New structured syntax for bash commands

[Ph0enixKM]

There are some quirks that make the command syntax hard to use. For instance one must remember to properly escape strings like $cmd "{text}"$ and also think about other stuff... Maybe we should introduce some command builder syntax? Something like:

sh cmd_name(text, 12) {
    --arg: name,
    --out: nameof variable,
    -q: true,
    -r: false // This will not render as an argument
}

or even we can remove the -- for standard field arguments:

sh cmd_name(text, 12) {
    arg: name,
    out: nameof variable,
    -q: true,
    -r: false
}

What are your thoughts on this?

[Mte90]

I would like that in comments so it is more easy to extract them and just keep the code as code

[KrosFire]

I'm not sure how the given example should work, could you explain it like I'm 5 yo? XD

As for the idea I see one big issue. In bash if you're not using basic codeish functionalities, then you're using third party apps. Those apps can exist on a machine or not. Different versions of given app, can have different API (command names, args names, order of them etc.). While you're using $$ syntax you know that you're in bash and it has it's own rules, and you're aware it's a sticky situation. If you have some app with CLI and you want to create API in amber for this CLI, so that people can easily use it, there needs to be a good way to for you to cover all your bases, so that you know it's safe to use.

By that I mean some good syntax for creating different versions of functions, for different versions of application. Amber would need to have the knowledge how to get and extract app version to execute correct code. In different versions some functions could exist and in others there would not. We need to account for that and tell it somehow to a user. We could give this responsibility to programmers and say to them "create different versions of your lib, for different versions of the app", but I think this approach is limiting. One Amber script can run on many different machines and making people choose correct version of the script for correct machine is worse than having one script that we incrementally improve for all machines. This is the part where Amber could really shine and be a good alternative for bash.

With current syntax this is pretty easy. We could do something like this:

let version = $my-app --version$?

if {
  version == "2.6.0" {
   // do something...
  }
  version == "1.6.0" {
  // do something else...
 }
  else {
  // do default thing
 }
}

It works but it's cumbersome. If we will create this command builder we would need to think this through thoroughly, so that it's intuitive and safe.

[Ph0enixKM]

I see now. I agree that commands are too unpredictable to handle them in a form of some syntax. Whatever we implement - some of them might be incompatible with amber out of the box. I think that ultimately we’ll create a package registry where there will be a package/library for any command maintained by either the owner of the package or the community.

[Ph0enixKM]

We should do some good versioning built-into the language (or the ecosystem) itself