Writing Portable Scripts (Bash-isms vs POSIX sh)

#!/usr/bin/env bash
echo "found bash via PATH lookup, running fine"

#!/bin/bash
fruits=("apple" "banana")
if [[ ${#fruits[@]} -eq 2 ]]; then
    echo "arrays and [[ ]] are bash-specific features working here"
fi

text="hello-world"
echo "${text//-/ }"

#!/bin/bash
value="banana"
case $value in
    apple) echo "it's an apple" ;;
    banana) echo "it's a banana (POSIX-compatible case statement)" ;;
    *) echo "unknown fruit" ;;
esac

if [ "$value" = "banana" ]; then
    echo "single-bracket test also works everywhere"
fi

1. Using #!/bin/sh as the shebang while still relying on bash-only features like arrays or [[ ]], which fails on systems where /bin/sh is a different, stricter shell.
2. Assuming [[ ]], arrays, local, and += are available everywhere; they are bash extensions, not part of POSIX sh.
3. Hardcoding #!/bin/bash and assuming bash is always at that exact path; #!/usr/bin/env bash is more portable across systems where bash lives elsewhere.

• #!/usr/bin/env bash finds bash wherever it is on the PATH, which is more portable than hardcoding #!/bin/bash.
• Arrays, [[ ]], local, string manipulation like ${var//a/b}, and namerefs are bash-isms not guaranteed in POSIX sh.
• If a script must run under plain sh, stick to [ ] (single bracket), avoid arrays, and use case instead of bash-only pattern matching.
• Running shellcheck with the correct --shell target helps catch accidental bash-isms in a script meant to be POSIX-compliant.