BASH(1)

📦

version 5.3 · 4 versions

5.3 · ubuntu resolute, ubuntu stonking (latest)
5.2.37 · ubuntu questing
5.1 · ubuntu jammy
5.2.37 · debian bookworm, debian trixie, ubuntu noble

👑 BASH(1) 👑

so this is the manual page for bash — the GNU Bourne-Again SHell — from the General Commands Manual section, bestie. yes THE shell. the one you've been typing into this whole time. sit down, hydrate, we're doing this 💅

🎀 NAME 🎀

bash — GNU Bourne-Again SHell 💖

✨ SYNOPSIS ✨

bash [options] [command_string | file]

that's it. that's the whole vibe. options, then either a string of commands or a file. iconic. 🦋

💜 COPYRIGHT 💜

🌸 DESCRIPTION 🌸

bash is a command language interpreter — it runs commands it reads from the standard input, from a string, or from a file. It's a reimplementation and an extension of the Bourne shell, aka the historical Unix command language interpreter (the OG). And it also stole some very useful features from the Korn and C shells (ksh and csh) because why not 💅

okay so POSIX is the name for a whole family of computing standards based on Unix. Bash is intended to be a conformant implementation of the Shell and Utilities portion of the IEEE POSIX specification (IEEE Standard 1003.1).

Now — Bash POSIX mode (from here on out we call it posix mode) is the setting that changes the shell's behavior wherever the default operation differs from the standard, so that it strictly conforms to the standard. Wanna know exactly how posix mode messes with bash's behavior? There's a whole document for that — check SEE ALSO below for the reference. And yes, bash can be configured to be POSIX-conformant by default. she contains multitudes ✨

💅 OPTIONS 💅

First things first, bestie: all of the single-character shell options documented in the description of the set builtin command — including -o — can be used as options when the shell is invoked. That's a lot of options. we love her for that.

On top of that, bash interprets the following options when it is invoked:

-c — If the -c option is present, then commands are read from the first non-option argument command_string. If there are arguments after the command_string, the first argument is assigned to $0 and any remaining arguments are assigned to the positional parameters. That assignment to $0 sets the name of the shell, which is what shows up in warning and error messages. (so yes, you can gaslight your own error messages, kind of) 🙈
-i — If the -i option is present, the shell is interactive. 💬
-l — Make bash act as if it had been invoked as a login shell (peep INVOCATION below). 🚪
-r — If the -r option is present, the shell becomes restricted (see RESTRICTED SHELL below). 🔒
-s — If the -s option is present, or if no arguments remain after option processing, the shell reads commands from the standard input. This option lets the positional parameters be set when you're invoking an interactive shell or when you're reading input through a pipe. 🚰
-v — Print shell input lines as they are read. 👀
-x — Print commands and their arguments as they are executed. 🔍
-D — Print a list of all double-quoted strings preceded by $ on the standard output. These are the strings that are subject to language translation when the current locale is not C or POSIX. This implies the -n option; no commands will be executed. 🌍
[-+]O [shopt_option] — shopt_option is one of the shell options accepted by the shopt builtin (see SHELL BUILTIN COMMANDS below). If shopt_option is present, -O sets the value of that option; +O unsets it. If shopt_option is not supplied, bash prints the names and values of the shell options accepted by shopt on the standard output. And if the invocation option is +O, that output is displayed in a format that may be reused as input. 🔁
-- — A -- signals the end of options and disables further option processing. Any arguments after the -- are treated as a shell script filename (see below) and arguments passed to that script. An argument of - is equivalent to --. ✋

🌈 the multi-character ones 🌈

Bash also interprets a number of multi-character options. Heads up though: these options must appear on the command line before the single-character options to be recognized. order matters, bestie, it's giving choreography 🩰

--debugger — Arrange for the debugger profile to be executed before the shell starts. Turns on extended debugging mode (see the description of the extdebug option to the shopt builtin below). 🐛
--dump-po-strings — Equivalent to -D, but the output is in the GNU gettext "po" (portable object) file format. 📦
--dump-strings — Equivalent to -D. 📦
--help — Display a usage message on standard output and exit successfully. 🆘
--init-file file — Execute commands from file instead of the system wide initialization file /etc/bash.bashrc and the standard personal initialization file ~/.bashrc if the shell is interactive (see INVOCATION below).
--rcfile file — Execute commands from file instead of the system wide initialization file /etc/bash.bashrc and the standard personal initialization file ~/.bashrc if the shell is interactive (see INVOCATION below).
--login — Equivalent to -l. 🚪
--noediting — Do not use the GNU readline library to read command lines when the shell is interactive. 🚫⌨️
--noprofile — Do not read either the system-wide startup file /etc/profile or any of the personal initialization files ~/.bash_profile, ~/.bash_login, or ~/.profile. By default, bash reads these files when it is invoked as a login shell (see INVOCATION below). 🚫📄
--norc — Do not read and execute the system wide initialization file /etc/bash.bashrc and the personal initialization file ~/.bashrc if the shell is interactive. This option is on by default if the shell is invoked as sh. 🚫📄
--posix — Enable posix mode; change the behavior of bash where the default operation differs from the POSIX standard to match the standard. 📏
--restricted — The shell becomes restricted (see RESTRICTED SHELL below). 🔒
--verbose — Equivalent to -v. 👀
--version — Show version information for this instance of bash on the standard output and exit successfully. 🏷️

🦋 ARGUMENTS 🦋

Okay so if arguments remain after option processing, and neither the -c nor the -s option has been supplied, then the first argument is treated as the name of a file containing shell commands (a shell script). When bash is invoked in this fashion:

$0 is set to the name of the file
the positional parameters are set to the remaining arguments

Bash reads and executes commands from this file, then exits. Bash's exit status is the exit status of the last command executed in the script. And if no commands are executed, the exit status is 0. 🫶

Oh and where does it look for the file? Bash first attempts to open the file in the current directory, and if no file is found, it searches the directories in PATH for the script. 🔎

🌟 INVOCATION 🌟

A login shell is one whose first character of argument zero is a -, or one started with the --login option. that's it, that's the definition 💅

An interactive shell is one started without non-option arguments (unless -s is specified) and without the -c option, and whose standard input and standard error are both connected to terminals (as determined by isatty(3)), or one started with the -i option. Bash sets PS1 and $- includes i if the shell is interactive — so a shell script or a startup file can test this state. 🕵️♀️

🧚 how bash runs its startup files 🧚

The following paragraphs describe how bash executes its startup files. And listen: if any of the files exist but cannot be read, bash reports an error. Tildes are expanded in filenames as described below under Tilde Expansion in the EXPANSION section.

When bash is invoked as an interactive login shell, or as a non-interactive shell with the --login option, it first reads and executes commands from the file /etc/profile, if that file exists. After reading that file, it looks for ~/.bash_profile, ~/.bash_login, and ~/.profile, in that order, and reads and executes commands from the first one that exists and is readable. The --noprofile option may be used when the shell is started to inhibit this behavior. 🚫

When an interactive login shell exits, or a non-interactive login shell executes the exit builtin command, bash reads and executes commands from the file ~/.bash_logout, if it exists. 👋

When an interactive shell that is not a login shell is started, bash reads and executes commands from etc/bash.bashrc and ~/.bashrc, if that file exists. The --norc option inhibits this behavior. The --rcfile file option causes bash to use file instead of etc/bash.bashrc and ~/.bashrc. 🔄

When bash is started non-interactively, to run a shell script for example, it looks for the variable BASH_ENV in the environment, expands its value if it appears there, and uses the expanded value as the name of a file to read and execute. Bash behaves as if the following command were executed:

if [ -n "$BASH_ENV" ]; then . "$BASH_ENV"; fi

...but it does not use the value of the PATH variable to search for the filename. nope. no path searching. 🚫🔎

💀 invoked as `sh` 💀

If bash is invoked with the name sh, it tries to mimic the startup behavior of historical versions of sh as closely as possible, while conforming to the POSIX standard as well. Two-for-one behavior, we love a versatile queen 👑

When invoked as an interactive login shell, or a non-interactive shell with the --login option, it first attempts to read and execute commands from /etc/profile and ~/.profile, in that order. The --noprofile option inhibits this behavior.
When invoked as an interactive shell with the name sh, bash looks for the variable ENV, expands its value if it is defined, and uses the expanded value as the name of a file to read and execute.
Since a shell invoked as sh does not attempt to read and execute commands from any other startup files, the --rcfile option has no effect. (literally nothing. it does nothing here. 💀)
A non-interactive shell invoked with the name sh does not attempt to read any other startup files.

When invoked as sh, bash enters posix mode after reading the startup files. 📏

📏 posix mode startup 📏

When bash is started in posix mode, as with the --posix command line option, it follows the POSIX standard for startup files. In this mode, interactive shells expand the ENV variable and read and execute commands from the file whose name is the expanded value. No other startup files are read. that's the whole menu, bestie 🍽️

🌐 the network connection thing 🌐

Bash attempts to determine when it is being run with its standard input connected to a network connection — like when it's executed by the historical and rarely-seen remote shell daemon, usually rshd, or the secure shell daemon sshd. If bash determines it is being run non-interactively in this fashion, it reads and executes commands from etc/bash.bashrc and ~/.bashrc, if these files exist and are readable. 🛰️

Bash does not read this file if invoked as sh. The --norc option inhibits this behavior, and the --rcfile option makes bash use a different file instead of ~/.bashrc — but neither rshd nor sshd generally invoke the shell with those options or allow them to be specified. so like. good luck with that 🙈

🔐 the setuid / setgid situation 🔐

listen carefully because this one is a real security-flavored caveat, no cap:

If the shell is started with the effective user (group) id not equal to the real user (group) id, and the -p option is not supplied, then:

no startup files are read 📄🚫
shell functions are not inherited from the environment 🚫
the SHELLOPTS, BASHOPTS, CDPATH, and GLOBIGNORE variables, if they appear in the environment, are ignored 🙈
and the effective user id is set to the real user id 🔁

If the -p option is supplied at invocation, the startup behavior is the same, but the effective user id is not reset. 💅

okay bestie that's part 1, and there is so much more coming — we're not even close to done. stay tuned 🫶✨

🦋 DEFINITIONS 🦋

okay bestie, these definitions get used alllll throughout the rest of this doc, so lock in 💅

blank — a space or tab. literally just a space or tab, no cap.
whitespace — a character that belongs to the space character class in the current locale, or one for which isspace(3) returns true ✨
word — a sequence of characters considered as a single unit by the shell. also known as a token 🎀
name — a word made only of alphanumeric characters and underscores, and it has to begin with an alphabetic character or an underscore. also referred to as an identifier 💖
metacharacter — a character that, when unquoted, separates words. one of the following:

|  & ; ( ) < > space tab newline

control operator — a token that performs a control function. it is one of the following symbols:

|| & && ; ;; ;& ;;& ( ) | |& <newline>

👑 RESERVED WORDS 👑

reserved words are words that have a special meaning to the shell, periodt. the following words are recognized as reserved when they're unquoted and either the first word of a command (see SHELL GRAMMAR below), the third word of a case or select command (only in is valid), or the third word of a for command (only in and do are valid):

! case  coproc  do done elif else esac fi for function if in select then
until while { } time [[ ]]

🌈 SHELL GRAMMAR 🌈

this section describes the syntax of the various forms of shell commands. buckle up, bestie 🎀

Simple Commands

a simple command is a sequence of optional variable assignments followed by blank-separated words and redirections, and terminated by a control operator. the first word specifies the command to be executed, and is passed as argument zero. the remaining words are passed as arguments to the invoked command.

the return value of a simple command is its exit status, or 128+n if the command is terminated by signal n.

Pipelines

a pipeline is a sequence of one or more commands separated by one of the control operators | or |&. the format for a pipeline is:

[time [-p]] [ ! ] command1 [ [|⎪|&] command2 ... ]

the standard output of command1 is connected via a pipe to the standard input of command2. this connection is performed before any redirections specified by the command1 (see REDIRECTION below). if |& is the pipeline operator, command1's standard error, in addition to its standard output, is connected to command2's standard input through the pipe; it is shorthand for 2>&1 |. this implicit redirection of the standard error to the standard output is performed after any redirections specified by command1.

the return status of a pipeline is the exit status of the last command, unless the pipefail option is enabled. if pipefail is enabled, the pipeline's return status is the value of the last (rightmost) command to exit with a non-zero status, or zero if all commands exit successfully. if the reserved word ! precedes a pipeline, the exit status of that pipeline is the logical negation of the exit status as described above. if a pipeline is executed synchronously, the shell waits for all commands in the pipeline to terminate before returning a value.

if the time reserved word precedes a pipeline, the shell reports the elapsed as well as user and system time consumed by its execution when the pipeline terminates. the -p option changes the output format to that specified by POSIX. when the shell is in posix mode, it does not recognize time as a reserved word if the next token begins with a “-”. the value of the TIMEFORMAT variable is a format string that specifies how the timing information should be displayed; see the description of TIMEFORMAT below under Shell Variables.

when the shell is in posix mode, time may appear by itself as the only word in a simple command. in this case, the shell displays the total user and system time consumed by the shell and its children. the TIMEFORMAT variable specifies the format of the time information.

each command in a multi-command pipeline, where pipes are created, is executed in a subshell, which is a separate process. see COMMAND EXECUTION ENVIRONMENT for a description of subshells and a subshell environment. if the lastpipe option is enabled using the shopt builtin (see the description of shopt below), and job control is not active, the last element of a pipeline may be run by the shell process.

Lists

a list is a sequence of one or more pipelines separated by one of the operators ;, &, &&, or ||, and optionally terminated by one of ;, &, or <newline>.

of these list operators, && and || have equal precedence, followed by ; and &, which have equal precedence.

a sequence of one or more newlines may appear in a list instead of a semicolon to delimit commands.

if a command is terminated by the control operator &, the shell executes the command in the background in a subshell. the shell does not wait for the command to finish, and the return status is 0. these are referred to as asynchronous commands. commands separated by a ; are executed sequentially; the shell waits for each command to terminate in turn. the return status is the exit status of the last command executed.

AND and OR lists are sequences of one or more pipelines separated by the && and || control operators, respectively. AND and OR lists are executed with left associativity. an AND list has the form

command1 && command2

command2 is executed if, and only if, command1 returns an exit status of zero (success).

an OR list has the form

command1 || command2

command2 is executed if, and only if, command1 returns a non-zero exit status. the return status of AND and OR lists is the exit status of the last command executed in the list.

Compound Commands

a compound command is one of the following. in most cases a list in a command's description may be separated from the rest of the command by one or more newlines, and may be followed by a newline in place of a semicolon.

(list) — list is executed in a subshell (see COMMAND EXECUTION ENVIRONMENT below for a description of a subshell environment). variable assignments and builtin commands that affect the shell's environment do not remain in effect after the command completes. the return status is the exit status of list.
{ list; } — list is executed in the current shell environment. list must be terminated with a newline or semicolon. this is known as a group command. the return status is the exit status of list.

note that unlike the metacharacters ( and ), { and } are reserved words and must occur where a reserved word is permitted to be recognized. since they do not cause a word break, they must be separated from list by whitespace or another shell metacharacter.

((expression)) — the arithmetic expression is evaluated according to the rules described below under ARITHMETIC EVALUATION. if the value of the expression is non-zero, the return status is 0; otherwise the return status is 1. the expression undergoes the same expansions as if it were within double quotes, but unescaped double quote characters in expression are not treated specially and are removed. since this can potentially result in empty strings, this command treats those as expressions that evaluate to 0.
[[ expression ]] — evaluate the conditional expression expression and return a status of zero (true) or non-zero (false). expressions are composed of the primaries described below under CONDITIONAL EXPRESSIONS. the words between the [[ and ]] do not undergo word splitting and pathname expansion. the shell performs tilde expansion, parameter and variable expansion, arithmetic expansion, command substitution, process substitution, and quote removal on those words. conditional operators such as -f must be unquoted to be recognized as primaries.

when used with [[, the < and > operators sort lexicographically using the current locale.

see the description of the test builtin command (section SHELL BUILTIN COMMANDS below) for the handling of parameters (i.e. missing parameters).

when the == and != operators are used, the string to the right of the operator is considered a pattern and matched according to the rules described below under Pattern Matching, as if the extglob shell option were enabled. the = operator is equivalent to ==. if the nocasematch shell option is enabled, the match is performed without regard to the case of alphabetic characters. the return value is 0 if the string matches (==) or does not match (!=) the pattern, and 1 otherwise. if any part of the pattern is quoted, the quoted portion is matched as a string: every character in the quoted portion matches itself, instead of having any special pattern matching meaning.

an additional binary operator, =~, is available, with the same precedence as == and !=. when it is used, the string to the right of the operator is considered a POSIX extended regular expression and matched accordingly (using the POSIX regcomp and regexec interfaces usually described in regex(3)). the return value is 0 if the string matches the pattern, and 1 otherwise. if the regular expression is syntactically incorrect, the conditional expression's return value is 2. if the nocasematch shell option is enabled, the match is performed without regard to the case of alphabetic characters.

if any part of the pattern is quoted, the quoted portion is matched literally, as above. if the pattern is stored in a shell variable, quoting the variable expansion forces the entire pattern to be matched literally. treat bracket expressions in regular expressions carefully, since normal quoting and pattern characters lose their meanings between brackets.

the match succeeds if the pattern matches any part of the string. anchor the pattern using the ^ and $ regular expression operators to force it to match the entire string.

The array variable BASH_REMATCH is the one keeping receipts 📝 — it records which parts of the string matched the pattern. The element of BASH_REMATCH with index 0 holds the portion of the string matching the entire regular expression. Substrings matched by parenthesized subexpressions inside the regular expression get saved in the remaining BASH_REMATCH indices. The element of BASH_REMATCH with index n is the portion of the string matching the *n*th parenthesized subexpression. And bestie, hear me: Bash sets BASH_REMATCH in the global scope — declaring it as a local variable will lead to unexpected results. do not do that 💀

Expressions may be combined using the following operators, listed in decreasing order of precedence (so the top one wins hardest, iykyk):

( expression ) — Returns the value of expression. This may be used to override the normal precedence of operators. parentheses said "i'm the main character now" 💅
! expression — True if expression is false.
expression1 && expression2 — True if both expression1 and expression2 are true.
expression1 || expression2 — True if either expression1 or expression2 is true.

The && and || operators do not evaluate expression2 if the value of expression1 is sufficient to determine the return value of the entire conditional expression. short-circuit behavior, she's efficient like that ✨

for name [ [ in word ... ] ; ] do list ; done

First, expand the list of words following in, generating a list of items. Then the variable name is set to each element of this list in turn, and list is executed each time. If the in word is omitted, the for command executes list once for each positional parameter that is set (see PARAMETERS below). The return status is the exit status of the last command that executes. If the expansion of the items following in results in an empty list, no commands are executed, and the return status is 0. no items, no problem, we just vibe 🦋

for (( expr1 ; expr2 ; expr3 )) [;] do list ; done

First, evaluate the arithmetic expression expr1 according to the rules described below under ARITHMETIC EVALUATION. Then, repeatedly evaluate the arithmetic expression expr2 until it evaluates to zero. Each time expr2 evaluates to a non-zero value, execute list and evaluate the arithmetic expression expr3. If any expression is omitted, it behaves as if it evaluates to 1. The return value is the exit status of the last command in list that is executed, or non-zero if any of the expressions is invalid. the C-style for loop said "i'm literally not like other loops" 🔥

Use the break and continue builtins (see SHELL BUILTIN COMMANDS below) to control loop execution. your escape hatch, bestie 🫶

select name [ in word ] ; do list ; done

First, expand the list of words following in, generating a list of items, and print the set of expanded words to the standard error, each preceded by a number. If the in word is omitted, print the positional parameters (see PARAMETERS below). select then displays the PS3 prompt and reads a line from the standard input. If the line consists of a number corresponding to one of the displayed words, then select sets the value of name to that word. If the line is empty, select displays the words and prompt again. If EOF is read, select completes and returns 1. Any other value sets name to null. The line read is saved in the variable REPLY. The list is executed after each selection until a break command is executed. The exit status of select is the exit status of the last command executed in list, or zero if no commands were executed. interactive menu energy, we love a choose-your-own-adventure 🎀

case word in [ [(] pattern [ | pattern ] ... ) list ;; ] ... esac

A case command first expands word, and tries to match it against each pattern in turn, proceeding from first to last, using the matching rules described under Pattern Matching below. A pattern list is a set of one or more patterns separated by |, and the ) operator terminates the pattern list. The word is expanded using tilde expansion, parameter and variable expansion, arithmetic expansion, command substitution, process substitution and quote removal. Each pattern examined is expanded using tilde expansion, parameter and variable expansion, arithmetic expansion, command substitution, process substitution, and quote removal. If the nocasematch shell option is enabled, the match is performed without regard to the case of alphabetic characters. A clause is a pattern list and an associated list. structure. elegance. the pattern matching ate ✨

When a match is found, case executes the corresponding list. If the ;; operator terminates the case clause, the case command completes after the first match. Using ;& in place of ;; causes execution to continue with the list associated with the next pattern list. Using ;;& in place of ;; causes the shell to test the next pattern list in the statement, if any, and execute any associated list if the match succeeds, continuing the case statement execution as if the pattern list had not matched. The exit status is zero if no pattern matches.

Otherwise, it is the exit status of the last command executed in the last list executed. three different terminators, three different personalities, pick your fighter 💜

if list; then list; [ elif list; then list; ] ... [ else list; ] fi

The if list is executed. If its exit status is zero, the then list is executed. Otherwise, each elif list is executed in turn, and if its exit status is zero, the corresponding then list is executed and the command completes. Otherwise, the else list is executed, if present. The exit status is the exit status of the last command executed, or zero if no condition tested true. zero means yes, non-zero means no, that's just the law of the land 🌸

while list-1; do list-2; done until list-1; do list-2; done

The while command continuously executes the list list-2 as long as the last command in the list list-1 returns an exit status of zero. The until command is identical to the while command, except that the test is negated: list-2 is executed as long as the last command in list-1 returns a non-zero exit status. The exit status of the while and until commands is the exit status of the last command executed in list-2, or zero if none was executed. until is just while with the vibes flipped, fr fr 🦋

🧚 Coprocesses 🧚

A coprocess is a shell command preceded by the coproc reserved word. A coprocess is executed asynchronously in a subshell, as if the command had been terminated with the & control operator, with a two-way pipe established between the executing shell and the coprocess. two-way pipe, two-way communication, it's giving healthy relationship 💖

The syntax for a coprocess is:

coproc [NAME] command [redirections]

This creates a coprocess named NAME. command may be either a simple command or a compound command (see above). NAME is a shell variable name. If NAME is not supplied, the default name is COPROC.

The recommended form to use for a coprocess is

coproc NAME { command [redirections]; }

This form is preferred because simple commands result in the coprocess always being named COPROC, and it is simpler to use and more complete than the other compound commands. bestie the docs literally said "use this one", listen to them 👑

If command is a compound command, NAME is optional. The word following coproc determines whether that word is interpreted as a variable name: it is interpreted as NAME if it is not a reserved word that introduces a compound command. If command is a simple command, NAME is not allowed; this is to avoid confusion between NAME and the first word of the simple command. ambiguity said "not today" ✨

When the coprocess is executed, the shell creates an array variable (see Arrays below) named NAME in the context of the executing shell. The standard output of command is connected via a pipe to a file descriptor in the executing shell, and that file descriptor is assigned to NAME[0]. The standard input of command is connected via a pipe to a file descriptor in the executing shell, and that file descriptor is assigned to NAME[1]. This pipe is established before any redirections specified by the command (see REDIRECTION below). The file descriptors can be utilized as arguments to shell commands and redirections using standard word expansions. Other than those created to execute command and process substitutions, the file descriptors are not available in subshells. not available in subshells, i repeat, not available 😭

The process ID of the shell spawned to execute the coprocess is available as the value of the variable NAME_PID. The wait builtin may be used to wait for the coprocess to terminate.

Since the coprocess is created as an asynchronous command, the coproc command always returns success. The return status of a coprocess is the exit status of command. coproc itself always says "slay, success", but the real tea is in command's exit status 💜

🎀 Shell Function Definitions 🎀

A shell function is an object that is called like a simple command and executes a compound command with a new set of positional parameters. Shell functions are declared as follows:

fname () compound-command [redirection]
function fname [()] compound-command [redirection]

This defines a function named fname. The reserved word function is optional. If the function reserved word is supplied, the parentheses are optional. The body of the function is the compound command compound-command (see Compound Commands above). That command is usually a list of commands between { and }, but may be any command listed under Compound Commands above. If the function reserved word is used, but the parentheses are not supplied, the braces are recommended. compound-command is executed whenever fname is specified as the name of a simple command. When in posix mode, fname must be a valid shell name and may not be the name of one of the POSIX special builtins. In default mode, a function name can be any unquoted shell word that does not contain $. no dollar signs in the name, she's picky like that 💅

Any redirections (see REDIRECTION below) specified when a function is defined are performed when the function is executed. define-time redirections, run-time performance, iconic behavior ✨

The exit status of a function definition is zero unless a syntax error occurs or a readonly function with the same name already exists. When executed, the exit status of a function is the exit status of the last command executed in the body. (See FUNCTIONS below.) last command in the body decides everything, no pressure or anything 🫶

💬 COMMENTS 💬

okay bestie, let's talk comments 💅 In a non-interactive shell, or an interactive shell in which the interactive_comments option to the shopt builtin is enabled (see SHELL BUILTIN COMMANDS below), a word beginning with # introduces a comment. A word begins at the beginning of a line, after unquoted whitespace, or after an operator. The comment causes that word and all remaining characters on that line to be ignored. An interactive shell without the interactive_comments option enabled does not allow comments. The interactive_comments option is enabled by default in interactive shells. ✨

💅 QUOTING 💅

Quoting is used to remove the special meaning of certain characters or words to the shell. Quoting can be used to disable special treatment for special characters, to prevent reserved words from being recognized as such, and to prevent parameter expansion. It's giving "respect my boundaries" fr fr.

Each of the metacharacters listed above under DEFINITIONS has special meaning to the shell and must be quoted if it is to represent itself.

When the command history expansion facilities are being used (see HISTORY EXPANSION below), the history expansion character, usually !, must be quoted to prevent history expansion. Don't let ! catch you slippin' 👀

There are four quoting mechanisms: the escape character, single quotes, double quotes, and dollar-single quotes. Four whole vibes, and they all matter.

A non-quoted backslash (\) is the escape character. It preserves the literal value of the next character that follows, removing any special meaning it has, with the exception of <newline>. If a \<newline> pair appears, and the backslash is not itself quoted, the \<newline> is treated as a line continuation (that is, it is removed from the input stream and effectively ignored). bye bye backslash-newline 👋

Enclosing characters in single quotes preserves the literal value of each character within the quotes. A single quote may not occur between single quotes, even when preceded by a backslash. So no nesting, bestie, it's giving "read the room" 💀

Enclosing characters in double quotes preserves the literal value of all characters within the quotes, with the exception of $, `, \, and, when history expansion is enabled, !. When the shell is in posix mode, the ! has no special meaning within double quotes, even when history expansion is enabled. The characters $ and ` retain their special meaning within double quotes. The backslash retains its special meaning only when followed by one of the following characters: $, `, ", \, or <newline>. Backslashes preceding characters without a special meaning are left unmodified.

A double quote may be quoted within double quotes by preceding it with a backslash. If enabled, history expansion will be performed unless an ! appearing in double quotes is escaped using a backslash. The backslash preceding the ! is not removed. Not me learning that the hard way 😭

The special parameters * and @ have special meaning when in double quotes (see PARAMETERS below).

Character sequences of the form $'string' are treated as a special variant of single quotes. The sequence expands to string, with backslash-escaped characters in string replaced as specified by the ANSI C standard. Backslash escape sequences, if present, are decoded as follows:

\a — alert (bell)
\b — backspace
\e — an escape character
\E — an escape character
\f — form feed
\n — new line
\r — carriage return
\t — horizontal tab
\v — vertical tab
\\ — backslash
\' — single quote
\" — double quote
\? — question mark
\nnn — the eight-bit character whose value is the octal value nnn (one to three octal digits).
\xHH — the eight-bit character whose value is the hexadecimal value HH (one or two hex digits).
\uHHHH — the Unicode (ISO/IEC 10646) character whose value is the hexadecimal value HHHH (one to four hex digits).
\UHHHHHHHH — the Unicode (ISO/IEC 10646) character whose value is the hexadecimal value HHHHHHHH (one to eight hex digits).
\cx — a control-x character.

The expanded result is single-quoted, as if the dollar sign had not been present. slay.

🌍 Translating Strings 🌍

A double-quoted string preceded by a dollar sign ($"string") causes the string to be translated according to the current locale. The gettext infrastructure performs the lookup and translation, using the LC_MESSAGES, TEXTDOMAINDIR, and TEXTDOMAIN shell variables. If the current locale is C or POSIX, if there are no translations available, or if the string is not translated, the dollar sign is ignored, and the string is treated as double-quoted as described above. This is a form of double quoting, so the string remains double-quoted by default, whether or not it is translated and replaced. If the noexpand_translation option is enabled using the shopt builtin, translated strings are single-quoted instead of double-quoted. See the description of shopt below under SHELL BUILTIN COMMANDS.

🧸 PARAMETERS 🧸

A parameter is an entity that stores values. It can be a name, a number, or one of the special characters listed below under Special Parameters. A variable is a parameter denoted by a name. A variable has a value and zero or more attributes. Attributes are assigned using the declare builtin command (see declare below in SHELL BUILTIN COMMANDS). The export and readonly builtins assign specific attributes.

A parameter is set if it has been assigned a value. The null string is a valid value. Once a variable is set, it may be unset only by using the unset builtin command (see SHELL BUILTIN COMMANDS below).

A variable is assigned to using a statement of the form

name=[value]

If value is not given, the variable is assigned the null string. All values undergo tilde expansion, parameter and variable expansion, command substitution, arithmetic expansion, and quote removal (see EXPANSION below). If the variable has its integer attribute set, then value is evaluated as an arithmetic expression even if the $((...)) expansion is not used (see Arithmetic Expansion below). Word splitting and pathname expansion are not performed. Assignment statements may also appear as arguments to the alias, declare, typeset, export, readonly, and local builtin commands (declaration commands). When in posix mode, these builtins may appear in a command after one or more instances of the command builtin and retain these assignment statement properties.

In the context where an assignment statement is assigning a value to a shell variable or array index, the += operator appends to or adds to the variable's previous value. This includes arguments to declaration commands such as declare that accept assignment statements. When += is applied to a variable for which the integer attribute has been set, the variable's current value and value are each evaluated as arithmetic expressions, and the sum of the results is assigned as the variable's value. The current value is usually an integer constant, but may be an expression. When += is applied to an array variable using compound assignment (see Arrays below), the variable's value is not unset (as it is when using =), and new values are appended to the array beginning at one greater than the array's maximum index (for indexed arrays) or added as additional key-value pairs in an associative array. When applied to a string-valued variable, value is expanded and appended to the variable's value.

A variable can be assigned the nameref attribute using the -n option to the declare or local builtin commands (see the descriptions of declare and local below) to create a nameref, or a reference to another variable. This allows variables to be manipulated indirectly. Whenever the nameref variable is referenced, assigned to, unset, or has its attributes modified (other than using or changing the nameref attribute itself), the operation is actually performed on the variable specified by the nameref variable's value. A nameref is commonly used within shell functions to refer to a variable whose name is passed as an argument to the function. For instance, if a variable name is passed to a shell function as its first argument, running

declare -n ref=$1

inside the function creates a local nameref variable ref whose value is the variable name passed as the first argument. References and assignments to ref, and changes to its attributes, are treated as references, assignments, and attribute modifications to the variable whose name was passed as $1. If the control variable in a for loop has the nameref attribute, the list of words can be a list of shell variables, and a name reference is established for each word in the list, in turn, when the loop is executed. Array variables cannot be given the nameref attribute. However, nameref variables can reference array variables and subscripted array variables. Namerefs can be unset using the -n option to the unset builtin. Otherwise, if unset is executed with the name of a nameref variable as an argument, the variable referenced by the nameref variable is unset.

When the shell starts, it reads its environment and creates a shell variable from each environment variable that has a valid name, as described below (see ENVIRONMENT).

📍 Positional Parameters 📍

A positional parameter is a parameter denoted by one or more digits, other than the single digit 0. Positional parameters are assigned from the shell's arguments when it is invoked, and may be reassigned using the set builtin command. Positional parameters may not be assigned to with assignment statements. The positional parameters are temporarily replaced when a shell function is executed (see FUNCTIONS below).

When a positional parameter consisting of more than a single digit is expanded, it must be enclosed in braces (see EXPANSION below). Without braces, a digit following $ can only refer to one of the first nine positional parameters ($1-$9) or the special parameter $0 (see the next section).

✨ Special Parameters ✨

The shell treats several parameters specially. These parameters may only be referenced; assignment to them is not allowed. Special parameters are denoted by one of the following characters.

* ($*) — bestie this one expands to all the positional parameters, starting from one 💫. When it's not chilling inside double quotes, every positional parameter expands into its own separate word. And in contexts where word expansions happen, those words get hit with further word splitting and pathname expansion (yeah, the whole little gauntlet 🙈). But when the expansion happens inside double quotes? It collapses into a single word where each parameter's value is separated by the first character of the IFS variable. So like, "$*" is equivalent to "$1c$2c...", where c is that first character of IFS's value. If IFS is unset, the params get separated by spaces. And if IFS is null? They're just joined together with no separators at all, no cap 🫶.
@ ($@) — also expands to the positional parameters, starting from one ✨. In contexts where word splitting is performed, it expands each positional parameter to a separate word; and if it's not inside double quotes, those words are subject to word splitting. In contexts where word splitting is not performed — like the value portion of an assignment statement — it expands to a single word with each positional parameter separated by a space. When the expansion occurs within double quotes and word splitting is performed, each parameter expands to its own separate word. So "$@" is equivalent to "$1" "$2" .... Now here's the cute little detail: if the double-quoted expansion happens within a word, the expansion of the first parameter gets joined with the expansion of the beginning part of the original word, and the expansion of the last parameter gets joined with the expansion of the last part of the original word. And when there are no positional parameters at all? "$@" and $@ expand to nothing — like, they're just removed. Gone. Poof 💀.
# ($#) — expands to the number of positional parameters, in decimal 🔢. That's it, that's the whole vibe.
? ($?) — expands to the exit status of the most recently executed command 👀. Your little success/fail receipt, fr fr.
- ($-) — expands to the current option flags, as specified upon invocation, by the set builtin command, or those set by the shell itself (like the -i option) 🎀.
$ ($$) — expands to the process ID of the shell 💅. And yes, in a subshell it expands to the process ID of the parent shell, not the subshell. Sneaky but iconic.
! ($!) — expands to the process ID of the job most recently placed into the background, whether that happened as an asynchronous command or by using the bg builtin (peep JOB CONTROL below, iykyk) 🌈.
0 ($0) — expands to the name of the shell or shell script 🦋. This gets set at shell initialization. If bash is invoked with a file of commands, $0 is set to the name of that file. If bash is started with the -c option, then $0 is set to the first argument after the string to be executed — if one is present. Otherwise? It's set to the filename used to invoke bash, as given by argument zero. Main character origin story, honestly 👑.

Shell Variables 🌸

The shell sets following variables:

_ — ($_, an underscore) bestie this one is a total shapeshifter 💅 it has a whole bunch of meanings depending on context. At shell startup, _ is set to the pathname used to invoke the shell or shell script being executed, as passed in the environment or argument list. After that? It expands to the last argument of the previous simple command executed in the foreground, after expansion. It's also set to the full pathname used to invoke each command executed, and it gets placed in the environment exported to that command. And when checking mail, $_ expands to the name of the mail file currently being checked. 📬
BASH — Expands to the full filename used to invoke this instance of bash. ✨
BASHOPTS — A colon-separated list of enabled shell options. Each word in the list is a valid argument for the -s option to the shopt builtin command (peep SHELL BUILTIN COMMANDS below 👀). The options showing up in BASHOPTS are the ones reported as on by shopt. If this variable is in the environment when bash starts up, the shell enables each option in the list before reading any startup files. If this variable is exported, child shells will enable each option in the list. This variable is read-only — no touchy 🚫.
BASHPID — Expands to the process ID of the current bash process. This differs from $$ under certain circumstances, like subshells that do not require bash to be re-initialized. Assignments to BASHPID have no effect. If BASHPID is unset, it loses its special properties, even if it is subsequently reset. 😭
BASH_ALIASES — An associative array variable whose members correspond to the internal list of aliases as maintained by the alias builtin. Elements added to this array appear in the alias list; however, unsetting array elements currently does not remove aliases from the alias list. If BASH_ALIASES is unset, it loses its special properties, even if it is subsequently reset. 🎀
BASH_ARGC — An array variable whose values are the number of parameters in each frame of the current bash execution call stack. The number of parameters to the current subroutine (shell function or script executed with . or source) is at the top of the stack. When a subroutine is executed, the number of parameters passed is pushed onto BASH_ARGC. The shell sets BASH_ARGC only when in extended debugging mode (see the description of the extdebug option to the shopt builtin below). Setting extdebug after the shell has started to execute a script, or referencing this variable when extdebug is not set, may result in inconsistent values. Assignments to BASH_ARGC have no effect, and it may not be unset. 💀
BASH_ARGV — An array variable containing all of the parameters in the current bash execution call stack. The final parameter of the last subroutine call is at the top of the stack; the first parameter of the initial call is at the bottom. When a subroutine is executed, the shell pushes the supplied parameters onto BASH_ARGV. The shell sets BASH_ARGV only when in extended debugging mode (see the description of the extdebug option to the shopt builtin below). Setting extdebug after the shell has started to execute a script, or referencing this variable when extdebug is not set, may result in inconsistent values. Assignments to BASH_ARGV have no effect, and it may not be unset. 🔥
BASH_ARGV0 — When referenced, this variable expands to the name of the shell or shell script (identical to $0; see the description of special parameter 0 above). Assigning a value to BASH_ARGV0 sets $0 to the same value. If BASH_ARGV0 is unset, it loses its special properties, even if it is subsequently reset. 🦋
BASH_CMDS — An associative array variable whose members correspond to the internal hash table of commands as maintained by the hash builtin. Adding elements to this array makes them appear in the hash table; however, unsetting array elements currently does not remove command names from the hash table. If BASH_CMDS is unset, it loses its special properties, even if it is subsequently reset. 🌸
BASH_COMMAND — Expands to the command currently being executed or about to be executed, unless the shell is executing a command as the result of a trap, in which case it is the command executing at the time of the trap. If BASH_COMMAND is unset, it loses its special properties, even if it is subsequently reset. 👑
BASH_EXECUTION_STRING — The command argument to the -c invocation option. 💅
BASH_LINENO — An array variable whose members are the line numbers in source files where each corresponding member of FUNCNAME was invoked. ${BASH_LINENO[$i]} is the line number in the source file (${BASH_SOURCE[$i+1]}) where ${FUNCNAME[$i]} was called (or ${BASH_LINENO[$i-1]} if referenced within another shell function). Use LINENO to obtain the current line number. Assignments to BASH_LINENO have no effect, and it may not be unset. 😭
BASH_LOADABLES_PATH — A colon-separated list of directories in which the enable command looks for dynamically loadable builtins. 🧚
BASH_MONOSECONDS — Each time this variable is referenced, it expands to the value returned by the system's monotonic clock, if one is available. If there is no monotonic clock, this is equivalent to EPOCHSECONDS. If BASH_MONOSECONDS is unset, it loses its special properties, even if it is subsequently reset. ⏰
BASH_REMATCH — An array variable whose members are assigned by the =~ binary operator to the [[ conditional command. The element with index 0 is the portion of the string matching the entire regular expression. The element with index n is the portion of the string matching the *n*th parenthesized subexpression. 🔍
BASH_SOURCE — An array variable whose members are the source filenames where the corresponding shell function names in the FUNCNAME array variable are defined. The shell function ${FUNCNAME[$i]} is defined in the file ${BASH_SOURCE[$i]} and called from ${BASH_SOURCE[$i+1]}. Assignments to BASH_SOURCE have no effect, and it may not be unset. 📂
BASH_SUBSHELL — Incremented by one within each subshell or subshell environment when the shell begins executing in that environment. The initial value is 0. If BASH_SUBSHELL is unset, it loses its special properties, even if it is subsequently reset. 🌀
BASH_TRAPSIG — Set to the signal number corresponding to the trap action being executed during its execution. See the description of trap under SHELL BUILTIN COMMANDS below for information about signal numbers and trap execution. 🚨
BASH_VERSINFO — A readonly array variable whose members hold version information for this instance of bash. The values assigned to the array members are:
- BASH_VERSINFO[0] — The major version number (the release).
- BASH_VERSINFO[1] — The minor version number (the version).
- BASH_VERSINFO[2] — The patch level.
- BASH_VERSINFO[3] — The build version.
- BASH_VERSINFO[4] — The release status (e.g., beta).
- BASH_VERSINFO[5] — The value of MACHTYPE. ✨ iconic ✨
BASH_VERSION — Expands to a string describing the version of this instance of bash (e.g., 5.2.37(3)-release). 🌈
COMP_CWORD — An index into ${COMP_WORDS} of the word containing the current cursor position. This variable is available only in shell functions invoked by the programmable completion facilities (see Programmable Completion below). 🎯
COMP_KEY — The key (or final key of a key sequence) used to invoke the current completion function. This variable is available only in shell functions and external commands invoked by the programmable completion facilities (see Programmable Completion below). 🔑
COMP_LINE — The current command line. This variable is available only in shell functions and external commands invoked by the programmable completion facilities (see Programmable Completion below). 📝
COMP_POINT — The index of the current cursor position relative to the beginning of the current command. If the current cursor position is at the end of the current command, the value of this variable is equal to ${#COMP_LINE}. This variable is available only in shell functions and external commands invoked by the programmable completion facilities (see Programmable Completion below). 📍
COMP_TYPE — Set to an integer value corresponding to the type of attempted completion that caused a completion function to be called: TAB, for normal completion, ?, for listing completions after successive tabs, !, for listing alternatives on partial word completion, @, to list completions if the word is not unmodified, or %, for menu completion. This variable is available only in shell functions and external commands invoked by the programmable completion facilities (see Programmable Completion below). 🎛️
COMP_WORDBREAKS — The set of characters that the readline library treats as word separators when performing word completion. If COMP_WORDBREAKS is unset, it loses its special properties, even if it is subsequently reset. ✂️
COMP_WORDS — An array variable (see Arrays below) consisting of the individual words in the current command line. The line is split into words as readline would split it, using COMP_WORDBREAKS as described above. This variable is available only in shell functions invoked by the programmable completion facilities (see Programmable Completion below). 🧩
COPROC — An array variable (see Arrays below) created to hold the file descriptors for output from and input to an unnamed coprocess (see Coprocesses above). 🎭
DIRSTACK — An array variable (see Arrays below) containing the current contents of the directory stack. Directories appear in the stack in the order they are displayed by the dirs builtin. Assigning to members of this array variable may be used to modify directories already in the stack, but the pushd and popd builtins must be used to add and remove directories. Assigning to this

…so setting that variable does not change your current directory, bestie, it's just a list. And if DIRSTACK gets unset, it loses all its special properties — even if you set it again later. 🙈

EPOCHREALTIME — every single time you reference this one, it expands to the number of seconds since the Unix Epoch (peep time(3)) as a floating-point value with micro-second granularity. Assignments to EPOCHREALTIME? Ignored. Nope. And if EPOCHREALTIME is unset, it loses its special properties, even if it is subsequently reset. 💅
EPOCHSECONDS — every time it's referenced, it expands to the number of seconds since the Unix Epoch (again, see time(3)). Assignments to EPOCHSECONDS are ignored. If EPOCHSECONDS is unset, it loses its special properties, even if it is subsequently reset. ✨
EUID — expands to the effective user ID of the current user, initialized at shell startup. This variable is readonly. Don't even try it. 🔒
FUNCNAME — an array variable holding the names of all shell functions currently chilling in the execution call stack. The element with index 0 is the name of any currently-executing shell function, and the bottom-most element (the one with the highest index) is main. This variable exists only when a shell function is executing. Assignments to FUNCNAME have no effect. If FUNCNAME is unset, it loses its special properties, even if it is subsequently reset.

You can use this variable with BASH_LINENO and BASH_SOURCE. Each element of FUNCNAME has corresponding elements in BASH_LINENO and BASH_SOURCE to describe the call stack. For instance, ${FUNCNAME[$i]} was called from the file ${BASH_SOURCE[$i+1]} at line number ${BASH_LINENO[$i]}. The caller builtin displays the current call stack using this information. 🫶

GROUPS — an array variable containing the list of groups of which the current user is a member. Assignments to GROUPS have no effect. If GROUPS is unset, it loses its special properties, even if it is subsequently reset. 👑
HISTCMD — the history number, aka the index in the history list, of the current command. Assignments to HISTCMD have no effect. If HISTCMD is unset, it loses its special properties, even if it is subsequently reset. 📜
HOSTNAME — automatically set to the name of the current host. It just knows. 🖥️
HOSTTYPE — automatically set to a string that uniquely describes the type of machine on which bash is executing. The default is system-dependent. 🌈
LINENO — each time this parameter is referenced, the shell substitutes a decimal number representing the current sequential line number (starting with 1) within a script or function. When you're not in a script or function, the value substituted is not guaranteed to be meaningful. If LINENO is unset, it loses its special properties, even if it is subsequently reset. 🔢
MACHTYPE — automatically set to a string that fully describes the system type on which bash is executing, in the standard GNU cpu-company-system format. The default is system-dependent. ✨
MAPFILE — an array variable (see Arrays below) created to hold the text read by the mapfile builtin when no variable name is supplied. 🗂️
OLDPWD — the previous working directory as set by the cd command. Your lil "where was I" bookmark. 📍
OPTARG — the value of the last option argument processed by the getopts builtin command (see SHELL BUILTIN COMMANDS below). 🎀
OPTIND — the index of the next argument to be processed by the getopts builtin command (see SHELL BUILTIN COMMANDS below). 🎀
OSTYPE — automatically set to a string that describes the operating system on which bash is executing. The default is system-dependent. 🦋
PIPESTATUS — an array variable (see Arrays below) containing a list of exit status values from the commands in the most-recently-executed foreground pipeline, which may consist of only a simple command (see SHELL GRAMMAR above). bash sets PIPESTATUS after executing multi-element pipelines, timed and negated pipelines, simple commands, subshells created with the ( operator, the [[ and (( compound commands, and after error conditions that result in the shell aborting command execution. So many moments. 🔥
PPID — the process ID of the shell's parent. This variable is readonly. 🚫
PWD — the current working directory as set by the cd command. 📂
RANDOM — each time this parameter is referenced, it expands to a random integer between 0 and 32767. Assigning a value to RANDOM initializes (seeds) the sequence of random numbers. Seeding the random number generator with the same constant value produces the same sequence of values. If RANDOM is unset, it loses its special properties, even if it is subsequently reset. 🎲
READLINE_ARGUMENT — any numeric argument given to a readline command that was defined using bind -x (see SHELL BUILTIN COMMANDS below) when it was invoked. ⌨️
READLINE_LINE — the contents of the readline line buffer, for use with bind -x (see SHELL BUILTIN COMMANDS below). ⌨️
READLINE_MARK — the position of the mark (saved insertion point) in the readline line buffer, for use with bind -x (see SHELL BUILTIN COMMANDS below). The characters between the insertion point and the mark are often called the region. 🎯
READLINE_POINT — the position of the insertion point in the readline line buffer, for use with bind -x (see SHELL BUILTIN COMMANDS below). ⌨️
REPLY — set to the line of input read by the read builtin command when no arguments are supplied. 💬
SECONDS — each time this parameter is referenced, it expands to the number of seconds since shell invocation. If a value is assigned to SECONDS, the value returned upon subsequent references is the number of seconds since the assignment plus the value assigned. The number of seconds at shell invocation and the current time are always determined by querying the system clock at one-second resolution. If SECONDS is unset, it loses its special properties, even if it is subsequently reset. ⏱️
SHELLOPTS — a colon-separated list of enabled shell options. Each word in the list is a valid argument for the -o option to the set builtin command (see SHELL BUILTIN COMMANDS below). The options appearing in SHELLOPTS are those reported as on by set -o. If this variable is in the environment when bash starts up, the shell enables each option in the list before reading any startup files. If this variable is exported, child shells will enable each option in the list. This variable is read-only. 🔒
SHLVL — incremented by one each time an instance of bash is started. Nesting, but make it official. 🪆
SRANDOM — each time it is referenced, this variable expands to a 32-bit pseudo-random number. The random number generator is not linear on systems that support /dev/urandom or arc4random(3), so each returned number has no relationship to the numbers preceding it. The random number generator cannot be seeded, so assignments to this variable have no effect. If SRANDOM is unset, it loses its special properties, even if it is subsequently reset. 🎲✨
UID — expands to the user ID of the current user, initialized at shell startup. This variable is readonly. 👤

The shell uses the following variables. In some cases, bash assigns a default value to a variable; these cases are noted below. 💖

`BASH_COMPAT` ✨

okay so this one sets the shell's compatibility level 💅. wanna know what all the levels actually do? scroll down to SHELL COMPATIBILITY MODE below, bestie, it's all laid out there.

the value can be a decimal number (like 4.2) or an integer (like 42) matching whatever compatibility level you're chasing 🎀.

if BASH_COMPAT is unset or set to the empty string → compatibility level goes to the default for the current version
if you set it to a value that is not one of the valid compatibility levels → the shell prints an error message and sets the compatibility level to the default for the current version (it's giving "nice try sweetie" 😭)

a subset of the valid values correspond to the compatibility levels described below under SHELL COMPATIBILITY MODE. for example, 4.2 and 42 are valid values that correspond to the compat42 shopt option and set the compatibility level to 42. the current version is also a valid value 🫶.

`BASH_ENV` 🌸

if this parameter is set when bash is executing a shell script, its expanded value is interpreted as a filename containing commands to initialize the shell before it reads and executes commands from the script ✨.

the value of BASH_ENV is subjected to parameter expansion, command substitution, and arithmetic expansion before being interpreted as a filename. and ngl, PATH is not used to search for the resultant filename 👑

`BASH_XTRACEFD` 💜

set this to an integer corresponding to a valid file descriptor and bash will write the trace output generated when set -x is enabled to that file descriptor, instead of the standard error 🔥.

the file descriptor is closed when BASH_XTRACEFD is unset or assigned a new value. unsetting BASH_XTRACEFD or assigning it the empty string causes the trace output to be sent to the standard error.

⚠️ heads up: setting BASH_XTRACEFD to 2 (the standard error file descriptor) and then unsetting it will result in the standard error being closed. not the vibe you want, bestie 💀

`CDPATH` 🧚

the search path for the cd command. it's a colon-separated list of directories where the shell looks for directories specified as arguments to the cd command. a sample value is .:~:/usr 💖

`CHILD_MAX` ⭐

sets the number of exited child status values for the shell to remember.

bash will not allow this value to be decreased below a POSIX-mandated minimum, and there is a maximum value (currently 8192) that this may not exceed. the minimum value is system-dependent 🎀

`COLUMNS` 🌈

used by the select compound command to determine the terminal width when printing selection lists.

automatically set if the checkwinsize option is enabled or in an interactive shell upon receipt of a SIGWINCH ✨

`COMPREPLY` 🍓

an array variable from which bash reads the possible completions generated by a shell function invoked by the programmable completion facility (see Programmable Completion below).

each array element contains one possible completion 💅

`EMACS` 🦋

if bash finds this variable in the environment when the shell starts with value t, it assumes that the shell is running in an Emacs shell buffer and disables line editing 😌

`ENV` 👑

expanded and executed similarly to BASH_ENV (see INVOCATION above) when an interactive shell is invoked in posix mode.

`EXECIGNORE` 💖

a colon-separated list of shell patterns (see Pattern Matching) defining the set of filenames to be ignored by command search using PATH.

files whose full pathnames match one of these patterns are not considered executable files for the purposes of completion and command execution via PATH lookup 🫶.

some deets, no cap:

this does not affect the behavior of the [, test, and [[ commands
full pathnames in the command hash table are not subject to EXECIGNORE
use this variable to ignore shared library files that have the executable bit set, but are not executable files 🎀

the pattern matching honors the setting of the extglob shell option ✨

`FCEDIT` 💜

the default editor for the fc builtin command.

`FIGNORE` 🌸

a colon-separated list of suffixes to ignore when performing filename completion (see READLINE below).

a filename whose suffix matches one of the entries in FIGNORE is excluded from the list of matched filenames 📝.

a sample value is ".o:~" — and fr fr, quoting is needed when assigning a value to this variable, which contains tildes 🙈

`FUNCNEST` 🔥

if set to a numeric value greater than 0, defines a maximum function nesting level.

function invocations that exceed this nesting level cause the current command to abort 💀

`GLOBIGNORE` 🎀

a colon-separated list of patterns defining the set of file names to be ignored by pathname expansion.

if a file name matched by a pathname expansion pattern also matches one of the patterns in GLOBIGNORE, it is removed from the list of matches ✨

the pattern matching honors the setting of the extglob shell option 👑

`GLOBSORT` 🌈

controls how the results of pathname expansion are sorted. the value of this variable specifies the sort criteria and sort order for the results of pathname expansion 🌸.

if this variable is unset or set to the null string, pathname expansion uses the historical behavior of sorting by name, in ascending lexicographic order as determined by the LC_COLLATE shell variable 💅.

if set, a valid value begins with an optional +, which is ignored, or -, which reverses the sort order from ascending to descending, followed by a sort specifier 🦋.

the valid sort specifiers are:

name — sorts the files on name
numeric — sorts names in numeric rather than lexicographic order
size — sorts on file size
mtime — sorts on modification time
atime — sorts on access time
ctime — sorts on inode change time
blocks — sorts on number of blocks

if any of the non-name keys compare as equal (e.g., if two files are the same size), sorting uses the name as a secondary sort key ⭐

for example, a value of -mtime sorts the results in descending order by modification time (newest first) 🔥

the numeric specifier treats names consisting solely of digits as numbers and sorts them using their numeric value (so 2 sorts before 10, for example) 🍓. when using numeric, names containing non-digits sort after all the all-digit names and are sorted by name using the traditional behavior.

a sort specifier of nosort disables sorting completely; bash returns the results in the order they are read from the file system, ignoring any leading - 💀

…and picking up exactly where we left off: if the sort specifier is missing, it defaults to name, so a value of + is equivalent to the null string, and a value of - sorts by name in descending order. Any invalid value restores the historical sorting behavior. 💅

HISTCONTROL — bestie this one is a colon-separated list of values that controls how commands get saved onto the history list. Let's break it down:
- ignorespace → lines that start with a space character are not saved in the history list.
- ignoredups → lines that match the previous history entry don't get saved.
- ignoreboth → shorthand for ignorespace and ignoredups together.
- erasedups → causes all previous lines matching the current line to be removed from the history list before that line is saved.

Any value that's not in that list is just ignored, no drama. If HISTCONTROL is unset, or doesn't include a valid value, bash saves all lines read by the shell parser on the history list, subject to the value of HISTIGNORE. 🫶

And here's the multi-line tea: if the first line of a multi-line compound command was saved, the second and subsequent lines are not tested, and are added to the history regardless of the value of HISTCONTROL. If the first line was not saved, the second and subsequent lines of the command are not saved either. Periodt.

HISTFILE — the name of the file where command history is saved (see HISTORY below). Bash assigns a default value of ~/.bash_history. If HISTFILE is unset or null, the shell does not save the command history when it exits. 😭
HISTFILESIZE — the maximum number of lines contained in the history file. When you assign this variable a value, the history file gets truncated, if necessary, to contain no more than the number of history entries that total no more than that number of lines — by removing the oldest entries. If the history list contains multi-line entries, the history file may contain more lines than this maximum, to avoid leaving partial history entries. The history file is also truncated to this size after writing it when a shell exits or by the history builtin. If the value is 0, the history file is truncated to zero size. Non-numeric values and numeric values less than zero inhibit truncation. The shell sets the default value to the value of HISTSIZE after reading any startup files. ✨
HISTIGNORE — a colon-separated list of patterns used to decide which command lines should be saved on the history list. If a command line matches one of the patterns in the value of HISTIGNORE, it is not saved on the history list. Heads up on the deets:
- Each pattern is anchored at the beginning of the line and must match the complete line (bash does not implicitly append a *).
- Each pattern is tested against the line after the checks specified by HISTCONTROL are applied.
- In addition to the normal shell pattern matching characters, & matches the previous history line. A backslash escapes the &; the backslash is removed before attempting a match.
- If the first line of a multi-line compound command was saved, the second and subsequent lines are not tested, and are added to the history regardless of the value of HISTIGNORE. If the first line was not saved, the second and subsequent lines of the command are not saved either.
- The pattern matching honors the setting of the extglob shell option.

HISTIGNORE subsumes some of the function of HISTCONTROL. A pattern of & is identical to ignoredups, and a pattern of [ ]* is identical to ignorespace. Combine those two patterns, separating them with a colon, and boom — you've got the functionality of ignoreboth. 🎀

HISTSIZE — the number of commands to remember in the command history (see HISTORY below). If the value is 0, commands are not saved in the history list. Numeric values less than zero result in every command being saved on the history list (there is no limit, go feral 🔥). The shell sets the default value to 500 after reading any startup files.
HISTTIMEFORMAT — if this variable is set and not null, its value is used as a format string for strftime(3) to print the time stamp associated with each history entry displayed by the history builtin. If this variable is set, the shell writes time stamps to the history file so they may be preserved across shell sessions. This uses the history comment character to distinguish timestamps from other history lines. ⭐
HOME — the home directory of the current user; the default argument for the cd builtin command. The value of this variable is also used when performing tilde expansion. 🏠
HOSTFILE — contains the name of a file in the same format as /etc/hosts that should be read when the shell needs to complete a hostname. The list of possible hostname completions may be changed while the shell is running; the next time hostname completion is attempted after the value is changed, bash adds the contents of the new file to the existing list. If HOSTFILE is set, but has no value, or does not name a readable file, bash attempts to read /etc/hosts to obtain the list of possible hostname completions. When HOSTFILE is unset, bash clears the hostname list. 🌈
IFS — the Internal Field Separator, used for word splitting after expansion and to split lines into words with the read builtin command. Word splitting is described below under EXPANSION. The default value is "<space><tab><newline>". 💖
IGNOREEOF — controls the action of an interactive shell on receipt of an EOF character as the sole input. If set, the value is the number of consecutive EOF characters which must be typed as the first characters on an input line before bash exits. If the variable is set but does not have a numeric value, or the value is null, the default value is 10. If it is unset, EOF signifies the end of input to the shell. 🦋
INPUTRC — the filename for the readline startup file, overriding the default of ~/.inputrc (see READLINE below).
INSIDE_EMACS — if this variable appears in the environment when the shell starts, bash assumes that it is running inside an Emacs shell buffer and may disable line editing, depending on the value of TERM.
LANG — used to determine the locale category for any category not specifically selected with a variable starting with LC_.
LC_ALL — this variable overrides the value of LANG and any other LC_ variable specifying a locale category. 👑
LC_COLLATE — this variable determines the collation order used when sorting the results of pathname expansion, and determines the behavior of range expressions, equivalence classes, and collating sequences within pathname expansion and pattern matching.
LC_CTYPE — this variable determines the interpretation of characters and the behavior of character classes within pathname expansion and pattern matching.
LC_MESSAGES — this variable determines the locale used to translate double-quoted strings preceded by a $.
LC_NUMERIC — this variable determines the locale category used for number formatting.
LC_TIME — this variable determines the locale category used for data and time formatting. 🕰️
LINES — used by the select compound command to determine the column length for printing selection lists. Automatically set if the checkwinsize option is enabled or in an interactive shell upon receipt of a SIGWINCH.
MAIL — if the value is set to a file or directory name and the MAILPATH variable is not set, bash informs the user of the arrival of mail in the specified file or Maildir-format directory. 📬
MAILCHECK — specifies how often (in seconds) bash checks for mail. The default is 60 seconds. When it is time to check for mail, the shell does so before displaying the primary prompt. If this variable is unset, or set to a value that is not a number greater than or equal to zero, the shell disables mail checking.
MAILPATH — a colon-separated list of filenames to be checked for mail. The message to be printed when mail arrives in a particular file may be specified by separating the filename from the message with a ?. When used in the text of the message, $_ expands to the name of the current mailfile. For example:

MAILPATH='/var/mail/bfox?"You have mail":~/shell-mail?"$_ has mail!"'

Bash can be configured to supply a default value for this variable (there is no value by default), but the location of the user mail files that it uses is system dependent (e.g., /var/mail/$USER). 🍓

OPTERR — if set to the value 1, bash displays error messages generated by the getopts builtin command (see SHELL BUILTIN COMMANDS below). OPTERR is initialized to 1 each time the shell is invoked or a shell script is executed.
PATH — the search path for commands. It is a colon-separated list of directories in which the shell looks for commands (see COMMAND EXECUTION below). A zero-length (null) directory name in the value of PATH indicates the current directory. A null directory name may appear as two adjacent colons, or as an initial or trailing colon. The default path is system-dependent, and is set by the administrator who installs bash. A common value is

/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin:/bin:/sbin

POSIXLY_CORRECT — if this variable is in the environment when bash starts, the shell enters posix mode before reading the startup files, as if the --posix invocation option had been supplied. If it is set while the shell is running, bash enables posix mode, as if the command set -o posix had been executed. When the shell enters posix mode, it sets this variable if it was not already set.
PROMPT_COMMAND — if this variable is set, and is an array, the value of each set element is executed as a command prior to issuing each primary prompt. If this is set but not an array variable, its value is used as a command to execute instead. 💜
PROMPT_DIRTRIM

If it's set to a number bigger than zero, that number is how many trailing directory components bash keeps around when it expands the \w and \W prompt string escapes (peep PROMPTING below for the deets on those). Whatever gets trimmed off? Replaced with an ellipsis. ✂️

PS0 — this one gets expanded (again, PROMPTING below) and then displayed by interactive shells after they've read a command but before that command actually runs. A lil pre-show moment. ✨
PS1 — expanded (PROMPTING below) and used as the primary prompt string. The default is \s-\v\$ . That's the one you stare at all day, bestie, she's the main character. 👑
PS2 — expanded exactly like PS1 is, and used as the secondary prompt string. Default is > . You know her — she shows up when you leave a quote hanging. 😭
PS3 — its value is the prompt for the select command (SHELL GRAMMAR above has the lore). 🌸
PS4 — expanded same way as PS1, and the value gets printed before every command bash displays during an execution trace. The first character of the expanded PS4 is replicated a bunch of times, as many as needed, to show you multiple levels of indirection. Default is + . 🔥
SHELL — expands to the full pathname to the shell. And if it's not already set when the shell starts up, bash goes ahead and assigns it the full pathname of the current user's login shell. She fills in the blanks herself, no cap. 💅
TIMEFORMAT — this value is a format string that decides how the timing info for pipelines prefixed with the time reserved word gets displayed. The % character kicks off an escape sequence that expands to a time value or some other info. Here's the whole menu, and btw the brackets mean those portions are optional:
- %% — a literal %.
- %[p][l]R — the elapsed time in seconds.
- %[p][l]U — the number of CPU seconds spent in user mode.
- %[p][l]S — the number of CPU seconds spent in system mode.
- %P — the CPU percentage, computed as (%U + %S) / %R.

That optional p is a digit that sets the precision — like, how many fractional digits come after the decimal point. If you set it to 0, no decimal point and no fraction get output at all. time prints at most six digits after the decimal point, so if you hand it a p greater than 6, it gets changed to 6. And if you don't specify p at all, time prints three digits after the decimal point.

The optional l gives you a longer format that includes minutes, shaped like MMmSS.FFs. Whether the fraction shows up depends on the value of p.

If this variable isn't set, bash acts like it had the value $'\nreal\t%3lR\nuser\t%3lU\nsys\t%3lS'. If the value is null, bash straight up doesn't display any timing information at all. And a trailing newline gets added whenever the format string is displayed. 🕐

TMOUT — if you set it to a value greater than zero, the read builtin uses that value as its default timeout. The select command will terminate if input doesn't arrive after TMOUT seconds, but only when the input is coming from a terminal. In an interactive shell, the value is interpreted as the number of seconds to wait for a line of input after the primary prompt has been issued. Bash terminates after waiting that many seconds if a complete line of input doesn't show up. ⏰
TMPDIR — if it's set, bash uses its value as the name of a directory where bash creates temporary files for the shell's own use. 🗂️
auto_resume — this variable controls how the shell interacts with the user and with job control. If it's set, simple commands that consist of only a single word, with no redirections, get treated as candidates for resuming an existing stopped job. There is no ambiguity allowed here — if there's more than one job beginning with or containing that word, it selects the most recently accessed job. The name of a stopped job, in this context, is the command line that was used to start it, as displayed by jobs. If you set it to the value exact, the word has to match the name of a stopped job exactly; if you set it to substring, the word just needs to match a substring of a stopped job's name. That substring value gives you functionality analogous to the %? job identifier (JOB CONTROL below has more). And if you set it to any other value (like, say, prefix), the word has to be a prefix of a stopped job's name — which gives you functionality analogous to the %string job identifier. 🔍
histchars — the two or three characters that control history expansion, quick substitution, and tokenization (HISTORY EXPANSION below, bestie). The first character is the history expansion character — the one that begins a history expansion, normally !. The second character is the quick substitution character, normally ^. When it shows up as the first character on the line, history substitution repeats the previous command, swapping one string out for another. The optional third character is the history comment character, normally #, and it signals that the rest of the line is a comment when it appears as the first character of a word. The history comment character disables history substitution for the remaining words on the line — but heads up, it does not necessarily make the shell parser treat the rest of the line as a comment. 🙈

Arrays 🦋

Bash gives you one-dimensional indexed arrays and associative arrays, periodt.

Any variable can be used as an indexed array, and the declare builtin is how you explicitly declare one. There's no maximum limit on how big an array can be, and no rule saying the members have to be indexed or assigned contiguously — she's flexible like that. 💅

Indexed arrays are referenced using arithmetic expressions that must expand to an integer (ARITHMETIC EVALUATION below), and they're zero-based. Associative arrays are referenced using arbitrary strings. And unless it says otherwise somewhere, indexed array indices have to be non-negative integers.

The shell performs parameter and variable expansion, arithmetic expansion, command substitution, and quote removal on indexed array subscripts. Since that can potentially leave you with empty strings, subscript indexing treats those as expressions that evaluate to 0. 🫶

The shell performs tilde expansion, parameter and variable expansion, arithmetic expansion, command substitution, and quote removal on associative array subscripts. And nope — empty strings cannot be used as associative array keys. 🚫

Bash automatically creates an indexed array if any variable gets assigned to using this syntax:

name[subscript]=value .

The subscript is treated as an arithmetic expression that must evaluate to a number greater than or equal to zero. If you want to explicitly declare an indexed array, use:

declare -a name

(see SHELL BUILTIN COMMANDS below). This next one is also accepted:

declare -a name[subscript]

— the subscript just gets ignored. 😌

Associative arrays are created using:

declare -A name

Attributes can be specified for an array variable using the declare and readonly builtins, and here's the thing: each attribute applies to all members of the array. All of them. The whole squad. 👑

Arrays get assigned using compound assignments shaped like name=(value1 ... valuen), where each value may be of the form [subscript]=string. Indexed array assignments don't need anything but string. Every value in the list is expanded using the shell expansions described below under EXPANSION — but values that are valid variable assignments, brackets and subscript included, do not undergo brace expansion and word splitting, same as with individual variable assignments.

When you're assigning to indexed arrays, if the optional brackets and subscript are supplied, that index is what gets assigned to; otherwise the index of the element being assigned is the last index assigned to by the statement plus one. Indexing starts at zero. 🍩

When you're assigning to an associative array, the words in a compound assignment can be either assignment statements — for which the subscript is required — or a list of words that gets interpreted as a sequence of alternating keys and values: name=( key1 value1 key2 value2 ...). These two are treated identically:

name=( [key1]=value1 [key2]=value2 ...)

The first word in the list decides how the remaining words get interpreted, and every assignment in a list has to be the same type. When you're doing key/value pairs, the keys can't be missing or empty — and a final missing value gets treated like the empty string.

This syntax is also accepted by the declare builtin. Individual array elements can be assigned to using the name[subscript]=value syntax we just introduced above. ✨

When assigning to an indexed array, if name is subscripted by a negative number, that number is interpreted as relative to one greater than the maximum index of name. So negative indices count back from the end of the array, and an index of -1 references the last element. 🌈

The += operator appends to an array variable when you're assigning using the compound assignment syntax — see PARAMETERS above.

An array element is referenced using ${name[subscript]}. Those braces are required, to avoid conflicts with pathname expansion. If subscript is @ or *, the word expands to all members of name, unless it's noted otherwise in the description of a builtin or word expansion.

These two subscripts differ only when the word appears within double quotes:

If the word is double-quoted, ${name[*]} expands to a single word with the value of each array member separated by the first character of the IFS special variable.
${name[@]} expands each element of name to a separate word.

When there are no array members, ${name[@]} expands to nothing. And if the double-quoted expansion happens within a word, the expansion of the first parameter gets joined with the beginning part of the expansion of the original word, and the expansion of the last parameter gets joined with the last part of the expansion of the original word. This is analogous to the expansion of the special parameters * and @ (see Special Parameters above). 💖

${#name[subscript]} expands to the length of ${name[subscript]}. And if subscript is * or @, the expansion is the number of elements in the array. 🧮

If the subscript used to reference an element of an indexed array evaluates to a number less than zero, it gets interpreted as relative to one greater than the maximum index of the array — so negative indices count back from the end of the array, and an index of -1 references the last element.

Referencing an array variable without a subscript is equivalent to referencing the array with a subscript of 0. And any reference to a variable using a valid subscript is valid; bash creates an array if it needs one. No cap. 🎀

An array variable counts as set if a subscript has been assigned a value — and bestie, the null string is a totally valid value, don't let anyone tell you otherwise. 💅

It's also possible to grab the keys (indices) of an array and not just the values. ${!name[@]} and ${!name[*]} expand to the indices assigned in array variable name. The treatment when you're in double quotes is similar to how the special parameters @ and * expand inside double quotes.

The unset builtin is your array-destroying bestie. unset name[subscript] unsets the array element at index subscript, for both indexed and associative arrays. Negative subscripts to indexed arrays get interpreted as described above. Unsetting the last element of an array variable does not unset the variable — the array just sits there, empty, thriving. unset name, where name is an array, removes the entire array. And unset name[subscript] behaves differently depending on whether name is an indexed or associative array when subscript is * or @. If name is an associative array, this unsets the element with subscript * or @. If name is an indexed array, unset removes all of the elements but does not remove the array itself.

Real talk: when you use a variable name with a subscript as an argument to a command — like with unset — without using the word expansion syntax described above (e.g., unset a[4]), the argument is subject to pathname expansion. Quote the argument if pathname expansion is not desired (e.g., unset 'a[4]'). 🌸

The declare, local, and readonly builtins each accept a -a option to specify an indexed array and a -A option to specify an associative array. If both options are supplied, -A takes precedence. The read builtin accepts a -a option to assign a list of words read from the standard input to an array. The set and declare builtins display array values in a way that allows them to be reused as assignments. Other builtins accept array name arguments as well (e.g., mapfile); see the descriptions of individual builtins below for details. The shell provides a number of builtin array variables. 🎀

✨ EXPANSION ✨

Expansion is performed on the command line after it has been split into words. The shell performs these expansions: brace expansion, tilde expansion, parameter and variable expansion, command substitution, arithmetic expansion, word splitting, pathname expansion, and quote removal.

The order of expansions is: brace expansion; tilde expansion, parameter and variable expansion, arithmetic expansion, and command substitution (done in a left-to-right fashion); word splitting; pathname expansion; and quote removal.

On systems that can support it, there is an additional expansion available: process substitution. This is performed at the same time as tilde, parameter, variable, and arithmetic expansion and command substitution.

Quote removal is always performed last. It removes quote characters present in the original word, not ones resulting from one of the other expansions, unless they have been quoted themselves. Sneaky little guys. 🙈

Only brace expansion, word splitting, and pathname expansion can increase the number of words of the expansion; other expansions expand a single word to a single word. The only exceptions to this are the expansions of "$@" and "${name[@]}", and, in most cases, $* and ${name[*]} as explained above (see PARAMETERS). 👑

🦋 Brace Expansion 🦋

Brace expansion is a mechanism to generate arbitrary strings sharing a common prefix and suffix, either of which can be empty. This mechanism is similar to pathname expansion, but the filenames generated need not exist — it's giving imaginary-friend energy and we love that.

Patterns to be brace expanded are formed from an optional preamble, followed by either a series of comma-separated strings or a sequence expression between a pair of braces, followed by an optional postscript. The preamble is prefixed to each string contained within the braces, and the postscript is then appended to each resulting string, expanding left to right.

Brace expansions may be nested. The results of each expanded string are not sorted; brace expansion preserves left to right order. For example, a{d,c,b}e expands into “ade ace abe”.

A sequence expression takes the form x..y[..incr], where x and y are either integers or single letters, and incr, an optional increment, is an integer. When integers are supplied, the expression expands to each number between x and y, inclusive. If either x or y begins with a zero, each generated term will contain the same number of digits, zero-padding where necessary. When letters are supplied, the expression expands to each character lexicographically between x and y, inclusive, using the C locale. Note that both x and y must be of the same type (integer or letter). When the increment is supplied, it is used as the difference between each term. The default increment is 1 or -1 as appropriate.

Brace expansion is performed before any other expansions, and any characters special to other expansions are preserved in the result. It is strictly textual. Bash does not apply any syntactic interpretation to the context of the expansion or the text between the braces.

A correctly-formed brace expansion must contain unquoted opening and closing braces, and at least one unquoted comma or a valid sequence expression. Any incorrectly formed brace expansion is left unchanged. No notes. 💀

A { or , may be quoted with a backslash to prevent its being considered part of a brace expression. To avoid conflicts with parameter expansion, the string ${ is not considered eligible for brace expansion, and inhibits brace expansion until the closing }.

This construct is typically used as shorthand when the common prefix of the strings to be generated is longer than in the above example:

mkdir /usr/local/src/bash/{old,new,dist,bugs}

chown root /usr/{ucb/{ex,edit},lib/{ex?.?*,how_ex}}

Brace expansion introduces a slight incompatibility with historical versions of sh. sh does not treat opening or closing braces specially when they appear as part of a word, and preserves them in the output. Bash removes braces from words as a consequence of brace expansion. For example, a word entered to sh as “file{1,2}” appears identically in the output. Bash outputs that word as “file1 file2” after brace expansion. Start bash with the +B option or disable brace expansion with the +B option to the set command (see SHELL BUILTIN COMMANDS below) for strict sh compatibility.

🌸 Tilde Expansion 🌸

If a word begins with an unquoted tilde character (~), all of the characters preceding the first unquoted slash (or all characters, if there is no unquoted slash) are considered a tilde-prefix. If none of the characters in the tilde-prefix are quoted, the characters in the tilde-prefix following the tilde are treated as a possible login name. If this login name is the null string, the tilde is replaced with the value of the shell parameter HOME. If HOME is unset, the tilde expands to the home directory of the user executing the shell instead. Otherwise, the tilde-prefix is replaced with the home directory associated with the specified login name.

If the tilde-prefix is a ~+, the value of the shell variable PWD replaces the tilde-prefix. If the tilde-prefix is a ~-, the shell substitutes the value of the shell variable OLDPWD, if it is set. If the characters following the tilde in the tilde-prefix consist of a number N, optionally prefixed by a + or a -, the tilde-prefix is replaced with the corresponding element from the directory stack, as it would be displayed by the dirs builtin invoked with the characters following the tilde in the tilde-prefix as an argument. If the characters following the tilde in the tilde-prefix consist of a number without a leading + or -, tilde expansion assumes +.

The results of tilde expansion are treated as if they were quoted, so the replacement is not subject to word splitting and pathname expansion. Iconic behaviour honestly. ✨

If the login name is invalid, or the tilde expansion fails, the tilde-prefix is unchanged.

Bash checks each variable assignment for unquoted tilde-prefixes immediately following a : or the first =, and performs tilde expansion in these cases. Consequently, one may use filenames with tildes in assignments to PATH, MAILPATH, and CDPATH, and the shell assigns the expanded value.

Bash also performs tilde expansion on words satisfying the conditions of variable assignments (as described above under PARAMETERS) when they appear as arguments to simple commands. Bash does not do this, except for the declaration commands listed above, when in posix mode.

💜 Parameter Expansion 💜

The $ character introduces parameter expansion, command substitution, or arithmetic expansion. The parameter name or symbol to be expanded may be enclosed in braces, which are optional but serve to protect the variable to be expanded from characters immediately following it which could be interpreted as part of the name.

When braces are used, the matching ending brace is the first } not escaped by a backslash or within a quoted string, and not within an embedded arithmetic expansion, command substitution, or parameter expansion.

The basic form of parameter expansion is

${parameter}

which substitutes the value of parameter. The braces are required when parameter is a positional parameter with more than one digit, or when parameter is followed by a character which is not to be interpreted as part of its name. The parameter is a shell parameter as described above (PARAMETERS) or an array reference (Arrays).

If the first character of parameter is an exclamation point (!), and parameter is not a nameref, it introduces a level of indirection. Bash uses the value formed by expanding the rest of parameter as the new parameter; this new parameter is then expanded and that value is used in the rest of the expansion, rather than the expansion of the original parameter. This is known as indirect expansion. The value is subject to tilde expansion, parameter expansion, command substitution, and arithmetic expansion. If parameter is a nameref, this expands to the name of the parameter referenced by parameter instead of performing the complete indirect expansion, for compatibility. The exceptions to this are the expansions of ${!prefix*} and ${!name[@]} described below. The exclamation point must immediately follow the left brace in order to introduce indirection. No exceptions, bestie. 💅

In each of the cases below, word is subject to tilde expansion, parameter expansion, command substitution, and arithmetic expansion.

When not performing substring expansion, using the forms documented below (e.g., :-), bash tests for a parameter that is unset or null. Omitting the colon tests only for a parameter that is unset.

`${parameter:-word}`

Use Default Values, bestie! If parameter is unset or null, the expansion of word is substituted. Otherwise, the value of parameter is substituted. It's giving safety net energy. 💖

`${parameter:=word}`

Assign Default Values. If parameter is unset or null, the expansion of word is assigned to parameter, and the expansion is the final value of parameter. Positional parameters and special parameters may not be assigned in this way. So don't even try it. 🚫

`${parameter:?word}`

Display Error if Null or Unset. If parameter is null or unset, the shell writes the expansion of word (or a message to that effect if word is not present) to the standard error and, if it is not interactive, exits with a non-zero status. An interactive shell does not exit, but does not execute the command associated with the expansion. Otherwise, the value of parameter is substituted. So basically, it's a vibe check that can kill your script. 😱

`${parameter:+word}`

Use Alternate Value. If parameter is null or unset, nothing is substituted, otherwise the expansion of word is substituted. The value of parameter is not used. It's like a little switch. ✨

Substring Expansion

${parameter:offset}
${parameter:offset:length}

Substring Expansion. Expands to up to length characters of the value of parameter starting at the character specified by offset. If parameter is @ or *, an indexed array subscripted by @ or *, or an associative array name, the results differ as described below. If :length is omitted (the first form above), this expands to the substring of the value of parameter starting at the character specified by offset and extending to the end of the value. If offset is omitted, it is treated as 0. If length is omitted, but the colon after offset is present, it is treated as 0. length and offset are arithmetic expressions (see ARITHMETIC EVALUATION below). 🧮

If offset evaluates to a number less than zero, the value is used as an offset in characters from the end of the value of parameter. If length evaluates to a number less than zero, it is interpreted as an offset in characters from the end of the value of parameter rather than a number of characters, and the expansion is the characters between offset and that result.

⚠️ Note that a negative offset must be separated from the colon by at least one space to avoid being confused with the :- expansion.

If parameter is @ or *, the result is length positional parameters beginning at offset. A negative offset is taken relative to one greater than the greatest positional parameter, so an offset of -1 evaluates to the last positional parameter (or 0 if there are no positional parameters). It is an expansion error if length evaluates to a number less than zero. 😬

If parameter is an indexed array name subscripted by @ or *, the result is the length members of the array beginning with ${parameter[offset]}. A negative offset is taken relative to one greater than the maximum index of the specified array. It is an expansion error if length evaluates to a number less than zero. 🍓

Substring expansion applied to an associative array produces undefined results. 🙈

Substring indexing is zero-based unless the positional parameters are used, in which case the indexing starts at 1 by default. If offset is 0, and the positional parameters are used, $0 is prefixed to the list. 👑

`${!prefix*}` / `${!prefix@}`

Names matching prefix. Expands to the names of variables whose names begin with prefix, separated by the first character of the IFS special variable. When @ is used and the expansion appears within double quotes, each variable name expands to a separate word. 🦋

`${!name[@]}` / `${!name[*]}`

List of array keys. If name is an array variable, expands to the list of array indices (keys) assigned in name. If name is not an array, expands to 0 if name is set and null otherwise. When @ is used and the expansion appears within double quotes, each key expands to a separate word. 🔑

`${#parameter}`

Parameter length. Substitutes the length in characters of the expanded value of parameter. If parameter is * or @, the value substituted is the number of positional parameters. If parameter is an array name subscripted by * or @, the value substituted is the number of elements in the array. If parameter is an indexed array name subscripted by a negative number, that number is interpreted as relative to one greater than the maximum index of parameter, so negative indices count back from the end of the array, and an index of -1 references the last element. It's giving measuring tape energy. 📏

Remove Matching Prefix Pattern

${parameter#word}
${parameter##word}

Remove matching prefix pattern. The word is expanded to produce a pattern just as in pathname expansion, and matched against the expanded value of parameter using the rules described under Pattern Matching below. If the pattern matches the beginning of the value of parameter, then the result of the expansion is the expanded value of parameter with the shortest matching pattern (the # case) or the longest matching pattern (the ## case) deleted. If parameter is @ or *, the pattern removal operation is applied to each positional parameter in turn, and the expansion is the resultant list. If parameter is an array variable subscripted with @ or *, the pattern removal operation is applied to each member of the array in turn, and the expansion is the resultant list. ✂️

Remove Matching Suffix Pattern

${parameter%word}
${parameter%%word}

Remove matching suffix pattern. The word is expanded to produce a pattern just as in pathname expansion, and matched against the expanded value of parameter using the rules described under Pattern Matching below. If the pattern matches a trailing portion of the expanded value of parameter, then the result of the expansion is the expanded value of parameter with the shortest matching pattern (the % case) or the longest matching pattern (the %% case) deleted. If parameter is @ or *, the pattern removal operation is applied to each positional parameter in turn, and the expansion is the resultant list. If parameter is an array variable subscripted with @ or *, the pattern removal operation is applied to each member of the array in turn, and the expansion is the resultant list. 🎀

Pattern Substitution

${parameter/pattern/string}
${parameter//pattern/string}
${parameter/#pattern/string}
${parameter/%pattern/string}

Pattern substitution. The pattern is expanded to produce a pattern and matched against the expanded value of parameter as described under Pattern Matching below. The longest match of pattern in the expanded value is replaced with string. string undergoes tilde expansion, parameter and variable expansion, arithmetic expansion, command and process substitution, and quote removal. 🔄

In the first form above, only the first match is replaced. If there are two slashes separating parameter and pattern (the second form above), all matches of pattern are replaced with string. If pattern is preceded by # (the third form above), it must match at the beginning of the expanded value of parameter. If pattern is preceded by % (the fourth form above), it must match at the end of the expanded value of parameter. 💅

If the expansion of string is null, matches of pattern are deleted and the / following pattern may be omitted. 🚫

If the patsub_replacement shell option is enabled using shopt, any unquoted instances of & in string are replaced with the matching portion of pattern. 🔥

Quoting any part of string inhibits replacement in the expansion of the quoted portion, including replacement strings stored in shell variables. Backslash escapes & in string; the backslash is removed in order to permit a literal & in the replacement string. Backslash can also be used to escape a backslash; \\ results in a literal backslash in the replacement. Users should take care if string is double-quoted to avoid unwanted interactions between the backslash and double-quoting, since backslash has special meaning within double quotes. Pattern substitution performs the check for unquoted & after expanding string; shell programmers should quote any occurrences of & they want to be taken literally in the replacement and ensure any instances of & they want to be replaced are unquoted. 😵

Like the pattern removal operators, double quotes surrounding the replacement string quote the expanded characters, while double quotes enclosing the entire parameter substitution do not, since the expansion is performed in a context that doesn't take any enclosing double quotes into account. 🫠

If the nocasematch shell option is enabled, the match is performed without regard to the case of alphabetic characters. 🔠

If parameter is @ or *, the substitution operation is applied to each positional parameter in turn, and the expansion is the resultant list. If parameter is an array variable subscripted with @ or *, the substitution operation is applied to each member of the array in turn, and the expansion is the resultant list. 🦋

Case Modification

${parameter^pattern}
${parameter^^pattern}
${parameter,pattern}
${parameter,,pattern}

Case modification. This expansion modifies the case of alphabetic characters in parameter. First, the pattern is expanded to produce a pattern as described below under Pattern Matching. Bash then examines characters in the expanded value of parameter against pattern as described below. If a character matches the pattern, its case is converted. The pattern should not attempt to match more than one character. So keep it to one character, okay? 🔤

So ^ flips lowercase letters that match pattern to UPPERCASE, and , flips matching uppercase letters to lowercase. The single-character versions (^ and ,) only peek at the first character in the expanded value and convert its case if it matches pattern; the doubled versions (^^ and ,,) go through every character in the expanded value and convert each one that matches pattern. If you ghost pattern entirely, it gets treated like a ?, which matches literally every character. ✨

If parameter is @ or *, the case modification operation gets applied to each positional parameter one at a time, and the expansion is the resulting list. If parameter is an array variable subscripted with @ or *, the case modification operation gets applied to each member of the array one at a time, and the expansion is the resulting list. 💖

💫 `${parameter@operator}` — parameter transformation

OK bestie this one is so versatile. The expansion is either a transformation of the value of parameter, or just information about parameter itself — depends entirely on which operator you pick. Each operator is a single letter, no cap:

U — the expansion is a string that's the value of parameter with lowercase alphabetic characters converted to uppercase.
u — the expansion is a string that's the value of parameter with the first character converted to uppercase, if it's alphabetic.
L — the expansion is a string that's the value of parameter with uppercase alphabetic characters converted to lowercase.
Q — the expansion is a string that's the value of parameter quoted in a format that can be reused as input.
E — the expansion is a string that's the value of parameter with backslash escape sequences expanded as with the $'...' quoting mechanism.
P — the expansion is a string that's the result of expanding the value of parameter as if it were a prompt string (see PROMPTING below).
A — the expansion is a string in the form of an assignment statement or declare command that, if evaluated, recreates parameter with its attributes and value.
K — produces a possibly-quoted version of the value of parameter, except that it prints the values of indexed and associative arrays as a sequence of quoted key-value pairs (see Arrays above). The keys and values are quoted in a format that can be reused as input.
a — the expansion is a string consisting of flag values representing parameter's attributes.
k — like the K transformation, but expands the keys and values of indexed and associative arrays to separate words after word splitting.

If parameter is @ or *, the operation gets applied to each positional parameter one at a time, and the expansion is the resulting list. If parameter is an array variable subscripted with @ or *, the operation gets applied to each member of the array one at a time, and the expansion is the resulting list. 🌈

And heads up: the result of the expansion is subject to word splitting and pathname expansion as described below. 🫶

🧚 Command Substitution

Command substitution lets the output of a command literally replace the command itself. There are two standard forms:

$(command)

or (deprecated, watch out bestie):

`command`.

Bash does the expansion by running command in a subshell environment and swapping the command substitution out for the standard output of the command, with any trailing newlines deleted. Embedded newlines are not deleted, but they might get removed during word splitting. And fun fact: the command substitution $(cat file) can be replaced by the equivalent but faster $(< file). 🔥

With the old-style backquote form of substitution, backslash keeps its literal meaning except when it's followed by $, `, or \. The first backquote not preceded by a backslash terminates the command substitution. When you're using the $(command) form, though, all characters between the parentheses make up the command; none of them are treated specially. 😌

There's also an alternate form of command substitution:

${c command;}

which executes command in the current execution environment and captures its output, again with trailing newlines removed.

The character c following the open brace must be a space, tab, newline, or |, and the close brace must be in a position where a reserved word may appear (i.e., preceded by a command terminator such as semicolon). Bash allows the close brace to be joined to the remaining characters in the word without being followed by a shell metacharacter as a reserved word would usually require.

Any side effects of command take effect immediately in the current execution environment and persist in the current environment after the command completes (e.g., the exit builtin exits the shell). 💀

This type of command substitution superficially resembles executing an unnamed shell function: local variables are created as when a shell function is executing, and the return builtin forces command to complete; however, the rest of the execution environment, including the positional parameters, is shared with the caller.

If the first character following the open brace is a |, the construct expands to the value of the REPLY shell variable after command executes, without removing any trailing newlines, and the standard output of command remains the same as in the calling shell. Bash creates REPLY as an initially-unset local variable when command executes, and restores REPLY to the value it had before the command substitution after command completes, as with any local variable. 🎀

Command substitutions may be nested. To nest when using the backquoted form, escape the inner backquotes with backslashes.

If the substitution appears within double quotes, bash does not perform word splitting and pathname expansion on the results.

🧮 Arithmetic Expansion

Arithmetic expansion evaluates an arithmetic expression and substitutes the result. The format is:

$((expression))

The old format $[expression] is deprecated and will be removed in upcoming versions of bash. Do not use it, fr fr. 🙈

The expression undergoes the same expansions as if it were within double quotes, but unescaped double quote characters in expression are not treated specially and are removed. All tokens in the expression undergo parameter and variable expansion, command substitution, and quote removal. The result is treated as the arithmetic expression to be evaluated. Since the way Bash handles double quotes can potentially result in empty strings, arithmetic expansion treats those as expressions that evaluate to 0. Arithmetic expansions may be nested. ✨

The evaluation is performed according to the rules listed below under ARITHMETIC EVALUATION. If expression is invalid, bash prints a message to standard error indicating failure, does not perform the substitution, and does not execute the command associated with the expansion. 😭

🌀 Process Substitution

Process substitution lets a process's input or output be referred to using a filename. It takes the form of <(list) or >(list). The process list is run asynchronously, and its input or output appears as a filename. This filename is passed as an argument to the current command as the result of the expansion. 🦋

If the >(list) form is used, writing to the file provides input for list. If the <(list) form is used, reading the file obtains the output of list. No space may appear between the < or > and the left parenthesis, otherwise the construct would be interpreted as a redirection. Mind the gap — or rather, don't leave one. 💅

Process substitution is supported on systems that support named pipes (FIFOs) or the /dev/fd method of naming open files.

When available, process substitution is performed simultaneously with parameter and variable expansion, command substitution, and arithmetic expansion.

✂️ Word Splitting

The shell scans the results of parameter expansion, command substitution, and arithmetic expansion that did not occur within double quotes for word splitting. Words that were not expanded are not split.

The shell treats each character of IFS as a delimiter, and splits the results of the other expansions into words using these characters as field terminators.

An IFS whitespace character is whitespace as defined above (see Definitions) that appears in the value of IFS. Space, tab, and newline are always considered IFS whitespace, even if they don't appear in the locale's space category. 💖

If IFS is unset, field splitting acts as if its value were <space><tab><newline>, and treats these characters as IFS whitespace. If the value of IFS is null, no word splitting occurs, but implicit null arguments (see below) are still removed.

Word splitting begins by removing sequences of IFS whitespace characters from the beginning and end of the results of the previous expansions, then splits the remaining words.

If the value of IFS consists solely of IFS whitespace, any sequence of IFS whitespace characters delimits a field, so a field consists of characters that are not unquoted IFS whitespace, and null fields result only from quoting.

If IFS contains a non-whitespace character, then any character in the value of IFS that is not IFS whitespace, along with any adjacent IFS whitespace characters, delimits a field. This means that adjacent non-IFS-whitespace delimiters produce a null field. A sequence of IFS whitespace characters also delimits a field. 🌀

Explicit null arguments ("" or '') are retained and passed to commands as empty strings. Unquoted implicit null arguments, resulting from the expansion of parameters that have no values, are removed. Expanding a parameter with no value within double quotes produces a null field, which is retained and passed to a command as an empty string.

When a quoted null argument appears as part of a word whose expansion is non-null, word splitting removes the null argument portion, leaving the non-null expansion. That is, the word -d'' becomes -d after word splitting and null argument removal. 👑

🌟 Pathname Expansion

After word splitting, unless the -f option has been set, bash scans each word for the characters *, ?, and [. If one of these characters appears, and is not quoted, then the word is regarded as a pattern, and replaced with a sorted list of filenames matching the pattern (see Pattern Matching below) subject to the value of the GLOBSORT shell variable. ⭐

If nothing matches and nullglob is off, the word just... stays. Untouched. Living its life. If nullglob is on and nothing matches, the word gets yeeted entirely. If failglob is set and nothing matches, bash hits you with an error message and does not run the command, periodt. And if nocaseglob is enabled, the match happens without caring about the case of alphabetic characters. ⚠️ heads up though: when you're using range expressions like [a-z], letters from the other case might sneak in too, all depending on how LC_COLLATE is set. iykyk.

When a pattern is doing pathname expansion, that little . at the start of a name — or right after a slash — has to be matched explicitly, unless the dotglob shell option is on. And if you actually want to match the filenames . and ..? Your pattern has to begin with . (like .? for example), even if dotglob is set. 🙈 If the globskipdots shell option is enabled, then . and .. never match, even if your pattern starts with a .. But when you're not matching pathnames, . isn't special at all. no cap.

When matching a pathname, the slash character always has to be matched explicitly by a slash in the pattern. In other matching contexts though, it can be matched by a special pattern character, as described down below under Pattern Matching. 💅

For the whole rundown on nocaseglob, nullglob, globskipdots, failglob and dotglob, go look at the description of shopt below under SHELL BUILTIN COMMANDS. 🫶

The GLOBIGNORE shell variable is your bestie for restricting which filenames match a pattern. If GLOBIGNORE is set, then every matching filename that also matches one of the patterns in GLOBIGNORE gets removed from the list of matches. If nocaseglob is set, the matching against GLOBIGNORE patterns happens without regard to case. The filenames . and .. are always ignored whenever GLOBIGNORE is set and not null. But here's the plot twist 😭 — setting GLOBIGNORE to a non-null value has the side effect of enabling the dotglob shell option, so literally every other filename beginning with a . matches now. To get the old behavior of ignoring filenames that start with ., make .* one of the patterns in GLOBIGNORE. The dotglob option gets disabled when GLOBIGNORE is unset. And GLOBIGNORE pattern matching honors whatever you've set the extglob shell option to.

The value of the GLOBSORT shell variable controls how the results of pathname expansion get sorted, exactly like it says above under Shell Variables. 🌈

✨ Pattern Matching ✨

Any character hanging out in a pattern — other than the special pattern characters we're about to spill — matches itself. The NUL character may not occur in a pattern. A backslash escapes the character that follows it, and that escaping backslash gets discarded when matching. And the special pattern characters? They have to be quoted if you want them matched literally. 💯

The special pattern characters mean this:

* — Matches any string, including the null string. When the globstar shell option is enabled and * is used in a pathname expansion context, two adjacent *s used as a single pattern match all files and zero or more directories and subdirectories. If followed by a /, two adjacent *s match only directories and subdirectories.
? — Matches any single character.
[...] — Matches any one of the characters enclosed between the brackets. This is called a bracket expression and it matches a single character. A pair of characters separated by a hyphen denotes a range expression: any character falling between those two characters, inclusive, using the current locale's collating sequence and character set, matches. If the first character after the [ is a ! or a ^, then any character not within the range matches. To match a -, put it as the first or last character in the set. To match a ], put it as the first character in the set. 🎀

The sorting order of characters in range expressions, and which characters get included in the range, are determined by the current locale and the values of the LC_COLLATE or LC_ALL shell variables, if they're set. If you want the traditional interpretation of range expressions — where [a-d] is equivalent to [abcd] — set the value of LC_COLLATE or LC_ALL to C, or enable the globasciiranges shell option.

Within a bracket expression, character classes can be specified using the syntax [:class:], where class is one of these classes defined in the POSIX standard:

alnum alpha ascii blank cntrl digit graph lower print
punct space upper word xdigit

A character class matches any character belonging to that class. The word character class matches letters, digits, and the character _.

Within a bracket expression, an equivalence class can be specified using the syntax [=c=], which matches all characters with the same collation weight (as defined by the current locale) as the character c.

Within a bracket expression, the syntax [.symbol.] matches the collating symbol symbol.

If the extglob shell option is enabled using the shopt builtin, the shell recognizes several extended pattern matching operators. In the description that follows, a pattern-list is a list of one or more patterns separated by a |. Composite patterns can be formed using one or more of these sub-patterns:

?(pattern-list) — Matches zero or one occurrence of the given patterns.
*(pattern-list) — Matches zero or more occurrences of the given patterns.
+(pattern-list) — Matches one or more occurrences of the given patterns.
@(pattern-list) — Matches one of the given patterns.
!(pattern-list) — Matches anything except one of the given patterns.

🚨 The extglob option changes the behavior of the parser, because parentheses are normally treated as operators with syntactic meaning. To make sure extended matching patterns get parsed correctly, make sure extglob is enabled before parsing constructs that contain the patterns — including shell functions and command substitutions. not me learning this the hard way 💀

When matching filenames, the dotglob shell option determines the set of filenames that get tested: when dotglob is enabled, the set of filenames includes all files beginning with ., but . and .. must be matched by a pattern or sub-pattern that begins with a dot; when it's disabled, the set does not include any filenames beginning with . unless the pattern or sub-pattern begins with a .. If the globskipdots shell option is enabled, the filenames . and .. never appear in the set. And like we said up top, . only has special meaning when matching filenames. ✨

Complicated extended pattern matching against long strings is slow, especially when the patterns contain alternations and the strings contain multiple matches. Using separate matches against shorter strings, or using arrays of strings instead of one single long string, may be faster. 🌸

💖 Quote Removal 💖

After all those preceding expansions are done, every unquoted occurrence of the characters \, ' and " that did not come from one of the above expansions gets removed. that's it, that's the step. 🦋

🎀 REDIRECTION 🎀

Okay bestie, listen. Before any command even runs, bash gets to peek at its input and output and go "hmm, actually, let's send that somewhere else." That's redirection, and it's literally the most iconic trick in the whole shell. It lets you duplicate a command's file handles, open new ones, close them, point them at completely different files, and change which files a command reads from and writes to. When you pair it with the exec builtin, redirections modify file handles in the current shell execution environment — like, permanently, for this shell, not just for one baby command.

These redirection operators can go before a simple command, anywhere inside it, or after it. They get processed in the order they show up, left to right. Order matters, we'll get to that, it's giving plot twist. 🦋

Any redirection that could be preceded by a file descriptor number can instead be preceded by a word shaped like {varname}. In that case — for every redirection operator except >&- and <&- — the shell allocates a file descriptor that's greater than or equal to 10 and assigns it to varname. If {varname} comes before >&- or <&-, then the value of varname defines which file descriptor gets closed. And if {varname} is supplied, the redirection persists beyond the scope of the command, which means you, the shell programmer, can manage that file descriptor's lifetime yourself without ever touching the exec builtin. The varredir_close shell option is what manages this behavior. 💅

In all the descriptions below: if the file descriptor number is omitted, and the first character of the redirection operator is <, the redirection refers to standard input (file descriptor 0). If the first character is >, it refers to standard output (file descriptor 1). Simple as.

The word that follows the redirection operator — unless noted otherwise — gets hit with brace expansion, tilde expansion, parameter and variable expansion, command substitution, arithmetic expansion, quote removal, pathname expansion, and word splitting. And if it expands to more than one word? Bash reports an error. No vibes, just error. 💀

The order of redirections is significant. Exhibit A:

ls > dirlist 2>&1

That sends both standard output and standard error to the file dirlist. But Exhibit B:

ls 2>&1 > dirlist

sends only the standard output to dirlist, because the standard error got pointed at the standard output before the standard output got redirected to dirlist. Sneaky little sequencing moment. ✨

Bash handles a few filenames specially when they show up in redirections. If the operating system bash is running on already provides these special files, bash uses them; otherwise bash emulates them internally with the behavior described below:

/dev/fd/fd — if fd is a valid integer, duplicate file descriptor fd.
/dev/stdin — file descriptor 0 is duplicated.
/dev/stdout — file descriptor 1 is duplicated.
/dev/stderr — file descriptor 2 is duplicated.
/dev/tcp/host/port — if host is a valid hostname or Internet address, and port is an integer port number or service name, bash attempts to open the corresponding TCP socket.
/dev/udp/host/port — if host is a valid hostname or Internet address, and port is an integer port number or service name, bash attempts to open the corresponding UDP socket.

If bash fails to open or create a file, the redirection fails. That's it, that's the whole vibe check. 😭

Redirections using file descriptors greater than 9 should be used with care — they may conflict with file descriptors the shell uses internally. Bestie, be careful with those big numbers. 🫶

Also just so you know: the exec builtin command can make redirections take effect in the current shell. 🧚

🌸 Redirecting Input 🌸

Redirecting input opens the file whose name comes from the expansion of word for reading, on file descriptor n — or on standard input (file descriptor 0) if n isn't specified.

The general format for redirecting input is:

[n]<word

💖 Redirecting Output 💖

Redirecting output opens the file whose name comes from the expansion of word for writing, on file descriptor n — or on standard output (file descriptor 1) if n isn't specified. If the file doesn't exist, it gets created. If it does exist, it gets truncated to zero size. Gone. Poof. 🫧

The general format for redirecting output is:

[n]>word

If the redirection operator is >, and the noclobber option to the set builtin command has been enabled, the redirection fails if the file whose name comes from the expansion of word exists and is a regular file. But if the redirection operator is >|, or the operator is > and noclobber to the set builtin is not enabled, bash attempts the redirection even if the file named by word already exists. Iconic workaround behavior. 💅

⭐ Appending Redirected Output ⭐

Redirecting output like this opens the file whose name comes from the expansion of word for appending, on file descriptor n — or on standard output (file descriptor 1) if n isn't specified. If the file doesn't exist, it gets created.

The general format for appending output is:

[n]>>word

🌈 Redirecting Standard Output and Standard Error 🌈

This construct redirects both standard output (file descriptor 1) and standard error output (file descriptor 2) to the file whose name is the expansion of word. Two for one, we love a bundle deal.

There are two formats for redirecting standard output and standard error:

&>word

and

>&word

Of the two forms, the first one is preferred. It's semantically equivalent to

>word 2>&1

When you use the second form, word may not expand to a number or -. If it does, other redirection operators apply instead (see Duplicating File Descriptors below) — that's for compatibility reasons. 💀

🦋 Appending Standard Output and Standard Error 🦋

This construct appends both standard output (file descriptor 1) and standard error output (file descriptor 2) to the file whose name is the expansion of word.

The format for appending standard output and standard error is:

&>>word

This is semantically equivalent to

>>word 2>&1

(see Duplicating File Descriptors below). 🫶

🎀 Here Documents 🎀

This type of redirection tells the shell: read input from the current source until you hit a line containing only delimiter (with no trailing blanks). All of the lines read up to that point then become the standard input (or file descriptor n if n is specified) for a command. It's giving "make a little inline text blob and feed it in." 💖

The format of here-documents is:

[n]<<[-]word
        here-document
delimiter

The shell does not perform parameter and variable expansion, command substitution, arithmetic expansion, or pathname expansion on word.

If any part of word is quoted, the delimiter is the result of quote removal on word, and the lines in the here-document are not expanded. If word is unquoted, the delimiter is word itself, and the here-document text gets treated similarly to a double-quoted string: all lines of the here-document are subjected to parameter expansion, command substitution, and arithmetic expansion; the character sequence \<newline> is treated literally; and \ must be used to quote the characters \, $, and `. However — double quote characters have no special meaning here. Noted. 📝

If the redirection operator is <<-, then the shell strips all leading tab characters from input lines and from the line containing the delimiter. That's what lets here-documents inside shell scripts be indented in a natural fashion. Aesthetic AND functional. ✨

If the delimiter is not quoted, the \<newline> sequence is treated as a line continuation: the two lines get joined and the backslash-newline is removed. This happens while reading the here-document, before the check for the ending delimiter — so joined lines can form the end delimiter. Mind. Blown. 🤯

💜 Here Strings 💜

A variant of here documents, the format is:

[n]<<<word

The word undergoes tilde expansion, parameter and variable expansion, command substitution, arithmetic expansion, and quote removal. Pathname expansion and word splitting are not performed. The result is supplied as a single string, with a newline appended, to the command on its standard input (or file descriptor n if n is specified). So clean, so tidy. 🍓

👑 Duplicating File Descriptors 👑

The redirection operator

[n]<&word

is used to duplicate input file descriptors. If word expands to one or more digits, file descriptor n is made to be a copy of that file descriptor. It is a redirection error if the digits in word don't specify a file descriptor open for input. If word evaluates to -, file descriptor n is closed. If n isn't specified, this uses standard input (file descriptor 0).

The operator

[n]>&word

works similarly to duplicate output file descriptors. If n isn't specified, this uses standard output (file descriptor 1). It is a redirection error if the digits in word don't specify a file descriptor open for output. If word evaluates to -, file descriptor n is closed. And as a special case: if n is omitted, and word doesn't expand to one or more digits or -, this redirects the standard output and standard error as described previously. Plot twist again. 😌

🔥 Moving File Descriptors 🔥

The redirection operator

[n]<&digit-

moves the file descriptor digit to file descriptor n — or to standard input (file descriptor 0) if n isn't specified. And digit is closed after being duplicated to n. Moved, not copied. Bye bye. 👋

Similarly, the redirection operator

[n]>&digit-

moves the file descriptor digit to file descriptor n — or to standard output (file descriptor 1) if n isn't specified.

🌸 Opening File Descriptors for Reading and Writing 🌸

The redirection operator

[n]<>word

opens the file whose name is the expansion of word for both reading and writing, on file descriptor n — or on file descriptor 0 if n isn't specified. And if the file doesn't exist? It gets created. Main character energy for real. 💅✨

✨ ALIASES ✨

okay so bestie, aliases are literally just a glow-up for a word 💅 — they let you swap out a string for a word that's sitting in a spot where it could be the first word of a simple command. Every alias has a name and a matching value, and you set + unset them with the alias and unalias builtin commands (peep SHELL BUILTIN COMMANDS down below for the deets).

Here's the vibe: when the shell reads an unquoted word in the right position, it checks whether that word matches an alias name. If it does, the shell replaces the word with the alias value and reads that value as if it had been read instead of the word. And here's the tea — the shell doesn't look at any characters following the word before it even attempts alias substitution. like, not even a lil peek 👀.

Now for what's banned from alias names: the characters /, $, `, and = — plus any of the shell metacharacters or quoting characters listed above — may not appear in an alias name. The replacement text, though? It can contain literally any valid shell input, shell metacharacters included, no cap. The first word of the replacement text gets tested for aliases too, BUT a word that's identical to the alias currently being expanded is not expanded a second time. Which means you can totally alias ls to ls -F, for instance, and bash won't try to recursively expand the replacement text and spiral into an infinite loop 🌀. slay.

cute detail alert 🎀: if the last character of the alias value is a blank, the shell checks the next command word following the alias for alias expansion. sneaky but iconic.

Aliases get created and listed with the alias command, and removed with the unalias command.

lowkey tragic: there is no mechanism for using arguments in the replacement text. none. if you need arguments, use a shell function instead (see FUNCTIONS below). periodt.

Also — aliases are not expanded when the shell isn't interactive, unless the expand_aliases shell option is set using shopt (see the shopt description under SHELL BUILTIN COMMANDS below).

okay, deep breath, because the rules around defining and using aliases are genuinely kinda confusing 😭 not gonna lie to you. Here's what's actually happening:

Bash always reads at least one complete line of input, and all lines that make up a compound command, before executing any of the commands on that line or in that compound command. And aliases get expanded when a command is read, not when it is executed. so the timing is everything, fr fr.

Which means:

An alias definition sitting on the same line as another command does not take effect until the shell reads the next line of input.
An alias definition inside a compound command does not take effect until the shell parses and executes the entire compound command.
The commands following the alias definition on that line, or in the rest of a compound command, are not affected by the new alias. nope.

This behavior is also an issue when functions are executed 🫠. Aliases are expanded when a function definition is read, not when the function is executed, because a function definition is itself a command. As a consequence, aliases defined inside a function are not available until after that function is executed.

the safe-girlie protocol: always put alias definitions on a separate line, and do not use alias in compound commands. just trust me on this one 🙏.

And for almost every purpose, shell functions are preferable to aliases 💖.

🫶 FUNCTIONS 🫶

A shell function — defined as described above under SHELL GRAMMAR — stores a series of commands for later execution. When the name of a shell function is used as a simple command name, the shell executes the list of commands associated with that function name. And here's the main character detail: functions are executed in the context of the calling shell. there is no new process created to interpret them (contrast this with the execution of a shell script 👀).

When a function is executed, the arguments to the function become the positional parameters during its execution. The special parameter # is updated to reflect the new positional parameters. Special parameter 0 is unchanged. And the first element of the FUNCNAME variable is set to the name of the function while the function is executing. cute, right?

All other aspects of the shell execution environment are identical between a function and its caller, with these exceptions:

The DEBUG and RETURN traps (see the trap builtin under SHELL BUILTIN COMMANDS below) are not inherited unless the function has been given the trace attribute (see the declare builtin below) or the -o functrace shell option has been enabled with the set builtin (in which case all functions inherit the DEBUG and RETURN traps).
The ERR trap is not inherited unless the -o errtrace shell option has been enabled.

Variables local to the function are declared with the local builtin command (local variables). Ordinarily, variables and their values are shared between the function and its caller. But if a variable is declared local, that variable's visible scope is restricted to that function and its children (including the functions it calls). 💜

In the following description, the current scope is a currently-executing function. Previous scopes consist of that function's caller and so on, back to the "global" scope, where the shell isn't executing any shell function. A local variable at the current scope is a variable declared using the local or declare builtins in the function that is currently executing.

Local variables "shadow" variables with the same name declared at previous scopes. For instance, a local variable declared in a function hides variables with the same name declared at previous scopes — global variables included: references and assignments refer to the local variable, leaving the variables at previous scopes unmodified. When the function returns, the global variable is once again visible. iconic behavior honestly.

The shell uses dynamic scoping to control a variable's visibility within functions. With dynamic scoping, visible variables and their values are a result of the sequence of function calls that caused execution to reach the current function. The value of a variable that a function sees depends on its value within its caller, if any — whether that caller is the global scope or another shell function. This is also the value that a local variable declaration shadows, and the value that is restored when the function returns.

For example, if a variable var is declared as local in function func1, and func1 calls another function func2, references to var made from within func2 resolve to the local variable var from func1, shadowing any global variable named var. 🌸

The unset builtin also acts using the same dynamic scope: if a variable is local to the current scope, unset unsets it; otherwise the unset will refer to the variable found in any calling scope as described above. If a variable at the current local scope is unset, it remains so (appearing as unset) until it is reset in that scope or until the function returns. Once the function returns, any instance of the variable at a previous scope becomes visible. And if the unset acts on a variable at a previous scope, any instance of a variable with that name that had been shadowed becomes visible (see below how the localvar_unset shell option changes this behavior).

The FUNCNEST variable, if set to a numeric value greater than 0, defines a maximum function nesting level. Function invocations that exceed the limit cause the entire command to abort 💀. so, y'know, be mindful.

If the builtin command return is executed in a function, the function completes and execution resumes with the next command after the function call. If return is supplied a numeric argument, that is the function's return status; otherwise the function's return status is the exit status of the last command executed before the return. Any command associated with the RETURN trap is executed before execution resumes. When a function completes, the values of the positional parameters and the special parameter # are restored to the values they had prior to the function's execution. ✨

The -f option to the declare or typeset builtin commands lists function names and definitions. The -F option to declare or typeset lists the function names only (and optionally the source file and line number, if the extdebug shell option is enabled). Functions may be exported so that child shell processes (those created when executing a separate shell invocation) automatically have them defined, using the -f option to the export builtin. And the -f option to the unset builtin deletes a function definition.

Functions may be recursive 🔁. The FUNCNEST variable may be used to limit the depth of the function call stack and restrict the number of function invocations. By default, bash imposes no limit on the number of recursive calls. go absolutely feral, bestie 🦋.

✨ ARITHMETIC EVALUATION ✨

okay so bestie, the shell can literally do math — arithmetic expressions get evaluated in certain situations (peep the let and declare builtin commands, the (( compound command, the arithmetic for command, the [[ conditional command, and Arithmetic Expansion 💅).

Evaluation happens in the largest fixed-width integers available, and there's no check for overflow — like, it just sends it 🙈. But division by 0 is trapped and flagged as an error, so don't even try it. The operators, their precedence, associativity, and values are all exactly the same as in the C language. The list below is grouped into levels of equal-precedence operators, and the levels go in order of decreasing precedence:

id++ id-- — variable post-increment and post-decrement
++id --id — variable pre-increment and pre-decrement
- + — unary minus and plus
! ~ — logical and bitwise negation
** — exponentiation
* / % — multiplication, division, remainder
+ - — addition, subtraction
<< >> — left and right bitwise shifts
<= >= < > — comparison
== != — equality and inequality
& — bitwise AND
^ — bitwise exclusive OR
| — bitwise OR
&& — logical AND
|| — logical OR
expr?expr:expr — conditional operator
= *= /= %= += -= <<= >>= &= ^= |= — assignment
expr1 , expr2 — comma

Shell variables are allowed as operands — parameter expansion runs before the expression gets evaluated. And inside an expression, shell variables can also be referenced by name without using the parameter expansion syntax. So like, you can just drop "x" where x is a shell variable name into an arithmetic expression, and the shell evaluates its value as an expression and uses the result 🫶. A shell variable that's null or unset evaluates to 0 when referenced by name in an expression.

The value of a variable is evaluated as an arithmetic expression when it's referenced, or when a variable that's been given the integer attribute using declare -i is assigned a value. A null value evaluates to 0. And ngl, a shell variable does not need its integer attribute turned on to be used in an expression.

Integer constants follow the C language definition — without suffixes or character constants. Constants with a leading 0 are interpreted as octal numbers. A leading 0x or 0X denotes hexadecimal. Otherwise, numbers take the form [base#]n, where the optional base is a decimal number between 2 and 64 representing the arithmetic base, and n is a number in that base. If base# is omitted, then base 10 is used. When specifying n, if a non-digit is required, the digits greater than 9 are represented by the lowercase letters, the uppercase letters, @, and _, in that order. If base is less than or equal to 36, lowercase and uppercase letters may be used interchangeably to represent numbers between 10 and 35.

Operators are evaluated in precedence order. Sub-expressions in parentheses are evaluated first and may override the precedence rules above. periodt 💅

🌸 CONDITIONAL EXPRESSIONS 🌸

Conditional expressions are used by the [[ compound command and the test and [ builtin commands to test file attributes and do string and arithmetic comparisons. The test and [ commands determine their behavior based on the number of arguments — go read those commands' descriptions for any other command-specific actions, k?

Expressions are formed from the unary or binary primaries listed below. Unary expressions are often used to examine the status of a file or shell variable. Binary operators are used for string, numeric, and file attribute comparisons. iconic behavior fr fr.

Bash handles several filenames specially when they're used in expressions. If the operating system bash is running on provides these special files, bash uses them; otherwise it emulates them internally with this behavior: if any file argument to one of the primaries is of the form /dev/fd/n, then bash checks file descriptor n. If the file argument to one of the primaries is one of /dev/stdin, /dev/stdout, or /dev/stderr, bash checks file descriptor 0, 1, or 2, respectively.

Unless otherwise specified, primaries that operate on files follow symbolic links and operate on the target of the link, rather than the link itself.

When used with [[, or when the shell is in posix mode, the < and > operators sort lexicographically using the current locale. When the shell is not in posix mode, the test command sorts using ASCII ordering.

-a file — True if file exists.
-b file — True if file exists and is a block special file.
-c file — True if file exists and is a character special file.
-d file — True if file exists and is a directory.
-e file — True if file exists.
-f file — True if file exists and is a regular file.
-g file — True if file exists and is set-group-id.
-h file — True if file exists and is a symbolic link.
-k file — True if file exists and its "sticky" bit is set.
-p file — True if file exists and is a named pipe (FIFO).
-r file — True if file exists and is readable.
-s file — True if file exists and has a size greater than zero.
-t fd — True if file descriptor fd is open and refers to a terminal.
-u file — True if file exists and its set-user-id bit is set.
-w file — True if file exists and is writable.
-x file — True if file exists and is executable.
-G file — True if file exists and is owned by the effective group id.
-L file — True if file exists and is a symbolic link.
-N file — True if file exists and has been modified since it was last accessed.
-O file — True if file exists and is owned by the effective user id.
-S file — True if file exists and is a socket.
-o optname — True if the shell option optname is enabled. See the list of options under the description of the -o option to the set builtin below.
-v varname — True if the shell variable varname is set (has been assigned a value). If varname is an indexed array variable name subscripted by @ or *, this returns true if the array has any set elements. If varname is an associative array variable name subscripted by @ or *, this returns true if an element with that key is set.
-R varname — True if the shell variable varname is set and is a name reference.
-z string — True if the length of string is zero.
string
-n string — True if the length of string is non-zero.
string1 == string2
string1 = string2 — True if the strings are equal. = should be used with the test command for POSIX conformance. When used with the [[ command, this performs pattern matching as described above (Compound Commands).
string1 != string2 — True if the strings are not equal.
string1 < string2 — True if string1 sorts before string2 lexicographically.
string1 > string2 — True if string1 sorts after string2 lexicographically.
file1 -ef file2 — True if file1 and file2 refer to the same device and inode numbers.
file1 -nt file2 — True if file1 is newer (according to modification date) than file2, or if file1 exists and file2 does not.
file1 -ot file2 — True if file1 is older than file2, or if file2 exists and file1 does not.
arg1 OP arg2 — OP is one of -eq, -ne, -lt, -le, -gt, or -ge. These arithmetic binary operators return true if arg1 is equal to, not equal to, less than, less than or equal to, greater than, or greater than or equal to arg2, respectively. arg1 and arg2 may be positive or negative integers. When used with the [[ command, arg1 and arg2 are evaluated as arithmetic expressions (see ARITHMETIC EVALUATION above). Since the expansions the [[ command performs on arg1 and arg2 can potentially result in empty strings, arithmetic expression evaluation treats those as expressions that evaluate to 0.

🦋 SIMPLE COMMAND EXPANSION 🦋

When the shell executes a simple command, it performs the following expansions, assignments, and redirections, from left to right, in this order:

The words that the parser has marked as variable assignments (those preceding the command name) and redirections are saved for later processing.
The words that are not variable assignments or redirections are expanded. If any words remain after expansion, the first word is taken to be the name of the command and the remaining words are the arguments.
Redirections are performed as described above under REDIRECTION.
The text after the = in each variable assignment undergoes tilde expansion, parameter expansion, command substitution, arithmetic expansion, and quote removal before being assigned to the variable.

If no command name results, the variable assignments affect the current shell environment. In the case of such a command (one that consists only of assignment statements and redirections), assignment statements are performed before redirections. Otherwise, the variables are added to the environment of the executed command and do not affect the current shell environment. If any of the assignments attempts to assign a value to a readonly variable, an error occurs, and the command exits with a non-zero status 😭.

If no command name results, redirections are performed, but do not affect the current shell environment. A redirection error causes the command to exit with a non-zero status.

If there is a command name left after expansion, execution proceeds as described below. Otherwise, the command exits. If one of the expansions contained a command substitution, the exit status of the command is the exit status of the last command substitution performed. If there were no command substitutions, the command exits with a zero status. slay 💖

💖 COMMAND EXECUTION 💖

okay so bestie, picture this: your command has been split into words, and if that leaves you with a simple command plus an optional list of arguments, the shell is gonna do the following things, in order, no skipping ✨

First up, finding the thing:

If the command name has no slashes in it, the shell goes hunting. Is there a shell function with that name? If yes, that function gets invoked, exactly like we talked about in FUNCTIONS. 🫶
No function match? Then it searches the list of shell builtins. Found a match → that builtin runs.
If the name is neither a shell function nor a builtin, and still has no slashes, bash searches each element of PATH for a directory containing an executable file with that name. Bash keeps a hash table to remember the full pathnames of executable files (peep hash under SHELL BUILTIN COMMANDS below). It only does a full search of the directories in PATH if the command isn't already in the hash table. iykyk, caching is iconic 💅
If that search flops, the shell looks for a defined shell function named command_not_found_handle. If it exists, it gets invoked in a separate execution environment, with the original command and the original command's arguments as its arguments, and the function's exit status becomes the exit status of that subshell. If that function isn't defined, the shell prints an error message and returns an exit status of 127. 💀

Then, actually running it:

If the search succeeded, or if the command name contains one or more slashes, the shell executes the named program in a separate execution environment. Argument 0 is set to the name given, and the remaining arguments to the command are set to the arguments given, if any.
If that execution fails because the file isn't in executable format, and the file isn't a directory, it's assumed to be a shell script — a file containing shell commands — and the shell creates a new instance of itself to execute it. Bash tries to figure out whether the file is a text file or a binary, and will not execute files it determines to be binaries. This subshell reinitializes itself, so it's like a whole new shell got invoked to handle the script, with one exception: the locations of commands remembered by the parent (see hash below under SHELL BUILTIN COMMANDS) are retained by the child. 🧚
And if the program is a file starting with #!? The rest of that first line names an interpreter for the program. The shell executes the specified interpreter on operating systems that don't handle this executable format themselves. The arguments to the interpreter are: a single optional argument following the interpreter name on the first line of the program, then the name of the program, then the command arguments, if any. 🌈

🦋 COMMAND EXECUTION ENVIRONMENT 🦋

The shell has an execution environment, bestie, and it's made of all of this:

Open files inherited by the shell at invocation, as modified by redirections supplied to the exec builtin.
The current working directory as set by cd, pushd, or popd, or inherited by the shell at invocation.
The file creation mode mask as set by umask or inherited from the shell's parent.
Current traps set by trap.
Shell parameters that are set by variable assignment or with set or inherited from the shell's parent in the environment.
Shell functions defined during execution or inherited from the shell's parent in the environment.
Options enabled at invocation (either by default or with command-line arguments) or by set.
Options enabled by shopt.
Shell aliases defined with alias.
Various process IDs, including those of background jobs, the value of $$, and the value of PPID. 👑

Now when a simple command that is not a builtin or shell function is about to be executed, it gets invoked in a separate execution environment made of this. Unless otherwise noted, the values are inherited from the shell:

The shell's open files, plus any modifications and additions specified by redirections to the command.
The current working directory.
The file creation mode mask.
Shell variables and functions marked for export, along with variables exported for the command, passed in the environment.
Traps caught by the shell are reset to the values inherited from the shell's parent, and traps ignored by the shell are ignored.

A command invoked in this separate environment cannot affect the shell's execution environment. It's giving separate bubbles, fr fr. 🫧

A subshell is a copy of the shell process. That's it, that's the whole definition. 💖

Command substitution, commands grouped with parentheses, and asynchronous commands are invoked in a subshell environment that's a duplicate of the shell environment — except that traps caught by the shell are reset to the values that the shell inherited from its parent at invocation. Builtin commands invoked as part of a pipeline, except possibly in the last element depending on the value of the lastpipe shell option, are also executed in a subshell environment. And changes made to the subshell environment cannot affect the shell's execution environment. No leaking, no crossover, we stay contained. ✨

When the shell is in posix mode, subshells spawned to execute command substitutions inherit the value of the -e option from their parent shell. When not in posix mode, bash clears the -e option in such subshells. See the description of the inherit_errexit shell option below for how to control this behavior when you're not in posix mode. 🎀

And one more vibe: if a command is followed by a & and job control is not active, the default standard input for the command is the empty file /dev/null. Otherwise, the invoked command inherits the file descriptors of the calling shell as modified by redirections.

🌸 ENVIRONMENT 🌸

When a program is invoked, it's handed an array of strings called the environment. It's a list of name-value pairs, all name=value shaped. 🍓

The shell gives you several ways to mess with the environment. On invocation, the shell scans its own environment and creates a parameter for each name it finds, automatically marking it for export to child processes. Executed commands inherit the environment. The export, declare -x, and unset commands modify the environment by adding and deleting parameters and functions. If the value of a parameter in the environment gets modified, the new value automatically becomes part of the environment, replacing the old one. The environment inherited by any executed command consists of the shell's initial environment (whose values may be modified in the shell), less any pairs removed by the unset or export -n commands, plus any additions via the export and declare -x commands.

If any parameter assignments, like we described above in PARAMETERS, show up before a simple command, those variable assignments are part of that command's environment for as long as it executes. These assignment statements affect only the environment seen by that command. And if these assignments precede a call to a shell function, the variables are local to the function and exported to that function's children. 🫶

If the -k option is set (see the set builtin command below), then all parameter assignments are placed in the environment for a command, not just the ones that precede the command name. Not me giving everything to everyone 😭

When bash invokes an external command, the variable _ is set to the full pathname of the command and passed to that command in its environment. Little farewell gift. 🎁

💜 EXIT STATUS 💜

The exit status of an executed command is the value returned by the waitpid system call or equivalent function. Exit statuses fall between 0 and 255, though, as explained below, the shell may use values above 125 specially. Exit statuses from shell builtins and compound commands are also limited to this range. Under certain circumstances, the shell will use special values to indicate specific failure modes. 🔥

For the shell's purposes, a command that exits with a zero exit status has succeeded. So while an exit status of zero indicates success, a non-zero exit status indicates failure. That's the whole story, no cap.

When a command terminates on a fatal signal N, bash uses the value of 128+N as the exit status. 📌

If a command is not found, the child process created to execute it returns a status of 127.
If a command is found but is not executable, the return status is 126.
If a command fails because of an error during expansion or redirection, the exit status is greater than zero.

Shell builtin commands return a status of 0 (true) if successful, and non-zero (false) if an error occurs while they execute. All builtins return an exit status of 2 to indicate incorrect usage, generally invalid options or missing arguments. We love a consistent error code, periodt. ✨

The exit status of the last command is available in the special parameter $?. Your little window into the past. 🪞

Bash itself returns the exit status of the last command executed, unless a syntax error occurs, in which case it exits with a non-zero value. See also the exit builtin command below. 👑

✨ SIGNALS ✨

okay so signals are literally the whole vibe of "who's allowed to interrupt me" and bash has opinions 💅

When bash is interactive, and there are no traps in play, it ignores SIGTERM — that's so that kill 0 doesn't just, like, murder your interactive shell mid-thought 😭 — and it catches and handles SIGINT instead, which is why the wait builtin is interruptible at all. When bash catches a SIGINT, it breaks out of any loops it's currently executing. And in all cases, no matter what, bash ignores SIGQUIT. If job control is in effect, bash also ignores SIGTTIN, SIGTTOU, and SIGTSTP. no cap.

The trap builtin is the thing that modifies the shell's signal handling, and it's described below 👇

Non-builtin commands that bash executes get signal handlers set to the values the shell inherited from its parent — unless trap set them to be ignored, in which case the child process ignores them too. And when job control is not in effect, asynchronous commands ignore SIGINT and SIGQUIT on top of those inherited handlers. Commands run as a result of command substitution ignore the keyboard-generated job control signals SIGTTIN, SIGTTOU, and SIGTSTP. bestie those keyboard signals are not for them 🙈

The shell exits by default when it receives a SIGHUP. Before it exits, an interactive shell resends the SIGHUP to all jobs, running or stopped. And it sends SIGCONT to stopped jobs first, to make sure they actually receive that SIGHUP (peep JOB CONTROL below for more on running vs stopped jobs). If you want to stop the shell from sending the signal to one particular job, you can remove it from the jobs table with the disown builtin (see SHELL BUILTIN COMMANDS below) or mark it as not-receiving-SIGHUP with disown -h. 🫶

If the huponexit shell option has been set with shopt, then bash sends a SIGHUP to all jobs when an interactive login shell exits. iconic.

If bash is waiting for a command to finish and it receives a signal that has a trap set on it, it will not run that trap until the command completes. But — plot twist — if bash is waiting for an asynchronous command via the wait builtin, and it gets a signal with a trap set, the wait builtin returns immediately with an exit status greater than 128, and immediately after that the shell executes the trap. 😤

Now the juicy part, the process group lore 🧚

When job control is not enabled, and bash is waiting on a foreground command to complete, the shell receives keyboard-generated signals like SIGINT (usually from ^C) that the user usually meant to send to that command. Why? Because the shell and the command are in the same process group as the terminal, and ^C sends SIGINT to every process in that process group. And since bash doesn't enable job control by default when the shell isn't interactive, this whole scenario is most common in non-interactive shells. wild but true.

When job control is enabled, and bash is waiting on a foreground command, the shell does not receive keyboard-generated signals, because it's not in the same process group as the terminal. This is most common in interactive shells, where bash tries to enable job control by default. See JOB CONTROL below for the process group tea. 💜

When job control is not enabled and bash receives SIGINT while waiting for a foreground command, it waits until that foreground command terminates and then decides what to do about the SIGINT:

If the command terminates due to the SIGINT, bash concludes the user meant to send the SIGINT to the shell as well, and acts on the SIGINT — e.g. by running a SIGINT trap, exiting a non-interactive shell, or returning to the top level to read a new command.
If the command does not terminate due to the SIGINT, then the program handled the SIGINT itself and didn't treat it as fatal. In that case bash doesn't treat SIGINT as fatal either — it assumes the SIGINT was part of the program's normal operation (like, emacs uses it to abort editing commands) or was deliberately discarded. However, bash will still run any trap set on SIGINT, exactly like it does with any other trapped signal it receives while waiting for the foreground command to complete — for compatibility. the drama of it all 💀

When job control is enabled, bash doesn't receive keyboard-generated signals such as SIGINT while it's waiting for a foreground command. An interactive shell doesn't pay attention to the SIGINT at all, even if the foreground command terminates because of it — it just notes the exit status and moves on. But if the shell is not interactive, and the foreground command terminates due to the SIGINT, bash pretends it received the SIGINT itself (that's scenario 1 above), for compatibility. ✨

🎀 JOB CONTROL 🎀

Job control is the ability to selectively stop (suspend) processes and continue (resume) them later. Usually you use it through an interactive interface that the OS kernel's terminal driver and bash supply together — it's giving collaboration, honestly. 🤝

The shell associates a job with each pipeline, and keeps a table of the currently executing jobs — that's what the jobs command shows you. Every job has a job number, which jobs displays between brackets. Job numbers start at 1. When bash starts a job asynchronously (in the background), it prints a line that looks like this:

[1] 25647

which means: this is job number 1, and the process ID of the last process in the pipeline associated with this job is 25647. All of the processes in a single pipeline are members of the same job. Bash uses the job abstraction as the foundation for job control. 👑

To make the whole user interface of job control possible, each process has a process group ID, and the operating system keeps track of a current terminal process group ID. That terminal process group ID is associated with the controlling terminal.

Processes with the same process group ID are said to be part of the same process group. Members of the foreground process group — the processes whose process group ID equals the current terminal process group ID — receive keyboard-generated signals like SIGINT. Processes in the foreground process group are called foreground processes. Background processes are the ones whose process group ID differs from the controlling terminal's; those babies are immune to keyboard-generated signals. 🛡️ Only foreground processes are allowed to read from — or, if the user specifies it with stty tostop, write to — the controlling terminal. The system sends a SIGTTIN (SIGTTOU) signal to background processes that try to read from (write to, when tostop is in effect) the terminal, and unless that signal is caught, it suspends the process.

If the operating system bash is running on supports job control, bash contains facilities to use it. Typing the suspend character (typically ^Z, Control-Z) while a process is running stops that process and hands control back to bash. Typing the delayed suspend character (typically ^Y, Control-Y) makes the process stop when it tries to read input from the terminal, and returns control to bash. Then you get to mess with the job's state: bg to continue it in the background, fg to continue it in the foreground, or kill to kill it. The suspend character takes effect immediately, and has the bonus side effect of discarding any pending output and typeahead. If you need to force a background process to stop, or stop a process that's not associated with the current terminal session, send it the SIGSTOP signal using kill. 😌

There are a bunch of ways to refer to a job in the shell. The % character introduces a job specification (jobspec).

Job number n can be referred to as %n. You can also refer to a job using a prefix of the name used to start it, or using a substring that shows up in its command line. For example, %ce refers to a job whose command name begins with ce. Whereas %?ce refers to any job containing the string ce in its command line. If that prefix or substring matches more than one job, bash reports an error — pick a lane bestie 😭

The symbols %% and %+ refer to the shell's notion of the current job. A single % (with no jobspec after it) also refers to the current job. %- refers to the previous job. When a job starts in the background, or a job stops while in the foreground, or a job is resumed in the background, it becomes the current job. The job that was the current job becomes the previous job. When the current job terminates, the previous job becomes the current job. If there's only a single job, both %+ and %- can be used to refer to it. And in output pertaining to jobs (like the output of the jobs command), the current job is always marked with a +, and the previous job with a -. 💖

Just naming a job brings it into the foreground: %1 is a synonym for fg %1, pulling job 1 from the background into the foreground. Similarly, %1 & resumes job 1 in the background, equivalent to bg %1. so convenient fr fr

The shell finds out immediately whenever a job changes state. Normally, bash waits until it's about to print a prompt before telling you about a job's status change, so it doesn't interrupt any other output — though it will notify you about status changes after a foreground command in a list completes, before executing the next command in the list. If the -b option to the set builtin is enabled, bash reports status changes immediately. And bash executes any trap on SIGCHLD for each child that terminates.

When a job terminates and bash notifies you about it, bash removes that job from the table. It won't show up in jobs output anymore, but wait will still report its exit status — as long as you hand it the process ID associated with the job as an argument. When the table is empty, job numbers start over at 1. 🔄

If a user tries to exit bash while jobs are stopped (or, if the checkjobs shell option has been enabled with the shopt builtin, while jobs are running), the shell prints a warning message, and if the checkjobs option is enabled, it lists the jobs and their statuses. Then you can use the jobs command to inspect their status. If the user immediately tries to exit again, with no command in between, bash does not print another warning, and it terminates any stopped jobs. 🫡

When the shell is waiting for a job or process using the wait builtin, and job control is enabled, wait returns when the job changes state. The -f option makes wait wait until the job or process actually terminates before returning. patience is a virtue or whatever ✨

✨ PROMPTING ✨

okay so when you're running bash interactively (like, actually typing at it, bestie, not scripting), bash shows the primary prompt PS1 whenever it's ready for you to type a command 💖. And when it needs more input to finish a command? That's the secondary prompt PS2 stepping in 🫶.

Bash peeks at the array variable PROMPT_COMMAND right before it prints every primary prompt 👀. If any of the elements in PROMPT_COMMAND are set and non-null, bash runs each one — in numeric order, no cap — exactly as if you'd typed it on the command line 🔥. Also, bash displays PS0 after it reads a command but before it executes it 👑.

And PS4? Bash displays that as described above before tracing each command when the -x option is enabled 💅.

Bash lets you customize the prompt strings PS0, PS1, PS2, and PS4 by dropping in a bunch of backslash-escaped special characters, which decode like this 🌈:

\a — an ASCII bell character (07). 🔔
\d — the date in "Weekday Month Date" format (like, "Tue May 26"). 📅
\D{format} — the format gets passed to strftime(3) and the result gets inserted into the prompt string; an empty format gives you a locale-specific time representation. The braces are required, bestie, don't skip them. ⏰
\e — an ASCII escape character (033).
\h — the hostname up to the first "." 🏠
\H — the hostname.
\j — the number of jobs currently managed by the shell. 💼
\l — the basename of the shell's terminal device name (like "ttys0").
\n — a newline.
\r — a carriage return.
\s — the name of the shell: the basename of $0 (the portion following the final slash).
\t — the current time in 24-hour HH:MM:SS format. 🕐
\T — the current time in 12-hour HH:MM:SS format.
\@ — the current time in 12-hour am/pm format.
\A — the current time in 24-hour HH:MM format.
\u — the username of the current user. 🫶
\v — the bash version (like, 2.00).
\V — the bash release, version + patch level (like, 2.00.0).
\w — the value of the PWD shell variable ($PWD), with $HOME abbreviated with a tilde (uses the value of the PROMPT_DIRTRIM variable).
\W — the basename of $PWD, with $HOME abbreviated with a tilde.
\! — the history number of this command.
\# — the command number of this command.
\$ — if the effective UID is 0, a #, otherwise a $.
\nnn — the character corresponding to the octal number nnn.
\\ — a backslash.
\[ — begin a sequence of non-printing characters, which could be used to embed a terminal control sequence into the prompt.
\] — end a sequence of non-printing characters.

The command number and the history number are usually different, and this trips people up so listen up 🫶: the history number of a command is its position in the history list, which may include commands restored from the history file (check out HISTORY below), while the command number is the position in the sequence of commands executed during the current shell session. After the string is decoded, it gets expanded via parameter expansion, command substitution, arithmetic expansion, and quote removal — subject to the value of the promptvars shell option (see the description of the shopt command under SHELL BUILTIN COMMANDS below).

⚠️ This can have unwanted side effects if escaped portions of the string appear within command substitution or contain characters special to word expansion. So like, be careful, fr fr. 💀

📖 READLINE 📖

This is the library that handles reading input when you're using an interactive shell — unless you passed the --noediting option at shell invocation 🚫. Line editing is also used when you use the -e option to the read builtin. By default, the line editing commands are similar to those of emacs; a vi-style line editing interface is also available 💅. You can enable line editing at literally any time using the -o emacs or -o vi options to the set builtin (see SHELL BUILTIN COMMANDS below). And to turn line editing off after the shell is running, use the +o emacs or +o vi options to the set builtin.

Readline Notation 🎀

This section uses Emacs-style editing concepts and its notation for keystrokes ✨. Control keys are denoted by C-key, so like C-n means Control-N. Similarly, meta keys are denoted by M-key, so M-x means Meta-X. The Meta key is often labeled "Alt" or "Option" 🧚.

On keyboards without a Meta key, M-x means ESC x — that is, press and release the Escape key, then press and release the x key, in sequence. This makes ESC the meta prefix 👑. The combination M-C-x means ESC Control-x: press and release the Escape key, then press and hold the Control key while pressing the x key, then release both.

On some keyboards, the Meta key modifier produces characters with the eighth bit (0200) set. You can use the enable-meta-key variable to control whether or not it does this, if the keyboard allows it. On many others, the terminal or terminal emulator converts the metafied key to a key sequence beginning with ESC, as described in the preceding paragraph 🔍.

If your Meta key produces a key sequence with the ESC meta prefix, you can make M-key key bindings you specify (see Readline Key Bindings below) do the same thing by setting the force-meta-prefix variable 💖.

Readline commands may be given numeric arguments, which normally act as a repeat count. Sometimes, however, it's the sign of the argument that's significant 👀. Passing a negative argument to a command that acts in the forward direction (like kill-line) makes that command act in a backward direction. Commands whose behavior with arguments deviates from this are noted below, don't worry.

The point is the current cursor position, and mark refers to a saved cursor position. The text between the point and mark is referred to as the region 🌸. Readline has the concept of an active region: when the region is active, readline redisplay highlights the region using the value of the active-region-start-color variable. The enable-active-region variable turns this on and off. Several commands set the region to active; those are noted below ✨.

When a command is described as killing text, the text deleted is saved for possible future retrieval (yanking). The killed text is saved in a kill ring 💀. Consecutive kills accumulate the deleted text into one unit, which can be yanked all at once. Commands which do not kill text separate the chunks of text on the kill ring.

Readline Initialization 🦋

Readline gets customized by putting commands in an initialization file (the inputrc file) 📝. The name of this file is taken from the value of the INPUTRC shell variable. If that variable is unset, the default is ~/.inputrc. If that file does not exist or can't be read, readline looks for /etc/inputrc 🔍. When a program that uses the readline library starts up, readline reads the initialization file and sets the key bindings and variables found there, before reading any user input.

There are only a few basic constructs allowed in the inputrc file:

Blank lines are ignored.
Lines beginning with a # are comments.
Lines beginning with a $ indicate conditional constructs.
Other lines denote key bindings and variable settings.

The default key-bindings in this section may be changed using key binding commands in the inputrc file. Programs that use the readline library, including bash, may add their own commands and bindings 👑.

For example, placing

M-Control-u: universal-argument

C-Meta-u: universal-argument

into the inputrc would make M-C-u execute the readline command universal-argument 🔥.

Key bindings may contain the following symbolic character names: DEL, ESC, ESCAPE, LFD, NEWLINE, RET, RETURN, RUBOUT (a destructive backspace), SPACE, SPC, and TAB ✨.

In addition to command names, readline allows keys to be bound to a string that gets inserted when the key is pressed (a macro) 🎀. The difference between a macro and a command is that a macro is enclosed in single or double quotes.

Readline Key Bindings 💅

The syntax for controlling key bindings in the inputrc file is simple, bestie. All you need is the name of the command or the text of a macro, plus a key sequence to bind it to 💖. The key sequence may be specified in one of two ways: as a symbolic key name, possibly with Meta- or Control- prefixes, or as a key sequence composed of one or more characters enclosed in double quotes. The key sequence and name are separated by a colon. And there can be no whitespace between the name and the colon 🚫.

When using the form keyname:function-name or macro, keyname is the name of a key spelled out in English. For example:

Control-u: universal-argument
Meta-Rubout: backward-kill-word
Control-o: "> output"

In the above example, C-u is bound to the function universal-argument, M-DEL is bound to the function backward-kill-word, and C-o is bound to run the macro expressed on the right hand side (that is, to insert the text "> output" into the line) ✨.

In the second form, "keyseq":function-name or macro, keyseq differs from keyname above in that strings denoting an entire key sequence may be specified by placing the sequence within double quotes 👀. Some GNU Emacs style key escapes can be used, as in the following example, but none of the symbolic character names are recognized:

"\C-u": universal-argument
"\C-x\C-r": re-read-init-file
"\e[11~": "Function Key 1"

In this example, C-u is again bound to the function universal-argument. C-x C-r is bound to the function re-read-init-file, and ESC [ 1 1 ~ is bound to insert the text "Function Key 1" 🔥.

The full set of GNU Emacs style escape sequences available when specifying key sequences is:

\C- — a control prefix.
\M- — adding the meta prefix or converting the following character to a meta character, as described below under force-meta-prefix.
\e — an escape character.
\\ — backslash.
\" — literal ", a double quote.
\' — literal ', a single quote.

In addition to the GNU Emacs style escape sequences, a second set of backslash escapes is available 🌈:

\a — alert (bell) 🔔
\b — backspace
\d — delete
\f — form feed
\n — newline
\r — carriage return
\t — horizontal tab
\v — vertical tab
\nnn — the eight-bit character whose value is the octal value nnn (one to three digits).
\xHH — the eight-bit character whose value is the hexadecimal value HH (one or two hex digits).

Okay so when you're typing in the text of a macro, you have to use single or double quotes to show it's a macro definition. If you leave it unquoted, readline just assumes you're naming a function. 🙈 And those backslash escapes we described above? They get expanded inside the macro body. A backslash also quotes literally any other character in the macro text — " and ' included. So the backslash really is that girl. 💅

Bash will display or modify your current readline key bindings with the bind builtin command. And the -o emacs or -o vi options to the set builtin (peep SHELL BUILTIN COMMANDS below) are what change the editing mode while you're in interactive use. 🔥

✨ Readline Variables ✨

Readline has variables, bestie — they exist so you can customize its behavior even further. You can set one in the inputrc file with a statement that looks like

set variable-name value

or by using the bind builtin command (yep, SHELL BUILTIN COMMANDS below, iykyk).

Unless it says otherwise, readline variables can take the values On or Off — and case does not matter, like at all. Unrecognized variable names? Ignored. 💀 And when readline reads a variable value, empty or null values, on (case-insensitive), and 1 are all equivalent to On. Every other value is equivalent to Off.

The bind -V command lists the current readline variable names and their values (see SHELL BUILTIN COMMANDS below). 👑

The variables and their default values are: 🎀

active-region-start-color — a string variable that controls the text color and background when readline is displaying text in the active region (peep the description of enable-active-region down below 👇). Big rule bestie: this string must not take up any physical character positions on the display, so it should be made of terminal escape sequences and literally nothing else. It gets output to the terminal before the active-region text shows up. It's reset to its default whenever the terminal type changes. The default is the string that puts your terminal in standout mode, pulled straight from the terminal's terminfo description. A sample value might be “\e[01;33m”. ✨
active-region-end-color — the string that “undoes” everything active-region-start-color did, restoring “normal” terminal display appearance after the active-region text is done. Same deal as its bestie above: no physical character positions, only terminal escape sequences. It's output to the terminal after the active-region text. Also reset to default whenever the terminal type changes. Default is the string that restores the terminal from standout mode, again from terminfo. A sample value might be “\e[0m”. 🫶
bell-style (default: audible) — controls what readline does when it wants to ring the terminal bell. Set it to none and readline will never ring the bell. Set it to visible and readline uses a visible bell if one's available. Set it to audible and readline attempts to ring the terminal's actual bell. 🔔
bind-tty-special-chars (default: On) — if it's On, readline attempts to bind the control characters that the kernel's terminal driver treats specially to their readline equivalents. These override the default readline bindings described here. Type stty -a at a bash prompt to see your current terminal settings, including the special control characters (usually cchars). iykyk 💅
blink-matching-paren (default: Off) — if On, readline attempts to briefly move the cursor to an opening parenthesis when a closing parenthesis gets inserted. Cute little cursor gymnastics. 🎀
colored-completion-prefix (default: Off) — if On, when listing completions, readline displays the common prefix of the set of possible completions using a different color. The color definitions come from the value of the LS_COLORS environment variable. And if there's a color definition in $LS_COLORS for the custom suffix “.readline-colored-completion-prefix”, readline uses that color for the common prefix instead of its default. obsessed. 🌈
colored-stats (default: Off) — if On, readline displays possible completions using different colors to indicate their file type. Color definitions come from the value of the LS_COLORS environment variable. 👑
comment-begin (default: “#”) — the string that the readline insert-comment command inserts. That command is bound to M-# in emacs mode and to # in vi command mode. 📝
completion-display-width (default: -1) — the number of screen columns used to display possible matches when performing completion. The value is ignored if it's less than 0 or greater than the terminal screen width. A value of 0 makes matches display one per line. Default is -1.
completion-ignore-case (default: Off) — if On, readline performs filename matching and completion in a case-insensitive fashion. no more yelling at your files fr fr 🔥
completion-map-case (default: Off) — if On, and completion-ignore-case is enabled, readline treats hyphens (-) and underscores (_) as equivalent when doing case-insensitive filename matching and completion. iconic behaviour tbh 💜
completion-prefix-display-length (default: 0) — the maximum length in characters of the common prefix of a list of possible completions that gets displayed without modification. Set it to anything greater than zero and readline replaces common prefixes longer than that value with an ellipsis when displaying possible completions. Oh and if a completion begins with a period, and readline is completing filenames, it uses three underscores instead of an ellipsis. 😭
completion-query-items (default: 100) — this is the one that decides when the user gets asked about viewing the number of possible completions generated by the possible-completions command. It can be any integer value greater than or equal to zero. If the number of possible completions is greater than or equal to the value of this variable, readline asks whether or not you wish to view them; otherwise readline just lists them on the terminal. A zero value means readline should never ask; negative values are treated as zero. 🙈
convert-meta (default: On) — if On, readline converts characters it reads that have the eighth bit set to an ASCII key sequence by clearing the eighth bit and prefixing it with an escape character (aka converting the character to have the meta prefix). Default is On, but readline sets it to Off if the locale contains characters whose encodings may include bytes with the eighth bit set. This variable depends on the LC_CTYPE locale category, and may change if the locale changes. It also affects key bindings — see the description of force-meta-prefix below. 👀
disable-completion (default: Off) — if On, readline inhibits word completion. Completion characters get inserted into the line as if they had been mapped to self-insert. 💀
echo-control-characters (default: On) — when On, on operating systems that indicate they support it, readline echoes a character corresponding to a signal generated from the keyboard.
editing-mode (default: emacs) — controls whether readline uses a set of key bindings similar to Emacs or vi. editing-mode can be set to either emacs or vi. pick your fighter ⚔️
emacs-mode-string (default: @) — if the show-mode-in-prompt variable is enabled, this string is displayed immediately before the last line of the primary prompt when emacs editing mode is active. The value is expanded like a key binding, so the standard set of meta- and control- prefixes and backslash escape sequences are all available to you. The \1 and \2 escapes begin and end sequences of non-printing characters, which you can use to embed a terminal control sequence into the mode string. fancy 💅
enable-active-region (default: On) — when this variable is On, readline allows certain commands to designate the region as active. When the region is active, readline highlights the text in it using the value of the active-region-start-color variable, which defaults to the string that enables the terminal's standout mode. The active region shows the text inserted by bracketed-paste and any matching text found by incremental and non-incremental history searches. ✨
enable-bracketed-paste (default: On) — when On, readline configures the terminal to insert each paste into the editing buffer as a single string of characters, instead of treating each character as if it had been read from the keyboard. This is called bracketed-paste mode, and it stops readline from executing any editing commands bound to key sequences that happen to appear in the pasted text. genuinely a lifesaver 🫶
enable-keypad (default: Off) — when On, readline tries to enable the application keypad when it's called. Some systems need this to enable the arrow keys. ➡️
enable-meta-key (default: On) — when On, readline tries to enable any meta modifier key the terminal claims to support. On many terminals the Meta key is used to send eight-bit characters; this variable checks for the terminal capability that indicates the terminal can enable and disable a mode that sets the eighth bit of a character (0200) if the Meta key is held down when the character is typed (a meta character).
expand-tilde (default: Off) — if On, readline performs tilde expansion when it attempts word completion. ~ girlies rise 🌸
force-meta-prefix (default: Off) — if On, readline modifies its behavior when binding key sequences containing \M- or Meta- (see Key Bindings above) by converting a key sequence of the form \M-C or Meta-C to the two-character sequence ESC C (adding the meta prefix). If force-meta-prefix is set to Off (the default), readline uses the value of the convert-meta variable to determine whether to perform this conversion: if convert-meta is On, readline performs the conversion described above; if it's Off, readline converts C to a meta character by setting the eighth bit (0200). 😵‍💫
history-preserve-point (default: Off) — if On, the history code attempts to place point at the same location on each history line retrieved with previous-history or next-history.
history-size (default: unset) — sets the maximum number of history entries saved in the history list. Set it to zero and any existing history entries get deleted and no new entries are saved. Set it to a value less than zero and the number of history entries is not limited. By default, bash sets the maximum number of history entries to the value of the HISTSIZE shell variable. And setting history-size to a non-numeric value will set the maximum number of history entries to 500. 💀
horizontal-scroll-mode (default: Off) — setting this variable to On makes readline use a single line for display, scrolling the input horizontally on a single screen line when it becomes longer than the screen width rather than wrapping to a new line. This setting is automatically enabled for terminals of height 1. 🦋
input-meta (default: Off) — if On, readline enables eight-bit input (that is, it does not clear the eighth bit in the characters it reads), regardless of what the terminal claims it can support. The default is Off, but readline sets it to On if the locale contains characters whose encodings may include bytes with the eighth bit set. This variable depends on the LC_CTYPE locale category, and its value may change if the locale changes. The name meta-flag is a synonym for input-meta. 💜
isearch-terminators (default: “C-[C-j”)
The string of characters that should terminate an incremental search without subsequently executing that character as a command. If this variable hasn't been given a value, the characters ESC and C-j are what terminate an incremental search. 🛑
keymap (emacs) — sets the current readline keymap. The set of valid keymap names is emacs, emacs-standard, emacs-meta, emacs-ctlx, vi, vi-command, and vi-insert. Heads up: vi is equivalent to vi-command, and emacs is equivalent to emacs-standard. Default value is emacs, and the value of editing-mode also affects the default keymap. 👑
keyseq-timeout (500) — how long readline waits for a character when it's reading an ambiguous key sequence, i.e. one that could form a complete key sequence using the input read so far, or could take additional input to complete a longer key sequence. If readline doesn't receive any input within the timeout, it just uses the shorter but complete key sequence. Value is in milliseconds, so 1000 means readline waits one whole second for extra input. If this variable is set to a value less than or equal to zero, or to a non-numeric value, readline waits until another key is pressed to decide which key sequence to complete. ⏳
mark-directories (On) — if set to On, completed directory names get a slash appended. 📁
mark-modified-lines (Off) — if set to On, readline displays history lines that have been modified with a preceding asterisk (*).
mark-symlinked-directories (Off) — if set to On, completed names which are symbolic links to directories have a slash appended, subject to the value of mark-directories. 🔗
match-hidden-files (On) — when set to On, this variable forces readline to match files whose names begin with a "." (hidden files) when performing filename completion. If set to Off, the user has to include that leading "." in the filename to be completed. 🙈
menu-complete-display-prefix (Off) — if set to On, menu completion displays the common prefix of the list of possible completions (which may be empty) before cycling through the list.
output-meta (Off) — if set to On, readline displays characters with the eighth bit set directly rather than as a meta-prefixed escape sequence. Default is Off, but readline sets it to On if the locale contains characters whose encodings may include bytes with the eighth bit set. This variable is dependent on the LC_CTYPE locale category, and its value may change if the locale changes. 🌈
page-completions (On) — if set to On, readline uses an internal pager resembling more(1) to display a screenful of possible completions at a time.
prefer-visible-bell — see bell-style. 🔔
print-completions-horizontally (Off) — if set to On, readline displays completions with matches sorted horizontally in alphabetical order, rather than down the screen.
revert-all-at-newline (Off) — if set to On, readline will undo all changes to history lines before returning when executing accept-line. By default, history lines may be modified and retain individual undo lists across calls to readline.
search-ignore-case (Off) — if set to On, readline performs incremental and non-incremental history list searches in a case-insensitive fashion.
show-all-if-ambiguous (Off) — this alters the default behavior of the completion functions. If set to On, words which have more than one possible completion cause the matches to be listed immediately instead of ringing the bell. 🔔➡️📜
show-all-if-unmodified (Off) — this alters the default behavior of the completion functions in a fashion similar to show-all-if-ambiguous. If set to On, words which have more than one possible completion without any possible partial completion (meaning the possible completions don't share a common prefix) cause the matches to be listed immediately instead of ringing the bell.
show-mode-in-prompt (Off) — if set to On, add a string to the beginning of the prompt indicating the editing mode: emacs, vi command, or vi insertion. The mode strings are user-settable (e.g., emacs-mode-string).
skip-completed-text (Off) — if set to On, this alters the default completion behavior when inserting a single match into the line. It's only active when performing completion in the middle of a word. If enabled, readline does not insert characters from the completion that match characters after point in the word being completed, so portions of the word following the cursor don't get duplicated. ✨
vi-cmd-mode-string ((cmd)) — if the show-mode-in-prompt variable is enabled, this string is displayed immediately before the last line of the primary prompt when vi editing mode is active and in command mode. The value is expanded like a key binding, so the standard set of meta- and control- prefixes and backslash escape sequences is available. The \1 and \2 escapes begin and end sequences of non-printing characters, which can be used to embed a terminal control sequence into the mode string. 💅
vi-ins-mode-string ((ins)) — if the show-mode-in-prompt variable is enabled, this string is displayed immediately before the last line of the primary prompt when vi editing mode is active and in insertion mode. The value is expanded like a key binding, so the standard set of meta- and control- prefixes and backslash escape sequences is available. The \1 and \2 escapes begin and end sequences of non-printing characters, which can be used to embed a terminal control sequence into the mode string. 🦋
visible-stats (Off) — if set to On, a character denoting a file's type as reported by stat(2) is appended to the filename when listing possible completions. ⭐

🎀 Readline Conditional Constructs 🎀

Readline implements a facility similar in spirit to the conditional compilation features of the C preprocessor, which lets key bindings and variable settings be performed as the result of tests. There are four parser directives available. 🔥

$if — the $if construct allows bindings to be made based on the editing mode, the terminal being used, or the application using readline. The text of the test, after any comparison operator, extends to the end of the line; unless otherwise noted, no characters are required to isolate it.
- mode — the mode= form of the $if directive is used to test whether readline is in emacs or vi mode. This may be used in conjunction with the set keymap command, for instance, to set bindings in the emacs-standard and emacs-ctlx keymaps only if readline is starting out in emacs mode.
- term — the term= form may be used to include terminal-specific key bindings, perhaps to bind the key sequences output by the terminal's function keys. The word on the right side of the = is tested against both the full name of the terminal and the portion of the terminal name before the first -. This allows xterm to match both xterm and xterm-256color, for instance. 🖥️
- version — the version test may be used to perform comparisons against specific readline versions. The version expands to the current readline version. The set of comparison operators includes =, (and ==), !=, <=, >=, <, and >. The version number supplied on the right side of the operator consists of a major version number, an optional decimal point, and an optional minor version (e.g., 7.1). If the minor version is omitted, it defaults to 0. The operator may be separated from the string version and from the version number argument by whitespace.
- application — the application construct is used to include application-specific settings. Each program using the readline library sets the application name, and an initialization file can test for a particular value. This could be used to bind key sequences to functions useful for a specific program. For instance, the following command adds a key sequence that quotes the current or previous word in bash:

$if Bash
# Quote the current or previous word
"\C-xq": "\eb\"\ef\""
$endif

variable — the variable construct provides simple equality tests for readline variables and values. The permitted comparison operators are =, ==, and !=. The variable name must be separated from the comparison operator by whitespace; the operator may be separated from the value on the right hand side by whitespace. String and boolean variables may be tested. Boolean variables must be tested against the values on and off.

$else — commands in this branch of the $if directive are executed if the test fails. 🙃
$endif — this command, as seen in the previous example, terminates an $if command.
$include — this directive takes a single filename as an argument and reads commands and key bindings from that file. For example, the following directive would read /etc/inputrc:

$include  /etc/inputrc

🔍 Searching 🔍

Readline provides commands for searching through the command history (see HISTORY below) for lines containing a specified string. There are two search modes: incremental and non-incremental. 🕵️‍♀️

okay so bestie, incremental searches? they start before you've even finished typing the search string 💅. literally as each character of the search string is typed, readline displays the next entry from the history matching the string typed so far. the whole point is efficiency — an incremental search requires only as many characters as needed to find the desired history entry. so few. so slay.

when you're in emacs editing mode, type C-r to search backward in the history for a particular string. typing C-s searches forward through the history. the characters present in the value of the isearch-terminators variable are used to terminate an incremental search. if that variable has not been assigned a value, ESC and C-j terminate an incremental search. C-g aborts an incremental search and restores the original line. and when the search is terminated, the history entry containing the search string becomes the current line. ✨

to find other matching entries in the history list, type C-r or C-s as appropriate — this searches backward or forward in the history for the next entry matching the search string typed so far. any other key sequence bound to a readline command terminates the search and executes that command. for instance, a newline terminates the search and accepts the line, thereby executing the command from the history list. a movement command will terminate the search, make the last line found the current line, and begin editing. 🦋

readline remembers the last incremental search string. so if two C-rs are typed without any intervening characters defining a new search string, readline uses any remembered search string. it's giving memory, periodt.

non-incremental searches, on the other hand, read the entire search string before starting to search for matching history entries. the search string may be typed by the user or be part of the contents of the current line. 💖

🌈 Readline Command Names 🌈

here's the whole list of the names of the commands and the default key sequences to which they're bound. command names without an accompanying key sequence are unbound by default — yes, they just be floating out there, unbonded, single, iykyk 🙈.

and real quick, vocab check so we're all on the same page: point refers to the current cursor position, and mark refers to a cursor position saved by the set-mark command. the text between the point and mark is referred to as the region. readline has the concept of an active region: when the region is active, readline redisplay highlights the region using the value of the active-region-start-color variable. the enable-active-region readline variable turns this on and off. several commands set the region to active — those are noted below. 👑

🎀 Commands for Moving 🎀

beginning-of-line (C-a) — move to the start of the current line. this may also be bound to the Home key on some keyboards. 🏠
end-of-line (C-e) — move to the end of the line. this may also be bound to the End key on some keyboards.
forward-char (C-f) — move forward a character. this may also be bound to the right arrow key on some keyboards. ➡️
backward-char (C-b) — move back a character. this may also be bound to the left arrow key on some keyboards. ⬅️
forward-word (M-f) — move forward to the end of the next word. words are composed of alphanumeric characters (letters and digits).
backward-word (M-b) — move back to the start of the current or previous word. words are composed of alphanumeric characters (letters and digits).
shell-forward-word (M-C-f) — move forward to the end of the next word. words are delimited by non-quoted shell metacharacters. 🐚
shell-backward-word (M-C-b) — move back to the start of the current or previous word. words are delimited by non-quoted shell metacharacters. 🐚
previous-screen-line — attempt to move point to the same physical screen column on the previous physical screen line. ngl, this will not have the desired effect if the current readline line does not take up more than one physical line, or if point is not greater than the length of the prompt plus the screen width. 💀
next-screen-line — attempt to move point to the same physical screen column on the next physical screen line. again, this will not have the desired effect if the current readline line does not take up more than one physical line, or if the length of the current readline line is not greater than the length of the prompt plus the screen width. 💀
clear-display (M-C-l) — clear the screen and, if possible, the terminal's scrollback buffer, then redraw the current line, leaving the current line at the top of the screen. the scrollback ate. 🧹
clear-screen (C-l) — clear the screen, then redraw the current line, leaving the current line at the top of the screen. with a numeric argument, refresh the current line without clearing the screen. 🌸
redraw-current-line — refresh the current line. ✨

💖 Commands for Manipulating the History 💖

ok so these are the readline commands that play with your history list, aka the ✨archive of everything you've ever typed✨. buckle up bestie:

accept-line (Newline, Return) — takes the line, periodt, doesn't matter where your cursor is chilling. if the line isn't empty, it gets yeeted into the history list based on whatever HISTCONTROL and HISTIGNORE are set to. and if the line was a modified history line? it gets restored back to its original state, no cap.
previous-history (C-p) — pulls the previous command out of the history list, stepping back through it. lowkey might also be bound to the up arrow key on some keyboards.
next-history (C-n) — grabs the next command from the history list, stepping forward. and yeah, this one might be bound to the down arrow on some keyboards too.
beginning-of-history (M-<) — zoom to the very first line in the history.
end-of-history (M->) — jump to the end of the input history, meaning the line you're literally typing right now.
operate-and-get-next (C-o) — accepts the current line for execution as if you'd hit newline, then fetches the next line relative to the current one from history for editing. if you hand it a numeric argument, that specifies which history entry to use instead of the current line. iconic for re-running stuff.
fetch-history — give it a numeric argument and it fetches that entry from the history list, making it your current line. no argument? it just moves back to the first entry in the history list.
reverse-search-history (C-r) — searches backward, starting at the current line and moving "up" through the history as needed. this is an incremental search, bestie. it sets the region to the matched text and activates the region.
forward-search-history (C-s) — searches forward, starting at the current line and moving "down" through the history as needed. also incremental. also sets the region to the matched text and activates it.
non-incremental-reverse-search-history (M-p) — searches backward through the history from the current line using a non-incremental search for a string you supply. the search string can match anywhere in a history line.
non-incremental-forward-search-history (M-n) — searches forward through the history with a non-incremental search for a string you supply. again, the string can match anywhere in a history line.
history-search-backward — searches backward through the history for the string of characters sitting between the start of the current line and the point. here the search string must match at the beginning of a history line. it's a non-incremental search. might be bound to the Page Up key on some keyboards. 👑
history-search-forward — searches forward through the history for the string between the start of the current line and the point. string must match at the beginning of a history line. non-incremental. might be bound to Page Down on some keyboards.
history-substring-search-backward — searches backward through the history for the string of characters between the start of the current line and the point. here the search string may match anywhere in a history line. non-incremental search, fr fr.
history-substring-search-forward — searches forward through the history for the string between the start of the current line and the point. string may match anywhere in a history line. non-incremental.
yank-nth-arg (M-C-y) — drops the first argument of the previous command (usually the second word on the previous line) in at point. with an argument n, it inserts the nth word from the previous command — and heads up, the words in the previous command start counting at word 0. a negative argument inserts the nth word from the end of the previous command. once that argument n is figured out, it uses the history expansion facilities to extract the *n*th word, exactly as if you'd typed the !n history expansion. 💅
yank-last-arg (M-., M-_) — inserts the last argument of the previous command (the last word of the previous history entry). with a numeric argument it behaves exactly like yank-nth-arg. successive calls to yank-last-arg move back through the history list, inserting the last word (or the word specified by the argument to the first call) of each line in turn. any numeric argument you supply to these successive calls decides which direction to move through the history. a negative argument flips the direction (back or forward). this uses the history expansion facilities to extract the last word, as if the !$ history expansion had been specified.
shell-expand-line (M-C-e) — expands the line by performing shell word expansions. that means it does alias and history expansion, $'string' and $"string" quoting, tilde expansion, parameter and variable expansion, arithmetic expansion, command and process substitution, word splitting, and quote removal. an explicit argument suppresses command and process substitution. peep HISTORY EXPANSION below for the history expansion deets.
history-expand-line (M-^) — performs history expansion on the current line. see HISTORY EXPANSION below for what that actually means.
magic-space — performs history expansion on the current line and then inserts a space. see HISTORY EXPANSION below.
alias-expand-line — performs alias expansion on the current line. see ALIASES above for the whole alias expansion situation.
history-and-alias-expand-line — performs both history and alias expansion on the current line. two-for-one, slay.
insert-last-argument (M-., M-_) — literally just a synonym for yank-last-arg. same energy, different name.
edit-and-execute-command (C-x C-e) — invokes an editor on the current command line and executes the result as shell commands. bash tries to invoke $VISUAL, then $EDITOR, then emacs as the editor, in that order. 🦋

✨ Commands for Changing Text ✨

now we're in our text-editing era, bestie. these ones move, flip, and rewrite the characters on your line:

end-of-file (usually C-d) — the character indicating end-of-file, as set for example by stty(1). if this character gets read when there are no characters on the line and point is at the beginning of the line, readline interprets it as the end of input and returns EOF. 🫶
delete-char (C-d) — deletes the character at point. if this function is bound to the same character as the tty EOF character — which C-d commonly is — then see above for the effects, because things get spicy. this may also be bound to the Delete key on some keyboards.
backward-delete-char (Rubout) — deletes the character behind the cursor. when you give it a numeric argument, it saves the deleted text on the kill ring. recycling, we love to see it.
forward-backward-delete-char — deletes the character under the cursor, unless the cursor is at the end of the line, in which case it deletes the character behind the cursor instead. flexible queen.
quoted-insert (C-q, C-v) — adds the next character you type to the line verbatim. this is how you insert characters like C-q, for example.
tab-insert (C-v TAB) — inserts a tab character. that's it, that's the whole job.
self-insert (a, b, A, 1, !, ...) — inserts the character typed. the most basic main-character move there is.
bracketed-paste-begin — this function is intended to be bound to the "bracketed paste" escape sequence that some terminals send, and such a binding is assigned by default. it lets readline insert the pasted text as a single unit instead of treating every character as if it had been read from the keyboard. the pasted characters get inserted as if each one were bound to self-insert rather than executing any editing commands. bracketed paste sets the region to the inserted text and activates the region. no more cursed paste chaos, we're healing. 🌸
transpose-chars (C-t) — drags the character before point forward over the character at point, moving point forward too. if point is at the end of the line, it transposes the two characters before point. negative arguments do nothing at all. 💀
transpose-words (M-t) — drags the word before point past the word after point, moving point past that word as well. if point is at the end of the line, it transposes the last two words on the line.
shell-transpose-words (M-C-t) — drags the word before point past the word after point, moving point past that word as well. if the insertion point is at the end of the line, it transposes the last two words on the line. word boundaries are the same as shell-forward-word and shell-backward-word. 🐚
upcase-word (M-u) — uppercases the current (or following) word. with a negative argument, it uppercases the previous word but does not move point. LOUD.
downcase-word (M-l) — lowercases the current (or following) word. with a negative argument, it lowercases the previous word but does not move point. quiet.
capitalize-word (M-c) — capitalizes the current (or following) word. with a negative argument, it capitalizes the previous word but does not move point. Fancy.
overwrite-mode — toggles overwrite mode. with an explicit positive numeric argument, it switches to overwrite mode. with an explicit non-positive numeric argument, it switches to insert mode. this command affects only emacs mode — vi mode does overwrite differently, so don't come for me. each call to readline() starts in insert mode. in overwrite mode, characters bound to self-insert replace the text at point rather than pushing the text to the right. characters bound to backward-delete-char replace the character before point with a space. by default this command is unbound, but it may be bound to the Insert key on some keyboards.

ngl the overwrite-mode space-replacement thing trips people up every single time, so clock it: it's a space, not a delete. 💅

💀 Killing and Yanking 💀

okay bestie, let's talk about killing and yanking — but like, in the readline sense, not true crime 💀. This is all about cutting and pasting text in your terminal like a main character. ✨

kill-line (C-k) — kills (aka cuts) everything from point (that's your cursor, bestie) to the end of the current line. If you give it a negative numeric argument, it kills backward from the cursor to the beginning of the line. 💅
backward-kill-line (C-x Rubout) — kills backward to the beginning of the current line. With a negative numeric argument, it kills forward from the cursor to the end of the line. 🔥
unix-line-discard (C-u) — kills backward from point to the beginning of the line, and saves the killed text on the kill-ring. 🎀
kill-whole-line — kills all characters on the current line, no matter where point is. periodt. 💀
kill-word (M-d) — kills from point to the end of the current word, or if you're between words, to the end of the next word. Word boundaries are the same as those used by forward-word. 👑
backward-kill-word (M-Rubout) — kills the word behind point. Word boundaries are the same as those used by backward-word. 🦋
shell-kill-word (M-C-d) — kills from point to the end of the current word, or if between words, to the end of the next word. Word boundaries are the same as those used by shell-forward-word. 🌸
shell-backward-kill-word — kills the word behind point. Word boundaries are the same as those used by shell-backward-word. 💜
unix-word-rubout (C-w) — kills the word behind point, using white space as a word boundary, saving the killed text on the kill-ring. ✨
unix-filename-rubout — kills the word behind point, using white space and the slash character as the word boundaries, saving the killed text on the kill-ring. 🍓
delete-horizontal-space (M-\) — deletes all spaces and tabs around point. no cap. 🧚
kill-region — kills the text in the current region. 💖
copy-region-as-kill — copies the text in the region to the kill buffer, so it can be yanked immediately. iconic. 🎀
copy-backward-word — copies the word before point to the kill buffer. The word boundaries are the same as backward-word. 🌈
copy-forward-word — copies the word following point to the kill buffer. The word boundaries are the same as forward-word. ⭐
yank (C-y) — yanks the top of the kill ring into the buffer at point. slay. 💅
yank-pop (M-y) — rotates the kill ring, and yanks the new top. Only works following yank or yank-pop. lowkey magic. ✨

🔢 Numeric Arguments 🔢

digit-argument (M-0, M-1, ..., M--) — adds this digit to the argument already accumulating, or starts a new argument. M-- starts a negative argument. 💀
universal-argument — this is another way to specify an argument. If this command is followed by one or more digits, optionally with a leading minus sign, those digits define the argument. If the command is followed by digits, executing universal-argument again ends the numeric argument, but is otherwise ignored. As a special case, if this command is immediately followed by a character that is neither a digit nor minus sign, the argument count for the next command is multiplied by four. The argument count is initially one, so executing this function the first time makes the argument count four, a second time makes the argument count sixteen, and so on. it's giving exponential growth, bestie. 🔥

✨ Completing ✨

complete (TAB) — attempts to perform completion on the text before point. Bash attempts completion by first checking for any programmable completions for the command word (see Programmable Completion below), otherwise treating the text as a variable (if the text begins with $), username (if the text begins with ~), hostname (if the text begins with @), or command (including aliases, functions, and builtins) in turn. If none of these produces a match, it falls back to filename completion. the multitasker we stan. 👑
possible-completions (M-?) — lists the possible completions of the text before point. When displaying completions, readline sets the number of columns used for display to the value of completion-display-width, the value of the shell variable COLUMNS, or the screen width, in that order. 🌸
insert-completions (M-*) — inserts all completions of the text before point that would have been generated by possible-completions, separated by a space. ✨
menu-complete — similar to complete, but replaces the word to be completed with a single match from the list of possible completions. Repeatedly executing menu-complete steps through the list of possible completions, inserting each match in turn. At the end of the list of completions, menu-complete rings the bell (subject to the setting of bell-style) and restores the original text. An argument of n moves n positions forward in the list of matches; a negative argument moves backward through the list. This command is intended to be bound to TAB, but is unbound by default. 💅
menu-complete-backward — identical to menu-complete, but moves backward through the list of possible completions, as if menu-complete had been given a negative argument. This command is unbound by default. 🦋
export-completions — performs completion on the word before point as described above and writes the list of possible completions to readline's output stream using the following format, writing information on separate lines:
- the number of matches N;
- the word being completed;
- S:E, where S and E are the start and end offsets of the word in the readline line buffer; then
- each match, one per line

If there are no matches, the first line will be “0”, and this command does not print any output after the S:E. If there is only a single match, this prints a single line containing it. If there is more than one match, this prints the common prefix of the matches, which may be empty, on the first line after the S:E, then the matches on subsequent lines. In this case, N will include the first line with the common prefix.

The user or application should be able to accommodate the possibility of a blank line. The intent is that the user or application reads N lines after the line containing S:E to obtain the match list. This command is unbound by default. 💜

delete-char-or-list — deletes the character under the cursor if not at the beginning or end of the line (like delete-char). At the end of the line, it behaves identically to possible-completions. This command is unbound by default. 💀
complete-filename (M-/) — attempts filename completion on the text before point. 🍓
possible-filename-completions (C-x /) — lists the possible completions of the text before point, treating it as a filename. 🌈
complete-username (M-~) — attempts completion on the text before point, treating it as a username. ✨
possible-username-completions (C-x ~) — lists the possible completions of the text before point, treating it as a username. 🎀
complete-variable (M-$) — attempts completion on the text before point, treating it as a shell variable. 💖
possible-variable-completions (C-x $) — lists the possible completions of the text before point, treating it as a shell variable. 🌸
complete-hostname (M-@) — attempts completion on the text before point, treating it as a hostname. 🦋
possible-hostname-completions (C-x @) — lists the possible completions of the text before point, treating it as a hostname. ⭐
complete-command (M-!) — attempts completion on the text before point, treating it as a command name. Command completion attempts to match the text against aliases, reserved words, shell functions, shell builtins, and finally executable filenames, in that order. 👑
possible-command-completions (C-x !) — lists the possible completions of the text before point, treating it as a command name. 🔥
dynamic-complete-history (M-TAB) — attempts completion on the text before point, comparing the text against history list entries for possible completion matches. 💅
dabbrev-expand — attempts menu completion on the text before point, comparing the text against lines from the history list for possible completion matches. ✨
complete-into-braces (M-{) — performs filename completion and inserts the list of possible completions enclosed within braces so the list is available to the shell (see Brace Expansion above). 🎀

⌨️ Keyboard Macros ⌨️

start-kbd-macro (C-x () — begins saving the characters typed into the current keyboard macro. 💖
end-kbd-macro (C-x )) — stops saving the characters typed into the current keyboard macro and stores the definition. 🌸
call-last-kbd-macro (C-x e) — re-executes the last keyboard macro defined, by making the characters in the macro appear as if typed at the keyboard. 🦋
print-last-kbd-macro () — prints the last keyboard macro defined in a format suitable for the inputrc file. ⭐

Miscellaneous 💫

okay bestie, this is the junk-drawer corner of readline and honestly? it ate. these are the commands that didn't fit anywhere else but still serve 💅

re-read-init-file (C-x C-r) — reads the contents of the inputrc file back in, and folds in any bindings or variable assignments it finds in there. like a lil refresh button for your vibes 🔄
abort (C-g) — ngl this one just bails on whatever editing command is happening right now and rings the terminal's bell (subject to whatever bell-style is set to). dramatic exit, love that for it 💀
do-lowercase-version (M-A, M-B, M-x, ...) — if the metafied character x is uppercase, it runs whatever command is bound to the matching metafied lowercase character. > heads up bestie: the behavior is undefined if x is already lowercase. so like. don't. 🙈
prefix-meta (ESC) — metafies the next character you type. so ESC f is equivalent to Meta-f. yes it's giving keyboard sorcery, that's the point ✨
undo (C-_, C-x C-u) — incremental undo, and it's remembered separately for each line. your mistakes are per-line, no crossover 🫶
revert-line (M-r) — undoes all the changes you made to this line. it's basically running undo enough times that the line goes all the way back to its initial state. the ultimate "we don't talk about that" 🌸
tilde-expand (M-&) — performs tilde expansion on the current word. ~ girlies rise 🎀
set-mark (C-@, M-<space>) — sets the mark to the point. if you give it a numeric argument, it sets the mark to that position instead 📍
exchange-point-and-mark (C-x C-x) — swaps the point with the mark. current cursor position goes to the saved position, then the mark becomes the old cursor position. a whole swap, iconic 💖
character-search (C-]) — reads a character, then moves point to the next occurrence of that character. give it a negative argument and it searches backwards for previous occurrences 🔍
character-search-backward (M-C-]) — reads a character and moves point to the previous occurrence of it. negative argument? it searches for subsequent occurrences instead. mirror mirror 🪞
skip-csi-sequence — reads enough characters to consume a multi-key sequence, like the ones defined for keys such as Home and End. CSI sequences begin with a Control Sequence Indicator (CSI), usually ESC [. if this sequence is bound to “\e[”, keys that produce CSI sequences have no effect unless they're explicitly bound to a readline command — instead of dumping stray characters into your editing buffer. it's unbound by default, but it's usually bound to ESC [. lowkey the unsung hero of not having garbage in your line 🧹
insert-comment (M-#) — with no numeric argument, it inserts the value of the readline comment-begin variable at the beginning of the current line. with a numeric argument it toggles: if the characters at the start of the line don't match comment-begin, insert the value; otherwise delete those comment-begin characters from the front of the line. either way the line is accepted as if you'd typed a newline. the default value of comment-begin makes this command turn your current line into a shell comment. and if a numeric argument causes the comment character to be removed, the line will be executed by the shell. so like, be aware bestie ⚠️
spell-correct-word (C-x s) — does spelling correction on the current word, treating it as a directory or filename, exactly the same way the cdspell shell option does. word boundaries are the same ones used by shell-forward-word 🔤
glob-complete-word (M-g) — treats the word before point as a pattern for pathname expansion, with an asterisk implicitly appended, then uses that pattern to generate a list of matching file names for possible completions 🌈
glob-expand-word (C-x *) — treats the word before point as a pattern for pathname expansion and inserts the list of matching file names, replacing the word. numeric argument? it appends a * before pathname expansion 💥
glob-list-expansions (C-x g) — displays the list of expansions that glob-expand-word would have generated, and redisplays the line. numeric argument appends a * before pathname expansion 👀
dump-functions — prints all of the functions and their key bindings to the readline output stream. numeric argument makes the output formatted so it can be dropped straight into an inputrc file 📋
dump-variables — prints all of the settable readline variables and their values to the readline output stream. numeric argument → formatted so it can be part of an inputrc file 📝
dump-macros — prints all of the readline key sequences bound to macros, plus the strings they output, to the readline output stream. numeric argument → formatted so it can be part of an inputrc file 🧾
execute-named-command (M-x) — reads a bindable readline command name from the input and executes the function it's bound to, as if the key sequence it was bound to had appeared in the input. if you give it a numeric argument, it passes that argument along to the function it executes 🎯
display-shell-version (C-x C-v) — displays version information about the current instance of bash. say hi to your bash bestie 💜

Programmable Completion 🧚

so here's the plot: when a user tries word completion for a command, or for an argument to a command, and a completion specification (a compspec) has been defined for it using the complete builtin (peep SHELL BUILTIN COMMANDS below), readline invokes the programmable completion facilities. it's giving custom completion couture ✨

First, bash identifies the command name. if a compspec has been defined for that command, that compspec is used to generate the list of possible completions for the word. if the command word is the empty string (completion attempted at the beginning of an empty line), bash uses any compspec defined with the -E option to complete. the -I option to complete indicates that the command word is the first non-assignment word on the line, or after a command delimiter such as ; or |. this usually indicates command name completion 💅

If the command word is a full pathname, bash searches for a compspec for the full pathname first. if there's no compspec for the full pathname, bash tries to find a compspec for the portion following the final slash. if those searches don't turn up a compspec, or if there's no compspec for the command word, bash uses any compspec defined with the -D option to complete as the default. and if there's no default compspec, bash performs alias expansion on the command word as a final resort, and tries to find a compspec for the command word that results from any successful expansion 🦋

If a compspec is not found, bash performs its default completion as described above under Completing. otherwise, once a compspec has been found, bash uses it to generate the list of matching words 🔥

First, bash performs the actions specified by the compspec. this only returns matches which are prefixes of the word being completed. when the -f or -d option is used for filename or directory name completion, bash uses the shell variable FIGNORE to filter the matches 🌸

Next, programmable completion generates matches specified by a pathname expansion pattern supplied as an argument to the -G option. the words generated by the pattern need not match the word being completed. bash uses the FIGNORE variable to filter the matches, but does not use the GLOBIGNORE shell variable 👑 (know the difference, bestie)

Next, completion considers the string specified as the argument to the -W option. the string is first split using the characters in the IFS special variable as delimiters. this honors shell quoting within the string, in order to provide a mechanism for the words to contain shell metacharacters or characters in the value of IFS. each word is then expanded using brace expansion, tilde expansion, parameter and variable expansion, command substitution, and arithmetic expansion, as described above under EXPANSION. the results are split using the rules described above under Word Splitting. the results of the expansion are prefix-matched against the word being completed, and the matching words become possible completions 💖

After these matches have been generated, bash executes any shell function or command specified with the -F and -C options. when the command or function is invoked, bash assigns values to the COMP_LINE, COMP_POINT, COMP_KEY, and COMP_TYPE variables as described above under Shell Variables. if a shell function is being invoked, bash also sets the COMP_WORDS and COMP_CWORD variables. when the function or command is invoked, the first argument ($1) is the name of the command whose arguments are being completed, the second argument ($2) is the word being completed, and the third argument ($3) is the word preceding the word being completed on the current command line. there is no filtering of the generated completions against the word being completed; the function or command has complete freedom in generating the matches and they don't need to match a prefix of the word 🎀

Any function specified with -F is invoked first. the function may use any of the shell facilities, including the compgen and compopt builtins described below, to generate the matches. it must put the possible completions in the COMPREPLY array variable, one per array element 📦

Next, any command specified with the -C option is invoked in an environment equivalent to command substitution. it should print a list of completions, one per line, to the standard output. backslash will escape a newline, if necessary. these are added to the set of possible completions 📃

After generating all of the possible completions, bash applies any filter specified with the -X option to the completions in the list. the filter is a pattern as used for pathname expansion; a & in the pattern is replaced with the text of the word being completed. a literal & may be escaped with a backslash; the backslash is removed before attempting a match. any completion that matches the pattern is removed from the list. a leading ! negates the pattern; in that case bash removes any completion that does not match the pattern. if the nocasematch shell option is enabled, bash performs the match without regard to the case of alphabetic characters 🔍

Finally, programmable completion adds any prefix and suffix specified with the -P and -S options, respectively, to each completion, and returns the result to readline as the list of possible completions 🎁

If the previously-applied actions do not generate any matches, and the -o dirnames option was supplied to complete when the compspec was defined, bash attempts directory name completion 📁

If the -o plusdirs option was supplied to complete when the compspec was defined, bash attempts directory name completion and adds any matches to the set of possible completions 📂

Okay so bestie, default behaviour time: if a compspec is found, literally whatever it generates gets handed back to the completion code as the full set of possible completions. And the default bash completions plus readline's default filename completion? disabled. Gone. Not invited. 💅

But wait — if you passed the -o bashdefault option to complete back when the compspec got defined, and the compspec generates no matches, then bash goes "fine, I'll do it myself" and attempts its default completions. And if the compspec and those default bash completions (if it even tried them) both come up empty, and you'd supplied -o default to complete when defining the compspec, then programmable completion just performs readline's default completion. The chain of fallbacks is real, no cap. 🌈

The options you hand to complete and compopt can control how readline treats the completions. Like for instance, -o fullquote tells readline to quote the matches as if they were filenames. See the description of complete below for the deets — it's all down there, iykyk.

When a compspec says it wants directory name completion, the programmable completion functions force readline to append a slash to completed names that are symbolic links to directories, and that's subject to the value of the mark-directories readline variable — regardless of how mark-symlinked-directories is set. Yes, it just overrides it. Main character behaviour. 👑

✨ dynamically modifying completions ✨

There's some support for dynamically modifying completions, and honestly? It's giving genius. It's most useful when combined with a default completion specified with complete -D. Here's the tea: shell functions executed as completion functions can signal that completion should be retried by returning an exit status of 124. If a shell function returns 124 and changes the compspec associated with the command on which completion is being attempted (which gets supplied as the first argument when the function is executed), programmable completion restarts from the beginning and tries to find a brand new compspec for that command. So you can build a set of completions dynamically as completion is attempted instead of loading them all at once. Lazy loading but make it slay. 🦋

For instance, say you've got a library of compspecs, each one kept in a file matching the name of the command — this default completion function would load completions dynamically:

_completion_loader()
{
  . "/etc/bash_completion.d/$1.sh" \
    >/dev/null 2>&1 && return 124
}
complete -D -F _completion_loader \
    -o bashdefault -o default

📜 HISTORY ✨

When the -o history option to the set builtin is enabled, the shell gives you access to the command history — that's the list of commands you previously typed. The value of the HISTSIZE variable is used as the number of commands to save in the history list: the shell saves the text of the last HISTSIZE commands, and the default is 500. 🎀

The shell stores each command in the history list prior to parameter and variable expansion (see EXPANSION above) but after history expansion is performed, and all of that is subject to the values of the shell variables HISTIGNORE and HISTCONTROL.

On startup, bash initializes the history list by reading history entries from the file named by the HISTFILE variable (default ~/.bash_history). That file is referred to as the history file. The history file gets truncated, if necessary, to contain no more than the number of history entries specified by the value of the HISTFILESIZE variable. But if HISTFILESIZE is unset, or set to null, a non-numeric value, or a numeric value less than zero, then the history file is not truncated. Not even a little. 🙈

When the history file is read, lines beginning with the history comment character followed immediately by a digit are interpreted as timestamps for the following history line. Those timestamps are optionally displayed depending on the value of the HISTTIMEFORMAT variable. And when they're present, history timestamps delimit history entries, which makes multi-line entries possible. Fancy. 💖

When a shell with history enabled exits, bash copies the last $HISTSIZE entries from the history list to $HISTFILE. If the histappend shell option is enabled (see the description of shopt under SHELL BUILTIN COMMANDS below), bash appends the entries to the history file — otherwise it overwrites the history file. And if HISTFILE is unset or null, or if the history file is unwritable, then the history is not saved. Rip. After saving the history, bash truncates the history file to contain no more than HISTFILESIZE lines, as described above.

If the HISTTIMEFORMAT variable is set, the shell writes the timestamp information associated with each history entry into the history file, marked with the history comment character, so timestamps are preserved across shell sessions. It uses the history comment character to tell timestamps apart from the other history lines. And as above, when you're using HISTTIMEFORMAT, the timestamps delimit multi-line history entries. 🕰️

The fc builtin command (see SHELL BUILTIN COMMANDS below) will list or edit and re-execute a portion of the history list. The history builtin can display or modify the history list and manipulate the history file. And when you're using command-line editing, search commands are available in each editing mode that provide access to the history list. Obsessed, honestly. 🌸

The shell lets you control which commands actually get saved on the history list — the HISTCONTROL and HISTIGNORE variables are used to save only a subset of the commands you enter. If the cmdhist shell option is enabled, the shell attempts to save each line of a multi-line command in the same history entry, adding semicolons where necessary to preserve syntactic correctness. The lithist shell option modifies cmdhist by saving the command with embedded newlines instead of semicolons. And for info on setting and unsetting shell options, see the description of the shopt builtin below under SHELL BUILTIN COMMANDS. 💜

✨ HISTORY EXPANSION ✨

okay so bestie, bash has a whole history expansion feature and it is giving csh energy, because it's basically the same idea. this section is the tour of what syntax you get. 💖

History expansion is on by default for interactive shells, and you can turn it off with the +H option to the set builtin command (peep SHELL BUILTIN COMMANDS below, it's right there). Non-interactive shells do not do history expansion by default, but you can absolutely turn it on with set -H. ✨

What it does: history expansions pull words from the history list into your input stream. so it's easy to repeat commands, insert the arguments to a previous command into the current input line, or fix errors in previous commands fast. lowkey iconic. 🎀

Timing matters here: history expansion happens immediately after a complete line is read, before the shell breaks it into words, and it happens on each line individually. The shell also attempts to inform the history expansion functions about quoting still in effect from previous lines.

It goes down in two parts:

Figure out which history list entry to use during substitution.
Select portions of that entry to include into the current one.

The entry selected from the history is the event, and the portions of that entry that get acted upon are words. Various modifiers are available to manipulate the selected words. The entry gets split into words in the same fashion as when reading input, so several metacharacter-separated words surrounded by quotes count as one word. The event designator selects the event, the optional word designator selects words from the event, and the various optional modifiers manipulate the selected words. 🦋

History expansions are introduced by the appearance of the history expansion character, which is ! by default. They may appear anywhere in the input, but they do not nest. 🚫

Only backslash (\) and single quotes can quote the history expansion character — but heads up, the history expansion character is also treated as quoted if it immediately precedes the closing double quote in a double-quoted string.

Several characters inhibit history expansion if found immediately following the history expansion character, even if it's unquoted: space, tab, newline, carriage return, =, and the other shell metacharacters defined above. 💅

There's also a special abbreviation for substitution, active when the quick substitution character (described above under histchars) is the first character on the line. It selects the previous history list entry, using an event designator equivalent to !!, and substitutes one string for another in that entry. It's described below under Event Designators. This is the only history expansion that does not begin with the history expansion character. no cap.

Several shell options settable with the shopt builtin will modify history expansion behavior (see the description of the shopt builtin below). If the histverify shell option is enabled, and readline is being used, history substitutions are not immediately passed to the shell parser. Instead, the expanded line is reloaded into the readline editing buffer for further modification. And if readline is being used, and the histreedit shell option is enabled, a failed history substitution gets reloaded into the readline editing buffer for correction. so forgiving, we love that. 🫶

The -p option to the history builtin command shows what a history expansion will do before you use it. The -s option to the history builtin will add commands to the end of the history list without actually executing them, so that they're available for subsequent recall.

The shell allows control of the various characters used by the history expansion mechanism (see the description of histchars above under Shell Variables). And the shell uses the history comment character to mark history timestamps when writing the history file. ⭐

🎯 Event Designators

An event designator is a reference to an entry in the history list. The event designator consists of the portion of the word beginning with the history expansion character and ending with the word designator if present, or the end of the word. Unless the reference is absolute, events are relative to the current position in the history list. 💜

! — Start a history substitution, except when followed by a blank, newline, carriage return, =, or, when the extglob shell option is enabled using the shopt builtin, (.
!n — Refer to history list entry n.
!-n — Refer to the current entry minus n.
!! — Refer to the previous entry. This is a synonym for !-1.
!string — Refer to the most recent command preceding the current position in the history list starting with string.
!?string[?] — Refer to the most recent command preceding the current position in the history list containing string. The trailing ? may be omitted if string is followed immediately by a newline. If string is missing, this uses the string from the most recent search; it is an error if there is no previous search string. 😭
^string1^string2^ — Quick substitution. Repeat the previous command, replacing string1 with string2. Equivalent to !!:s^string1^string2^ (see Modifiers below).
!# — The entire command line typed so far. main character energy fr.

🔤 Word Designators

Word designators are used to select desired words from the event. They're optional — if the word designator isn't supplied, the history expansion uses the entire event. A : separates the event specification from the word designator. It may be omitted if the word designator begins with a ^, $, *, -, or %. Words are numbered from the beginning of the line, with the first word being denoted by 0 (zero). Words are inserted into the current line separated by single spaces. ✨

0 (zero) — The zeroth word. For the shell, this is the command word.
n — The *n*th word.
^ — The first argument: word 1.
$ — The last word. This is usually the last argument, but will expand to the zeroth word if there is only one word in the line.
% — The first word matched by the most recent ?string? search, if the search string begins with a character that is part of a word. By default, searches begin at the end of each line and proceed to the beginning, so the first word matched is the one closest to the end of the line.
x-y — A range of words; -y abbreviates 0-y.
* — All of the words but the zeroth. This is a synonym for 1-$. It is not an error to use * if there is just one word in the event; it expands to the empty string in that case.
x* — Abbreviates x-$.
x- — Abbreviates x-$ like x*, but omits the last word. If x is missing, it defaults to 0.

If a word designator is supplied without an event specification, the previous command is used as the event, equivalent to !!. 💅

💫 Modifiers

After the optional word designator, the expansion may include a sequence of one or more of the following modifiers, each preceded by a :. These modify, or edit, the word or words selected from the history event. 🎀

h — Remove a trailing pathname component, leaving only the head.
t — Remove all leading pathname components, leaving the tail.
r — Remove a trailing suffix of the form .xxx, leaving the basename.
e — Remove all but the trailing suffix.
p — Print the new command but do not execute it.
q — Quote the substituted words, escaping further substitutions.
x — Quote the substituted words as with q, but break into words at blanks and newlines. The q and x modifiers are mutually exclusive; expansion uses the last one supplied.
s/old/new/ — Substitute new for the first occurrence of old in the event line. Any character may be used as the delimiter in place of /. The final delimiter is optional if it is the last character of the event line. A single backslash quotes the delimiter in old and new. If & appears in new, it is replaced with old. A single backslash quotes the &. If old is null, it is set to the last old substituted, or, if no previous history substitutions took place, the last string in a !?string[?] search. If new is null, each matching old is deleted.
& — Repeat the previous substitution.
g — Cause changes to be applied over the entire event line. This is used in conjunction with :s (e.g., :gs/old/new/) or :&. If used with :s, any delimiter can be used in place of /, and the final delimiter is optional if it is the last character of the event line. An a may be used as a synonym for g.
G — Apply the following s or & modifier once to each word in the event line.

👑 SHELL BUILTIN COMMANDS

Unless otherwise noted, each builtin command documented in this section as accepting options preceded by - accepts -- to signify the end of the options. 🫶

But listen — there are exceptions, and they matter:

The :, true, false, and test/[ builtins do not accept options and do not treat -- specially.
The exit, logout, return, break, continue, let, and shift builtins accept and process arguments beginning with - without requiring --.
Other builtins that accept arguments but are not specified as accepting options interpret arguments beginning with - as invalid options and require -- to prevent this interpretation. 💀

`:` — the do-nothing icon

: [arguments]

No effect; the command does nothing beyond expanding arguments and performing any specified redirections. The return status is zero. she's just standing there. iconic. 🎀

`.` / `source` — the "run this in MY shell" bestie

. [-p path] filename [arguments]
source [-p path] filename [arguments]

The . command (source) reads and execute commands from filename in the current shell environment and returns the exit status of the last command executed from filename. 🌸

If filename does not contain a slash, . searches for it. If the -p option is supplied, . treats path as a colon-separated list of directories in which to find filename; otherwise, . uses the entries in PATH to find the directory containing filename. filename does not need to be executable. When bash is not in posix mode, it searches the current directory if filename is not found in PATH, but does not search the current directory if -p is supplied. And if the sourcepath option to the shopt builtin command is turned off, . does not search PATH. 👑

If any arguments are supplied, they become the positional parameters when filename is executed. Otherwise the positional parameters are unchanged. 💜

…and okay bestie, ONE more thing about . before we move on, because the DEBUG trap lore is real. If the -T option is enabled, . inherits any trap on DEBUG. If it's not enabled, any DEBUG trap string gets saved and restored around the call to ., and . unsets the DEBUG trap while it executes. And if -T is not set, and the sourced file changes the DEBUG trap, that new value persists after . completes. The return status is the status of the last command executed from filename (0 if no commands are executed), and non-zero if filename is not found or cannot be read. 💅

`alias` 🌈

alias [-p] [name[=value] ...]

With no arguments, or with the -p option, alias prints the list of aliases in the form alias name=value on standard output. When arguments are supplied, it defines an alias for each name whose value is given. A trailing space in value causes the next word to be checked for alias substitution when the alias is expanded during command parsing. For each name in the argument list for which no value is supplied, it prints the name and value of the alias name. alias returns true unless a name is given (without a corresponding =value) for which no alias has been defined.

`bg` 🦋

bg [jobspec ...]

Resume each suspended job jobspec in the background, as if it had been started with &. If jobspec is not present, the shell uses its notion of the current job. bg jobspec returns 0 unless run when job control is disabled or, when run with job control enabled, any specified jobspec was not found or was started without job control.

`bind` ✨

bind [-m keymap] [-lsvSVX]
bind [-m keymap] [-q function] [-u function] [-r keyseq]
bind [-m keymap] -f filename
bind [-m keymap] -x keyseq[:] shell-command
bind [-m keymap] keyseq:function-name
bind [-m keymap] -p|-P [readline-command]
bind [-m keymap] keyseq:readline-command
bind readline-command-line

Display current readline key and function bindings, bind a key sequence to a readline function or macro or to a shell command, or set a readline variable. Each non-option argument is a key binding or command as it would appear in a readline initialization file such as .inputrc, but each binding or command must be passed as a separate argument; e.g., '"\C-x\C-r": re-read-init-file'. In the following descriptions, output available to be re-read is formatted as commands that would appear in a readline initialization file or that would be supplied as individual arguments to a bind command.

Options, if supplied, have the following meanings:

-m keymap — use keymap as the keymap to be affected by the subsequent bindings. The acceptable keymap names are emacs, emacs-standard, emacs-meta, emacs-ctlx, vi, vi-move, vi-command, and vi-insert. vi is equivalent to vi-command (vi-move is also a synonym); emacs is equivalent to emacs-standard.
-l — list the names of all readline functions.
-p — display readline function names and bindings in such a way that they can be used as an argument to a subsequent bind command or in a readline initialization file. If arguments remain after option processing, bind treats them as readline command names and restricts output to those names.
-P — list current readline function names and bindings. If arguments remain after option processing, bind treats them as readline command names and restricts output to those names.
-s — display readline key sequences bound to macros and the strings they output in such a way that they can be used as an argument to a subsequent bind command or in a readline initialization file.
-S — display readline key sequences bound to macros and the strings they output.
-v — display readline variable names and values in such a way that they can be used as an argument to a subsequent bind command or in a readline initialization file.
-V — list current readline variable names and values.
-f filename — read key bindings from filename.
-q function — display key sequences that invoke the named readline function.
-u function — unbind all key sequences bound to the named readline function.
-r keyseq — remove any current binding for keyseq.
-x keyseq[: ]shell-command — cause shell-command to be executed whenever keyseq is entered. The separator between keyseq and shell-command is either whitespace or a colon optionally followed by whitespace. If the separator is whitespace, shell-command must be enclosed in double quotes and readline expands any of its special backslash-escapes in shell-command before saving it. If the separator is a colon, any enclosing double quotes are optional, and readline does not expand the command string before saving it. Since the entire key binding expression must be a single argument, it should be enclosed in single quotes. When shell-command is executed, the shell sets the READLINE_LINE variable to the contents of the readline line buffer and the READLINE_POINT and READLINE_MARK variables to the current location of the insertion point and the saved insertion point (the mark), respectively. The shell assigns any numeric argument the user supplied to the READLINE_ARGUMENT variable. If there was no argument, that variable is not set. If the executed command changes the value of any of READLINE_LINE, READLINE_POINT, or READLINE_MARK, those new values will be reflected in the editing state.
-X — list all key sequences bound to shell commands and the associated commands in a format that can be reused as an argument to a subsequent bind command.

The return value is 0 unless an unrecognized option is supplied or an error occurred. 🎀

`break` 💀

break [n]

Exit from within a for, while, until, or select loop. If n is specified, break exits n enclosing loops. n must be ≥ 1. If n is greater than the number of enclosing loops, all enclosing loops are exited. The return value is 0 unless n is not greater than or equal to 1.

`builtin` 👑

builtin shell-builtin [arguments]

Execute the specified shell builtin shell-builtin, passing it arguments, and return its exit status. This is useful when defining a function whose name is the same as a shell builtin, retaining the functionality of the builtin within the function. The cd builtin is commonly redefined this way. The return status is false if shell-builtin is not a shell builtin command.

`caller` 🧚

caller [expr]

Returns the context of any active subroutine call (a shell function or a script executed with the . or source builtins).

Without expr, caller displays the line number and source filename of the current subroutine call. If a non-negative integer is supplied as expr, caller displays the line number, subroutine name, and source file corresponding to that position in the current execution call stack. This extra information may be used, for example, to print a stack trace. The current frame is frame 0.

The return value is 0 unless the shell is not executing a subroutine call or expr does not correspond to a valid position in the call stack.

`cd` 🌸

cd [-L] [-@] [dir]
cd -P [-e] [-@] [dir]

Change the current directory to dir. If dir is not supplied, the value of the HOME shell variable is used as dir. If the variable CDPATH exists, and dir does not begin with a slash (/), cd uses it as a search path: the shell searches each directory name in CDPATH for dir. Alternative directory names in CDPATH are separated by a colon (:). A null directory name in CDPATH is the same as the current directory, i.e., ..

The -P option causes cd to use the physical directory structure by resolving symbolic links while traversing dir and before processing instances of .. in dir (see also the -P option to the set builtin command).

The -L option forces cd to follow symbolic links by resolving the link after processing instances of .. in dir. If .. appears in dir, cd processes it by removing the immediately previous pathname component from dir, back to a slash or the beginning of dir, and verifying that the portion of dir it has processed to that point is still a valid directory name after removing the pathname component. If it is not a valid directory name, cd returns a non-zero status. If neither -L nor -P is supplied, cd behaves as if -L had been supplied.

If the -e option is supplied with -P, and cd cannot successfully determine the current working directory after a successful directory change, it returns a non-zero status.

On systems that support it, the -@ option presents the extended attributes associated with a file as a directory.

An argument of - is converted to $OLDPWD before attempting the directory change.

If cd uses a non-empty directory name from CDPATH, or if - is the first argument, and the directory change is successful, cd writes the absolute pathname of the new working directory to the standard output.

If the directory change is successful, cd sets the value of the PWD environment variable to the new directory name, and sets the OLDPWD environment variable to the value of the current working directory before the change.

The return value is true if the directory was successfully changed; false otherwise. 🫶

🎀 `command` 🎀

command [-pVv] command [arg ...]

okay bestie, this one runs command with its args, but the whole ✨point✨ is that it suppresses the normal shell function lookup for command. So if you've got a function with the same name lurking around trying to steal the spotlight, it does not get to perform here. Only builtin commands or commands found in the PATH named command get executed. periodt.

-p — the search for command gets done using a default value for PATH that's guaranteed to find all of the standard utilities. like, the safe little starter pack 🫶
-V or -v — command prints a description of command. The -v version shows you a single word indicating the command or filename used to invoke command; the -V version gives you the more verbose, yappier description 💅

Now the exit-status tea, and this is important so listen up 🍓:

if you passed -V or -v, the exit status is zero if command was found, and non-zero if it wasn't.
if you passed neither option and an error occurred or command cannot be found, the exit status is 127.
otherwise, the exit status of the command builtin is the exit status of command. (yes queen, it passes the vibe straight through 💫)

🌸 `compgen` 🌸

compgen [-V varname] [option] [word]

Generate possible completion matches for word according to the options — and those options may be any option accepted by the complete builtin, with the exceptions of -p, -r, -D, -E, and -I. Then it writes the matches to standard output. iconic.

-V varname — instead of dumping the completions to standard output, compgen stores the generated completions into the indexed array variable varname. little storage queen 🎀

when you're using the -F or -C options, heads up: the various shell variables set by the programmable completion facilities, while available, will not have useful values. no cap. 🙈

the matches get generated in the exact same way as if the programmable completion code had generated them directly from a completion specification with the same flags. and if word is specified, only those completions matching word will be displayed or stored. 🦋

The return value is true unless an invalid option is supplied, or no matches were generated.

👑 `complete` 👑

complete [-abcdefgjksuv] [-o comp-option] [-DEI] [-A action]
         [-G globpat] [-W wordlist] [-F function] [-C command]
         [-X filterpat] [-P prefix] [-S suffix] name [name ...]
complete -pr [-DEI] [name ...]

This is the one that specifies how arguments to each name should be completed. main character energy fr fr 💅

-p — or if you supply no options or names at all — print existing completion specifications in a way that allows them to be reused as input. so like, screenshot-ready ✨
-r — removes a completion specification for each name, or, if no names are supplied, all completion specifications. bye bye 🫶

Now the special little trio:

-D — other supplied options and actions apply to the “default” command completion; that is, completion attempted on a command for which no completion has previously been defined.
-E — other supplied options and actions apply to “empty” command completion; that is, completion attempted on a blank line.
-I — other supplied options and actions apply to completion on the initial non-assignment word on the line, or after a command delimiter such as ; or |, which is usually command name completion.

if multiple of these are supplied, -D takes precedence over -E, and both of those take precedence over -I. and if any of -D, -E, or -I are supplied, any other name arguments are ignored — these completions only apply to the case specified by the option. no crossover, bestie.

the process of applying these completion specifications when attempting word completion is described above under Programmable Completion. go reread that part if you need the lore 📖

Other options, if specified, have the following meanings. The arguments to the -G, -W, and -X options (and, if necessary, the -P and -S options) should be quoted to protect them from expansion before the complete builtin is invoked. ⚠️

-o comp-option — the comp-option controls several aspects of the compspec's behavior beyond the simple generation of completions. comp-option may be one of:
- bashdefault — Perform the rest of the default bash completions if the compspec generates no matches.
- default — Use readline's default filename completion if the compspec generates no matches.
- dirnames — Perform directory name completion if the compspec generates no matches.
- filenames — Tell readline that the compspec generates filenames, so it can perform any filename-specific processing (such as adding a slash to directory names, quoting special characters, or suppressing trailing spaces). This is intended to be used with shell functions.
- fullquote — Tell readline to quote all the completed words even if they are not filenames.
- noquote — Tell readline not to quote the completed words if they are filenames (quoting filenames is the default).
- nosort — Tell readline not to sort the list of possible completions alphabetically.
- nospace — Tell readline not to append a space (the default) to words completed at the end of the line.
- plusdirs — After generating any matches defined by the compspec, attempt directory name completion and add any matches to the results of the other actions.
-A action — the action may be one of the following to generate a list of possible completions:
- alias — Alias names. May also be specified as -a.
- arrayvar — Array variable names.
- binding — Readline key binding names.
- builtin — Names of shell builtin commands. May also be specified as -b.
- command — Command names. May also be specified as -c.
- directory — Directory names. May also be specified as -d.
- disabled — Names of disabled shell builtins.
- enabled — Names of enabled shell builtins.
- export — Names of exported shell variables. May also be specified as -e.
- file — File and directory names, similar to readline's filename completion. May also be specified as -f.
- function — Names of shell functions.
- group — Group names. May also be specified as -g.
- helptopic — Help topics as accepted by the help builtin.
- hostname — Hostnames, as taken from the file specified by the HOSTFILE shell variable.
- job — Job names, if job control is active. May also be specified as -j.
- keyword — Shell reserved words. May also be specified as -k.
- running — Names of running jobs, if job control is active.
- service — Service names. May also be specified as -s.
- setopt — Valid arguments for the -o option to the set builtin.
- shopt — Shell option names as accepted by the shopt builtin.
- signal — Signal names.
- stopped — Names of stopped jobs, if job control is active.
- user — User names. May also be specified as -u.
- variable — Names of all shell variables. May also be specified as -v.
-C command — command is executed in a subshell environment, and its output is used as the possible completions. Arguments are passed as with the -F option.
-F function — the shell function function is executed in the current shell environment. When the function is executed, the first argument ($1) is the name of the command whose arguments are being completed, the second argument ($2) is the word being completed, and the third argument ($3) is the word preceding the word being completed on the current command line. When function finishes, programmable completion retrieves the possible completions from the value of the COMPREPLY array variable. 👑
-G globpat — expand the pathname expansion pattern globpat to generate the possible completions.
-P prefix — add prefix to the beginning of each possible completion after all other options have been applied.
-S suffix — append suffix to each possible completion after all other options have been applied.
-W wordlist — split the wordlist using the characters in the IFS special variable as delimiters, and expand each resulting word. Shell quoting is honored within wordlist, in order to provide a mechanism for the words to contain shell metacharacters or characters in the value of IFS. The possible completions are the members of the resultant list which match a prefix of the word being completed. 🧚
-X filterpat — filterpat is a pattern as used for pathname expansion. It is applied to the list of possible completions generated by the preceding options and arguments, and each completion matching filterpat is removed from the list. A leading ! in filterpat negates the pattern; in this case, any completion not matching filterpat is removed. 🚫

...and the tea on the return value: it's true unless you hand it an invalid option, or you pass an option that isn't -p, -r, -D, -E, or -I without a name argument, or you try to remove a completion spec for a name that literally has no spec (embarrassing, bestie), or something explodes while adding a completion specification. 💀

✨ `compopt` ✨

compopt [-o option] [-DEI] [+o option] [name]

Okay so compopt is the glow-up artist for your completion options 💅 — it modifies the completion options for each name according to whatever options you gave it, or for the completion that's currently executing if you didn't hand it any names. And if you supply no options at all? It just displays the completion options for each name, or for the current completion. The possible values of option are exactly the same ones that are valid for the complete builtin up above 👆 (iykyk).

Now the three special flags, and listen carefully because the vibes matter:

-D — says "hey, apply these other options to the default command completion."
-E — says "apply these other options to the empty command completion."
-I — says "apply these other options to completion on the initial word on the line."

These get figured out the exact same way as the complete builtin does it. 🎯

And if you stack multiple of them? Precedence time, no drama: -D wins over -E, and both of them win over -I. Periodt.

Return value tea ☕: it's true unless an invalid option is supplied, or you try to modify the options for a name that has no completion specification (nothing to modify, girl), or an output error happens. 😭

🌸 `continue` 🌸

continue [n]

continue is the "okay next one" of loops ✨ — it resumes the next iteration of the enclosing for, while, until, or select loop. If you give it n, bash hops up and resumes the nth enclosing loop. And n must be ≥ 1 — that's the rule, no cap. If n is bigger than the number of enclosing loops, the shell just resumes the last enclosing loop (the "top-level" loop, the main character 👑).

Return value is 0 unless n is not greater than or equal to 1. Simple as that. 💖

🎀 `declare` / `typeset` 🎀

declare [-aAfFgiIlnrtux] [-p] [name[=value] ...]
typeset [-aAfFgiIlnrtux] [-p] [name[=value] ...]

These two are twins — declare and typeset do the same thing, so don't stress about which one you grab. They declare variables and/or give them attributes. If you pass no names, they display the values of variables or functions.

-p — displays the attributes and values of each name. And when -p is used with name arguments, any additional options other than -f and -F are just... ignored, bestie, they don't get a say.

When you use -p without any name arguments, declare will show you the attributes and values of all variables that have the attributes specified by your additional options. If you gave -p with no other options, it displays the attributes and values of all shell variables. 🌈 And -f restricts that display to shell functions only.

-F — inhibits the display of function definitions; only the function name and attributes get printed. But plot twist 🫢: if the extdebug shell option is turned on using shopt, then the source file name and line number where each name is defined get displayed too. And -F implies -f.
-g — forces variables to be created or modified at the global scope, even when declare is running inside a shell function. (It gets ignored when declare isn't executed in a shell function, so don't worry about it.)
-I — makes local variables inherit the attributes (except the nameref attribute) and value of any existing variable with the same name in a surrounding scope. If there's no existing variable? The local variable starts out unset. 🦋

Now here's the menu of options that restrict output to variables with a certain attribute, or give variables attributes. Read carefully, this is the important part:

-a — each name is an indexed array variable (peep Arrays above).
-A — each name is an associative array variable (again, Arrays above 👀).
-f — each name refers to a shell function.
-i — the variable is treated as an integer; arithmetic evaluation (see ARITHMETIC EVALUATION above) happens whenever the variable gets assigned a value.
-l — when the variable is assigned a value, all upper-case characters get converted to lower-case. The upper-case attribute is disabled.
-n — gives each name the nameref attribute, making it a name reference to another variable. That other variable is defined by the value of name. All references, assignments, and attribute modifications to name — except the ones using or changing the -n attribute itself — are performed on the variable that name's value points to. 🪞 Important: the nameref attribute cannot be applied to array variables.
-r — makes names readonly. After that, these names cannot be assigned values by subsequent assignment statements or unset. Locked, throw away the key. 🔒
-t — gives each name the trace attribute. Traced functions inherit the DEBUG and RETURN traps from the calling shell. The trace attribute has no special meaning for variables, just so you know.
-u — when the variable is assigned a value, all lower-case characters get converted to upper-case. The lower-case attribute is disabled.
-x — marks each name for export to subsequent commands via the environment. ✈️

And here's the sneaky bit: using + instead of - turns off the specified attribute instead. But there are exceptions, and they matter — +a and +A may not be used to destroy array variables, and +r will not remove the readonly attribute. (Once you're readonly you're readonly, bestie. It's giving permanent.)

When used in a function, declare and typeset make each name local, just like the local command — unless you supplied -g. If a variable name is followed by =value, the variable's value gets set to value. And heads up 🙈: when you use -a or -A with the compound assignment syntax to create array variables, any additional attributes do not take effect until subsequent assignments.

Return value: 0 unless you hit an invalid option, or you try to define a function using "-f foo=bar" (nope), or you try to assign a value to a readonly variable, or you try to assign a value to an array variable without using the compound assignment syntax (see Arrays above), or one of the names isn't a valid shell variable name, or you try to turn off readonly status for a readonly variable, or you try to turn off array status for an array variable, or you try to display a non-existent function with -f. That's a whole list, I know, but now you know. 💀

🦋 `dirs` 🦋

dirs [-clpv] [+n] [-n]

No options? It displays the list of currently remembered directories. The default display is on a single line with directory names separated by spaces. ✨

How do directories even get in there? You add them with the pushd command, and the popd command removes entries from the list. And the current directory is always the first directory in the stack — main character energy, always at the front 👑.

Options, if you supply them, mean this:

-c — clears the directory stack by deleting all of the entries. Poof. 💨
-l — produces a listing using full pathnames; the default listing format uses a tilde to denote the home directory.
-p — prints the directory stack with one entry per line.
-v — prints the directory stack with one entry per line, prefixing each entry with its index in the stack.
+n — displays the *n*th entry counting from the left of the list shown by dirs when invoked without options, starting with zero.
-n — displays the *n*th entry counting from the right of the list shown by dirs when invoked without options, starting with zero.

Return value is 0 unless an invalid option is supplied, or n indexes beyond the end of the directory stack. (Out of bounds, girlie. 😭)

💅 `disown` 💅

disown [-ar] [-h] [id ...]

No options = remove each id from the table of active jobs. Each id can be a job specification jobspec or a process ID pid; and if id is a pid, disown uses the job containing pid as jobspec. 🎯

-h — plot twist: disown does not remove the jobs corresponding to each id from the jobs table. Instead it marks them so the shell doesn't send SIGHUP to the job if the shell itself receives a SIGHUP. Protective bestie behavior 🫶.

And if you supply no id at all:

-a means remove or mark all jobs.
-r (with no id argument) removes or marks running jobs.
If no id is supplied and neither -a nor -r is given, disown removes or marks the current job.

Return value is 0 unless an id doesn't specify a valid job. 🚫

⭐ `echo` ⭐

echo [-neE] [arg ...]

Output the args, separated by spaces, followed by a newline. Return status is 0 unless a write error occurs. And if -n is specified, the trailing newline is not printed (classic move, very useful).

If -e is given, echo interprets the following backslash-escaped characters. The -E option disables interpretation of these escape characters — even on systems where they get interpreted by default. And the xpg_echo shell option determines whether or not echo interprets any options and expands these escape characters.

Heads up 🫢: echo does not interpret -- to mean the end of options. It just doesn't. Don't come for me.

Here's the full escape sequence menu echo understands:

\a — alert (bell)
\b — backspace
\c — suppress further output
\e, \E — an escape character
\f — form feed
\n — new line
\r — carriage return
\t — horizontal tab
\v — vertical tab
\\ — backslash
\0nnn — the eight-bit character whose value is the octal value nnn (zero to three octal digits).
\xHH — the eight-bit character whose value is the hexadecimal value HH (one or two hex digits).
\uHHHH — the Unicode (ISO/IEC 10646) character whose value is the hexadecimal value HHHH (one to four hex digits).
\UHHHHHHHH — the Unicode (ISO/IEC 10646) character whose value is the hexadecimal value HHHHHHHH (one to eight hex digits).

And echo writes any unrecognized backslash-escaped characters unchanged — it's not gonna guess what you meant, bestie. 💖

🧚 `enable` 🧚

enable [-a] [-dnps] [-f filename] [name ...]

Enable and disable builtin shell commands. 🎛️ And here's why that matters: disabling a builtin allows an executable file that has the same name as a shell builtin to be executed without specifying a full pathname — even though the shell normally searches for builtins before files. So yeah, it's a whole priority flip. ✨

the -n flag is giving disable energy — every name you hand over gets switched off. no -n? then those names are enabled, bestie. periodt. 💅

like, say you wanna use the test binary you found chillin' in your PATH instead of the shell builtin version — you literally just run enable -n test and boom, done. ✨

now, if you pass no name arguments at all, or if you toss in -p, it prints a whole list of shell builtins:

no other option arguments? the list is all the enabled ones.
-n? it prints only the disabled builtins. 💀
-a? the list includes every builtin, with a lil' indication of whether each one is enabled or not.
-s? that restricts the output to the POSIX special builtins. iconic.

and then there's -f, which loads a brand new builtin command name from a shared object filename — on systems that support dynamic loading, iykyk. if filename has no slash in it, Bash goes and uses the value of the BASH_LOADABLES_PATH variable, which is a colon-separated list of directories to search for filename in. 🦋 the default for BASH_LOADABLES_PATH is system-dependent, and it may include . to force a search of the current directory. the -d option will delete a builtin you previously loaded with -f. and if you use -s together with -f, the new builtin becomes a POSIX special builtin. slay.

if you supply no options and the name isn't a shell builtin, enable will attempt to load name from a shared object named name, as if the command were enable -f name name. cute, right?

the return value is 0 unless a name is not a shell builtin, or there's an error loading a new builtin from a shared object. 🫶

`eval`

eval [arg ...]

concatenate the args together into one single command, separating them with spaces. then Bash reads and executes that command, and hands you back its exit status as the return status of eval. no args, or only null arguments? eval returns 0. that's it, that's the whole vibe. ✨

`exec`

exec [-cl] [-a name] [command [arguments]]

if command is specified, it replaces the shell without creating a new process. 😳 and heads up — command cannot be a shell builtin or function. the arguments become the arguments to command.

-l? the shell places a dash at the beginning of the zeroth argument passed to command. this is literally what login(1) does.
-c? causes command to be executed with an empty environment.
-a? the shell passes name as the zeroth argument to the executed command.

if command cannot be executed for some reason, a non-interactive shell exits — unless the execfail shell option is enabled. in that case, it returns a non-zero status. an interactive shell returns a non-zero status if the file cannot be executed. and a subshell exits unconditionally if exec fails. no takebacks, bestie. 💀

if command is not specified, any redirections take effect in the current shell, and the return status is 0. if there's a redirection error, the return status is 1.

`exit`

exit [n]

cause the shell to exit with a status of n. if n is omitted, the exit status is that of the last command executed. any trap on EXIT gets executed before the shell terminates. so like, say your goodbyes first. 👑

`export`

export [-fn] [name[=value]] ...
export -p [-f]

the names you hand over get marked for automatic export to the environment of all the commands executed after this. if -f is given, the names refer to functions instead.

-n unexports — aka removes the export attribute — from each name. if you give no names, or if you only supply -p, export displays a list of the names of all exported variables on standard output. using -p and -f together displays exported functions. and -p shows the output in a form that may be reused as input. 🎀

export also lets you set the value of a variable right when it's exported or unexported, by following the variable name with =value. this sets the value of the variable to value while modifying the export attribute. it returns an exit status of 0 unless:

an invalid option is encountered,
one of the names is not a valid shell variable name, or
-f is supplied with a name that is not a function.

`false`

does nothing. returns a non-zero status. that's the whole personality. 💀 literally just existing to be negative.

`fc`

fc [-e ename] [-lnr] [first] [last]
fc -s [pat=rep] [cmd]

the first form selects a range of commands from first to last out of the history list, and then displays them, or edits and re-executes them. first and last can be a string (to locate the last command beginning with that string) or a number (an index into the history list, where a negative number is used as an offset from the current command number). 🌸

when listing, a first or last of 0 is equivalent to -1, and -0 is equivalent to the current command (usually the fc command itself). otherwise, 0 is equivalent to -1 and -0 is invalid. if last isn't specified, it's set to the current command for listing (so fc -l -10 prints the last 10 commands) and to first otherwise. if first isn't specified, it's set to the previous command for editing, and to -16 for listing.

-l? the commands are listed on standard output.
-n? suppresses the command numbers when listing.
-r? reverses the order of the commands.

otherwise, fc invokes the editor named by ename on a file containing those commands. if ename isn't supplied, fc uses the value of the FCEDIT variable, and the value of EDITOR if FCEDIT isn't set. if neither variable is set, fc uses vi. when editing is complete, fc reads the file containing the edited commands and echoes and executes them. ✨

in the second form, fc re-executes command after replacing each instance of pat with rep. command is interpreted the same as first above.

a useful alias to use with fc is r="fc -s", so that typing r cc runs the last command beginning with cc, and typing r re-executes the last command. obsessed. 🔥

if the first form is used, the return value is zero unless an invalid option is encountered, or first or last specify history lines out of range. when editing and re-executing a file of commands, the return value is the value of the last command executed, or failure if an error occurs with the temporary file. if the second form is used, the return status is that of the re-executed command — unless cmd doesn't specify a valid history entry, in which case fc returns a non-zero status.

`fg`

fg [jobspec]

resume jobspec in the foreground, and make it the current job. if jobspec isn't present, fg uses the shell's notion of the current job. the return value is that of the command placed into the foreground, or failure if it's run when job control is disabled, or — when run with job control enabled — if jobspec doesn't specify a valid job, or if jobspec specifies a job that was started without job control. 👑

`getopts`

getopts optstring name [arg ...]

getopts is what shell scripts and functions use to parse positional parameters and grab options and their arguments. optstring holds the option characters to recognize; if a character is followed by a colon, that option is expected to have an argument, which should be separated from it by white space. and bestie, the colon and question mark characters may not be used as option characters. 🙈

every time it's invoked, getopts places the next option in the shell variable name — initializing name if it doesn't exist — and puts the index of the next argument to be processed into the variable OPTIND. OPTIND is initialized to 1 each time the shell or a shell script is invoked. when an option requires an argument, getopts places that argument into the variable OPTARG.

the shell does not reset OPTIND automatically. you have to manually reset it between multiple calls to getopts within the same shell invocation if you wanna use a new set of parameters. 💅

when it reaches the end of the options, getopts exits with a return value greater than zero. OPTIND is set to the index of the first non-option argument, and name is set to ?.

getopts normally parses the positional parameters, but if more arguments are supplied as arg values, it parses those instead.

getopts can report errors in two ways. if the first character of optstring is a colon, getopts uses silent error reporting. in normal operation, getopts prints diagnostic messages when it hits invalid options or missing option arguments. if the variable OPTERR is set to 0, getopts doesn't display any error messages — even if the first character of optstring is not a colon. 🦋

if getopts detects an invalid option, it places ? into name and, if not silent, prints an error message and unsets OPTARG. if getopts is silent, it assigns the option character found to OPTARG and does not print a diagnostic message.
if a required argument is not found, and getopts is not silent, it sets the value of name to a question mark (?), unsets OPTARG, and prints a diagnostic message. if getopts is silent, it sets the value of name to a colon (:) and sets OPTARG to the option character found.

getopts returns true if an option — specified or unspecified — is found. it returns false if the end of options is encountered or an error occurs. 🌈

`hash`

hash [-lr] [-p filename] [-dt] [name]

each time hash is invoked, it remembers the full pathname of the command name as determined by searching the directories in $PATH. any previously-remembered pathname associated with name gets discarded. if the -p option is supplied, hash uses filename as the full pathname of the command. ✨

-r? causes the shell to forget all remembered locations. (also: assigning to the PATH variable clears all hashed filenames too, sneaky.)
-d? causes the shell to forget the remembered location of each name.
-t? hash prints the full pathname corresponding to each name. if multiple name arguments are supplied with -t, hash prints the name before the corresponding hashed full pathname.
-l? displays output in a format that may be reused as input. 🎀

Okay so if you hand hash no arguments at all, or literally just -l, it prints out the tea on the commands it's been remembering. 💅 And listen — the -t, -d, and -p options (the ones that actually mess with the name arguments) are mutually exclusive, bestie. Only ONE of them is active at a time. If you're chaotic and supply more than one, here's the priority order: -t beats -p, and both of them beat -d. 🫶

The return status is zero unless a name isn't found, or you supplied an invalid option.

✨ `help` ✨

help [-dms] [pattern]

This one displays the helpful info about builtin commands. 🎀 If you give it a pattern, help gives you the detailed help on all commands matching that pattern (more on how that works below). Otherwise? It just dumps a list of ALL the builtins and shell compound commands. It's giving directory energy.

Options, if supplied, have the following meanings:

-d — Display a short description of each pattern.
-m — Display the description of each pattern in a manpage-like format.
-s — Display only a short usage synopsis for each pattern.

So if pattern contains pattern matching characters (peep the Pattern Matching section above 👀), it gets treated as a shell pattern, and help prints the description of every help topic matching pattern.

If it doesn't, and pattern exactly matches the name of a help topic, help prints the description tied to that topic. Otherwise, help does prefix matching and prints the descriptions of all the matching help topics.

The return status is 0 unless no command matches pattern.

✨ `history` ✨

history [n]
history -c
history -d offset
history -d start-end
history -anrw [filename]
history -p arg [arg ...]
history -s arg [arg ...]

With no options, it shows you the command history list, numbered. 📜 Entries prefixed with a * have been modified, iykyk. An argument of n lists only the last n entries. If the shell variable HISTTIMEFORMAT is set and not null, it gets used as a format string for strftime(3) to display the time stamp attached to each displayed history entry. And heads up — if history uses HISTTIMEFORMAT, it does not print an intervening space between the formatted time stamp and the history entry. Periodt.

If filename is supplied, history uses it as the name of the history file; if not, it uses the value of HISTFILE. If filename isn't supplied and HISTFILE is unset or null, then the -a, -n, -r, and -w options have no effect. 💀

Options, if supplied, have the following meanings:

-c — Clear the history list by deleting all the entries. This can be used with the other options to replace the history list.
-d offset — Delete the history entry at position offset. If offset is negative, it's interpreted as relative to one greater than the last history position, so negative indices count back from the end of the history, and an index of -1 refers to the current history -d command.
-d start-end — Delete the range of history entries between positions start and end, inclusive. Positive and negative values for start and end are interpreted as described above.
-a — Append the "new" history lines to the history file. These are history lines entered since the beginning of the current bash session, but not already appended to the history file.
-n — Read the history lines not already read from the history file and add them to the current history list. These are lines appended to the history file since the beginning of the current bash session.
-r — Read the history file and append its contents to the current history list.
-w — Write the current history list to the history file, overwriting the history file.
-p — Perform history substitution on the following args and display the result on the standard output, without storing the results in the history list. Each arg must be quoted to disable normal history expansion.
-s — Store the args in the history list as a single entry. The last command in the history list is removed before adding the args.

If the HISTTIMEFORMAT variable is set, history writes the time stamp information associated with each history entry to the history file, marked with the history comment character as described above. When the history file is read, lines beginning with the history comment character followed immediately by a digit are interpreted as timestamps for the following history entry. 🕰️

The return value is 0 unless an invalid option is encountered, an error occurs while reading or writing the history file, an invalid offset or range is supplied as an argument to -d, or the history expansion supplied as an argument to -p fails.

✨ `jobs` ✨

jobs [-lnprs] [ jobspec ... ]
jobs -x command [ args ... ]

The first form lists the active jobs. The options have the following meanings:

-l — List process IDs in addition to the normal information.
-n — Display information only about jobs that have changed status since the user was last notified of their status.
-p — List only the process ID of the job's process group leader.
-r — Display only running jobs.
-s — Display only stopped jobs.

If jobspec is supplied, jobs restricts output to information about that job. The return status is 0 unless an invalid option is encountered or an invalid jobspec is supplied.

If the -x option is supplied, jobs replaces any jobspec found in command or args with the corresponding process group ID, and executes command, passing it args, returning its exit status. 🔥

✨ `kill` ✨

kill [-s sigspec | -n signum | -sigspec] id [ ... ]
kill -l|-L [sigspec | exit_status]

Send the signal specified by sigspec or signum to the processes named by each id. Each id may be a job specification jobspec or a process ID pid. sigspec is either a case-insensitive signal name such as SIGKILL (with or without the SIG prefix) or a signal number; signum is a signal number. If sigspec is not supplied, then kill sends SIGTERM. ⚰️

The -l option lists the signal names. If any arguments are supplied when -l is given, kill lists the names of the signals corresponding to the arguments, and the return status is 0. The exit_status argument to -l is a number specifying either a signal number or the exit status of a process terminated by a signal; if it is supplied, kill prints the name of the signal that caused the process to terminate. kill assumes that process exit statuses are greater than 128; anything less than that is a signal number. The -L option is equivalent to -l.

kill returns true if at least one signal was successfully sent, or false if an error occurs or an invalid option is encountered.

✨ `let` ✨

let arg [arg ...]

Each arg is evaluated as an arithmetic expression (go peek at ARITHMETIC EVALUATION above 🧮). If the last arg evaluates to 0, let returns 1; otherwise let returns 0.

✨ `local` ✨

local [option] [name[=value] ... | - ]

For each argument, create a local variable named name and assign it value. The option can be any of the options accepted by declare. When local is used within a function, it causes the variable name to have a visible scope restricted to that function and its children. It is an error to use local when not within a function. 🚨

If name is -, it makes the set of shell options local to the function in which local is invoked: any shell options changed using the set builtin inside the function after the call to local are restored to their original values when the function returns. The restore is performed as if a series of set commands were executed to restore the values that were in place before the function. 🦋

With no operands, local writes a list of local variables to the standard output.

The return status is 0 unless local is used outside a function, an invalid name is supplied, or name is a readonly variable.

✨ `logout` ✨

logout [n]

Exit a login shell, returning a status of n to the shell's parent. 👋 Byeee.

✨ `mapfile` ✨

mapfile [-d delim] [-n count] [-O origin] [-s count] [-t] [-u fd] [-C
callback] [-c quantum] [array]

✨ `readarray` ✨

readarray [-d delim] [-n count] [-O origin] [-s count] [-t] [-u fd] [-C
callback] [-c quantum] [array]

Read lines from the standard input, or from file descriptor fd if the -u option is supplied, into the indexed array variable array. The variable MAPFILE is the default array. 📚 Options, if supplied, have the following meanings:

-d — Use the first character of delim to terminate each input line, rather than newline. If delim is the empty string, mapfile will terminate a line when it reads a NUL character.
-n — Copy at most count lines. If count is 0, copy all lines.
-O — Begin assigning to array at index origin. The default index is 0.
-s — Discard the first count lines read.
-t — Remove a trailing delim (default newline) from each line read.
-u — Read lines from file descriptor fd instead of the standard input.
-C — Evaluate callback each time quantum lines are read. The -c option specifies quantum.
-c — Specify the number of lines read between each call to callback.

If -C is specified without -c, the default quantum is 5000. When callback is evaluated, it is supplied the index of the next array element to be assigned and the line to be assigned to that element as additional arguments. callback is evaluated after the line is read but before the array element is assigned. ✨

If not supplied with an explicit origin, mapfile will clear array before assigning to it.

mapfile returns zero unless an invalid option or option argument is supplied, array is invalid or unassignable, or if array is not an indexed array. 💜

`popd` 💅

popd [-n] [+n] [-n]

Okay so popd is the ✨undo✨ button for your directory stack. It removes entries from the directory stack, bestie. The elements are numbered from 0 starting at the first directory listed by dirs, so plain popd is literally the same thing as “popd +0.” No cap. With no arguments, popd removes the top directory from the stack, and changes to the new top directory. 🎀

Arguments, if you supply them, mean this:

-n — suppress the normal change of directory when removing directories from the stack; only manipulate the stack. Like, touch the stack but don't move me. 🙈
+n — remove the *n*th entry counting from the left of the list shown by dirs, starting with zero, from the stack. For example: “popd +0” removes the first directory, “popd +1” the second. Count from the left, babe. ⭐
-n — remove the *n*th entry counting from the right of the list shown by dirs, starting with zero. For example: “popd -0” removes the last directory, “popd -1” the next to last. Count from the right this time. 🦋

If the top element of the directory stack is modified, and the -n option was not supplied, popd uses the cd builtin to change to the directory at the top of the stack. If the cd fails, popd returns a non-zero value. 💀

Otherwise, popd returns false if an invalid option is supplied, the directory stack is empty, or n specifies a non-existent directory stack entry.

If the popd command is successful, bash runs dirs to show the final contents of the directory stack, and the return status is 0. Slay. ✨

`printf` 🖨️💖

printf [-v var] format [arguments]

Write the formatted arguments to the standard output, under the control of the format. The -v option assigns the output to the variable var rather than printing it to the standard output. So like, it goes in the box instead of on the screen. 📦

The format is a character string which contains three types of objects:

plain characters — which are simply copied to standard output,
character escape sequences — which are converted and copied to the standard output,
format specifications — each of which causes printing of the next successive argument.

In addition to the standard printf(3) format characters cCsSndiouxXeEfFgGaA, printf interprets the following additional format specifiers:

%b — causes printf to expand backslash escape sequences in the corresponding argument in the same way as echo -e.
%q — causes printf to output the corresponding argument in a format that can be reused as shell input. %q and %Q use the $'' quoting style if any characters in the argument string require it, and backslash quoting otherwise. If the format string uses the printf alternate form, these two formats quote the argument string using single quotes.
%Q — like %q, but applies any supplied precision to the argument before quoting it. 👑
%(datefmt)T — causes printf to output the date-time string resulting from using datefmt as a format string for strftime(3). The corresponding argument is an integer representing the number of seconds since the epoch. This format specifier recognizes two special argument values: -1 represents the current time, and -2 represents the time the shell was invoked. If no argument is specified, conversion behaves as if -1 had been supplied. This is an exception to the usual printf behavior. ⏰

The %b, %q, and %T format specifiers all use the field width and precision arguments from the format specification and write that many bytes from (or use that wide a field for) the expanded argument, which usually contains more characters than the original. iykyk.

The %n format specifier accepts a corresponding argument that is treated as a shell variable name.

The %s and %c format specifiers accept an l (long) modifier, which forces them to convert the argument string to a wide-character string and apply any supplied field width and precision in terms of characters, not bytes. The %S and %C format specifiers are equivalent to %ls and %lc, respectively.

Arguments to non-string format specifiers are treated as C constants, except that a leading plus or minus sign is allowed, and if the leading character is a single or double quote, the value is the numeric value of the following character, using the current locale.

The format is reused as necessary to consume all of the arguments. If the format requires more arguments than are supplied, the extra format specifications behave as if a zero value or null string, as appropriate, had been supplied. The return value is zero on success, non-zero if an invalid option is supplied or a write or assignment error occurs. 🫶

`pushd` 🎀

pushd [-n] [+n] [-n]
pushd [-n] [dir]

Add a directory to the top of the directory stack, or rotate the stack, making the new top of the stack the current working directory. With no arguments, pushd exchanges the top two elements of the directory stack. Arguments, if supplied, have the following meanings:

-n — suppress the normal change of directory when rotating or adding directories to the stack; only manipulate the stack. Just stack things, don't move me. 💅
+n — rotate the stack so that the *n*th directory (counting from the left of the list shown by dirs, starting with zero) is at the top.
-n — rotates the stack so that the *n*th directory (counting from the right of the list shown by dirs, starting with zero) is at the top.
dir — adds dir to the directory stack at the top. 🌸

After the stack has been modified, if the -n option was not supplied, pushd uses the cd builtin to change to the directory at the top of the stack. If the cd fails, pushd returns a non-zero value.

Otherwise, if no arguments are supplied, pushd returns zero unless the directory stack is empty. When rotating the directory stack, pushd returns zero unless the directory stack is empty or n specifies a non-existent directory stack element.

If the pushd command is successful, bash runs dirs to show the final contents of the directory stack. ✨

`pwd` 📍

pwd [-LP]

Print the absolute pathname of the current working directory. Where am I?? 🗺️

The pathname printed contains no symbolic links if the -P option is supplied or the -o physical option to the set builtin command is enabled. If the -L option is used, the pathname printed may contain symbolic links.

The return status is 0 unless an error occurs while reading the name of the current directory or an invalid option is supplied. 🦋

`read` 📖💜

read [-Eers] [-a aname] [-d delim] [-i text] [-n nchars] [-N nchars] [-p prompt] [-t timeout] [-u fd] [name ...]

Read one line from the standard input, or from the file descriptor fd supplied as an argument to the -u option, split it into words as described above under Word Splitting, and assign the first word to the first name, the second word to the second name, and so on. It's giving very "pass the notes around class" energy, honestly. 💌

If there are more words than names, the remaining words and their intervening delimiters are assigned to the last name. If there are fewer words read from the input stream than names, the remaining names are assigned empty values. The characters in the value of the IFS variable are used to split the line into words using the same rules the shell uses for expansion (described above under Word Splitting).

The backslash character (\) removes any special meaning for the next character read and is used for line continuation. 🧚

-a aname — bestie, this one drops all the words into the array variable aname, filling it up index by index starting at 0 💖. Heads up: aname gets unset before any new values land in it, so nothing old survives. Any other name arguments you passed? Ignored. Sorry not sorry.
-d delim — the very first character of delim is what ends the input line, instead of a newline 🎀. And if delim is the empty string, read terminates a line the moment it reads a NUL character. Iconic little trick.
-e — if standard input is coming from a terminal, read grabs the line using readline (peep READLINE above ✨). Readline uses the current editing settings (or the default ones, if line editing wasn't already active), but it uses readline's default filename completion.
-E — same terminal situation, read uses readline again, current-or-default editing settings again, but this time it's bash's default completion, programmable completion and all 👑.
-i text — if readline is the one reading the line, read tucks text into the editing buffer before editing even starts, so you can edit it 🦋.
-n nchars — read comes back after it's read nchars characters instead of waiting around for a whole line, unless it hits EOF or times out 🙈. But it still honors a delimiter if it reads fewer than nchars characters before that delimiter shows up.
-N nchars — read returns after reading exactly nchars characters, not waiting for a full line, unless it hits EOF or times out. Any delimiter characters in the input? Not special at all, and they do not make read return until it has read nchars characters 💅. Also the result does not get split on the characters in IFS — the whole point is that the variable gets assigned exactly the characters read (except for backslash; see the -r option below, iykyk).
-p prompt — show prompt on standard error, with no trailing newline, before trying to read any input — but only if the input is coming from a terminal 🍓.
-r — backslash stops being an escape character 💀. It's just part of the line now. In particular, a backslash-newline pair can't be used as a line continuation anymore.
-s — silent mode 🌸. If input is coming from a terminal, the characters are not echoed. Sneaky typing, no witnesses.
-t timeout — makes read time out and fail if it doesn't read a complete line of input (or the specified number of characters) within timeout seconds ⏰. timeout can be a decimal number with a fractional part after the decimal point. This only does anything if read is reading from a terminal, a pipe, or some other special file — it has no effect when reading from regular files. If read times out, it saves whatever partial input it got into the specified variable name, and the exit status is greater than 128. And if timeout is 0, read returns immediately without even trying to read data — in that case the exit status is 0 if input is available on the specified file descriptor, or the read will return EOF, and non-zero otherwise.
-u fd — read input from file descriptor fd instead of standard input 🌈.

Other than the case where delim is the empty string, read ignores any NUL characters in the input. Noted, moving on 💜.

If you supply no names at all, read assigns the line it read — without the ending delimiter, but otherwise unmodified — to the variable REPLY ✨.

The exit status is zero, unless end-of-file is encountered, read times out (in which case the status is greater than 128), a variable assignment error happens (like assigning to a readonly variable 😭), or you handed it an invalid file descriptor as the argument to -u.

🫶 `readonly` 🫶

readonly [-aAf] [-p] [name[=word] ...]

The given names get marked readonly — meaning their values cannot be changed by any later assignment or unset 🔒. If you pass -f, then each name refers to a shell function instead. The -a option restricts the variables to indexed arrays, and -A restricts them to associative arrays. If you supply both, -A takes precedence 👑. If you supply no name arguments, or if you pass -p, it prints a list of all readonly names — and the other options can be used to narrow that output down to a subset of the readonly names. The -p option displays output in a format that can be reused as input 💅.

readonly also lets you set a variable's value at the same time you change the readonly attribute, by following the variable name with =value. That sets the value of the variable to value while it's modifying the readonly attribute ✨.

The return status is 0 unless an invalid option shows up, one of the names isn't a valid shell variable name, or -f is given with a name that isn't a function 💀.

💫 `return` 💫

return [n]

Stop executing a shell function or a sourced file, and hand the value specified by n back to whoever called it 🌸. If n is omitted, the return status is that of the last command executed. If return is executed by a trap handler, the last command used to determine the status is the last command executed before the trap handler. And if return is executed during a DEBUG trap, the last command used to determine the status is the last command executed by the trap handler before return was invoked 🫶.

When return is used to terminate a script that's being executed by the . (source) command, it makes the shell stop executing that script and return either n or the exit status of the last command executed inside the script as the exit status of the script. If n is supplied, the return value is its least significant 8 bits 🎀.

Any command associated with the RETURN trap gets executed before execution resumes after the function or script 👑.

The return status is non-zero if return is given a non-numeric argument, or if it's used outside a function and not during execution of a script by . or source 🙈.

⭐ `set` ⭐

set [-abefhkmnptuvxBCEHPT] [-o option-name] [--] [-] [arg ...]
set [+abefhkmnptuvxBCEHPT] [+o option-name] [--] [-] [arg ...]
set -o
set +o

With no options, this displays the name and value of every shell variable, in a format you can reuse as input to set or reset the currently-set variables ✨. Read-only variables can't be reset, periodt. In posix mode, only shell variables get listed. The output is sorted according to the current locale 🌍. When options are specified, they set or unset shell attributes. Any arguments left over after option processing get treated as values for the positional parameters and are assigned, in order, to $1, $2, ..., $n 💅.

Options, if specified, mean the following:

-a — every variable or function that gets created or modified is given the export attribute and marked for export to the environment of subsequent commands 🌈.
-b — report the status of terminated background jobs immediately, instead of before the next primary prompt or after a foreground command terminates. This only does anything when job control is enabled ⚡.
-e — exit immediately if a pipeline (which can be a single simple command), a list, or a compound command (peep SHELL GRAMMAR above) exits with a non-zero status 💀. But wait, bestie — the shell does not exit if the failing command is: part of the command list immediately following a while or until reserved word; part of the test following the if or elif reserved words; part of any command executed in a && or || list except the command following the final && or ||; any command in a pipeline but the last one (this is subject to the state of the pipefail shell option); or if the command's return value is being inverted with ! 🙈. If a compound command other than a subshell returns a non-zero status because a command failed while -e was being ignored, the shell does not exit. A trap on ERR, if set, is executed before the shell exits. This option applies to the shell environment and each subshell environment separately (see COMMAND EXECUTION ENVIRONMENT above), and it may cause subshells to exit before they've executed all the commands in the subshell 😭.

If a compound command or a shell function is running somewhere that -e is being ignored, then literally none of the commands inside that compound command or function body get touched by -e — even if -e is set and a command comes back with a failure status. 😤 And if a compound command or shell function sets -e while it's executing in a context where -e is ignored? That setting does absolutely nothing until the compound command, or the command containing the function call, finishes. no cap.

-f — turns off pathname expansion. bye globbing bestie 🦋
-h — remembers where commands are located as they get looked up for execution. This one is on by default. 💅
-k — every argument shaped like an assignment statement gets dropped into the environment for a command, not just the ones sitting in front of the command name.
-m — monitor mode, aka job control is enabled. This is on by default for interactive shells on systems that support it (peep JOB CONTROL above). Every process runs in its own separate process group. And when a background job finishes, the shell prints a line with its exit status. ✨
-n — reads commands but does not execute them. Perfect for vibe-checking a shell script for syntax errors. Interactive shells just ignore this one. 🙈
-o option-name — and option-name can be any of these:
- allexport — same as -a.
- braceexpand — same as -B.
- emacs — use an emacs-style command line editing interface. On by default when the shell is interactive, unless the shell got started with the --noediting option. This also affects the editing interface used for read -e.
- errexit — same as -e.
- errtrace — same as -E.
- functrace — same as -T.
- hashall — same as -h.
- histexpand — same as -H.
- history — enables command history, exactly like it says up in HISTORY. This is on by default in interactive shells.
- ignoreeof — the effect is as if the shell command IGNOREEOF=10 had been executed (check Shell Variables above 👀).
- keyword — same as -k.
- monitor — same as -m.
- noclobber — same as -C.
- noexec — same as -n.
- noglob — same as -f.
- nolog — currently ignored. 💀
- notify — same as -b.
- nounset — same as -u.
- onecmd — same as -t.
- physical — same as -P.
- pipefail — if set, the return value of a pipeline is the value of the last (rightmost) command to exit with a non-zero status, or zero if every command in the pipeline exited successfully. Disabled by default.
- posix — enables posix mode, which changes bash's behavior wherever the default operation differs from the POSIX standard so it matches the standard instead. See SEE ALSO below for a reference to a document that spills all the tea on how posix mode affects bash's behavior.
- privileged — same as -p.
- verbose — same as -v.
- vi — use a vi-style command line editing interface. This also affects the editing interface used for read -e.
- xtrace — same as -x.

If -o is supplied with no option-name, set prints the current shell option settings. And if +o is supplied with no option-name, set prints a whole series of set commands to the standard output so you can recreate the current option settings. iconic.

-p — turns on privileged mode. In this mode the shell does not read the $ENV and $BASH_ENV files, shell functions are not inherited from the environment, and the SHELLOPTS, BASHOPTS, CDPATH, and GLOBIGNORE variables — if they show up in the environment — are ignored. If the shell is started with the effective user (group) id not equal to the real user (group) id, and -p was not supplied, these actions get taken and the effective user id is set to the real user id. If -p was supplied at startup, the effective user id is not reset. Turning this option off sets the effective user and group ids to the real user and group ids.
-r — enables restricted shell mode. ⚠️ This option cannot be unset once it has been set. Like, ever.
-t — exits after reading and executing one command.
-u — treats unset variables and parameters other than the special parameters @ and *, or array variables subscripted with @ or *, as an error when performing parameter expansion. If expansion gets attempted on an unset variable or parameter, the shell prints an error message and, if it's not interactive, exits with a non-zero status. 😭
-v — prints shell input lines as they are read.
-x — after expanding each simple command, for command, case command, select command, or arithmetic for command, displays the expanded value of PS4, followed by the command and its expanded arguments or associated word list, all to the standard error. 🔍
-B — the shell performs brace expansion (see Brace Expansion above). On by default.
-C — if set, bash does not overwrite an existing file with the >, >&, and <> redirection operators. Using the redirection operator >| instead of > overrides this and forces the creation of an output file. 💅
-E — if set, any trap on ERR is inherited by shell functions, command substitutions, and commands executed in a subshell environment. Normally the ERR trap is not inherited in those cases.
-H — enables ! style history substitution. On by default when the shell is interactive.
-P — if set, the shell does not resolve symbolic links when executing commands such as cd that change the current working directory. It uses the physical directory structure instead. By default, bash follows the logical chain of directories when performing commands that change the current directory.
-T — if set, any traps on DEBUG and RETURN are inherited by shell functions, command substitutions, and commands executed in a subshell environment. Normally the DEBUG and RETURN traps are not inherited in those cases.
-- — if no arguments follow this option, it unsets the positional parameters. Otherwise, it sets the positional parameters to the args, even if some of them begin with a -.
- — signals the end of options and assigns all remaining args to the positional parameters. The -x and -v options get turned off. If there are no args, the positional parameters stay exactly as they were.

The options are off by default unless otherwise noted. Using + instead of - turns these options off. The options can also be specified as arguments to an invocation of the shell. You can always find the current set of options in $-. The return status is always zero unless an invalid option gets encountered.

🌸 shift 🌸

shift [n]

Renames the positional parameters from n+1 ... to $1 .... The parameters represented by the numbers $# down to $#-n+1 get unset. n has to be a non-negative number less than or equal to $#. If n is 0, no parameters change at all. If n isn't given, it's assumed to be 1. And if n is greater than $#, the positional parameters don't change. The return status is greater than zero if n is greater than $# or less than zero; otherwise it's 0. 🫶

🌸 shopt 🌸

shopt [-pqsu] [-o] [optname ...]

Toggles the values of settings that control optional shell behavior. The settings can be the ones listed below, or — if the -o option is used — the ones available with the -o option to the set builtin command.

With no options, or with the -p option, it displays a list of all settable options, with an indication of whether or not each one is set. If any optnames are supplied, the output gets restricted to just those options. The -p option displays output in a form that may be reused as input. 💖

The other options mean this:

-s — enables (sets) each optname.
-u — disables (unsets) each optname.
-q — suppresses normal output (quiet mode, bestie 🤫); the return status tells you whether the optname is set or unset. If multiple optname arguments are supplied with -q, the return status is zero if all optnames are enabled; non-zero otherwise.
-o — restricts the values of optname to those defined for the -o option to the set builtin.

If either -s or -u is used with no optname arguments, shopt shows only those options which are set or unset, respectively. Unless otherwise noted, the shopt options are disabled (unset) by default.

The return status when listing options is zero if all optnames are enabled, non-zero otherwise. When setting or unsetting options, the return status is zero unless an optname is not a valid shell option. 👑

The list of shopt options is:

array_expand_once — when this is set, the shell stops double-evaluating associative and indexed array subscripts 💅 — this kicks in during arithmetic expression evaluation, while running builtins that can perform variable assignments, and while running builtins that do array dereferencing. no repeated side-eyes, periodt.
assoc_expand_once — deprecated bestie 🫶 it's literally just a synonym for array_expand_once. we love a legacy alias.
autocd — if you set this, and you type a command name that happens to be the name of a directory, it gets executed as if it were the argument to the cd command. so basically you can just type the folder name and walk in. this one is only used by interactive shells 🚪✨
bash_source_fullpath — if set, the filenames that get added to the BASH_SOURCE array variable are converted to full pathnames (peep Shell Variables above for the tea ☕).
cdable_vars — if set, then when you hand cd an argument that isn't a directory, bash just assumes it's the name of a variable whose value is the directory you actually wanna change to. so cd myvar works if myvar holds a path. iconic shortcut energy 💖
cdspell — if set, cd tries to fix lil typos in the spelling of a directory component 🔤. the "minor errors" it forgives are transposed characters, a missing character, and one extra character. if cd manages to correct the directory name, it prints the corrected filename and the command just proceeds like nothing happened. only used by interactive shells, bestie.
checkhash — if set, bash double-checks that a command it found in the hash table actually still exists before trying to run it 👀. if a hashed command has vanished, bash just does a normal path search instead. no ghost commands here.
checkjobs — if set, bash lists the status of any stopped and running jobs before it exits an interactive shell. if any jobs are still running, bash defers the exit until you try to exit a second time without any command in between (see JOB CONTROL above). and ngl the shell always postpones exiting if any jobs are stopped — it's not letting you leave your kids behind 😭
checkwinsize — if set, bash checks the window size after every external (non-builtin) command and, if needed, updates the values of LINES and COLUMNS, using the file descriptor associated with standard error if that's a terminal. this one is enabled by default 📐✨
cmdhist — if set, bash tries to save all the lines of a multiple-line command into the same history entry, which makes re-editing multi-line commands so much easier 🎀. enabled by default, but it only does anything if command history is actually enabled, as described up top under HISTORY.
compat31, compat32, compat40, compat41, compat42, compat43, compat44 — these all control different aspects of the shell's compatibility mode (see SHELL COMPATIBILITY MODE below). the nostalgia collection, fr fr 🕰️
complete_fullquote — if set, bash quotes all shell metacharacters in filenames and directory names when it's performing completion. if not set, bash removes metacharacters like the dollar sign from the set of characters that get quoted in completed filenames — but only when those metacharacters show up in shell variable references inside the words being completed. what that means: dollar signs in variable names that expand to directories won't be quoted; however, any dollar signs actually appearing in filenames won't be quoted either. this is only active when bash is using backslashes to quote completed filenames. it's set by default, which is the default bash behavior in versions through 4.2 👑
direxpand — if set, bash swaps directory names out for the results of word expansion when performing filename completion, which changes what's sitting in the readline editing buffer. if it's not set, bash tries to preserve whatever you actually typed ✍️
dirspell — if set, bash attempts spelling correction on directory names during word completion if the directory name you initially gave it doesn't exist. auto-correct but make it directories 💅
dotglob — if set, bash includes filenames that begin with a "." in the results of pathname expansion 👻. but heads up: the filenames . and .. must always be matched explicitly, even if dotglob is set.
execfail — if set, a non-interactive shell will not exit if it can't execute the file specified as an argument to the exec builtin. (an interactive shell doesn't exit if exec fails anyway.) 🚪
expand_aliases — if set, aliases get expanded exactly as described above under ALIASES. this option is enabled by default for interactive shells ✨
extdebug — okay this one's got layers, stay with me 🧠✨. if it's set at shell invocation, or in a shell startup file, it arranges to execute the debugger profile before the shell starts — identical to the --debugger option. if it's set after invocation, then the behavior intended for use by debuggers gets enabled:
1. The -F option to the declare builtin displays the source file name and line number corresponding to each function name supplied as an argument.
2. If the command run by the DEBUG trap returns a non-zero value, the next command is skipped and not executed.
3. If the command run by the DEBUG trap returns a value of 2, and the shell is executing in a subroutine (a shell function or a shell script executed by the . or source builtins), the shell simulates a call to return.
4. BASH_ARGC and BASH_ARGV are updated as described in their descriptions above).
5. Function tracing is enabled: command substitution, shell functions, and subshells invoked with ( command ) inherit the DEBUG and RETURN traps.
6. Error tracing is enabled: command substitution, shell functions, and subshells invoked with ( command ) inherit the ERR trap.
extglob — if set, it enables the extended pattern matching features described above under Pathname Expansion 🦋
extquote — if set, $'string' and $"string" quoting is performed within ${parameter} expansions that are enclosed in double quotes. enabled by default 💜
failglob — if set, any pattern that fails to match filenames during pathname expansion results in an expansion error 💀 no silent misses.
force_fignore — if set, the suffixes specified by the FIGNORE shell variable cause words to be ignored when performing word completion even if the ignored words are the only possible completions. see Shell Variables above for a description of FIGNORE. this option is enabled by default 🙈
globasciiranges — if set, range expressions used in pattern matching bracket expressions (see Pattern Matching above) behave as if they were in the traditional C locale when doing comparisons. basically: pattern matching doesn't take the current locale's collating sequence into account, so b will not collate between A and B, and upper-case and lower-case ASCII characters will collate together 🌈
globskipdots — if set, pathname expansion will never match the filenames . and .., even if the pattern begins with a "." 🚫👻. this option is enabled by default.
globstar — if set, the pattern ** used in a pathname expansion context will match all files and zero or more directories and subdirectories. and if the pattern is followed by a /, then only directories and subdirectories match 🔥 the double-star glow-up.
gnu_errfmt — if set, shell error messages are written in the standard GNU error message format 📝
histappend — if set, the history list gets appended to the file named by the value of the HISTFILE variable when the shell exits, rather than overwriting that file. no more losing your history, we stan 💖
histreedit — if set, and readline is being used, you get the opportunity to re-edit a failed history substitution ✍️
histverify — if set, and readline is being used, the results of history substitution are not immediately passed to the shell parser. instead, the resulting line is loaded into the readline editing buffer, letting you modify it further before it actually runs 👀
hostcomplete — if set, and readline is being used, bash will attempt to perform hostname completion when a word containing an @ is being completed (see Completing under READLINE above). this is enabled by default 🌐
huponexit — if set, bash will send SIGHUP to all jobs when an interactive login shell exits 💀 dramatic but fair.
inherit_errexit — if set, command substitution inherits the value of the errexit option, instead of unsetting it in the subshell environment. this option is enabled when posix mode is enabled 👑
interactive_comments — in an interactive shell, a word beginning with # causes that word and all the remaining characters on that line to be ignored, exactly like in a non-interactive shell (see COMMENTS above). this option is enabled by default 💬
lastpipe — if set, and job control is not active, the shell runs the last command of a pipeline that's not executed in the background in the current shell environment 🫶
lithist — if set, and the cmdhist option is enabled, multi-line

commands are saved to the history with embedded newlines rather than using semicolon separators where possible. 💅

localvar_inherit — turn this on and your local variables will inherit the value and the attributes of a same-named variable from a previous scope, all before anything new gets assigned to them. One exception bestie: the nameref attribute does not get inherited. 🙈
localvar_unset — if set, calling unset on local variables that live in previous function scopes marks them so that every lookup after that finds them unset, right up until that function returns. It's literally the exact same behavior as unsetting local variables in the current function scope. 🫶
login_shell — the shell sets this itself when it's started as a login shell (peep INVOCATION above). You cannot change this value. periodt. 🔒
mailwarn — if set, and a file that bash is checking for mail has been accessed since the last time it got checked, bash hits you with the message “The mail in mailfile has been read”. 👀
no_empty_cmd_completion — if set, and readline is in the mix, bash will not search PATH for possible completions when you try to complete on an empty line. 🚫
nocaseglob — if set, bash matches filenames in a case-insensitive fashion when it's doing pathname expansion (see Pathname Expansion above). ✨
nocasematch — if set, bash matches patterns in a case-insensitive fashion when it's matching while executing case or [[ conditional commands, when it's doing pattern substitution word expansions, or when it's filtering possible completions as part of programmable completion. 🌸
noexpand_translation — if set, bash wraps the translated results of $"..." quoting in single quotes instead of double quotes. If the string doesn't get translated, this does absolutely nothing. 💀
nullglob — if set, pathname expansion patterns that match no files at all (see Pathname Expansion above) expand to nothing and get removed, instead of just expanding to themselves. 👑
patsub_replacement — if set, bash expands occurrences of & in the replacement string of pattern substitution to the text that the pattern matched, exactly as described under Parameter Expansion above. This one is enabled by default. 🔥
progcomp — if set, it enables the programmable completion facilities (see Programmable Completion above). Also enabled by default. 💖
progcomp_alias — if set, and programmable completion is enabled, bash treats a command name that has no completions as a possible alias and attempts alias expansion. If it does have an alias, bash attempts programmable completion using the command word that comes out of the expanded alias. 🦋
promptvars — if set, prompt strings go through parameter expansion, command substitution, arithmetic expansion, and quote removal after being expanded as described in PROMPTING above. Yep — enabled by default. ✨
restricted_shell — the shell sets this if it's started in restricted mode (see RESTRICTED SHELL below). The value may not be changed. It's not reset when the startup files are executed either, which lets those startup files figure out whether or not a shell is restricted. 🎀
shift_verbose — if set, the shift builtin prints an error message when the shift count exceeds the number of positional parameters. 😭
sourcepath — if set, the . (source) builtin uses the value of PATH to find the directory containing the file you handed it as an argument, whenever the -p option isn't supplied. Enabled by default, bestie. 💜
varredir_close — if set, the shell automatically closes the file descriptors assigned using the {varname} redirection syntax (see REDIRECTION above) instead of leaving them hanging open when the command finishes. 🧚
xpg_echo — if set, the echo builtin expands backslash-escape sequences by default. And if the posix shell option is also enabled, echo does not interpret any options at all. 🌈

✨ `suspend [-f]` ✨

Pauses the execution of this shell until it catches a SIGCONT signal. 😴

Real talk though: a login shell, or a shell that doesn't have job control enabled, cannot be suspended. The -f option overrides that and forces the suspension anyway. 💅

The return status is 0 unless the shell is a login shell or job control isn't enabled and -f wasn't supplied.

✨ `test expr` ✨

✨ `[ expr ]` ✨

Returns a status of 0 (true) or 1 (false), depending on how the conditional expression expr evaluates. Every single operator and operand has to be its own separate argument, ok? Expressions are built out of the primaries described above under CONDITIONAL EXPRESSIONS.

test does not accept any options, and it does not accept-and-ignore an argument of -- as a way of saying "options are over". Nope. 🚫

Expressions can be combined using these operators, listed in decreasing order of precedence. The evaluation depends on how many arguments you have — see below. test uses operator precedence when there are five or more arguments.

! expr — true if expr is false.
( expr ) — returns the value of expr. You can use this to override normal operator precedence. 💅
expr1 -a expr2 — true if both expr1 and expr2 are true.
expr1 -o expr2 — true if either expr1 or expr2 is true.

test and [ evaluate conditional expressions using a set of rules based on the number of arguments. 🧮

0 arguments

The expression is false.

1 argument

The expression is true if and only if the argument is not null.

2 arguments

If the first argument is !, the expression is true if and only if the second argument is null. If the first argument is one of the unary conditional operators listed above under CONDITIONAL EXPRESSIONS, the expression is true if the unary test is true. If the first argument is not a valid unary conditional operator, the expression is false.

3 arguments

The following conditions get applied in the order they're listed. If the second argument is one of the binary conditional operators listed above under CONDITIONAL EXPRESSIONS, the result of the expression is the result of that binary test using the first and third arguments as operands. The -a and -o operators count as binary operators when there are three arguments. If the first argument is !, the value is the negation of the two-argument test using the second and third arguments. If the first argument is exactly ( and the third argument is exactly ), the result is the one-argument test of the second argument. Otherwise, the expression is false.

4 arguments

Again, applied in the order listed. If the first argument is !, the result is the negation of the three-argument expression composed of the remaining arguments. If the first argument is exactly ( and the fourth argument is exactly ), the result is the two-argument test of the second and third arguments. Otherwise, the expression is parsed and evaluated according to precedence using the rules listed above.

5 or more arguments

The expression is parsed and evaluated according to precedence using the rules listed above.

When the shell is in posix mode, or if the expression is part of the [[ command, the < and > operators sort using the current locale. If the shell is not in posix mode, the test and [ commands sort lexicographically using ASCII ordering. 🌍

⚠️ Okay this part matters: the historical operator-precedence parsing with 4 or more arguments can lead to ambiguities when it runs into strings that look like primaries. The POSIX standard has deprecated the -a and -o primaries and enclosing expressions inside parentheses. Scripts should not be using them anymore. It is much more reliable to keep your test invocations restricted to a single primary, and to replace your -a and -o usage with the shell's && and || list operators. fr fr. 🙏

✨ `times` ✨

Prints the accumulated user and system times for the shell itself and for processes run from the shell. The return status is 0. ⏰

✨ `trap [-lpP] [[action] sigspec ...]` ✨

The action is a command that gets read and executed whenever the shell receives any of the signals sigspec. 🚨

If action is absent (and there's a single sigspec) or is -, each specified sigspec gets reset to the value it had when the shell was started. If action is the null string, the signal specified by each sigspec is ignored by the shell and by the commands it invokes. 🤫

okay bestie, we're still deep in trap land, so let's keep going 💖

If you supply no arguments at all, trap just shows you the actions tied to every trapped signal, formatted as a set of trap commands you could literally paste back in as shell input to restore your current signal dispositions. ✨ If you give it -p and there's no action present, then trap displays the actions for each sigspec, or — if you didn't hand it any sigspec — for all trapped signals, again as reusable-as-shell-input trap commands that restore the current signal dispositions. The -P option acts basically the same, but it only shows the actions for each sigspec argument you pass it. Heads up tho: -P requires at least one sigspec argument, no cap.

Both -P and -p can be used inside a subshell environment (like, say, a command substitution 👀) and — as long as you use them before trap has been used to change a signal's handling — they'll show you the state of its parent's traps. inheritance but make it signals 💅

The -l option prints a list of signal names with their corresponding numbers. Each sigspec is either a signal name defined in <signal.h> or a signal number. Signal names are case insensitive and the SIG prefix is optional (so int, INT, SIGINT — all the same girl 🫶). If you give -l with no sigspec arguments, it prints a list of valid signal names.

If a sigspec is EXIT (0), the action runs on exit from the shell. If a sigspec is DEBUG, the action runs before every simple command, for command, case command, select command, (( arithmetic command, [[ conditional command, arithmetic for command, and before the first command executes in a shell function (peep SHELL GRAMMAR above for that). Go read the description of the extdebug shell option (see shopt above) for the details of how it messes with the DEBUG trap. And if a sigspec is RETURN, the action runs each time a shell function or a script executed with the . or source builtins finishes executing. the way it just knows when things end 🦋

If a sigspec is ERR, the action runs whenever a pipeline (which might be just a single simple command), a list, or a compound command returns a non-zero exit status — but subject to these conditions, bestie, pay attention:

The ERR trap does not run if the failed command is part of the command list immediately following a while or until reserved word
part of the test in an if statement
part of a command executed in a && or || list, except the command following the final && or ||
any command in a pipeline but the last (subject to the state of the pipefail shell option)
or if the command's return value is being inverted using !

These are literally the same conditions obeyed by the errexit (-e) option. siblings fr fr 👑

When the shell is not interactive, signals ignored upon entry to the shell cannot be trapped or reset. Interactive shells do permit trapping signals ignored on entry. Trapped signals that aren't being ignored get reset to their original values in a subshell or subshell environment when one is created. The return status is false if any sigspec is invalid; otherwise trap returns true.

✨ `true` ✨

Does nothing. Returns a 0 status. That's it, that's the whole girl. zero thoughts, zero errors 🎀

✨ `type` ✨

type [-aftpP] name [name ...]

Tells you how each name would be interpreted if you used it as a command name. vibe check for commands 💜

With -t, type prints a string that's one of alias, keyword, function, builtin, or file — depending on whether name is an alias, a shell reserved word, a function, a builtin, or an executable file, respectively. If the name isn't found, type prints nothing and returns a non-zero exit status. silent treatment 😭

With -p, type either returns the pathname of the executable file that would be found by searching $PATH for name, or nothing if type -t name wouldn't return file. The -P option forces a PATH search for each name, even if type -t name wouldn't return file. And if name is present in the table of hashed commands, -p and -P print the hashed value — which is not necessarily the file that appears first in PATH. 👀

With -a, type prints all the places that contain a command named name. That includes aliases, reserved words, functions, and builtins — but you can supply the path search options (-p and -P) to restrict the output to executable files. When you use -a with -p, type does not consult the table of hashed commands, and only performs a PATH search for name.

The -f option suppresses shell function lookup, same as with the command builtin. type returns true if all of the arguments are found, false if any are not found.

✨ `ulimit` ✨

ulimit [-HS] -a
ulimit [-HS] [-bcdefiklmnpqrstuvxPRT [limit]]

Gives you control over the resources available to the shell and to the processes it starts, on systems that allow such control. the resource manager era 🌸

The -H and -S options say whether the hard or soft limit is being set for the given resource. A hard limit cannot be increased by a non-root user once it's set; a soft limit may be increased up to the value of the hard limit. If neither -H nor -S is given, ulimit sets both the soft and hard limits.

The value of limit can be a number in the unit specified for the resource, or one of the special values hard, soft, or unlimited — which stand for the current hard limit, the current soft limit, and no limit, respectively. If limit is omitted, ulimit prints the current value of the soft limit of the resource, unless -H is given. When more than one resource is specified, the limit name and unit (if appropriate) are printed before the value. The other options are interpreted like this:

-a — report all current limits; no limits are set.
-b — the maximum socket buffer size.
-c — the maximum size of core files created.
-d — the maximum size of a process's data segment.
-e — the maximum scheduling priority ("nice").
-f — the maximum size of files written by the shell and its children.
-i — the maximum number of pending signals.
-k — the maximum number of kqueues that may be allocated.
-l — the maximum size that may be locked into memory.
-m — the maximum resident set size (many systems do not honor this limit).
-n — the maximum number of open file descriptors (most systems do not allow this value to be set).
-p — the pipe size in 512-byte blocks (this may not be set).
-q — the maximum number of bytes in POSIX message queues.
-r — the maximum real-time scheduling priority.
-s — the maximum stack size.
-t — the maximum amount of cpu time in seconds.
-u — the maximum number of processes available to a single user.
-v — the maximum amount of virtual memory available to the shell and, on some systems, to its children.
-x — the maximum number of file locks.
-P — the maximum number of pseudoterminals.
-R — the maximum time a real-time process can run before blocking, in microseconds.
-T — the maximum number of threads.

If limit is supplied, and -a isn't used, limit is the new value of the specified resource. If no option is supplied, then -f is assumed. default behavior said what it said 💅

Values are in 1024-byte increments, except for:

-t, which is in seconds
-R, which is in microseconds
-p, which is in units of 512-byte blocks
-P, -T, -b, -k, -n, and -u, which are unscaled values
and, when in posix mode, -c and -f, which are in 512-byte increments

The return status is 0 unless an invalid option or argument is supplied, or an error occurs while setting a new limit.

✨ `umask` ✨

umask [-p] [-S] [mode]

Sets the user file-creation mask to mode. If mode starts with a digit, it's interpreted as an octal number; otherwise it's interpreted as a symbolic mode mask similar to the one chmod(1) accepts. If mode is omitted, umask prints the current value of the mask. The -S option without a mode argument prints the mask in a symbolic format; the default output is an octal number. If -p is supplied and mode is omitted, the output is in a form you can reuse as input. reusable, eco-friendly, iconic 🦋

The return status is zero if the mode was successfully changed or if no mode argument was supplied, and non-zero otherwise.

✨ `unalias` ✨

unalias [-a] [name ...]

Removes each name from the list of defined aliases. If -a is supplied, all alias definitions get removed. The return value is true unless a supplied name isn't a defined alias. bye felicia 👋

✨ `unset` ✨

unset [-fv] [-n] [name ...]

For each name, removes the corresponding variable or function. If -v is given, each name refers to a shell variable and that variable is removed. If -f is specified, each name refers to a shell function and the function definition is removed. If -n is supplied and name is a variable with the nameref attribute, name will be unset rather than the variable it references. -n has no effect if -f is supplied.

Read-only variables and functions may not be unset. don't even try it bestie, it's not happening 💀 When variables or functions are removed, they're also removed from the environment passed to subsequent commands. If no options are supplied, each name refers to a variable; if there's no variable by that name, a function with that name (if any) is unset. Some shell variables may not be unset.

🚨 Important: if any of BASH_ALIASES, BASH_ARGV0, BASH_CMDS, BASH_COMMAND, BASH_SUBSHELL, BASHPID, COMP_WORDBREAKS, DIRSTACK, EPOCHREALTIME, EPOCHSECONDS, FUNCNAME, GROUPS, HISTCMD, LINENO, RANDOM, SECONDS, or SRANDOM are unset, they lose their special properties, even if they are subsequently reset. 😭

The exit status is true unless a name is readonly or may not be unset.

✨ `wait` ✨

wait [-fn] [-p varname] [id ...]

Waits for each specified child process id and returns the termination status of the last id. Each id may be a process ID pid or a job specification jobspec; if a jobspec is supplied, wait waits for all processes in the job. patience is a virtue, apparently 🌈

If no options or *id*s are supplied, wait waits for all running background jobs and the last-executed process substitution (if its process id is the same as $!), and the return status is zero.

Okay so bestie, for wait with -n? If you pass -n, wait waits for any one of the given ids — or, if you didn't give any ids, any job or process substitution — to complete, and then it returns that exit status. And if none of the supplied ids is a child of the shell, or if no ids are supplied and the shell has no unwaited-for children, the exit status is 127. 💀

If you pass -p, wait assigns the process or job identifier of the job for which the exit status is returned to the variable varname named by the option argument. That variable can't be readonly, and it will be unset initially, before any assignment. This is useful only when used with the -n option. 🫶

Supplying the -f option, when job control is enabled, forces wait to wait for each id to terminate before returning its status, instead of returning when it changes status. So like, no jumping the gun. ⏳

If none of the ids specify one of the shell's active child processes, the return status is 127. If wait is interrupted by a signal, any varname will remain unset, and the return status will be greater than 128, as described under SIGNALS above. Otherwise, the return status is the exit status of the last id. Periodt. 💅

✨ SHELL COMPATIBILITY MODE ✨

Bash-4.0 introduced the concept of a shell compatibility level, specified as a set of options to the shopt builtin (compat31, compat32, compat40, compat41, and so on). There is only one current compatibility level — each option is mutually exclusive. The compatibility level is intended to allow users to select behavior from previous versions that is incompatible with newer versions while they migrate scripts to use current features and behavior. It's intended to be a temporary solution, bestie. Not a forever vibe. 🦋

This section does not mention behavior that is standard for a particular version (e.g., setting compat32 means that quoting the right hand side of the regexp matching operator quotes special regexp characters in the word, which is default behavior in bash-3.2 and subsequent versions).

If a user enables, say, compat32, it may affect the behavior of other compatibility levels up to and including the current compatibility level. The idea is that each compatibility level controls behavior that changed in that version of bash, but that behavior may have been present in earlier versions. For instance, the change to use locale-based comparisons with the [[ command came in bash-4.1, and earlier versions used ASCII-based comparisons, so enabling compat32 will enable ASCII-based comparisons as well. That granularity may not be sufficient for all uses, and as a result users should employ compatibility levels carefully. Read the documentation for a particular feature to find out the current behavior. 👑

Bash-4.3 introduced a new shell variable: BASH_COMPAT. The value assigned to this variable (a decimal version number like 4.2, or an integer corresponding to the compatNN option, like 42) determines the compatibility level.

Starting with bash-4.4, bash began deprecating older compatibility levels. Eventually, the options will be removed in favor of BASH_COMPAT.

Bash-5.0 was the final version for which there was an individual shopt option for the previous version. BASH_COMPAT is the only mechanism to control the compatibility level in versions newer than bash-5.0.

The following table describes the behavior changes controlled by each compatibility level setting. The compatNN tag is used as shorthand for setting the compatibility level to NN using one of the following mechanisms. For versions prior to bash-5.0, the compatibility level may be set using the corresponding compatNN shopt option. For bash-4.3 and later versions, the BASH_COMPAT variable is preferred, and it is required for bash-5.1 and later versions.

💖 compat31

Quoting the rhs of the [[ command's regexp matching operator (=~) has no special effect.

💖 compat32

The < and > operators to the [[ command do not consider the current locale when comparing strings; they use ASCII ordering.

💖 compat40

The < and > operators to the [[ command do not consider the current locale when comparing strings; they use ASCII ordering. Bash versions prior to bash-4.1 use ASCII collation and strcmp(3); bash-4.1 and later use the current locale's collation sequence and strcoll(3).

💖 compat41

In posix mode, time may be followed by options and still be recognized as a reserved word (this is POSIX interpretation 267).
In posix mode, the parser requires that an even number of single quotes occur in the word portion of a double-quoted parameter expansion and treats them specially, so that characters within the single quotes are considered quoted (this is POSIX interpretation 221).

💖 compat42

The replacement string in double-quoted pattern substitution does not undergo quote removal, as it does in versions after bash-4.2.
In posix mode, single quotes are considered special when expanding the word portion of a double-quoted parameter expansion and can be used to quote a closing brace or other special character (this is part of POSIX interpretation 221); in later versions, single quotes are not special within double-quoted word expansions.

💖 compat43

Word expansion errors are considered non-fatal errors that cause the current command to fail, even in posix mode (the default behavior is to make them fatal errors that cause the shell to exit).
When executing a shell function, the loop state (while/until/etc.) is not reset, so break or continue in that function will break or continue loops in the calling context. Bash-4.4 and later reset the loop state to prevent this.

💖 compat44

The shell sets up the values used by BASH_ARGV and BASH_ARGC so they can expand to the shell's positional parameters even if extended debugging mode is not enabled.
A subshell inherits loops from its parent context, so break or continue will cause the subshell to exit. Bash-5.0 and later reset the loop state to prevent the exit
Variable assignments preceding builtins like export and readonly that set attributes continue to affect variables with the same name in the calling environment even if the shell is not in posix mode.

💖 compat50

Bash-5.1 changed the way $RANDOM is generated to introduce slightly more randomness. If the shell compatibility level is set to 50 or lower, it reverts to the method from bash-5.0 and previous versions, so seeding the random number generator by assigning a value to RANDOM will produce the same sequence as in bash-5.0.
If the command hash table is empty, bash versions prior to bash-5.1 printed an informational message to that effect, even when producing output that can be reused as input. Bash-5.1 suppresses that message when the -l option is supplied.

💖 compat51

The unset builtin treats attempts to unset array subscripts @ and * differently depending on whether the array is indexed or associative, and differently than in previous versions.
Arithmetic commands ( ((...)) ) and the expressions in an arithmetic for statement can be expanded more than once.
Expressions used as arguments to arithmetic operators in the [[ conditional command can be expanded more than once.
The expressions in substring parameter brace expansion can be expanded more than once.
The expressions in the $((...)) word expansion can be expanded more than once.
Arithmetic expressions used as indexed array subscripts can be expanded more than once.
test -v, when given an argument of A[@], where A is an existing associative array, will return true if the array has any set elements. Bash-5.2 will look for and report on a key named @.
The ${parameter[:]=value} word expansion will return value, before any variable-specific transformations have been performed (e.g., converting to lowercase). Bash-5.2 will return the final value assigned to the variable.
Parsing command substitutions will behave as if extended globbing (see the description of the shopt builtin above) is enabled, so that parsing a command substitution containing an extglob pattern (say, as part of a shell function) will not fail. This assumes the intent is to enable extglob before the command is executed and word expansions are performed. It will fail at word expansion time if extglob hasn't been enabled by the time the command is executed.

💖 compat52

The test builtin uses its historical algorithm to parse parenthesized subexpressions when given five or more arguments.
If the -p or -P option is supplied to the bind builtin, bind treats any arguments remaining after option processing as bindable command names, and displays any key sequences bound to those commands, instead of treating the arguments as key sequences to bind.

🔒 RESTRICTED SHELL 🔒

Okay bestie, so if bash gets started under the name rbash, OR if you hand it the -r option when you invoke it, the shell goes restricted. A restricted shell exists so you can set up an environment that's way more locked-down than your standard shell. It behaves identically to bash except for the stuff below, which is either disallowed or just not performed at all:

Changing directories with cd. 💀
Setting or unsetting the values of SHELL, PATH, HISTFILE, ENV, or BASH_ENV.
Specifying command names containing /.
Specifying a filename containing a / as an argument to the . builtin command.
Using the -p option to the . builtin command to specify a search path.
Specifying a filename containing a slash as an argument to the history builtin command.
Specifying a filename containing a slash as an argument to the -p option to the hash builtin command.
Importing function definitions from the shell environment at startup.
Parsing the values of BASHOPTS and SHELLOPTS from the shell environment at startup.
Redirecting output using the >, >|, <>, >&, &>, and >> redirection operators.
Using the exec builtin command to replace the shell with another command.
Adding or deleting builtin commands with the -f and -d options to the enable builtin command.
Using the enable builtin command to enable disabled shell builtins.
Specifying the -p option to the command builtin command.
Turning off restricted mode with set +r or shopt -u restricted_shell.

Heads up: these restrictions get enforced after any startup files are read. So the startup files still happen, then the cage door slams. 🚪✨

And here's a lil plot twist: when a command that turns out to be a shell script gets executed (peep COMMAND EXECUTION above 👀), rbash turns off any restrictions in the shell that gets spawned to actually run that script. Iconic behaviour honestly.

👀 SEE ALSO 👀

Bash Reference Manual, Brian Fox and Chet Ramey
The Gnu Readline Library, Brian Fox and Chet Ramey
The Gnu History Library, Brian Fox and Chet Ramey
Portable Operating System Interface (POSIX) Part 2: Shell and Utilities, IEEE — http://pubs.opengroup.org/onlinepubs/9799919799/
http://tiswww.case.edu/~chet/bash/POSIX — a description of posix mode
sh(1), ksh(1), csh(1)
emacs(1), vi(1)
readline(3)

📁 FILES 📁

/bin/bash — The bash executable
/etc/profile — The systemwide initialization file, executed for login shells
/etc/bash.bashrc — The systemwide per-interactive-shell startup file
/etc/bash.bash.logout — The systemwide login shell cleanup file, executed when a login shell exits
~/.bash_profile — The personal initialization file, executed for login shells
~/.bashrc — The individual per-interactive-shell startup file
~/.bash_logout — The individual login shell cleanup file, executed when a login shell exits
~/.bash_history — The default value of HISTFILE, the file in which bash saves the command history
~/.inputrc — Individual readline initialization file

👑 AUTHORS 👑

Brian Fox, Free Software Foundation bfox@gnu.org

Chet Ramey, Case Western Reserve University chet.ramey@case.edu

🐛 BUG REPORTS 🐛

So like, if you find a bug in bash, you should actually report it. But bestie — first make sure it really is a bug, and that it shows up in the latest version of bash. The latest version is always available from ftp://ftp.gnu.org/pub/gnu/bash/ and http://git.savannah.gnu.org/cgit/bash.git/snapshot/bash-master.tar.gz.

Once you've confirmed a bug genuinely exists, use the bashbug command to submit a bug report. And if you've got a fix? Please mail that too!! You may also send suggestions and “philosophical” bug reports to bug-bash@gnu.org, or post them to the Usenet newsgroup gnu.bash.bug.

ALL bug reports should include:

The version number of bash
The hardware and operating system
The compiler used to compile
A description of the bug behavior
A short script or “recipe” which exercises the bug

bashbug inserts the first three items automatically into the template it provides for filing a bug report. 🎀

Comments and bug reports concerning this manual page should be directed to chet.ramey@case.edu.

💀 BUGS 💀

It's too big and too slow.
There are some subtle differences between bash and traditional versions of sh, mostly because of the POSIX specification.
Aliases are confusing in some uses.
Shell builtin commands and functions are not stoppable/restartable.
Compound commands and command lists of the form “a ; b ; c” are not handled gracefully when combined with process suspension. When a process is stopped, the shell immediately executes the next command in the list or breaks out of any existing loops. It suffices to enclose the command in parentheses to force it into a subshell, which may be stopped as a unit, or to start the command in the background and immediately bring it into the foreground.
Array variables may not (yet) be exported.

and that's a wrap on the whole entire BASH(1) page, bestie!! 🎉🫶 you read all of it. that's main character energy fr fr. this edition is GNU Bash 5.3, dated 2025 April 7 — so go forth, chmod +x your dreams, and may your exit status always be 0. 💖✨🌈

📦 where this page comes from

ubuntu resolute: bash-static 5.3-2ubuntu1 · GPL-3.0-or-later
ubuntu stonking: bash-static 5.3-3ubuntu1 · GPL-3.0-or-later
💖 girlified by deepseek-ai/DeepSeek-V4.1-Flash on 2026-10-13: a language model's rewrite of the original, under its terms

BASH(1)

👑 BASH(1) 👑

🎀 NAME 🎀

✨ SYNOPSIS ✨

💜 COPYRIGHT 💜

🌸 DESCRIPTION 🌸

💅 OPTIONS 💅

🌈 the multi-character ones 🌈

🦋 ARGUMENTS 🦋

🌟 INVOCATION 🌟

🧚 how bash runs its startup files 🧚

💀 invoked as sh 💀

📏 posix mode startup 📏

🌐 the network connection thing 🌐

🔐 the setuid / setgid situation 🔐

🦋 DEFINITIONS 🦋

👑 RESERVED WORDS 👑

🌈 SHELL GRAMMAR 🌈

Simple Commands

Pipelines

Lists

Compound Commands

🧚 Coprocesses 🧚

🎀 Shell Function Definitions 🎀

💬 COMMENTS 💬

💅 QUOTING 💅

🌍 Translating Strings 🌍

🧸 PARAMETERS 🧸

📍 Positional Parameters 📍

✨ Special Parameters ✨

Shell Variables 🌸

BASH_COMPAT ✨

BASH_ENV 🌸

BASH_XTRACEFD 💜

CDPATH 🧚

CHILD_MAX ⭐

COLUMNS 🌈

COMPREPLY 🍓

EMACS 🦋

ENV 👑

EXECIGNORE 💖

FCEDIT 💜

FIGNORE 🌸

FUNCNEST 🔥

GLOBIGNORE 🎀

GLOBSORT 🌈

Arrays 🦋

✨ EXPANSION ✨

🦋 Brace Expansion 🦋

🌸 Tilde Expansion 🌸

💜 Parameter Expansion 💜

${parameter:-word}

${parameter:=word}

${parameter:?word}

${parameter:+word}

Substring Expansion

${!prefix*} / ${!prefix@}

${!name[@]} / ${!name[*]}

${#parameter}

Remove Matching Prefix Pattern

Remove Matching Suffix Pattern

Pattern Substitution

Case Modification

💫 ${parameter@operator} — parameter transformation

🧚 Command Substitution

🧮 Arithmetic Expansion

🌀 Process Substitution

✂️ Word Splitting

🌟 Pathname Expansion

✨ Pattern Matching ✨

💖 Quote Removal 💖

🎀 REDIRECTION 🎀

🌸 Redirecting Input 🌸

💖 Redirecting Output 💖

⭐ Appending Redirected Output ⭐

🌈 Redirecting Standard Output and Standard Error 🌈

🦋 Appending Standard Output and Standard Error 🦋

🎀 Here Documents 🎀

💜 Here Strings 💜

👑 Duplicating File Descriptors 👑

💀 invoked as `sh` 💀

`BASH_COMPAT` ✨

`BASH_ENV` 🌸

`BASH_XTRACEFD` 💜

`CDPATH` 🧚

`CHILD_MAX` ⭐

`COLUMNS` 🌈

`COMPREPLY` 🍓

`EMACS` 🦋

`ENV` 👑

`EXECIGNORE` 💖

`FCEDIT` 💜

`FIGNORE` 🌸

`FUNCNEST` 🔥

`GLOBIGNORE` 🎀

`GLOBSORT` 🌈

`${parameter:-word}`

`${parameter:=word}`

`${parameter:?word}`

`${parameter:+word}`

`${!prefix*}` / `${!prefix@}`

`${!name[@]}` / `${!name[*]}`

`${#parameter}`

💫 `${parameter@operator}` — parameter transformation

`:` — the do-nothing icon

`.` / `source` — the "run this in MY shell" bestie

`alias` 🌈

`bg` 🦋

`bind` ✨

`break` 💀

`builtin` 👑

`caller` 🧚

`cd` 🌸

🎀 `command` 🎀

🌸 `compgen` 🌸

👑 `complete` 👑

✨ `compopt` ✨

🌸 `continue` 🌸

🎀 `declare` / `typeset` 🎀

🦋 `dirs` 🦋

💅 `disown` 💅

⭐ `echo` ⭐

🧚 `enable` 🧚

`eval`

`exec`

`exit`

`export`

`false`

`fc`

`fg`

`getopts`

`hash`

✨ `help` ✨

✨ `history` ✨

✨ `jobs` ✨

✨ `kill` ✨