Decide what precedence .. should have

[ftxqxd]

Currently, 1i + 1i..1i + 1i is parsed as 1i + (1i..(1i + 1i)). This seems problematic. I would expect it to be parsed as (1i + 1i)..(1i + 1i) (to match with how the current slicing syntax behaves), or possibly as 1i + (1i..1i) + 1i.

[nagisa]

I think cc @nick29581

[nrc]

I agree they should be parsed as (1i + 1i)..(1i + 1i)

[nrc]

I think 1..2..3 should be an error rather than resolving either way, since there is not an obvious right answer

[sinistersnare]

@nick29581 could 1..2..3 work if there was a Range<Range<int,int>, int> implementation? (random question)

[nrc]

@sinistersnare We could make it work in one direction or the other, but I'm not sure I'd want to - (1..2)..3 is not too much hassle and it's a lot clearer

[sinistersnare]

Sure, makes sense :)

[dgrunwald]

So did we decide on a specific precedence level? I'd argue it should be lower than arithmetic operators (which is now fixed); but higher than comparison operators (range == 1..2 should be valid, but currently isn't).