'\" te
|
.TH pixz 1
|
.SH NAME
|
pixz \- parallel, indexing version of XZ
|
|
.SH SYNOPSIS
|
.LP
|
.nf
|
\fB/usr/bin/pixz\fR input output.pxz
|
.fi
|
|
.LP
|
.nf
|
\fB/usr/bin/pixz\fR -d input.pxz output
|
.fi
|
|
.LP
|
.nf
|
\fB/usr/bin/pixz\fR -l input.tpxz
|
.fi
|
|
.LP
|
.nf
|
\fB/usr/bin/pixz\fR -x /path/to/file < input.tpxz
|
.fi
|
|
.SH DESCRIPTION
|
.sp
|
.LP
|
The existing XZ Utils ( http://tukaani.org/xz/ ) provide great compression
|
in the .xz file format, but they have two significant problems:
|
.sp
|
.LP
|
.RS +4
|
.TP
|
.ie t \(bu
|
.el o
|
They are single-threaded, while most users nowadays have multi-core computers.
|
.RE
|
.RS +4
|
.TP
|
.ie t \(bu
|
.el o
|
The .xz files they produce are just one big block of compressed data, rather than a collection of smaller blocks. This makes random access to the original data impossible.
|
.RE
|
.sp
|
.LP
|
With pixz, both these problems are solved.
|
|
.SH EXAMPLES
|
.sp
|
.LP
|
Specifying input and output:
|
.sp
|
.ne 2
|
.mk
|
.na
|
$ pixz < foo.tar > foo.tpxz
|
.LP
|
|
.sp .6
|
.RS 4n
|
Same as 'pixz foo.tar foo.tpxz'
|
.RE
|
|
.sp
|
.ne 2
|
.mk
|
.na
|
$ pixz -i foo.tar -o foo.tpxz
|
.LP
|
|
.sp .6
|
.RS 4n
|
Ditto. These both work for -x, -d and -l too, eg:
|
.RE
|
|
.sp
|
.ne 2
|
.mk
|
.na
|
$ pixz -x -i foo.tpxz -o foo.tar file1 file2 ...
|
.LP
|
|
.sp .6
|
.RS 4n
|
# Extract the files from foo.tpxz into foo.tar
|
.RE
|
|
.sp
|
.ne 2
|
.mk
|
.na
|
$ pixz foo.tar
|
.LP
|
|
.sp .6
|
.RS 4n
|
Compress it to foo.tpxz, removing the original
|
.RE
|
|
.sp
|
.ne 2
|
.mk
|
.na
|
$ pixz -d foo.tpxz
|
.LP
|
|
.sp .6
|
.RS 4n
|
Extract it to foo.tar, removing the original
|
.RE
|
|
.SH OPTIONS
|
.TP
|
.B -1
|
Faster, worse compression.
|
.TP
|
.B -9
|
Better, slower compression.
|
.TP
|
.B -p <number>
|
Cap the number of threads at <number>.
|
.TP
|
.B -t
|
Compress but don't treat it as a tarball (don't index it).
|
.TP
|
.B -d
|
Decompress, don't check that contents match index.
|
.TP
|
.B -l
|
List the xz blocks instead of files.
|
|
.SH WARNING
|
Running pixz without the -t flag will cause it to treat the input
|
as a tarball, as long as it looks vaguely tarball-like. This means if the
|
file starts with at least 1024 zero bytes, pixz will assume it's empty, and
|
truncate the output! If your input files aren't tarballs, run with -t or
|
face possible data-loss.
|
