echox

echox is a framework for writing web applications in shell languages. It manages a number of shell processes and executes functions in them in response to web requests.

Why?

I am not sure. Because I can, I guess.

Should I use this for my next "microservice"?

Absolutely not.

Basic Configuration

Configuration can be supplied from a file or to stdin. The configuration language is simple and only uses a few commands:

include [file] - Loads a source file into all shell processes
[method] [path] [func] - Requests for the given method and path invoke the given function

For example, this server will return the current date:

echo "get / date" | echox

Or, you can load a function from another file

# hello.sh
say_hello() {
    now=$(date)
    echo "Hello! It is $now"
}

echo "include hello.sh; get / say_hello" | echox

Semicolons or newlines can be used as separators. An equivalent config file could look like this:

# example.txt
include hello.sh

get / say_hello

echox example.txt

Request Handling

Your function will be invoked every time an HTTP request matches the supplied path. Shell processes are locked when executing, so you only have to worry about dealing with one request at a time.

Route Parameters

Route parameters are expanded into shell variables. Using them is trivial:

# greet.sh
greet_by_name() {
    echo "Hello, $name!"
}

echo "include greet.sh; get /greet/:name greet_by_name" | echox

Headers

Request headers are supplied in the headers variable. Each header is separated by a newline, and header names and values are separated by a :. This allows them to be interpreted with coreutils.

# headers.sh
get_user_agent() {
    for h in $headers; do
        name=$(echo $h | cut -d ':' -f 1)
        if [ "$name" = "User-Agent" ]; then
            agent=$(echo $h | cut -d ':' -f 2)
            echo "Your user agent is $agent"
        fi
    done
}

tmlbl / echox