Understanding Module.sh

What Module.sh does is quite simple. The complexity comes alone from making it work under different conditions. So it should be easy to understand.

The basic Logic

Whenever we write module "NAME" "https://example.com/ABC.sh" Module.sh replaces that line with the content from the URL, plus a thin wrapper.

That wrapper makes it available under the name and creates a sub-shell to separate the scope:

$name() {(
# [download content]
)}

A simple Example

Let us use the hello-world module

#!/bin/sh

printf 'Hello World\n'

and use it in a normal script (demo.sh):

#!/bin/sh
set -eEuo pipefail

eval "$(curl -fsL
"https://mdl.sh/development/module/online/module-static-0.9.21.sh")"
module "helloWorld" "https://mdl.sh/misc/hello-world/hello-world-1.0.1.sh"

helloWorld

Now we will download the compiler and make it executable:

curl -O https://mdl.sh/development/tools/compiler/compiler-static-0.9.26.sh
chmod +x compiler-0.9.26.sh

To compile the demo.sh we use the following command:

./compiler-0.9.0.sh demo.sh demo-static.sh

Afterwards we can see the compiled result (demo-static.sh):

#!/bin/sh
set -eEuo pipefail

# start module https://mdl.sh/misc/hello-world/hello-world-1.0.1.sh
helloWorld() {(
printf 'Hello World\n'
)}
# end module https://mdl.sh/misc/hello-world/hello-world-1.0.1.sh

helloWorld

So basically it looks like we expected it. The lines for loading module.sh and loading the hello-world module got replaced by the content of the hello-world module plus the wrapper (and some additional comments).