home *** CD-ROM | disk | FTP | other *** search
- From: Peter Jannesen <peter@ncs.nl>
- Subject: v02i026: ncompress - (Ver. 4.2.3) an improved file compressor, Part02/02
- Newsgroups: comp.sources.reviewed
- Approved: csr@calvin.dgbt.doc.ca
-
- Submitted-by: Peter Jannesen <peter@ncs.nl>
- Posting-number: Volume 2, Issue 26
- Archive-name: ncompress/part02
-
- #! /bin/sh
- # This is a shell archive. Remove anything before this line, then unpack
- # it by saving it into a file and typing "sh file". To overwrite existing
- # files, type "sh file -c". You can also feed this as standard input via
- # unshar, or by typing "sh <file", e.g.. If this archive is complete, you
- # will see the following message at the end:
- # "End of archive 2 (of 2)."
- # Contents: compress42.c
- # Wrapped by csr@calvin.dgbt.doc.ca on Wed Aug 5 18:27:12 1992
- PATH=/bin:/usr/bin:/usr/ucb ; export PATH
- if test -f 'compress42.c' -a "${1}" != "-c" ; then
- echo shar: Will not clobber existing file \"'compress42.c'\"
- else
- echo shar: Extracting \"'compress42.c'\" \(48463 characters\)
- sed "s/^X//" >'compress42.c' <<'END_OF_FILE'
- X/* (N)compress42.c - File compression ala IEEE Computer, Mar 1992.
- X *
- X * Authors:
- X * Spencer W. Thomas (decvax!harpo!utah-cs!utah-gr!thomas)
- X * Jim McKie (decvax!mcvax!jim)
- X * Steve Davies (decvax!vax135!petsd!peora!srd)
- X * Ken Turkowski (decvax!decwrl!turtlevax!ken)
- X * James A. Woods (decvax!ihnp4!ames!jaw)
- X * Joe Orost (decvax!vax135!petsd!joe)
- X * Dave Mack (csu@alembic.acs.com)
- X * Peter Jannesen, Network Communication Systems
- X * (peter@ncs.nl)
- X *
- X * Revision 4.2.3 92/03/14 peter@ncs.nl
- X * Optimise compress and decompress function and a lot of cleanups.
- X * New fast hash algoritme added (if more than 800Kb available).
- X *
- X * Revision 4.1 91/05/26 csu@alembic.acs.com
- X * Modified to recursively compress directories ('r' flag). As a side
- X * effect, compress will no longer attempt to compress things that
- X * aren't "regular" files. See Changes.
- X *
- X * Revision 4.0 85/07/30 12:50:00 joe
- X * Removed ferror() calls in output routine on every output except first.
- X * Prepared for release to the world.
- X *
- X * Revision 3.6 85/07/04 01:22:21 joe
- X * Remove much wasted storage by overlaying hash table with the tables
- X * used by decompress: tab_suffix[1<<BITS], stack[8000]. Updated USERMEM
- X * computations. Fixed dump_tab() DEBUG routine.
- X *
- X * Revision 3.5 85/06/30 20:47:21 jaw
- X * Change hash function to use exclusive-or. Rip out hash cache. These
- X * speedups render the megamemory version defunct, for now. Make decoder
- X * stack global. Parts of the RCS trunks 2.7, 2.6, and 2.1 no longer apply.
- X *
- X * Revision 3.4 85/06/27 12:00:00 ken
- X * Get rid of all floating-point calculations by doing all compression ratio
- X * calculations in fixed point.
- X *
- X * Revision 3.3 85/06/24 21:53:24 joe
- X * Incorporate portability suggestion for M_XENIX. Got rid of text on #else
- X * and #endif lines. Cleaned up #ifdefs for vax and interdata.
- X *
- X * Revision 3.2 85/06/06 21:53:24 jaw
- X * Incorporate portability suggestions for Z8000, IBM PC/XT from mailing list.
- X * Default to "quiet" output (no compression statistics).
- X *
- X * Revision 3.1 85/05/12 18:56:13 jaw
- X * Integrate decompress() stack speedups (from early pointer mods by McKie).
- X * Repair multi-file USERMEM gaffe. Unify 'force' flags to mimic semantics
- X * of SVR2 'pack'. Streamline block-compress table clear logic. Increase
- X * output byte count by magic number size.
- X *
- X * Revision 3.0 84/11/27 11:50:00 petsd!joe
- X * Set HSIZE depending on BITS. Set BITS depending on USERMEM. Unrolled
- X * loops in clear routines. Added "-C" flag for 2.0 compatibility. Used
- X * unsigned compares on Perkin-Elmer. Fixed foreground check.
- X *
- X * Revision 2.7 84/11/16 19:35:39 ames!jaw
- X * Cache common hash codes based on input statistics; this improves
- X * performance for low-density raster images. Pass on #ifdef bundle
- X * from Turkowski.
- X *
- X * Revision 2.6 84/11/05 19:18:21 ames!jaw
- X * Vary size of hash tables to reduce time for small files.
- X * Tune PDP-11 hash function.
- X *
- X * Revision 2.5 84/10/30 20:15:14 ames!jaw
- X * Junk chaining; replace with the simpler (and, on the VAX, faster)
- X * double hashing, discussed within. Make block compression standard.
- X *
- X * Revision 2.4 84/10/16 11:11:11 ames!jaw
- X * Introduce adaptive reset for block compression, to boost the rate
- X * another several percent. (See mailing list notes.)
- X *
- X * Revision 2.3 84/09/22 22:00:00 petsd!joe
- X * Implemented "-B" block compress. Implemented REVERSE sorting of tab_next.
- X * Bug fix for last bits. Changed fwrite to putchar loop everywhere.
- X *
- X * Revision 2.2 84/09/18 14:12:21 ames!jaw
- X * Fold in news changes, small machine typedef from thomas,
- X * #ifdef interdata from joe.
- X *
- X * Revision 2.1 84/09/10 12:34:56 ames!jaw
- X * Configured fast table lookup for 32-bit machines.
- X * This cuts user time in half for b <= FBITS, and is useful for news batching
- X * from VAX to PDP sites. Also sped up decompress() [fwrite->putc] and
- X * added signal catcher [plus beef in write_error()] to delete effluvia.
- X *
- X * Revision 2.0 84/08/28 22:00:00 petsd!joe
- X * Add check for foreground before prompting user. Insert maxbits into
- X * compressed file. Force file being uncompressed to end with ".Z".
- X * Added "-c" flag and "zcat". Prepared for release.
- X *
- X * Revision 1.10 84/08/24 18:28:00 turtlevax!ken
- X * Will only compress regular files (no directories), added a magic number
- X * header (plus an undocumented -n flag to handle old files without headers),
- X * added -f flag to force overwriting of possibly existing destination file,
- X * otherwise the user is prompted for a response. Will tack on a .Z to a
- X * filename if it doesn't have one when decompressing. Will only replace
- X * file if it was compressed.
- X *
- X * Revision 1.9 84/08/16 17:28:00 turtlevax!ken
- X * Removed scanargs(), getopt(), added .Z extension and unlimited number of
- X * filenames to compress. Flags may be clustered (-Ddvb12) or separated
- X * (-D -d -v -b 12), or combination thereof. Modes and other status is
- X * copied with copystat(). -O bug for 4.2 seems to have disappeared with
- X * 1.8.
- X *
- X * Revision 1.8 84/08/09 23:15:00 joe
- X * Made it compatible with vax version, installed jim's fixes/enhancements
- X *
- X * Revision 1.6 84/08/01 22:08:00 joe
- X * Sped up algorithm significantly by sorting the compress chain.
- X *
- X * Revision 1.5 84/07/13 13:11:00 srd
- X * Added C version of vax asm routines. Changed structure to arrays to
- X * save much memory. Do unsigned compares where possible (faster on
- X * Perkin-Elmer)
- X *
- X * Revision 1.4 84/07/05 03:11:11 thomas
- X * Clean up the code a little and lint it. (Lint complains about all
- X * the regs used in the asm, but I'm not going to "fix" this.)
- X *
- X * Revision 1.3 84/07/05 02:06:54 thomas
- X * Minor fixes.
- X *
- X * Revision 1.2 84/07/05 00:27:27 thomas
- X * Add variable bit length output.
- X *
- X */
- X#include <stdio.h>
- X#include <fcntl.h>
- X#include <ctype.h>
- X#include <signal.h>
- X#include <sys/types.h>
- X#include <sys/stat.h>
- X#include <errno.h>
- X
- X#ifdef DIRENT
- X# include <dirent.h>
- X# define RECURSIVE 1
- X# undef SYSDIR
- X#endif
- X#ifdef SYSDIR
- X# include <sys/dir.h>
- X# define RECURSIVE 1
- X#endif
- X#ifdef UTIME_H
- X# include <utime.h>
- X#else
- X struct utimbuf {
- X time_t actime;
- X time_t modtime;
- X };
- X#endif
- X
- X#ifdef __STDC__
- X# define ARGS(a) a
- X#else
- X# define ARGS(a) ()
- X#endif
- X
- X#define LARGS(a) () /* Relay on include files for libary func defs. */
- X
- X#ifndef SIG_TYPE
- X# define SIG_TYPE void (*)()
- X#endif
- X
- X#ifndef NOFUNCDEF
- X extern void *malloc LARGS((int));
- X extern void free LARGS((void *));
- X#ifndef _IBMR2
- X extern int open LARGS((char const *,int,...));
- X#endif
- X extern int close LARGS((int));
- X extern int read LARGS((int,void *,int));
- X extern int write LARGS((int,void const *,int));
- X extern int chmod LARGS((char const *,int));
- X extern int unlink LARGS((char const *));
- X extern int chown LARGS((char const *,int,int));
- X extern int utime LARGS((char const *,struct utimbuf const *));
- X extern char *strcpy LARGS((char *,char const *));
- X extern char *strcat LARGS((char *,char const *));
- X extern int strcmp LARGS((char const *,char const *));
- X extern unsigned strlen LARGS((char const *));
- X extern void *memset LARGS((void *,char,unsigned int));
- X extern void *memcpy LARGS((void *,void const *,unsigned int));
- X extern int atoi LARGS((char const *));
- X extern void exit LARGS((int));
- X extern int isatty LARGS((int));
- X#endif
- X
- X#define MARK(a) { asm(" .globl M.a"); asm("M.a:"); }
- X
- X#ifdef DEF_ERRNO
- X extern int errno;
- X#endif
- X
- X#include "patchlevel.h"
- X
- X#undef min
- X#define min(a,b) ((a>b) ? b : a)
- X
- X#ifndef IBUFSIZ
- X# define IBUFSIZ BUFSIZ /* Defailt input buffer size */
- X#endif
- X#ifndef OBUFSIZ
- X# define OBUFSIZ BUFSIZ /* Default output buffer size */
- X#endif
- X
- X#define MAXPATHLEN 1024 /* MAXPATHLEN - maximum length of a pathname we allow */
- X#define SIZE_INNER_LOOP 256 /* Size of the inter (fast) compress loop */
- X
- X /* Defines for third byte of header */
- X#define MAGIC_1 (char_type)'\037'/* First byte of compressed file */
- X#define MAGIC_2 (char_type)'\235'/* Second byte of compressed file */
- X#define BIT_MASK 0x1f /* Mask for 'number of compresssion bits' */
- X /* Masks 0x20 and 0x40 are free. */
- X /* I think 0x20 should mean that there is */
- X /* a fourth header byte (for expansion). */
- X#define BLOCK_MODE 0x80 /* Block compresssion if table is full and */
- X /* compression rate is dropping flush tables */
- X
- X /* the next two codes should not be changed lightly, as they must not */
- X /* lie within the contiguous general code space. */
- X#define FIRST 257 /* first free entry */
- X#define CLEAR 256 /* table clear output code */
- X
- X#define INIT_BITS 9 /* initial number of bits/code */
- X
- X#ifndef SACREDMEM
- X /*
- X * SACREDMEM is the amount of physical memory saved for others; compress
- X * will hog the rest.
- X */
- X# define SACREDMEM 0
- X#endif
- X
- X#ifndef USERMEM
- X /*
- X * Set USERMEM to the maximum amount of physical user memory available
- X * in bytes. USERMEM is used to determine the maximum BITS that can be used
- X * for compression.
- X */
- X# define USERMEM 450000 /* default user memory */
- X#endif
- X
- X#ifndef BYTEORDER
- X# define BYTEORDER 0000
- X#endif
- X
- X#ifndef NOALLIGN
- X# define NOALLIGN 0
- X#endif
- X
- X/*
- X * machine variants which require cc -Dmachine: pdp11, z8000, DOS
- X */
- X
- X#ifdef interdata /* Perkin-Elmer */
- X# define SIGNED_COMPARE_SLOW /* signed compare is slower than unsigned */
- X#endif
- X
- X#ifdef pdp11 /* PDP11: don't forget to compile with -i */
- X# define BITS 12 /* max bits/code for 16-bit machine */
- X# define NO_UCHAR /* also if "unsigned char" functions as signed char */
- X#endif /* pdp11 */
- X
- X#ifdef z8000 /* Z8000: */
- X# define BITS 12 /* 16-bits processor max 12 bits */
- X# undef vax /* weird preprocessor */
- X#endif /* z8000 */
- X
- X#ifdef DOS /* PC/XT/AT (8088) processor */
- X# define BITS 16 /* 16-bits processor max 12 bits */
- X# if BITS == 16
- X# define MAXSEG_64K
- X# endif
- X# undef BYTEORDER
- X# define BYTEORDER 4321
- X# undef NOALLIGN
- X# define NOALLIGN 1
- X# define COMPILE_DATE __DATE__
- X#endif /* DOS */
- X
- X#ifndef O_BINARY
- X# define O_BINARY 0 /* System has no binary mode */
- X#endif
- X
- X#ifdef M_XENIX /* Stupid compiler can't handle arrays with */
- X# if BITS == 16 /* more than 65535 bytes - so we fake it */
- X# define MAXSEG_64K
- X# else
- X# if BITS > 13 /* Code only handles BITS = 12, 13, or 16 */
- X# define BITS 13
- X# endif
- X# endif
- X#endif
- X
- X#ifndef BITS /* General processor calculate BITS */
- X# if USERMEM >= (800000+SACREDMEM)
- X# define FAST
- X# else
- X# if USERMEM >= (433484+SACREDMEM)
- X# define BITS 16
- X# else
- X# if USERMEM >= (229600+SACREDMEM)
- X# define BITS 15
- X# else
- X# if USERMEM >= (127536+SACREDMEM)
- X# define BITS 14
- X# else
- X# if USERMEM >= (73464+SACREDMEM)
- X# define BITS 13
- X# else
- X# define BITS 12
- X# endif
- X# endif
- X# endif
- X# endif
- X# endif
- X#endif /* BITS */
- X
- X#ifdef FAST
- X# define HBITS 17 /* 50% occupancy */
- X# define HSIZE (1<<HBITS)
- X# define HMASK (HSIZE-1)
- X# define HPRIME 9941
- X# define BITS 16
- X# undef MAXSEG_64K
- X#else
- X# if BITS == 16
- X# define HSIZE 69001 /* 95% occupancy */
- X# endif
- X# if BITS == 15
- X# define HSIZE 35023 /* 94% occupancy */
- X# endif
- X# if BITS == 14
- X# define HSIZE 18013 /* 91% occupancy */
- X# endif
- X# if BITS == 13
- X# define HSIZE 9001 /* 91% occupancy */
- X# endif
- X# if BITS <= 12
- X# define HSIZE 5003 /* 80% occupancy */
- X# endif
- X#endif
- X
- X#define CHECK_GAP 10000
- X
- Xtypedef long int code_int;
- X
- X#ifdef SIGNED_COMPARE_SLOW
- X typedef unsigned long int count_int;
- X typedef unsigned short int count_short;
- X typedef unsigned long int cmp_code_int; /* Cast to make compare faster */
- X#else
- X typedef long int count_int;
- X typedef long int cmp_code_int;
- X#endif
- X
- Xtypedef unsigned char char_type;
- X
- X#define ARGVAL() (*++(*argv) || (--argc && *++argv))
- X
- X#define MAXCODE(n) (1L << (n))
- X
- X#ifndef REGISTERS
- X# define REGISTERS 2
- X#endif
- X#define REG1
- X#define REG2
- X#define REG3
- X#define REG4
- X#define REG5
- X#define REG6
- X#define REG7
- X#define REG8
- X#define REG9
- X#define REG10
- X#define REG11
- X#define REG12
- X#define REG13
- X#define REG14
- X#define REG15
- X#define REG16
- X#if REGISTERS >= 1
- X# undef REG1
- X# define REG1 register
- X#endif
- X#if REGISTERS >= 2
- X# undef REG2
- X# define REG2 register
- X#endif
- X#if REGISTERS >= 3
- X# undef REG3
- X# define REG3 register
- X#endif
- X#if REGISTERS >= 4
- X# undef REG4
- X# define REG4 register
- X#endif
- X#if REGISTERS >= 5
- X# undef REG5
- X# define REG5 register
- X#endif
- X#if REGISTERS >= 6
- X# undef REG6
- X# define REG6 register
- X#endif
- X#if REGISTERS >= 7
- X# undef REG7
- X# define REG7 register
- X#endif
- X#if REGISTERS >= 8
- X# undef REG8
- X# define REG8 register
- X#endif
- X#if REGISTERS >= 9
- X# undef REG9
- X# define REG9 register
- X#endif
- X#if REGISTERS >= 10
- X# undef REG10
- X# define REG10 register
- X#endif
- X#if REGISTERS >= 11
- X# undef REG11
- X# define REG11 register
- X#endif
- X#if REGISTERS >= 12
- X# undef REG12
- X# define REG12 register
- X#endif
- X#if REGISTERS >= 13
- X# undef REG13
- X# define REG13 register
- X#endif
- X#if REGISTERS >= 14
- X# undef REG14
- X# define REG14 register
- X#endif
- X#if REGISTERS >= 15
- X# undef REG15
- X# define REG15 register
- X#endif
- X#if REGISTERS >= 16
- X# undef REG16
- X# define REG16 register
- X#endif
- X
- X
- Xunion bytes
- X{
- X long word;
- X struct
- X {
- X#if BYTEORDER == 4321
- X char_type b1;
- X char_type b2;
- X char_type b3;
- X char_type b4;
- X#else
- X#if BYTEORDER == 1234
- X char_type b4;
- X char_type b3;
- X char_type b2;
- X char_type b1;
- X#else
- X# undef BYTEORDER
- X int dummy;
- X#endif
- X#endif
- X } bytes;
- X} ;
- X#if BYTEORDER == 4321 && NOALLIGN == 1
- X#define output(b,o,c,n) { \
- X *(long *)&((b)[(o)>>3]) |= ((long)(c))<<((o)&0x7);\
- X (o) += (n); \
- X }
- X#else
- X#ifdef BYTEORDER
- X#define output(b,o,c,n) { REG1 char_type *p = &(b)[(o)>>3]; \
- X union bytes i; \
- X i.word = ((long)(c))<<((o)&0x7); \
- X p[0] |= i.bytes.b1; \
- X p[1] |= i.bytes.b2; \
- X p[2] |= i.bytes.b3; \
- X (o) += (n); \
- X }
- X#else
- X#define output(b,o,c,n) { REG1 char_type *p = &(b)[(o)>>3]; \
- X REG2 long i = ((long)(c))<<((o)&0x7); \
- X p[0] |= (char_type)(i); \
- X p[1] |= (char_type)(i>>8); \
- X p[2] |= (char_type)(i>>16); \
- X (o) += (n); \
- X }
- X#endif
- X#endif
- X#if BYTEORDER == 4321 && NOALLIGN == 1
- X#define input(b,o,c,n,m){ \
- X (c) = (*(long *)(&(b)[(o)>>3])>>((o)&0x7))&(m); \
- X (o) += (n); \
- X }
- X#else
- X#define input(b,o,c,n,m){ REG1 char_type *p = &(b)[(o)>>3]; \
- X (c) = ((((long)(p[0]))|((long)(p[1])<<8)| \
- X ((long)(p[2])<<16))>>((o)&0x7))&(m); \
- X (o) += (n); \
- X }
- X#endif
- X
- Xchar *progname; /* Program name */
- Xint silent = 0; /* don't tell me about errors */
- Xint quiet = 1; /* don't tell me about compression */
- Xint do_decomp = 0; /* Decompress mode */
- Xint force = 0; /* Force overwrite of files and links */
- Xint nomagic = 0; /* Use a 3-byte magic number header, */
- X /* unless old file */
- Xint block_mode = BLOCK_MODE;/* Block compress mode -C compatible with 2.0*/
- Xint maxbits = BITS; /* user settable max # bits/code */
- Xint zcat_flg = 0; /* Write output on stdout, suppress messages */
- Xint recursive = 0; /* compress directories */
- Xint exit_code = -1; /* Exitcode of compress (-1 no file compressed) */
- X
- Xchar_type inbuf[IBUFSIZ+64]; /* Input buffer */
- Xchar_type outbuf[OBUFSIZ+2048];/* Output buffer */
- X
- Xstruct stat infstat; /* Input file status */
- Xchar *ifname; /* Input filename */
- Xint remove_ofname = 0; /* Remove output file on a error */
- Xchar ofname[MAXPATHLEN]; /* Output filename */
- Xint fgnd_flag = 0; /* Running in background (SIGINT=SIGIGN) */
- X
- Xlong bytes_in; /* Total number of byte from input */
- Xlong bytes_out; /* Total number of byte to output */
- X
- X/*
- X * 8086 & 80286 Has a problem with array bigger than 64K so fake the array
- X * For processors with a limited address space and segments.
- X */
- X/*
- X * To save much memory, we overlay the table used by compress() with those
- X * used by decompress(). The tab_prefix table is the same size and type
- X * as the codetab. The tab_suffix table needs 2**BITS characters. We
- X * get this from the beginning of htab. The output stack uses the rest
- X * of htab, and contains characters. There is plenty of room for any
- X * possible stack (stack used to be 8000 characters).
- X */
- X#ifdef MAXSEG_64K
- X count_int htab0[8192];
- X count_int htab1[8192];
- X count_int htab2[8192];
- X count_int htab3[8192];
- X count_int htab4[8192];
- X count_int htab5[8192];
- X count_int htab6[8192];
- X count_int htab7[8192];
- X count_int htab8[HSIZE-65536];
- X count_int * htab[9] = {htab0,htab1,htab2,htab3,htab4,htab5,htab6,htab7,htab8};
- X
- X unsigned short code0tab[16384];
- X unsigned short code1tab[16384];
- X unsigned short code2tab[16384];
- X unsigned short code3tab[16384];
- X unsigned short code4tab[16384];
- X unsigned short * codetab[5] = {code0tab,code1tab,code2tab,code3tab,code4tab};
- X
- X# define htabof(i) (htab[(i) >> 13][(i) & 0x1fff])
- X# define codetabof(i) (codetab[(i) >> 14][(i) & 0x3fff])
- X# define tab_prefixof(i) codetabof(i)
- X# define tab_suffixof(i) ((char_type *)htab[(i)>>15])[(i) & 0x7fff]
- X# define de_stack ((char_type *)(&htab2[8191]))
- X void clear_htab()
- X {
- X memset(htab0, -1, sizeof(htab0));
- X memset(htab1, -1, sizeof(htab1));
- X memset(htab2, -1, sizeof(htab2));
- X memset(htab3, -1, sizeof(htab3));
- X memset(htab4, -1, sizeof(htab4));
- X memset(htab5, -1, sizeof(htab5));
- X memset(htab6, -1, sizeof(htab6));
- X memset(htab7, -1, sizeof(htab7));
- X memset(htab8, -1, sizeof(htab8));
- X }
- X# define clear_tab_prefixof() memset(code0tab, 0, 256);
- X#else /* Normal machine */
- X count_int htab[HSIZE];
- X unsigned short codetab[HSIZE];
- X
- X# define htabof(i) htab[i]
- X# define codetabof(i) codetab[i]
- X# define tab_prefixof(i) codetabof(i)
- X# define tab_suffixof(i) ((char_type *)(htab))[i]
- X# define de_stack ((char_type *)&(htab[HSIZE-1]))
- X# define clear_htab() memset(htab, -1, sizeof(htab))
- X# define clear_tab_prefixof() memset(codetab, 0, 256);
- X#endif /* MAXSEG_64K */
- X
- X#ifdef FAST
- X int primetab[256] = /* Special secudary hash table. */
- X {
- X 1013, -1061, 1109, -1181, 1231, -1291, 1361, -1429,
- X 1481, -1531, 1583, -1627, 1699, -1759, 1831, -1889,
- X 1973, -2017, 2083, -2137, 2213, -2273, 2339, -2383,
- X 2441, -2531, 2593, -2663, 2707, -2753, 2819, -2887,
- X 2957, -3023, 3089, -3181, 3251, -3313, 3361, -3449,
- X 3511, -3557, 3617, -3677, 3739, -3821, 3881, -3931,
- X 4013, -4079, 4139, -4219, 4271, -4349, 4423, -4493,
- X 4561, -4639, 4691, -4783, 4831, -4931, 4973, -5023,
- X 5101, -5179, 5261, -5333, 5413, -5471, 5521, -5591,
- X 5659, -5737, 5807, -5857, 5923, -6029, 6089, -6151,
- X 6221, -6287, 6343, -6397, 6491, -6571, 6659, -6709,
- X 6791, -6857, 6917, -6983, 7043, -7129, 7213, -7297,
- X 7369, -7477, 7529, -7577, 7643, -7703, 7789, -7873,
- X 7933, -8017, 8093, -8171, 8237, -8297, 8387, -8461,
- X 8543, -8627, 8689, -8741, 8819, -8867, 8963, -9029,
- X 9109, -9181, 9241, -9323, 9397, -9439, 9511, -9613,
- X 9677, -9743, 9811, -9871, 9941,-10061,10111,-10177,
- X 10259,-10321,10399,-10477,10567,-10639,10711,-10789,
- X 10867,-10949,11047,-11113,11173,-11261,11329,-11423,
- X 11491,-11587,11681,-11777,11827,-11903,11959,-12041,
- X 12109,-12197,12263,-12343,12413,-12487,12541,-12611,
- X 12671,-12757,12829,-12917,12979,-13043,13127,-13187,
- X 13291,-13367,13451,-13523,13619,-13691,13751,-13829,
- X 13901,-13967,14057,-14153,14249,-14341,14419,-14489,
- X 14557,-14633,14717,-14767,14831,-14897,14983,-15083,
- X 15149,-15233,15289,-15359,15427,-15497,15583,-15649,
- X 15733,-15791,15881,-15937,16057,-16097,16189,-16267,
- X 16363,-16447,16529,-16619,16691,-16763,16879,-16937,
- X 17021,-17093,17183,-17257,17341,-17401,17477,-17551,
- X 17623,-17713,17791,-17891,17957,-18041,18097,-18169,
- X 18233,-18307,18379,-18451,18523,-18637,18731,-18803,
- X 18919,-19031,19121,-19211,19273,-19381,19429,-19477
- X } ;
- X#endif
- X
- Xvoid main ARGS((int,char **));
- Xvoid Usage ARGS((void));
- Xvoid comprexx ARGS((char **));
- Xvoid compdir ARGS((char *));
- Xvoid compress ARGS((int,int));
- Xvoid decompress ARGS((int,int));
- Xchar *rindex ARGS((char *,int));
- Xvoid read_error ARGS((void));
- Xvoid write_error ARGS((void));
- Xvoid abort_compress ARGS((void));
- Xvoid prratio ARGS((FILE *,long,long));
- Xvoid about ARGS((void));
- X
- X/*****************************************************************
- X * TAG( main )
- X *
- X * Algorithm from "A Technique for High Performance Data Compression",
- X * Terry A. Welch, IEEE Computer Vol 17, No 6 (June 1984), pp 8-19.
- X *
- X * Usage: compress [-dfvc] [-b bits] [file ...]
- X * Inputs:
- X * -d: If given, decompression is done instead.
- X *
- X * -c: Write output on stdout, don't remove original.
- X *
- X * -b: Parameter limits the max number of bits/code.
- X *
- X * -f: Forces output file to be generated, even if one already
- X * exists, and even if no space is saved by compressing.
- X * If -f is not used, the user will be prompted if stdin is
- X * a tty, otherwise, the output file will not be overwritten.
- X *
- X * -v: Write compression statistics
- X *
- X * -r: Recursive. If a filename is a directory, descend
- X * into it and compress everything in it.
- X *
- X * file ...:
- X * Files to be compressed. If none specified, stdin is used.
- X * Outputs:
- X * file.Z: Compressed form of file with same mode, owner, and utimes
- X * or stdout (if stdin used as input)
- X *
- X * Assumptions:
- X * When filenames are given, replaces with the compressed version
- X * (.Z suffix) only if the file decreases in size.
- X *
- X * Algorithm:
- X * Modified Lempel-Ziv method (LZW). Basically finds common
- X * substrings and replaces them with a variable size code. This is
- X * deterministic, and can be done on the fly. Thus, the decompression
- X * procedure needs no input table, but tracks the way the table was built.
- X */
- Xvoid
- Xmain(argc, argv)
- X REG1 int argc;
- X REG2 char *argv[];
- X {
- X REG3 char **filelist;
- X REG4 char **fileptr;
- X
- X if (fgnd_flag = (signal(SIGINT, SIG_IGN) != SIG_IGN))
- X signal(SIGINT, (SIG_TYPE)abort_compress);
- X
- X signal(SIGTERM, (SIG_TYPE)abort_compress);
- X#ifndef DOS
- X signal(SIGHUP, (SIG_TYPE)abort_compress);
- X#endif
- X
- X#ifdef COMPATIBLE
- X nomagic = 1; /* Original didn't have a magic number */
- X#endif
- X
- X filelist = fileptr = (char **)malloc(argc*sizeof(char *));
- X *filelist = NULL;
- X
- X if((progname = rindex(argv[0], '/')) != 0)
- X progname++;
- X else
- X progname = argv[0];
- X
- X if (strcmp(progname, "uncompress") == 0)
- X do_decomp = 1;
- X else
- X if (strcmp(progname, "zcat") == 0)
- X do_decomp = zcat_flg = 1;
- X
- X /* Argument Processing
- X * All flags are optional.
- X * -V => print Version; debug verbose
- X * -d => do_decomp
- X * -v => unquiet
- X * -f => force overwrite of output file
- X * -n => no header: useful to uncompress old files
- X * -b maxbits => maxbits. If -b is specified, then maxbits MUST be given also.
- X * -c => cat all output to stdout
- X * -C => generate output compatible with compress 2.0.
- X * -r => recursively compress directories
- X * if a string is left, must be an input filename.
- X */
- X
- X for (argc--, argv++; argc > 0; argc--, argv++)
- X {
- X if (**argv == '-')
- X {/* A flag argument */
- X while (*++(*argv))
- X {/* Process all flags in this arg */
- X switch (**argv)
- X {
- X case 'V':
- X about();
- X break;
- X
- X case 's':
- X silent = 1;
- X quiet = 1;
- X break;
- X
- X case 'v':
- X silent = 0;
- X quiet = 0;
- X break;
- X
- X case 'd':
- X do_decomp = 1;
- X break;
- X
- X case 'f':
- X case 'F':
- X force = 1;
- X break;
- X
- X case 'n':
- X nomagic = 1;
- X break;
- X
- X case 'C':
- X block_mode = 0;
- X break;
- X
- X case 'b':
- X if (!ARGVAL())
- X {
- X fprintf(stderr, "Missing maxbits\n");
- X Usage();
- X }
- X
- X maxbits = atoi(*argv);
- X goto nextarg;
- X
- X case 'c':
- X do_decomp = zcat_flg = 1;
- X break;
- X
- X case 'q':
- X quiet = 1;
- X break;
- X case 'r':
- X case 'R':
- X#ifdef RECURSIVE
- X recursive = 1;
- X#else
- X fprintf(stderr, "%s -r not availble (du to missing directory functions)\n", **argv);
- X#endif
- X break;
- X
- X default:
- X fprintf(stderr, "Unknown flag: '%c'; ", **argv);
- X Usage();
- X }
- X }
- X }
- X else
- X {
- X *fileptr++ = *argv; /* Build input file list */
- X *fileptr = NULL;
- X }
- X
- Xnextarg: continue;
- X }
- X
- X if (maxbits < INIT_BITS) maxbits = INIT_BITS;
- X if (maxbits > BITS) maxbits = BITS;
- X
- X if (*filelist != NULL)
- X {
- X for (fileptr = filelist; *fileptr; fileptr++)
- X comprexx(fileptr);
- X }
- X else
- X {/* Standard input */
- X ifname = "";
- X exit_code = 0;
- X remove_ofname = 0;
- X
- X if (do_decomp == 0)
- X {
- X compress(0, 1);
- X
- X if (zcat_flg == 0 && !quiet)
- X {
- X fprintf(stderr, "Compression: ");
- X prratio(stderr, bytes_in-bytes_out, bytes_in);
- X fprintf(stderr, "\n");
- X }
- X
- X if (bytes_out >= bytes_in && !(force))
- X exit_code = 2;
- X }
- X else
- X decompress(0, 1);
- X }
- X
- X exit((exit_code== -1) ? 1:exit_code);
- X }
- X
- Xvoid
- XUsage()
- X {
- X fprintf(stderr, "\
- XUsage: %s [-dfvcVr] [-b maxbits] [file ...]\n\
- X -d If given, decompression is done instead.\n\
- X -c Write output on stdout, don't remove original.\n\
- X -b Parameter limits the max number of bits/code.\n", progname);
- X fprintf(stderr, "\
- X -f Forces output file to be generated, even if one already.\n\
- X exists, and even if no space is saved by compressing.\n\
- X If -f is not used, the user will be prompted if stdin is.\n\
- X a tty, otherwise, the output file will not be overwritten.\n\
- X -v Write compression statistics.\n\
- X -V Output vesion and compile options.\n\
- X -r Recursive. If a filename is a directory, descend\n\
- X into it and compress everything in it.\n");
- X
- X exit(1);
- X }
- X
- Xvoid
- Xcomprexx(fileptr)
- X char **fileptr;
- X {
- X int fdin;
- X int fdout;
- X char tempname[MAXPATHLEN];
- X
- X strcpy(tempname,*fileptr);
- X errno = 0;
- X
- X#ifdef LSTAT
- X if (lstat(tempname,&infstat) == -1)
- X#else
- X if (stat(tempname,&infstat) == -1)
- X#endif
- X {
- X if (do_decomp)
- X {
- X switch (errno)
- X {
- X case ENOENT: /* file doesn't exist */
- X /*
- X ** if the given name doesn't end with .Z, try appending one
- X ** This is obviously the wrong thing to do if it's a
- X ** directory, but it shouldn't do any harm.
- X */
- X if (strcmp(tempname + strlen(tempname) - 2, ".Z") != 0)
- X {
- X strcat(tempname,".Z");
- X errno = 0;
- X#ifdef LSTAT
- X if (lstat(tempname,&infstat) == -1)
- X#else
- X if (stat(tempname,&infstat) == -1)
- X#endif
- X {
- X perror(tempname);
- X exit_code = 1;
- X return;
- X }
- X
- X if ((infstat.st_mode & S_IFMT) != S_IFREG)
- X {
- X fprintf(stderr, "%s: Not a regular file.\n", tempname);
- X exit_code = 1;
- X return ;
- X }
- X }
- X else
- X {
- X perror(tempname);
- X exit_code = 1;
- X return;
- X }
- X
- X break;
- X
- X default:
- X perror(tempname);
- X exit_code = 1;
- X return;
- X }
- X }
- X else
- X {
- X perror(tempname);
- X exit_code = 1;
- X return;
- X }
- X }
- X
- X switch (infstat.st_mode & S_IFMT)
- X {
- X case S_IFDIR: /* directory */
- X#ifdef RECURSIVE
- X if (recursive)
- X compdir(tempname);
- X else
- X#endif
- X if (!quiet)
- X fprintf(stderr,"%s is a directory -- ignored\n", tempname);
- X break;
- X
- X case S_IFREG: /* regular file */
- X if (do_decomp != 0)
- X {/* DECOMPRESSION */
- X if (!zcat_flg)
- X {
- X if (strcmp(tempname + strlen(tempname) - 2, ".Z") != 0)
- X {
- X if (!quiet)
- X fprintf(stderr,"%s - no .Z suffix\n",tempname);
- X
- X return;
- X }
- X }
- X
- X strcpy(ofname, tempname);
- X
- X /* Strip of .Z suffix */
- X
- X if (strcmp(tempname + strlen(tempname) - 2, ".Z") == 0)
- X ofname[strlen(tempname) - 2] = '\0';
- X }
- X else
- X {/* COMPRESSION */
- X if (strcmp(tempname + strlen(tempname) - 2, ".Z") == 0)
- X {
- X fprintf(stderr, "%s: already has .Z suffix -- no change\n", tempname);
- X return;
- X }
- X
- X if (infstat.st_nlink > 1 && (!force))
- X {
- X fprintf(stderr, "%s has %d other links: unchanged\n",
- X tempname, infstat.st_nlink - 1);
- X exit_code = 1;
- X return;
- X }
- X
- X strcpy(ofname, tempname);
- X strcat(ofname, ".Z");
- X }
- X
- X if ((fdin = open(ifname = tempname, O_RDONLY|O_BINARY)) == -1)
- X {
- X perror(tempname);
- X exit_code = 1;
- X return;
- X }
- X
- X if (zcat_flg == 0)
- X {
- X int c;
- X int s;
- X struct stat statbuf;
- X struct stat statbuf2;
- X
- X if (stat(ofname, &statbuf) == 0)
- X {
- X if ((s = strlen(ofname)) > 8)
- X {
- X c = ofname[s-1];
- X ofname[s-1] = '\0';
- X
- X statbuf2 = statbuf;
- X
- X if (!stat(ofname, &statbuf2) &&
- X statbuf.st_mode == statbuf2.st_mode &&
- X statbuf.st_ino == statbuf2.st_ino &&
- X statbuf.st_dev == statbuf2.st_dev &&
- X statbuf.st_uid == statbuf2.st_uid &&
- X statbuf.st_gid == statbuf2.st_gid &&
- X statbuf.st_size == statbuf2.st_size &&
- X statbuf.st_atime == statbuf2.st_atime &&
- X statbuf.st_mtime == statbuf2.st_mtime &&
- X statbuf.st_ctime == statbuf2.st_ctime)
- X {
- X fprintf(stderr, "%s: filename too long to tack on .Z\n", tempname);
- X exit_code = 1;
- X return;
- X }
- X
- X ofname[s-1] = (char)c;
- X }
- X
- X if (!force)
- X {
- X inbuf[0] = 'n';
- X
- X fprintf(stderr, "%s already exists.\n", ofname);
- X
- X if (fgnd_flag && isatty(0))
- X {
- X fprintf(stderr, "Do you wish to overwrite %s (y or n)? ", ofname);
- X fflush(stderr);
- X
- X if (read(0, inbuf, 1) > 0)
- X {
- X if (inbuf[0] != '\n')
- X {
- X do
- X {
- X if (read(0, inbuf+1, 1) <= 0)
- X {
- X perror("stdin");
- X break;
- X }
- X }
- X while (inbuf[1] != '\n');
- X }
- X }
- X else
- X perror("stdin");
- X }
- X
- X if (inbuf[0] != 'y')
- X {
- X fprintf(stderr, "%s not overwritten\n", ofname);
- X exit_code = 1;
- X return;
- X }
- X }
- X
- X if (unlink(ofname))
- X {
- X fprintf(stderr, "Can't remove old output file\n");
- X perror(ofname);
- X exit_code = 1;
- X return ;
- X }
- X }
- X
- X if ((fdout = open(ofname, O_WRONLY|O_CREAT|O_EXCL|O_BINARY,0600)) == -1)
- X {
- X perror(tempname);
- X return;
- X }
- X
- X if ((s = strlen(ofname)) > 8)
- X {
- X if (fstat(fdout, &statbuf))
- X {
- X fprintf(stderr, "Can't get status op output file\n");
- X perror(ofname);
- X exit_code = 1;
- X return ;
- X }
- X
- X c = ofname[s-1];
- X ofname[s-1] = '\0';
- X statbuf2 = statbuf;
- X
- X if (!stat(ofname, &statbuf2) &&
- X statbuf.st_mode == statbuf2.st_mode &&
- X statbuf.st_ino == statbuf2.st_ino &&
- X statbuf.st_dev == statbuf2.st_dev &&
- X statbuf.st_uid == statbuf2.st_uid &&
- X statbuf.st_gid == statbuf2.st_gid &&
- X statbuf.st_size == statbuf2.st_size &&
- X statbuf.st_atime == statbuf2.st_atime &&
- X statbuf.st_mtime == statbuf2.st_mtime &&
- X statbuf.st_ctime == statbuf2.st_ctime)
- X {
- X fprintf(stderr, "%s: filename too long to tack on .Z\n", tempname);
- X
- X if (unlink(ofname))
- X {
- X fprintf(stderr, "can't remove bad output file\n");
- X perror(ofname);
- X }
- X exit_code = 1;
- X return;
- X }
- X
- X ofname[s-1] = (char)c;
- X }
- X
- X if(!quiet)
- X fprintf(stderr, "%s: ", tempname);
- X
- X remove_ofname = 1;
- X }
- X else
- X {
- X fdout = 1;
- X ofname[0] = '\0';
- X remove_ofname = 0;
- X }
- X
- X if (do_decomp == 0)
- X compress(fdin, fdout);
- X else
- X decompress(fdin, fdout);
- X
- X close(fdin);
- X
- X if (fdout != 1 && close(fdout))
- X write_error();
- X
- X if (bytes_in == 0)
- X {
- X if (remove_ofname)
- X {
- X if (unlink(ofname)) /* Remove input file */
- X {
- X fprintf(stderr, "\nunlink error (ignored) ");
- X perror(ofname);
- X exit_code = 1;
- X }
- X
- X remove_ofname = 0;
- X }
- X }
- X else
- X if (zcat_flg == 0)
- X {
- X struct utimbuf timep;
- X
- X if (!do_decomp && bytes_out >= bytes_in && (!force))
- X {/* No compression: remove file.Z */
- X if(!quiet)
- X fprintf(stderr, "No compression -- %s unchanged\n", ifname);
- X
- X if (unlink(ofname))
- X {
- X fprintf(stderr, "unlink error (ignored) ");
- X perror(ofname);
- X }
- X
- X remove_ofname = 0;
- X exit_code = 2;
- X }
- X else
- X {/* ***** Successful Compression ***** */
- X if(!quiet)
- X {
- X fprintf(stderr, " -- replaced with %s",ofname);
- X
- X if (!do_decomp)
- X {
- X fprintf(stderr, " Compression: ");
- X prratio(stderr, bytes_in-bytes_out, bytes_in);
- X }
- X
- X fprintf(stderr, "\n");
- X }
- X
- X if (chmod(ofname, infstat.st_mode & 07777)) /* Copy modes */
- X {
- X fprintf(stderr, "\nchmod error (ignored) ");
- X perror(ofname);
- X exit_code = 1;
- X }
- X#ifndef DOS
- X chown(ofname, infstat.st_uid, infstat.st_gid); /* Copy ownership */
- X#endif
- X
- X timep.actime = infstat.st_atime;
- X timep.modtime = infstat.st_mtime;
- X
- X if (utime(ofname, &timep))
- X {
- X fprintf(stderr, "\nutime error (ignored) ");
- X perror(ofname);
- X exit_code = 1;
- X }
- X
- X remove_ofname = 0;
- X
- X if (unlink(ifname)) /* Remove input file */
- X {
- X fprintf(stderr, "\nunlink error (ignored) ");
- X perror(ifname);
- X exit_code = 1;
- X }
- X }
- X }
- X
- X if (exit_code == -1)
- X exit_code = 0;
- X
- X break;
- X
- X default:
- X fprintf(stderr,"%s is not a directory or a regular file - ignored\n",
- X tempname);
- X break;
- X }
- X }
- X
- X#ifdef RECURSIVE
- Xvoid
- Xcompdir(dir)
- X REG3 char *dir;
- X {
- X#ifndef DIRENT
- X REG1 struct direct *dp;
- X#else
- X REG1 struct dirent *dp;
- X#endif
- X REG2 DIR *dirp;
- X char nbuf[MAXPATHLEN];
- X char *nptr = nbuf;
- X
- X dirp = opendir(dir);
- X
- X if (dirp == NULL)
- X {
- X printf("%s unreadable\n", dir); /* not stderr! */
- X return ;
- X }
- X /*
- X ** WARNING: the following algorithm will occasionally cause
- X ** compress to produce error warnings of the form "<filename>.Z
- X ** already has .Z suffix - ignored". This occurs when the
- X ** .Z output file is inserted into the directory below
- X ** readdir's current pointer.
- X ** These warnings are harmless but annoying. The alternative
- X ** to allowing this would be to store the entire directory
- X ** list in memory, then compress the entries in the stored
- X ** list. Given the depth-first recursive algorithm used here,
- X ** this could use up a tremendous amount of memory. I don't
- X ** think it's worth it. -- Dave Mack
- X */
- X
- X while (dp = readdir(dirp))
- X {
- X if (dp->d_ino == 0)
- X continue;
- X
- X if (strcmp(dp->d_name,".") == 0 || strcmp(dp->d_name,"..") == 0)
- X continue;
- X
- X if ((strlen(dir)+strlen(dp->d_name)+1) < (MAXPATHLEN - 1))
- X {
- X strcpy(nbuf,dir);
- X strcat(nbuf,"/");
- X strcat(nbuf,dp->d_name);
- X comprexx(&nptr);
- X }
- X else
- X fprintf(stderr,"Pathname too long: %s/%s\n", dir, dp->d_name);
- X }
- X
- X closedir(dirp);
- X
- X return;
- X }
- X#endif
- X/*
- X * compress fdin to fdout
- X *
- X * Algorithm: use open addressing double hashing (no chaining) on the
- X * prefix code / next character combination. We do a variant of Knuth's
- X * algorithm D (vol. 3, sec. 6.4) along with G. Knott's relatively-prime
- X * secondary probe. Here, the modular division first probe is gives way
- X * to a faster exclusive-or manipulation. Also do block compression with
- X * an adaptive reset, whereby the code table is cleared when the compression
- X * ratio decreases, but after the table fills. The variable-length output
- X * codes are re-sized at this point, and a special CLEAR code is generated
- X * for the decompressor. Late addition: construct the table according to
- X * file size for noticeable speed improvement on small files. Please direct
- X * questions about this implementation to ames!jaw.
- X */
- Xvoid
- Xcompress(fdin, fdout)
- X int fdin;
- X int fdout;
- X {
- X REG2 long hp;
- X REG3 int rpos;
- X#if REGISTERS >= 5
- X REG5 long fc;
- X#endif
- X REG6 int outbits;
- X REG7 int rlop;
- X REG8 int rsize;
- X REG9 int stcode;
- X REG10 code_int free_ent;
- X REG11 int boff;
- X REG12 int n_bits;
- X REG13 int ratio;
- X REG14 long checkpoint;
- X REG15 code_int extcode;
- X union
- X {
- X long code;
- X struct
- X {
- X char_type c;
- X unsigned short ent;
- X } e;
- X } fcode;
- X
- X ratio = 0;
- X checkpoint = CHECK_GAP;
- X extcode = MAXCODE(n_bits = INIT_BITS)+1;
- X stcode = 1;
- X free_ent = FIRST;
- X
- X memset(outbuf, 0, sizeof(outbuf));
- X bytes_out = 0; bytes_in = 0;
- X outbuf[0] = MAGIC_1;
- X outbuf[1] = MAGIC_2;
- X outbuf[2] = (char)(maxbits | block_mode);
- X boff = outbits = (3<<3);
- X fcode.code = 0;
- X
- X clear_htab();
- X
- X while ((rsize = read(fdin, inbuf, IBUFSIZ)) > 0)
- X {
- X if (bytes_in == 0)
- X {
- X fcode.e.ent = inbuf[0];
- X rpos = 1;
- X }
- X else
- X rpos = 0;
- X
- X rlop = 0;
- X
- X do
- X {
- X if (free_ent >= extcode && fcode.e.ent < FIRST)
- X {
- X if (n_bits < maxbits)
- X {
- X boff = outbits = (outbits-1)+((n_bits<<3)-
- X ((outbits-boff-1+(n_bits<<3))%(n_bits<<3)));
- X if (++n_bits < maxbits)
- X extcode = MAXCODE(n_bits)+1;
- X else
- X extcode = MAXCODE(n_bits);
- X }
- X else
- X {
- X extcode = MAXCODE(16)+OBUFSIZ;
- X stcode = 0;
- X }
- X }
- X
- X if (!stcode && bytes_in >= checkpoint && fcode.e.ent < FIRST)
- X {
- X REG1 long int rat;
- X
- X checkpoint = bytes_in + CHECK_GAP;
- X
- X if (bytes_in > 0x007fffff)
- X { /* shift will overflow */
- X rat = (bytes_out+(outbits>>3)) >> 8;
- X
- X if (rat == 0) /* Don't divide by zero */
- X rat = 0x7fffffff;
- X else
- X rat = bytes_in / rat;
- X }
- X else
- X rat = (bytes_in << 8) / (bytes_out+(outbits>>3)); /* 8 fractional bits */
- X if (rat >= ratio)
- X ratio = (int)rat;
- X else
- X {
- X ratio = 0;
- X clear_htab();
- X output(outbuf,outbits,CLEAR,n_bits);
- X boff = outbits = (outbits-1)+((n_bits<<3)-
- X ((outbits-boff-1+(n_bits<<3))%(n_bits<<3)));
- X extcode = MAXCODE(n_bits = INIT_BITS)+1;
- X free_ent = FIRST;
- X stcode = 1;
- X }
- X }
- X
- X if (outbits >= (OBUFSIZ<<3))
- X {
- X if (write(fdout, outbuf, OBUFSIZ) != OBUFSIZ)
- X write_error();
- X
- X outbits -= (OBUFSIZ<<3);
- X boff = -(((OBUFSIZ<<3)-boff)%(n_bits<<3));
- X bytes_out += OBUFSIZ;
- X
- X memcpy(outbuf, outbuf+OBUFSIZ, (outbits>>3)+1);
- X memset(outbuf+(outbits>>3)+1, '\0', OBUFSIZ);
- X }
- X
- X {
- X REG1 int i;
- X
- X i = rsize-rlop;
- X
- X if ((code_int)i > extcode-free_ent) i = (int)(extcode-free_ent);
- X if (i > ((sizeof(outbuf) - 32)*8 - outbits)/n_bits)
- X i = ((sizeof(outbuf) - 32)*8 - outbits)/n_bits;
- X
- X if (!stcode && (long)i > checkpoint-bytes_in)
- X i = (int)(checkpoint-bytes_in);
- X
- X rlop += i;
- X bytes_in += i;
- X }
- X
- X goto next;
- Xhfound: fcode.e.ent = codetabof(hp);
- Xnext: if (rpos >= rlop)
- X goto endlop;
- Xnext2: fcode.e.c = inbuf[rpos++];
- X#ifndef FAST
- X {
- X REG1 code_int i;
- X#if REGISTERS >= 5
- X fc = fcode.code;
- X#else
- X# define fc fcode.code
- X#endif
- X hp = (((long)(fcode.e.c)) << (BITS-8)) ^ (long)(fcode.e.ent);
- X
- X if ((i = htabof(hp)) == fc)
- X goto hfound;
- X
- X if (i != -1)
- X {
- X REG4 long disp;
- X
- X disp = (HSIZE - hp)-1; /* secondary hash (after G. Knott) */
- X
- X do
- X {
- X if ((hp -= disp) < 0) hp += HSIZE;
- X
- X if ((i = htabof(hp)) == fc)
- X goto hfound;
- X }
- X while (i != -1);
- X }
- X }
- X#else
- X {
- X REG1 long i;
- X REG4 long p;
- X#if REGISTERS >= 5
- X fc = fcode.code;
- X#else
- X# define fc fcode.code
- X#endif
- X hp = ((((long)(fcode.e.c)) << (HBITS-8)) ^ (long)(fcode.e.ent));
- X
- X if ((i = htabof(hp)) == fc) goto hfound;
- X if (i == -1) goto out;
- X
- X p = primetab[fcode.e.c];
- Xlookup: hp = (hp+p)&HMASK;
- X if ((i = htabof(hp)) == fc) goto hfound;
- X if (i == -1) goto out;
- X hp = (hp+p)&HMASK;
- X if ((i = htabof(hp)) == fc) goto hfound;
- X if (i == -1) goto out;
- X hp = (hp+p)&HMASK;
- X if ((i = htabof(hp)) == fc) goto hfound;
- X if (i == -1) goto out;
- X goto lookup;
- X }
- Xout: ;
- X#endif
- X output(outbuf,outbits,fcode.e.ent,n_bits);
- X
- X {
- X#if REGISTERS < 5
- X# undef fc
- X REG1 long fc;
- X fc = fcode.code;
- X#endif
- X fcode.e.ent = fcode.e.c;
- X
- X
- X if (stcode)
- X {
- X codetabof(hp) = (unsigned short)free_ent++;
- X htabof(hp) = fc;
- X }
- X }
- X
- X goto next;
- X
- Xendlop: if (fcode.e.ent >= FIRST && rpos < rsize)
- X goto next2;
- X
- X if (rpos > rlop)
- X {
- X bytes_in += rpos-rlop;
- X rlop = rpos;
- X }
- X }
- X while (rlop < rsize);
- X }
- X
- X if (rsize < 0)
- X read_error();
- X
- X if (bytes_in > 0)
- X output(outbuf,outbits,fcode.e.ent,n_bits);
- X
- X if (write(fdout, outbuf, (outbits+7)>>3) != (outbits+7)>>3)
- X write_error();
- X
- X bytes_out += (outbits+7)>>3;
- X
- X return;
- X }
- X
- X/*
- X * Decompress stdin to stdout. This routine adapts to the codes in the
- X * file building the "string" table on-the-fly; requiring no table to
- X * be stored in the compressed file. The tables used herein are shared
- X * with those of the compress() routine. See the definitions above.
- X */
- X
- Xvoid
- Xdecompress(fdin, fdout)
- X int fdin;
- X int fdout;
- X {
- X REG2 char_type *stackp;
- X REG3 code_int code;
- X REG4 int finchar;
- X REG5 code_int oldcode;
- X REG6 code_int incode;
- X REG7 int inbits;
- X REG8 int posbits;
- X REG9 int outpos;
- X REG10 int insize;
- X REG11 int bitmask;
- X REG12 code_int free_ent;
- X REG13 code_int maxcode;
- X REG14 code_int maxmaxcode;
- X REG15 int n_bits;
- X REG16 int rsize;
- X
- X bytes_in = 0;
- X bytes_out = 0;
- X insize = 0;
- X
- X while (insize < 3 && (rsize = read(fdin, inbuf+insize, IBUFSIZ)) > 0)
- X insize += rsize;
- X
- X if (insize < 3 || inbuf[0] != MAGIC_1 || inbuf[1] != MAGIC_2)
- X {
- X if (rsize < 0)
- X read_error();
- X
- X if (insize > 0)
- X {
- X fprintf(stderr, "%s: not in compressed format\n",
- X (ifname[0] != '\0'? ifname : "stdin"));
- X exit_code = 1;
- X }
- X
- X return ;
- X }
- X
- X maxbits = inbuf[2] & BIT_MASK;
- X block_mode = inbuf[2] & BLOCK_MODE;
- X maxmaxcode = MAXCODE(maxbits);
- X
- X if (maxbits > BITS)
- X {
- X fprintf(stderr,
- X "%s: compressed with %d bits, can only handle %d bits\n",
- X (*ifname != '\0' ? ifname : "stdin"), maxbits, BITS);
- X exit_code = 4;
- X return;
- X }
- X
- X bytes_in = insize;
- X maxcode = MAXCODE(n_bits = INIT_BITS)-1;
- X bitmask = (1<<n_bits)-1;
- X oldcode = -1;
- X finchar = 0;
- X outpos = 0;
- X posbits = 3<<3;
- X
- X free_ent = ((block_mode) ? FIRST : 256);
- X
- X clear_tab_prefixof(); /* As above, initialize the first
- X 256 entries in the table. */
- X
- X for (code = 255 ; code >= 0 ; --code)
- X tab_suffixof(code) = (char_type)code;
- X
- X do
- X {
- Xresetbuf: ;
- X {
- X REG1 int i;
- X int e;
- X int o;
- X
- X e = insize-(o = (posbits>>3));
- X
- X for (i = 0 ; i < e ; ++i)
- X inbuf[i] = inbuf[i+o];
- X
- X insize = e;
- X posbits = 0;
- X }
- X
- X if (insize < sizeof(inbuf)-IBUFSIZ)
- X {
- X if ((rsize = read(fdin, inbuf+insize, IBUFSIZ)) < 0)
- X read_error();
- X
- X insize += rsize;
- X }
- X
- X inbits = ((rsize > 0) ? (insize - insize%n_bits)<<3 :
- X (insize<<3)-(n_bits-1));
- X
- X while (inbits > posbits)
- X {
- X if (free_ent > maxcode)
- X {
- X posbits = ((posbits-1) + ((n_bits<<3) -
- X (posbits-1+(n_bits<<3))%(n_bits<<3)));
- X
- X ++n_bits;
- X if (n_bits == maxbits)
- X maxcode = maxmaxcode;
- X else
- X maxcode = MAXCODE(n_bits)-1;
- X
- X bitmask = (1<<n_bits)-1;
- X goto resetbuf;
- X }
- X
- X input(inbuf,posbits,code,n_bits,bitmask);
- X
- X if (oldcode == -1)
- X {
- X outbuf[outpos++] = (char_type)(finchar = (int)(oldcode = code));
- X continue;
- X }
- X
- X if (code == CLEAR && block_mode)
- X {
- X clear_tab_prefixof();
- X free_ent = FIRST - 1;
- X posbits = ((posbits-1) + ((n_bits<<3) -
- X (posbits-1+(n_bits<<3))%(n_bits<<3)));
- X maxcode = MAXCODE(n_bits = INIT_BITS)-1;
- X bitmask = (1<<n_bits)-1;
- X goto resetbuf;
- X }
- X
- X incode = code;
- X stackp = de_stack;
- X
- X if (code >= free_ent) /* Special case for KwKwK string. */
- X {
- X if (code > free_ent)
- X {
- X REG1 char_type *p;
- X
- X posbits -= n_bits;
- X p = &inbuf[posbits>>3];
- X
- X fprintf(stderr, "insize:%d posbits:%d inbuf:%02X %02X %02X %02X %02X (%d)\n", insize, posbits,
- X p[-1],p[0],p[1],p[2],p[3], (posbits&07));
- X fprintf(stderr, "uncompress: corrupt input\n");
- X abort_compress();
- X }
- X
- X *--stackp = (char_type)finchar;
- X code = oldcode;
- X }
- X
- X while ((cmp_code_int)code >= (cmp_code_int)256)
- X { /* Generate output characters in reverse order */
- X *--stackp = tab_suffixof(code);
- X code = tab_prefixof(code);
- X }
- X
- X *--stackp = (char_type)(finchar = tab_suffixof(code));
- X
- X /* And put them out in forward order */
- X
- X {
- X REG1 int i;
- X
- X if (outpos+(i = (de_stack-stackp)) >= OBUFSIZ)
- X {
- X do
- X {
- X if (i > OBUFSIZ-outpos) i = OBUFSIZ-outpos;
- X
- X if (i > 0)
- X {
- X memcpy(outbuf+outpos, stackp, i);
- X outpos += i;
- X }
- X
- X if (outpos >= OBUFSIZ)
- X {
- X if (write(fdout, outbuf, outpos) != outpos)
- X write_error();
- X
- X outpos = 0;
- X }
- X stackp+= i;
- X }
- X while ((i = (de_stack-stackp)) > 0);
- X }
- X else
- X {
- X memcpy(outbuf+outpos, stackp, i);
- X outpos += i;
- X }
- X }
- X
- X if ((code = free_ent) < maxmaxcode) /* Generate the new entry. */
- X {
- X tab_prefixof(code) = (unsigned short)oldcode;
- X tab_suffixof(code) = (char_type)finchar;
- X free_ent = code+1;
- X }
- X
- X oldcode = incode; /* Remember previous code. */
- X }
- X
- X bytes_in += rsize;
- X }
- X while (rsize > 0);
- X
- X if (outpos > 0 && write(fdout, outbuf, outpos) != outpos)
- X write_error();
- X }
- X
- Xchar *
- Xrindex(s, c) /* For those who don't have it in libc.a */
- X REG1 char *s;
- X REG2 int c;
- X {
- X char *p;
- X
- X for (p = NULL; *s; s++)
- X if (*s == (char)c)
- X p = s;
- X
- X return(p);
- X }
- X
- Xvoid
- Xread_error()
- X {
- X fprintf(stderr, "\nread error on");
- X perror((ifname[0] != '\0') ? ifname : "stdin");
- X abort_compress();
- X }
- X
- Xvoid
- Xwrite_error()
- X {
- X fprintf(stderr, "\nwrite error on");
- X perror((ofname[0] != '\0') ? ofname : "stdout");
- X abort_compress();
- X }
- X
- Xvoid
- Xabort_compress()
- X {
- X if (remove_ofname)
- X unlink(ofname);
- X
- X exit(1);
- X }
- X
- Xvoid
- Xprratio(stream, num, den)
- X FILE *stream;
- X long int num;
- X long int den;
- X {
- X REG1 int q; /* Doesn't need to be long */
- X
- X if (den > 0)
- X {
- X if (num > 214748L)
- X q = (int)(num/(den/10000L)); /* 2147483647/10000 */
- X else
- X q = (int)(10000L*num/den); /* Long calculations, though */
- X }
- X else
- X q = 10000;
- X
- X if (q < 0)
- X {
- X putc('-', stream);
- X q = -q;
- X }
- X
- X fprintf(stream, "%d.%02d%%", q / 100, q % 100);
- X }
- X
- Xvoid
- Xabout()
- X {
- X fprintf(stderr, "Compress version: %s, compiled: %s\n", version_id, COMPILE_DATE);
- X fprintf(stderr, "Compile options:\n ");
- X#if BYTEORDER == 4321 && NOALLIGN == 1
- X fprintf(stderr, "USE_BYTEORDER, ");
- X#endif
- X#ifdef FAST
- X fprintf(stderr, "FAST, ");
- X#endif
- X#ifdef vax
- X fprintf(stderr, "vax, ");
- X#endif
- X#ifdef DIRENT
- X fprintf(stderr,"DIRENT, ");
- X#endif
- X#ifdef SYSDIR
- X fprintf(stderr,"SYSDIR, ");
- X#endif
- X#ifdef NO_UCHAR
- X fprintf(stderr, "NO_UCHAR, ");
- X#endif
- X#ifdef SIGNED_COMPARE_SLOW
- X fprintf(stderr, "SIGNED_COMPARE_SLOW, ");
- X#endif
- X#ifdef MAXSEG_64K
- X fprintf(stderr, "MAXSEG_64K, ");
- X#endif
- X#ifdef DOS
- X fprintf(stderr, "DOS, ");
- X#endif
- X#ifdef DEBUG
- X fprintf(stderr, "DEBUG, ");
- X#endif
- X#ifdef LSTAT
- X fprintf(stderr, "LSTAT, ");
- X#endif
- X fprintf(stderr, "\n REGISTERS=%d IBUFSIZ=%d, OBUFSIZ=%d, BITS=%d\n",
- X REGISTERS, IBUFSIZ, OBUFSIZ, BITS);
- X
- X fprintf(stderr, "\n\
- XAuthor version 4.2 (Speed improvement & source cleanup):\n\
- X Peter Jannesen (peter@ncs.nl)\n\
- X\n\
- XAuthor version 4.1 (Added recursive directory compress):\n\
- X Dave Mack (csu@alembic.acs.com)\n\
- X\n\
- XAuthors version 4.0 (World release in 1985):\n\
- X Spencer W. Thomas, Jim McKie, Steve Davies,\n\
- X Ken Turkowski, James A. Woods, Joe Orost\n");
- X
- X exit(0);
- X }
- END_OF_FILE
- if test 48463 -ne `wc -c <'compress42.c'`; then
- echo shar: \"'compress42.c'\" unpacked with wrong size!
- fi
- # end of 'compress42.c'
- fi
- echo shar: End of archive 2 \(of 2\).
- cp /dev/null ark2isdone
- MISSING=""
- for I in 1 2 ; do
- if test ! -f ark${I}isdone ; then
- MISSING="${MISSING} ${I}"
- fi
- done
- if test "${MISSING}" = "" ; then
- echo You have unpacked both archives.
- rm -f ark[1-9]isdone
- else
- echo You still need to unpack the following archives:
- echo " " ${MISSING}
- fi
- ## End of shell archive.
- exit 0
- exit 0 # Just in case...
-