1218decompress() { 1219 1220#ifdef BIG 1221 register char_type far *stackp; 1222#else 1223 register char_type *stackp; 1224#endif 1225 1226 register int finchar; 1227 register code_int code, oldcode, incode; 1228 1229 /* 1230 * As above, initialize the first 256 entries in the table. 1231 */ 1232 maxcode = MAXCODE(n_bits = INIT_BITS); 1233 for ( code = 255; code >= 0; code-- ) { 1234 tab_prefixof(code) = 0; 1235 tab_suffixof(code) = (char_type)code; 1236 } 1237 free_ent = ((block_compress) ? FIRST : 256 ); 1238 1239 finchar = oldcode = getcode(); 1240 if(oldcode == -1) /* EOF already? */ 1241 return; /* Get out of here */ 1242 putchar( (char)finchar ); /* first code must be 8 bits = char */ 1243 if(ferror(stdout)) /* Crash if can't write */ 1244 writeerr(); 1245 stackp = de_stack; 1246 1247 while ( (code = getcode()) > -1 ) { 1248 1249 if ( (code == CLEAR) && block_compress ) { 1250 for ( code = 255; code >= 0; code-- ) 1251 tab_prefixof(code) = 0; 1252 clear_flg = 1; 1253 free_ent = FIRST - 1; 1254 if ( (code = getcode ()) == -1 ) /* O, untimely death! */ 1255 break; 1256 } 1257 incode = code; 1258 /* 1259 * Special case for KwKwK string. 1260 */ 1261 if ( code >= free_ent ) { 1262 *stackp++ = finchar; 1263 code = oldcode; 1264 } 1265 1266 /* 1267 * Generate output characters in reverse order 1268 */ 1269#ifdef SIGNED_COMPARE_SLOW 1270 while ( ((unsigned long)code) >= ((unsigned long)256) ) { 1271#else 1272 while ( code >= 256 ) { 1273#endif 1274 *stackp++ = tab_suffixof(code); 1275 code = tab_prefixof(code); 1276 } 1277 *stackp++ = finchar = tab_suffixof(code); 1278 1279 /* 1280 * And put them out in forward order 1281 */ 1282 do 1283 putchar ( *--stackp ); 1284 while ( stackp > de_stack ); 1285 1286 /* 1287 * Generate the new entry. 1288 */ 1289 if ( (code=free_ent) < maxmaxcode ) { 1290 tab_prefixof(code) = (unsigned short)oldcode; 1291 tab_suffixof(code) = finchar; 1292 free_ent = code+1; 1293 } 1294 /* 1295 * Remember previous code. 1296 */ 1297 oldcode = incode; 1298 } 1299 fflush( stdout ); 1300 if(ferror(stdout)) 1301 writeerr(); 1302} 1303 1304/***************************************************************** 1305 * TAG( getcode ) 1306 * 1307 * Read one code from the standard input. If EOF, return -1. 1308 * Inputs: 1309 * stdin 1310 * Outputs: 1311 * code or -1 is returned. 1312 */ 1313 1314code_int 1315getcode() { 1316 /* 1317 * On the VAX, it is important to have the register declarations 1318 * in exactly the order given, or the asm will break. 1319 */ 1320 register code_int code; 1321 static int offset = 0, size = 0; 1322 static char_type buf[BITS]; 1323 register int r_off, bits; 1324 register char_type *bp = buf; 1325 1326 if ( clear_flg > 0 || offset >= size || free_ent > maxcode ) { 1327 /* 1328 * If the next entry will be too big for the current code 1329 * size, then we must increase the size. This implies reading 1330 * a new buffer full, too. 1331 */ 1332 if ( free_ent > maxcode ) { 1333 n_bits++; 1334 if ( n_bits == maxbits ) 1335 maxcode = maxmaxcode; /* won't get any bigger now */ 1336 else 1337 maxcode = MAXCODE(n_bits); 1338 } 1339 if ( clear_flg > 0) { 1340 maxcode = MAXCODE (n_bits = INIT_BITS); 1341 clear_flg = 0; 1342 } 1343 size = fread( buf, 1, n_bits, stdin ); 1344 if ( size <= 0 ) 1345 return -1; /* end of file */ 1346 offset = 0; 1347 /* Round size down to integral number of codes */ 1348 size = (size << 3) - (n_bits - 1); 1349 } 1350 r_off = offset; 1351 bits = n_bits; 1352#ifdef vax 1353 asm( "extzv r10,r9,(r8),r11" ); 1354#else /* not a vax */ 1355 /* 1356 * Get to the first byte. 1357 */ 1358 bp += (r_off >> 3); 1359 r_off &= 7; 1360 /* Get first part (low order bits) */ 1361#ifdef NO_UCHAR 1362 code = ((*bp++ >> r_off) & rmask[8 - r_off]) & 0xff; 1363#else 1364 code = (*bp++ >> r_off); 1365#endif /* NO_UCHAR */ 1366 bits -= (8 - r_off); 1367 r_off = 8 - r_off; /* now, offset into code word */ 1368 /* Get any 8 bit parts in the middle (<=1 for up to 16 bits). */ 1369 if ( bits >= 8 ) { 1370#ifdef NO_UCHAR 1371 code |= (*bp++ & 0xff) << r_off; 1372#else 1373 code |= *bp++ << r_off; 1374#endif /* NO_UCHAR */ 1375 r_off += 8; 1376 bits -= 8; 1377 } 1378 /* high order bits. */ 1379 code |= (*bp & rmask[bits]) << r_off; 1380#endif /* vax */ 1381 offset += n_bits; 1382 1383 return code; 1384} 1385 1386char * 1387rindex(s, c) /* For those who don't have it in libc.a */ 1388register char *s, c; 1389{ 1390 char *p; 1391 for (p = NULL; *s; s++) 1392 if (*s == c) 1393 p = s; 1394 return(p); 1395} 1396 1397#ifdef DEBUG 1398printcodes() 1399{ 1400 /* 1401 * Just print out codes from input file. For debugging. 1402 */ 1403 code_int code; 1404 int col = 0, bits; 1405 1406 bits = n_bits = INIT_BITS; 1407 maxcode = MAXCODE(n_bits); 1408 free_ent = ((block_compress) ? FIRST : 256 ); 1409 while ( ( code = getcode() ) >= 0 ) { 1410 if ( (code == CLEAR) && block_compress ) { 1411 free_ent = FIRST - 1; 1412 clear_flg = 1; 1413 } 1414 else if ( free_ent < maxmaxcode ) 1415 free_ent++; 1416 if ( bits != n_bits ) { 1417 fprintf(stderr, "\nChange to %d bits\n", n_bits ); 1418 bits = n_bits; 1419 col = 0; 1420 } 1421 fprintf(stderr, "%5d%c", code, (col+=6) >= 74 ? (col = 0, '\n') : ' ' ); 1422 } 1423 putc( '\n', stderr ); 1424 exit( 0 ); 1425} 1426 1427code_int sorttab[1<<BITS]; /* sorted pointers into htab */ 1428 1429dump_tab() /* dump string table */ 1430{ 1431 register int i, first; 1432 register ent; 1433#define STACK_SIZE 15000 1434 int stack_top = STACK_SIZE; 1435 register c; 1436 1437 if(do_decomp == 0) { /* compressing */ 1438 register int flag = 1; 1439 1440 for(i=0; i<hsize; i++) { /* build sort pointers */ 1441 if((long)htabof(i) >= 0) { 1442 sorttab[codetabof(i)] = i; 1443 } 1444 } 1445 first = block_compress ? FIRST : 256; 1446 for(i = first; i < free_ent; i++) { 1447 fprintf(stderr, "%5d: \"", i); 1448 de_stack[--stack_top] = '\n'; 1449 de_stack[--stack_top] = '"'; 1450 stack_top = in_stack((htabof(sorttab[i])>>maxbits)&0xff, 1451 stack_top); 1452 for(ent=htabof(sorttab[i]) & ((1<<maxbits)-1); 1453 ent > 256; 1454 ent=htabof(sorttab[ent]) & ((1<<maxbits)-1)) { 1455 stack_top = in_stack(htabof(sorttab[ent]) >> maxbits, 1456 stack_top); 1457 } 1458 stack_top = in_stack(ent, stack_top); 1459 fwrite( &de_stack[stack_top], 1, STACK_SIZE-stack_top, stderr); 1460 stack_top = STACK_SIZE; 1461 } 1462 } else if(!debug) { /* decompressing */ 1463 1464 for ( i = 0; i < free_ent; i++ ) { 1465 ent = i; 1466 c = tab_suffixof(ent); 1467 if ( isascii(c) && isprint(c) ) 1468 fprintf( stderr, "%5d: %5d/'%c' \"", 1469 ent, tab_prefixof(ent), c ); 1470 else 1471 fprintf( stderr, "%5d: %5d/\\%03o \"", 1472 ent, tab_prefixof(ent), c ); 1473 de_stack[--stack_top] = '\n'; 1474 de_stack[--stack_top] = '"'; 1475 for ( ; ent != NULL; 1476 ent = (ent >= FIRST ? tab_prefixof(ent) : NULL) ) { 1477 stack_top = in_stack(tab_suffixof(ent), stack_top); 1478 } 1479 fwrite( &de_stack[stack_top], 1, STACK_SIZE - stack_top, stderr ); 1480 stack_top = STACK_SIZE; 1481 } 1482 } 1483} 1484 1485int 1486in_stack(c, stack_top) 1487 register c, stack_top; 1488{ 1489 if ( (isascii(c) && isprint(c) && c != '\\') || c == ' ' ) { 1490 de_stack[--stack_top] = c; 1491 } else { 1492 switch( c ) { 1493 case '\n': de_stack[--stack_top] = 'n'; break; 1494 case '\t': de_stack[--stack_top] = 't'; break; 1495 case '\b': de_stack[--stack_top] = 'b'; break; 1496 case '\f': de_stack[--stack_top] = 'f'; break; 1497 case '\r': de_stack[--stack_top] = 'r'; break; 1498 case '\\': de_stack[--stack_top] = '\\'; break; 1499 default: 1500 de_stack[--stack_top] = '0' + c % 8; 1501 de_stack[--stack_top] = '0' + (c / 8) % 8; 1502 de_stack[--stack_top] = '0' + c / 64; 1503 break; 1504 } 1505 de_stack[--stack_top] = '\\'; 1506 } 1507 return stack_top; 1508} 1509#endif /* DEBUG */ 1510 1511writeerr() 1512{ 1513 perror ( ofname ); 1514 unlink ( ofname ); 1515 exit ( 1 ); 1516} 1517 1518copystat(ifname, ofname) 1519char *ifname, *ofname; 1520{ 1521 struct stat statbuf; 1522 int mode; 1523 struct utimbuf timep; 1524 1525#ifdef MSDOS 1526 if (_osmajor < 3) freopen("CON","at",stdout); else /* MS-DOS 2.xx bug */ 1527#endif 1528 1529 fclose(stdout); 1530 if (stat(ifname, &statbuf)) { /* Get stat on input file */ 1531 perror(ifname); 1532 return; 1533 } 1534 1535#ifndef MSDOS 1536 if ((statbuf.st_mode & S_IFMT/*0170000*/) != S_IFREG/*0100000*/) { 1537 if(quiet) 1538 fprintf(stderr, "%s: ", ifname); 1539 fprintf(stderr, " -- not a regular file: unchanged"); 1540 exit_stat = 1; 1541 } else if (statbuf.st_nlink > 1) { 1542 if(quiet) 1543 fprintf(stderr, "%s: ", ifname); 1544 fprintf(stderr, " -- has %d other links: unchanged", 1545 statbuf.st_nlink - 1); 1546 exit_stat = 1; 1547 } else if (exit_stat == 2 && (!force)) { /* No compression: remove file.Z */ 1548#else 1549 if (exit_stat == 2 && (!force)) { /* No compression: remove file.Z */ 1550#endif /* MSDOS */ 1551 1552 if(!quiet) 1553 fprintf(stderr, " -- file unchanged"); 1554 } else { /* ***** Successful Compression ***** */ 1555 exit_stat = 0; 1556 mode = statbuf.st_mode & 07777; 1557 if (chmod(ofname, mode)) /* Copy modes */ 1558 perror(ofname); 1559 1560#ifndef MSDOS 1561 chown(ofname, statbuf.st_uid, statbuf.st_gid); /* Copy ownership */ 1562#endif 1563 1564 timep.actime = statbuf.st_atime; 1565 timep.modtime = statbuf.st_mtime; 1566 utime(ofname, &timep); /* Update last accessed and modified times */ 1567 precious = 1; 1568 fclose(stdin); 1569 if (unlink(ifname)) { /* Remove input file */ 1570 perror(ifname); 1571 } 1572 if(!quiet) 1573 fprintf(stderr, " -- replaced with %s", ofname); 1574 return; /* Successful return */ 1575 } 1576 1577 /* Unsuccessful return -- one of the tests failed */ 1578 if (unlink(ofname)) { 1579 perror(ofname); 1580 } 1581} 1582 1583#ifndef MSDOS 1584/* 1585 * This routine returns 1 if we are running in the foreground and stderr 1586 * is a tty. 1587 */ 1588foreground() 1589{ 1590 if(bgnd_flag != SIG_DFL) { /* background? */ 1591 return(0); 1592 } else { /* foreground */ 1593 if(isatty(2)) { /* and stderr is a tty */ 1594 return(1); 1595 } else { 1596 return(0); 1597 } 1598 } 1599} 1600#endif 1601 1602void 1603onintr ( ) 1604{ 1605 if (!precious) 1606 unlink ( ofname ); 1607 exit ( 1 ); 1608} 1609 1610#ifndef MSDOS 1611void 1612oops ( ) /* wild pointer -- assume bad input */ 1613{ 1614 if ( do_decomp == 1 ) 1615 fprintf ( stderr, "uncompress: corrupt input\n" ); 1616 unlink ( ofname ); 1617 exit ( 1 ); 1618} 1619#endif /* MSDOS */ 1620 1621cl_block () /* table clear for block compress */ 1622{ 1623 register long int rat; 1624 1625 checkpoint = in_count + CHECK_GAP; 1626#ifdef DEBUG 1627 if ( debug ) { 1628 fprintf ( stderr, "count: %ld, ratio: ", in_count ); 1629 prratio ( stderr, in_count, bytes_out ); 1630 fprintf ( stderr, "\n"); 1631 } 1632#endif /* DEBUG */ 1633 1634 if(in_count > 0x007fffff) { /* shift will overflow */ 1635 rat = bytes_out >> 8; 1636 if(rat == 0) { /* Don't divide by zero */ 1637 rat = 0x7fffffff; 1638 } else { 1639 rat = in_count / rat; 1640 } 1641 } else { 1642 rat = (in_count << 8) / bytes_out; /* 8 fractional bits */ 1643 } 1644 if ( rat > ratio ) { 1645 ratio = rat; 1646 } else { 1647 ratio = 0; 1648#ifdef DEBUG 1649 if(verbose) 1650 dump_tab(); /* dump string table */ 1651#endif 1652 cl_hash ( (count_int) hsize ); 1653 free_ent = FIRST; 1654 clear_flg = 1; 1655 output ( (code_int) CLEAR ); 1656#ifdef DEBUG 1657 if(debug) 1658 fprintf ( stderr, "clear\n" ); 1659#endif /* DEBUG */ 1660 } 1661} 1662 1663cl_hash(hsize) /* reset code table */ 1664 register count_int hsize; 1665{ 1666 1667#ifdef XENIX_16 1668 register j; 1669 register long k = hsize; 1670 1671# ifdef MSDOS 1672 register count_int far *htab_p; 1673# else 1674 register count_int *htab_p; 1675# endif /* MSDOS */ 1676 1677#else /* Normal machine */ 1678 register count_int *htab_p = htab+hsize; 1679#endif /* XENIX_16 */ 1680 1681 register long i; 1682 register long m1 = -1; 1683 1684#ifdef XENIX_16 1685 for(j=0; j<=8 && k>=0; j++,k-=8192) { 1686 i = 8192; 1687 if(k < 8192) { 1688 i = k; 1689 } 1690 htab_p = &(htab[j][i]); 1691 i -= 16; 1692 if(i > 0) { 1693#else 1694 i = hsize - 16; 1695#endif 1696 do { /* might use Sys V memset(3) here */ 1697 *(htab_p-16) = m1; 1698 *(htab_p-15) = m1; 1699 *(htab_p-14) = m1; 1700 *(htab_p-13) = m1; 1701 *(htab_p-12) = m1; 1702 *(htab_p-11) = m1; 1703 *(htab_p-10) = m1; 1704 *(htab_p-9) = m1; 1705 *(htab_p-8) = m1; 1706 *(htab_p-7) = m1; 1707 *(htab_p-6) = m1; 1708 *(htab_p-5) = m1; 1709 *(htab_p-4) = m1; 1710 *(htab_p-3) = m1; 1711 *(htab_p-2) = m1; 1712 *(htab_p-1) = m1; 1713 htab_p -= 16; 1714 } while ((i -= 16) >= 0); 1715#ifdef XENIX_16 1716 } 1717 } 1718#endif 1719 for ( i += 16; i > 0; i-- ) 1720 *--htab_p = m1; 1721} 1722 1723prratio(stream, num, den) 1724FILE *stream; 1725long int num, den; 1726{ 1727 1728#ifdef DEBUG 1729 register long q; /* permits |result| > 655.36% */ 1730#else 1731 register int q; /* Doesn't need to be long */ 1732#endif 1733 1734 if(num > 214748L) { /* 2147483647/10000 */ 1735 q = num / (den / 10000L); 1736 } else { 1737 q = 10000L * num / den; /* Long calculations, though */ 1738 } 1739 if (q < 0) { 1740 putc('-', stream); 1741 q = -q; 1742 } 1743 fprintf(stream, "%d.%02d%%", (int)(q / 100), (int)(q % 100)); 1744} 1745 1746version() 1747{ 1748 fprintf(stderr, "%s\n", rcs_ident); 1749 fprintf(stderr, "Options: "); 1750#ifdef vax 1751 fprintf(stderr, "vax, "); 1752#endif 1753#ifdef NO_UCHAR 1754 fprintf(stderr, "NO_UCHAR, "); 1755#endif 1756#ifdef SIGNED_COMPARE_SLOW 1757 fprintf(stderr, "SIGNED_COMPARE_SLOW, "); 1758#endif 1759#ifdef MSDOS 1760 fprintf(stderr, "MSDOS, "); 1761#endif 1762#ifdef XENIX_16 1763 fprintf(stderr, "XENIX_16, "); 1764#endif 1765#ifdef COMPATIBLE 1766 fprintf(stderr, "COMPATIBLE, "); 1767#endif 1768#ifdef DEBUG 1769 fprintf(stderr, "DEBUG, "); 1770#endif 1771#ifdef BSD4_2 1772 fprintf(stderr, "BSD4_2, "); 1773#endif 1774 fprintf(stderr, "BITS = %d\n", BITS); 1775} 1776