Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



A suite of Unix command line tools for englightened pipelines.


The name "zip" was already taken. This interlaces the contents of a variadic list of files, especially handy with subshell file descriptor replacement. The following shows the opposing sides of a six sided die. The first example works in BSD variants, and the second in SysV/Linux variants::

xip <(jot 6) <(jot 6 6 1) | xargs -n 2
xip <(seq 6) <(seq 6 -1 1) | xargs -n 2


Buffers stdin into memory until it closes, then writes to a given file. This is handy for pipeline loops, where normally redirecting to and from the same file would result in premature truncation. The following are equivalent::

sort file
cat file | sort | dog file


Shuffles the lines from standard input and writes them to standard output. If the input stream is indefinite, the input can be shuffled indefinitely by specifying a pool size as a second argument.


Enquotes every line from standard input and writes it to standard output. Handy for pipelines that end in xargs, when file -print0 and xargs -0 aren't really an option. The following are pretty close to equivalent:

find . -print0 | xargs -0 tar cf -
find . | enquote | xargs tar cf -


Reads all of standard input and writes it back, repeatedly, to standard output. The first pass gets written immediately, so an indifinite stream will only cause memory bloat at the rate at which cycle's standard output is consumed::

find . -name '*.mp3' \
| cycle \
| shuffle 10000 \
| enquote \
| xargs -n 1 mpg123


Convenience command for find . -name "<file_pattern>" -print0 | xargs -0 grep "<grep_pattern>". Significantly faster than grep -r.

findall "<pattern1>"

is equivalent to

find . -name "*" -print0 | xargs -0 grep "<pattern1>"


findall "<pattern1>" "<pattern2>"

expands to

find . -name "<pattern1>" -print0 | xargs -0 grep "<pattern2>"


Writes a random sample of the lines from standard input back out to standard output. Specify a sample size as an argument. The sample is in stable order. The input stream may safely be of indefinite size.

This program is called samp so as not to collide with bsd's random or OS X's sample.


Reads lines from standard input and writes the first of each unique line. This is similar to the behavior of the highly optimized uniq. uniq suppresses the output of sequentially redundant lines, so to create a truly unique set, the common idiom is to sort the input. However, this does not preserve the order. In this sense, unique is a "stable" operation.

The following are equivalent:

cat input | sort | uniq | dog output
cat input | unique | sort | dog output

Example, removing redundant elements from $PATH:

export PATH=$(
    echo $PATH \
    | tr ':' '\n' \
    | unique \
    | paste -sd: -


Source this script to add this project to your PATH.

. bin/activate


After activating this project, use mkroot to replicate it.

mkroot foo

See https://github.com/kriskowal/mkroot for details.