Linux – Bad interpreter error when executing shell script with the correct shebang line

linuxshell-script

When executing a script from the path (scriptname, rather than "sh scriptname") I receive the following error:

-bash: ./gitup: /bin/bash/: bad interpreter: Not a directory

All of the other scripts in the same directory execute without issue, and I do not encounter this issue when the script is executed manually (sh scriptname), only when executed by itself. So the script, itself, seems to not be the issue as it does execute normally, and did so until I opened a new shell session and encountered this error for this script alone.

The script is executable, includes all requisite lines to execute in the way I've described (it worked properly until earlier today) and is able to be executed by my user, so I am unsure of what the issue could be.

The shebang line reads #!/bin/bash. I used the command "export PATH=$PATH:repos/utilities
" to direct to the directory storing these scripts, all of which, besides this one, execute simply by calling the script name (i.e. gitup).

Best Answer

There are a few things I can think of.

You said your script is in your $PATH, yet you call it with its full path (./gitup as opposed to just gitup). Could it be that you have edited one version of the script (the one in your $PATH, say ~/bin/gitup) and have then executed another (./gitup)?
Could there be multiple versions in your $PATH? You can check with
```
which -a gitup
```
You may have inadvertently entered a strange invisible character while editing the script. Check the shebang line using od:
```
head -1 ./gitup | od -c
```
Make sure there is nothing after #!/bin/bash but the newline character (\n). It should look like this:
```
0000000   #   !   /   b   i   n   /   b   a   s   h  \n
0000014
```

Related Solutions

Linux – #!/usr/bin/env Interpreter Arguments — portable scripts with arguments for the interpreter

Here's an inline solution to work around the portability problem for ZSH.

#! /bin/sh

if [ -z "$IN_ZSH" ]; then
  export IN_ZSH=1
  exec zsh -f "$0" "$@"
fi

## Your ZSH script here

Some other methods to try include

Passing env variables via /usr/bin/env which modify the behavior the same as --options such as doing

    #! /usr/bin/env POSIXLY_CORRECT=1 bash

Using clever commenting tricks, such as how this script starts out using SH but invokes TCL on the same script

    #! /bin/sh
    # \
    TCLBIN=/usr/bin/tclsh; \
    exec $TCLBIN "$0" "$@"
    # Execute the rest via tclsh
    set argc

Setting the option once in the interpreter, if the --options being passed in do not affect the load behavior

    #! /usr/bin/env bash
    # Exit if any error detected
    set -e

For perl, if you're able to use newer versions, this may work in lieu of -w:

    #! /usr/bin/env perl
    use warnings;

Using a bootstrap invoke.sh script instead of /usr/bin/env to use your PATH, calling with /path/to/invoke.sh script with your script starting with #! zsh -f

    #! /bin/sh

    SCRIPT=$1
    shift 1
    cmd=`sed -n -e 's:#! \?::' -e '1p' $SCRIPT`
    exec $cmd $SCRIPT

Should I put .sh and .rb file extensions on all the scripts

You can do wildcard commands like ls *.rb or cp *.sh if you want to organize your scripts in the future.

Start early or regret later, in my opinion.

Editors like vim will also be able to apply the correct syntax highlighting based on shebang or file extension.

This can also be accomplished by using modelines in various editors. E.g. for vim:

# vim: ft=sh

Best Answer

Related Solutions

Linux – #!/usr/bin/env Interpreter Arguments — portable scripts with arguments for the interpreter

Should I put *.sh and *.rb file extensions on all the scripts

Related Question

Should I put .sh and .rb file extensions on all the scripts