Differences between revisions 10 and 12 (spanning 2 versions)
Revision 10 as of 2014-04-28 22:43:24
Size: 4357
Editor: ormaaj
Comment: Exported function pitfalls
Revision 12 as of 2019-06-06 17:29:00
Size: 4485
Editor: GreyCat
Comment:
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:
You can use the shell's `-c` option to run the shell with the sole purpose of executing a short bit of script:
Line 4: Line 5:
You can use the shell's `-c` option to run the shell with the sole purpose of executing a short bit of script:
{{{
{{{#!highlight bash
Line 10: Line 10:
{{{ {{{#!highlight bash
Line 17: Line 17:
{{{ {{{#!highlight bash
Line 35: Line 35:
{{{ {{{#!highlight bash
Line 42: Line 42:
{{{ {{{#!highlight bash
Line 50: Line 50:
{{{ {{{#!highlight bash
Line 58: Line 58:
{{{ {{{#!highlight bash
Line 76: Line 76:
{{{ {{{#!highlight bash

How do I invoke a shell command from a non-shell application?

You can use the shell's -c option to run the shell with the sole purpose of executing a short bit of script:

   1 sh -c 'echo "Hi!  This is a short script."'

This is usually pretty useless without a means of passing data to it. The best way to pass bits of data to your shell is to pass them as positional arguments:

   1 sh -c 'echo "Hi! This short script was run with the arguments: $@"' -- "foo" "bar"

Notice the -- before the actual positional parameters. The first argument you pass to the shell process (that isn't the argument to the -c option) will be placed in $0. Positional parameters start at $1, so we put a little placeholder in $0. This can be anything you like; in the example, we use the generic --.

This technique is used often in shell scripting, when trying to have a non-shell CLI utility execute some bash code, such as with find(1):

   1 find /foo -name '*.bar' -exec bash -c 'mv "$1" "${1%.bar}.jpg"' -- {} \;

Here, we ask find to run the bash command for every *.bar file it finds, passing it to the bash process as the first positional parameter. The bash process runs the mv command after doing some Parameter Expansion on the first positional parameter in order to rename our file's extension from bar to jpg.

Alternatively, if your non-shell application allows you to set environment variables, you can do that, and then read them using normal variables of the same name.

Similarly, suppose a program (e.g. a file manager) lets you define an external command that an argument will be appended to, but you need that argument somewhere in the middle. In that case:

#!/bin/sh
sh -c 'command foo "$1" bar' -- "$@"

Calling shell functions

Only a shell can call a shell function. So constructs like this won't work:

   1 # This won't work!
   2 find . -type f -exec my_bash_function {} +

If your shell function is defined in a file, you can invoke a shell which sources that file, and then calls the function:

   1 find . -type f -exec bash -c 'source /my/bash/function; my_bash_function "$@"' _ {} +

(See UsingFind for explanations.)

Bash also permits function definitions to be exported through the environment. So, if your function is defined within your current shell, you can export it to make it available to the new shell which find invokes:

   1 # Bash
   2 export -f my_bash_function
   3 find . -type f -exec bash -c 'my_bash_function "$@"' _ {} +

This works ONLY in bash, and isn't without problems. The maximum length of the function code cannot exceed the max size of an environment variable, which is platform-specific. Functions from the environment can be a security risk as well because bash simply scans environment variables for values that fit the form of a shell function, and has no way of knowing who exported the function, or whether a value that happens to look like a function really is one. It is generally a better idea to retrieve the function definition and put it directly into the code. This technique is also more portable.

   1 # Bash/ksh/zsh
   2 
   3 function someFunc {
   4     :
   5 }
   6 
   7 bash -s <<EOF
   8 $(typeset -f someFunc)
   9 
  10 someFunc
  11 EOF

Note that ksh93 uses a completely different approach for sourcing functions. See the manpage for the FPATH variable. Bash doesn't use FPATH, but the source tree includes 3 different examples of how to emulate it: http://git.savannah.gnu.org/cgit/bash.git/tree/examples/functions/autoload.v3

This technique is also necessary when calling the shell function on a remote system (e.g. over ssh). Sourcing a file containing the function definition on the remote system will work, if such a file is available. If no such file is available, the only viable approach is to ask the current shell to spit out the function definition, and feed that to the remote shell over the ssh channel:

   1 {
   2     declare -f my_bash_function
   3     echo "my_bash_function foo 'bar bar'"
   4 } | ssh -T user@host bash

Care must be taken when writing a script to send through ssh. Ssh works like eval, with the same concerns; see that FAQ for details.


CategoryShell

BashFAQ/012 (last edited 2019-06-06 17:29:00 by GreyCat)