Differences between revisions 43 and 47 (spanning 4 versions)
Revision 43 as of 2013-02-24 03:05:45
Size: 9385
Editor: 201-51-47-84
Comment: Co-processes can be used to simulate some functions of Process Substitution. by (@antonio_cfc)
Revision 47 as of 2013-03-18 18:15:05
Size: 10450
Editor: ormaaj
Comment: Add #, ##, %, %% PE
Deletions are marked like this. Additions are marked like this.
Line 2: Line 2:
This page is an attempt to list some of the most common bashisms, i.e. features not defined by POSIX (won't work in dash, or general `/bin/sh`). It probably won't be exhaustive. Note also we talk about "bashism" because this wiki is largely bash-centric but a number of these extensions work in other shells like ksh or zsh. This page is an attempt to list some of the most common bashisms, i.e. features not defined by POSIX (won't work in dash, or general `/bin/sh`). It probably won't be exhaustive. Note also we talk about "bashism" because this wiki is largely bash-centric but a number (almost all) of these extensions work in at least some other shells like ksh or zsh with perhaps some differences in the details, as most of Bash's scripting features are derived from ksh. POSIX has simply required a much smaller number of them.
Line 9: Line 9:
||expand sequences ||`echo $'hello\tworld'` ||`printf "hello\tworld\n"` ||`$' '` is not defined by POSIX and is bash-specific || ||expand sequences ||`echo $'hello\tworld'` ||`printf "hello\tworld\n"` || Historically `$' '` was not defined by POSIX. However, `$' '` has since been added to POSIX. http://austingroupbugs.net/view.php?id=249 ||
Line 15: Line 15:
 * Brace Expansion, eg `{a,b,c}` or `{1..10}` is not defined by POSIX. Both forms are present in ksh93, and the first form in older ksh.  * Brace Expansion, eg `{a,b,c}` or `{1..10}` is not defined by POSIX. Both forms are present in ksh93, and the first form in older ksh and mksh.
Line 17: Line 17:
 * The `<( )` and `>( )` process substitutions are not defined by POSIX, but can be simulated with FIFOs: instead of `foo <(bar)`, write `mkfifo fifo; bar > fifo & foo fifo` or `bar &| foo <&p` (this is basically how process substitution is implemented on OSes that don't have a mechanism like `/dev/fd/` to refer to unnamed pipes with filenames). It is present in ksh93.  * The `<( )` and `>( )` process substitutions are not defined by POSIX, but can be simulated with FIFOs: instead of `foo <(bar)`, write `mkfifo fifo; bar > fifo & foo fifo` (this is basically how process substitution is implemented on OSes that don't have a mechanism like `/dev/fd/` to refer to unnamed pipes with filenames). It is present in ksh93.
Line 24: Line 24:
 * `${name/ }` -- you can use `$(printf '%s\n' "$name" | sed 's/foo/bar/')`, after changing shell patterns to regular expressions. This form is present in ksh93.  * `${name/ }` -- you can use `$(printf '%s\n' "$name" | sed 's/foo/bar/')`, after changing shell patterns to regular expressions. This form is present in ksh93, mksh, and zsh, but ksh93's substitution expansion differs from Bash's.
Line 27: Line 27:

 * The `#`, `##`, `%`, and `%%` operators are unspecified by POSIX and the ksh88 manual when used together with the `@` or `*` parameters. Dash applies the trimming to the flattened result. mksh/pdksh treats it as a bad expansion.
Line 57: Line 59:
||`(( ))` ||`(( ))` (without the `$`) acts like a command on its own ||For simple comparison: `[ -lt ] (and -ne -gt -ge)` or `[ "$(( 3+1 < 5))" -eq 0 ]`. ||present in ksh || ||`(( ))` ||`(( ))` (without the `$`) acts like a command on its own ||For simple comparison: `[ -lt ] (and -ne -gt -ge)` or `[ $((3 + 1 < 5)) -ne 0 ]`. ||present in ksh ||
Line 71: Line 73:
||pre/pos increment/decrement ||`++ --` ||`i=$((i+1))` ||- ||
||- ||`let` ||`: $((i=i+1))` ||The `:` command can be used to peform side effects with an expansion ||
||pre/pos increment/decrement ||`++ --` ||`i=$((i+1))` or `: $((i+=1))` ||- ||
||comma operator ||`,` ||`: $((...))`; cmd $((...))` || The comma operator is ''widely'' supported by almost everything except dash -- even posh and Busybox. ||
||- ||`let` or `((...))` ||`[ $((...)) -ne 0 ]` ||Because of the above comma restriction, `let` can't be simulated exactly without a loop. ||
Line 77: Line 80:
||redirect both stdout and stderr ||`>&` and `&>` ||`command > file 2>&1 or commnd 2>&1 | othercommand` ||- ||
|| ||`|&` (bash4) ||`command 2>&1 | othercommand` ||- ||
||redirect both stdout and stderr ||`>&` and `&>` ||`command > file 2>&1 or command 2>&1 | othercommand` ||- ||
|| ||`|&` (bash4) ||`command 2>&1 | othercommand` ||Conflicts with ksh. Not recommended, even in Bash. Just use `2>&1`. ||
Line 86: Line 89:
 * `printf -v` is not defined by POSIX. Also the `%q` format is not defined by POSIX.  * `printf -v` is not defined by POSIX, and only Bash supports it. The `%q` and `%()T` formats are not defined by POSIX but supported by ksh93 and Bash. The a, A, e, E, f, F, g, and G formats are not required by POSIX for `printf(1)`, but dash appears to support `%f`, `%e`, `%E`, `%g`, and `%G`.

How to make bash scripts work in dash

This page is an attempt to list some of the most common bashisms, i.e. features not defined by POSIX (won't work in dash, or general /bin/sh). It probably won't be exhaustive. Note also we talk about "bashism" because this wiki is largely bash-centric but a number (almost all) of these extensions work in at least some other shells like ksh or zsh with perhaps some differences in the details, as most of Bash's scripting features are derived from ksh. POSIX has simply required a much smaller number of them.

Syntax

Works in bash

Change to for dash

Comment

defining functions

function f { echo hello world; }

f() { echo hello world; }

"function" is not defined by POSIX, only "name ()" is. In ksh both forms are present, but functions defined with "function" work slightly differently.

case

;;& ;& etc

None. Duplicate the case (use a function to avoid code duplication)

;;& ;& in bash4 is not defined by POSIX. ksh93 has ;& but not ;;&

numeric C-like for loop

for ((i=0; i<3; i++)); do
 echo $i
done

i=0 ; while test $i -lt 3 ; do
 echo $i ; i=$(($i+1))
done

this syntax is not defined by POSIX. Present in ksh93.

expand sequences

echo $'hello\tworld'

printf "hello\tworld\n"

Historically $' ' was not defined by POSIX. However, $' ' has since been added to POSIX. http://austingroupbugs.net/view.php?id=249

extended glob

+( ) @( ) !( ) *( )

not always possible, sometimes you can use several globs, sometimes you can use find(1)

not defined by POSIX. Present in ksh.

select

select

some ideas: implement the menu yourself, use a command like dialog

not defined by POSIX. Present in ksh.

Expansions

  • Brace Expansion, eg {a,b,c} or {1..10} is not defined by POSIX. Both forms are present in ksh93, and the first form in older ksh and mksh.

  • The <( ) and >( ) process substitutions are not defined by POSIX, but can be simulated with FIFOs: instead of foo <(bar), write mkfifo fifo; bar > fifo & foo fifo (this is basically how process substitution is implemented on OSes that don't have a mechanism like /dev/fd/ to refer to unnamed pipes with filenames). It is present in ksh93.

Parameter Expansions

List of expansions not defined by POSIX:

  • ${name:n:l} -- if the variable contains no newlines, you can use $(printf '%s\n' "$name" | cut -c $n-$((n+l-1))). If there are newlines, you can use $(printf %s "$name" | dd bs=1 skip=$n count=$l 2>/dev/null), but printf may report a (harmless) broken pipe if the variable contents are large enough. This form is present in ksh93.

  • ${name/ } -- you can use $(printf '%s\n' "$name" | sed 's/foo/bar/'), after changing shell patterns to regular expressions. This form is present in ksh93, mksh, and zsh, but ksh93's substitution expansion differs from Bash's.

  • ${!name} -- bash-specific; it is possible, but dangerous, to use eval to achieve similar effects; see BashFAQ/006.

  • The #, ##, %, and %% operators are unspecified by POSIX and the ksh88 manual when used together with the @ or * parameters. Dash applies the trimming to the flattened result. mksh/pdksh treats it as a bad expansion.

Note that using $( ) has the side-effect of removing trailing newlines from the results. Furthermore, since many standard Posix utilities, such as sed, require text files as input, you should ensure their input ends in a newline.

Arrays

Arrays are not defined by POSIX (but are present in ksh); there is no easy general workaround for arrays. Here are some hints:

  • The positional parameters are a kind of array (only one array):

#build a command dynamically see [[BashFAQ/050]]
set -- 'mycommand' 'needs some complex' 'args'
"$@"
#access the i'th param
set -- one two three
i=2
eval var=\$$i #take care if i comes from some user input see below
echo "$var"

Conditionals

Works in bash

Change to for dash

Comment

simple test

[[

use [ and use double quotes around the expansions [ "$var" = "" ]

[[ is not defined by POSIX, but is present in ksh

pattern matching

[[ foo = *glov ]]

use case or grep

see BashFAQ/041

equality with test

==

use = instead

only = is defined by POSIX, = works also in bash

compare lexicographically.

< >

no change

present in dash and ksh, but not defined by POSIX. See note below for possible workarounds.

compare modification times

[[ file1 -nt file2 ]] or -ot

[ "$(find 'file1' -prune -newer 'file2')" ] or [ "file1" -nt "file2" ]

-prune is required to avoid recursion; present in ksh. -nt and -ot aren't specified by POSIX.

check if 2 files are the same hardlink

[[ file1 -ef file2 ]]

[ "file1" -ef "file2" ]

-ef is not defined by POSIX, but is present in ksh and Dash.

(( ))

(( )) (without the $) acts like a command on its own

For simple comparison: [ -lt ] (and -ne -gt -ge) or [ $((3 + 1 < 5)) -ne 0 ].

present in ksh

To assign a variable var=$((3+1))

Note: several standard POSIX utilities can be used for lexical comparisons. The examples below return a true (zero) exit status if the content of $a sorts before $b.

  • awk -v v1="$a" -v v2="$b" 'BEGIN { exit !(v1 "" < "" v2) }'

  • expr "x$a" "<" "x$b" >/dev/null

  • printf "%s\n" "x$a" "x$b" | sort -c >/dev/null 2>&1 (also returns true if $a and $b are equal)

Arithmetic

Works in bash

Change to for dash

Comment

pre/pos increment/decrement

++ --

i=$((i+1)) or : $((i+=1))

-

comma operator

,

: $((...)); cmd $((...))`

The comma operator is widely supported by almost everything except dash -- even posh and Busybox.

-

let or ((...))

[ $((...)) -ne 0 ]

Because of the above comma restriction, let can't be simulated exactly without a loop.

Redirections

Works in bash

Change to for dash

Comment

redirect both stdout and stderr

>& and &>

command > file 2>&1 or command 2>&1 | othercommand

-

|& (bash4)

command 2>&1 | othercommand

Conflicts with ksh. Not recommended, even in Bash. Just use 2>&1.

duplicate and close

m>&n- m<&n-

m>&n n>&-

not defined by POSIX

herestring

<<<"string"

echo | command, or a here document to avoid a subshell (<<EOF)

-

Builtins

  • echo -n or -e -- POSIX doesn't define any options, and furthermore allows echo -e to be the default behavior. Instead use printf "%s\n" (for normal echo) or printf "%b\n" (for echo -e); leave off the \n to simulate echo -n.

  • printf -v is not defined by POSIX, and only Bash supports it. The %q and %()T formats are not defined by POSIX but supported by ksh93 and Bash. The a, A, e, E, f, F, g, and G formats are not required by POSIX for printf(1), but dash appears to support %f, %e, %E, %g, and %G.

  • read -- the only option defined by POSIX is -r; ksh has a different set of options that only partially overlaps with bash.

  • shopt, and therefore all the options it provides (extglob, nullglob, dotglob, etc.) are not defined by POSIX and are bash-specific

  • local -- there is no POSIX equivalent. You can use $funcname_varname to reduce the likelihood of conflicts, but even that is not enough for recursive functions. You can ensure that recursive calls occur in subshell environments (so there is a "local" copy of all variables), or pass all "local variables" as parameters (because the positional parameters $@, $1, $2, etc are truly local). dash explicitly supports local as a non-Posix extension; ksh uses typeset instead, which works like bash's declare.

Special Variables

Works in bash

Change to for dash

Comment

keep track of the times

SECONDS

before=$(date +%s) ....seconds=$(( $(date +%s) - $before))

date +%s is not POSIX; see this faq for more info. Present in ksh

Generate a random number

RANDOM

random=$(awk 'BEGIN{srand(); printf "%d\n",(rand()*256)}') gives a number between 0 and 256
random=$(hexdump -n 1 -e '/1 "%u"' /dev/urandom) and random=$(od -A n -N 1 -t u1 /dev/urandom) give a timer-independent number between 0 and 256
random=$(hexdump -n 2 -e '/2 "%u"' /dev/urandom) and random=$(od -A n -N 2 -t u2 /dev/urandom) give a timer-independent number between 0 and 65535

Be sure to learn what srand() and rand() do, ie this method fails if you call awk several times rapidly. Instead generate all the numbers you need inside awk. Some systems also provide /dev/random and /dev/urandom , but this is not necessarily mandated by the POSIX standard. ksh has RANDOM

Get the status of all the commands in a pipeline

PIPESTATUS

Simplest solution:
mkfifo fifo; command2 <fifo & command1 >fifo; echo $?
see NamedPipes

bash-specific; see this faq and this script pipe status for POSIX shell

Get the name of all / the current function name(s)

FUNCNAME

??

bash-specific see stackoverflow question

More

  • The bash manual has a list of the differences between bash running in POSIX mode and a normal bash.

Note that bash in POSIX mode is only guaranteed to run a shell written according to the POSIX specification. It doesn't mean that it will fail if you use bashisms in your scripts.

  • There is a handy perl script checkbashisms which is part of the debian devscripts package which can help point out bashisms in a particular script.


CategoryShell

Bashism (last edited 2022-10-20 23:13:29 by larryv)