Add syntax sugar for maps

[macabeus]

Closes #137

This code

var set1 = {1, 2, 3}
var set2 = {"foo", "bar": "baz"}

now is the same of:

var set1 = {1: true, 2: true, 3: true}
var set2 = {"foo": true, "bar": "baz"}

[mhermier]

Not a big fan of the default true value assignment. Duplicating the key to the value sounds more logical to me. But it has other problems/implications...

[macabeus]

@mhermier What do you suggest? I think that it's a simple and useful implementation to give a set-like type in Wren.

true is a good choice in order to do something like:

> var x = {1, 2, 3}
> if (x[1]) { System.print("foo") } //> "foo"
> if (x[5]) { System.print("foo") } //> nil

[mhermier]

I suggest that `var a = { 1, 2, 3}` is equivalent to `var a = { 1: 1, 2: 2, 3: 3}`. Since most values are evalued to true, it should be equivalent to your suggestion. Le 10 juin 2018 8:16 PM, "Bruno Macabeus" <notifications@github.com> a écrit : @mhermier <https://github.com/mhermier> What do you suggest? I think that it's a simple and useful implementation to give a set-like feature in Wren. true is a good choice in order to do something like:

var x = {1, 2, 3} if (x[1]) { System.print("foo") } //> "foo" if (x[5]) { System.print("foo") } //> nil

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#558 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AY7blbK9y_mjvwfifP2WUU3-nGp3uj60ks5t7WJkgaJpZM4UhkT5> .

[Akaricchi]

But then you can't put false in a "set" like this, which is a bit counter-intuitive. What exactly is wrong with true?

[mhermier]

There is no problems with true, there are advantage of having the key back has values. When the keys match, there is no guarantee that it is the same object, so you have to traverse the list of keys to get the key back and use that instance.

[Akaricchi]

https://github.com/munificent/wren/blob/master/src/vm/wren_value.c#L389

Every object that can be hashed is immutable. Why would user code ever need to care whether two immutable objects of equal value are actually the same object or not?

[mhermier]

Because this immutability is an implementation detail of the map. It should not restrict language design.

[ruby0x1]

Thanks for the PR!
Right now, I don't feel like this fits just right with Wren as is, so I'm gonna close it.