BASH(1)\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 General Commands Manual\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 BASH(1)<\/p>\n
NAME<\/p>\n
bash – GNU Bourne-Again SHell<\/p>\n
SYNOPSIS<\/p>\n
bash [options] [command_string | file]<\/p>\n
COPYRIGHT<\/p>\n
Bash is Copyright (C) 1989-2013 by the Free Software Foundation, Inc.<\/p>\n
DESCRIPTION<\/p>\n
Bash is an sh-compatible command language interpreter that executes
\ncommands read from the standard input or from a file.\u00a0 Bash also
\nincorporates useful features from the Korn and C shells (ksh and
\ncsh).<\/p>\n
Bash is intended to be a conformant implementation of the Shell and
\nUtilities portion of the IEEE POSIX specification (IEEE Standard
\n1003.1).\u00a0 Bash can be configured to be POSIX-conformant by default.<\/p>\n
OPTIONS<\/p>\n
All of the\u00a0 single-character shell options documented in the
\ndescription of the set builtin command can be used as options when
\nthe shell is invoked.\u00a0 In addition, bash interprets the following
\noptions when it is invoked:<\/p>\n
-c\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 If the -c option is present, then commands are read from
\nthe first non-option argument command_string.\u00a0 If there are
\narguments after the command_string, they are assigned to
\nthe positional parameters, starting with $0.
\n-i\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 If the -i option is present, the shell is interactive.
\n-l\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Make bash act as if it had been invoked as a login shell
\n(see INVOCATION below).
\n-r\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 If the -r option is present, the shell becomes restricted
\n(see RESTRICTED SHELL below).
\n-s\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 If the -s option is present, or if no arguments remain
\nafter option processing, then commands are read from the
\nstandard input.\u00a0 This option allows the positional
\nparameters to be set when invoking an interactive shell.
\n-D\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 A list of all double-quoted strings preceded by $ is
\nprinted on the standard output.\u00a0 These are the strings that
\nare subject to language translation when the current locale
\nis not C or POSIX.\u00a0 This implies the -n option; no commands
\nwill be executed.
\n[-+]O [shopt_option]
\nshopt_option is one of the shell options accepted by the
\nshopt builtin (see SHELL BUILTIN COMMANDS below).\u00a0 If
\nshopt_option is present, -O sets the value of that option;
\n+O unsets it.\u00a0 If shopt_option is not supplied, the names
\nand values of the shell options accepted by shopt are
\nprinted on the standard output.\u00a0 If the invocation option
\nis +O, the output is displayed in a format that may be
\nreused as input.
\n—\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 A — signals the end of options and disables further option
\nprocessing.\u00a0 Any arguments after the — are treated as
\nfilenames and arguments.\u00a0 An argument of – is equivalent to
\n–.<\/p>\n
Bash also interprets a number of multi-character options.\u00a0 These
\noptions must appear on the command line before the single-character
\noptions to be recognized.<\/p>\n
–debugger
\nArrange for the debugger profile to be executed before the
\nshell starts.\u00a0 Turns on extended debugging mode (see the
\ndescription of the extdebug option to the shopt builtin
\nbelow).
\n–dump-po-strings
\nEquivalent to -D, but the output is in the GNU gettext po
\n(portable object) file format.
\n–dump-strings
\nEquivalent to -D.
\n–help Display a usage message on standard output and exit
\nsuccessfully.
\n–init-file file
\n–rcfile file
\nExecute commands from file instead of the standard personal
\ninitialization file ~\/.bashrc if the shell is interactive (see
\nINVOCATION below).<\/p>\n
–login
\nEquivalent to -l.<\/p>\n
–noediting
\nDo not use the GNU readline library to read command lines when
\nthe shell is interactive.<\/p>\n
–noprofile
\nDo not read either the system-wide startup file \/etc\/profile
\nor any of the personal initialization files ~\/.bash_profile,
\n~\/.bash_login, or ~\/.profile.\u00a0 By default, bash reads these
\nfiles when it is invoked as a login shell (see INVOCATION
\nbelow).<\/p>\n
–norc Do not read and execute the personal initialization file
\n~\/.bashrc if the shell is interactive.\u00a0 This option is on by
\ndefault if the shell is invoked as sh.<\/p>\n
–posix
\nChange the behavior of bash where the default operation
\ndiffers from the POSIX standard to match the standard (posix
\nmode).\u00a0 See SEE ALSO below for a reference to a document that
\ndetails how posix mode affects bash’s behavior.<\/p>\n
–restricted
\nThe shell becomes restricted (see RESTRICTED SHELL below).<\/p>\n
–verbose
\nEquivalent to\u00a0 -v.<\/p>\n
–version
\nShow version information for this instance of bash on the
\nstandard output and exit successfully.<\/p>\n
ARGUMENTS<\/p>\n
If arguments remain after option processing, and neither the -c nor
\nthe -s option has been supplied, the first argument is assumed to be
\nthe name of a file containing shell commands.\u00a0 If bash is invoked in
\nthis fashion, $0 is set to the name of the file, and the positional
\nparameters are set to the remaining arguments.\u00a0 Bash reads and
\nexecutes commands from this file, then exits.\u00a0 Bash’s exit status is
\nthe exit status of the last command executed in the script.\u00a0 If no
\ncommands are executed, the exit status is 0.\u00a0 An attempt is first
\nmade to open the file in the current directory, and, if no file is
\nfound, then the shell searches the directories in PATH for the
\nscript.<\/p>\n
INVOCATION<\/p>\n
A login shell is one whose first character of argument zero is a -,
\nor one started with the –login option.<\/p>\n
An interactive shell is one started without non-option arguments and
\nwithout the -c option whose standard input and error are both
\nconnected to terminals (as determined by isatty(3)), or one started
\nwith the -i option.\u00a0 PS1 is set and $- includes i if bash is
\ninteractive, allowing a shell script or a startup file to test this
\nstate.<\/p>\n
The following paragraphs describe how bash executes its startup
\nfiles.\u00a0 If any of the files exist but cannot be read, bash reports an
\nerror.\u00a0 Tildes are expanded in filenames as described below under
\nTilde Expansion in the EXPANSION section.<\/p>\n
When bash is invoked as an interactive login shell, or as a non-
\ninteractive shell with the –login option, it first reads and
\nexecutes commands from the file \/etc\/profile, if that file exists.
\nAfter reading that file, it looks for ~\/.bash_profile, ~\/.bash_login,
\nand ~\/.profile, in that order, and reads and executes commands from
\nthe first one that exists and is readable.\u00a0 The –noprofile option
\nmay be used when the shell is started to inhibit this behavior.<\/p>\n
When a login shell exits, bash reads and executes commands from the
\nfile ~\/.bash_logout, if it exists.<\/p>\n
When an interactive shell that is not a login shell is started, bash
\nreads and executes commands from ~\/.bashrc, if that file exists.
\nThis may be inhibited by using the –norc option.\u00a0 The –rcfile file
\noption will force bash to read and execute commands from file instead
\nof ~\/.bashrc.<\/p>\n
When bash is started non-interactively, to run a shell script, for
\nexample, it looks for the variable BASH_ENV in the environment,
\nexpands its value if it appears there, and uses the expanded value as
\nthe name of a file to read and execute.\u00a0 Bash behaves as if the
\nfollowing command were executed:
\nif [ -n “$BASH_ENV” ]; then . “$BASH_ENV”; fi
\nbut the value of the PATH variable is not used to search for the
\nfilename.<\/p>\n
If bash is invoked with the name sh, it tries to mimic the startup
\nbehavior of historical versions of sh as closely as possible, while
\nconforming to the POSIX standard as well.\u00a0 When invoked as an
\ninteractive login shell, or a non-interactive shell with the –login
\noption, it first attempts to read and execute commands from
\n\/etc\/profile and ~\/.profile, in that order.\u00a0 The –noprofile option
\nmay be used to inhibit this behavior.\u00a0 When invoked as an interactive
\nshell with the name sh, bash looks for the variable ENV, expands its
\nvalue if it is defined, and uses the expanded value as the name of a
\nfile to read and execute.\u00a0 Since a shell invoked as sh does not
\nattempt to read and execute commands from any other startup files,
\nthe –rcfile option has no effect.\u00a0 A non-interactive shell invoked
\nwith the name sh does not attempt to read any other startup files.
\nWhen invoked as sh, bash enters posix mode after the startup files
\nare read.<\/p>\n
When bash is started in posix mode, as with the –posix command line
\noption, it follows the POSIX standard for startup files.\u00a0 In this
\nmode, interactive shells expand the ENV variable and commands are
\nread and executed from the file whose name is the expanded value.\u00a0 No
\nother startup files are read.<\/p>\n
Bash attempts to determine when it is being run with its standard
\ninput connected to a network connection, as when executed by the
\nremote shell daemon, usually rshd, or the secure shell daemon sshd.
\nIf bash determines it is being run in this fashion, it reads and
\nexecutes commands from ~\/.bashrc, if that file exists and is
\nreadable.\u00a0 It will not do this if invoked as sh.\u00a0 The –norc option
\nmay be used to inhibit this behavior, and the –rcfile option may be
\nused to force another file to be read, but neither rshd nor sshd
\ngenerally invoke the shell with those options or allow them to be
\nspecified.<\/p>\n
If the shell is started with the effective user (group) id not equal
\nto the real user (group) id, and the -p option is not supplied, no
\nstartup files are read, shell functions are not inherited from the
\nenvironment, the SHELLOPTS, BASHOPTS, CDPATH, and GLOBIGNORE
\nvariables, if they appear in the environment, are ignored, and the
\neffective user id is set to the real user id.\u00a0 If the -p option is
\nsupplied at invocation, the startup behavior is the same, but the
\neffective user id is not reset.<\/p>\n
DEFINITIONS<\/p>\n
The following definitions are used throughout the rest of this
\ndocument.
\nblank\u00a0 A space or tab.
\nword\u00a0\u00a0 A sequence of characters considered as a single unit by the
\nshell.\u00a0 Also known as a token.
\nname\u00a0\u00a0 A word consisting only of alphanumeric characters and
\nunderscores, and beginning with an alphabetic character or an
\nunderscore.\u00a0 Also referred to as an identifier.
\nmetacharacter
\nA character that, when unquoted, separates words.\u00a0 One of the
\nfollowing:
\n|\u00a0 & ; ( ) < > space tab
\ncontrol operator
\nA token that performs a control function.\u00a0 It is one of the
\nfollowing symbols:
\n|| & && ; ;; ( ) | |& <newline><\/p>\n
RESERVED WORDS<\/p>\n
Reserved words are words that have a special meaning to the shell.
\nThe following words are recognized as reserved when unquoted and
\neither the first word of a simple command (see SHELL GRAMMAR below)
\nor the third word of a case or for command:<\/p>\n
! case\u00a0 coproc\u00a0 do done elif else esac fi for function if in select
\nthen until while { } time [[ ]]<\/p>\n
SHELL GRAMMAR<\/p>\n
Simple Commands
\nA simple command is a sequence of optional variable assignments
\nfollowed by blank-separated words and redirections, and terminated by
\na control operator.\u00a0 The first word specifies the command to be
\nexecuted, and is passed as argument zero.\u00a0 The remaining words are
\npassed as arguments to the invoked command.<\/p>\n
The return value of a simple command is its exit status, or 128+n if
\nthe command is terminated by signal n.<\/p>\n
Pipelines
\nA pipeline is a sequence of one or more commands separated by one of
\nthe control operators | or |&.\u00a0 The format for a pipeline is:<\/p>\n
[time [-p]] [ ! ] command [ [|\u23aa|&] command2 … ]<\/p>\n
The standard output of command is connected via a pipe to the
\nstandard input of command2.\u00a0 This connection is performed before any
\nredirections specified by the command (see REDIRECTION below).\u00a0 If |&
\nis used, command’s standard error, in addition to its standard
\noutput, is connected to command2’s standard input through the pipe;
\nit is shorthand for 2>&1 |.\u00a0 This implicit redirection of the
\nstandard error to the standard output is performed after any
\nredirections specified by the command.<\/p>\n
The return status of a pipeline is the exit status of the last
\ncommand, unless the pipefail option is enabled.\u00a0 If pipefail is
\nenabled, the pipeline’s return status is the value of the last
\n(rightmost) command to exit with a non-zero status, or zero if all
\ncommands exit successfully.\u00a0 If the reserved word !\u00a0 precedes a
\npipeline, the exit status of that pipeline is the logical negation of
\nthe exit status as described above.\u00a0 The shell waits for all commands
\nin the pipeline to terminate before returning a value.<\/p>\n
If the time reserved word precedes a pipeline, the elapsed as well as
\nuser and system time consumed by its execution are reported when the
\npipeline terminates.\u00a0 The -p option changes the output format to that
\nspecified by POSIX.\u00a0 When the shell is in posix mode, it does not
\nrecognize time as a reserved word if the next token begins with a
\n`-‘.\u00a0 The TIMEFORMAT variable may be set to a format string that
\nspecifies how the timing information should be displayed; see the
\ndescription of TIMEFORMAT under Shell Variables below.<\/p>\n
When the shell is in posix mode, time may be followed by a newline.
\nIn this case, the shell displays the total user and system time
\nconsumed by the shell and its children.\u00a0 The TIMEFORMAT variable may
\nbe used to specify the format of the time information.<\/p>\n
Each command in a pipeline is executed as a separate process (i.e.,
\nin a subshell).<\/p>\n
Lists
\nA list is a sequence of one or more pipelines separated by one of the
\noperators ;, &, &&, or ||, and optionally terminated by one of ;, &,
\nor <newline>.<\/p>\n
Of these list operators, && and || have equal precedence, followed by
\n; and &, which have equal precedence.<\/p>\n
A sequence of one or more newlines may appear in a list instead of a
\nsemicolon to delimit commands.<\/p>\n
If a command is terminated by the control operator &, the shell
\nexecutes the command in the background in a subshell.\u00a0 The shell does
\nnot wait for the command to finish, and the return status is 0.
\nCommands separated by a ; are executed sequentially; the shell waits
\nfor each command to terminate in turn.\u00a0 The return status is the exit
\nstatus of the last command executed.<\/p>\n
AND and OR lists are sequences of one of more pipelines separated by
\nthe && and || control operators, respectively.\u00a0 AND and OR lists are
\nexecuted with left associativity.\u00a0 An AND list has the form<\/p>\n
command1 && command2<\/p>\n
command2 is executed if, and only if, command1 returns an exit status
\nof zero.<\/p>\n
An OR list has the form<\/p>\n
command1 || command2<\/p>\n
command2 is executed if and only if command1 returns a non-zero exit
\nstatus.\u00a0 The return status of AND and OR lists is the exit status of
\nthe last command executed in the list.<\/p>\n
Compound Commands
\nA compound command is one of the following.\u00a0 In most cases a list in
\na command’s description may be separated from the rest of the command
\nby one or more newlines, and may be followed by a newline in place of
\na semicolon.<\/p>\n
(list) list is executed in a subshell environment (see COMMAND
\nEXECUTION ENVIRONMENT below).\u00a0 Variable assignments and
\nbuiltin commands that affect the shell’s environment do not
\nremain in effect after the command completes.\u00a0 The return
\nstatus is the exit status of list.<\/p>\n
{ list; }
\nlist is simply executed in the current shell environment.
\nlist must be terminated with a newline or semicolon.\u00a0 This is
\nknown as a group command.\u00a0 The return status is the exit
\nstatus of list.\u00a0 Note that unlike the metacharacters ( and ),
\n{ and } are reserved words and must occur where a reserved
\nword is permitted to be recognized.\u00a0 Since they do not cause a
\nword break, they must be separated from list by whitespace or
\nanother shell metacharacter.<\/p>\n
((expression))
\nThe expression is evaluated according to the rules described
\nbelow under ARITHMETIC EVALUATION.\u00a0 If the value of the
\nexpression is non-zero, the return status is 0; otherwise the
\nreturn status is 1.\u00a0 This is exactly equivalent to let
\n“expression”.<\/p>\n
[[ expression ]]
\nReturn a status of 0 or 1 depending on the evaluation of the
\nconditional expression expression.\u00a0 Expressions are composed
\nof the primaries described below under CONDITIONAL
\nEXPRESSIONS.\u00a0 Word splitting and pathname expansion are not
\nperformed on the words between the [[ and ]]; tilde expansion,
\nparameter and variable expansion, arithmetic expansion,
\ncommand substitution, process substitution, and quote removal
\nare performed.\u00a0 Conditional operators such as -f must be
\nunquoted to be recognized as primaries.<\/p>\n
When used with [[, the < and > operators sort
\nlexicographically using the current locale.<\/p>\n
When the == and != operators are used, the string to the right
\nof the operator is considered a pattern and matched according
\nto the rules described below under Pattern Matching, as if the
\nextglob shell option were enabled.\u00a0 The = operator is
\nequivalent to ==.\u00a0 If the shell option nocasematch is enabled,
\nthe match is performed without regard to the case of
\nalphabetic characters.\u00a0 The return value is 0 if the string
\nmatches (==) or does not match (!=) the pattern, and 1
\notherwise.\u00a0 Any part of the pattern may be quoted to force the
\nquoted portion to be matched as a string.<\/p>\n
An additional binary operator, =~, is available, with the same
\nprecedence as == and !=.\u00a0 When it is used, the string to the
\nright of the operator is considered an extended regular
\nexpression and matched accordingly (as in regex(3)).\u00a0 The
\nreturn value is 0 if the string matches the pattern, and 1
\notherwise.\u00a0 If the regular expression is syntactically
\nincorrect, the conditional expression’s return value is 2.\u00a0 If
\nthe shell option nocasematch is enabled, the match is
\nperformed without regard to the case of alphabetic characters.
\nAny part of the pattern may be quoted to force the quoted
\nportion to be matched as a string.\u00a0 Bracket expressions in
\nregular expressions must be treated carefully, since normal
\nquoting characters lose their meanings between brackets.\u00a0 If
\nthe pattern is stored in a shell variable, quoting the
\nvariable expansion forces the entire pattern to be matched as
\na string.\u00a0 Substrings matched by parenthesized subexpressions
\nwithin the regular expression are saved in the array variable
\nBASH_REMATCH.\u00a0 The element of BASH_REMATCH with index 0 is the
\nportion of the string matching the entire regular expression.
\nThe element of BASH_REMATCH with index n is the portion of the
\nstring matching the nth parenthesized subexpression.<\/p>\n
Expressions may be combined using the following operators,
\nlisted in decreasing order of precedence:<\/p>\n
( expression )
\nReturns the value of expression.\u00a0 This may be used to
\noverride the normal precedence of operators.
\n! expression
\nTrue if expression is false.
\nexpression1 && expression2
\nTrue if both expression1 and expression2 are true.
\nexpression1 || expression2
\nTrue if either expression1 or expression2 is true.<\/p>\n
The && and || operators do not evaluate expression2 if the
\nvalue of expression1 is sufficient to determine the return
\nvalue of the entire conditional expression.<\/p>\n
for name [ [ in [ word … ] ] ; ] do list ; done
\nThe list of words following in is expanded, generating a list
\nof items.\u00a0 The variable name is set to each element of this
\nlist in turn, and list is executed each time.\u00a0 If the in word
\nis omitted, the for command executes list once for each
\npositional parameter that is set (see PARAMETERS below).\u00a0 The
\nreturn status is the exit status of the last command that
\nexecutes.\u00a0 If the expansion of the items following in results
\nin an empty list, no commands are executed, and the return
\nstatus is 0.<\/p>\n
for (( expr1 ; expr2 ; expr3 )) ; do list ; done
\nFirst, the arithmetic expression expr1 is evaluated according
\nto the rules described below under ARITHMETIC EVALUATION.\u00a0 The
\narithmetic expression expr2 is then evaluated repeatedly until
\nit evaluates to zero.\u00a0 Each time expr2 evaluates to a non-zero
\nvalue, list is executed and the arithmetic expression expr3 is
\nevaluated.\u00a0 If any expression is omitted, it behaves as if it
\nevaluates to 1.\u00a0 The return value is the exit status of the
\nlast command in list that is executed, or false if any of the
\nexpressions is invalid.<\/p>\n
select name [ in word ] ; do list ; done
\nThe list of words following in is expanded, generating a list
\nof items.\u00a0 The set of expanded words is printed on the
\nstandard error, each preceded by a number.\u00a0 If the in word is
\nomitted, the positional parameters are printed (see PARAMETERS
\nbelow).\u00a0 The PS3 prompt is then displayed and a line read from
\nthe standard input.\u00a0 If the line consists of a number
\ncorresponding to one of the displayed words, then the value of
\nname is set to that word.\u00a0 If the line is empty, the words and
\nprompt are displayed again.\u00a0 If EOF is read, the command
\ncompletes.\u00a0 Any other value read causes name to be set to
\nnull.\u00a0 The line read is saved in the variable REPLY.\u00a0 The list
\nis executed after each selection until a break command is
\nexecuted.\u00a0 The exit status of select is the exit status of the
\nlast command executed in list, or zero if no commands were
\nexecuted.<\/p>\n
case word in [ [(] pattern [ | pattern ] … ) list ;; ] … esac
\nA case command first expands word, and tries to match it
\nagainst each pattern in turn, using the same matching rules as
\nfor pathname expansion (see Pathname Expansion below).\u00a0 The
\nword is expanded using tilde expansion, parameter and variable
\nexpansion, arithmetic substitution, command substitution,
\nprocess substitution and quote removal.\u00a0 Each pattern examined
\nis expanded using tilde expansion, parameter and variable
\nexpansion, arithmetic substitution, command substitution, and
\nprocess substitution.\u00a0 If the shell option nocasematch is
\nenabled, the match is performed without regard to the case of
\nalphabetic characters.\u00a0 When a match is found, the
\ncorresponding list is executed.\u00a0 If the ;; operator is used,
\nno subsequent matches are attempted after the first pattern
\nmatch.\u00a0 Using ;& in place of ;; causes execution to continue
\nwith the list associated with the next set of patterns.\u00a0 Using
\n;;& in place of ;; causes the shell to test the next pattern
\nlist in the statement, if any, and execute any associated list
\non a successful match.\u00a0 The exit status is zero if no pattern
\nmatches.\u00a0 Otherwise, it is the exit status of the last command
\nexecuted in list.<\/p>\n
if list; then list; [ elif list; then list; ] … [ else list; ] fi
\nThe if list is executed.\u00a0 If its exit status is zero, the then
\nlist is executed.\u00a0 Otherwise, each elif list is executed in
\nturn, and if its exit status is zero, the corresponding then
\nlist is executed and the command completes.\u00a0 Otherwise, the
\nelse list is executed, if present.\u00a0 The exit status is the
\nexit status of the last command executed, or zero if no
\ncondition tested true.<\/p>\n
while list-1; do list-2; done
\nuntil list-1; do list-2; done
\nThe while command continuously executes the list list-2 as
\nlong as the last command in the list list-1 returns an exit
\nstatus of zero.\u00a0 The until command is identical to the while
\ncommand, except that the test is negated; list-2 is executed
\nas long as the last command in list-1 returns a non-zero exit
\nstatus.\u00a0 The exit status of the while and until commands is
\nthe exit status of the last command executed in list-2, or
\nzero if none was executed.<\/p>\n
Coprocesses
\nA coprocess is a shell command preceded by the coproc reserved word.
\nA coprocess is executed asynchronously in a subshell, as if the
\ncommand had been terminated with the & control operator, with a two-
\nway pipe established between the executing shell and the coprocess.<\/p>\n
The format for a coprocess is:<\/p>\n
coproc [NAME] command [redirections]<\/p>\n
This creates a coprocess named NAME.\u00a0 If NAME is not supplied, the
\ndefault name is COPROC.\u00a0 NAME must not be supplied if command is a
\nsimple command (see above); otherwise, it is interpreted as the first
\nword of the simple command.\u00a0 When the coprocess is executed, the
\nshell creates an array variable (see Arrays below) named NAME in the
\ncontext of the executing shell.\u00a0 The standard output of command is
\nconnected via a pipe to a file descriptor in the executing shell, and
\nthat file descriptor is assigned to NAME[0].\u00a0 The standard input of
\ncommand is connected via a pipe to a file descriptor in the executing
\nshell, and that file descriptor is assigned to NAME[1].\u00a0 This pipe is
\nestablished before any redirections specified by the command (see
\nREDIRECTION below).\u00a0 The file descriptors can be utilized as
\narguments to shell commands and redirections using standard word
\nexpansions.\u00a0 The file descriptors are not available in subshells.
\nThe process ID of the shell spawned to execute the coprocess is
\navailable as the value of the variable NAME_PID.\u00a0 The wait builtin
\ncommand may be used to wait for the coprocess to terminate.<\/p>\n
Since the coprocess is created as an asynchronous command, the coproc
\ncommand always returns success.\u00a0 The return status of a coprocess is
\nthe exit status of command.<\/p>\n
Shell Function Definitions
\nA shell function is an object that is called like a simple command
\nand executes a compound command with a new set of positional
\nparameters.\u00a0 Shell functions are declared as follows:<\/p>\n
name () compound-command [redirection]
\nfunction name [()] compound-command [redirection]
\nThis defines a function named name.\u00a0 The reserved word
\nfunction is optional.\u00a0 If the function reserved word is
\nsupplied, the parentheses are optional.\u00a0 The body of the
\nfunction is the compound command compound-command (see
\nCompound Commands above).\u00a0 That command is usually a list of
\ncommands between { and }, but may be any command listed under
\nCompound Commands above.\u00a0 compound-command is executed
\nwhenever name is specified as the name of a simple command.
\nWhen in posix mode, name may not be the name of one of the
\nPOSIX special builtins.\u00a0 Any redirections (see REDIRECTION
\nbelow) specified when a function is defined are performed when
\nthe function is executed.\u00a0 The exit status of a function
\ndefinition is zero unless a syntax error occurs or a readonly
\nfunction with the same name already exists.\u00a0 When executed,
\nthe exit status of a function is the exit status of the last
\ncommand executed in the body.\u00a0 (See FUNCTIONS below.)<\/p>\n
COMMENTS<\/p>\n
In a non-interactive shell, or an interactive shell in which the
\ninteractive_comments option to the shopt builtin is enabled (see
\nSHELL BUILTIN COMMANDS below), a word beginning with # causes that
\nword and all remaining characters on that line to be ignored.\u00a0 An
\ninteractive shell without the interactive_comments option enabled
\ndoes not allow comments.\u00a0 The interactive_comments option is on by
\ndefault in interactive shells.<\/p>\n
QUOTING<\/p>\n
Quoting is used to remove the special meaning of certain characters
\nor words to the shell.\u00a0 Quoting can be used to disable special
\ntreatment for special characters, to prevent reserved words from
\nbeing recognized as such, and to prevent parameter expansion.<\/p>\n
Each of the metacharacters listed above under DEFINITIONS has special
\nmeaning to the shell and must be quoted if it is to represent itself.<\/p>\n
When the command history expansion facilities are being used (see
\nHISTORY EXPANSION below), the history expansion character, usually !,
\nmust be quoted to prevent history expansion.<\/p>\n
There are three quoting mechanisms: the escape character, single
\nquotes, and double quotes.<\/p>\n
A non-quoted backslash (\\) is the escape character.\u00a0 It preserves the
\nliteral value of the next character that follows, with the exception
\nof <newline>.\u00a0 If a \\<newline> pair appears, and the backslash is not
\nitself quoted, the \\<newline> is treated as a line continuation (that
\nis, it is removed from the input stream and effectively ignored).<\/p>\n
Enclosing characters in single quotes preserves the literal value of
\neach character within the quotes.\u00a0 A single quote may not occur
\nbetween single quotes, even when preceded by a backslash.<\/p>\n
Enclosing characters in double quotes preserves the literal value of
\nall characters within the quotes, with the exception of $, `, \\, and,
\nwhen history expansion is enabled, !.\u00a0 The characters $ and ` retain
\ntheir special meaning within double quotes.\u00a0 The backslash retains
\nits special meaning only when followed by one of the following
\ncharacters: $, `, “, \\, or <newline>.\u00a0 A double quote may be quoted
\nwithin double quotes by preceding it with a backslash.\u00a0 If enabled,
\nhistory expansion will be performed unless an !\u00a0 appearing in double
\nquotes is escaped using a backslash.\u00a0 The backslash preceding the !
\nis not removed.<\/p>\n
The special parameters * and @ have special meaning when in double
\nquotes (see PARAMETERS below).<\/p>\n
Words of the form $’string’ are treated specially.\u00a0 The word expands
\nto string, with backslash-escaped characters replaced as specified by
\nthe ANSI C standard.\u00a0 Backslash escape sequences, if present, are
\ndecoded as follows:
\n\\a\u00a0\u00a0\u00a0\u00a0 alert (bell)
\n\\b\u00a0\u00a0\u00a0\u00a0 backspace
\n\\e
\n\\E\u00a0\u00a0\u00a0\u00a0 an escape character
\n\\f\u00a0\u00a0\u00a0\u00a0 form feed
\n\\n\u00a0\u00a0\u00a0\u00a0 new line
\n\\r\u00a0\u00a0\u00a0\u00a0 carriage return
\n\\t\u00a0\u00a0\u00a0\u00a0 horizontal tab
\n\\v\u00a0\u00a0\u00a0\u00a0 vertical tab
\n\\\\\u00a0\u00a0\u00a0\u00a0 backslash
\n\\’\u00a0\u00a0\u00a0\u00a0 single quote
\n\\”\u00a0\u00a0\u00a0\u00a0 double quote
\n\\nnn\u00a0\u00a0 the eight-bit character whose value is the octal value
\nnnn (one to three digits)
\n\\xHH\u00a0\u00a0 the eight-bit character whose value is the hexadecimal
\nvalue HH (one or two hex digits)
\n\\uHHHH the Unicode (ISO\/IEC 10646) character whose value is
\nthe hexadecimal value HHHH (one to four hex digits)
\n\\UHHHHHHHH
\nthe Unicode (ISO\/IEC 10646) character whose value is
\nthe hexadecimal value HHHHHHHH (one to eight hex
\ndigits)
\n\\cx\u00a0\u00a0\u00a0 a control-x character<\/p>\n
The expanded result is single-quoted, as if the dollar sign had not
\nbeen present.<\/p>\n
A double-quoted string preceded by a dollar sign ($”string”) will
\ncause the string to be translated according to the current locale.
\nIf the current locale is C or POSIX, the dollar sign is ignored.\u00a0 If
\nthe string is translated and replaced, the replacement is double-
\nquoted.<\/p>\n
PARAMETERS<\/p>\n
A parameter is an entity that stores values.\u00a0 It can be a name, a
\nnumber, or one of the special characters listed below under Special
\nParameters.\u00a0 A variable is a parameter denoted by a name.\u00a0 A variable
\nhas a value and zero or more attributes.\u00a0 Attributes are assigned
\nusing the declare builtin command (see declare below in SHELL BUILTIN
\nCOMMANDS).<\/p>\n
A parameter is set if it has been assigned a value.\u00a0 The null string
\nis a valid value.\u00a0 Once a variable is set, it may be unset only by
\nusing the unset builtin command (see SHELL BUILTIN COMMANDS below).<\/p>\n
A variable may be assigned to by a statement of the form<\/p>\n
name=[value]<\/p>\n
If value is not given, the variable is assigned the null string.\u00a0 All
\nvalues undergo tilde expansion, parameter and variable expansion,
\ncommand substitution, arithmetic expansion, and quote removal (see
\nEXPANSION below).\u00a0 If the variable has its integer attribute set,
\nthen value is evaluated as an arithmetic expression even if the
\n$((…)) expansion is not used (see Arithmetic Expansion below).
\nWord splitting is not performed, with the exception of “$@” as
\nexplained below under Special Parameters.\u00a0 Pathname expansion is not
\nperformed.\u00a0 Assignment statements may also appear as arguments to the
\nalias, declare, typeset, export, readonly, and local builtin
\ncommands.\u00a0 When in posix mode, these builtins may appear in a command
\nafter one or more instances of the command builtin and retain these
\nassignment statement properties.<\/p>\n
In the context where an assignment statement is assigning a value to
\na shell variable or array index, the += operator can be used to
\nappend to or add to the variable’s previous value.\u00a0 When += is
\napplied to a variable for which the integer attribute has been set,
\nvalue is evaluated as an arithmetic expression and added to the
\nvariable’s current value, which is also evaluated.\u00a0 When += is
\napplied to an array variable using compound assignment (see Arrays
\nbelow), the variable’s value is not unset (as it is when using =),
\nand new values are appended to the array beginning at one greater
\nthan the array’s maximum index (for indexed arrays) or added as
\nadditional key-value pairs in an associative array.\u00a0 When applied to
\na string-valued variable, value is expanded and appended to the
\nvariable’s value.<\/p>\n
A variable can be assigned the nameref attribute using the -n option
\nto the declare or local builtin commands (see the descriptions of
\ndeclare and local below) to create a nameref, or a reference to
\nanother variable.\u00a0 This allows variables to be manipulated
\nindirectly.\u00a0 Whenever the nameref variable is referenced or assigned
\nto, the operation is actually performed on the variable specified by
\nthe nameref variable’s value.\u00a0 A nameref is commonly used within
\nshell functions to refer to a variable whose name is passed as an
\nargument to the function.\u00a0 For instance, if a variable name is passed
\nto a shell function as its first argument, running
\ndeclare -n ref=$1
\ninside the function creates a nameref variable ref whose value is the
\nvariable name passed as the first argument.\u00a0 References and
\nassignments to ref are treated as references and assignments to the
\nvariable whose name was passed as $1.\u00a0 If the control variable in a
\nfor loop has the nameref attribute, the list of words can be a list
\nof shell variables, and a name reference will be established for each
\nword in the list, in turn, when the loop is executed.\u00a0 Array
\nvariables cannot be given the -n attribute.\u00a0 However, nameref
\nvariables can reference array variables and subscripted array
\nvariables.\u00a0 Namerefs can be unset using the -n option to the unset
\nbuiltin.\u00a0 Otherwise, if unset is executed with the name of a nameref
\nvariable as an argument, the variable referenced by the nameref
\nvariable will be unset.<\/p>\n
Positional Parameters
\nA positional parameter is a parameter denoted by one or more digits,
\nother than the single digit 0.\u00a0 Positional parameters are assigned
\nfrom the shell’s arguments when it is invoked, and may be reassigned
\nusing the set builtin command.\u00a0 Positional parameters may not be
\nassigned to with assignment statements.\u00a0 The positional parameters
\nare temporarily replaced when a shell function is executed (see
\nFUNCTIONS below).<\/p>\n
When a positional parameter consisting of more than a single digit is
\nexpanded, it must be enclosed in braces (see EXPANSION below).<\/p>\n
Special Parameters
\nThe shell treats several parameters specially.\u00a0 These parameters may
\nonly be referenced; assignment to them is not allowed.
\n*\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the positional parameters, starting from one.\u00a0 When
\nthe expansion is not within double quotes, each positional
\nparameter expands to a separate word.\u00a0 In contexts where it is
\nperformed, those words are subject to further word splitting
\nand pathname expansion.\u00a0 When the expansion occurs within
\ndouble quotes, it expands to a single word with the value of
\neach parameter separated by the first character of the IFS
\nspecial variable.\u00a0 That is, “$*” is equivalent to “$1c$2c…”,
\nwhere c is the first character of the value of the IFS
\nvariable.\u00a0 If IFS is unset, the parameters are separated by
\nspaces.\u00a0 If IFS is null, the parameters are joined without
\nintervening separators.
\n@\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the positional parameters, starting from one.\u00a0 When
\nthe expansion occurs within double quotes, each parameter
\nexpands to a separate word.\u00a0 That is, “$@” is equivalent to
\n“$1” “$2” …\u00a0 If the double-quoted expansion occurs within a
\nword, the expansion of the first parameter is joined with the
\nbeginning part of the original word, and the expansion of the
\nlast parameter is joined with the last part of the original
\nword.\u00a0 When there are no positional parameters, “$@” and\u00a0$@
\nexpand to nothing (i.e., they are removed).
\n#\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the number of positional parameters in decimal.
\n?\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the exit status of the most recently executed
\nforeground pipeline.
\n–\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the current option flags as specified upon
\ninvocation, by the set builtin command, or those set by the
\nshell itself (such as the -i option).
\n$\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the process ID of the shell.\u00a0 In a () subshell, it
\nexpands to the process ID of the current shell, not the
\nsubshell.
\n!\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the process ID of the job most recently placed into
\nthe background, whether executed as an asynchronous command or
\nusing the bg builtin (see JOB CONTROL below).
\n0\u00a0\u00a0\u00a0\u00a0\u00a0 Expands to the name of the shell or shell script.\u00a0 This is set
\nat shell initialization.\u00a0 If bash is invoked with a file of
\ncommands, $0 is set to the name of that file.\u00a0 If bash is
\nstarted with the -c option, then $0 is set to the first
\nargument after the string to be executed, if one is present.
\nOtherwise, it is set to the filename used to invoke bash, as
\ngiven by argument zero.
\n_\u00a0\u00a0\u00a0\u00a0\u00a0 At shell startup, set to the absolute pathname used to invoke
\nthe shell or shell script being executed as passed in the
\nenvironment or argument list.\u00a0 Subsequently, expands to the
\nlast argument to the previous command, after expansion.\u00a0 Also
\nset to the full pathname used to invoke each command executed
\nand placed in the environment exported to that command.\u00a0 When
\nchecking mail, this parameter holds the name of the mail file
\ncurrently being checked.<\/p>\n
Shell Variables
\nThe following variables are set by the shell:<\/p>\n
BASH\u00a0\u00a0 Expands to the full filename used to invoke this instance of
\nbash.
\nBASHOPTS
\nA colon-separated list of enabled shell options.\u00a0 Each word in
\nthe list is a valid argument for the -s option to the shopt
\nbuiltin command (see SHELL BUILTIN COMMANDS below).\u00a0 The
\noptions appearing in BASHOPTS are those reported as on by
\nshopt.\u00a0 If this variable is in the environment when bash
\nstarts up, each shell option in the list will be enabled
\nbefore reading any startup files.\u00a0 This variable is read-only.
\nBASHPID
\nExpands to the process ID of the current bash process.\u00a0 This
\ndiffers from $$ under certain circumstances, such as subshells
\nthat do not require bash to be re-initialized.
\nBASH_ALIASES
\nAn associative array variable whose members correspond to the
\ninternal list of aliases as maintained by the alias builtin.
\nElements added to this array appear in the alias list;
\nunsetting array elements cause aliases to be removed from the
\nalias list.
\nBASH_ARGC
\nAn array variable whose values are the number of parameters in
\neach frame of the current bash execution call stack.\u00a0 The
\nnumber of parameters to the current subroutine (shell function
\nor script executed with . or source) is at the top of the
\nstack.\u00a0 When a subroutine is executed, the number of
\nparameters passed is pushed onto BASH_ARGC.\u00a0 The shell sets
\nBASH_ARGC only when in extended debugging mode (see the
\ndescription of the extdebug option to the shopt builtin below)
\nBASH_ARGV
\nAn array variable containing all of the parameters in the
\ncurrent bash execution call stack.\u00a0 The final parameter of the
\nlast subroutine call is at the top of the stack; the first
\nparameter of the initial call is at the bottom.\u00a0 When a
\nsubroutine is executed, the parameters supplied are pushed
\nonto BASH_ARGV.\u00a0 The shell sets BASH_ARGV only when in
\nextended debugging mode (see the description of the extdebug
\noption to the shopt builtin below)
\nBASH_CMDS
\nAn associative array variable whose members correspond to the
\ninternal hash table of commands as maintained by the hash
\nbuiltin.\u00a0 Elements added to this array appear in the hash
\ntable; unsetting array elements cause commands to be removed
\nfrom the hash table.
\nBASH_COMMAND
\nThe command currently being executed or about to be executed,
\nunless the shell is executing a command as the result of a
\ntrap, in which case it is the command executing at the time of
\nthe trap.
\nBASH_EXECUTION_STRING
\nThe command argument to the -c invocation option.
\nBASH_LINENO
\nAn array variable whose members are the line numbers in source
\nfiles where each corresponding member of FUNCNAME was invoked.
\n${BASH_LINENO[$i]} is the line number in the source file
\n(${BASH_SOURCE[$i+1]}) where ${FUNCNAME[$i]} was called (or
\n${BASH_LINENO[$i-1]} if referenced within another shell
\nfunction).\u00a0 Use LINENO to obtain the current line number.
\nBASH_REMATCH
\nAn array variable whose members are assigned by the =~ binary
\noperator to the [[ conditional command.\u00a0 The element with
\nindex 0 is the portion of the string matching the entire
\nregular expression.\u00a0 The element with index n is the portion
\nof the string matching the nth parenthesized subexpression.
\nThis variable is read-only.
\nBASH_SOURCE
\nAn array variable whose members are the source filenames where
\nthe corresponding shell function names in the FUNCNAME array
\nvariable are defined.\u00a0 The shell function ${FUNCNAME[$i]} is
\ndefined in the file ${BASH_SOURCE[$i]} and called from
\n${BASH_SOURCE[$i+1]}.
\nBASH_SUBSHELL
\nIncremented by one within each subshell or subshell
\nenvironment when the shell begins executing in that
\nenvironment.\u00a0 The initial value is 0.
\nBASH_VERSINFO
\nA readonly array variable whose members hold version
\ninformation for this instance of bash.\u00a0 The values assigned to
\nthe array members are as follows:
\nBASH_VERSINFO[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The major version number (the
\nrelease).
\nBASH_VERSINFO[1]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The minor version number (the
\nversion).
\nBASH_VERSINFO[2]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The patch level.
\nBASH_VERSINFO[3]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The build version.
\nBASH_VERSINFO[4]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The release status (e.g., beta1).
\nBASH_VERSINFO[5]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The value of MACHTYPE.
\nBASH_VERSION
\nExpands to a string describing the version of this instance of
\nbash.
\nCOMP_CWORD
\nAn index into ${COMP_WORDS} of the word containing the current
\ncursor position.\u00a0 This variable is available only in shell
\nfunctions invoked by the programmable completion facilities
\n(see Programmable Completion below).
\nCOMP_KEY
\nThe key (or final key of a key sequence) used to invoke the
\ncurrent completion function.
\nCOMP_LINE
\nThe current command line.\u00a0 This variable is available only in
\nshell functions and external commands invoked by the
\nprogrammable completion facilities (see Programmable
\nCompletion below).
\nCOMP_POINT
\nThe index of the current cursor position relative to the
\nbeginning of the current command.\u00a0 If the current cursor
\nposition is at the end of the current command, the value of
\nthis variable is equal to ${#COMP_LINE}.\u00a0 This variable is
\navailable only in shell functions and external commands
\ninvoked by the programmable completion facilities (see
\nProgrammable Completion below).
\nCOMP_TYPE
\nSet to an integer value corresponding to the type of
\ncompletion attempted that caused a completion function to be
\ncalled: TAB, for normal completion, ?, for listing completions
\nafter successive tabs, !, for listing alternatives on partial
\nword completion, @, to list completions if the word is not
\nunmodified, or %, for menu completion.\u00a0 This variable is
\navailable only in shell functions and external commands
\ninvoked by the programmable completion facilities (see
\nProgrammable Completion below).
\nCOMP_WORDBREAKS
\nThe set of characters that the readline library treats as word
\nseparators when performing word completion.\u00a0 If
\nCOMP_WORDBREAKS is unset, it loses its special properties,
\neven if it is subsequently reset.
\nCOMP_WORDS
\nAn array variable (see Arrays below) consisting of the
\nindividual words in the current command line.\u00a0 The line is
\nsplit into words as readline would split it, using
\nCOMP_WORDBREAKS as described above.\u00a0 This variable is
\navailable only in shell functions invoked by the programmable
\ncompletion facilities (see Programmable Completion below).
\nCOPROC An array variable (see Arrays below) created to hold the file
\ndescriptors for output from and input to an unnamed coprocess
\n(see Coprocesses above).
\nDIRSTACK
\nAn array variable (see Arrays below) containing the current
\ncontents of the directory stack.\u00a0 Directories appear in the
\nstack in the order they are displayed by the dirs builtin.
\nAssigning to members of this array variable may be used to
\nmodify directories already in the stack, but the pushd and
\npopd builtins must be used to add and remove directories.
\nAssignment to this variable will not change the current
\ndirectory.\u00a0 If DIRSTACK is unset, it loses its special
\nproperties, even if it is subsequently reset.
\nEUID\u00a0\u00a0 Expands to the effective user ID of the current user,
\ninitialized at shell startup.\u00a0 This variable is readonly.
\nFUNCNAME
\nAn array variable containing the names of all shell functions
\ncurrently in the execution call stack.\u00a0 The element with index
\n0 is the name of any currently-executing shell function.\u00a0 The
\nbottom-most element (the one with the highest index) is
\n“main”.\u00a0 This variable exists only when a shell function is
\nexecuting.\u00a0 Assignments to FUNCNAME have no effect and return
\nan error status.\u00a0 If FUNCNAME is unset, it loses its special
\nproperties, even if it is subsequently reset.<\/p>\n
This variable can be used with BASH_LINENO and BASH_SOURCE.
\nEach element of FUNCNAME has corresponding elements in
\nBASH_LINENO and BASH_SOURCE to describe the call stack.\u00a0 For
\ninstance, ${FUNCNAME[$i]} was called from the file
\n${BASH_SOURCE[$i+1]} at line number ${BASH_LINENO[$i]}.\u00a0 The
\ncaller builtin displays the current call stack using this
\ninformation.
\nGROUPS An array variable containing the list of groups of which the
\ncurrent user is a member.\u00a0 Assignments to GROUPS have no
\neffect and return an error status.\u00a0 If GROUPS is unset, it
\nloses its special properties, even if it is subsequently
\nreset.
\nHISTCMD
\nThe history number, or index in the history list, of the
\ncurrent command.\u00a0 If HISTCMD is unset, it loses its special
\nproperties, even if it is subsequently reset.
\nHOSTNAME
\nAutomatically set to the name of the current host.
\nHOSTTYPE
\nAutomatically set to a string that uniquely describes the type
\nof machine on which bash is executing.\u00a0 The default is system-
\ndependent.
\nLINENO Each time this parameter is referenced, the shell substitutes
\na decimal number representing the current sequential line
\nnumber (starting with 1) within a script or function.\u00a0 When
\nnot in a script or function, the value substituted is not
\nguaranteed to be meaningful.\u00a0 If LINENO is unset, it loses its
\nspecial properties, even if it is subsequently reset.
\nMACHTYPE
\nAutomatically set to a string that fully describes the system
\ntype on which bash is executing, in the standard GNU cpu-
\ncompany-system format.\u00a0 The default is system-dependent.
\nMAPFILE
\nAn array variable (see Arrays below) created to hold the text
\nread by the mapfile builtin when no variable name is supplied.
\nOLDPWD The previous working directory as set by the cd command.
\nOPTARG The value of the last option argument processed by the getopts
\nbuiltin command (see SHELL BUILTIN COMMANDS below).
\nOPTIND The index of the next argument to be processed by the getopts
\nbuiltin command (see SHELL BUILTIN COMMANDS below).
\nOSTYPE Automatically set to a string that describes the operating
\nsystem on which bash is executing.\u00a0 The default is system-
\ndependent.
\nPIPESTATUS
\nAn array variable (see Arrays below) containing a list of exit
\nstatus values from the processes in the most-recently-executed
\nforeground pipeline (which may contain only a single command).
\nPPID\u00a0\u00a0 The process ID of the shell’s parent.\u00a0 This variable is
\nreadonly.
\nPWD\u00a0\u00a0\u00a0 The current working directory as set by the cd command.
\nRANDOM Each time this parameter is referenced, a random integer
\nbetween 0 and 32767 is generated.\u00a0 The sequence of random
\nnumbers may be initialized by assigning a value to RANDOM.\u00a0 If
\nRANDOM is unset, it loses its special properties, even if it
\nis subsequently reset.
\nREADLINE_LINE
\nThe contents of the readline line buffer, for use with “bind
\n-x” (see SHELL BUILTIN COMMANDS below).
\nREADLINE_POINT
\nThe position of the insertion point in the readline line
\nbuffer, for use with “bind -x” (see SHELL BUILTIN COMMANDS
\nbelow).
\nREPLY\u00a0 Set to the line of input read by the read builtin command when
\nno arguments are supplied.
\nSECONDS
\nEach time this parameter is referenced, the number of seconds
\nsince shell invocation is returned.\u00a0 If a value is assigned to
\nSECONDS, the value returned upon subsequent references is the
\nnumber of seconds since the assignment plus the value
\nassigned.\u00a0 If SECONDS is unset, it loses its special
\nproperties, even if it is subsequently reset.
\nSHELLOPTS
\nA colon-separated list of enabled shell options.\u00a0 Each word in
\nthe list is a valid argument for the -o option to the set
\nbuiltin command (see SHELL BUILTIN COMMANDS below).\u00a0 The
\noptions appearing in SHELLOPTS are those reported as on by set
\n-o.\u00a0 If this variable is in the environment when bash starts
\nup, each shell option in the list will be enabled before
\nreading any startup files.\u00a0 This variable is read-only.
\nSHLVL\u00a0 Incremented by one each time an instance of bash is started.
\nUID\u00a0\u00a0\u00a0 Expands to the user ID of the current user, initialized at
\nshell startup.\u00a0 This variable is readonly.<\/p>\n
The following variables are used by the shell.\u00a0 In some cases, bash
\nassigns a default value to a variable; these cases are noted below.<\/p>\n
BASH_COMPAT
\nThe value is used to set the shell’s compatibility level.\u00a0 See
\nthe description of the shopt builtin below under SHELL BUILTIN
\nCOMMANDS for a description of the various compatibility levels
\nand their effects.\u00a0 The value may be a decimal number (e.g.,
\n4.2) or an integer (e.g., 42) corresponding to the desired
\ncompatibility level.\u00a0 If BASH_COMPAT is unset or set to the
\nempty string, the compatibility level is set to the default
\nfor the current version.\u00a0 If BASH_COMPAT is set to a value
\nthat is not one of the valid compatibility levels, the shell
\nprints an error message and sets the compatibility level to
\nthe default for the current version.\u00a0 The valid compatibility
\nlevels correspond to the compatibility options accepted by the
\nshopt builtin described below (for example, compat42 means
\nthat 4.2 and 42 are valid values).\u00a0 The current version is
\nalso a valid value.
\nBASH_ENV
\nIf this parameter is set when bash is executing a shell
\nscript, its value is interpreted as a filename containing
\ncommands to initialize the shell, as in ~\/.bashrc.\u00a0 The value
\nof BASH_ENV is subjected to parameter expansion, command
\nsubstitution, and arithmetic expansion before being
\ninterpreted as a filename.\u00a0 PATH is not used to search for the
\nresultant filename.
\nBASH_XTRACEFD
\nIf set to an integer corresponding to a valid file descriptor,
\nbash will write the trace output generated when set -x is
\nenabled to that file descriptor.\u00a0 The file descriptor is
\nclosed when BASH_XTRACEFD is unset or assigned a new value.
\nUnsetting BASH_XTRACEFD or assigning it the empty string
\ncauses the trace output to be sent to the standard error.
\nNote that setting BASH_XTRACEFD to 2 (the standard error file
\ndescriptor) and then unsetting it will result in the standard
\nerror being closed.
\nCDPATH The search path for the cd command.\u00a0 This is a colon-separated
\nlist of directories in which the shell looks for destination
\ndirectories specified by the cd command.\u00a0 A sample value is
\n“.:~:\/usr”.
\nCHILD_MAX
\nSet the number of exited child status values for the shell to
\nremember.\u00a0 Bash will not allow this value to be decreased
\nbelow a POSIX-mandated minimum, and there is a maximum value
\n(currently 8192) that this may not exceed.\u00a0 The minimum value
\nis system-dependent.
\nCOLUMNS
\nUsed by the select compound command to determine the terminal
\nwidth when printing selection lists.\u00a0 Automatically set if the
\ncheckwinsize option is enabled or in an interactive shell upon
\nreceipt of a SIGWINCH.
\nCOMPREPLY
\nAn array variable from which bash reads the possible
\ncompletions generated by a shell function invoked by the
\nprogrammable completion facility (see Programmable Completion
\nbelow).\u00a0 Each array element contains one possible completion.
\nEMACS\u00a0 If bash finds this variable in the environment when the shell
\nstarts with value “t”, it assumes that the shell is running in
\nan Emacs shell buffer and disables line editing.
\nENV\u00a0\u00a0\u00a0 Similar to BASH_ENV; used when the shell is invoked in POSIX
\nmode.
\nFCEDIT The default editor for the fc builtin command.
\nFIGNORE
\nA colon-separated list of suffixes to ignore when performing
\nfilename completion (see READLINE below).\u00a0 A filename whose
\nsuffix matches one of the entries in FIGNORE is excluded from
\nthe list of matched filenames.\u00a0 A sample value is “.o:~”.
\nFUNCNEST
\nIf set to a numeric value greater than 0, defines a maximum
\nfunction nesting level.\u00a0 Function invocations that exceed this
\nnesting level will cause the current command to abort.
\nGLOBIGNORE
\nA colon-separated list of patterns defining the set of
\nfilenames to be ignored by pathname expansion.\u00a0 If a filename
\nmatched by a pathname expansion pattern also matches one of
\nthe patterns in GLOBIGNORE, it is removed from the list of
\nmatches.
\nHISTCONTROL
\nA colon-separated list of values controlling how commands are
\nsaved on the history list.\u00a0 If the list of values includes
\nignorespace, lines which begin with a space character are not
\nsaved in the history list.\u00a0 A value of ignoredups causes lines
\nmatching the previous history entry to not be saved.\u00a0 A value
\nof ignoreboth is shorthand for ignorespace and ignoredups.\u00a0 A
\nvalue of erasedups causes all previous lines matching the
\ncurrent line to be removed from the history list before that
\nline is saved.\u00a0 Any value not in the above list is ignored.
\nIf HISTCONTROL is unset, or does not include a valid value,
\nall lines read by the shell parser are saved on the history
\nlist, subject to the value of HISTIGNORE.\u00a0 The second and
\nsubsequent lines of a multi-line compound command are not
\ntested, and are added to the history regardless of the value
\nof HISTCONTROL.
\nHISTFILE
\nThe name of the file in which command history is saved (see
\nHISTORY below).\u00a0 The default value is ~\/.bash_history.\u00a0 If
\nunset, the command history is not saved when a shell exits.
\nHISTFILESIZE
\nThe maximum number of lines contained in the history file.
\nWhen this variable is assigned a value, the history file is
\ntruncated, if necessary, to contain no more than that number
\nof lines by removing the oldest entries.\u00a0 The history file is
\nalso truncated to this size after writing it when a shell
\nexits.\u00a0 If the value is 0, the history file is truncated to
\nzero size.\u00a0 Non-numeric values and numeric values less than
\nzero inhibit truncation.\u00a0 The shell sets the default value to
\nthe value of HISTSIZE after reading any startup files.
\nHISTIGNORE
\nA colon-separated list of patterns used to decide which
\ncommand lines should be saved on the history list.\u00a0 Each
\npattern is anchored at the beginning of the line and must
\nmatch the complete line (no implicit `*’ is appended).\u00a0 Each
\npattern is tested against the line after the checks specified
\nby HISTCONTROL are applied.\u00a0 In addition to the normal shell
\npattern matching characters, `&’ matches the previous history
\nline.\u00a0 `&’ may be escaped using a backslash; the backslash is
\nremoved before attempting a match.\u00a0 The second and subsequent
\nlines of a multi-line compound command are not tested, and are
\nadded to the history regardless of the value of HISTIGNORE.
\nHISTSIZE
\nThe number of commands to remember in the command history (see
\nHISTORY below).\u00a0 If the value is 0, commands are not saved in
\nthe history list.\u00a0 Numeric values less than zero result in
\nevery command being saved on the history list (there is no
\nlimit).\u00a0 The shell sets the default value to 500 after reading
\nany startup files.
\nHISTTIMEFORMAT
\nIf this variable is set and not null, its value is used as a
\nformat string for strftime(3) to print the time stamp
\nassociated with each history entry displayed by the history
\nbuiltin.\u00a0 If this variable is set, time stamps are written to
\nthe history file so they may be preserved across shell
\nsessions.\u00a0 This uses the history comment character to
\ndistinguish timestamps from other history lines.
\nHOME\u00a0\u00a0 The home directory of the current user; the default argument
\nfor the cd builtin command.\u00a0 The value of this variable is
\nalso used when performing tilde expansion.
\nHOSTFILE
\nContains the name of a file in the same format as \/etc\/hosts
\nthat should be read when the shell needs to complete a
\nhostname.\u00a0 The list of possible hostname completions may be
\nchanged while the shell is running; the next time hostname
\ncompletion is attempted after the value is changed, bash adds
\nthe contents of the new file to the existing list.\u00a0 If
\nHOSTFILE is set, but has no value, or does not name a readable
\nfile, bash attempts to read \/etc\/hosts to obtain the list of
\npossible hostname completions.\u00a0 When HOSTFILE is unset, the
\nhostname list is cleared.
\nIFS\u00a0\u00a0\u00a0 The Internal Field Separator that is used for word splitting
\nafter expansion and to split lines into words with the read
\nbuiltin command.\u00a0 The default value is
\n“<space><tab><newline>”.
\nIGNOREEOF
\nControls the action of an interactive shell on receipt of an
\nEOF character as the sole input.\u00a0 If set, the value is the
\nnumber of consecutive EOF characters which must be typed as
\nthe first characters on an input line before bash exits.\u00a0 If
\nthe variable exists but does not have a numeric value, or has
\nno value, the default value is 10.\u00a0 If it does not exist, EOF
\nsignifies the end of input to the shell.
\nINPUTRC
\nThe filename for the readline startup file, overriding the
\ndefault of ~\/.inputrc (see READLINE below).
\nLANG\u00a0\u00a0 Used to determine the locale category for any category not
\nspecifically selected with a variable starting with LC_.
\nLC_ALL This variable overrides the value of LANG and any other LC_
\nvariable specifying a locale category.
\nLC_COLLATE
\nThis variable determines the collation order used when sorting
\nthe results of pathname expansion, and determines the behavior
\nof range expressions, equivalence classes, and collating
\nsequences within pathname expansion and pattern matching.
\nLC_CTYPE
\nThis variable determines the interpretation of characters and
\nthe behavior of character classes within pathname expansion
\nand pattern matching.
\nLC_MESSAGES
\nThis variable determines the locale used to translate double-
\nquoted strings preceded by a $.
\nLC_NUMERIC
\nThis variable determines the locale category used for number
\nformatting.
\nLINES\u00a0 Used by the select compound command to determine the column
\nlength for printing selection lists.\u00a0 Automatically set if the
\ncheckwinsize option is enabled or in an interactive shell upon
\nreceipt of a SIGWINCH.
\nMAIL\u00a0\u00a0 If this parameter is set to a file or directory name and the
\nMAILPATH variable is not set, bash informs the user of the
\narrival of mail in the specified file or Maildir-format
\ndirectory.
\nMAILCHECK
\nSpecifies how often (in seconds) bash checks for mail.\u00a0 The
\ndefault is 60 seconds.\u00a0 When it is time to check for mail, the
\nshell does so before displaying the primary prompt.\u00a0 If this
\nvariable is unset, or set to a value that is not a number
\ngreater than or equal to zero, the shell disables mail
\nchecking.
\nMAILPATH
\nA colon-separated list of filenames to be checked for mail.
\nThe message to be printed when mail arrives in a particular
\nfile may be specified by separating the filename from the
\nmessage with a `?’.\u00a0 When used in the text of the message, $_
\nexpands to the name of the current mailfile.\u00a0 Example:
\nMAILPATH=’\/var\/mail\/bfox?”You have mail”:~\/shell-mail?”$_ has
\nmail!”‘
\nBash supplies a default value for this variable, but the
\nlocation of the user mail files that it uses is system
\ndependent (e.g., \/var\/mail\/$USER).
\nOPTERR If set to the value 1, bash displays error messages generated
\nby the getopts builtin command (see SHELL BUILTIN COMMANDS
\nbelow).\u00a0 OPTERR is initialized to 1 each time the shell is
\ninvoked or a shell script is executed.
\nPATH\u00a0\u00a0 The search path for commands.\u00a0 It is a colon-separated list of
\ndirectories in which the shell looks for commands (see COMMAND
\nEXECUTION below).\u00a0 A zero-length (null) directory name in the
\nvalue of PATH indicates the current directory.\u00a0 A null
\ndirectory name may appear as two adjacent colons, or as an
\ninitial or trailing colon.\u00a0 The default path is system-
\ndependent, and is set by the administrator who installs bash.
\nA common value is
\n“\/usr\/local\/bin:\/usr\/local\/sbin:\/usr\/bin:\/usr\/sbin:\/bin:\/sbin”.
\nPOSIXLY_CORRECT
\nIf this variable is in the environment when bash starts, the
\nshell enters posix mode before reading the startup files, as
\nif the –posix invocation option had been supplied.\u00a0 If it is
\nset while the shell is running, bash enables posix mode, as if
\nthe command set -o posix had been executed.
\nPROMPT_COMMAND
\nIf set, the value is executed as a command prior to issuing
\neach primary prompt.
\nPROMPT_DIRTRIM
\nIf set to a number greater than zero, the value is used as the
\nnumber of trailing directory components to retain when
\nexpanding the \\w and \\W prompt string escapes (see PROMPTING
\nbelow).\u00a0 Characters removed are replaced with an ellipsis.
\nPS1\u00a0\u00a0\u00a0 The value of this parameter is expanded (see PROMPTING below)
\nand used as the primary prompt string.\u00a0 The default value is
\n“\\s-\\v\\$ ”.
\nPS2\u00a0\u00a0\u00a0 The value of this parameter is expanded as with PS1 and used
\nas the secondary prompt string.\u00a0 The default is “> ”.
\nPS3\u00a0\u00a0\u00a0 The value of this parameter is used as the prompt for the
\nselect command (see SHELL GRAMMAR above).
\nPS4\u00a0\u00a0\u00a0 The value of this parameter is expanded as with PS1 and the
\nvalue is printed before each command bash displays during an
\nexecution trace.\u00a0 The first character of PS4 is replicated
\nmultiple times, as necessary, to indicate multiple levels of
\nindirection.\u00a0 The default is “+ ”.
\nSHELL\u00a0 The full pathname to the shell is kept in this environment
\nvariable.\u00a0 If it is not set when the shell starts, bash
\nassigns to it the full pathname of the current user’s login
\nshell.
\nTIMEFORMAT
\nThe value of this parameter is used as a format string
\nspecifying how the timing information for pipelines prefixed
\nwith the time reserved word should be displayed.\u00a0 The %
\ncharacter introduces an escape sequence that is expanded to a
\ntime value or other information.\u00a0 The escape sequences and
\ntheir meanings are as follows; the braces denote optional
\nportions.
\n%%\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 A literal %.
\n%[p][l]R\u00a0 The elapsed time in seconds.
\n%[p][l]U\u00a0 The number of CPU seconds spent in user mode.
\n%[p][l]S\u00a0 The number of CPU seconds spent in system mode.
\n%P\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 The CPU percentage, computed as (%U + %S) \/ %R.<\/p>\n
The optional p is a digit specifying the precision, the number
\nof fractional digits after a decimal point.\u00a0 A value of 0
\ncauses no decimal point or fraction to be output.\u00a0 At most
\nthree places after the decimal point may be specified; values
\nof p greater than 3 are changed to 3.\u00a0 If p is not specified,
\nthe value 3 is used.<\/p>\n
The optional l specifies a longer format, including minutes,
\nof the form MMmSS.FFs.\u00a0 The value of p determines whether or
\nnot the fraction is included.<\/p>\n
If this variable is not set, bash acts as if it had the value
\n$’\\nreal\\t%3lR\\nuser\\t%3lU\\nsys\\t%3lS’.\u00a0 If the value is null,
\nno timing information is displayed.\u00a0 A trailing newline is
\nadded when the format string is displayed.
\nTMOUT\u00a0 If set to a value greater than zero, TMOUT is treated as the
\ndefault timeout for the read builtin.\u00a0 The select command
\nterminates if input does not arrive after TMOUT seconds when
\ninput is coming from a terminal.\u00a0 In an interactive shell, the
\nvalue is interpreted as the number of seconds to wait for a
\nline of input after issuing the primary prompt.\u00a0 Bash
\nterminates after waiting for that number of seconds if a
\ncomplete line of input does not arrive.
\nTMPDIR If set, bash uses its value as the name of a directory in
\nwhich bash creates temporary files for the shell’s use.
\nauto_resume
\nThis variable controls how the shell interacts with the user
\nand job control.\u00a0 If this variable is set, single word simple
\ncommands without redirections are treated as candidates for
\nresumption of an existing stopped job.\u00a0 There is no ambiguity
\nallowed; if there is more than one job beginning with the
\nstring typed, the job most recently accessed is selected.\u00a0 The
\nname of a stopped job, in this context, is the command line
\nused to start it.\u00a0 If set to the value exact, the string
\nsupplied must match the name of a stopped job exactly; if set
\nto substring, the string supplied needs to match a substring
\nof the name of a stopped job.\u00a0 The substring value provides
\nfunctionality analogous to the %?\u00a0 job identifier (see JOB
\nCONTROL below).\u00a0 If set to any other value, the supplied
\nstring must be a prefix of a stopped job’s name; this provides
\nfunctionality analogous to the %string job identifier.
\nhistchars
\nThe two or three characters which control history expansion
\nand tokenization (see HISTORY EXPANSION below).\u00a0 The first
\ncharacter is the history expansion character, the character
\nwhich signals the start of a history expansion, normally `!’.
\nThe second character is the quick substitution character,
\nwhich is used as shorthand for re-running the previous command
\nentered, substituting one string for another in the command.
\nThe default is `^’.\u00a0 The optional third character is the
\ncharacter which indicates that the remainder of the line is a
\ncomment when found as the first character of a word, normally
\n`#’.\u00a0 The history comment character causes history
\nsubstitution to be skipped for the remaining words on the
\nline.\u00a0 It does not necessarily cause the shell parser to treat
\nthe rest of the line as a comment.<\/p>\n
Arrays
\nBash provides one-dimensional indexed and associative array
\nvariables.\u00a0 Any variable may be used as an indexed array; the declare
\nbuiltin will explicitly declare an array.\u00a0 There is no maximum limit
\non the size of an array, nor any requirement that members be indexed
\nor assigned contiguously.\u00a0 Indexed arrays are referenced using
\nintegers (including arithmetic expressions)\u00a0 and are zero-based;
\nassociative arrays are referenced using arbitrary strings.\u00a0 Unless
\notherwise noted, indexed array indices must be non-negative integers.<\/p>\n
An indexed array is created automatically if any variable is assigned
\nto using the syntax name[subscript]=value.\u00a0 The subscript is treated
\nas an arithmetic expression that must evaluate to a number.\u00a0 To
\nexplicitly declare an indexed array, use declare -a name (see SHELL
\nBUILTIN COMMANDS below).\u00a0 declare -a name[subscript] is also
\naccepted; the subscript is ignored.<\/p>\n
Associative arrays are created using declare -A name.<\/p>\n
Attributes may be specified for an array variable using the declare
\nand readonly builtins.\u00a0 Each attribute applies to all members of an
\narray.<\/p>\n
Arrays are assigned to using compound assignments of the form
\nname=(value1 … valuen), where each value is of the form
\n[subscript]=string.\u00a0 Indexed array assignments do not require
\nanything but string.\u00a0 When assigning to indexed arrays, if the
\noptional brackets and subscript are supplied, that index is assigned
\nto; otherwise the index of the element assigned is the last index
\nassigned to by the statement plus one.\u00a0 Indexing starts at zero.<\/p>\n
When assigning to an associative array, the subscript is required.<\/p>\n
This syntax is also accepted by the declare builtin.\u00a0 Individual
\narray elements may be assigned to using the name[subscript]=value
\nsyntax introduced above.\u00a0 When assigning to an indexed array, if name
\nis subscripted by a negative number, that number is interpreted as
\nrelative to one greater than the maximum index of name, so negative
\nindices count back from the end of the array, and an index of -1
\nreferences the last element.<\/p>\n
Any element of an array may be referenced using ${name[subscript]}.
\nThe braces are required to avoid conflicts with pathname expansion.
\nIf subscript is @ or *, the word expands to all members of name.
\nThese subscripts differ only when the word appears within double
\nquotes.\u00a0 If the word is double-quoted, ${name[*]} expands to a single
\nword with the value of each array member separated by the first
\ncharacter of the IFS special variable, and ${name[@]} expands each
\nelement of name to a separate word.\u00a0 When there are no array members,
\n${name[@]} expands to nothing.\u00a0 If the double-quoted expansion occurs
\nwithin a word, the expansion of the first parameter is joined with
\nthe beginning part of the original word, and the expansion of the
\nlast parameter is joined with the last part of the original word.
\nThis is analogous to the expansion of the special parameters * and @
\n(see Special Parameters above).\u00a0 ${#name[subscript]} expands to the
\nlength of ${name[subscript]}.\u00a0 If subscript is * or @, the expansion
\nis the number of elements in the array.\u00a0 Referencing an array
\nvariable without a subscript is equivalent to referencing the array
\nwith a subscript of 0.\u00a0 If the subscript used to reference an element
\nof an indexed array evaluates to a number less than zero, it is
\ninterpreted as relative to one greater than the maximum index of the
\narray, so negative indices count back from the end of the array, and
\nan index of -1 references the last element.<\/p>\n
An array variable is considered set if a subscript has been assigned
\na value.\u00a0 The null string is a valid value.<\/p>\n
It is possible to obtain the keys (indices) of an array as well as
\nthe values.\u00a0 ${!name[@]} and ${!name[*]} expand to the indices
\nassigned in array variable name.\u00a0 The treatment when in double quotes
\nis similar to the expansion of the special parameters @ and * within
\ndouble quotes.<\/p>\n
The unset builtin is used to destroy arrays.\u00a0 unset name[subscript]
\ndestroys the array element at index subscript.\u00a0 Negative subscripts
\nto indexed arrays are interpreted as described above.\u00a0 Care must be
\ntaken to avoid unwanted side effects caused by pathname expansion.
\nunset name, where name is an array, or unset name[subscript], where
\nsubscript is * or @, removes the entire array.<\/p>\n
The declare, local, and readonly builtins each accept a -a option to
\nspecify an indexed array and a -A option to specify an associative
\narray.\u00a0 If both options are supplied, -A takes precedence.\u00a0 The read
\nbuiltin accepts a -a option to assign a list of words read from the
\nstandard input to an array.\u00a0 The set and declare builtins display
\narray values in a way that allows them to be reused as assignments.<\/p>\n
EXPANSION<\/p>\n
Expansion is performed on the command line after it has been split
\ninto words.\u00a0 There are seven kinds of expansion performed: brace
\nexpansion, tilde expansion, parameter and variable expansion, command
\nsubstitution, arithmetic expansion, word splitting, and pathname
\nexpansion.<\/p>\n
The order of expansions is: brace expansion; tilde expansion,
\nparameter and variable expansion, arithmetic expansion, and command
\nsubstitution (done in a left-to-right fashion); word splitting; and
\npathname expansion.<\/p>\n
On systems that can support it, there is an additional expansion
\navailable: process substitution.\u00a0 This is performed at the same time
\nas tilde, parameter, variable, and arithmetic expansion and command
\nsubstitution.<\/p>\n
Only brace expansion, word splitting, and pathname expansion can
\nchange the number of words of the expansion; other expansions expand
\na single word to a single word.\u00a0 The only exceptions to this are the
\nexpansions of “$@” and “${name[@]}” as explained above (see
\nPARAMETERS).<\/p>\n
Brace Expansion
\nBrace expansion is a mechanism by which arbitrary strings may be
\ngenerated.\u00a0 This mechanism is similar to pathname expansion, but the
\nfilenames generated need not exist.\u00a0 Patterns to be brace expanded
\ntake the form of an optional preamble, followed by either a series of
\ncomma-separated strings or a sequence expression between a pair of
\nbraces, followed by an optional postscript.\u00a0 The preamble is prefixed
\nto each string contained within the braces, and the postscript is
\nthen appended to each resulting string, expanding left to right.<\/p>\n
Brace expansions may be nested.\u00a0 The results of each expanded string
\nare not sorted; left to right order is preserved.\u00a0 For example,
\na{d,c,b}e expands into `ade ace abe’.<\/p>\n
A sequence expression takes the form {x..y[..incr]}, where x and y
\nare either integers or single characters, and incr, an optional
\nincrement, is an integer.\u00a0 When integers are supplied, the expression
\nexpands to each number between x and y, inclusive.\u00a0 Supplied integers
\nmay be prefixed with 0 to force each term to have the same width.
\nWhen either x or y begins with a zero, the shell attempts to force
\nall generated terms to contain the same number of digits, zero-
\npadding where necessary.\u00a0 When characters are supplied, the
\nexpression expands to each character lexicographically between x and
\ny, inclusive, using the default C locale.\u00a0 Note that both x and y
\nmust be of the same type.\u00a0 When the increment is supplied, it is used
\nas the difference between each term.\u00a0 The default increment is 1 or
\n-1 as appropriate.<\/p>\n
Brace expansion is performed before any other expansions, and any
\ncharacters special to other expansions are preserved in the result.
\nIt is strictly textual.\u00a0 Bash does not apply any syntactic
\ninterpretation to the context of the expansion or the text between
\nthe braces.<\/p>\n
A correctly-formed brace expansion must contain unquoted opening and
\nclosing braces, and at least one unquoted comma or a valid sequence
\nexpression.\u00a0 Any incorrectly formed brace expansion is left
\nunchanged.\u00a0 A { or , may be quoted with a backslash to prevent its
\nbeing considered part of a brace expression.\u00a0 To avoid conflicts with
\nparameter expansion, the string ${ is not considered eligible for
\nbrace expansion.<\/p>\n
This construct is typically used as shorthand when the common prefix
\nof the strings to be generated is longer than in the above example:<\/p>\n
mkdir \/usr\/local\/src\/bash\/{old,new,dist,bugs}
\nor
\nchown root \/usr\/{ucb\/{ex,edit},lib\/{ex?.?*,how_ex}}<\/p>\n
Brace expansion introduces a slight incompatibility with historical
\nversions of sh.\u00a0 sh does not treat opening or closing braces
\nspecially when they appear as part of a word, and preserves them in
\nthe output.\u00a0 Bash removes braces from words as a consequence of brace
\nexpansion.\u00a0 For example, a word entered to sh as file{1,2} appears
\nidentically in the output.\u00a0 The same word is output as file1 file2
\nafter expansion by bash.\u00a0 If strict compatibility with sh is desired,
\nstart bash with the +B option or disable brace expansion with the +B
\noption to the set command (see SHELL BUILTIN COMMANDS below).<\/p>\n
Tilde Expansion
\nIf a word begins with an unquoted tilde character (`~’), all of the
\ncharacters preceding the first unquoted slash (or all characters, if
\nthere is no unquoted slash) are considered a tilde-prefix.\u00a0 If none
\nof the characters in the tilde-prefix are quoted, the characters in
\nthe tilde-prefix following the tilde are treated as a possible login
\nname.\u00a0 If this login name is the null string, the tilde is replaced
\nwith the value of the shell parameter HOME.\u00a0 If HOME is unset, the
\nhome directory of the user executing the shell is substituted
\ninstead.\u00a0 Otherwise, the tilde-prefix is replaced with the home
\ndirectory associated with the specified login name.<\/p>\n
If the tilde-prefix is a `~+’, the value of the shell variable PWD
\nreplaces the tilde-prefix.\u00a0 If the tilde-prefix is a `~-‘, the value
\nof the shell variable OLDPWD, if it is set, is substituted.\u00a0 If the
\ncharacters following the tilde in the tilde-prefix consist of a
\nnumber N, optionally prefixed by a `+’ or a `-‘, the tilde-prefix is
\nreplaced with the corresponding element from the directory stack, as
\nit would be displayed by the dirs builtin invoked with the tilde-
\nprefix as an argument.\u00a0 If the characters following the tilde in the
\ntilde-prefix consist of a number without a leading `+’ or `-‘, `+’ is
\nassumed.<\/p>\n
If the login name is invalid, or the tilde expansion fails, the word
\nis unchanged.<\/p>\n
Each variable assignment is checked for unquoted tilde-prefixes
\nimmediately following a : or the first =.\u00a0 In these cases, tilde
\nexpansion is also performed.\u00a0 Consequently, one may use filenames
\nwith tildes in assignments to PATH, MAILPATH, and CDPATH, and the
\nshell assigns the expanded value.<\/p>\n
Parameter Expansion
\nThe `$’ character introduces parameter expansion, command
\nsubstitution, or arithmetic expansion.\u00a0 The parameter name or symbol
\nto be expanded may be enclosed in braces, which are optional but
\nserve to protect the variable to be expanded from characters
\nimmediately following it which could be interpreted as part of the
\nname.<\/p>\n
When braces are used, the matching ending brace is the first `}’ not
\nescaped by a backslash or within a quoted string, and not within an
\nembedded arithmetic expansion, command substitution, or parameter
\nexpansion.<\/p>\n
${parameter}
\nThe value of parameter is substituted.\u00a0 The braces are
\nrequired when parameter is a positional parameter with more
\nthan one digit, or when parameter is followed by a character
\nwhich is not to be interpreted as part of its name.\u00a0 The
\nparameter is a shell parameter as described above PARAMETERS)
\nor an array reference (Arrays).<\/p>\n
If the first character of parameter is an exclamation point (!), it
\nintroduces a level of variable indirection.\u00a0 Bash uses the value of
\nthe variable formed from the rest of parameter as the name of the
\nvariable; this variable is then expanded and that value is used in
\nthe rest of the substitution, rather than the value of parameter
\nitself.\u00a0 This is known as indirect expansion.\u00a0 The exceptions to this
\nare the expansions of ${!prefix*} and ${!name[@]} described below.
\nThe exclamation point must immediately follow the left brace in order
\nto introduce indirection.<\/p>\n
In each of the cases below, word is subject to tilde expansion,
\nparameter expansion, command substitution, and arithmetic expansion.<\/p>\n
When not performing substring expansion, using the forms documented
\nbelow (e.g., :-), bash tests for a parameter that is unset or null.
\nOmitting the colon results in a test only for a parameter that is
\nunset.<\/p>\n
${parameter:-word}
\nUse Default Values.\u00a0 If parameter is unset or null, the
\nexpansion of word is substituted.\u00a0 Otherwise, the value of
\nparameter is substituted.
\n${parameter:=word}
\nAssign Default Values.\u00a0 If parameter is unset or null, the
\nexpansion of word is assigned to parameter.\u00a0 The value of
\nparameter is then substituted.\u00a0 Positional parameters and
\nspecial parameters may not be assigned to in this way.
\n${parameter:?word}
\nDisplay Error if Null or Unset.\u00a0 If parameter is null or
\nunset, the expansion of word (or a message to that effect if
\nword is not present) is written to the standard error and the
\nshell, if it is not interactive, exits.\u00a0 Otherwise, the value
\nof parameter is substituted.
\n${parameter:+word}
\nUse Alternate Value.\u00a0 If parameter is null or unset, nothing
\nis substituted, otherwise the expansion of word is
\nsubstituted.
\n${parameter:offset}
\n${parameter:offset:length}
\nSubstring Expansion.\u00a0 Expands to up to length characters of
\nthe value of parameter starting at the character specified by
\noffset.\u00a0 If parameter is @, an indexed array subscripted by @
\nor *, or an associative array name, the results differ as
\ndescribed below.\u00a0 If length is omitted, expands to the
\nsubstring of the value of parameter starting at the character
\nspecified by offset and extending to the end of the value.
\nlength and offset are arithmetic expressions (see ARITHMETIC
\nEVALUATION below).<\/p>\n
If offset evaluates to a number less than zero, the value is
\nused as an offset in characters from the end of the value of
\nparameter.\u00a0 If length evaluates to a number less than zero, it
\nis interpreted as an offset in characters from the end of the
\nvalue of parameter rather than a number of characters, and the
\nexpansion is the characters between offset and that result.
\nNote that a negative offset must be separated from the colon
\nby at least one space to avoid being confused with the :-
\nexpansion.<\/p>\n
If parameter is @, the result is length positional parameters
\nbeginning at offset.\u00a0 A negative offset is taken relative to
\none greater than the greatest positional parameter, so an
\noffset of -1 evaluates to the last positional parameter.\u00a0 It
\nis an expansion error if length evaluates to a number less
\nthan zero.<\/p>\n
If parameter is an indexed array name subscripted by @ or *,
\nthe result is the length members of the array beginning with
\n${parameter[offset]}.\u00a0 A negative offset is taken relative to
\none greater than the maximum index of the specified array.\u00a0 It
\nis an expansion error if length evaluates to a number less
\nthan zero.<\/p>\n
Substring expansion applied to an associative array produces
\nundefined results.<\/p>\n
Substring indexing is zero-based unless the positional
\nparameters are used, in which case the indexing starts at 1 by
\ndefault.\u00a0 If offset is 0, and the positional parameters are
\nused, $0 is prefixed to the list.<\/p>\n
${!prefix*}
\n${!prefix@}
\nNames matching prefix.\u00a0 Expands to the names of variables
\nwhose names begin with prefix, separated by the first
\ncharacter of the IFS special variable.\u00a0 When @ is used and the
\nexpansion appears within double quotes, each variable name
\nexpands to a separate word.<\/p>\n
${!name[@]}
\n${!name[*]}
\nList of array keys.\u00a0 If name is an array variable, expands to
\nthe list of array indices (keys) assigned in name.\u00a0 If name is
\nnot an array, expands to 0 if name is set and null otherwise.
\nWhen @ is used and the expansion appears within double quotes,
\neach key expands to a separate word.<\/p>\n
${#parameter}
\nParameter length.\u00a0 The length in characters of the value of
\nparameter is substituted.\u00a0 If parameter is * or @, the value
\nsubstituted is the number of positional parameters.\u00a0 If
\nparameter is an array name subscripted by * or @, the value
\nsubstituted is the number of elements in the array.\u00a0 If
\nparameter is an indexed array name subscripted by a negative
\nnumber, that number is interpreted as relative to one greater
\nthan the maximum index of parameter, so negative indices count
\nback from the end of the array, and an index of -1 references
\nthe last element.<\/p>\n
${parameter#word}
\n${parameter##word}
\nRemove matching prefix pattern.\u00a0 The word is expanded to
\nproduce a pattern just as in pathname expansion.\u00a0 If the
\npattern matches the beginning of the value of parameter, then
\nthe result of the expansion is the expanded value of parameter
\nwith the shortest matching pattern (the “#” case) or the
\nlongest matching pattern (the “##” case) deleted.\u00a0 If
\nparameter is @ or *, the pattern removal operation is applied
\nto each positional parameter in turn, and the expansion is the
\nresultant list.\u00a0 If parameter is an array variable subscripted
\nwith @ or *, the pattern removal operation is applied to each
\nmember of the array in turn, and the expansion is the
\nresultant list.<\/p>\n
${parameter%word}
\n${parameter%%word}
\nRemove matching suffix pattern.\u00a0 The word is expanded to
\nproduce a pattern just as in pathname expansion.\u00a0 If the
\npattern matches a trailing portion of the expanded value of
\nparameter, then the result of the expansion is the expanded
\nvalue of parameter with the shortest matching pattern (the
\n“%” case) or the longest matching pattern (the “%%” case)
\ndeleted.\u00a0 If parameter is @ or *, the pattern removal
\noperation is applied to each positional parameter in turn, and
\nthe expansion is the resultant list.\u00a0 If parameter is an array
\nvariable subscripted with @ or *, the pattern removal
\noperation is applied to each member of the array in turn, and
\nthe expansion is the resultant list.<\/p>\n
${parameter\/pattern\/string}
\nPattern substitution.\u00a0 The pattern is expanded to produce a
\npattern just as in pathname expansion.\u00a0 Parameter is expanded
\nand the longest match of pattern against its value is replaced
\nwith string.\u00a0 If pattern begins with \/, all matches of pattern
\nare replaced with string.\u00a0 Normally only the first match is
\nreplaced.\u00a0 If pattern begins with #, it must match at the
\nbeginning of the expanded value of parameter.\u00a0 If pattern
\nbegins with %, it must match at the end of the expanded value
\nof parameter.\u00a0 If string is null, matches of pattern are
\ndeleted and the \/ following pattern may be omitted.\u00a0 If
\nparameter is @ or *, the substitution operation is applied to
\neach positional parameter in turn, and the expansion is the
\nresultant list.\u00a0 If parameter is an array variable subscripted
\nwith @ or *, the substitution operation is applied to each
\nmember of the array in turn, and the expansion is the
\nresultant list.<\/p>\n
${parameter^pattern}
\n${parameter^^pattern}
\n${parameter,pattern}
\n${parameter,,pattern}
\nCase modification.\u00a0 This expansion modifies the case of
\nalphabetic characters in parameter.\u00a0 The pattern is expanded
\nto produce a pattern just as in pathname expansion.\u00a0 Each
\ncharacter in the expanded value of parameter is tested against
\npattern, and, if it matches the pattern, its case is
\nconverted.\u00a0 The pattern should not attempt to match more than
\none character.\u00a0 The ^ operator converts lowercase letters
\nmatching pattern to uppercase; the , operator converts
\nmatching uppercase letters to lowercase.\u00a0 The ^^ and ,,
\nexpansions convert each matched character in the expanded
\nvalue; the ^ and , expansions match and convert only the first
\ncharacter in the expanded value.\u00a0 If pattern is omitted, it is
\ntreated like a ?, which matches every character.\u00a0 If parameter
\nis @ or *, the case modification operation is applied to each
\npositional parameter in turn, and the expansion is the
\nresultant list.\u00a0 If parameter is an array variable subscripted
\nwith @ or *, the case modification operation is applied to
\neach member of the array in turn, and the expansion is the
\nresultant list.<\/p>\n
Command Substitution
\nCommand substitution allows the output of a command to replace the
\ncommand name.\u00a0 There are two forms:<\/p>\n
$(command)
\nor
\n`command`<\/p>\n
Bash performs the expansion by executing command and replacing the
\ncommand substitution with the standard output of the command, with
\nany trailing newlines deleted.\u00a0 Embedded newlines are not deleted,
\nbut they may be removed during word splitting.\u00a0 The command
\nsubstitution $(cat file) can be replaced by the equivalent but faster
\n$(< file).<\/p>\n
When the old-style backquote form of substitution is used, backslash
\nretains its literal meaning except when followed by $, `, or \\.\u00a0 The
\nfirst backquote not preceded by a backslash terminates the command
\nsubstitution.\u00a0 When using the $(command) form, all characters between
\nthe parentheses make up the command; none are treated specially.<\/p>\n
Command substitutions may be nested.\u00a0 To nest when using the
\nbackquoted form, escape the inner backquotes with backslashes.<\/p>\n
If the substitution appears within double quotes, word splitting and
\npathname expansion are not performed on the results.<\/p>\n
Arithmetic Expansion
\nArithmetic expansion allows the evaluation of an arithmetic
\nexpression and the substitution of the result.\u00a0 The format for
\narithmetic expansion is:<\/p>\n
$((expression))<\/p>\n
The expression is treated as if it were within double quotes, but a
\ndouble quote inside the parentheses is not treated specially.\u00a0 All
\ntokens in the expression undergo parameter and variable expansion,
\ncommand substitution, and quote removal.\u00a0 The result is treated as
\nthe arithmetic expression to be evaluated.\u00a0 Arithmetic expansions may
\nbe nested.<\/p>\n
The evaluation is performed according to the rules listed below under
\nARITHMETIC EVALUATION.\u00a0 If expression is invalid, bash prints a
\nmessage indicating failure and no substitution occurs.<\/p>\n
Process Substitution
\nProcess substitution is supported on systems that support named pipes
\n(FIFOs) or the \/dev\/fd method of naming open files.\u00a0 It takes the
\nform of <(list) or >(list).\u00a0 The process list is run with its input
\nor output connected to a FIFO or some file in \/dev\/fd.\u00a0 The name of
\nthis file is passed as an argument to the current command as the
\nresult of the expansion.\u00a0 If the >(list) form is used, writing to the
\nfile will provide input for list.\u00a0 If the <(list) form is used, the
\nfile passed as an argument should be read to obtain the output of
\nlist.<\/p>\n
When available, process substitution is performed simultaneously with
\nparameter and variable expansion, command substitution, and
\narithmetic expansion.<\/p>\n
Word Splitting
\nThe shell scans the results of parameter expansion, command
\nsubstitution, and arithmetic expansion that did not occur within
\ndouble quotes for word splitting.<\/p>\n
The shell treats each character of IFS as a delimiter, and splits the
\nresults of the other expansions into words using these characters as
\nfield terminators.\u00a0 If IFS is unset, or its value is exactly
\n<space><tab><newline>, the default, then sequences of <space>, <tab>,
\nand <newline> at the beginning and end of the results of the previous
\nexpansions are ignored, and any sequence of IFS characters not at the
\nbeginning or end serves to delimit words.\u00a0 If IFS has a value other
\nthan the default, then sequences of the whitespace characters space
\nand tab are ignored at the beginning and end of the word, as long as
\nthe whitespace character is in the value of IFS (an IFS whitespace
\ncharacter).\u00a0 Any character in IFS that is not IFS whitespace, along
\nwith any adjacent IFS whitespace characters, delimits a field.\u00a0 A
\nsequence of IFS whitespace characters is also treated as a delimiter.
\nIf the value of IFS is null, no word splitting occurs.<\/p>\n
Explicit null arguments (“” or ”) are retained.\u00a0 Unquoted implicit
\nnull arguments, resulting from the expansion of parameters that have
\nno values, are removed.\u00a0 If a parameter with no value is expanded
\nwithin double quotes, a null argument results and is retained.<\/p>\n
Note that if no expansion occurs, no splitting is performed.<\/p>\n
Pathname Expansion
\nAfter word splitting, unless the -f option has been set, bash scans
\neach word for the characters *, ?, and [.\u00a0 If one of these characters
\nappears, then the word is regarded as a pattern, and replaced with an
\nalphabetically sorted list of filenames matching the pattern (see
\nPattern Matching below).\u00a0 If no matching filenames are found, and the
\nshell option nullglob is not enabled, the word is left unchanged.\u00a0 If
\nthe nullglob option is set, and no matches are found, the word is
\nremoved.\u00a0 If the failglob shell option is set, and no matches are
\nfound, an error message is printed and the command is not executed.
\nIf the shell option nocaseglob is enabled, the match is performed
\nwithout regard to the case of alphabetic characters.\u00a0 When a pattern
\nis used for pathname expansion, the character “.”\u00a0 at the start of
\na name or immediately following a slash must be matched explicitly,
\nunless the shell option dotglob is set.\u00a0 When matching a pathname,
\nthe slash character must always be matched explicitly.\u00a0 In other
\ncases, the “.”\u00a0 character is not treated specially.\u00a0 See the
\ndescription of shopt below under SHELL BUILTIN COMMANDS for a
\ndescription of the nocaseglob, nullglob, failglob, and dotglob shell
\noptions.<\/p>\n
The GLOBIGNORE shell variable may be used to restrict the set of
\nfilenames matching a pattern.\u00a0 If GLOBIGNORE is set, each matching
\nfilename that also matches one of the patterns in GLOBIGNORE is
\nremoved from the list of matches.\u00a0 The filenames “.”\u00a0 and “..”
\nare always ignored when GLOBIGNORE is set and not null.\u00a0 However,
\nsetting GLOBIGNORE to a non-null value has the effect of enabling the
\ndotglob shell option, so all other filenames beginning with a “.”
\nwill match.\u00a0 To get the old behavior of ignoring filenames beginning
\nwith a “.”, make “.*”\u00a0 one of the patterns in GLOBIGNORE.\u00a0 The
\ndotglob option is disabled when GLOBIGNORE is unset.<\/p>\n
Pattern Matching<\/p>\n
Any character that appears in a pattern, other than the special
\npattern characters described below, matches itself.\u00a0 The NUL
\ncharacter may not occur in a pattern.\u00a0 A backslash escapes the
\nfollowing character; the escaping backslash is discarded when
\nmatching.\u00a0 The special pattern characters must be quoted if they are
\nto be matched literally.<\/p>\n
The special pattern characters have the following meanings:<\/p>\n
*\u00a0\u00a0\u00a0\u00a0\u00a0 Matches any string, including the null string.\u00a0 When
\nthe globstar shell option is enabled, and * is used in
\na pathname expansion context, two adjacent *s used as a
\nsingle pattern will match all files and zero or more
\ndirectories and subdirectories.\u00a0 If followed by a \/,
\ntwo adjacent *s will match only directories and
\nsubdirectories.
\n?\u00a0\u00a0\u00a0\u00a0\u00a0 Matches any single character.
\n[…]\u00a0 Matches any one of the enclosed characters.\u00a0 A pair of
\ncharacters separated by a hyphen denotes a range
\nexpression; any character that falls between those two
\ncharacters, inclusive, using the current locale’s
\ncollating sequence and character set, is matched.\u00a0 If
\nthe first character following the [ is a !\u00a0 or a ^ then
\nany character not enclosed is matched.\u00a0 The sorting
\norder of characters in range expressions is determined
\nby the current locale and the values of the LC_COLLATE
\nor LC_ALL shell variables, if set.\u00a0 To obtain the
\ntraditional interpretation of range expressions, where
\n[a-d] is equivalent to [abcd], set value of the LC_ALL
\nshell variable to C, or enable the globasciiranges
\nshell option.\u00a0 A – may be matched by including it as
\nthe first or last character in the set.\u00a0 A ] may be
\nmatched by including it as the first character in the
\nset.<\/p>\n
Within [ and ], character classes can be specified
\nusing the syntax [:class:], where class is one of the
\nfollowing classes defined in the POSIX standard:
\nalnum alpha ascii blank cntrl digit graph lower print
\npunct space upper word xdigit
\nA character class matches any character belonging to
\nthat class.\u00a0 The word character class matches letters,
\ndigits, and the character _.<\/p>\n
Within [ and ], an equivalence class can be specified
\nusing the syntax [=c=], which matches all characters
\nwith the same collation weight (as defined by the
\ncurrent locale) as the character c.<\/p>\n
Within [ and ], the syntax [.symbol.] matches the
\ncollating symbol symbol.<\/p>\n
If the extglob shell option is enabled using the shopt builtin,
\nseveral extended pattern matching operators are recognized.\u00a0 In the
\nfollowing description, a pattern-list is a list of one or more
\npatterns separated by a |.\u00a0 Composite patterns may be formed using
\none or more of the following sub-patterns:<\/p>\n
?(pattern-list)
\nMatches zero or one occurrence of the given patterns
\n*(pattern-list)
\nMatches zero or more occurrences of the given patterns
\n+(pattern-list)
\nMatches one or more occurrences of the given patterns
\n@(pattern-list)
\nMatches one of the given patterns
\n!(pattern-list)
\nMatches anything except one of the given patterns<\/p>\n
Quote Removal
\nAfter the preceding expansions, all unquoted occurrences of the
\ncharacters \\, ‘, and ” that did not result from one of the above
\nexpansions are removed.<\/p>\n
REDIRECTION<\/p>\n
Before a command is executed, its input and output may be redirected
\nusing a special notation interpreted by the shell.\u00a0 Redirection
\nallows commands’ file handles to be duplicated, opened, closed, made
\nto refer to different files, and can change the files the command
\nreads from and writes to.\u00a0 Redirection may also be used to modify
\nfile handles in the current shell execution environment.\u00a0 The
\nfollowing redirection operators may precede or appear anywhere within
\na simple command or may follow a command.\u00a0 Redirections are processed
\nin the order they appear, from left to right.<\/p>\n
Each redirection that may be preceded by a file descriptor number may
\ninstead be preceded by a word of the form {varname}.\u00a0 In this case,
\nfor each redirection operator except >&- and <&-, the shell will
\nallocate a file descriptor greater than or equal to 10 and assign it
\nto varname.\u00a0 If >&- or <&- is preceded by {varname}, the value of
\nvarname defines the file descriptor to close.<\/p>\n
In the following descriptions, if the file descriptor number is
\nomitted, and the first character of the redirection operator is <,
\nthe redirection refers to the standard input (file descriptor 0).\u00a0 If
\nthe first character of the redirection operator is >, the redirection
\nrefers to the standard output (file descriptor 1).<\/p>\n
The word following the redirection operator in the following
\ndescriptions, unless otherwise noted, is subjected to brace
\nexpansion, tilde expansion, parameter and variable expansion, command
\nsubstitution, arithmetic expansion, quote removal, pathname
\nexpansion, and word splitting.\u00a0 If it expands to more than one word,
\nbash reports an error.<\/p>\n
Note that the order of redirections is significant.\u00a0 For example, the
\ncommand<\/p>\n
ls > dirlist 2>&1<\/p>\n
directs both standard output and standard error to the file dirlist,
\nwhile the command<\/p>\n
ls 2>&1 > dirlist<\/p>\n
directs only the standard output to file dirlist, because the
\nstandard error was duplicated from the standard output before the
\nstandard output was redirected to dirlist.<\/p>\n
Bash handles several filenames specially when they are used in
\nredirections, as described in the following table:<\/p>\n
\/dev\/fd\/fd
\nIf fd is a valid integer, file descriptor fd is
\nduplicated.
\n\/dev\/stdin
\nFile descriptor 0 is duplicated.
\n\/dev\/stdout
\nFile descriptor 1 is duplicated.
\n\/dev\/stderr
\nFile descriptor 2 is duplicated.
\n\/dev\/tcp\/host\/port
\nIf host is a valid hostname or Internet address, and
\nport is an integer port number or service name, bash
\nattempts to open the corresponding TCP socket.
\n\/dev\/udp\/host\/port
\nIf host is a valid hostname or Internet address, and
\nport is an integer port number or service name, bash
\nattempts to open the corresponding UDP socket.<\/p>\n
A failure to open or create a file causes the redirection to fail.<\/p>\n
Redirections using file descriptors greater than 9 should be used
\nwith care, as they may conflict with file descriptors the shell uses
\ninternally.<\/p>\n
Redirecting Input
\nRedirection of input causes the file whose name results from the
\nexpansion of word to be opened for reading on file descriptor n, or
\nthe standard input (file descriptor 0) if n is not specified.<\/p>\n
The general format for redirecting input is:<\/p>\n
[n]<word<\/p>\n
Redirecting Output
\nRedirection of output causes the file whose name results from the
\nexpansion of word to be opened for writing on file descriptor n, or
\nthe standard output (file descriptor 1) if n is not specified.\u00a0 If
\nthe file does not exist it is created; if it does exist it is
\ntruncated to zero size.<\/p>\n
The general format for redirecting output is:<\/p>\n
[n]>word<\/p>\n
If the redirection operator is >, and the noclobber option to the set
\nbuiltin has been enabled, the redirection will fail if the file whose
\nname results from the expansion of word exists and is a regular file.
\nIf the redirection operator is >|, or the redirection operator is >
\nand the noclobber option to the set builtin command is not enabled,
\nthe redirection is attempted even if the file named by word exists.<\/p>\n
Appending Redirected Output
\nRedirection of output in this fashion causes the file whose name
\nresults from the expansion of word to be opened for appending on file
\ndescriptor n, or the standard output (file descriptor 1) if n is not
\nspecified.\u00a0 If the file does not exist it is created.<\/p>\n
The general format for appending output is:<\/p>\n
[n]>>word<\/p>\n
Redirecting Standard Output and Standard Error
\nThis construct allows both the standard output (file descriptor 1)
\nand the standard error output (file descriptor 2) to be redirected to
\nthe file whose name is the expansion of word.<\/p>\n
There are two formats for redirecting standard output and standard
\nerror:<\/p>\n
&>word
\nand
\n>&word<\/p>\n
Of the two forms, the first is preferred.\u00a0 This is semantically
\nequivalent to<\/p>\n
>word 2>&1<\/p>\n
When using the second form, word may not expand to a number or -.\u00a0 If
\nit does, other redirection operators apply (see Duplicating File
\nDescriptors below) for compatibility reasons.<\/p>\n
Appending Standard Output and Standard Error
\nThis construct allows both the standard output (file descriptor 1)
\nand the standard error output (file descriptor 2) to be appended to
\nthe file whose name is the expansion of word.<\/p>\n
The format for appending standard output and standard error is:<\/p>\n
&>>word<\/p>\n
This is semantically equivalent to<\/p>\n
>>word 2>&1<\/p>\n
(see Duplicating File Descriptors below).<\/p>\n
Here Documents
\nThis type of redirection instructs the shell to read input from the
\ncurrent source until a line containing only delimiter (with no
\ntrailing blanks) is seen.\u00a0 All of the lines read up to that point are
\nthen used as the standard input for a command.<\/p>\n
The format of here-documents is:<\/p>\n
<<[-]word
\nhere-document
\ndelimiter<\/p>\n
No parameter and variable expansion, command substitution, arithmetic
\nexpansion, or pathname expansion is performed on word.\u00a0 If any
\ncharacters in word are quoted, the delimiter is the result of quote
\nremoval on word, and the lines in the here-document are not expanded.
\nIf word is unquoted, all lines of the here-document are subjected to
\nparameter expansion, command substitution, and arithmetic expansion,
\nthe character sequence \\<newline> is ignored, and \\ must be used to
\nquote the characters \\, $, and `.<\/p>\n
If the redirection operator is <<-, then all leading tab characters
\nare stripped from input lines and the line containing delimiter.
\nThis allows here-documents within shell scripts to be indented in a
\nnatural fashion.<\/p>\n
Here Strings
\nA variant of here documents, the format is:<\/p>\n
<<<word<\/p>\n
The word undergoes brace expansion, tilde expansion, parameter and
\nvariable expansion, command substitution, arithmetic expansion, and
\nquote removal.\u00a0 Pathname expansion and word splitting are not
\nperformed.\u00a0 The result is supplied as a single string to the command
\non its standard input.<\/p>\n
Duplicating File Descriptors
\nThe redirection operator<\/p>\n
[n]<&word<\/p>\n
is used to duplicate input file descriptors.\u00a0 If word expands to one
\nor more digits, the file descriptor denoted by n is made to be a copy
\nof that file descriptor.\u00a0 If the digits in word do not specify a file
\ndescriptor open for input, a redirection error occurs.\u00a0 If word
\nevaluates to -, file descriptor n is closed.\u00a0 If n is not specified,
\nthe standard input (file descriptor 0) is used.<\/p>\n
The operator<\/p>\n
[n]>&word<\/p>\n
is used similarly to duplicate output file descriptors.\u00a0 If n is not
\nspecified, the standard output (file descriptor 1) is used.\u00a0 If the
\ndigits in word do not specify a file descriptor open for output, a
\nredirection error occurs.\u00a0 If word evaluates to -, file descriptor n
\nis closed.\u00a0 As a special case, if n is omitted, and word does not
\nexpand to one or more digits or -, the standard output and standard
\nerror are redirected as described previously.<\/p>\n
Moving File Descriptors
\nThe redirection operator<\/p>\n
[n]<&digit-<\/p>\n
moves the file descriptor digit to file descriptor n, or the standard
\ninput (file descriptor 0) if n is not specified.\u00a0 digit is closed
\nafter being duplicated to n.<\/p>\n
Similarly, the redirection operator<\/p>\n
[n]>&digit-<\/p>\n
moves the file descriptor digit to file descriptor n, or the standard
\noutput (file descriptor 1) if n is not specified.<\/p>\n
Opening File Descriptors for Reading and Writing
\nThe redirection operator<\/p>\n
[n]<>word<\/p>\n
causes the file whose name is the expansion of word to be opened for
\nboth reading and writing on file descriptor n, or on file descriptor
\n0 if n is not specified.\u00a0 If the file does not exist, it is created.<\/p>\n
ALIASES<\/p>\n
Aliases allow a string to be substituted for a word when it is used
\nas the first word of a simple command.\u00a0 The shell maintains a list of
\naliases that may be set and unset with the alias and unalias builtin
\ncommands (see SHELL BUILTIN COMMANDS below).\u00a0 The first word of each
\nsimple command, if unquoted, is checked to see if it has an alias.
\nIf so, that word is replaced by the text of the alias.\u00a0 The
\ncharacters \/, $, `, and = and any of the shell metacharacters or
\nquoting characters listed above may not appear in an alias name.\u00a0 The
\nreplacement text may contain any valid shell input, including shell
\nmetacharacters.\u00a0 The first word of the replacement text is tested for
\naliases, but a word that is identical to an alias being expanded is
\nnot expanded a second time.\u00a0 This means that one may alias ls to ls
\n-F, for instance, and bash does not try to recursively expand the
\nreplacement text.\u00a0 If the last character of the alias value is a
\nblank, then the next command word following the alias is also checked
\nfor alias expansion.<\/p>\n
Aliases are created and listed with the alias command, and removed
\nwith the unalias command.<\/p>\n
There is no mechanism for using arguments in the replacement text.
\nIf arguments are needed, a shell function should be used (see
\nFUNCTIONS below).<\/p>\n
Aliases are not expanded when the shell is not interactive, unless
\nthe expand_aliases shell option is set using shopt (see the
\ndescription of shopt under SHELL BUILTIN COMMANDS below).<\/p>\n
The rules concerning the definition and use of aliases are somewhat
\nconfusing.\u00a0 Bash always reads at least one complete line of input
\nbefore executing any of the commands on that line.\u00a0 Aliases are
\nexpanded when a command is read, not when it is executed.\u00a0 Therefore,
\nan alias definition appearing on the same line as another command
\ndoes not take effect until the next line of input is read.\u00a0 The
\ncommands following the alias definition on that line are not affected
\nby the new alias.\u00a0 This behavior is also an issue when functions are
\nexecuted.\u00a0 Aliases are expanded when a function definition is read,
\nnot when the function is executed, because a function definition is
\nitself a compound command.\u00a0 As a consequence, aliases defined in a
\nfunction are not available until after that function is executed.\u00a0 To
\nbe safe, always put alias definitions on a separate line, and do not
\nuse alias in compound commands.<\/p>\n
For almost every purpose, aliases are superseded by shell functions.<\/p>\n
FUNCTIONS<\/p>\n
A shell function, defined as described above under SHELL GRAMMAR,
\nstores a series of commands for later execution.\u00a0 When the name of a
\nshell function is used as a simple command name, the list of commands
\nassociated with that function name is executed.\u00a0 Functions are
\nexecuted in the context of the current shell; no new process is
\ncreated to interpret them (contrast this with the execution of a
\nshell script).\u00a0 When a function is executed, the arguments to the
\nfunction become the positional parameters during its execution.\u00a0 The
\nspecial parameter # is updated to reflect the change.\u00a0 Special
\nparameter 0 is unchanged.\u00a0 The first element of the FUNCNAME variable
\nis set to the name of the function while the function is executing.<\/p>\n
All other aspects of the shell execution environment are identical
\nbetween a function and its caller with these exceptions:\u00a0 the DEBUG
\nand RETURN traps (see the description of the trap builtin under SHELL
\nBUILTIN COMMANDS below) are not inherited unless the function has
\nbeen given the trace attribute (see the description of the declare
\nbuiltin below) or the -o functrace shell option has been enabled with
\nthe set builtin (in which case all functions inherit the DEBUG and
\nRETURN traps), and the ERR trap is not inherited unless the -o
\nerrtrace shell option has been enabled.<\/p>\n
Variables local to the function may be declared with the local
\nbuiltin command.\u00a0 Ordinarily, variables and their values are shared
\nbetween the function and its caller.<\/p>\n
The FUNCNEST variable, if set to a numeric value greater than 0,
\ndefines a maximum function nesting level.\u00a0 Function invocations that
\nexceed the limit cause the entire command to abort.<\/p>\n
If the builtin command return is executed in a function, the function
\ncompletes and execution resumes with the next command after the
\nfunction call.\u00a0 Any command associated with the RETURN trap is
\nexecuted before execution resumes.\u00a0 When a function completes, the
\nvalues of the positional parameters and the special parameter # are
\nrestored to the values they had prior to the function’s execution.<\/p>\n
Function names and definitions may be listed with the -f option to
\nthe declare or typeset builtin commands.\u00a0 The -F option to declare or
\ntypeset will list the function names only (and optionally the source
\nfile and line number, if the extdebug shell option is enabled).
\nFunctions may be exported so that subshells automatically have them
\ndefined with the -f option to the export builtin.\u00a0 A function
\ndefinition may be deleted using the -f option to the unset builtin.
\nNote that shell functions and variables with the same name may result
\nin multiple identically-named entries in the environment passed to
\nthe shell’s children.\u00a0 Care should be taken in cases where this may
\ncause a problem.<\/p>\n
Functions may be recursive.\u00a0 The FUNCNEST variable may be used to
\nlimit the depth of the function call stack and restrict the number of
\nfunction invocations.\u00a0 By default, no limit is imposed on the number
\nof recursive calls.<\/p>\n
ARITHMETIC EVALUATION<\/p>\n
The shell allows arithmetic expressions to be evaluated, under
\ncertain circumstances (see the let and declare builtin commands and
\nArithmetic Expansion).\u00a0 Evaluation is done in fixed-width integers
\nwith no check for overflow, though division by 0 is trapped and
\nflagged as an error.\u00a0 The operators and their precedence,
\nassociativity, and values are the same as in the C language.\u00a0 The
\nfollowing list of operators is grouped into levels of equal-
\nprecedence operators.\u00a0 The levels are listed in order of decreasing
\nprecedence.<\/p>\n
id++ id–
\nvariable post-increment and post-decrement
\n++id –id
\nvariable pre-increment and pre-decrement
\n– +\u00a0\u00a0\u00a0 unary minus and plus
\n! ~\u00a0\u00a0\u00a0 logical and bitwise negation
\n**\u00a0\u00a0\u00a0\u00a0 exponentiation
\n* \/ %\u00a0 multiplication, division, remainder
\n+ –\u00a0\u00a0\u00a0 addition, subtraction
\n<< >>\u00a0 left and right bitwise shifts
\n<= >= < >
\ncomparison
\n== !=\u00a0 equality and inequality
\n&\u00a0\u00a0\u00a0\u00a0\u00a0 bitwise AND
\n^\u00a0\u00a0\u00a0\u00a0\u00a0 bitwise exclusive OR
\n|\u00a0\u00a0\u00a0\u00a0\u00a0 bitwise OR
\n&&\u00a0\u00a0\u00a0\u00a0 logical AND
\n||\u00a0\u00a0\u00a0\u00a0 logical OR
\nexpr?expr:expr
\nconditional operator
\n= *= \/= %= += -= <<= >>= &= ^= |=
\nassignment
\nexpr1 , expr2
\ncomma<\/p>\n
Shell variables are allowed as operands; parameter expansion is
\nperformed before the expression is evaluated.\u00a0 Within an expression,
\nshell variables may also be referenced by name without using the
\nparameter expansion syntax.\u00a0 A shell variable that is null or unset
\nevaluates to 0 when referenced by name without using the parameter
\nexpansion syntax.\u00a0 The value of a variable is evaluated as an
\narithmetic expression when it is referenced, or when a variable which
\nhas been given the integer attribute using declare -i is assigned a
\nvalue.\u00a0 A null value evaluates to 0.\u00a0 A shell variable need not have
\nits integer attribute turned on to be used in an expression.<\/p>\n
Constants with a leading 0 are interpreted as octal numbers.\u00a0 A
\nleading 0x or 0X denotes hexadecimal.\u00a0 Otherwise, numbers take the
\nform [base#]n, where the optional base is a decimal number between 2
\nand 64 representing the arithmetic base, and n is a number in that
\nbase.\u00a0 If base# is omitted, then base 10 is used.\u00a0 When specifying n,
\nthe digits greater< than 9 are represented by the lowercase letters,
\nthe uppercase letters, @, and _, in that order.\u00a0 If base is less than
\nor equal to 36, lowercase and uppercase letters may be used
\ninterchangeably to represent numbers between 10 and 35.<\/p>\n
Operators are evaluated in order of precedence.\u00a0 Sub-expressions in
\nparentheses are evaluated first and may override the precedence rules
\nabove.<\/p>\n
CONDITIONAL EXPRESSIONS<\/p>\n
Conditional expressions are used by the [[ compound command and the
\ntest and [ builtin commands to test file attributes and perform
\nstring and arithmetic comparisons.\u00a0 Expressions are formed from the
\nfollowing unary or binary primaries.\u00a0 If any file argument to one of
\nthe primaries is of the form \/dev\/fd\/n, then file descriptor n is
\nchecked.\u00a0 If the file argument to one of the primaries is one of
\n\/dev\/stdin, \/dev\/stdout, or \/dev\/stderr, file descriptor 0, 1, or 2,
\nrespectively, is checked.<\/p>\n
Unless otherwise specified, primaries that operate on files follow
\nsymbolic links and operate on the target of the link, rather than the
\nlink itself.<\/p>\n
When used with [[, the < and > operators sort lexicographically using
\nthe current locale.\u00a0 The test command sorts using ASCII ordering.<\/p>\n
-a file
\nTrue if file exists.
\n-b file
\nTrue if file exists and is a block special file.
\n-c file
\nTrue if file exists and is a character special file.
\n-d file
\nTrue if file exists and is a directory.
\n-e file
\nTrue if file exists.
\n-f file
\nTrue if file exists and is a regular file.
\n-g file
\nTrue if file exists and is set-group-id.
\n-h file
\nTrue if file exists and is a symbolic link.
\n-k file
\nTrue if file exists and its “sticky” bit is set.
\n-p file
\nTrue if file exists and is a named pipe (FIFO).
\n-r file
\nTrue if file exists and is readable.
\n-s file
\nTrue if file exists and has a size greater than zero.
\n-t fd\u00a0 True if file descriptor fd is open and refers to a terminal.
\n-u file
\nTrue if file exists and its set-user-id bit is set.
\n-w file
\nTrue if file exists and is writable.
\n-x file
\nTrue if file exists and is executable.
\n-G file
\nTrue if file exists and is owned by the effective group id.
\n-L file
\nTrue if file exists and is a symbolic link.
\n-N file
\nTrue if file exists and has been modified since it was last
\nread.
\n-O file
\nTrue if file exists and is owned by the effective user id.
\n-S file
\nTrue if file exists and is a socket.
\nfile1 -ef file2
\nTrue if file1 and file2 refer to the same device and inode
\nnumbers.
\nfile1 -nt file2
\nTrue if file1 is newer (according to modification date) than
\nfile2, or if file1 exists and file2 does not.
\nfile1 -ot file2
\nTrue if file1 is older than file2, or if file2 exists and
\nfile1 does not.
\n-o optname
\nTrue if the shell option optname is enabled.\u00a0 See the list of
\noptions under the description of the -o option to the set
\nbuiltin below.
\n-v varname
\nTrue if the shell variable varname is set (has been assigned a
\nvalue).
\n-R varname
\nTrue if the shell variable varname is set and is a name
\nreference.
\n-z string
\nTrue if the length of string is zero.
\nstring
\n-n string
\nTrue if the length of string is non-zero.<\/p>\n
string1 == string2
\nstring1 = string2
\nTrue if the strings are equal.\u00a0 = should be used with the test
\ncommand for POSIX conformance.\u00a0 When used with the [[ command,
\nthis performs pattern matching as described above (Compound
\nCommands).<\/p>\n
string1 != string2
\nTrue if the strings are not equal.<\/p>\n
string1 < string2
\nTrue if string1 sorts before string2 lexicographically.<\/p>\n
string1 > string2
\nTrue if string1 sorts after string2 lexicographically.<\/p>\n
arg1 OP arg2
\nOP is one of -eq, -ne, -lt, -le, -gt, or -ge.\u00a0 These
\narithmetic binary operators return true if arg1 is equal to,
\nnot equal to, less than, less than or equal to, greater than,
\nor greater than or equal to arg2, respectively.\u00a0 Arg1 and arg2
\nmay be positive or negative integers.<\/p>\n
SIMPLE COMMAND EXPANSION<\/p>\n
When a simple command is executed, the shell performs the following
\nexpansions, assignments, and redirections, from left to right.<\/p>\n
1.\u00a0\u00a0\u00a0\u00a0 The words that the parser has marked as variable assignments
\n(those preceding the command name) and redirections are saved
\nfor later processing.<\/p>\n
2.\u00a0\u00a0\u00a0\u00a0 The words that are not variable assignments or redirections
\nare expanded.\u00a0 If any words remain after expansion, the first
\nword is taken to be the name of the command and the remaining
\nwords are the arguments.<\/p>\n
3.\u00a0\u00a0\u00a0\u00a0 Redirections are performed as described above under
\nREDIRECTION.<\/p>\n
4.\u00a0\u00a0\u00a0\u00a0 The text after the = in each variable assignment undergoes
\ntilde expansion, parameter expansion, command substitution,
\narithmetic expansion, and quote removal before being assigned
\nto the variable.<\/p>\n
If no command name results, the variable assignments affect the
\ncurrent shell environment.\u00a0 Otherwise, the variables are added to the
\nenvironment of the executed command and do not affect the current
\nshell environment.\u00a0 If any of the assignments attempts to assign a
\nvalue to a readonly variable, an error occurs, and the command exits
\nwith a non-zero status.<\/p>\n
If no command name results, redirections are performed, but do not
\naffect the current shell environment.\u00a0 A redirection error causes the
\ncommand to exit with a non-zero status.<\/p>\n
If there is a command name left after expansion, execution proceeds
\nas described below.\u00a0 Otherwise, the command exits.\u00a0 If one of the
\nexpansions contained a command substitution, the exit status of the
\ncommand is the exit status of the last command substitution
\nperformed.\u00a0 If there were no command substitutions, the command exits
\nwith a status of zero.<\/p>\n
COMMAND EXECUTION<\/p>\n
After a command has been split into words, if it results in a simple
\ncommand and an optional list of arguments, the following actions are
\ntaken.<\/p>\n
If the command name contains no slashes, the shell attempts to locate
\nit.\u00a0 If there exists a shell function by that name, that function is
\ninvoked as described above in FUNCTIONS.\u00a0 If the name does not match
\na function, the shell searches for it in the list of shell builtins.
\nIf a match is found, that builtin is invoked.<\/p>\n
If the name is neither a shell function nor a builtin, and contains
\nno slashes, bash searches each element of the PATH for a directory
\ncontaining an executable file by that name.\u00a0 Bash uses a hash table
\nto remember the full pathnames of executable files (see hash under
\nSHELL BUILTIN COMMANDS below).\u00a0 A full search of the directories in
\nPATH is performed only if the command is not found in the hash table.
\nIf the search is unsuccessful, the shell searches for a defined shell
\nfunction named command_not_found_handle.\u00a0 If that function exists, it
\nis invoked with the original command and the original command’s
\narguments as its arguments, and the function’s exit status becomes
\nthe exit status of the shell.\u00a0 If that function is not defined, the
\nshell prints an error message and returns an exit status of 127.<\/p>\n
If the search is successful, or if the command name contains one or
\nmore slashes, the shell executes the named program in a separate
\nexecution environment.\u00a0 Argument 0 is set to the name given, and the
\nremaining arguments to the command are set to the arguments given, if
\nany.<\/p>\n
If this execution fails because the file is not in executable format,
\nand the file is not a directory, it is assumed to be a shell script,
\na file containing shell commands.\u00a0 A subshell is spawned to execute
\nit.\u00a0 This subshell reinitializes itself, so that the effect is as if
\na new shell had been invoked to handle the script, with the exception
\nthat the locations of commands remembered by the parent (see hash
\nbelow under SHELL BUILTIN COMMANDS) are retained by the child.<\/p>\n
If the program is a file beginning with #!, the remainder of the
\nfirst line specifies an interpreter for the program.\u00a0 The shell
\nexecutes the specified interpreter on operating systems that do not
\nhandle this executable format themselves.\u00a0 The arguments to the
\ninterpreter consist of a single optional argument following the
\ninterpreter name on the first line of the program, followed by the
\nname of the program, followed by the command arguments, if any.<\/p>\n
COMMAND EXECUTION ENVIRONMENT<\/p>\n
The shell has an execution environment, which consists of the
\nfollowing:<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 open files inherited by the shell at invocation, as modified
\nby redirections supplied to the exec builtin<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 the current working directory as set by cd, pushd, or popd, or
\ninherited by the shell at invocation<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 the file creation mode mask as set by umask or inherited from
\nthe shell’s parent<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 current traps set by trap<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 shell parameters that are set by variable assignment or with
\nset or inherited from the shell’s parent in the environment<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 shell functions defined during execution or inherited from the
\nshell’s parent in the environment<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 options enabled at invocation (either by default or with
\ncommand-line arguments) or by set<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 options enabled by shopt<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 shell aliases defined with alias<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 various process IDs, including those of background jobs, the
\nvalue of $$, and the value of PPID<\/p>\n
When a simple command other than a builtin or shell function is to be
\nexecuted, it is invoked in a separate execution environment that
\nconsists of the following.\u00a0 Unless otherwise noted, the values are
\ninherited from the shell.<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 the shell’s open files, plus any modifications and additions
\nspecified by redirections to the command<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 the current working directory<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 the file creation mode mask<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 shell variables and functions marked for export, along with
\nvariables exported for the command, passed in the environment<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 traps caught by the shell are reset to the values inherited
\nfrom the shell’s parent, and traps ignored by the shell are
\nignored<\/p>\n
A command invoked in this separate environment cannot affect the
\nshell’s execution environment.<\/p>\n
Command substitution, commands grouped with parentheses, and
\nasynchronous commands are invoked in a subshell environment that is a
\nduplicate of the shell environment, except that traps caught by the
\nshell are reset to the values that the shell inherited from its
\nparent at invocation.\u00a0 Builtin commands that are invoked as part of a
\npipeline are also executed in a subshell environment.\u00a0 Changes made
\nto the subshell environment cannot affect the shell’s execution
\nenvironment.<\/p>\n
Subshells spawned to execute command substitutions inherit the value
\nof the -e option from the parent shell.\u00a0 When not in posix mode, bash
\nclears the -e option in such subshells.<\/p>\n
If a command is followed by a & and job control is not active, the
\ndefault standard input for the command is the empty file \/dev\/null.
\nOtherwise, the invoked command inherits the file descriptors of the
\ncalling shell as modified by redirections.<\/p>\n
ENVIRONMENT<\/p>\n
When a program is invoked it is given an array of strings called the
\nenvironment.\u00a0 This is a list of name-value pairs, of the form
\nname=value.<\/p>\n
The shell provides several ways to manipulate the environment.\u00a0 On
\ninvocation, the shell scans its own environment and creates a
\nparameter for each name found, automatically marking it for export to
\nchild processes.\u00a0 Executed commands inherit the environment.\u00a0 The
\nexport and declare -x commands allow parameters and functions to be
\nadded to and deleted from the environment.\u00a0 If the value of a
\nparameter in the environment is modified, the new value becomes part
\nof the environment, replacing the old.\u00a0 The environment inherited by
\nany executed command consists of the shell’s initial environment,
\nwhose values may be modified in the shell, less any pairs removed by
\nthe unset command, plus any additions via the export and declare -x
\ncommands.<\/p>\n
The environment for any simple command or function may be augmented
\ntemporarily by prefixing it with parameter assignments, as described
\nabove in PARAMETERS.\u00a0 These assignment statements affect only the
\nenvironment seen by that command.<\/p>\n
If the -k option is set (see the set builtin command below), then all
\nparameter assignments are placed in the environment for a command,
\nnot just those that precede the command name.<\/p>\n
When bash invokes an external command, the variable _ is set to the
\nfull filename of the command and passed to that command in its
\nenvironment.<\/p>\n
EXIT STATUS<\/p>\n
The exit status of an executed command is the value returned by the
\nwaitpid system call or equivalent function.\u00a0 Exit statuses fall
\nbetween 0 and 255, though, as explained below, the shell may use
\nvalues above 125 specially.\u00a0 Exit statuses from shell builtins and
\ncompound commands are also limited to this range. Under certain
\ncircumstances, the shell will use special values to indicate specific
\nfailure modes.<\/p>\n
For the shell’s purposes, a command which exits with a zero exit
\nstatus has succeeded.\u00a0 An exit status of zero indicates success.\u00a0 A
\nnon-zero exit status indicates failure.\u00a0 When a command terminates on
\na fatal signal N, bash uses the value of 128+N as the exit status.<\/p>\n
If a command is not found, the child process created to execute it
\nreturns a status of 127.\u00a0 If a command is found but is not
\nexecutable, the return status is 126.<\/p>\n
If a command fails because of an error during expansion or
\nredirection, the exit status is greater than zero.<\/p>\n
Shell builtin commands return a status of 0 (true) if successful, and
\nnon-zero (false) if an error occurs while they execute.\u00a0 All builtins
\nreturn an exit status of 2 to indicate incorrect usage.<\/p>\n
Bash itself returns the exit status of the last command executed,
\nunless a syntax error occurs, in which case it exits with a non-zero
\nvalue.\u00a0 See also the exit builtin command below.<\/p>\n
SIGNALS<\/p>\n
When bash is interactive, in the absence of any traps, it ignores
\nSIGTERM (so that kill 0 does not kill an interactive shell), and
\nSIGINT is caught and handled (so that the wait builtin is
\ninterruptible).\u00a0 In all cases, bash ignores SIGQUIT.\u00a0 If job control
\nis in effect, bash ignores SIGTTIN, SIGTTOU, and SIGTSTP.<\/p>\n
Non-builtin commands run by bash have signal handlers set to the
\nvalues inherited by the shell from its parent.\u00a0 When job control is
\nnot in effect, asynchronous commands ignore SIGINT and SIGQUIT in
\naddition to these inherited handlers.\u00a0 Commands run as a result of
\ncommand substitution ignore the keyboard-generated job control
\nsignals SIGTTIN, SIGTTOU, and SIGTSTP.<\/p>\n
The shell exits by default upon receipt of a SIGHUP.\u00a0 Before exiting,
\nan interactive shell resends the SIGHUP to all jobs, running or
\nstopped.\u00a0 Stopped jobs are sent SIGCONT to ensure that they receive
\nthe SIGHUP.\u00a0 To prevent the shell from sending the signal to a
\nparticular job, it should be removed from the jobs table with the
\ndisown builtin (see SHELL BUILTIN COMMANDS below) or marked to not
\nreceive SIGHUP using disown -h.<\/p>\n
If the huponexit shell option has been set with shopt, bash sends a
\nSIGHUP to all jobs when an interactive login shell exits.<\/p>\n
If bash is waiting for a command to complete and receives a signal
\nfor which a trap has been set, the trap will not be executed until
\nthe command completes.\u00a0 When bash is waiting for an asynchronous
\ncommand via the wait builtin, the reception of a signal for which a
\ntrap has been set will cause the wait builtin to return immediately
\nwith an exit status greater than 128, immediately after which the
\ntrap is executed.<\/p>\n
JOB CONTROL<\/p>\n
Job control refers to the ability to selectively stop (suspend) the
\nexecution of processes and continue (resume) their execution at a
\nlater point.\u00a0 A user typically employs this facility via an
\ninteractive interface supplied jointly by the operating system
\nkernel’s terminal driver and bash.<\/p>\n
The shell associates a job with each pipeline.\u00a0 It keeps a table of
\ncurrently executing jobs, which may be listed with the jobs command.
\nWhen bash starts a job asynchronously (in the background), it prints
\na line that looks like:<\/p>\n
[1] 25647<\/p>\n
indicating that this job is job number 1 and that the process ID of
\nthe last process in the pipeline associated with this job is 25647.
\nAll of the processes in a single pipeline are members of the same
\njob.\u00a0 Bash uses the job abstraction as the basis for job control.<\/p>\n
To facilitate the implementation of the user interface to job
\ncontrol, the operating system maintains the notion of a current
\nterminal process group ID.\u00a0 Members of this process group (processes
\nwhose process group ID is equal to the current terminal process group
\nID) receive keyboard-generated signals such as SIGINT.\u00a0 These
\nprocesses are said to be in the foreground.\u00a0 Background processes are
\nthose whose process group ID differs from the terminal’s; such
\nprocesses are immune to keyboard-generated signals.\u00a0 Only foreground
\nprocesses are allowed to read from or, if the user so specifies with
\nstty tostop, write to the terminal.\u00a0 Background processes which
\nattempt to read from (write to when stty tostop is in effect) the
\nterminal are sent a SIGTTIN (SIGTTOU) signal by the kernel’s terminal
\ndriver, which, unless caught, suspends the process.<\/p>\n
If the operating system on which bash is running supports job
\ncontrol, bash contains facilities to use it.\u00a0 Typing the suspend
\ncharacter (typically ^Z, Control-Z) while a process is running causes
\nthat process to be stopped and returns control to bash.\u00a0 Typing the
\ndelayed suspend character (typically ^Y, Control-Y) causes the
\nprocess to be stopped when it attempts to read input from the
\nterminal, and control to be returned to bash.\u00a0 The user may then
\nmanipulate the state of this job, using the bg command to continue it
\nin the background, the fg command to continue it in the foreground,
\nor the kill command to kill it.\u00a0 A ^Z takes effect immediately, and
\nhas the additional side effect of causing pending output and
\ntypeahead to be discarded.<\/p>\n
There are a number of ways to refer to a job in the shell.\u00a0 The
\ncharacter % introduces a job specification (jobspec).\u00a0 Job number n
\nmay be referred to as %n.\u00a0 A job may also be referred to using a
\nprefix of the name used to start it, or using a substring that
\nappears in its command line.\u00a0 For example, %ce refers to a stopped ce
\njob.\u00a0 If a prefix matches more than one job, bash reports an error.
\nUsing %?ce, on the other hand, refers to any job containing the
\nstring ce in its command line.\u00a0 If the substring matches more than
\none job, bash reports an error.\u00a0 The symbols %% and %+ refer to the
\nshell’s notion of the current job, which is the last job stopped
\nwhile it was in the foreground or started in the background.\u00a0 The
\nprevious job may be referenced using %-.\u00a0 If there is only a single
\njob, %+ and %- can both be used to refer to that job.\u00a0 In output
\npertaining to jobs (e.g., the output of the jobs command), the
\ncurrent job is always flagged with a +, and the previous job with a
\n-.\u00a0 A single % (with no accompanying job specification) also refers
\nto the current job.<\/p>\n
Simply naming a job can be used to bring it into the foreground: %1
\nis a synonym for “fg %1”, bringing job 1 from the background into
\nthe foreground.\u00a0 Similarly, “%1 &” resumes job 1 in the background,
\nequivalent to “bg %1”.<\/p>\n
The shell learns immediately whenever a job changes state.\u00a0 Normally,
\nbash waits until it is about to print a prompt before reporting
\nchanges in a job’s status so as to not interrupt any other output.
\nIf the -b option to the set builtin command is enabled, bash reports
\nsuch changes immediately.\u00a0 Any trap on SIGCHLD is executed for each
\nchild that exits.<\/p>\n
If an attempt to exit bash is made while jobs are stopped (or, if the
\ncheckjobs shell option has been enabled using the shopt builtin,
\nrunning), the shell prints a warning message, and, if the checkjobs
\noption is enabled, lists the jobs and their statuses.\u00a0 The jobs
\ncommand may then be used to inspect their status.\u00a0 If a second
\nattempt to exit is made without an intervening command, the shell
\ndoes not print another warning, and any stopped jobs are terminated.<\/p>\n
PROMPTING<\/p>\n
When executing interactively, bash displays the primary prompt PS1
\nwhen it is ready to read a command, and the secondary prompt PS2 when
\nit needs more input to complete a command.\u00a0 Bash allows these prompt
\nstrings to be customized by inserting a number of backslash-escaped
\nspecial characters that are decoded as follows:
\n\\a\u00a0\u00a0\u00a0\u00a0 an ASCII bell character (07)
\n\\d\u00a0\u00a0\u00a0\u00a0 the date in “Weekday Month Date” format (e.g., “Tue May
\n26”)
\n\\D{format}
\nthe format is passed to strftime(3) and the result is
\ninserted into the prompt string; an empty format
\nresults in a locale-specific time representation.\u00a0 The
\nbraces are required
\n\\e\u00a0\u00a0\u00a0\u00a0 an ASCII escape character (033)
\n\\h\u00a0\u00a0\u00a0\u00a0 the hostname up to the first `.’
\n\\H\u00a0\u00a0\u00a0\u00a0 the hostname
\n\\j\u00a0\u00a0\u00a0\u00a0 the number of jobs currently managed by the shell
\n\\l\u00a0\u00a0\u00a0\u00a0 the basename of the shell’s terminal device name
\n\\n\u00a0\u00a0\u00a0\u00a0 newline
\n\\r\u00a0\u00a0\u00a0\u00a0 carriage return
\n\\s\u00a0\u00a0\u00a0\u00a0 the name of the shell, the basename of $0 (the portion
\nfollowing the final slash)
\n\\t\u00a0\u00a0\u00a0\u00a0 the current time in 24-hour HH:MM:SS format
\n\\T\u00a0\u00a0\u00a0\u00a0 the current time in 12-hour HH:MM:SS format
\n\\@\u00a0\u00a0\u00a0\u00a0 the current time in 12-hour am\/pm format
\n\\A\u00a0\u00a0\u00a0\u00a0 the current time in 24-hour HH:MM format
\n\\u\u00a0\u00a0\u00a0\u00a0 the username of the current user
\n\\v\u00a0\u00a0\u00a0\u00a0 the version of bash (e.g., 2.00)
\n\\V\u00a0\u00a0\u00a0\u00a0 the release of bash, version + patch level (e.g.,
\n2.00.0)
\n\\w\u00a0\u00a0\u00a0\u00a0 the current working directory, with $HOME abbreviated
\nwith a tilde (uses the value of the PROMPT_DIRTRIM
\nvariable)
\n\\W\u00a0\u00a0\u00a0\u00a0 the basename of the current working directory, with
\n$HOME abbreviated with a tilde
\n\\!\u00a0\u00a0\u00a0\u00a0 the history number of this command
\n\\#\u00a0\u00a0\u00a0\u00a0 the command number of this command
\n\\$\u00a0\u00a0\u00a0\u00a0 if the effective UID is 0, a #, otherwise a $
\n\\nnn\u00a0\u00a0 the character corresponding to the octal number nnn
\n\\\\\u00a0\u00a0\u00a0\u00a0 a backslash
\n\\[\u00a0\u00a0\u00a0\u00a0 begin a sequence of non-printing characters, which
\ncould be used to embed a terminal control sequence into
\nthe prompt
\n\\]\u00a0\u00a0\u00a0\u00a0 end a sequence of non-printing characters<\/p>\n
The command number and the history number are usually different: the
\nhistory number of a command is its position in the history list,
\nwhich may include commands restored from the history file (see
\nHISTORY below), while the command number is the position in the
\nsequence of commands executed during the current shell session.
\nAfter the string is decoded, it is expanded via parameter expansion,
\ncommand substitution, arithmetic expansion, and quote removal,
\nsubject to the value of the promptvars shell option (see the
\ndescription of the shopt command under SHELL BUILTIN COMMANDS below).<\/p>\n
READLINE<\/p>\n
This is the library that handles reading input when using an
\ninteractive shell, unless the –noediting option is given at shell
\ninvocation.\u00a0 Line editing is also used when using the -e option to
\nthe read builtin.\u00a0 By default, the line editing commands are similar
\nto those of Emacs.\u00a0 A vi-style line editing interface is also
\navailable.\u00a0 Line editing can be enabled at any time using the -o
\nemacs or -o vi options to the set builtin (see SHELL BUILTIN COMMANDS
\nbelow).\u00a0 To turn off line editing after the shell is running, use the
\n+o emacs or +o vi options to the set builtin.<\/p>\n
Readline Notation
\nIn this section, the Emacs-style notation is used to denote
\nkeystrokes.\u00a0 Control keys are denoted by C-key, e.g., C-n means
\nControl-N.\u00a0 Similarly, meta keys are denoted by M-key, so M-x means
\nMeta-X.\u00a0 (On keyboards without a meta key, M-x means ESC x, i.e.,
\npress the Escape key then the x key.\u00a0 This makes ESC the meta prefix.
\nThe combination M-C-x means ESC-Control-x, or press the Escape key
\nthen hold the Control key while pressing the x key.)<\/p>\n
Readline commands may be given numeric arguments, which normally act
\nas a repeat count.\u00a0 Sometimes, however, it is the sign of the
\nargument that is significant.\u00a0 Passing a negative argument to a
\ncommand that acts in the forward direction (e.g., kill-line) causes
\nthat command to act in a backward direction.\u00a0 Commands whose behavior
\nwith arguments deviates from this are noted below.<\/p>\n
When a command is described as killing text, the text deleted is
\nsaved for possible future retrieval (yanking).\u00a0 The killed text is
\nsaved in a kill ring.\u00a0 Consecutive kills cause the text to be
\naccumulated into one unit, which can be yanked all at once.\u00a0 Commands
\nwhich do not kill text separate the chunks of text on the kill ring.<\/p>\n
Readline Initialization
\nReadline is customized by putting commands in an initialization file
\n(the inputrc file).\u00a0 The name of this file is taken from the value of
\nthe INPUTRC variable.\u00a0 If that variable is unset, the default is
\n~\/.inputrc.\u00a0 When a program which uses the readline library starts
\nup, the initialization file is read, and the key bindings and
\nvariables are set.\u00a0 There are only a few basic constructs allowed in
\nthe readline initialization file.\u00a0 Blank lines are ignored.\u00a0 Lines
\nbeginning with a # are comments.\u00a0 Lines beginning with a $ indicate
\nconditional constructs.\u00a0 Other lines denote key bindings and variable
\nsettings.<\/p>\n
The default key-bindings may be changed with an inputrc file.\u00a0 Other
\nprograms that use this library may add their own commands and
\nbindings.<\/p>\n
For example, placing<\/p>\n
M-Control-u: universal-argument
\nor
\nC-Meta-u: universal-argument
\ninto the inputrc would make M-C-u execute the readline command
\nuniversal-argument.<\/p>\n
The following symbolic character names are recognized: RUBOUT, DEL,
\nESC, LFD, NEWLINE, RET, RETURN, SPC, SPACE, and TAB.<\/p>\n
In addition to command names, readline allows keys to be bound to a
\nstring that is inserted when the key is pressed (a macro).<\/p>\n
Readline Key Bindings
\nThe syntax for controlling key bindings in the inputrc file is
\nsimple.\u00a0 All that is required is the name of the command or the text
\nof a macro and a key sequence to which it should be bound. The name
\nmay be specified in one of two ways: as a symbolic key name, possibly
\nwith Meta- or Control- prefixes, or as a key sequence.<\/p>\n
When using the form keyname:function-name or macro, keyname is the
\nname of a key spelled out in English.\u00a0 For example:<\/p>\n
Control-u: universal-argument
\nMeta-Rubout: backward-kill-word
\nControl-o: “> output”<\/p>\n
In the above example, C-u is bound to the function
\nuniversal-argument, M-DEL is bound to the function
\nbackward-kill-word, and C-o is bound to run the macro expressed on
\nthe right hand side (that is, to insert the text “> output” into
\nthe line).<\/p>\n
In the second form, “keyseq”:function-name or macro, keyseq differs
\nfrom keyname above in that strings denoting an entire key sequence
\nmay be specified by placing the sequence within double quotes.\u00a0 Some
\nGNU Emacs style key escapes can be used, as in the following example,
\nbut the symbolic character names are not recognized.<\/p>\n
“\\C-u”: universal-argument
\n“\\C-x\\C-r”: re-read-init-file
\n“\\e[11~”: “Function Key 1″<\/p>\n
In this example, C-u is again bound to the function
\nuniversal-argument.\u00a0 C-x C-r is bound to the function
\nre-read-init-file, and ESC [ 1 1 ~ is bound to insert the text
\n“Function Key 1”.<\/p>\n
The full set of GNU Emacs style escape sequences is
\n\\C-\u00a0\u00a0\u00a0 control prefix
\n\\M-\u00a0\u00a0\u00a0 meta prefix
\n\\e\u00a0\u00a0\u00a0\u00a0 an escape character
\n\\\\\u00a0\u00a0\u00a0\u00a0 backslash
\n\\”\u00a0\u00a0\u00a0\u00a0 literal ”
\n\\’\u00a0\u00a0\u00a0\u00a0 literal ‘<\/p>\n
In addition to the GNU Emacs style escape sequences, a second set of
\nbackslash escapes is available:
\n\\a\u00a0\u00a0\u00a0\u00a0 alert (bell)
\n\\b\u00a0\u00a0\u00a0\u00a0 backspace
\n\\d\u00a0\u00a0\u00a0\u00a0 delete
\n\\f\u00a0\u00a0\u00a0\u00a0 form feed
\n\\n\u00a0\u00a0\u00a0\u00a0 newline
\n\\r\u00a0\u00a0\u00a0\u00a0 carriage return
\n\\t\u00a0\u00a0\u00a0\u00a0 horizontal tab
\n\\v\u00a0\u00a0\u00a0\u00a0 vertical tab
\n\\nnn\u00a0\u00a0 the eight-bit character whose value is the octal value
\nnnn (one to three digits)
\n\\xHH\u00a0\u00a0 the eight-bit character whose value is the hexadecimal
\nvalue HH (one or two hex digits)<\/p>\n
When entering the text of a macro, single or double quotes must be
\nused to indicate a macro definition.\u00a0 Unquoted text is assumed to be
\na function name.\u00a0 In the macro body, the backslash escapes described
\nabove are expanded.\u00a0 Backslash will quote any other character in the
\nmacro text, including ” and ‘.<\/p>\n
Bash allows the current readline key bindings to be displayed or
\nmodified with the bind builtin command.\u00a0 The editing mode may be
\nswitched during interactive use by using the -o option to the set
\nbuiltin command (see SHELL BUILTIN COMMANDS below).<\/p>\n
Readline Variables
\nReadline has variables that can be used to further customize its
\nbehavior.\u00a0 A variable may be set in the inputrc file with a statement
\nof the form<\/p>\n
set variable-name value<\/p>\n
Except where noted, readline variables can take the values On or Off
\n(without regard to case).\u00a0 Unrecognized variable names are ignored.
\nWhen a variable value is read, empty or null values, “on” (case-
\ninsensitive), and “1” are equivalent to On.\u00a0 All other values are
\nequivalent to Off.\u00a0 The variables and their default values are:<\/p>\n
bell-style (audible)
\nControls what happens when readline wants to ring the terminal
\nbell.\u00a0 If set to none, readline never rings the bell.\u00a0 If set
\nto visible, readline uses a visible bell if one is available.
\nIf set to audible, readline attempts to ring the terminal’s
\nbell.
\nbind-tty-special-chars (On)
\nIf set to On, readline attempts to bind the control characters
\ntreated specially by the kernel’s terminal driver to their
\nreadline equivalents.
\ncolored-stats (Off)
\nIf set to On, readline displays possible completions using
\ndifferent colors to indicate their file type.\u00a0 The color
\ndefinitions are taken from the value of the LS_COLORS
\nenvironment variable.
\ncomment-begin (“#”)
\nThe string that is inserted when the readline insert-comment
\ncommand is executed.\u00a0 This command is bound to M-# in emacs
\nmode and to # in vi command mode.
\ncompletion-ignore-case (Off)
\nIf set to On, readline performs filename matching and
\ncompletion in a case-insensitive fashion.
\ncompletion-prefix-display-length (0)
\nThe length in characters of the common prefix of a list of
\npossible completions that is displayed without modification.
\nWhen set to a value greater than zero, common prefixes longer
\nthan this value are replaced with an ellipsis when displaying
\npossible completions.
\ncompletion-query-items (100)
\nThis determines when the user is queried about viewing the
\nnumber of possible completions generated by the
\npossible-completions command.\u00a0 It may be set to any integer
\nvalue greater than or equal to zero.\u00a0 If the number of
\npossible completions is greater than or equal to the value of
\nthis variable, the user is asked whether or not he wishes to
\nview them; otherwise they are simply listed on the terminal.
\nconvert-meta (On)
\nIf set to On, readline will convert characters with the eighth
\nbit set to an ASCII key sequence by stripping the eighth bit
\nand prefixing an escape character (in effect, using escape as
\nthe meta prefix).
\ndisable-completion (Off)
\nIf set to On, readline will inhibit word completion.
\nCompletion characters will be inserted into the line as if
\nthey had been mapped to self-insert.
\nediting-mode (emacs)
\nControls whether readline begins with a set of key bindings
\nsimilar to Emacs or vi.\u00a0 editing-mode can be set to either
\nemacs or vi.
\necho-control-characters (On)
\nWhen set to On, on operating systems that indicate they
\nsupport it, readline echoes a character corresponding to a
\nsignal generated from the keyboard.
\nenable-keypad (Off)
\nWhen set to On, readline will try to enable the application
\nkeypad when it is called.\u00a0 Some systems need this to enable
\nthe arrow keys.
\nenable-meta-key (On)
\nWhen set to On, readline will try to enable any meta modifier
\nkey the terminal claims to support when it is called.\u00a0 On many
\nterminals, the meta key is used to send eight-bit characters.
\nexpand-tilde (Off)
\nIf set to On, tilde expansion is performed when readline
\nattempts word completion.
\nhistory-preserve-point (Off)
\nIf set to On, the history code attempts to place point at the
\nsame location on each history line retrieved with previous-
\nhistory or next-history.
\nhistory-size (0)
\nSet the maximum number of history entries saved in the history
\nlist.\u00a0 If set to zero, any existing history entries are
\ndeleted and no new entries are saved.\u00a0 If set to a value less
\nthan zero, the number of history entries is not limited.\u00a0 By
\ndefault, the number of history entries is not limited.
\nhorizontal-scroll-mode (Off)
\nWhen set to On, makes readline use a single line for display,
\nscrolling the input horizontally on a single screen line when
\nit becomes longer than the screen width rather than wrapping
\nto a new line.
\ninput-meta (Off)
\nIf set to On, readline will enable eight-bit input (that is,
\nit will not strip the high bit from the characters it reads),
\nregardless of what the terminal claims it can support.\u00a0 The
\nname meta-flag is a synonym for this variable.
\nisearch-terminators (“C-[C-J”)
\nThe string of characters that should terminate an incremental
\nsearch without subsequently executing the character as a
\ncommand.\u00a0 If this variable has not been given a value, the
\ncharacters ESC and C-J will terminate an incremental search.
\nkeymap (emacs)
\nSet the current readline keymap.\u00a0 The set of valid keymap
\nnames is emacs, emacs-standard, emacs-meta, emacs-ctlx, vi,
\nvi-command, and vi-insert.\u00a0 vi is equivalent to vi-command;
\nemacs is equivalent to emacs-standard.\u00a0 The default value is
\nemacs; the value of editing-mode also affects the default
\nkeymap.
\nkeyseq-timeout (500)
\nSpecifies the duration readline will wait for a character when
\nreading an ambiguous key sequence (one that can form a
\ncomplete key sequence using the input read so far, or can take
\nadditional input to complete a longer key sequence).\u00a0 If no
\ninput is received within the timeout, readline will use the
\nshorter but complete key sequence.\u00a0 The value is specified in
\nmilliseconds, so a value of 1000 means that readline will wait
\none second for additional input.\u00a0 If this variable is set to a
\nvalue less than or equal to zero, or to a non-numeric value,
\nreadline will wait until another key is pressed to decide
\nwhich key sequence to complete.
\nmark-directories (On)
\nIf set to On, completed directory names have a slash appended.
\nmark-modified-lines (Off)
\nIf set to On, history lines that have been modified are
\ndisplayed with a preceding asterisk (*).
\nmark-symlinked-directories (Off)
\nIf set to On, completed names which are symbolic links to
\ndirectories have a slash appended (subject to the value of
\nmark-directories).
\nmatch-hidden-files (On)
\nThis variable, when set to On, causes readline to match files
\nwhose names begin with a `.’ (hidden files) when performing
\nfilename completion.\u00a0 If set to Off, the leading `.’ must be
\nsupplied by the user in the filename to be completed.
\nmenu-complete-display-prefix (Off)
\nIf set to On, menu completion displays the common prefix of
\nthe list of possible completions (which may be empty) before
\ncycling through the list.
\noutput-meta (Off)
\nIf set to On, readline will display characters with the eighth
\nbit set directly rather than as a meta-prefixed escape
\nsequence.
\npage-completions (On)
\nIf set to On, readline uses an internal more-like pager to
\ndisplay a screenful of possible completions at a time.
\nprint-completions-horizontally (Off)
\nIf set to On, readline will display completions with matches
\nsorted horizontally in alphabetical order, rather than down
\nthe screen.
\nrevert-all-at-newline (Off)
\nIf set to On, readline will undo all changes to history lines
\nbefore returning when accept-line is executed.\u00a0 By default,
\nhistory lines may be modified and retain individual undo lists
\nacross calls to readline.
\nshow-all-if-ambiguous (Off)
\nThis alters the default behavior of the completion functions.
\nIf set to On, words which have more than one possible
\ncompletion cause the matches to be listed immediately instead
\nof ringing the bell.
\nshow-all-if-unmodified (Off)
\nThis alters the default behavior of the completion functions
\nin a fashion similar to show-all-if-ambiguous.\u00a0 If set to On,
\nwords which have more than one possible completion without any
\npossible partial completion (the possible completions don’t
\nshare a common prefix) cause the matches to be listed
\nimmediately instead of ringing the bell.
\nshow-mode-in-prompt (Off)
\nIf set to On, add a character to the beginning of the prompt
\nindicating the editing mode: emacs (@), vi command (:) or vi
\ninsertion (+).
\nskip-completed-text (Off)
\nIf set to On, this alters the default completion behavior when
\ninserting a single match into the line.\u00a0 It’s only active when
\nperforming completion in the middle of a word.\u00a0 If enabled,
\nreadline does not insert characters from the completion that
\nmatch characters after point in the word being completed, so
\nportions of the word following the cursor are not duplicated.
\nvisible-stats (Off)
\nIf set to On, a character denoting a file’s type as reported
\nby stat(2) is appended to the filename when listing possible
\ncompletions.<\/p>\n
Readline Conditional Constructs
\nReadline implements a facility similar in spirit to the conditional
\ncompilation features of the C preprocessor which allows key bindings
\nand variable settings to be performed as the result of tests.\u00a0 There
\nare four parser directives used.<\/p>\n
$if\u00a0\u00a0\u00a0 The $if construct allows bindings to be made based on the
\nediting mode, the terminal being used, or the application
\nusing readline.\u00a0 The text of the test extends to the end of
\nthe line; no characters are required to isolate it.<\/p>\n
mode\u00a0\u00a0 The mode= form of the $if directive is used to test
\nwhether readline is in emacs or vi mode.\u00a0 This may be
\nused in conjunction with the set keymap command, for
\ninstance, to set bindings in the emacs-standard and
\nemacs-ctlx keymaps only if readline is starting out in
\nemacs mode.<\/p>\n
term\u00a0\u00a0 The term= form may be used to include terminal-specific
\nkey bindings, perhaps to bind the key sequences output
\nby the terminal’s function keys.\u00a0 The word on the right
\nside of the = is tested against the both full name of
\nthe terminal and the portion of the terminal name
\nbefore the first -.\u00a0 This allows sun to match both sun
\nand sun-cmd, for instance.<\/p>\n
application
\nThe application construct is used to include
\napplication-specific settings.\u00a0 Each program using the
\nreadline library sets the application name, and an
\ninitialization file can test for a particular value.
\nThis could be used to bind key sequences to functions
\nuseful for a specific program.\u00a0 For instance, the
\nfollowing command adds a key sequence that quotes the
\ncurrent or previous word in bash:<\/p>\n
$if Bash
\n# Quote the current or previous word
\n“\\C-xq”: “\\eb\\”\\ef\\””
\n$endif<\/p>\n
$endif This command, as seen in the previous example, terminates an
\n$if command.<\/p>\n
$else\u00a0 Commands in this branch of the $if directive are executed if
\nthe test fails.<\/p>\n
$include
\nThis directive takes a single filename as an argument and
\nreads commands and bindings from that file.\u00a0 For example, the
\nfollowing directive would read \/etc\/inputrc:<\/p>\n
$include\u00a0 \/etc\/inputrc<\/p>\n
Searching
\nReadline provides commands for searching through the command history
\n(see HISTORY below) for lines containing a specified string.\u00a0 There
\nare two search modes: incremental and non-incremental.<\/p>\n
Incremental searches begin before the user has finished typing the
\nsearch string.\u00a0 As each character of the search string is typed,
\nreadline displays the next entry from the history matching the string
\ntyped so far.\u00a0 An incremental search requires only as many characters
\nas needed to find the desired history entry.\u00a0 The characters present
\nin the value of the isearch-terminators variable are used to
\nterminate an incremental search.\u00a0 If that variable has not been
\nassigned a value the Escape and Control-J characters will terminate
\nan incremental search.\u00a0 Control-G will abort an incremental search
\nand restore the original line.\u00a0 When the search is terminated, the
\nhistory entry containing the search string becomes the current line.<\/p>\n
To find other matching entries in the history list, type Control-S or
\nControl-R as appropriate.\u00a0 This will search backward or forward in
\nthe history for the next entry matching the search string typed so
\nfar.\u00a0 Any other key sequence bound to a readline command will
\nterminate the search and execute that command.\u00a0 For instance, a
\nnewline will terminate the search and accept the line, thereby
\nexecuting the command from the history list.<\/p>\n
Readline remembers the last incremental search string.\u00a0 If two
\nControl-Rs are typed without any intervening characters defining a
\nnew search string, any remembered search string is used.<\/p>\n
Non-incremental searches read the entire search string before
\nstarting to search for matching history lines.\u00a0 The search string may
\nbe typed by the user or be part of the contents of the current line.<\/p>\n
Readline Command Names
\nThe following is a list of the names of the commands and the default
\nkey sequences to which they are bound.\u00a0 Command names without an
\naccompanying key sequence are unbound by default.\u00a0 In the following
\ndescriptions, point refers to the current cursor position, and mark
\nrefers to a cursor position saved by the set-mark command.\u00a0 The text
\nbetween the point and mark is referred to as the region.<\/p>\n
Commands for Moving
\nbeginning-of-line (C-a)
\nMove to the start of the current line.
\nend-of-line (C-e)
\nMove to the end of the line.
\nforward-char (C-f)
\nMove forward a character.
\nbackward-char (C-b)
\nMove back a character.
\nforward-word (M-f)
\nMove forward to the end of the next word.\u00a0 Words are composed
\nof alphanumeric characters (letters and digits).
\nbackward-word (M-b)
\nMove back to the start of the current or previous word.\u00a0 Words
\nare composed of alphanumeric characters (letters and digits).
\nshell-forward-word
\nMove forward to the end of the next word.\u00a0 Words are delimited
\nby non-quoted shell metacharacters.
\nshell-backward-word
\nMove back to the start of the current or previous word.\u00a0 Words
\nare delimited by non-quoted shell metacharacters.
\nclear-screen (C-l)
\nClear the screen leaving the current line at the top of the
\nscreen.\u00a0 With an argument, refresh the current line without
\nclearing the screen.
\nredraw-current-line
\nRefresh the current line.<\/p>\n
Commands for Manipulating the History
\naccept-line (Newline, Return)
\nAccept the line regardless of where the cursor is.\u00a0 If this
\nline is non-empty, add it to the history list according to the
\nstate of the HISTCONTROL variable.\u00a0 If the line is a modified
\nhistory line, then restore the history line to its original
\nstate.
\nprevious-history (C-p)
\nFetch the previous command from the history list, moving back
\nin the list.
\nnext-history (C-n)
\nFetch the next command from the history list, moving forward
\nin the list.
\nbeginning-of-history (M-<)
\nMove to the first line in the history.
\nend-of-history (M->)
\nMove to the end of the input history, i.e., the line currently
\nbeing entered.
\nreverse-search-history (C-r)
\nSearch backward starting at the current line and moving `up’
\nthrough the history as necessary.\u00a0 This is an incremental
\nsearch.
\nforward-search-history (C-s)
\nSearch forward starting at the current line and moving `down’
\nthrough the history as necessary.\u00a0 This is an incremental
\nsearch.
\nnon-incremental-reverse-search-history (M-p)
\nSearch backward through the history starting at the current
\nline using a non-incremental search for a string supplied by
\nthe user.
\nnon-incremental-forward-search-history (M-n)
\nSearch forward through the history using a non-incremental
\nsearch for a string supplied by the user.
\nhistory-search-forward
\nSearch forward through the history for the string of
\ncharacters between the start of the current line and the
\npoint.\u00a0 This is a non-incremental search.
\nhistory-search-backward
\nSearch backward through the history for the string of
\ncharacters between the start of the current line and the
\npoint.\u00a0 This is a non-incremental search.
\nyank-nth-arg (M-C-y)
\nInsert the first argument to the previous command (usually the
\nsecond word on the previous line) at point.\u00a0 With an argument
\nn, insert the nth word from the previous command (the words in
\nthe previous command begin with word 0).\u00a0 A negative argument
\ninserts the nth word from the end of the previous command.
\nOnce the argument n is computed, the argument is extracted as
\nif the “!n” history expansion had been specified.
\nyank-last-arg (M-., M-_)
\nInsert the last argument to the previous command (the last
\nword of the previous history entry).\u00a0 With a numeric argument,
\nbehave exactly like yank-nth-arg.\u00a0 Successive calls to
\nyank-last-arg move back through the history list, inserting
\nthe last word (or the word specified by the argument to the
\nfirst call) of each line in turn.\u00a0 Any numeric argument
\nsupplied to these successive calls determines the direction to
\nmove through the history.\u00a0 A negative argument switches the
\ndirection through the history (back or forward).\u00a0 The history
\nexpansion facilities are used to extract the last word, as if
\nthe “!$” history expansion had been specified.
\nshell-expand-line (M-C-e)
\nExpand the line as the shell does.\u00a0 This performs alias and
\nhistory expansion as well as all of the shell word expansions.
\nSee HISTORY EXPANSION below for a description of history
\nexpansion.
\nhistory-expand-line (M-^)
\nPerform history expansion on the current line.\u00a0 See HISTORY
\nEXPANSION below for a description of history expansion.
\nmagic-space
\nPerform history expansion on the current line and insert a
\nspace.\u00a0 See HISTORY EXPANSION below for a description of
\nhistory expansion.
\nalias-expand-line
\nPerform alias expansion on the current line.\u00a0 See ALIASES
\nabove for a description of alias expansion.
\nhistory-and-alias-expand-line
\nPerform history and alias expansion on the current line.
\ninsert-last-argument (M-., M-_)
\nA synonym for yank-last-arg.
\noperate-and-get-next (C-o)
\nAccept the current line for execution and fetch the next line
\nrelative to the current line from the history for editing.
\nAny argument is ignored.
\nedit-and-execute-command (C-xC-e)
\nInvoke an editor on the current command line, and execute the
\nresult as shell commands.\u00a0 Bash attempts to invoke $VISUAL,
\n$EDITOR, and emacs as the editor, in that order.<\/p>\n
Commands for Changing Text
\nend-of-file (usually C-d)
\nThe character indicating end-of-file as set, for example, by
\n“stty”.\u00a0 If this character is read when there are no
\ncharacters on the line, and point is at the beginning of the
\nline, Readline interprets it as the end of input and returns
\nEOF.
\ndelete-char (C-d)
\nDelete the character at point.\u00a0 If this function is bound to
\nthe same character as the tty EOF character, as C-d commonly
\nis, see above for the effects.
\nbackward-delete-char (Rubout)
\nDelete the character behind the cursor.\u00a0 When given a numeric
\nargument, save the deleted text on the kill ring.
\nforward-backward-delete-char
\nDelete the character under the cursor, unless the cursor is at
\nthe end of the line, in which case the character behind the
\ncursor is deleted.
\nquoted-insert (C-q, C-v)
\nAdd the next character typed to the line verbatim.\u00a0 This is
\nhow to insert characters like C-q, for example.
\ntab-insert (C-v TAB)
\nInsert a tab character.
\nself-insert (a, b, A, 1, !, …)
\nInsert the character typed.
\ntranspose-chars (C-t)
\nDrag the character before point forward over the character at
\npoint, moving point forward as well.\u00a0 If point is at the end
\nof the line, then this transposes the two characters before
\npoint.\u00a0 Negative arguments have no effect.
\ntranspose-words (M-t)
\nDrag the word before point past the word after point, moving
\npoint over that word as well.\u00a0 If point is at the end of the
\nline, this transposes the last two words on the line.
\nupcase-word (M-u)
\nUppercase the current (or following) word.\u00a0 With a negative
\nargument, uppercase the previous word, but do not move point.
\ndowncase-word (M-l)
\nLowercase the current (or following) word.\u00a0 With a negative
\nargument, lowercase the previous word, but do not move point.
\ncapitalize-word (M-c)
\nCapitalize the current (or following) word.\u00a0 With a negative
\nargument, capitalize the previous word, but do not move point.
\noverwrite-mode
\nToggle overwrite mode.\u00a0 With an explicit positive numeric
\nargument, switches to overwrite mode.\u00a0 With an explicit non-
\npositive numeric argument, switches to insert mode.\u00a0 This
\ncommand affects only emacs mode; vi mode does overwrite
\ndifferently.\u00a0 Each call to readline() starts in insert mode.
\nIn overwrite mode, characters bound to self-insert replace the
\ntext at point rather than pushing the text to the right.
\nCharacters bound to backward-delete-char replace the character
\nbefore point with a space.\u00a0 By default, this command is
\nunbound.<\/p>\n
Killing and Yanking
\nkill-line (C-k)
\nKill the text from point to the end of the line.
\nbackward-kill-line (C-x Rubout)
\nKill backward to the beginning of the line.
\nunix-line-discard (C-u)
\nKill backward from point to the beginning of the line.\u00a0 The
\nkilled text is saved on the kill-ring.
\nkill-whole-line
\nKill all characters on the current line, no matter where point
\nis.
\nkill-word (M-d)
\nKill from point to the end of the current word, or if between
\nwords, to the end of the next word.\u00a0 Word boundaries are the
\nsame as those used by forward-word.
\nbackward-kill-word (M-Rubout)
\nKill the word behind point.\u00a0 Word boundaries are the same as
\nthose used by backward-word.
\nshell-kill-word (M-d)
\nKill from point to the end of the current word, or if between
\nwords, to the end of the next word.\u00a0 Word boundaries are the
\nsame as those used by shell-forward-word.
\nshell-backward-kill-word (M-Rubout)
\nKill the word behind point.\u00a0 Word boundaries are the same as
\nthose used by shell-backward-word.
\nunix-word-rubout (C-w)
\nKill the word behind point, using white space as a word
\nboundary.\u00a0 The killed text is saved on the kill-ring.
\nunix-filename-rubout
\nKill the word behind point, using white space and the slash
\ncharacter as the word boundaries.\u00a0 The killed text is saved on
\nthe kill-ring.
\ndelete-horizontal-space (M-\\)
\nDelete all spaces and tabs around point.
\nkill-region
\nKill the text in the current region.
\ncopy-region-as-kill
\nCopy the text in the region to the kill buffer.
\ncopy-backward-word
\nCopy the word before point to the kill buffer.\u00a0 The word
\nboundaries are the same as backward-word.
\ncopy-forward-word
\nCopy the word following point to the kill buffer.\u00a0 The word
\nboundaries are the same as forward-word.
\nyank (C-y)
\nYank the top of the kill ring into the buffer at point.
\nyank-pop (M-y)
\nRotate the kill ring, and yank the new top.\u00a0 Only works
\nfollowing yank or yank-pop.<\/p>\n
Numeric Arguments
\ndigit-argument (M-0, M-1, …, M–)
\nAdd this digit to the argument already accumulating, or start
\na new argument.\u00a0 M– starts a negative argument.
\nuniversal-argument
\nThis is another way to specify an argument.\u00a0 If this command
\nis followed by one or more digits, optionally with a leading
\nminus sign, those digits define the argument.\u00a0 If the command
\nis followed by digits, executing universal-argument again ends
\nthe numeric argument, but is otherwise ignored.\u00a0 As a special
\ncase, if this command is immediately followed by a character
\nthat is neither a digit or minus sign, the argument count for
\nthe next command is multiplied by four.\u00a0 The argument count is
\ninitially one, so executing this function the first time makes
\nthe argument count four, a second time makes the argument
\ncount sixteen, and so on.<\/p>\n
Completing
\ncomplete (TAB)
\nAttempt to perform completion on the text before point.\u00a0 Bash
\nattempts completion treating the text as a variable (if the
\ntext begins with $), username (if the text begins with ~),
\nhostname (if the text begins with @), or command (including
\naliases and functions) in turn.\u00a0 If none of these produces a
\nmatch, filename completion is attempted.
\npossible-completions (M-?)
\nList the possible completions of the text before point.
\ninsert-completions (M-*)
\nInsert all completions of the text before point that would
\nhave been generated by possible-completions.
\nmenu-complete
\nSimilar to complete, but replaces the word to be completed
\nwith a single match from the list of possible completions.
\nRepeated execution of menu-complete steps through the list of
\npossible completions, inserting each match in turn.\u00a0 At the
\nend of the list of completions, the bell is rung (subject to
\nthe setting of bell-style) and the original text is restored.
\nAn argument of n moves n positions forward in the list of
\nmatches; a negative argument may be used to move backward
\nthrough the list.\u00a0 This command is intended to be bound to
\nTAB, but is unbound by default.
\nmenu-complete-backward
\nIdentical to menu-complete, but moves backward through the
\nlist of possible completions, as if menu-complete had been
\ngiven a negative argument.\u00a0 This command is unbound by
\ndefault.
\ndelete-char-or-list
\nDeletes the character under the cursor if not at the beginning
\nor end of the line (like delete-char).\u00a0 If at the end of the
\nline, behaves identically to possible-completions.\u00a0 This
\ncommand is unbound by default.
\ncomplete-filename (M-\/)
\nAttempt filename completion on the text before point.
\npossible-filename-completions (C-x \/)
\nList the possible completions of the text before point,
\ntreating it as a filename.
\ncomplete-username (M-~)
\nAttempt completion on the text before point, treating it as a
\nusername.
\npossible-username-completions (C-x ~)
\nList the possible completions of the text before point,
\ntreating it as a username.
\ncomplete-variable (M-$)
\nAttempt completion on the text before point, treating it as a
\nshell variable.
\npossible-variable-completions (C-x $)
\nList the possible completions of the text before point,
\ntreating it as a shell variable.
\ncomplete-hostname (M-@)
\nAttempt completion on the text before point, treating it as a
\nhostname.
\npossible-hostname-completions (C-x @)
\nList the possible completions of the text before point,
\ntreating it as a hostname.
\ncomplete-command (M-!)
\nAttempt completion on the text before point, treating it as a
\ncommand name.\u00a0 Command completion attempts to match the text
\nagainst aliases, reserved words, shell functions, shell
\nbuiltins, and finally executable filenames, in that order.
\npossible-command-completions (C-x !)
\nList the possible completions of the text before point,
\ntreating it as a command name.
\ndynamic-complete-history (M-TAB)
\nAttempt completion on the text before point, comparing the
\ntext against lines from the history list for possible
\ncompletion matches.
\ndabbrev-expand
\nAttempt menu completion on the text before point, comparing
\nthe text against lines from the history list for possible
\ncompletion matches.
\ncomplete-into-braces (M-{)
\nPerform filename completion and insert the list of possible
\ncompletions enclosed within braces so the list is available to
\nthe shell (see Brace Expansion above).<\/p>\n
Keyboard Macros
\nstart-kbd-macro (C-x ()
\nBegin saving the characters typed into the current keyboard
\nmacro.
\nend-kbd-macro (C-x ))
\nStop saving the characters typed into the current keyboard
\nmacro and store the definition.
\ncall-last-kbd-macro (C-x e)
\nRe-execute the last keyboard macro defined, by making the
\ncharacters in the macro appear as if typed at the keyboard.
\nprint-last-kbd-macro ()
\nPrint the last keyboard macro defined in a format suitable for
\nthe inputrc file.<\/p>\n
Miscellaneous
\nre-read-init-file (C-x C-r)
\nRead in the contents of the inputrc file, and incorporate any
\nbindings or variable assignments found there.
\nabort (C-g)
\nAbort the current editing command and ring the terminal’s bell
\n(subject to the setting of bell-style).
\ndo-uppercase-version (M-a, M-b, M-x, …)
\nIf the metafied character x is lowercase, run the command that
\nis bound to the corresponding uppercase character.
\nprefix-meta (ESC)
\nMetafy the next character typed.\u00a0 ESC f is equivalent to
\nMeta-f.
\nundo (C-_, C-x C-u)
\nIncremental undo, separately remembered for each line.
\nrevert-line (M-r)
\nUndo all changes made to this line.\u00a0 This is like executing
\nthe undo command enough times to return the line to its
\ninitial state.
\ntilde-expand (M-&)
\nPerform tilde expansion on the current word.
\nset-mark (C-@, M-<space>)
\nSet the mark to the point.\u00a0 If a numeric argument is supplied,
\nthe mark is set to that position.
\nexchange-point-and-mark (C-x C-x)
\nSwap the point with the mark.\u00a0 The current cursor position is
\nset to the saved position, and the old cursor position is
\nsaved as the mark.
\ncharacter-search (C-])
\nA character is read and point is moved to the next occurrence
\nof that character.\u00a0 A negative count searches for previous
\noccurrences.
\ncharacter-search-backward (M-C-])
\nA character is read and point is moved to the previous
\noccurrence of that character.\u00a0 A negative count searches for
\nsubsequent occurrences.
\nskip-csi-sequence
\nRead enough characters to consume a multi-key sequence such as
\nthose defined for keys like Home and End.\u00a0 Such sequences
\nbegin with a Control Sequence Indicator (CSI), usually ESC-[.
\nIf this sequence is bound to “\\[“, keys producing such
\nsequences will have no effect unless explicitly bound to a
\nreadline command, instead of inserting stray characters into
\nthe editing buffer.\u00a0 This is unbound by default, but usually
\nbound to ESC-[.
\ninsert-comment (M-#)
\nWithout a numeric argument, the value of the readline
\ncomment-begin variable is inserted at the beginning of the
\ncurrent line.\u00a0 If a numeric argument is supplied, this command
\nacts as a toggle:\u00a0 if the characters at the beginning of the
\nline do not match the value of comment-begin, the value is
\ninserted, otherwise the characters in comment-begin are
\ndeleted from the beginning of the line.\u00a0 In either case, the
\nline is accepted as if a newline had been typed.\u00a0 The default
\nvalue of comment-begin causes this command to make the current
\nline a shell comment.\u00a0 If a numeric argument causes the
\ncomment character to be removed, the line will be executed by
\nthe shell.
\nglob-complete-word (M-g)
\nThe word before point is treated as a pattern for pathname
\nexpansion, with an asterisk implicitly appended.\u00a0 This pattern
\nis used to generate a list of matching filenames for possible
\ncompletions.
\nglob-expand-word (C-x *)
\nThe word before point is treated as a pattern for pathname
\nexpansion, and the list of matching filenames is inserted,
\nreplacing the word.\u00a0 If a numeric argument is supplied, an
\nasterisk is appended before pathname expansion.
\nglob-list-expansions (C-x g)
\nThe list of expansions that would have been generated by
\nglob-expand-word is displayed, and the line is redrawn.\u00a0 If a
\nnumeric argument is supplied, an asterisk is appended before
\npathname expansion.
\ndump-functions
\nPrint all of the functions and their key bindings to the
\nreadline output stream.\u00a0 If a numeric argument is supplied,
\nthe output is formatted in such a way that it can be made part
\nof an inputrc file.
\ndump-variables
\nPrint all of the settable readline variables and their values
\nto the readline output stream.\u00a0 If a numeric argument is
\nsupplied, the output is formatted in such a way that it can be
\nmade part of an inputrc file.
\ndump-macros
\nPrint all of the readline key sequences bound to macros and
\nthe strings they output.\u00a0 If a numeric argument is supplied,
\nthe output is formatted in such a way that it can be made part
\nof an inputrc file.
\ndisplay-shell-version (C-x C-v)
\nDisplay version information about the current instance of
\nbash.<\/p>\n
Programmable Completion
\nWhen word completion is attempted for an argument to a command for
\nwhich a completion specification (a compspec) has been defined using
\nthe complete builtin (see SHELL BUILTIN COMMANDS below), the
\nprogrammable completion facilities are invoked.<\/p>\n
First, the command name is identified.\u00a0 If the command word is the
\nempty string (completion attempted at the beginning of an empty
\nline), any compspec defined with the -E option to complete is used.
\nIf a compspec has been defined for that command, the compspec is used
\nto generate the list of possible completions for the word.\u00a0 If the
\ncommand word is a full pathname, a compspec for the full pathname is
\nsearched for first.\u00a0 If no compspec is found for the full pathname,
\nan attempt is made to find a compspec for the portion following the
\nfinal slash.\u00a0 If those searches do not result in a compspec, any
\ncompspec defined with the -D option to complete is used as the
\ndefault.<\/p>\n
Once a compspec has been found, it is used to generate the list of
\nmatching words.\u00a0 If a compspec is not found, the default bash
\ncompletion as described above under Completing is performed.<\/p>\n
First, the actions specified by the compspec are used.\u00a0 Only matches
\nwhich are prefixed by the word being completed are returned.\u00a0 When
\nthe -f or -d option is used for filename or directory name
\ncompletion, the shell variable FIGNORE is used to filter the matches.<\/p>\n
Any completions specified by a pathname expansion pattern to the -G
\noption are generated next.\u00a0 The words generated by the pattern need
\nnot match the word being completed.\u00a0 The GLOBIGNORE shell variable is
\nnot used to filter the matches, but the FIGNORE variable is used.<\/p>\n
Next, the string specified as the argument to the -W option is
\nconsidered.\u00a0 The string is first split using the characters in the
\nIFS special variable as delimiters.\u00a0 Shell quoting is honored.\u00a0 Each
\nword is then expanded using brace expansion, tilde expansion,
\nparameter and variable expansion, command substitution, and
\narithmetic expansion, as described above under EXPANSION.\u00a0 The
\nresults are split using the rules described above under Word
\nSplitting.\u00a0 The results of the expansion are prefix-matched against
\nthe word being completed, and the matching words become the possible
\ncompletions.<\/p>\n
After these matches have been generated, any shell function or
\ncommand specified with the -F and -C options is invoked.\u00a0 When the
\ncommand or function is invoked, the COMP_LINE, COMP_POINT, COMP_KEY,
\nand COMP_TYPE variables are assigned values as described above under
\nShell Variables.\u00a0 If a shell function is being invoked, the
\nCOMP_WORDS and COMP_CWORD variables are also set.\u00a0 When the function
\nor command is invoked, the first argument ($1) is the name of the
\ncommand whose arguments are being completed, the second argument ($2)
\nis the word being completed, and the third argument ($3) is the word
\npreceding the word being completed on the current command line.\u00a0 No
\nfiltering of the generated completions against the word being
\ncompleted is performed; the function or command has complete freedom
\nin generating the matches.<\/p>\n
Any function specified with -F is invoked first.\u00a0 The function may
\nuse any of the shell facilities, including the compgen builtin
\ndescribed below, to generate the matches.\u00a0 It must put the possible
\ncompletions in the COMPREPLY array variable, one per array element.<\/p>\n
Next, any command specified with the -C option is invoked in an
\nenvironment equivalent to command substitution.\u00a0 It should print a
\nlist of completions, one per line, to the standard output.\u00a0 Backslash
\nmay be used to escape a newline, if necessary.<\/p>\n
After all of the possible completions are generated, any filter
\nspecified with the -X option is applied to the list.\u00a0 The filter is a
\npattern as used for pathname expansion; a & in the pattern is
\nreplaced with the text of the word being completed.\u00a0 A literal & may
\nbe escaped with a backslash; the backslash is removed before
\nattempting a match.\u00a0 Any completion that matches the pattern will be
\nremoved from the list.\u00a0 A leading ! negates the pattern; in this case
\nany completion not matching the pattern will be removed.<\/p>\n
Finally, any prefix and suffix specified with the -P and -S options
\nare added to each member of the completion list, and the result is
\nreturned to the readline completion code as the list of possible
\ncompletions.<\/p>\n
If the previously-applied actions do not generate any matches, and
\nthe -o dirnames option was supplied to complete when the compspec was
\ndefined, directory name completion is attempted.<\/p>\n
If the -o plusdirs option was supplied to complete when the compspec
\nwas defined, directory name completion is attempted and any matches
\nare added to the results of the other actions.<\/p>\n
By default, if a compspec is found, whatever it generates is returned
\nto the completion code as the full set of possible completions.\u00a0 The
\ndefault bash completions are not attempted, and the readline default
\nof filename completion is disabled.\u00a0 If the -o bashdefault option was
\nsupplied to complete when the compspec was defined, the bash default
\ncompletions are attempted if the compspec generates no matches.\u00a0 If
\nthe -o default option was supplied to complete when the compspec was
\ndefined, readline’s default completion will be performed if the
\ncompspec (and, if attempted, the default bash completions) generate
\nno matches.<\/p>\n
When a compspec indicates that directory name completion is desired,
\nthe programmable completion functions force readline to append a
\nslash to completed names which are symbolic links to directories,
\nsubject to the value of the mark-directories readline variable,
\nregardless of the setting of the mark-symlinked-directories readline
\nvariable.<\/p>\n
There is some support for dynamically modifying completions.\u00a0 This is
\nmost useful when used in combination with a default completion
\nspecified with complete -D.\u00a0 It’s possible for shell functions
\nexecuted as completion handlers to indicate that completion should be
\nretried by returning an exit status of 124.\u00a0 If a shell function
\nreturns 124, and changes the compspec associated with the command on
\nwhich completion is being attempted (supplied as the first argument
\nwhen the function is executed), programmable completion restarts from
\nthe beginning, with an attempt to find a new compspec for that
\ncommand.\u00a0 This allows a set of completions to be built dynamically as
\ncompletion is attempted, rather than being loaded all at once.<\/p>\n
For instance, assuming that there is a library of compspecs, each
\nkept in a file corresponding to the name of the command, the
\nfollowing default completion function would load completions
\ndynamically:<\/p>\n
_completion_loader()
\n{
\n. “\/etc\/bash_completion.d\/$1.sh” >\/dev\/null 2>&1 && return 124
\n}
\ncomplete -D -F _completion_loader -o bashdefault -o default<\/p>\n
HISTORY<\/p>\n
When the -o history option to the set builtin is enabled, the shell
\nprovides access to the command history, the list of commands
\npreviously typed.\u00a0 The value of the HISTSIZE variable is used as the
\nnumber of commands to save in a history list.\u00a0 The text of the last
\nHISTSIZE commands (default 500) is saved.\u00a0 The shell stores each
\ncommand in the history list prior to parameter and variable expansion
\n(see EXPANSION above) but after history expansion is performed,
\nsubject to the values of the shell variables HISTIGNORE and
\nHISTCONTROL.<\/p>\n
On startup, the history is initialized from the file named by the
\nvariable HISTFILE (default ~\/.bash_history).\u00a0 The file named by the
\nvalue of HISTFILE is truncated, if necessary, to contain no more than
\nthe number of lines specified by the value of HISTFILESIZE.\u00a0 If
\nHISTFILESIZE is unset, or set to null, a non-numeric value, or a
\nnumeric value less than zero, the history file is not truncated.
\nWhen the history file is read, lines beginning with the history
\ncomment character followed immediately by a digit are interpreted as
\ntimestamps for the preceding history line.\u00a0 These timestamps are
\noptionally displayed depending on the value of the HISTTIMEFORMAT
\nvariable.\u00a0 When a shell with history enabled exits, the last
\n$HISTSIZE lines are copied from the history list to $HISTFILE.\u00a0 If
\nthe histappend shell option is enabled (see the description of shopt
\nunder SHELL BUILTIN COMMANDS below), the lines are appended to the
\nhistory file, otherwise the history file is overwritten.\u00a0 If HISTFILE
\nis unset, or if the history file is unwritable, the history is not
\nsaved.\u00a0 If the HISTTIMEFORMAT variable is set, time stamps are
\nwritten to the history file, marked with the history comment
\ncharacter, so they may be preserved across shell sessions.\u00a0 This uses
\nthe history comment character to distinguish timestamps from other
\nhistory lines.\u00a0 After saving the history, the history file is
\ntruncated to contain no more than HISTFILESIZE lines.\u00a0 If
\nHISTFILESIZE is unset, or set to null, a non-numeric value, or a
\nnumeric value less than zero, the history file is not truncated.<\/p>\n
The builtin command fc (see SHELL BUILTIN COMMANDS below) may be used
\nto list or edit and re-execute a portion of the history list.\u00a0 The
\nhistory builtin may be used to display or modify the history list and
\nmanipulate the history file.\u00a0 When using command-line editing, search
\ncommands are available in each editing mode that provide access to
\nthe history list.<\/p>\n
The shell allows control over which commands are saved on the history
\nlist.\u00a0 The HISTCONTROL and HISTIGNORE variables may be set to cause
\nthe shell to save only a subset of the commands entered.\u00a0 The cmdhist
\nshell option, if enabled, causes the shell to attempt to save each
\nline of a multi-line command in the same history entry, adding
\nsemicolons where necessary to preserve syntactic correctness.\u00a0 The
\nlithist shell option causes the shell to save the command with
\nembedded newlines instead of semicolons.\u00a0 See the description of the
\nshopt builtin below under SHELL BUILTIN COMMANDS for information on
\nsetting and unsetting shell options.<\/p>\n
HISTORY EXPANSION<\/p>\n
The shell supports a history expansion feature that is similar to the
\nhistory expansion in csh.\u00a0 This section describes what syntax
\nfeatures are available.\u00a0 This feature is enabled by default for
\ninteractive shells, and can be disabled using the +H option to the
\nset builtin command (see SHELL BUILTIN COMMANDS below).\u00a0 Non-
\ninteractive shells do not perform history expansion by default.<\/p>\n
History expansions introduce words from the history list into the
\ninput stream, making it easy to repeat commands, insert the arguments
\nto a previous command into the current input line, or fix errors in
\nprevious commands quickly.<\/p>\n
History expansion is performed immediately after a complete line is
\nread, before the shell breaks it into words.\u00a0 It takes place in two
\nparts.\u00a0 The first is to determine which line from the history list to
\nuse during substitution.\u00a0 The second is to select portions of that
\nline for inclusion into the current one.\u00a0 The line selected from the
\nhistory is the event, and the portions of that line that are acted
\nupon are words.\u00a0 Various modifiers are available to manipulate the
\nselected words.\u00a0 The line is broken into words in the same fashion as
\nwhen reading input, so that several metacharacter-separated words
\nsurrounded by quotes are considered one word.\u00a0 History expansions are
\nintroduced by the appearance of the history expansion character,
\nwhich is ! by default.\u00a0 Only backslash (\\) and single quotes can
\nquote the history expansion character.<\/p>\n
Several characters inhibit history expansion if found immediately
\nfollowing the history expansion character, even if it is unquoted:
\nspace, tab, newline, carriage return, and =.\u00a0 If the extglob shell
\noption is enabled, ( will also inhibit expansion.<\/p>\n
Several shell options settable with the shopt builtin may be used to
\ntailor the behavior of history expansion.\u00a0 If the histverify shell
\noption is enabled (see the description of the shopt builtin below),
\nand readline is being used, history substitutions are not immediately
\npassed to the shell parser.\u00a0 Instead, the expanded line is reloaded
\ninto the readline editing buffer for further modification.\u00a0 If
\nreadline is being used, and the histreedit shell option is enabled, a
\nfailed history substitution will be reloaded into the readline
\nediting buffer for correction.\u00a0 The -p option to the history builtin
\ncommand may be used to see what a history expansion will do before
\nusing it.\u00a0 The -s option to the history builtin may be used to add
\ncommands to the end of the history list without actually executing
\nthem, so that they are available for subsequent recall.<\/p>\n
The shell allows control of the various characters used by the
\nhistory expansion mechanism (see the description of histchars above
\nunder Shell Variables).\u00a0 The shell uses the history comment character
\nto mark history timestamps when writing the history file.<\/p>\n
Event Designators
\nAn event designator is a reference to a command line entry in the
\nhistory list.\u00a0 Unless the reference is absolute, events are relative
\nto the current position in the history list.<\/p>\n
!\u00a0\u00a0\u00a0\u00a0\u00a0 Start a history substitution, except when followed by a blank,
\nnewline, carriage return, = or ( (when the extglob shell
\noption is enabled using the shopt builtin).
\n!n\u00a0\u00a0\u00a0\u00a0 Refer to command line n.
\n!-n\u00a0\u00a0\u00a0 Refer to the current command minus n.
\n!!\u00a0\u00a0\u00a0\u00a0 Refer to the previous command.\u00a0 This is a synonym for `!-1′.
\n!string
\nRefer to the most recent command preceding the current
\nposition in the history list starting with string.
\n!?string[?]
\nRefer to the most recent command preceding the current
\nposition in the history list containing string.\u00a0 The trailing
\n? may be omitted if string is followed immediately by a
\nnewline.
\n^string1^string2^
\nQuick substitution.\u00a0 Repeat the previous command, replacing
\nstring1 with string2.\u00a0 Equivalent to “!!:s\/string1\/string2\/”
\n(see Modifiers below).
\n!#\u00a0\u00a0\u00a0\u00a0 The entire command line typed so far.<\/p>\n
Word Designators
\nWord designators are used to select desired words from the event.\u00a0 A
\n: separates the event specification from the word designator.\u00a0 It may
\nbe omitted if the word designator begins with a ^, $, *, -, or %.
\nWords are numbered from the beginning of the line, with the first
\nword being denoted by 0 (zero).\u00a0 Words are inserted into the current
\nline separated by single spaces.<\/p>\n
0 (zero)
\nThe zeroth word.\u00a0 For the shell, this is the command word.
\nn\u00a0\u00a0\u00a0\u00a0\u00a0 The nth word.
\n^\u00a0\u00a0\u00a0\u00a0\u00a0 The first argument.\u00a0 That is, word 1.
\n$\u00a0\u00a0\u00a0\u00a0\u00a0 The last word.\u00a0 This is usually the last argument, but will
\nexpand to the zeroth word if there is only one word in the
\nline.
\n%\u00a0\u00a0\u00a0\u00a0\u00a0 The word matched by the most recent `?string?’ search.
\nx-y\u00a0\u00a0\u00a0 A range of words; `-y’ abbreviates `0-y’.
\n*\u00a0\u00a0\u00a0\u00a0\u00a0 All of the words but the zeroth.\u00a0 This is a synonym for `1-$’.
\nIt is not an error to use * if there is just one word in the
\nevent; the empty string is returned in that case.
\nx*\u00a0\u00a0\u00a0\u00a0 Abbreviates x-$.
\nx-\u00a0\u00a0\u00a0\u00a0 Abbreviates x-$ like x*, but omits the last word.<\/p>\n
If a word designator is supplied without an event specification, the
\nprevious command is used as the event.<\/p>\n
Modifiers
\nAfter the optional word designator, there may appear a sequence of
\none or more of the following modifiers, each preceded by a `:’.<\/p>\n
h\u00a0\u00a0\u00a0\u00a0\u00a0 Remove a trailing filename component, leaving only the head.
\nt\u00a0\u00a0\u00a0\u00a0\u00a0 Remove all leading filename components, leaving the tail.
\nr\u00a0\u00a0\u00a0\u00a0\u00a0 Remove a trailing suffix of the form .xxx, leaving the
\nbasename.
\ne\u00a0\u00a0\u00a0\u00a0\u00a0 Remove all but the trailing suffix.
\np\u00a0\u00a0\u00a0\u00a0\u00a0 Print the new command but do not execute it.
\nq\u00a0\u00a0\u00a0\u00a0\u00a0 Quote the substituted words, escaping further substitutions.
\nx\u00a0\u00a0\u00a0\u00a0\u00a0 Quote the substituted words as with q, but break into words at
\nblanks and newlines.
\ns\/old\/new\/
\nSubstitute new for the first occurrence of old in the event
\nline.\u00a0 Any delimiter can be used in place of \/.\u00a0 The final
\ndelimiter is optional if it is the last character of the event
\nline.\u00a0 The delimiter may be quoted in old and new with a
\nsingle backslash.\u00a0 If & appears in new, it is replaced by old.
\nA single backslash will quote the &.\u00a0 If old is null, it is
\nset to the last old substituted, or, if no previous history
\nsubstitutions took place, the last string in a !?string[?]
\nsearch.
\n&\u00a0\u00a0\u00a0\u00a0\u00a0 Repeat the previous substitution.
\ng\u00a0\u00a0\u00a0\u00a0\u00a0 Cause changes to be applied over the entire event line.\u00a0 This
\nis used in conjunction with `:s’ (e.g., `:gs\/old\/new\/’) or
\n`:&’.\u00a0 If used with `:s’, any delimiter can be used in place
\nof \/, and the final delimiter is optional if it is the last
\ncharacter of the event line.\u00a0 An a may be used as a synonym
\nfor g.
\nG\u00a0\u00a0\u00a0\u00a0\u00a0 Apply the following `s’ modifier once to each word in the
\nevent line.<\/p>\n
SHELL BUILTIN COMMANDS\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 top<\/p>\n
Unless otherwise noted, each builtin command documented in this
\nsection as accepting options preceded by – accepts — to signify the
\nend of the options.\u00a0 The :, true, false, and test builtins do not
\naccept options and do not treat — specially.\u00a0 The exit, logout,
\nbreak, continue, let, and shift builtins accept and process arguments
\nbeginning with – without requiring –.\u00a0 Other builtins that accept
\narguments but are not specified as accepting options interpret
\narguments beginning with – as invalid options and require — to
\nprevent this interpretation.
\n: [arguments]
\nNo effect; the command does nothing beyond expanding arguments
\nand performing any specified redirections.\u00a0 A zero exit code
\nis returned.<\/p>\n
.\u00a0 filename [arguments]
\nsource filename [arguments]
\nRead and execute commands from filename in the current shell
\nenvironment and return the exit status of the last command
\nexecuted from filename.\u00a0 If filename does not contain a slash,
\nfilenames in PATH are used to find the directory containing
\nfilename.\u00a0 The file searched for in PATH need not be
\nexecutable.\u00a0 When bash is not in posix mode, the current
\ndirectory is searched if no file is found in PATH.\u00a0 If the
\nsourcepath option to the shopt builtin command is turned off,
\nthe PATH is not searched.\u00a0 If any arguments are supplied, they
\nbecome the positional parameters when filename is executed.
\nOtherwise the positional parameters are unchanged.\u00a0 The return
\nstatus is the status of the last command exited within the
\nscript (0 if no commands are executed), and false if filename
\nis not found or cannot be read.<\/p>\n
alias [-p] [name[=value] …]
\nAlias with no arguments or with the -p option prints the list
\nof aliases in the form alias name=value on standard output.
\nWhen arguments are supplied, an alias is defined for each name
\nwhose value is given.\u00a0 A trailing space in\u00a0 value causes the
\nnext word to be checked for alias substitution when the alias
\nis expanded.\u00a0 For each name in the argument list for which no
\nvalue is supplied, the name and value of the alias is printed.
\nAlias returns true unless a name is given for which no alias
\nhas been defined.<\/p>\n
bg [jobspec …]
\nResume each suspended job jobspec in the background, as if it
\nhad been started with &.\u00a0 If jobspec is not present, the
\nshell’s notion of the current job is used.\u00a0 bg jobspec returns
\n0 unless run when job control is disabled or, when run with
\njob control enabled, any specified jobspec was not found or
\nwas started without job control.<\/p>\n
bind [-m keymap] [-lpsvPSVX]
\nbind [-m keymap] [-q function] [-u function] [-r keyseq]
\nbind [-m keymap] -f filename
\nbind [-m keymap] -x keyseq:shell-command
\nbind [-m keymap] keyseq:function-name
\nbind readline-command
\nDisplay current readline key and function bindings, bind a key
\nsequence to a readline function or macro, or set a readline
\nvariable.\u00a0 Each non-option argument is a command as it would
\nappear in .inputrc, but each binding or command must be passed
\nas a separate argument; e.g., ‘”\\C-x\\C-r”: re-read-init-file’.
\nOptions, if supplied, have the following meanings:
\n-m keymap
\nUse keymap as the keymap to be affected by the
\nsubsequent bindings.\u00a0 Acceptable keymap names are
\nemacs, emacs-standard, emacs-meta, emacs-ctlx, vi,
\nvi-move, vi-command, and vi-insert.\u00a0 vi is equivalent
\nto vi-command; emacs is equivalent to emacs-standard.
\n-l\u00a0\u00a0\u00a0\u00a0 List the names of all readline functions.
\n-p\u00a0\u00a0\u00a0\u00a0 Display readline function names and bindings in such a
\nway that they can be re-read.
\n-P\u00a0\u00a0\u00a0\u00a0 List current readline function names and bindings.
\n-s\u00a0\u00a0\u00a0\u00a0 Display readline key sequences bound to macros and the
\nstrings they output in such a way that they can be re-
\nread.
\n-S\u00a0\u00a0\u00a0\u00a0 Display readline key sequences bound to macros and the
\nstrings they output.
\n-v\u00a0\u00a0\u00a0\u00a0 Display readline variable names and values in such a
\nway that they can be re-read.
\n-V\u00a0\u00a0\u00a0\u00a0 List current readline variable names and values.
\n-f filename
\nRead key bindings from filename.
\n-q function
\nQuery about which keys invoke the named function.
\n-u function
\nUnbind all keys bound to the named function.
\n-r keyseq
\nRemove any current binding for keyseq.
\n-x keyseq:shell-command
\nCause shell-command to be executed whenever keyseq is
\nentered.\u00a0 When shell-command is executed, the shell
\nsets the READLINE_LINE variable to the contents of the
\nreadline line buffer and the READLINE_POINT variable to
\nthe current location of the insertion point.\u00a0 If the
\nexecuted command changes the value of READLINE_LINE or
\nREADLINE_POINT, those new values will be reflected in
\nthe editing state.
\n-X\u00a0\u00a0\u00a0\u00a0 List all key sequences bound to shell commands and the
\nassociated commands in a format that can be reused as
\ninput.<\/p>\n
The return value is 0 unless an unrecognized option is given
\nor an error occurred.<\/p>\n
break [n]
\nExit from within a for, while, until, or select loop.\u00a0 If n is
\nspecified, break n levels.\u00a0 n must be \u2265 1.\u00a0 If n is greater
\nthan the number of enclosing loops, all enclosing loops are
\nexited.\u00a0 The return value is 0 unless n is not greater than or
\nequal to 1.<\/p>\n
builtin shell-builtin [arguments]
\nExecute the specified shell builtin, passing it arguments, and
\nreturn its exit status.\u00a0 This is useful when defining a
\nfunction whose name is the same as a shell builtin, retaining
\nthe functionality of the builtin within the function.\u00a0 The cd
\nbuiltin is commonly redefined this way.\u00a0 The return status is
\nfalse if shell-builtin is not a shell builtin command.<\/p>\n
caller [expr]
\nReturns the context of any active subroutine call (a shell
\nfunction or a script executed with the . or source builtins).
\nWithout expr, caller displays the line number and source
\nfilename of the current subroutine call.\u00a0 If a non-negative
\ninteger is supplied as expr, caller displays the line number,
\nsubroutine name, and source file corresponding to that
\nposition in the current execution call stack.\u00a0 This extra
\ninformation may be used, for example, to print a stack trace.
\nThe current frame is frame 0.\u00a0 The return value is 0 unless
\nthe shell is not executing a subroutine call or expr does not
\ncorrespond to a valid position in the call stack.<\/p>\n
cd [-L|[-P [-e]] [-@]] [dir]
\nChange the current directory to dir.\u00a0 if dir is not supplied,
\nthe value of the HOME shell variable is the default.\u00a0 Any
\nadditional arguments following dir are ignored.\u00a0 The variable
\nCDPATH defines the search path for the directory containing
\ndir: each directory name in CDPATH is searched for dir.
\nAlternative directory names in CDPATH are separated by a colon
\n(:).\u00a0 A null directory name in CDPATH is the same as the
\ncurrent directory, i.e., “.”.\u00a0 If dir begins with a slash
\n(\/), then CDPATH is not used. The -P option causes cd to use
\nthe physical directory structure by resolving symbolic links
\nwhile traversing dir and before processing instances of .. in
\ndir (see also the -P option to the set builtin command); the
\n-L option forces symbolic links to be followed by resolving
\nthe link after processing instances of .. in dir.\u00a0 If ..
\nappears in dir, it is processed by removing the immediately
\nprevious pathname component from dir, back to a slash or the
\nbeginning of dir.\u00a0 If the -e option is supplied with -P, and
\nthe current working directory cannot be successfully
\ndetermined after a successful directory change, cd will return
\nan unsuccessful status.\u00a0 On systems that support it, the\u00a0-@
\noption presents the extended attributes associated with a file
\nas a directory.\u00a0 An argument of – is converted to $OLDPWD
\nbefore the directory change is attempted.\u00a0 If a non-empty
\ndirectory name from CDPATH is used, or if – is the first
\nargument, and the directory change is successful, the absolute
\npathname of the new working directory is written to the
\nstandard output.\u00a0 The return value is true if the directory
\nwas successfully changed; false otherwise.<\/p>\n
command [-pVv] command [arg …]
\nRun command with args suppressing the normal shell function
\nlookup. Only builtin commands or commands found in the PATH
\nare executed.\u00a0 If the -p option is given, the search for
\ncommand is performed using a default value for PATH that is
\nguaranteed to find all of the standard utilities.\u00a0 If either
\nthe -V or -v option is supplied, a description of command is
\nprinted.\u00a0 The -v option causes a single word indicating the
\ncommand or filename used to invoke command to be displayed;
\nthe -V option produces a more verbose description.\u00a0 If the -V
\nor -v option is supplied, the exit status is 0 if command was
\nfound, and 1 if not.\u00a0 If neither option is supplied and an
\nerror occurred or command cannot be found, the exit status is
\n127.\u00a0 Otherwise, the exit status of the command builtin is the
\nexit status of command.<\/p>\n
compgen [option] [word]
\nGenerate possible completion matches for word according to the
\noptions, which may be any option accepted by the complete
\nbuiltin with the exception of -p and -r, and write the matches
\nto the standard output.\u00a0 When using the -F or -C options, the
\nvarious shell variables set by the programmable completion
\nfacilities, while available, will not have useful values.<\/p>\n
The matches will be generated in the same way as if the
\nprogrammable completion code had generated them directly from
\na completion specification with the same flags.\u00a0 If word is
\nspecified, only those completions matching word will be
\ndisplayed.<\/p>\n
The return value is true unless an invalid option is supplied,
\nor no matches were generated.<\/p>\n
complete [-abcdefgjksuv] [-o comp-option] [-DE] [-A action] [-G
\nglobpat] [-W wordlist] [-F function] [-C command]
\n[-X filterpat] [-P prefix] [-S suffix] name [name …]
\ncomplete -pr [-DE] [name …]
\nSpecify how arguments to each name should be completed.\u00a0 If
\nthe -p option is supplied, or if no options are supplied,
\nexisting completion specifications are printed in a way that
\nallows them to be reused as input.\u00a0 The -r option removes a
\ncompletion specification for each name, or, if no names are
\nsupplied, all completion specifications.\u00a0 The -D option
\nindicates that the remaining options and actions should apply
\nto the “default” command completion; that is, completion
\nattempted on a command for which no completion has previously
\nbeen defined.\u00a0 The -E option indicates that the remaining
\noptions and actions should apply to “empty” command
\ncompletion; that is, completion attempted on a blank line.<\/p>\n
The process of applying these completion specifications when
\nword completion is attempted is described above under
\nProgrammable Completion.<\/p>\n
Other options, if specified, have the following meanings.\u00a0 The
\narguments to the -G, -W, and -X options (and, if necessary,
\nthe -P and -S options) should be quoted to protect them from
\nexpansion before the complete builtin is invoked.
\n-o comp-option
\nThe comp-option controls several aspects of the
\ncompspec’s behavior beyond the simple generation of
\ncompletions.\u00a0 comp-option may be one of:
\nbashdefault
\nPerform the rest of the default bash
\ncompletions if the compspec generates no
\nmatches.
\ndefault Use readline’s default filename completion if
\nthe compspec generates no matches.
\ndirnames
\nPerform directory name completion if the
\ncompspec generates no matches.
\nfilenames
\nTell readline that the compspec generates
\nfilenames, so it can perform any
\nfilename-specific processing (like adding a
\nslash to directory names, quoting special
\ncharacters, or suppressing trailing spaces).
\nIntended to be used with shell functions.
\nnoquote Tell readline not to quote the completed words
\nif they are filenames (quoting filenames is
\nthe default).
\nnospace Tell readline not to append a space (the
\ndefault) to words completed at the end of the
\nline.
\nplusdirs
\nAfter any matches defined by the compspec are
\ngenerated, directory name completion is
\nattempted and any matches are added to the
\nresults of the other actions.
\n-A action
\nThe action may be one of the following to generate a
\nlist of possible completions:
\nalias\u00a0\u00a0 Alias names.\u00a0 May also be specified as -a.
\narrayvar
\nArray variable names.
\nbinding Readline key binding names.
\nbuiltin Names of shell builtin commands.\u00a0 May also be
\nspecified as -b.
\ncommand Command names.\u00a0 May also be specified as -c.
\ndirectory
\nDirectory names.\u00a0 May also be specified as -d.
\ndisabled
\nNames of disabled shell builtins.
\nenabled Names of enabled shell builtins.
\nexport\u00a0 Names of exported shell variables.\u00a0 May also
\nbe specified as -e.
\nfile\u00a0\u00a0\u00a0 File names.\u00a0 May also be specified as -f.
\nfunction
\nNames of shell functions.
\ngroup\u00a0\u00a0 Group names.\u00a0 May also be specified as -g.
\nhelptopic
\nHelp topics as accepted by the help builtin.
\nhostname
\nHostnames, as taken from the file specified by
\nthe HOSTFILE shell variable.
\njob\u00a0\u00a0\u00a0\u00a0 Job names, if job control is active.\u00a0 May also
\nbe specified as -j.
\nkeyword Shell reserved words.\u00a0 May also be specified
\nas -k.
\nrunning Names of running jobs, if job control is
\nactive.
\nservice Service names.\u00a0 May also be specified as -s.
\nsetopt\u00a0 Valid arguments for the -o option to the set
\nbuiltin.
\nshopt\u00a0\u00a0 Shell option names as accepted by the shopt
\nbuiltin.
\nsignal\u00a0 Signal names.
\nstopped Names of stopped jobs, if job control is
\nactive.
\nuser\u00a0\u00a0\u00a0 User names.\u00a0 May also be specified as -u.
\nvariable
\nNames of all shell variables.\u00a0 May also be
\nspecified as -v.
\n-C command
\ncommand is executed in a subshell environment, and its
\noutput is used as the possible completions.
\n-F function
\nThe shell function function is executed in the current
\nshell environment.\u00a0 When the function is executed, the
\nfirst argument ($1) is the name of the command whose
\narguments are being completed, the second argument
\n($2) is the word being completed, and the third
\nargument ($3) is the word preceding the word being
\ncompleted on the current command line.\u00a0 When it
\nfinishes, the possible completions are retrieved from
\nthe value of the COMPREPLY array variable.
\n-G globpat
\nThe pathname expansion pattern globpat is expanded to
\ngenerate the possible completions.
\n-P prefix
\nprefix is added at the beginning of each possible
\ncompletion after all other options have been applied.
\n-S suffix
\nsuffix is appended to each possible completion after
\nall other options have been applied.
\n-W wordlist
\nThe wordlist is split using the characters in the IFS
\nspecial variable as delimiters, and each resultant
\nword is expanded.\u00a0 The possible completions are the
\nmembers of the resultant list which match the word
\nbeing completed.
\n-X filterpat
\nfilterpat is a pattern as used for pathname expansion.
\nIt is applied to the list of possible completions
\ngenerated by the preceding options and arguments, and
\neach completion matching filterpat is removed from the
\nlist.\u00a0 A leading ! in filterpat negates the pattern;
\nin this case, any completion not matching filterpat is
\nremoved.<\/p>\n
The return value is true unless an invalid option is supplied,
\nan option other than -p or -r is supplied without a name
\nargument, an attempt is made to remove a completion
\nspecification for a name for which no specification exists, or
\nan error occurs adding a completion specification.<\/p>\n
compopt [-o option] [-DE] [+o option] [name]
\nModify completion options for each name according to the
\noptions, or for the currently-executing completion if no names
\nare supplied.\u00a0 If no options are given, display the completion
\noptions for each name or the current completion.\u00a0 The possible
\nvalues of option are those valid for the complete builtin
\ndescribed above.\u00a0 The -D option indicates that the remaining
\noptions should apply to the “default” command completion;
\nthat is, completion attempted on a command for which no
\ncompletion has previously been defined.\u00a0 The -E option
\nindicates that the remaining options should apply to “empty”
\ncommand completion; that is, completion attempted on a blank
\nline.<\/p>\n
The return value is true unless an invalid option is supplied,
\nan attempt is made to modify the options for a name for which
\nno completion specification exists, or an output error occurs.<\/p>\n
continue [n]
\nResume the next iteration of the enclosing for, while, until,
\nor select loop.\u00a0 If n is specified, resume at the nth
\nenclosing loop.\u00a0 n must be \u2265 1.\u00a0 If n is greater than the
\nnumber of enclosing loops, the last enclosing loop (the “top-
\nlevel” loop) is resumed.\u00a0 The return value is 0 unless n is
\nnot greater than or equal to 1.<\/p>\n
declare [-aAfFgilnrtux] [-p] [name[=value] …]
\ntypeset [-aAfFgilnrtux] [-p] [name[=value] …]
\nDeclare variables and\/or give them attributes.\u00a0 If no names
\nare given then display the values of variables.\u00a0 The -p option
\nwill display the attributes and values of each name.\u00a0 When -p
\nis used with name arguments, additional options, other than -f
\nand -F, are ignored.\u00a0 When -p is supplied without name
\narguments, it will display the attributes and values of all
\nvariables having the attributes specified by the additional
\noptions.\u00a0 If no other options are supplied with -p, declare
\nwill display the attributes and values of all shell variables.
\nThe -f option will restrict the display to shell functions.
\nThe -F option inhibits the display of function definitions;
\nonly the function name and attributes are printed.\u00a0 If the
\nextdebug shell option is enabled using shopt, the source file
\nname and line number where the function is defined are
\ndisplayed as well.\u00a0 The -F option implies -f.\u00a0 The -g option
\nforces variables to be created or modified at the global
\nscope, even when declare is executed in a shell function.\u00a0 It
\nis ignored in all other cases.\u00a0 The following options can be
\nused to restrict output to variables with the specified
\nattribute or to give variables attributes:
\n-a\u00a0\u00a0\u00a0\u00a0 Each name is an indexed array variable (see Arrays
\nabove).
\n-A\u00a0\u00a0\u00a0\u00a0 Each name is an associative array variable (see Arrays
\nabove).
\n-f\u00a0\u00a0\u00a0\u00a0 Use function names only.
\n-i\u00a0\u00a0\u00a0\u00a0 The variable is treated as an integer; arithmetic
\nevaluation (see ARITHMETIC EVALUATION above) is
\nperformed when the variable is assigned a value.
\n-l\u00a0\u00a0\u00a0\u00a0 When the variable is assigned a value, all upper-case
\ncharacters are converted to lower-case.\u00a0 The upper-case
\nattribute is disabled.
\n-n\u00a0\u00a0\u00a0\u00a0 Give each name the nameref attribute, making it a name
\nreference to another variable.\u00a0 That other variable is
\ndefined by the value of name.\u00a0 All references and
\nassignments to name, except for changing the -n
\nattribute itself, are performed on the variable
\nreferenced by name’s value.\u00a0 The -n attribute cannot be
\napplied to array variables.
\n-r\u00a0\u00a0\u00a0\u00a0 Make names readonly.\u00a0 These names cannot then be
\nassigned values by subsequent assignment statements or
\nunset.
\n-t\u00a0\u00a0\u00a0\u00a0 Give each name the trace attribute.\u00a0 Traced functions
\ninherit the DEBUG and RETURN traps from the calling
\nshell.\u00a0 The trace attribute has no special meaning for
\nvariables.
\n-u\u00a0\u00a0\u00a0\u00a0 When the variable is assigned a value, all lower-case
\ncharacters are converted to upper-case.\u00a0 The lower-case
\nattribute is disabled.
\n-x\u00a0\u00a0\u00a0\u00a0 Mark names for export to subsequent commands via the
\nenvironment.<\/p>\n
Using `+’ instead of `-‘ turns off the attribute instead, with
\nthe exceptions that +a may not be used to destroy an array
\nvariable and +r will not remove the readonly attribute.\u00a0 When
\nused in a function, declare and typeset make each name local,
\nas with the local command, unless the -g option is supplied.
\nIf a variable name is followed by =value, the value of the
\nvariable is set to value.\u00a0 When using -a or -A and the
\ncompound assignment syntax to create array variables,
\nadditional attributes do not take effect until subsequent
\nassignments.\u00a0 The return value is 0 unless an invalid option
\nis encountered, an attempt is made to define a function using
\n“-f foo=bar”, an attempt is made to assign a value to a
\nreadonly variable, an attempt is made to assign a value to an
\narray variable without using the compound assignment syntax
\n(see Arrays above), one of the names is not a valid shell
\nvariable name, an attempt is made to turn off readonly status
\nfor a readonly variable, an attempt is made to turn off array
\nstatus for an array variable, or an attempt is made to display
\na non-existent function with -f.<\/p>\n
dirs [-clpv] [+n] [-n]
\nWithout options, displays the list of currently remembered
\ndirectories.\u00a0 The default display is on a single line with
\ndirectory names separated by spaces.\u00a0 Directories are added to
\nthe list with the pushd command; the popd command removes
\nentries from the list.
\n-c\u00a0\u00a0\u00a0\u00a0 Clears the directory stack by deleting all of the
\nentries.
\n-l\u00a0\u00a0\u00a0\u00a0 Produces a listing using full pathnames; the default
\nlisting format uses a tilde to denote the home
\ndirectory.
\n-p\u00a0\u00a0\u00a0\u00a0 Print the directory stack with one entry per line.
\n-v\u00a0\u00a0\u00a0\u00a0 Print the directory stack with one entry per line,
\nprefixing each entry with its index in the stack.
\n+n\u00a0\u00a0\u00a0\u00a0 Displays the nth entry counting from the left of the
\nlist shown by dirs when invoked without options,
\nstarting with zero.
\n-n\u00a0\u00a0\u00a0\u00a0 Displays the nth entry counting from the right of the
\nlist shown by dirs when invoked without options,
\nstarting with zero.<\/p>\n
The return value is 0 unless an invalid option is supplied or
\nn indexes beyond the end of the directory stack.<\/p>\n
disown [-ar] [-h] [jobspec …]
\nWithout options, remove each jobspec from the table of active
\njobs.\u00a0 If jobspec is not present, and neither the -a nor the
\n-r option is supplied, the current job is used.\u00a0 If the -h
\noption is given, each jobspec is not removed from the table,
\nbut is marked so that SIGHUP is not sent to the job if the
\nshell receives a SIGHUP.\u00a0 If no jobspec is supplied, the -a
\noption means to remove or mark all jobs; the -r option without
\na jobspec argument restricts operation to running jobs.\u00a0 The
\nreturn value is 0 unless a jobspec does not specify a valid
\njob.<\/p>\n
echo [-neE] [arg …]
\nOutput the args, separated by spaces, followed by a newline.
\nThe return status is 0 unless a write error occurs.\u00a0 If -n is
\nspecified, the trailing newline is suppressed.\u00a0 If the -e
\noption is given, interpretation of the following backslash-
\nescaped characters is enabled.\u00a0 The -E option disables the
\ninterpretation of these escape characters, even on systems
\nwhere they are interpreted by default.\u00a0 The xpg_echo shell
\noption may be used to dynamically determine whether or not
\necho expands these escape characters by default.\u00a0 echo does
\nnot interpret — to mean the end of options.\u00a0 echo interprets
\nthe following escape sequences:
\n\\a\u00a0\u00a0\u00a0\u00a0 alert (bell)
\n\\b\u00a0\u00a0\u00a0\u00a0 backspace
\n\\c\u00a0\u00a0\u00a0\u00a0 suppress further output
\n\\e
\n\\E\u00a0\u00a0\u00a0\u00a0 an escape character
\n\\f\u00a0\u00a0\u00a0\u00a0 form feed
\n\\n\u00a0\u00a0\u00a0\u00a0 new line
\n\\r\u00a0\u00a0\u00a0\u00a0 carriage return
\n\\t\u00a0\u00a0\u00a0\u00a0 horizontal tab
\n\\v\u00a0\u00a0\u00a0\u00a0 vertical tab
\n\\\\\u00a0\u00a0\u00a0\u00a0 backslash
\n\\0nnn\u00a0 the eight-bit character whose value is the octal value
\nnnn (zero to three octal digits)
\n\\xHH\u00a0\u00a0 the eight-bit character whose value is the hexadecimal
\nvalue HH (one or two hex digits)
\n\\uHHHH the Unicode (ISO\/IEC 10646) character whose value is
\nthe hexadecimal value HHHH (one to four hex digits)
\n\\UHHHHHHHH
\nthe Unicode (ISO\/IEC 10646) character whose value is
\nthe hexadecimal value HHHHHHHH (one to eight hex
\ndigits)<\/p>\n
enable [-a] [-dnps] [-f filename] [name …]
\nEnable and disable builtin shell commands.\u00a0 Disabling a
\nbuiltin allows a disk command which has the same name as a
\nshell builtin to be executed without specifying a full
\npathname, even though the shell normally searches for builtins
\nbefore disk commands.\u00a0 If -n is used, each name is disabled;
\notherwise, names are enabled.\u00a0 For example, to use the test
\nbinary found via the PATH instead of the shell builtin
\nversion, run “enable -n test”.\u00a0 The -f option means to load
\nthe new builtin command name from shared object filename, on
\nsystems that support dynamic loading.\u00a0 The -d option will
\ndelete a builtin previously loaded with -f.\u00a0 If no name
\narguments are given, or if the -p option is supplied, a list
\nof shell builtins is printed.\u00a0 With no other option arguments,
\nthe list consists of all enabled shell builtins.\u00a0 If -n is
\nsupplied, only disabled builtins are printed.\u00a0 If -a is
\nsupplied, the list printed includes all builtins, with an
\nindication of whether or not each is enabled.\u00a0 If -s is
\nsupplied, the output is restricted to the POSIX special
\nbuiltins.\u00a0 The return value is 0 unless a name is not a shell
\nbuiltin or there is an error loading a new builtin from a
\nshared object.<\/p>\n
eval [arg …]
\nThe args are read and concatenated together into a single
\ncommand.\u00a0 This command is then read and executed by the shell,
\nand its exit status is returned as the value of eval.\u00a0 If
\nthere are no args, or only null arguments, eval returns 0.<\/p>\n
exec [-cl] [-a name] [command [arguments]]
\nIf command is specified, it replaces the shell.\u00a0 No new
\nprocess is created.\u00a0 The arguments become the arguments to
\ncommand.\u00a0 If the -l option is supplied, the shell places a
\ndash at the beginning of the zeroth argument passed to
\ncommand.\u00a0 This is what login(1) does.\u00a0 The -c option causes
\ncommand to be executed with an empty environment.\u00a0 If -a is
\nsupplied, the shell passes name as the zeroth argument to the
\nexecuted command.\u00a0 If command cannot be executed for some
\nreason, a non-interactive shell exits, unless the execfail
\nshell option is enabled.\u00a0 In that case, it returns failure.
\nAn interactive shell returns failure if the file cannot be
\nexecuted.\u00a0 If command is not specified, any redirections take
\neffect in the current shell, and the return status is 0.\u00a0 If
\nthere is a redirection error, the return status is 1.<\/p>\n
exit [n]
\nCause the shell to exit with a status of n.\u00a0 If n is omitted,
\nthe exit status is that of the last command executed.\u00a0 A trap
\non EXIT is executed before the shell terminates.<\/p>\n
export [-fn] [name[=word]] …
\nexport -p
\nThe supplied names are marked for automatic export to the
\nenvironment of subsequently executed commands.\u00a0 If the -f
\noption is given, the names refer to functions.\u00a0 If no names
\nare given, or if the -p option is supplied, a list of names of
\nall exported variables is printed.\u00a0 The -n option causes the
\nexport property to be removed from each name.\u00a0 If a variable
\nname is followed by =word, the value of the variable is set to
\nword.\u00a0 export returns an exit status of 0 unless an invalid
\noption is encountered, one of the names is not a valid shell
\nvariable name, or -f is supplied with a name that is not a
\nfunction.<\/p>\n
fc [-e ename] [-lnr] [first] [last]
\nfc -s [pat=rep] [cmd]
\nThe first form selects a range of commands from first to last
\nfrom the history list and displays or edits and re-executes
\nthem.\u00a0 First and last may be specified as a string (to locate
\nthe last command beginning with that string) or as a number
\n(an index into the history list, where a negative number is
\nused as an offset from the current command number).\u00a0 If last
\nis not specified it is set to the current command for listing
\n(so that “fc -l -10” prints the last 10 commands) and to
\nfirst otherwise.\u00a0 If first is not specified it is set to the
\nprevious command for editing and -16 for listing.<\/p>\n
The -n option suppresses the command numbers when listing.
\nThe -r option reverses the order of the commands.\u00a0 If the -l
\noption is given, the commands are listed on standard output.
\nOtherwise, the editor given by ename is invoked on a file
\ncontaining those commands.\u00a0 If ename is not given, the value
\nof the FCEDIT variable is used, and the value of EDITOR if
\nFCEDIT is not set.\u00a0 If neither variable is set, vi is used.
\nWhen editing is complete, the edited commands are echoed and
\nexecuted.<\/p>\n
In the second form, command is re-executed after each instance
\nof pat is replaced by rep.\u00a0 Command is intepreted the same as
\nfirst above.\u00a0 A useful alias to use with this is “r=”fc
\n-s””, so that typing “r cc” runs the last command beginning
\nwith “cc” and typing “r” re-executes the last command.<\/p>\n
If the first form is used, the return value is 0 unless an
\ninvalid option is encountered or first or last specify history
\nlines out of range.\u00a0 If the -e option is supplied, the return
\nvalue is the value of the last command executed or failure if
\nan error occurs with the temporary file of commands.\u00a0 If the
\nsecond form is used, the return status is that of the command
\nre-executed, unless cmd does not specify a valid history line,
\nin which case fc returns failure.<\/p>\n
fg [jobspec]
\nResume jobspec in the foreground, and make it the current job.
\nIf jobspec is not present, the shell’s notion of the current
\njob is used.\u00a0 The return value is that of the command placed
\ninto the foreground, or failure if run when job control is
\ndisabled or, when run with job control enabled, if jobspec
\ndoes not specify a valid job or jobspec specifies a job that
\nwas started without job control.<\/p>\n
getopts optstring name [args]
\ngetopts is used by shell procedures to parse positional
\nparameters.\u00a0 optstring contains the option characters to be
\nrecognized; if a character is followed by a colon, the option
\nis expected to have an argument, which should be separated
\nfrom it by white space.\u00a0 The colon and question mark
\ncharacters may not be used as option characters.\u00a0 Each time it
\nis invoked, getopts places the next option in the shell
\nvariable name, initializing name if it does not exist, and the
\nindex of the next argument to be processed into the variable
\nOPTIND.\u00a0 OPTIND is initialized to 1 each time the shell or a
\nshell script is invoked.\u00a0 When an option requires an argument,
\ngetopts places that argument into the variable OPTARG.\u00a0 The
\nshell does not reset OPTIND automatically; it must be manually
\nreset between multiple calls to getopts within the same shell
\ninvocation if a new set of parameters is to be used.<\/p>\n
When the end of options is encountered, getopts exits with a
\nreturn value greater than zero.\u00a0 OPTIND is set to the index of
\nthe first non-option argument, and name is set to ?.<\/p>\n
getopts normally parses the positional parameters, but if more
\narguments are given in args, getopts parses those instead.<\/p>\n
getopts can report errors in two ways.\u00a0 If the first character
\nof optstring is a colon, silent error reporting is used.\u00a0 In
\nnormal operation, diagnostic messages are printed when invalid
\noptions or missing option arguments are encountered.\u00a0 If the
\nvariable OPTERR is set to 0, no error messages will be
\ndisplayed, even if the first character of optstring is not a
\ncolon.<\/p>\n
If an invalid option is seen, getopts places ? into name and,
\nif not silent, prints an error message and unsets OPTARG.\u00a0 If
\ngetopts is silent, the option character found is placed in
\nOPTARG and no diagnostic message is printed.<\/p>\n
If a required argument is not found, and getopts is not
\nsilent, a question mark (?) is placed in name, OPTARG is
\nunset, and a diagnostic message is printed.\u00a0 If getopts is
\nsilent, then a colon (:) is placed in name and OPTARG is set
\nto the option character found.<\/p>\n
getopts returns true if an option, specified or unspecified,
\nis found.\u00a0 It returns false if the end of options is
\nencountered or an error occurs.<\/p>\n
hash [-lr] [-p filename] [-dt] [name]
\nEach time hash is invoked, the full pathname of the command
\nname is determined by searching the directories in $PATH and
\nremembered.\u00a0 Any previously-remembered pathname is discarded.
\nIf the -p option is supplied, no path search is performed, and
\nfilename is used as the full filename of the command.\u00a0 The -r
\noption causes the shell to forget all remembered locations.
\nThe -d option causes the shell to forget the remembered
\nlocation of each name.\u00a0 If the -t option is supplied, the full
\npathname to which each name corresponds is printed.\u00a0 If
\nmultiple name arguments are supplied with -t, the name is
\nprinted before the hashed full pathname.\u00a0 The -l option causes
\noutput to be displayed in a format that may be reused as
\ninput.\u00a0 If no arguments are given, or if only -l is supplied,
\ninformation about remembered commands is printed.\u00a0 The return
\nstatus is true unless a name is not found or an invalid option
\nis supplied.<\/p>\n
help [-dms] [pattern]
\nDisplay helpful information about builtin commands.\u00a0 If
\npattern is specified, help gives detailed help on all commands
\nmatching pattern; otherwise help for all the builtins and
\nshell control structures is printed.
\n-d\u00a0\u00a0\u00a0\u00a0 Display a short description of each pattern
\n-m\u00a0\u00a0\u00a0\u00a0 Display the description of each pattern in a manpage-
\nlike format
\n-s\u00a0\u00a0\u00a0\u00a0 Display only a short usage synopsis for each pattern<\/p>\n
The return status is 0 unless no command matches pattern.<\/p>\n
history [n]
\nhistory -c
\nhistory -d offset
\nhistory -anrw [filename]
\nhistory -p arg [arg …]
\nhistory -s arg [arg …]
\nWith no options, display the command history list with line
\nnumbers.\u00a0 Lines listed with a * have been modified.\u00a0 An
\nargument of n lists only the last n lines.\u00a0 If the shell
\nvariable HISTTIMEFORMAT is set and not null, it is used as a
\nformat string for strftime(3) to display the time stamp
\nassociated with each displayed history entry.\u00a0 No intervening
\nblank is printed between the formatted time stamp and the
\nhistory line.\u00a0 If filename is supplied, it is used as the name
\nof the history file; if not, the value of HISTFILE is used.
\nOptions, if supplied, have the following meanings:
\n-c\u00a0\u00a0\u00a0\u00a0 Clear the history list by deleting all the entries.
\n-d offset
\nDelete the history entry at position offset.
\n-a\u00a0\u00a0\u00a0\u00a0 Append the “new” history lines (history lines entered
\nsince the beginning of the current bash session) to the
\nhistory file.
\n-n\u00a0\u00a0\u00a0\u00a0 Read the history lines not already read from the
\nhistory file into the current history list.\u00a0 These are
\nlines appended to the history file since the beginning
\nof the current bash session.
\n-r\u00a0\u00a0\u00a0\u00a0 Read the contents of the history file and append them
\nto the current history list.
\n-w\u00a0\u00a0\u00a0\u00a0 Write the current history list to the history file,
\noverwriting the history file’s contents.
\n-p\u00a0\u00a0\u00a0\u00a0 Perform history substitution on the following args and
\ndisplay the result on the standard output.\u00a0 Does not
\nstore the results in the history list.\u00a0 Each arg must
\nbe quoted to disable normal history expansion.
\n-s\u00a0\u00a0\u00a0\u00a0 Store the args in the history list as a single entry.
\nThe last command in the history list is removed before
\nthe args are added.<\/p>\n
If the HISTTIMEFORMAT variable is set, the time stamp
\ninformation associated with each history entry is written to
\nthe history file, marked with the history comment character.
\nWhen the history file is read, lines beginning with the
\nhistory comment character followed immediately by a digit are
\ninterpreted as timestamps for the previous history line.\u00a0 The
\nreturn value is 0 unless an invalid option is encountered, an
\nerror occurs while reading or writing the history file, an
\ninvalid offset is supplied as an argument to -d, or the
\nhistory expansion supplied as an argument to -p fails.<\/p>\n
jobs [-lnprs] [ jobspec … ]
\njobs -x command [ args … ]
\nThe first form lists the active jobs.\u00a0 The options have the
\nfollowing meanings:
\n-l\u00a0\u00a0\u00a0\u00a0 List process IDs in addition to the normal information.
\n-n\u00a0\u00a0\u00a0\u00a0 Display information only about jobs that have changed
\nstatus since the user was last notified of their
\nstatus.
\n-p\u00a0\u00a0\u00a0\u00a0 List only the process ID of the job’s process group
\nleader.
\n-r\u00a0\u00a0\u00a0\u00a0 Display only running jobs.
\n-s\u00a0\u00a0\u00a0\u00a0 Display only stopped jobs.<\/p>\n
If jobspec is given, output is restricted to information about
\nthat job.\u00a0 The return status is 0 unless an invalid option is
\nencountered or an invalid jobspec is supplied.<\/p>\n
If the -x option is supplied, jobs replaces any jobspec found
\nin command or args with the corresponding process group ID,
\nand executes command passing it args, returning its exit
\nstatus.<\/p>\n
kill [-s sigspec | -n signum | -sigspec] [pid | jobspec] …
\nkill -l [sigspec | exit_status]
\nSend the signal named by sigspec or signum to the processes
\nnamed by pid or jobspec.\u00a0 sigspec is either a case-insensitive
\nsignal name such as SIGKILL (with or without the SIG prefix)
\nor a signal number; signum is a signal number.\u00a0 If sigspec is
\nnot present, then SIGTERM is assumed.\u00a0 An argument of -l lists
\nthe signal names.\u00a0 If any arguments are supplied when -l is
\ngiven, the names of the signals corresponding to the arguments
\nare listed, and the return status is 0.\u00a0 The exit_status
\nargument to -l is a number specifying either a signal number
\nor the exit status of a process terminated by a signal.\u00a0 kill
\nreturns true if at least one signal was successfully sent, or
\nfalse if an error occurs or an invalid option is encountered.<\/p>\n
let arg [arg …]
\nEach arg is an arithmetic expression to be evaluated (see
\nARITHMETIC EVALUATION above).\u00a0 If the last arg evaluates to 0,
\nlet returns 1; 0 is returned otherwise.<\/p>\n
local [option] [name[=value] …]
\nFor each argument, a local variable named name is created, and
\nassigned value.\u00a0 The option can be any of the options accepted
\nby declare.\u00a0 When local is used within a function, it causes
\nthe variable name to have a visible scope restricted to that
\nfunction and its children.\u00a0 With no operands, local writes a
\nlist of local variables to the standard output.\u00a0 It is an
\nerror to use local when not within a function.\u00a0 The return
\nstatus is 0 unless local is used outside a function, an
\ninvalid name is supplied, or name is a readonly variable.<\/p>\n
logout Exit a login shell.<\/p>\n
mapfile [-n count] [-O origin] [-s count] [-t] [-u fd] [-C callback]
\n[-c quantum] [array]
\nreadarray [-n count] [-O origin] [-s count] [-t] [-u fd] [-C
\ncallback] [-c quantum] [array]
\nRead lines from the standard input into the indexed array
\nvariable array, or from file descriptor fd if the -u option is
\nsupplied.\u00a0 The variable MAPFILE is the default array.
\nOptions, if supplied, have the following meanings:
\n-n\u00a0\u00a0\u00a0\u00a0 Copy at most count lines.\u00a0 If count is 0, all lines are
\ncopied.
\n-O\u00a0\u00a0\u00a0\u00a0 Begin assigning to array at index origin.\u00a0 The default
\nindex is 0.
\n-s\u00a0\u00a0\u00a0\u00a0 Discard the first count lines read.
\n-t\u00a0\u00a0\u00a0\u00a0 Remove a trailing newline from each line read.
\n-u\u00a0\u00a0\u00a0\u00a0 Read lines from file descriptor fd instead of the
\nstandard input.
\n-C\u00a0\u00a0\u00a0\u00a0 Evaluate callback each time quantum lines are read.
\nThe -c option specifies quantum.
\n-c\u00a0\u00a0\u00a0\u00a0 Specify the number of lines read between each call to
\ncallback.<\/p>\n
If -C is specified without -c, the default quantum is 5000.
\nWhen callback is evaluated, it is supplied the index of the
\nnext array element to be assigned and the line to be assigned
\nto that element as additional arguments.\u00a0 callback is
\nevaluated after the line is read but before the array element
\nis assigned.<\/p>\n
If not supplied with an explicit origin, mapfile will clear
\narray before assigning to it.<\/p>\n
mapfile returns successfully unless an invalid option or
\noption argument is supplied, array is invalid or unassignable,
\nor if array is not an indexed array.<\/p>\n
popd [-n] [+n] [-n]
\nRemoves entries from the directory stack.\u00a0 With no arguments,
\nremoves the top directory from the stack, and performs a cd to
\nthe new top directory.\u00a0 Arguments, if supplied, have the
\nfollowing meanings:
\n-n\u00a0\u00a0\u00a0\u00a0 Suppresses the normal change of directory when removing
\ndirectories from the stack, so that only the stack is
\nmanipulated.
\n+n\u00a0\u00a0\u00a0\u00a0 Removes the nth entry counting from the left of the
\nlist shown by dirs, starting with zero.\u00a0 For example:
\n“popd +0” removes the first directory, “popd +1”
\nthe second.
\n-n\u00a0\u00a0\u00a0\u00a0 Removes the nth entry counting from the right of the
\nlist shown by dirs, starting with zero.\u00a0 For example:
\n“popd -0” removes the last directory, “popd -1” the
\nnext to last.<\/p>\n
If the popd command is successful, a dirs is performed as
\nwell, and the return status is 0.\u00a0 popd returns false if an
\ninvalid option is encountered, the directory stack is empty, a
\nnon-existent directory stack entry is specified, or the
\ndirectory change fails.<\/p>\n
printf [-v var] format [arguments]
\nWrite the formatted arguments to the standard output under the
\ncontrol of the format.\u00a0 The -v option causes the output to be
\nassigned to the variable var rather than being printed to the
\nstandard output.<\/p>\n
The format is a character string which contains three types of
\nobjects: plain characters, which are simply copied to standard
\noutput, character escape sequences, which are converted and
\ncopied to the standard output, and format specifications, each
\nof which causes printing of the next successive argument.\u00a0 In
\naddition to the standard printf(1) format specifications,
\nprintf interprets the following extensions:
\n%b\u00a0\u00a0\u00a0\u00a0 causes printf to expand backslash escape sequences in
\nthe corresponding argument (except that \\c terminates
\noutput, backslashes in \\’, \\”, and \\? are not removed,
\nand octal escapes beginning with \\0 may contain up to
\nfour digits).
\n%q\u00a0\u00a0\u00a0\u00a0 causes printf to output the corresponding argument in a
\nformat that can be reused as shell input.
\n%(datefmt)T
\ncauses printf to output the date-time string resulting
\nfrom using datefmt as a format string for strftime(3).
\nThe corresponding argument is an integer representing
\nthe number of seconds since the epoch.\u00a0 Two special
\nargument values may be used: -1 represents the current
\ntime, and -2 represents the time the shell was invoked.
\nIf no argument is specified, conversion behaves as if
\n-1 had been given.\u00a0 This is an exception to the usual
\nprintf behavior.<\/p>\n
Arguments to non-string format specifiers are treated as C
\nconstants, except that a leading plus or minus sign is
\nallowed, and if the leading character is a single or double
\nquote, the value is the ASCII value of the following
\ncharacter.<\/p>\n
The format is reused as necessary to consume all of the
\narguments.\u00a0 If the format requires more arguments than are
\nsupplied, the extra format specifications behave as if a zero
\nvalue or null string, as appropriate, had been supplied.\u00a0 The
\nreturn value is zero on success, non-zero on failure.<\/p>\n
pushd [-n] [+n] [-n]
\npushd [-n] [dir]
\nAdds a directory to the top of the directory stack, or rotates
\nthe stack, making the new top of the stack the current working
\ndirectory.\u00a0 With no arguments, exchanges the top two
\ndirectories and returns 0, unless the directory stack is
\nempty.\u00a0 Arguments, if supplied, have the following meanings:
\n-n\u00a0\u00a0\u00a0\u00a0 Suppresses the normal change of directory when adding
\ndirectories to the stack, so that only the stack is
\nmanipulated.
\n+n\u00a0\u00a0\u00a0\u00a0 Rotates the stack so that the nth directory (counting
\nfrom the left of the list shown by dirs, starting with
\nzero) is at the top.
\n-n\u00a0\u00a0\u00a0\u00a0 Rotates the stack so that the nth directory (counting
\nfrom the right of the list shown by dirs, starting with
\nzero) is at the top.
\ndir\u00a0\u00a0\u00a0 Adds dir to the directory stack at the top, making it
\nthe new current working directory as if it had been
\nsupplied as the argument to the cd builtin.<\/p>\n
If the pushd command is successful, a dirs is performed as
\nwell.\u00a0 If the first form is used, pushd returns 0 unless the
\ncd to dir fails.\u00a0 With the second form, pushd returns 0 unless
\nthe directory stack is empty, a non-existent directory stack
\nelement is specified, or the directory change to the specified
\nnew current directory fails.<\/p>\n
pwd [-LP]
\nPrint the absolute pathname of the current working directory.
\nThe pathname printed contains no symbolic links if the -P
\noption is supplied or the -o physical option to the set
\nbuiltin command is enabled.\u00a0 If the -L option is used, the
\npathname printed may contain symbolic links.\u00a0 The return
\nstatus is 0 unless an error occurs while reading the name of
\nthe current directory or an invalid option is supplied.<\/p>\n
read [-ers] [-a aname] [-d delim] [-i text] [-n nchars] [-N nchars]
\n[-p prompt] [-t timeout] [-u fd] [name …]
\nOne line is read from the standard input, or from the file
\ndescriptor fd supplied as an argument to the -u option, and
\nthe first word is assigned to the first name, the second word
\nto the second name, and so on, with leftover words and their
\nintervening separators assigned to the last name.\u00a0 If there
\nare fewer words read from the input stream than names, the
\nremaining names are assigned empty values.\u00a0 The characters in
\nIFS are used to split the line into words using the same rules
\nthe shell uses for expansion (described above under Word
\nSplitting).\u00a0 The backslash character (\\) may be used to remove
\nany special meaning for the next character read and for line
\ncontinuation.\u00a0 Options, if supplied, have the following
\nmeanings:
\n-a aname
\nThe words are assigned to sequential indices of the
\narray variable aname, starting at 0.\u00a0 aname is unset
\nbefore any new values are assigned.\u00a0 Other name
\narguments are ignored.
\n-d delim
\nThe first character of delim is used to terminate the
\ninput line, rather than newline.
\n-e\u00a0\u00a0\u00a0\u00a0 If the standard input is coming from a terminal,
\nreadline (see READLINE above) is used to obtain the
\nline.\u00a0 Readline uses the current (or default, if line
\nediting was not previously active) editing settings.
\n-i text
\nIf readline is being used to read the line, text is
\nplaced into the editing buffer before editing begins.
\n-n nchars
\nread returns after reading nchars characters rather
\nthan waiting for a complete line of input, but honor a
\ndelimiter if fewer than nchars characters are read
\nbefore the delimiter.
\n-N nchars
\nread returns after reading exactly nchars characters
\nrather than waiting for a complete line of input,
\nunless EOF is encountered or read times out.\u00a0 Delimiter
\ncharacters encountered in the input are not treated
\nspecially and do not cause read to return until nchars
\ncharacters are read.
\n-p prompt
\nDisplay prompt on standard error, without a trailing
\nnewline, before attempting to read any input.\u00a0 The
\nprompt is displayed only if input is coming from a
\nterminal.
\n-r\u00a0\u00a0\u00a0\u00a0 Backslash does not act as an escape character.\u00a0 The
\nbackslash is considered to be part of the line.\u00a0 In
\nparticular, a backslash-newline pair may not be used as
\na line continuation.
\n-s\u00a0\u00a0\u00a0\u00a0 Silent mode.\u00a0 If input is coming from a terminal,
\ncharacters are not echoed.
\n-t timeout
\nCause read to time out and return failure if a complete
\nline of input (or a specified number of characters) is
\nnot read within timeout seconds.\u00a0 timeout may be a
\ndecimal number with a fractional portion following the
\ndecimal point.\u00a0 This option is only effective if read
\nis reading input from a terminal, pipe, or other
\nspecial file; it has no effect when reading from
\nregular files.\u00a0 If read times out, read saves any
\npartial input read into the specified variable name.
\nIf timeout is 0, read returns immediately, without
\ntrying to read any data.\u00a0 The exit status is 0 if input
\nis available on the specified file descriptor, non-zero
\notherwise.\u00a0 The exit status is greater than 128 if the
\ntimeout is exceeded.
\n-u fd\u00a0 Read input from file descriptor fd.<\/p>\n
If no names are supplied, the line read is assigned to the
\nvariable REPLY.\u00a0 The return code is zero, unless end-of-file
\nis encountered, read times out (in which case the return code
\nis greater than 128), a variable assignment error (such as
\nassigning to a readonly variable) occurs, or an invalid file
\ndescriptor is supplied as the argument to -u.<\/p>\n
readonly [-aAf] [-p] [name[=word] …]
\nThe given names are marked readonly; the values of these names
\nmay not be changed by subsequent assignment.\u00a0 If the -f option
\nis supplied, the functions corresponding to the names are so
\nmarked.\u00a0 The -a option restricts the variables to indexed
\narrays; the -A option restricts the variables to associative
\narrays.\u00a0 If both options are supplied, -A takes precedence.
\nIf no name arguments are given, or if the -p option is
\nsupplied, a list of all readonly names is printed.\u00a0 The other
\noptions may be used to restrict the output to a subset of the
\nset of readonly names.\u00a0 The -p option causes output to be
\ndisplayed in a format that may be reused as input.\u00a0 If a
\nvariable name is followed by =word, the value of the variable
\nis set to word.\u00a0 The return status is 0 unless an invalid
\noption is encountered, one of the names is not a valid shell
\nvariable name, or -f is supplied with a name that is not a
\nfunction.<\/p>\n
return [n]
\nCauses a function to stop executing and return the value
\nspecified by n to its caller.\u00a0 If n is omitted, the return
\nstatus is that of the last command executed in the function
\nbody.\u00a0 If return is used outside a function, but during
\nexecution of a script by the .\u00a0 (source) command, it causes
\nthe shell to stop executing that script and return either n or
\nthe exit status of the last command executed within the script
\nas the exit status of the script.\u00a0 If n is supplied, the
\nreturn value is its least significant 8 bits.\u00a0 The return
\nstatus is non-zero if return is supplied a non-numeric
\nargument, or is used outside a function and not during
\nexecution of a script by . or source.\u00a0 Any command associated
\nwith the RETURN trap is executed before execution resumes
\nafter the function or script.<\/p>\n
set [–abefhkmnptuvxBCEHPT] [-o option-name] [arg …]
\nset [+abefhkmnptuvxBCEHPT] [+o option-name] [arg …]
\nWithout options, the name and value of each shell variable are
\ndisplayed in a format that can be reused as input for setting
\nor resetting the currently-set variables.\u00a0 Read-only variables
\ncannot be reset.\u00a0 In posix mode, only shell variables are
\nlisted.\u00a0 The output is sorted according to the current locale.
\nWhen options are specified, they set or unset shell
\nattributes.\u00a0 Any arguments remaining after option processing
\nare treated as values for the positional parameters and are
\nassigned, in order, to $1, $2, …\u00a0 $n.\u00a0 Options, if
\nspecified, have the following meanings:
\n-a\u00a0\u00a0\u00a0\u00a0\u00a0 Automatically mark variables and functions which are
\nmodified or created for export to the environment of
\nsubsequent commands.
\n-b\u00a0\u00a0\u00a0\u00a0\u00a0 Report the status of terminated background jobs
\nimmediately, rather than before the next primary
\nprompt.\u00a0 This is effective only when job control is
\nenabled.
\n-e\u00a0\u00a0\u00a0\u00a0\u00a0 Exit immediately if a pipeline (which may consist of a
\nsingle simple command), a list, or a compound command
\n(see SHELL GRAMMAR above),\u00a0 exits with a non-zero
\nstatus.\u00a0 The shell does not exit if the command that
\nfails is part of the command list immediately
\nfollowing a while or until keyword, part of the test
\nfollowing the if or elif reserved words, part of any
\ncommand executed in a && or || list except the command
\nfollowing the final && or ||, any command in a
\npipeline but the last, or if the command’s return
\nvalue is being inverted with !.\u00a0 If a compound command
\nother than a subshell returns a non-zero status
\nbecause a command failed while -e was being ignored,
\nthe shell does not exit.\u00a0 A trap on ERR, if set, is
\nexecuted before the shell exits.\u00a0 This option applies
\nto the shell environment and each subshell environment
\nseparately (see COMMAND EXECUTION ENVIRONMENT above),
\nand may cause subshells to exit before executing all
\nthe commands in the subshell.<\/p>\n
If a compound command or shell function executes in a
\ncontext where -e is being ignored, none of the
\ncommands executed within the compound command or
\nfunction body will be affected by the -e setting, even
\nif -e is set and a command returns a failure status.
\nIf a compound command or shell function sets -e while
\nexecuting in a context where -e is ignored, that
\nsetting will not have any effect until the compound
\ncommand or the command containing the function call
\ncompletes.
\n-f\u00a0\u00a0\u00a0\u00a0\u00a0 Disable pathname expansion.
\n-h\u00a0\u00a0\u00a0\u00a0\u00a0 Remember the location of commands as they are looked
\nup for execution.\u00a0 This is enabled by default.
\n-k\u00a0\u00a0\u00a0\u00a0\u00a0 All arguments in the form of assignment statements are
\nplaced in the environment for a command, not just
\nthose that precede the command name.
\n-m\u00a0\u00a0\u00a0\u00a0\u00a0 Monitor mode.\u00a0 Job control is enabled.\u00a0 This option is
\non by default for interactive shells on systems that
\nsupport it (see JOB CONTROL above).\u00a0 All processes run
\nin a separate process group.\u00a0 When a background job
\ncompletes, the shell prints a line containing its exit
\nstatus.
\n-n\u00a0\u00a0\u00a0\u00a0\u00a0 Read commands but do not execute them.\u00a0 This may be
\nused to check a shell script for syntax errors.\u00a0 This
\nis ignored by interactive shells.
\n-o option-name
\nThe option-name can be one of the following:
\nallexport
\nSame as -a.
\nbraceexpand
\nSame as -B.
\nemacs\u00a0\u00a0 Use an emacs-style command line editing
\ninterface.\u00a0 This is enabled by default when
\nthe shell is interactive, unless the shell is
\nstarted with the –noediting option.\u00a0 This
\nalso affects the editing interface used for
\nread -e.
\nerrexit Same as -e.
\nerrtrace
\nSame as -E.
\nfunctrace
\nSame as -T.
\nhashall Same as -h.
\nhistexpand
\nSame as -H.
\nhistory Enable command history, as described above
\nunder HISTORY.\u00a0 This option is on by default
\nin interactive shells.
\nignoreeof
\nThe effect is as if the shell command
\n“IGNOREEOF=10” had been executed (see Shell
\nVariables above).
\nkeyword Same as -k.
\nmonitor Same as -m.
\nnoclobber
\nSame as -C.
\nnoexec\u00a0 Same as -n.
\nnoglob\u00a0 Same as -f.
\nnolog\u00a0\u00a0 Currently ignored.
\nnotify\u00a0 Same as -b.
\nnounset Same as -u.
\nonecmd\u00a0 Same as -t.
\nphysical
\nSame as -P.
\npipefail
\nIf set, the return value of a pipeline is the
\nvalue of the last (rightmost) command to exit
\nwith a non-zero status, or zero if all
\ncommands in the pipeline exit successfully.
\nThis option is disabled by default.
\nposix\u00a0\u00a0 Change the behavior of bash where the default
\noperation differs from the POSIX standard to
\nmatch the standard (posix mode).\u00a0 See SEE ALSO
\nbelow for a reference to a document that
\ndetails how posix mode affects bash’s
\nbehavior.
\nprivileged
\nSame as -p.
\nverbose Same as -v.
\nvi\u00a0\u00a0\u00a0\u00a0\u00a0 Use a vi-style command line editing interface.
\nThis also affects the editing interface used
\nfor read -e.
\nxtrace\u00a0 Same as -x.
\nIf -o is supplied with no option-name, the values of
\nthe current options are printed.\u00a0 If +o is supplied
\nwith no option-name, a series of set commands to
\nrecreate the current option settings is displayed on
\nthe standard output.
\n-p\u00a0\u00a0\u00a0\u00a0\u00a0 Turn on privileged mode.\u00a0 In this mode, the $ENV and
\n$BASH_ENV files are not processed, shell functions are
\nnot inherited from the environment, and the SHELLOPTS,
\nBASHOPTS, CDPATH, and GLOBIGNORE variables, if they
\nappear in the environment, are ignored.\u00a0 If the shell
\nis started with the effective user (group) id not
\nequal to the real user (group) id, and the -p option
\nis not supplied, these actions are taken and the
\neffective user id is set to the real user id.\u00a0 If the
\n-p option is supplied at startup, the effective user
\nid is not reset.\u00a0 Turning this option off causes the
\neffective user and group ids to be set to the real
\nuser and group ids.
\n-t\u00a0\u00a0\u00a0\u00a0\u00a0 Exit after reading and executing one command.
\n-u\u00a0\u00a0\u00a0\u00a0\u00a0 Treat unset variables and parameters other than the
\nspecial parameters “@” and “*” as an error when
\nperforming parameter expansion.\u00a0 If expansion is
\nattempted on an unset variable or parameter, the shell
\nprints an error message, and, if not interactive,
\nexits with a non-zero status.
\n-v\u00a0\u00a0\u00a0\u00a0\u00a0 Print shell input lines as they are read.
\n-x\u00a0\u00a0\u00a0\u00a0\u00a0 After expanding each simple command, for command, case
\ncommand, select command, or arithmetic for command,
\ndisplay the expanded value of PS4, followed by the
\ncommand and its expanded arguments or associated word
\nlist.
\n-B\u00a0\u00a0\u00a0\u00a0\u00a0 The shell performs brace expansion (see Brace
\nExpansion above).\u00a0 This is on by default.
\n-C\u00a0\u00a0\u00a0\u00a0\u00a0 If set, bash does not overwrite an existing file with
\nthe >, >&, and <> redirection operators.\u00a0 This may be
\noverridden when creating output files by using the
\nredirection operator >| instead of >.
\n-E\u00a0\u00a0\u00a0\u00a0\u00a0 If set, any trap on ERR is inherited by shell
\nfunctions, command substitutions, and commands
\nexecuted in a subshell environment.\u00a0 The ERR trap is
\nnormally not inherited in such cases.
\n-H\u00a0\u00a0\u00a0\u00a0\u00a0 Enable !\u00a0 style history substitution.\u00a0 This option is
\non by default when the shell is interactive.
\n-P\u00a0\u00a0\u00a0\u00a0\u00a0 If set, the shell does not resolve symbolic links when
\nexecuting commands such as cd that change the current
\nworking directory.\u00a0 It uses the physical directory
\nstructure instead.\u00a0 By default, bash follows the
\nlogical chain of directories when performing commands
\nwhich change the current directory.
\n-T\u00a0\u00a0\u00a0\u00a0\u00a0 If set, any traps on DEBUG and RETURN are inherited by
\nshell functions, command substitutions, and commands
\nexecuted in a subshell environment.\u00a0 The DEBUG and
\nRETURN traps are normally not inherited in such cases.
\n—\u00a0\u00a0\u00a0\u00a0\u00a0 If no arguments follow this option, then the
\npositional parameters are unset.\u00a0 Otherwise, the
\npositional parameters are set to the args, even if
\nsome of them begin with a -.
\n–\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Signal the end of options, cause all remaining args to
\nbe assigned to the positional parameters.\u00a0 The -x and
\n-v options are turned off.\u00a0 If there are no args, the
\npositional parameters remain unchanged.<\/p>\n
The options are off by default unless otherwise noted.\u00a0 Using
\n+ rather than – causes these options to be turned off.\u00a0 The
\noptions can also be specified as arguments to an invocation of
\nthe shell.\u00a0 The current set of options may be found in $-.
\nThe return status is always true unless an invalid option is
\nencountered.<\/p>\n
shift [n]
\nThe positional parameters from n+1 … are renamed to $1 ….
\nParameters represented by the numbers $# down to $#-n+1 are
\nunset.\u00a0 n must be a non-negative number less than or equal to
\n$#.\u00a0 If n is 0, no parameters are changed.\u00a0 If n is not given,
\nit is assumed to be 1.\u00a0 If n is greater than $#, the
\npositional parameters are not changed.\u00a0 The return status is
\ngreater than zero if n is greater than $# or less than zero;
\notherwise 0.<\/p>\n
shopt [-pqsu] [-o] [optname …]
\nToggle the values of settings controlling optional shell
\nbehavior.\u00a0 The settings can be either those listed below, or,
\nif the -o option is used, those available with the -o option
\nto the set builtin command.\u00a0 With no options, or with the -p
\noption, a list of all settable options is displayed, with an
\nindication of whether or not each is set.\u00a0 The -p option
\ncauses output to be displayed in a form that may be reused as
\ninput.\u00a0 Other options have the following meanings:
\n-s\u00a0\u00a0\u00a0\u00a0 Enable (set) each optname.
\n-u\u00a0\u00a0\u00a0\u00a0 Disable (unset) each optname.
\n-q\u00a0\u00a0\u00a0\u00a0 Suppresses normal output (quiet mode); the return
\nstatus indicates whether the optname is set or unset.
\nIf multiple optname arguments are given with -q, the
\nreturn status is zero if all optnames are enabled; non-
\nzero otherwise.
\n-o\u00a0\u00a0\u00a0\u00a0 Restricts the values of optname to be those defined for
\nthe -o option to the set builtin.<\/p>\n
If either -s or -u is used with no optname arguments, shopt
\nshows only those options which are set or unset, respectively.
\nUnless otherwise noted, the shopt options are disabled (unset)
\nby default.<\/p>\n
The return status when listing options is zero if all optnames
\nare enabled, non-zero otherwise.\u00a0 When setting or unsetting
\noptions, the return status is zero unless an optname is not a
\nvalid shell option.<\/p>\n
The list of shopt options is:<\/p>\n
autocd\u00a0 If set, a command name that is the name of a directory
\nis executed as if it were the argument to the cd
\ncommand.\u00a0 This option is only used by interactive
\nshells.
\ncdable_vars
\nIf set, an argument to the cd builtin command that is
\nnot a directory is assumed to be the name of a
\nvariable whose value is the directory to change to.
\ncdspell If set, minor errors in the spelling of a directory
\ncomponent in a cd command will be corrected.\u00a0 The
\nerrors checked for are transposed characters, a
\nmissing character, and one character too many.\u00a0 If a
\ncorrection is found, the corrected filename is
\nprinted, and the command proceeds.\u00a0 This option is
\nonly used by interactive shells.
\ncheckhash
\nIf set, bash checks that a command found in the hash
\ntable exists before trying to execute it.\u00a0 If a hashed
\ncommand no longer exists, a normal path search is
\nperformed.
\ncheckjobs
\nIf set, bash lists the status of any stopped and
\nrunning jobs before exiting an interactive shell.\u00a0 If
\nany jobs are running, this causes the exit to be
\ndeferred until a second exit is attempted without an
\nintervening command (see JOB CONTROL above).\u00a0 The
\nshell always postpones exiting if any jobs are
\nstopped.
\ncheckwinsize
\nIf set, bash checks the window size after each command
\nand, if necessary, updates the values of LINES and
\nCOLUMNS.
\ncmdhist If set, bash attempts to save all lines of a multiple-
\nline command in the same history entry.\u00a0 This allows
\neasy re-editing of multi-line commands.
\ncompat31
\nIf set, bash changes its behavior to that of version
\n3.1 with respect to quoted arguments to the [[
\nconditional command’s =~ operator and locale-specific
\nstring comparison when using the [[ conditional
\ncommand’s < and > operators.\u00a0 Bash versions prior to
\nbash-4.1 use ASCII collation and strcmp(3); bash-4.1
\nand later use the current locale’s collation sequence
\nand strcoll(3).
\ncompat32
\nIf set, bash changes its behavior to that of version
\n3.2 with respect to locale-specific string comparison
\nwhen using the [[ conditional command’s < and >
\noperators (see previous item).
\ncompat40
\nIf set, bash changes its behavior to that of version
\n4.0 with respect to locale-specific string comparison
\nwhen using the [[ conditional command’s < and >
\noperators (see description of compat31) and the effect
\nof interrupting a command list.\u00a0 Bash versions 4.0 and
\nlater interrupt the list as if the shell received the
\ninterrupt; previous versions continue with the next
\ncommand in the list.
\ncompat41
\nIf set, bash, when in posix mode, treats a single
\nquote in a double-quoted parameter expansion as a
\nspecial character.\u00a0 The single quotes must match (an
\neven number) and the characters between the single
\nquotes are considered quoted.\u00a0 This is the behavior of
\nposix mode through version 4.1.\u00a0 The default bash
\nbehavior remains as in previous versions.
\ncompat42
\nIf set, bash does not process the replacement string
\nin the pattern substitution word expansion using quote
\nremoval.
\ncomplete_fullquote
\nIf set, bash quotes all shell metacharacters in
\nfilenames and directory names when performing
\ncompletion.\u00a0 If not set, bash removes metacharacters
\nsuch as the dollar sign from the set of characters
\nthat will be quoted in completed filenames when these
\nmetacharacters appear in shell variable references in
\nwords to be completed.\u00a0 This means that dollar signs
\nin variable names that expand to directories will not
\nbe quoted; however, any dollar signs appearing in
\nfilenames will not be quoted, either.\u00a0 This is active
\nonly when bash is using backslashes to quote completed
\nfilenames.\u00a0 This variable is set by default, which is
\nthe default bash behavior in versions through 4.2.
\ndirexpand
\nIf set, bash replaces directory names with the results
\nof word expansion when performing filename completion.
\nThis changes the contents of the readline editing
\nbuffer.\u00a0 If not set, bash attempts to preserve what
\nthe user typed.
\ndirspell
\nIf set, bash attempts spelling correction on directory
\nnames during word completion if the directory name
\ninitially supplied does not exist.
\ndotglob If set, bash includes filenames beginning with a `.’
\nin the results of pathname expansion.
\nexecfail
\nIf set, a non-interactive shell will not exit if it
\ncannot execute the file specified as an argument to
\nthe exec builtin command.\u00a0 An interactive shell does
\nnot exit if exec fails.
\nexpand_aliases
\nIf set, aliases are expanded as described above under
\nALIASES.\u00a0 This option is enabled by default for
\ninteractive shells.
\nextdebug
\nIf set, behavior intended for use by debuggers is
\nenabled:
\n1.\u00a0\u00a0\u00a0\u00a0 The -F option to the declare builtin displays
\nthe source file name and line number
\ncorresponding to each function name supplied as
\nan argument.
\n2.\u00a0\u00a0\u00a0\u00a0 If the command run by the DEBUG trap returns a
\nnon-zero value, the next command is skipped and
\nnot executed.
\n3.\u00a0\u00a0\u00a0\u00a0 If the command run by the DEBUG trap returns a
\nvalue of 2, and the shell is executing in a
\nsubroutine (a shell function or a shell script
\nexecuted by the . or source builtins), a call
\nto return is simulated.
\n4.\u00a0\u00a0\u00a0\u00a0 BASH_ARGC and BASH_ARGV are updated as
\ndescribed in their descriptions above.
\n5.\u00a0\u00a0\u00a0\u00a0 Function tracing is enabled:\u00a0 command
\nsubstitution, shell functions, and subshells
\ninvoked with ( command ) inherit the DEBUG and
\nRETURN traps.
\n6.\u00a0\u00a0\u00a0\u00a0 Error tracing is enabled:\u00a0 command
\nsubstitution, shell functions, and subshells
\ninvoked with ( command ) inherit the ERR trap.
\nextglob If set, the extended pattern matching features
\ndescribed above under Pathname Expansion are enabled.
\nextquote
\nIf set, $’string’ and $”string” quoting is performed
\nwithin ${parameter} expansions enclosed in double
\nquotes.\u00a0 This option is enabled by default.
\nfailglob
\nIf set, patterns which fail to match filenames during
\npathname expansion result in an expansion error.
\nforce_fignore
\nIf set, the suffixes specified by the FIGNORE shell
\nvariable cause words to be ignored when performing
\nword completion even if the ignored words are the only
\npossible completions.\u00a0 See SHELL VARIABLES above for a
\ndescription of FIGNORE.\u00a0 This option is enabled by
\ndefault.
\nglobasciiranges
\nIf set, range expressions used in pattern matching
\nbracket expressions (see Pattern Matching above)
\nbehave as if in the traditional C locale when
\nperforming comparisons.\u00a0 That is, the current locale’s
\ncollating sequence is not taken into account, so b
\nwill not collate between A and B, and upper-case and
\nlower-case ASCII characters will collate together.
\nglobstar
\nIf set, the pattern ** used in a pathname expansion
\ncontext will match all files and zero or more
\ndirectories and subdirectories.\u00a0 If the pattern is
\nfollowed by a \/, only directories and subdirectories
\nmatch.
\ngnu_errfmt
\nIf set, shell error messages are written in the
\nstandard GNU error message format.
\nhistappend
\nIf set, the history list is appended to the file named
\nby the value of the HISTFILE variable when the shell
\nexits, rather than overwriting the file.
\nhistreedit
\nIf set, and readline is being used, a user is given
\nthe opportunity to re-edit a failed history
\nsubstitution.
\nhistverify
\nIf set, and readline is being used, the results of
\nhistory substitution are not immediately passed to the
\nshell parser.\u00a0 Instead, the resulting line is loaded
\ninto the readline editing buffer, allowing further
\nmodification.
\nhostcomplete
\nIf set, and readline is being used, bash will attempt
\nto perform hostname completion when a word containing
\na @ is being completed (see Completing under READLINE
\nabove).\u00a0 This is enabled by default.
\nhuponexit
\nIf set, bash will send SIGHUP to all jobs when an
\ninteractive login shell exits.
\ninteractive_comments
\nIf set, allow a word beginning with # to cause that
\nword and all remaining characters on that line to be
\nignored in an interactive shell (see COMMENTS above).
\nThis option is enabled by default.
\nlastpipe
\nIf set, and job control is not active, the shell runs
\nthe last command of a pipeline not executed in the
\nbackground in the current shell environment.
\nlithist If set, and the cmdhist option is enabled, multi-line
\ncommands are saved to the history with embedded
\nnewlines rather than using semicolon separators where
\npossible.
\nlogin_shell
\nThe shell sets this option if it is started as a login
\nshell (see INVOCATION above).\u00a0 The value may not be
\nchanged.
\nmailwarn
\nIf set, and a file that bash is checking for mail has
\nbeen accessed since the last time it was checked, the
\nmessage “The mail in mailfile has been read” is
\ndisplayed.
\nno_empty_cmd_completion
\nIf set, and readline is being used, bash will not
\nattempt to search the PATH for possible completions
\nwhen completion is attempted on an empty line.
\nnocaseglob
\nIf set, bash matches filenames in a case-insensitive
\nfashion when performing pathname expansion (see
\nPathname Expansion above).
\nnocasematch
\nIf set, bash matches patterns in a case-insensitive
\nfashion when performing matching while executing case
\nor [[ conditional commands.
\nnullglob
\nIf set, bash allows patterns which match no files (see
\nPathname Expansion above) to expand to a null string,
\nrather than themselves.
\nprogcomp
\nIf set, the programmable completion facilities (see
\nProgrammable Completion above) are enabled.\u00a0 This
\noption is enabled by default.
\npromptvars
\nIf set, prompt strings undergo parameter expansion,
\ncommand substitution, arithmetic expansion, and quote
\nremoval after being expanded as described in PROMPTING
\nabove.\u00a0 This option is enabled by default.
\nrestricted_shell
\nThe shell sets this option if it is started in
\nrestricted mode (see RESTRICTED SHELL below).\u00a0 The
\nvalue may not be changed.\u00a0 This is not reset when the
\nstartup files are executed, allowing the startup files
\nto discover whether or not a shell is restricted.
\nshift_verbose
\nIf set, the shift builtin prints an error message when
\nthe shift count exceeds the number of positional
\nparameters.
\nsourcepath
\nIf set, the source (.) builtin uses the value of PATH
\nto find the directory containing the file supplied as
\nan argument.\u00a0 This option is enabled by default.
\nxpg_echo
\nIf set, the echo builtin expands backslash-escape
\nsequences by default.<\/p>\n
suspend [-f]
\nSuspend the execution of this shell until it receives a
\nSIGCONT signal.\u00a0 A login shell cannot be suspended; the -f
\noption can be used to override this and force the suspension.
\nThe return status is 0 unless the shell is a login shell and
\n-f is not supplied, or if job control is not enabled.<\/p>\n
test expr
\n[ expr ]
\nReturn a status of 0 (true) or 1 (false) depending on the
\nevaluation of the conditional expression expr.\u00a0 Each operator
\nand operand must be a separate argument.\u00a0 Expressions are
\ncomposed of the primaries described above under CONDITIONAL
\nEXPRESSIONS.\u00a0 test does not accept any options, nor does it
\naccept and ignore an argument of — as signifying the end of
\noptions.<\/p>\n
Expressions may be combined using the following operators,
\nlisted in decreasing order of precedence.\u00a0 The evaluation
\ndepends on the number of arguments; see below.\u00a0 Operator
\nprecedence is used when there are five or more arguments.
\n! expr True if expr is false.
\n( expr )
\nReturns the value of expr.\u00a0 This may be used to
\noverride the normal precedence of operators.
\nexpr1 -a expr2
\nTrue if both expr1 and expr2 are true.
\nexpr1 -o expr2
\nTrue if either expr1 or expr2 is true.<\/p>\n
test and [ evaluate conditional expressions using a set of
\nrules based on the number of arguments.<\/p>\n
0 arguments
\nThe expression is false.
\n1 argument
\nThe expression is true if and only if the argument is
\nnot null.
\n2 arguments
\nIf the first argument is !, the expression is true if
\nand only if the second argument is null.\u00a0 If the first
\nargument is one of the unary conditional operators
\nlisted above under CONDITIONAL EXPRESSIONS, the
\nexpression is true if the unary test is true.\u00a0 If the
\nfirst argument is not a valid unary conditional
\noperator, the expression is false.
\n3 arguments
\nThe following conditions are applied in the order
\nlisted.\u00a0 If the second argument is one of the binary
\nconditional operators listed above under CONDITIONAL
\nEXPRESSIONS, the result of the expression is the result
\nof the binary test using the first and third arguments
\nas operands.\u00a0 The -a and -o operators are considered
\nbinary operators when there are three arguments.\u00a0 If
\nthe first argument is !, the value is the negation of
\nthe two-argument test using the second and third
\narguments.\u00a0 If the first argument is exactly ( and the
\nthird argument is exactly ), the result is the one-
\nargument test of the second argument.\u00a0 Otherwise, the
\nexpression is false.
\n4 arguments
\nIf the first argument is !, the result is the negation
\nof the three-argument expression composed of the
\nremaining arguments.\u00a0 Otherwise, the expression is
\nparsed and evaluated according to precedence using the
\nrules listed above.
\n5 or more arguments
\nThe expression is parsed and evaluated according to
\nprecedence using the rules listed above.<\/p>\n
When used with test or [, the < and > operators sort
\nlexicographically using ASCII ordering.<\/p>\n
times\u00a0 Print the accumulated user and system times for the shell and
\nfor processes run from the shell.\u00a0 The return status is 0.<\/p>\n
trap [-lp] [[arg] sigspec …]
\nThe command arg is to be read and executed when the shell
\nreceives signal(s) sigspec.\u00a0 If arg is absent (and there is a
\nsingle sigspec) or -, each specified signal is reset to its
\noriginal disposition (the value it had upon entrance to the
\nshell).\u00a0 If arg is the null string the signal specified by
\neach sigspec is ignored by the shell and by the commands it
\ninvokes.\u00a0 If arg is not present and -p has been supplied, then
\nthe trap commands associated with each sigspec are displayed.
\nIf no arguments are supplied or if only -p is given, trap
\nprints the list of commands associated with each signal.\u00a0 The
\n-l option causes the shell to print a list of signal names and
\ntheir corresponding numbers.\u00a0 Each sigspec is either a signal
\nname defined in <signal.h>, or a signal number.\u00a0 Signal names
\nare case insensitive and the SIG prefix is optional.<\/p>\n
If a sigspec is EXIT (0) the command arg is executed on exit
\nfrom the shell.\u00a0 If a sigspec is DEBUG, the command arg is
\nexecuted before every simple command, for command, case
\ncommand, select command, every arithmetic for command, and
\nbefore the first command executes in a shell function (see
\nSHELL GRAMMAR above).\u00a0 Refer to the description of the
\nextdebug option to the shopt builtin for details of its effect
\non the DEBUG trap.\u00a0 If a sigspec is RETURN, the command arg is
\nexecuted each time a shell function or a script executed with
\nthe . or source builtins finishes executing.<\/p>\n
If a sigspec is ERR, the command arg is executed whenever a a
\npipeline (which may consist of a single simple command), a
\nlist, or a compound command returns a non-zero exit status,
\nsubject to the following conditions.\u00a0 The ERR trap is not
\nexecuted if the failed command is part of the command list
\nimmediately following a while or until keyword, part of the
\ntest in an if statement, part of a command executed in a && or
\n|| list except the command following the final && or ||, any
\ncommand in a pipeline but the last, or if the command’s return
\nvalue is being inverted using !.\u00a0 These are the same
\nconditions obeyed by the errexit (-e) option.<\/p>\n
Signals ignored upon entry to the shell cannot be trapped or
\nreset.\u00a0 Trapped signals that are not being ignored are reset
\nto their original values in a subshell or subshell environment
\nwhen one is created.\u00a0 The return status is false if any
\nsigspec is invalid; otherwise trap returns true.<\/p>\n
type [-aftpP] name [name …]
\nWith no options, indicate how each name would be interpreted
\nif used as a command name.\u00a0 If the -t option is used, type
\nprints a string which is one of alias, keyword, function,
\nbuiltin, or file if name is an alias, shell reserved word,
\nfunction, builtin, or disk file, respectively.\u00a0 If the name is
\nnot found, then nothing is printed, and an exit status of
\nfalse is returned.\u00a0 If the -p option is used, type either
\nreturns the name of the disk file that would be executed if
\nname were specified as a command name, or nothing if “type -t
\nname” would not return file.\u00a0 The -P option forces a PATH
\nsearch for each name, even if “type -t name” would not
\nreturn file.\u00a0 If a command is hashed, -p and -P print the
\nhashed value, which is not necessarily the file that appears
\nfirst in PATH.\u00a0 If the -a option is used, type prints all of
\nthe places that contain an executable named name.\u00a0 This
\nincludes aliases and functions, if and only if the -p option
\nis not also used.\u00a0 The table of hashed commands is not
\nconsulted when using -a.\u00a0 The -f option suppresses shell
\nfunction lookup, as with the command builtin.\u00a0 type returns
\ntrue if all of the arguments are found, false if any are not
\nfound.<\/p>\n
ulimit [-HSTabcdefilmnpqrstuvx [limit]]
\nProvides control over the resources available to the shell and
\nto processes started by it, on systems that allow such
\ncontrol.\u00a0 The -H and -S options specify that the hard or soft
\nlimit is set for the given resource.\u00a0 A hard limit cannot be
\nincreased by a non-root user once it is set; a soft limit may
\nbe increased up to the value of the hard limit.\u00a0 If neither -H
\nnor -S is specified, both the soft and hard limits are set.
\nThe value of limit can be a number in the unit specified for
\nthe resource or one of the special values hard, soft, or
\nunlimited, which stand for the current hard limit, the current
\nsoft limit, and no limit, respectively.\u00a0 If limit is omitted,
\nthe current value of the soft limit of the resource is
\nprinted, unless the -H option is given.\u00a0 When more than one
\nresource is specified, the limit name and unit are printed
\nbefore the value.\u00a0 Other options are interpreted as follows:
\n-a\u00a0\u00a0\u00a0\u00a0 All current limits are reported
\n-b\u00a0\u00a0\u00a0\u00a0 The maximum socket buffer size
\n-c\u00a0\u00a0\u00a0\u00a0 The maximum size of core files created
\n-d\u00a0\u00a0\u00a0\u00a0 The maximum size of a process’s data segment
\n-e\u00a0\u00a0\u00a0\u00a0 The maximum scheduling priority (“nice”)
\n-f\u00a0\u00a0\u00a0\u00a0 The maximum size of files written by the shell and its
\nchildren
\n-i\u00a0\u00a0\u00a0\u00a0 The maximum number of pending signals
\n-l\u00a0\u00a0\u00a0\u00a0 The maximum size that may be locked into memory
\n-m\u00a0\u00a0\u00a0\u00a0 The maximum resident set size (many systems do not
\nhonor this limit)
\n-n\u00a0\u00a0\u00a0\u00a0 The maximum number of open file descriptors (most
\nsystems do not allow this value to be set)
\n-p\u00a0\u00a0\u00a0\u00a0 The pipe size in 512-byte blocks (this may not be set)
\n-q\u00a0\u00a0\u00a0\u00a0 The maximum number of bytes in POSIX message queues
\n-r\u00a0\u00a0\u00a0\u00a0 The maximum real-time scheduling priority
\n-s\u00a0\u00a0\u00a0\u00a0 The maximum stack size
\n-t\u00a0\u00a0\u00a0\u00a0 The maximum amount of cpu time in seconds
\n-u\u00a0\u00a0\u00a0\u00a0 The maximum number of processes available to a single
\nuser
\n-v\u00a0\u00a0\u00a0\u00a0 The maximum amount of virtual memory available to the
\nshell and, on some systems, to its children
\n-x\u00a0\u00a0\u00a0\u00a0 The maximum number of file locks
\n-T\u00a0\u00a0\u00a0\u00a0 The maximum number of threads<\/p>\n
If limit is given, and the -a option is not used, limit is the
\nnew value of the specified resource.\u00a0 If no option is given,
\nthen -f is assumed.\u00a0 Values are in 1024-byte increments,
\nexcept for -t, which is in seconds; -p, which is in units of
\n512-byte blocks; and -T, -b, -n, and -u, which are unscaled
\nvalues.\u00a0 The return status is 0 unless an invalid option or
\nargument is supplied, or an error occurs while setting a new
\nlimit.<\/p>\n
umask [-p] [-S] [mode]
\nThe user file-creation mask is set to mode.\u00a0 If mode begins
\nwith a digit, it is interpreted as an octal number; otherwise
\nit is interpreted as a symbolic mode mask similar to that
\naccepted by chmod(1).\u00a0 If mode is omitted, the current value
\nof the mask is printed.\u00a0 The -S option causes the mask to be
\nprinted in symbolic form; the default output is an octal
\nnumber.\u00a0 If the -p option is supplied, and mode is omitted,
\nthe output is in a form that may be reused as input.\u00a0 The
\nreturn status is 0 if the mode was successfully changed or if
\nno mode argument was supplied, and false otherwise.<\/p>\n
unalias [-a] [name …]
\nRemove each name from the list of defined aliases.\u00a0 If -a is
\nsupplied, all alias definitions are removed.\u00a0 The return value
\nis true unless a supplied name is not a defined alias.<\/p>\n
unset [-fv] [-n] [name …]
\nFor each name, remove the corresponding variable or function.
\nIf the -v option is given, each name refers to a shell
\nvariable, and that variable is removed.\u00a0 Read-only variables
\nmay not be unset.\u00a0 If -f is specified, each name refers to a
\nshell function, and the function definition is removed.\u00a0 If
\nthe -n option is supplied, and name is a variable with the
\nnameref attribute, name will be unset rather than the variable
\nit references.\u00a0 -n has no effect if the -f option is supplied.
\nIf no options are supplied, each name refers to a variable; if
\nthere is no variable by that name, any function with that name
\nis unset.\u00a0 Each unset variable or function is removed from the
\nenvironment passed to subsequent commands.\u00a0 If any of
\nCOMP_WORDBREAKS, RANDOM, SECONDS, LINENO, HISTCMD, FUNCNAME,
\nGROUPS, or DIRSTACK are unset, they lose their special
\nproperties, even if they are subsequently reset.\u00a0 The exit
\nstatus is true unless a name is readonly.<\/p>\n
wait [-n] [n …]
\nWait for each specified child process and return its
\ntermination status.\u00a0 Each n may be a process ID or a job
\nspecification; if a job spec is given, all processes in that
\njob’s pipeline are waited for.\u00a0 If n is not given, all
\ncurrently active child processes are waited for, and the
\nreturn status is zero.\u00a0 If the -n option is supplied, wait
\nwaits for any job to terminate and returns its exit status.
\nIf n specifies a non-existent process or job, the return
\nstatus is 127.\u00a0 Otherwise, the return status is the exit
\nstatus of the last process or job waited for.<\/p>\n
RESTRICTED SHELL<\/p>\n
If bash is started with the name rbash, or the -r option is supplied
\nat invocation, the shell becomes restricted.\u00a0 A restricted shell is
\nused to set up an environment more controlled than the standard
\nshell.\u00a0 It behaves identically to bash with the exception that the
\nfollowing are disallowed or not performed:<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 changing directories with cd<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 setting or unsetting the values of SHELL, PATH, ENV, or
\nBASH_ENV<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 specifying command names containing \/<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 specifying a filename containing a \/ as an argument to the .
\nbuiltin command<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 specifying a filename containing a slash as an argument to the
\n-p option to the hash builtin command<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 importing function definitions from the shell environment at
\nstartup<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 parsing the value of SHELLOPTS from the shell environment at
\nstartup<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 redirecting output using the >, >|, <>, >&, &>, and >>
\nredirection operators<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 using the exec builtin command to replace the shell with
\nanother command<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 adding or deleting builtin commands with the -f and -d options
\nto the enable builtin command<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 using the enable builtin command to enable disabled shell
\nbuiltins<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 specifying the -p option to the command builtin command<\/p>\n
\u00b7\u00a0\u00a0\u00a0\u00a0\u00a0 turning off restricted mode with set +r or set +o restricted.<\/p>\n
These restrictions are enforced after any startup files are read.<\/p>\n
When a command that is found to be a shell script is executed (see
\nCOMMAND EXECUTION above), rbash turns off any restrictions in the
\nshell spawned to execute the script.<\/p>\n
SEE ALSO<\/p>\n
Bash Reference Manual, Brian Fox and Chet Ramey
\nThe Gnu Readline Library, Brian Fox and Chet Ramey
\nThe Gnu History Library, Brian Fox and Chet Ramey
\nPortable Operating System Interface (POSIX) Part 2: Shell and
\nUtilities, IEEE —
\nhttp:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/
\nhttp:\/\/tiswww.case.edu\/~chet\/bash\/POSIX — a description of posix
\nmode
\nsh(1), ksh(1), csh(1)
\nemacs(1), vi(1)
\nreadline(3)<\/p>\n
FILES<\/p>\n
\/bin\/bash
\nThe bash executable
\n\/etc\/profile
\nThe systemwide initialization file, executed for login shells
\n~\/.bash_profile
\nThe personal initialization file, executed for login shells
\n~\/.bashrc
\nThe individual per-interactive-shell startup file
\n~\/.bash_logout
\nThe individual login shell cleanup file, executed when a login
\nshell exits
\n~\/.inputrc
\nIndividual readline initialization file<\/p>\n
AUTHORS<\/p>\n
Brian Fox, Free Software Foundation
\nbfox@gnu.org<\/p>\n
Chet Ramey, Case Western Reserve University
\nchet.ramey@case.edu<\/p>\n
BUG REPORTS<\/p>\n
If you find a bug in bash, you should report it.\u00a0 But first, you
\nshould make sure that it really is a bug, and that it appears in the
\nlatest version of bash.\u00a0 The latest version is always available from
\nftp:\/\/ftp.gnu.org\/pub\/gnu\/bash\/.<\/p>\n
Once you have determined that a bug actually exists, use the bashbug
\ncommand to submit a bug report.\u00a0 If you have a fix, you are
\nencouraged to mail that as well!\u00a0 Suggestions and `philosophical’ bug
\nreports may be mailed to\u00a0bug-bash@gnu.org\u00a0or posted to the Usenet
\nnewsgroup gnu.bash.bug.<\/p>\n
ALL bug reports should include:<\/p>\n
The version number of bash
\nThe hardware and operating system
\nThe compiler used to compile
\nA description of the bug behaviour
\nA short script or `recipe’ which exercises the bug<\/p>\n
bashbug inserts the first three items automatically into the template
\nit provides for filing a bug report.<\/p>\n
Comments and bug reports concerning this manual page should be
\ndirected to\u00a0chet.ramey@case.edu.<\/p>\n
BUGS<\/p>\n
It’s too big and too slow.<\/p>\n
There are some subtle differences between bash and traditional
\nversions of sh, mostly because of the POSIX specification.<\/p>\n
Aliases are confusing in some uses.<\/p>\n
Shell builtin commands and functions are not stoppable\/restartable.<\/p>\n
Compound commands and command sequences of the form `a ; b ; c’ are
\nnot handled gracefully when process suspension is attempted.\u00a0 When a
\nprocess is stopped, the shell immediately executes the next command
\nin the sequence.\u00a0 It suffices to place the sequence of commands
\nbetween parentheses to force it into a subshell, which may be stopped
\nas a unit.<\/p>\n
Array variables may not (yet) be exported.<\/p>\n
There may be only one active coprocess at a time.<\/p>\n
COLOPHON<\/p>\n
This page is part of the bash (Bourne again shell) project.
\nInformation about the project can be found at
\n\u27e8http:\/\/www.gnu.org\/software\/bash\/\u27e9.\u00a0 If you have a bug report for
\nthis manual page, see \u27e8http:\/\/www.gnu.org\/software\/bash\/\u27e9.\u00a0 This page
\nwas obtained from the project’s upstream Git repository
\n(git:\/\/git.savannah.gnu.org\/bash.git) on 2014-12-30.\u00a0 If you discover
\nany rendering problems in this HTML version of the page, or you
\nbelieve there is a better or more up-to-date source for the page, or
\nyou have corrections or improvements to the information in this
\nCOLOPHON (which is not part of the original manual page), send a mail
\nto\u00a0man-pages@man7.org<\/p>\n
GNU Bash 4.3\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 2014 February 2\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 BASH(1)<\/p>\n","protected":false},"excerpt":{"rendered":"
BASH(1)\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 General Commands Manual\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 BASH(1) NAME bash – GNU Bourne-Again SHell SYNOPSIS bash [options] [command_string | file] COPYRIGHT Bash is Copyright (C) 1989-2013 by the Free Software Foundation, Inc. DESCRIPTION Bash is an sh-compatible command language interpreter that executes commands read from the standard input or from a file.\u00a0 Bash also incorporates useful features from […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_sitemap_exclude":false,"_sitemap_priority":"","_sitemap_frequency":"","footnotes":""},"categories":[8],"tags":[],"class_list":["post-280","post","type-post","status-publish","format-standard","hentry","category-shell"],"a3_pvc":{"activated":false,"total_views":0,"today_views":0},"_links":{"self":[{"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=\/wp\/v2\/posts\/280","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=280"}],"version-history":[{"count":0,"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=\/wp\/v2\/posts\/280\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=280"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=280"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.linuxboxen.dk\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=280"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}