1.\" Copyright (c) 1986, 1990 The Regents of the University of California. 2.\" All rights reserved. 3.\" 4.\" This code is derived from software contributed to Berkeley by 5.\" James A. Woods, derived from original work by Spencer Thomas 6.\" and Joseph Orost. 7.\" 8.\" %sccs.include.redist.man% 9.\" 10.\" @(#)compress.1 6.7 (Berkeley) 06/11/90 11.\" 12.Dd 13.Dt COMPRESS 1 14.Os BSD 4.3 15.Sh NAME 16.Nm compress , 17.Nm uncompress , 18.Nm zcat 19.Nd compress and expand data 20.Sh SYNOPSIS 21.Nm compress 22.Op Fl f 23.Op Fl v 24.Op Fl c 25.Op Fl b Ar bits 26.Ar 27.br 28.Nm uncompress 29.Op Fl f 30.Op Fl v 31.Op Fl c 32.Ar 33.br 34.Nm zcat 35.Ar 36.Sh DESCRIPTION 37.Nm Compress 38reduces the size of the named files using adaptive Lempel-Ziv coding. 39Whenever possible, 40each 41.Ar file 42is replaced by one with the extension 43.Sy \&.Z , 44while keeping the same ownership modes, access and modification times. 45If no files are specified, the standard input is compressed to the 46standard output. 47Compressed files can be restored to their original form using 48.Nm uncompress 49or 50.Nm zcat 51.Tw Ds 52.Tp Fl f 53Force compression of 54.Ar file , 55even if it does not actually shrink 56or the corresponding 57.Ar file.Z 58file already exists. 59Except when run in the background under 60.Pa /bin/sh , 61if 62.Fl f 63is not given the user is prompted as to whether an existing 64.Ar file.Z 65file should be overwritten. 66.Pp 67.Tp Fl c 68(``cat'') makes 69.Nm compress/uncompress 70write to the standard output; no files are changed. 71The nondestructive behavior of 72.Nm zcat 73is identical to that of 74.Nm uncompress 75.Fl c. 76.Tp Fl b 77Specify 78.Ar bits 79code limit (see below). 80The 81.Tp Fl v 82option causes 83the printing of the percentage reduction of each file. 84.Tp 85.Pp 86.Nm Compress 87uses the modified Lempel-Ziv algorithm popularized in 88"A Technique for High Performance Data Compression", 89Terry A. Welch, 90.Em IEEE Computer , 91vol. 17, no. 6 (June 1984), pp. 8-19. 92Common substrings in the file are first replaced by 9-bit codes 257 and up. 93When code 512 is reached, the algorithm switches to 10-bit codes and 94continues to use more bits until the 95limit specified by the 96.Fl b 97flag is reached (default 16). 98.Ar Bits 99must be between 9 and 16. The default can be changed in the source to allow 100.Nm compress 101to be run on a smaller machine. 102.Pp 103After the 104.Ar bits 105limit is attained, 106.Nm compress 107periodically checks the compression ratio. If it is increasing, 108.Nm compress 109continues to use the existing code dictionary. However, 110if the compression ratio decreases, 111.Nm compress 112discards the table of substrings and rebuilds it from scratch. This allows 113the algorithm to adapt to the next "block" of the file. 114.Pp 115Note that the 116.Fl b 117flag is omitted for 118.Ar uncompress 119since the 120.Ar bits 121parameter specified during compression 122is encoded within the output, along with 123a magic number to ensure that neither decompression of random data nor 124recompression of compressed data is attempted. 125.Pp 126.ne 8 127The amount of compression obtained depends on the size of the 128input, the number of 129.Ar bits 130per code, and the distribution of common substrings. 131Typically, text such as source code or English 132is reduced by 50\-60%. 133Compression is generally much better than that achieved by 134Huffman coding (as used in 135.Xr pack ) , 136or adaptive Huffman coding (as 137used in 138.Xr compact ) , 139and takes less time to compute. 140.Pp 141If an error occurs, exit status is 1, else 142if the last file was not compressed because it became larger, the status 143is 2; else the status is 0. 144.Sh DIAGNOSTICS 145Usage: compress 146.Op Fl fvc 147.Op Fl b Ar maxbits 148.Ar 149.Dl Invalid options were specified on the command line. 150.Pp 151Missing maxbits 152.Df I 153Maxbits must follow 154.Fl b . 155.De 156.Pp 157.Ar file : 158not in compressed format 159.Df I 160The file specified to 161.Ar uncompress 162has not been compressed. 163.De 164.Pp 165.Ar file : 166compressed with 167.Ar xx 168bits, can only handle 169.Ar yy 170bits 171.Df I 172.Ar File 173was compressed by a program that could deal with 174more 175.Ar bits 176than the compress code on this machine. 177Recompress the file with smaller 178.Ar bits . 179.De 180.Pp 181.Ar file : 182already has .Z suffix -- no change 183.Df I 184The file is assumed to be already compressed. 185Rename the file and try again. 186.De 187.Pp 188.Ar file : 189filename too long to tack on .Z 190.Df I 191The file cannot be compressed because its name is longer than 19212 characters. 193Rename and try again. 194This message does not occur on BSD systems. 195.De 196.Pp 197.Ar file 198already exists; do you wish to overwrite (y or n)? 199.Df I 200Respond "y" if you want the output file to be replaced; "n" if not. 201.De 202.Pp 203uncompress: corrupt input 204.Df I 205A SIGSEGV violation was detected which usually means that the input file is 206corrupted. 207.De 208.Pp 209Compression: 210.Em xx.xx% 211.Df I 212Percentage of the input saved by compression. 213(Relevant only for 214.Fl v . ) 215.De 216.Pp 217-- not a regular file: unchanged 218.Df I 219When the input file is not a regular file, 220(e.g. a directory), it is 221left unaltered. 222.De 223.Pp 224-- has 225.Ar xx 226other links: unchanged 227.Df I 228The input file has links; it is left unchanged. See 229.Xr ln 1 230for more information. 231.De 232.Pp 233-- file unchanged 234.Df I 235No savings is achieved by 236compression. The input remains virgin. 237.De 238.Pp 239.Sh FILES 240.Tw file.Z 241.Tp Pa file.Z 242compressed file is file.Z 243.Tp 244.Sh HISTORY 245Appeared in 4.3 BSD. 246.Sh "BUGS" 247Although compressed files are compatible between machines with large memory, 248.Cx Fl b 249.Ar 12 250.Cx 251should be used for file transfer to architectures with 252a small process data space (64KB or less, as exhibited by the DEC PDP 253series, the Intel 80286, etc.) 254.Pp 255.Nm compress 256should be more flexible about the existence of the `.Z' suffix. 257