<@U5UU34LPK> I often resort to `!!` on `.min()` or...

# language-proposals

[poohbar]

@karelpeeters I often resort to !! on .min() or .max() calls where I know the collection is not empty. I kind of wish we had a NonEmptyList that would return non-nullable types from such functions. Do you guys have a better way to handle this?

[benleggiero]

A non-empty list type would be nice. Otherwise one cannot be sure that another thread has not emptied the collection by the time you call .min()

[dalexander]

They could always add support for dependent types 🧌 However I guess you could add extension methods for min/max that either accepted a fallback value if the list was empty, or returned null if the list was empty (then use ?: for your fallback value). Or you could stick with !! you’ll get some kind of exception if you ever do have an empty list, which I think is fine because you’re expecting something not empty?

[poohbar]

min(), max() already return null for empty lists AFAIK

[poohbar]

so you can do:

myList.min() ?: 0
// or
myList.min()!!

but both seem ugly when I know 100% that the list is not empty, maybe because it's in a local variable that I have just initialized with values one row above.

[benleggiero]

I personally dislike how the stdlib diverges from its common approach of .fooOrNull() here, but also personally, I wish "or null" was the default (like these) instead of the current "or crash"

[poohbar]

whatever the choice is I agree that it should at least be consistent

[gildor]

fun List<Int>.minOrZero() = min ?: 0

[gildor]

But agree that default implementation of NEL would be nice addition for stdlib

[diego-gomez-olvera]

#arrow has one http://arrow-kt.io/docs/datatypes/nonemptylist/