This isn't a good idea at all, but you can produce the C++ behavior by
overriding next() in a mutable fashion, e.g.
@groovy.transform.TupleConstructor
class Bar {
int value
Bar next() {
++value
this
}
}
Bar b = new Bar(0)
++(++(++b))
assert b.value == 3
But in reality, it's just eliminating the effects/ordering of re-assignment
by only keeping one instance of Bar in play.
-Joe
On Tue, Feb 14, 2017 at 12:10 PM, Jochen Theodorou <[email protected]>
wrote:
>
>
> On 13.02.2017 23:58, Thibault Kruse wrote:
>
>>
>>
>> On Feb 14, 2017 3:52 AM, "Jochen Theodorou" <[email protected]
>> <mailto:[email protected]>> wrote:
>>
>>
>> Options:
>>
>> (2) do allow these post- and prefix operators only on
>> VariableExpressions, making the code above no longer compile
>>
>>
>> Just to be sure, a VariableExpression is an expression consisting only
>> if a variable name?
>>
>
> yes, but yes, you are right, there are also PropertyExpressions and
> ArrayExpressions... so I was narrowing it down too much.
>
> And this applies to both the pre and postfix ++ operator?
>> So ++i++ would also stop compiling?
>>
>
> yes
>
> So the question could be rephrased as: is there any good usecase for ++
>> outside the increment of a variable? Like
>> x = (((a * b)++)/c)--
>>
>
> yes... and that even if considering we call next() for ++. I could imagine
> some crazy DSL to use this... but should we still consider doing this
> breaking change?
>
> bye Jochen
>
