tilde iife

[plexigras]

x~(->y) would be behave like (->y)(x) and be transformed to (function() { return y; })(x);

[wcjohnson]

Hm. The behavior you describe here would probably not work out as written. It would need to be a placeholder arg:

'use @oigroup/lightscript with placeholderArgs'
x~(-> _)

Aside from that, the tilde call syntax is specifically written to disallow arbitrary expressions on the RHS. I shall have to ask @rattrayalex why it is that way. There might be some reason I'm not aware of.

Another question: what do you see as the value and/or win here? Do you have any particular use cases in mind?

[plexigras]

you are right about it probably not working it would have to be something like x~(->y)() because you can't do x~y but have to do x~y()

to why i think when it would be usefull for example when you want to do a lodash less mapvalues you could do something like this

obj
  ~something()
  ~something()
  ~something()
  ~(x ->
    Object.keys(x).reduce((acc,key) ->
      acc[key] = x[val] + 1
      acc
    , {})
  )()
  ~something()
  ~something()
  ~something()

[rattrayalex]

Hmm. That seems a little ugly to me but I do see where it's coming from, and I can't presently think of a reason it should be illegal... I don't remember why I made it that way to begin with, other than generally thinking it'd be ridiculous/unreadable (which, looking at @plexigras 's example, would be an exaggeration).

So I don't think I'd be opposed to adding this feature.

[wcjohnson]

I will point out that this is potentially a use case for the pipe operator, which is now on the JS standards track. In fact,

x~(-> stuff)!

is actually just

x |> -> stuff

I also think the pipe operator reduces a little bit of ugliness from the syntax, as it doesn't require the ! or ().

'use @oigroup/lightscript with noEnforcedSubscriptIndentation`
obj
  ~something()
  ~something()
  ~something()
  |> x ->
    Object.keys(x).reduce((acc,key) ->
      acc[key] = x[val] + 1
      acc
    , {})
  ~something()
  ~something()
  ~something()

Incidentally, this is also a case study in why noEnforcedSubscriptIndentation is a good idea. Without it:

obj
  ~x()
  |> x -> x
    ~deepIndent()
      |> x -> x
         ~deeperIndent()