Roadmap for and state of single element tuple types?


(Slava Pestov) #41

Thanks for the writeup -- I didn't know canonical types were a @Douglas_Gregor invention :slight_smile:

Minor nitpick though:

I can't imagine that adding or removing a sugared type should have ABI impact -- mangling uses canonical types only. Or do you mean if we changed (T) to be a different type than T? That would be major breaking and quite odd from a language standpoint, I agree.


(Chris Lattner) #42

Yeah I mean that changing (T) to be different than T would be seriously source/ABI breaking. Similarly, any change from a sugar type to a canonical type would be a similar breaking change.