I certainly make that change more than once per day. I especially do when I write build.gradle files, where I'm particularly happy that Groovy has that feature. I'm going to be jumping over to build.gradle.kts files once the Kotlin scripts are more mature, but I'm really going to miss trailing commas, possibly more than anything else from Groovy

the fact that I can move these lines around without considering punctuation saves a lot of time. And I frequently reorganize these arrays when I add new dependencies to group them together logically

Short of having this supported in the language, maybe IDEA could simply remove last commas e.g. after a paste or when lines have just been moved around. Might be a bit jarring though

yeah, I'm not too crazy about the editor adding and removing non-whitespace

I am for automatic adding/removing comas in the event of deleting or commenting a line out. IDE already does that for imports

allowing trailing commas is a much simpler solution IMO

@kevinmost isn’t the job of a code editor to add and remove non-whitespace 🙂

I dunno... I guess, but I'm not convinced this is a better idea than just supporting something that makes sense, doesn't rely on editor specific features, and doesn't harm readability or ease of use in the slightest, while making things like long parameter lists and array literals much easier to write

it just looks wrong 🙂

Should a

?:

be evaluated on an argument of an infix function call, or on the call's result? From general:

>>> val m = mutableMapOf<String, String>()
>>> "a" to m["b"] ?: "was null"
(a, null)
>>> "a" to (m["b"] ?: "was null")
(a, was null)

I wasn't expecting the first result. Perhaps I should have but it seemed obvious to me that the

?: "was null"

would be applied to

m["b"]

and not

"a" to m["b"]

. The response on #general was that the first result was obvious, but is it?

in any case changing operator precedence rules would be a breaking change, and therefore is not going to happen

right

for any set of precedence rules, you can come up with examples that will look confusing

true enough

Perhaps an IJ intention then? 🙂 The type error in IJ was quite confusing...

required:

List<Pair<String, String>>

, found

List<Serializable>

that could work, yes

I'll create an issue, thanks!

@User uploaded a file:

It even has an intention “Remove useless elvis operator"

@paulblessing Try adding a

map

surrounding that expression. It will hide the error you see above...

OK, was just going based on the code as it was posted

Yeah, I didn't know about that existing intention because my code didn't show it

Interesting, seems even the

map

is not enough to cause what I was seeing:

@User uploaded a file:

Not sure what's different about my code...

Ah, here we go:

@User uploaded a file:

Yeah, in that case I just get the type mismatch with nothing about the useless elvis