Post History

How to write a bash function to sanitize filenames for Linux and Windows

I'm trying to write a bash function that can sanitize filenames to make them compatible with both Linux and Windows file systems. The function should perform the following operations:

1. Replace invalid characters with similar valid ones (e.g., replace `*` or `+` with `-` or `_`, `?` with `¿`, etc).
2. Remove leading and trailing spaces before the file extension.
3. Replace multiple consecutive spaces, carriage returns, etc., with a single space.
4. Replace single spaces with underscores (`_`).

Here's the function I've come up with:

```bash
sanitize_filename() {
    local input_str="$1"
    
    # Replace invalid characters
    input_str="${input_str//[^a-zA-Z0-9_\-\. ]/-}"
    
    # Remove spaces at the beginning and end before the extension
    input_str="${input_str#"${input_str%%[![:space:]]*[![:space:]]}"}"
    input_str="${input_str%"${input_str##*[![:space:]]}"}"
    
    # Replace multiple spaces, carriage returns, etc. with a single space
    input_str=$(echo "$input_str" | tr -s '[:space:]' | tr -s '\r')
    
    # Replace single spaces with underscore
    input_str="${input_str// /_}"
    
    echo "$input_str"
}

# Test cases
test_cases=(
    "my file 123.txt"
    "my+file+123?.txt"
    "file_with/special*characters.jpg"
    "   leading_spaces.docx"
    "trailing_spaces   .txt"
    "multiple    spaces.txt"
    "multiple  +  spaces.txt"
    "carriage\r\r return.txt"
)

# Expected sanitized filenames
expected=(
    "my_file_123.txt"
    "my-file-123¿.txt"
    "file_with_special-characters.jpg"
    "leading_spaces.docx"
    "trailing_spaces.txt"
    "multiple_spaces.txt"
    "multiple-spaces.txt"
    "carriage_return.txt"
)

# Test the function
for ((i=0; i<${#test_cases[@]}; i++)); do
    result=$(sanitize_filename "${test_cases[i]}")
    if [ "$result" == "${expected[i]}" ]; then
        echo "Test $i: PASSED"
    else
        echo "Test $i: FAILED"
        echo "Expected: ${expected[i]}"
        echo "Got: $result"
    fi
done
```

I get the following output:

```
❯ ./sanitize_filenames.sh
Test 0: PASSED
Test 1: FAILED
Expected: my-file-123¿.txt
Got: my-file-123-.txt
Test 2: FAILED
Expected: file_with_special-characters.jpg
Got: file_with-special-characters.jpg
Test 3: PASSED
Test 4: FAILED
Expected: trailing_spaces.txt
Got: trailing_spaces_.txt
Test 5: PASSED
Test 6: FAILED
Expected: multiple-spaces.txt
Got: multiple_-_spaces.txt
Test 7: FAILED
Expected: carriage_return.txt
Got: carriage-r-r_return.txt
```

How can I fix this function to handle all the test cases correctly?