Scripting

Command substitution

`some command` plugs the some command output into some other context. Currently backticks are deprecated in favour of $(some command) form which:

1. permits nesting $(some command $(some nested))
2. treats \\ differently

Process substitution

<(command_list) Feeds the output of processes into the stdin of another process (piping allows only one command output to be redirected into stdin of another process).

The process list is run with its input or output connected to a FIFO or some file in /dev/fd. The name of this file is passed as an argument to the current command as the result of the expansion.

Bash

Shell implementation

Variable expansion

Introduced by $ e.g. $param_name. Variable may be enclosed in braces in order to separate it from adjacent characters

Indirect expansion

todo

Quotes

Single quotes preserve everything. No expansion occurs

Built-in variables

Positional parameters

• $# number of command line arguments
• $* all of positional arguments (single word)

Special

• $? last command's exit code

Test constructs

if statement tests wheter exit status of command is equal to 0

if test command; then
   #sth
fi

bash has alias for test command and is called [ so the statement becomes:

if [ command ]; then
fi

With Bash 2.02 extended test statement was introduced:

if [[ command ]]; then
fi

Extended test statement is not compliant with POSIX. It features:

• No need to quote variables (in [ not quoted variables like "something with spaces" will yield error: too many arguments)
• Regular expression matching with =~ e.g. [[ $variable =~ .*string ]]
• Wildcard matching e.g. [[ $(lxc-ls) == *"desired_container"* ]]
• Chaining tests with and and or like: [[ ... && ... ]] (otherwise: [] && [])

Negate

To negate test condition: if ! [ ... ]

AWK

print the quote sign

awk 'BEGIN { print "Here is a single quote <'"'"'>" }' #this is concat of three strings '...' "'" '...'
awk 'BEGIN { print "Here is a single quote <'\''>" }'

simple examples:

#longest line in data file
awk '{ if (length($0) > max) max = length($0) }
     END { print max }' data
# Print every line that has at least one field
awk 'NF > 0' data

you can have multiple rules in awk (first goes first rule, then second, if line contains both patterns then will be processed two times)

awk '/12/ { print $0 }
     /21/ { print $0 }' data

awk env variables:

• AWKPATH - contains awk programs
• AWKLIBPATH - contains extensions (can be written in C/C++)

References

1. https://google.github.io/styleguide/shell.xml
2. https://github.com/robbyrussell/oh-my-zsh/wiki/Coding-style-guide
3. http://tldp.org/HOWTO/Bash-Prog-Intro-HOWTO.html
4. https://www.kernel.org/doc/Documentation/process/coding-style.rst
5. https://github.com/zsh-users/zsh-completions/blob/master/zsh-completions-howto.org
6. http://www.tldp.org/LDP/abs/html/internalvariables.html
7. http://mywiki.wooledge.org/BashFAQ
8. http://mywiki.wooledge.org/BashPitfalls
9. https://www.gnu.org/software/bash/manual/html_node/Shell-Parameter-Expansion.html