1 \input texinfo @c -*-texinfo-*-
5 @settitle Gzip User's Manual
11 * gzip: (gzip). The gzip command for compressing files.
15 This file documents the GNU `gzip' command for compressing files.
17 This manual is for Gzip
18 (version @value{VERSION}, @value{UPDATED}),
19 and documents commands for compressing and decompressing data.
21 Copyright @copyright{} 1998, 1999, 2001, 2002 Free Software
24 Copyright @copyright{} 1992, 1993 Jean-loup Gailly
27 Permission is granted to copy, distribute and/or modify this document
28 under the terms of the GNU Free Documentation License, Version 1.1 or
29 any later version published by the Free Software Foundation; with no
30 Invariant Sections, with the Front-Cover texts being ``A GNU Manual,''
31 and with the Back-Cover Texts as in (a) below. A copy of the
32 license is included in the section entitled ``GNU Free Documentation
35 (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
36 this GNU Manual, like GNU software. Copies published by the Free
37 Software Foundation raise funds for GNU development.''
43 @subtitle The data compression program
44 @subtitle for Gzip Version @value{VERSION}
45 @subtitle @value{UPDATED}
46 @author by Jean-loup Gailly
49 @vskip 0pt plus 1filll
57 @top Compressing Files
63 * Overview:: Preliminary information.
64 * Sample:: Sample output from @code{gzip}.
65 * Invoking gzip:: How to run @code{gzip}.
66 * Advanced usage:: Concatenated files.
67 * Environment:: The @code{GZIP} environment variable
68 * Tapes:: Using @code{gzip} on tapes.
69 * Problems:: Reporting bugs.
70 * Copying This Manual:: How to make copies of this manual.
71 * Concept Index:: Index of concepts.
74 @node Overview, Sample, , Top
78 @code{gzip} reduces the size of the named files using Lempel-Ziv coding
79 (LZ77). Whenever possible, each file is replaced by one with the
80 extension @samp{.gz}, while keeping the same ownership modes, access and
81 modification times. (The default extension is @samp{-gz} for VMS,
82 @samp{z} for MSDOS, OS/2 FAT and Atari.) If no files are specified or
83 if a file name is "-", the standard input is compressed to the standard
84 output. @code{gzip} will only attempt to compress regular files. In
85 particular, it will ignore symbolic links.
87 If the new file name is too long for its file system, @code{gzip}
88 truncates it. @code{gzip} attempts to truncate only the parts of the
89 file name longer than 3 characters. (A part is delimited by dots.) If
90 the name consists of small parts only, the longest parts are truncated.
91 For example, if file names are limited to 14 characters, gzip.msdos.exe
92 is compressed to gzi.msd.exe.gz. Names are not truncated on systems
93 which do not have a limit on file name length.
95 By default, @code{gzip} keeps the original file name and timestamp in
96 the compressed file. These are used when decompressing the file with the
97 @samp{-N} option. This is useful when the compressed file name was
98 truncated or when the time stamp was not preserved after a file
101 Compressed files can be restored to their original form using @samp{gzip -d}
102 or @code{gunzip} or @code{zcat}. If the original name saved in the
103 compressed file is not suitable for its file system, a new name is
104 constructed from the original one to make it legal.
106 @code{gunzip} takes a list of files on its command line and replaces
107 each file whose name ends with @samp{.gz}, @samp{.z}, @samp{.Z},
108 @samp{-gz}, @samp{-z} or @samp{_z} and which begins with the correct
109 magic number with an uncompressed file without the original extension.
110 @code{gunzip} also recognizes the special extensions @samp{.tgz} and
111 @samp{.taz} as shorthands for @samp{.tar.gz} and @samp{.tar.Z}
112 respectively. When compressing, @code{gzip} uses the @samp{.tgz}
113 extension if necessary instead of truncating a file with a @samp{.tar}
116 @code{gunzip} can currently decompress files created by @code{gzip},
117 @code{zip}, @code{compress} or @code{pack}. The detection of the input
118 format is automatic. When using the first two formats, @code{gunzip}
119 checks a 32 bit CRC (cyclic redundancy check). For @code{pack},
120 @code{gunzip} checks the uncompressed length. The @code{compress} format
121 was not designed to allow consistency checks. However @code{gunzip} is
122 sometimes able to detect a bad @samp{.Z} file. If you get an error when
123 uncompressing a @samp{.Z} file, do not assume that the @samp{.Z} file is
124 correct simply because the standard @code{uncompress} does not complain.
125 This generally means that the standard @code{uncompress} does not check
126 its input, and happily generates garbage output. The SCO @samp{compress
127 -H} format (@code{lzh} compression method) does not include a CRC but
128 also allows some consistency checks.
130 Files created by @code{zip} can be uncompressed by @code{gzip} only if
131 they have a single member compressed with the 'deflation' method. This
132 feature is only intended to help conversion of @code{tar.zip} files to
133 the @code{tar.gz} format. To extract a @code{zip} file with a single
134 member, use a command like @samp{gunzip <foo.zip} or @samp{gunzip -S
135 .zip foo.zip}. To extract @code{zip} files with several
136 members, use @code{unzip} instead of @code{gunzip}.
138 @code{zcat} is identical to @samp{gunzip -c}. @code{zcat}
139 uncompresses either a list of files on the command line or its standard
140 input and writes the uncompressed data on standard output. @code{zcat}
141 will uncompress files that have the correct magic number whether they
142 have a @samp{.gz} suffix or not.
144 @code{gzip} uses the Lempel-Ziv algorithm used in @code{zip} and PKZIP.
145 The amount of compression obtained depends on the size of the input and
146 the distribution of common substrings. Typically, text such as source
147 code or English is reduced by 60-70%. Compression is generally much
148 better than that achieved by LZW (as used in @code{compress}), Huffman
149 coding (as used in @code{pack}), or adaptive Huffman coding
152 Compression is always performed, even if the compressed file is slightly
153 larger than the original. The worst case expansion is a few bytes for
154 the @code{gzip} file header, plus 5 bytes every 32K block, or an expansion
155 ratio of 0.015% for large files. Note that the actual number of used
156 disk blocks almost never increases. @code{gzip} preserves the mode,
157 ownership and timestamps of files when compressing or decompressing.
159 The @code{gzip} file format is specified in P. Deutsch, @sc{gzip} file
160 format specification version 4.3,
161 @uref{ftp://ftp.isi.edu/in-notes/rfc1952.txt, Internet RFC 1952} (May
162 1996). The @code{zip} deflation format is specified in P. Deutsch,
163 @sc{deflate} Compressed Data Format Specification version 1.3,
164 @uref{ftp://ftp.isi.edu/in-notes/rfc1951.txt, Internet RFC 1951} (May
167 @node Sample, Invoking gzip, Overview, Top
168 @chapter Sample Output
171 Here are some realistic examples of running @code{gzip}.
173 This is the output of the command @samp{gzip -h}:
178 usage: gzip [-cdfhlLnNrtvV19] [-S suffix] [file ...]
179 -c --stdout write on standard output, keep original files unchanged
180 -d --decompress decompress
181 -f --force force overwrite of output file and compress links
182 -h --help give this help
183 -l --list list compressed file contents
184 -L --license display software license
185 -n --no-name do not save or restore the original name and time stamp
186 -N --name save or restore the original name and time stamp
187 -q --quiet suppress all warnings
188 -r --recursive operate recursively on directories
189 -S .suf --suffix .suf use suffix .suf on compressed files
190 -t --test test compressed file integrity
191 -v --verbose verbose mode
192 -V --version display version number
193 -1 --fast compress faster
194 -9 --best compress better
195 file... files to (de)compress. If none given, use standard input.
196 Report bugs to <bug-gzip@@gnu.org>.
199 This is the output of the command @samp{gzip -v texinfo.tex}:
202 texinfo.tex: 69.7% -- replaced with texinfo.tex.gz
205 The following command will find all @code{gzip} files in the current
206 directory and subdirectories, and extract them in place without
207 destroying the original:
210 find . -name '*.gz' -print | sed 's/^\(.*\)[.]gz$/gunzip < "&" > "\1"/' | sh
213 @node Invoking gzip, Advanced usage, Sample, Top
214 @chapter Invoking @code{gzip}
218 The format for running the @code{gzip} program is:
221 gzip @var{option} @dots{}
224 @code{gzip} supports the following options:
230 Write output on standard output; keep original files unchanged.
231 If there are several input files, the output consists of a sequence of
232 independently compressed members. To obtain better compression,
233 concatenate all input files before compressing them.
242 Force compression or decompression even if the file has multiple links
243 or the corresponding file already exists, or if the compressed data
244 is read from or written to a terminal. If the input data is not in
245 a format recognized by @code{gzip}, and if the option @samp{--stdout} is also
246 given, copy the input data without change to the standard output: let
247 @code{zcat} behave as @code{cat}. If @samp{-f} is not given, and
248 when not running in the background, @code{gzip} prompts to verify
249 whether an existing file should be overwritten.
253 Print an informative help message describing the options then quit.
257 For each compressed file, list the following fields:
260 compressed size: size of the compressed file
261 uncompressed size: size of the uncompressed file
262 ratio: compression ratio (0.0% if unknown)
263 uncompressed_name: name of the uncompressed file
266 The uncompressed size is given as @samp{-1} for files not in @code{gzip}
267 format, such as compressed @samp{.Z} files. To get the uncompressed size for
268 such a file, you can use:
274 In combination with the @samp{--verbose} option, the following fields are also
278 method: compression method (deflate,compress,lzh,pack)
279 crc: the 32-bit CRC of the uncompressed data
280 date & time: time stamp for the uncompressed file
283 The crc is given as ffffffff for a file not in gzip format.
285 With @samp{--verbose}, the size totals and compression ratio for all files
286 is also displayed, unless some sizes are unknown. With @samp{--quiet},
287 the title and totals lines are not displayed.
289 The @code{gzip} format represents the the input size modulo
290 @math{2^32}, so the uncompressed size and compression ratio are listed
291 incorrectly for uncompressed files 4 GB and larger. To work around
292 this problem, you can use the following command to discover a large
293 uncompressed file's true size:
301 Display the @code{gzip} license then quit.
305 When compressing, do not save the original file name and time stamp by
306 default. (The original name is always saved if the name had to be
307 truncated.) When decompressing, do not restore the original file name
308 if present (remove only the @code{gzip}
309 suffix from the compressed file name) and do not restore the original
310 time stamp if present (copy it from the compressed file). This option
311 is the default when decompressing.
315 When compressing, always save the original file name and time stamp; this
316 is the default. When decompressing, restore the original file name and
317 time stamp if present. This option is useful on systems which have
318 a limit on file name length or when the time stamp has been lost after
323 Suppress all warning messages.
327 Travel the directory structure recursively. If any of the file names
328 specified on the command line are directories, @code{gzip} will descend
329 into the directory and compress all the files it finds there (or
330 decompress them in the case of @code{gunzip}).
333 While compressing, synchronize the output occasionally based on the
334 input. This increases size by less than 1 percent most cases, but
335 means that the @code{rsync} program can take advantage of similarities
336 in the uncompressed input when syncronizing two files compressed with
337 this flag. @code{gunzip} cannot tell the difference between a
338 compressed file created with this option, and one created without it.
340 @item --suffix @var{suf}
342 Use suffix @samp{@var{suf}} instead of @samp{.gz}. Any suffix can be
343 given, but suffixes other than @samp{.z} and @samp{.gz} should be
344 avoided to avoid confusion when files are transferred to other systems.
345 A null suffix forces gunzip to try decompression on all given files
346 regardless of suffix, as in:
349 gunzip -S "" * (*.* for MSDOS)
352 Previous versions of gzip used the @samp{.z} suffix. This was changed to
353 avoid a conflict with @code{pack}.
357 Test. Check the compressed file integrity.
361 Verbose. Display the name and percentage reduction for each file compressed.
365 Version. Display the version number and compilation options, then quit.
370 Regulate the speed of compression using the specified digit @var{n},
371 where @samp{-1} or @samp{--fast} indicates the fastest compression
372 method (less compression) and @samp{--best} or @samp{-9} indicates the
373 slowest compression method (optimal compression). The default
374 compression level is @samp{-6} (that is, biased towards high compression at
378 @node Advanced usage, Environment, Invoking gzip, Top
379 @chapter Advanced usage
380 @cindex concatenated files
382 Multiple compressed files can be concatenated. In this case,
383 @code{gunzip} will extract all members at once. If one member is
384 damaged, other members might still be recovered after removal of the
385 damaged member. Better compression can be usually obtained if all
386 members are decompressed and then recompressed in a single step.
388 This is an example of concatenating @code{gzip} files:
391 gzip -c file1 > foo.gz
392 gzip -c file2 >> foo.gz
407 In case of damage to one member of a @samp{.gz} file, other members can
408 still be recovered (if the damaged member is removed). However,
409 you can get better compression by compressing all members at once:
412 cat file1 file2 | gzip > foo.gz
415 compresses better than
418 gzip -c file1 file2 > foo.gz
421 If you want to recompress concatenated files to get better compression, do:
424 zcat old.gz | gzip > new.gz
427 If a compressed file consists of several members, the uncompressed
428 size and CRC reported by the @samp{--list} option applies to the last member
429 only. If you need the uncompressed size for all members, you can use:
435 If you wish to create a single archive file with multiple members so
436 that members can later be extracted independently, use an archiver such
437 as @code{tar} or @code{zip}. GNU @code{tar} supports the @samp{-z}
438 option to invoke @code{gzip} transparently. @code{gzip} is designed as a
439 complement to @code{tar}, not as a replacement.
441 @node Environment, Tapes, Advanced usage, Top
445 The environment variable @code{GZIP} can hold a set of default options for
446 @code{gzip}. These options are interpreted first and can be overwritten by
447 explicit command line parameters. For example:
450 for sh: GZIP="-8v --name"; export GZIP
451 for csh: setenv GZIP "-8v --name"
452 for MSDOS: set GZIP=-8v --name
455 On Vax/VMS, the name of the environment variable is @code{GZIP_OPT}, to
456 avoid a conflict with the symbol set for invocation of the program.
458 @node Tapes, Problems, Environment, Top
459 @chapter Using @code{gzip} on tapes
462 When writing compressed data to a tape, it is generally necessary to pad
463 the output with zeroes up to a block boundary. When the data is read and
464 the whole block is passed to @code{gunzip} for decompression,
465 @code{gunzip} detects that there is extra trailing garbage after the
466 compressed data and emits a warning by default if the garbage contains
467 nonzero bytes. You have to use the
468 @samp{--quiet} option to suppress the warning. This option can be set in the
469 @code{GZIP} environment variable, as in:
472 for sh: GZIP="-q" tar -xfz --block-compress /dev/rst0
473 for csh: (setenv GZIP "-q"; tar -xfz --block-compress /dev/rst0)
476 In the above example, @code{gzip} is invoked implicitly by the @samp{-z}
477 option of GNU @code{tar}. Make sure that the same block size (@samp{-b}
478 option of @code{tar}) is used for reading and writing compressed data on
479 tapes. (This example assumes you are using the GNU version of
482 @node Problems, Copying This Manual, Tapes, Top
483 @chapter Reporting Bugs
486 If you find a bug in @code{gzip}, please send electronic mail to
487 @email{bug-gzip@@gnu.org}. Include the version number,
488 which you can find by running @w{@samp{gzip -V}}. Also include in your
489 message the hardware and operating system, the compiler used to compile
491 a description of the bug behavior, and the input to @code{gzip} that triggered
494 @node Copying This Manual, Concept Index, Problems, Top
495 @appendix Copying This Manual
498 * GNU Free Documentation License:: License for copying this manual.
503 @node Concept Index, , Copying This Manual, Top
504 @appendix Concept Index