interruptible(boolean) reads differently than its intent

[etspaceman]

I'm wondering if we should create an alias around this. For example, interruptible(false) reads like it isn't interruptible. Perhaps we should consider interruptibleOnce and interruptibleMany.

[vasilmkd]

I think adding interruptibleOnce and interruptibleMany is indeed the easiest way to introduce this change, without pushing the ergonomics of enums in Scala 2 (sealed trait InterruptibleMode and case object Many, case object Once). I do see this as a positive change and I think we should do it.

[djspiewak]

I generally agree this would be a good change, though I wish we could use interruptible and interruptibleMany instead. Technically we can it would just require a pretty major source compatibility breakage.

[vasilmkd]

I wish that too but I'm against source breakage and the hassle. Maybe later down the line. I would like for us to go ahead with interruptibleOnce and interruptibleMany for now.

[djspiewak]

I think the counter argument is that the "many" version is quite weird and unlikely to be commonly necessary, thus it would be better to carrot people to the asymmetrical default. It's also not totally clear to me how common the method is in the wild, so we may have a window of opportunity to improve it without hideous pain and suffering.

[vasilmkd]

Well put. I do agree with this very much.

[vasilmkd]

Do you want me to attempt a scalafix migration too?

[djspiewak]

Do you want me to attempt a scalafix migration too?

Very much yes! It seems nicely migrateable in most cases.