Feature request: { foo bar } style records syntax

[taoeffect]

In Janet, it can be very verbose to call functions that have a large number of keyword-based arguments. In JS this is somewhat simplified by the ability to write objects like so: { foo, bar, baz: 5 } which gets expanded to { foo: foo, bar: bar, baz: 5 }

It would be awesome if we had this for tables and structs in Janet too so that we could write { x y z } instead of { :x x :y y :z z }

In the meantime I'm living with this $: macro:

(defmacro $: [& args]
  (var pairs @[])
  (var i 0)

  (while (< i (length args))
    (let [x (args i)]
      (if (keyword? x)
        # Explicit key/value: :baz bazz
        (do
          (when (>= (+ i 1) (length args))
            (error (string/format "$: missing value for keyword %j" x)))
          (array/push pairs x)
          (array/push pairs (args (+ i 1)))
          (set i (+ i 2)))

        # Shorthand symbol: foo => :foo foo
        (do
          (when (not (symbol? x))
            (error (string/format "$: expected symbol or keyword, got %j" x)))
          (array/push pairs (keyword x))
          (array/push pairs x)
          (set i (+ i 1))))))

  ~(table ,;pairs))

[sogaiu]

At this stage of Janet's life I don't think making this kind of change would be a good idea, but perhaps other folks have different opinions.

Apart from that though, if you consider the case of an even number of elements:

{x y z a}

How is one supposed to tell that it doesn't actually mean the following?

# keys are from `x` and `z`, values from `y` and `a`
{x y
 z a}

and not shorthand for:

{:x x
 :y y
 :z z
 :a a}

That is, it seems ambiguous...though may be I am missing something.

[taoeffect]

I suppose it would be a breaking change in that regard. But that also seems like such a rare edge case…

[sogaiu]

If you haven't seen this interview with bakpakin, may be it's worth a look. Specifically, please see this bit:

With making things last, the idea is enabling forwards compatibility and non-breaking changes, making sure there are resources for future programmers even when we’ve all moved on. Documentation shouldn’t disappear, forum posting should at least be public if not well organized.

[taoeffect]

Yeah I can respect that. Maybe some new syntax or some built-in optimized version of $: macro could be added instead?

It works like so: ($: x y z) #=> @{:x x :y y :z z}

[sogaiu]

As mentioned in these issues:

• assoc macro... #1501 (comment)
• Add function to generate UUID v4 #1504 (comment)
• Convenient math operators #1507 (comment)

I think in most cases it's better to have these types of things to be vetted in spork first so that people can provide feedback and then modifications can be made if needed.

[hamidb80]

a rare edge case…

it's not a rare edge case :( it would break my code