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