Colin Rofls
Colin Rofls
I'm not going to say 'impossible to deal with', but ValueRecord has a unique combination of qualities: - it cannot be parsed naively, because you need to know the appropriate...
Your reasoning is correct. FWIW traversal currently only exists as a mechanism to generate reasonable impls of `Debug`, and it definitely has some warts/missing impls. It is possible to tell...
(oops, thought I'd replied, sorry!) So you're totally right, here, and this is a situation I've been hemming & hawing about for the past year. Basically: traversal's main justification is...
interesting, will take a look. :)
A rough outline of how I imagine this should be implemented: - there is already a step we perform for each record/table, after we have parsed all the fields: in...
I think we could do that eventually, but it's an unusual case and it isn't clear what code we would want to generate. I prefer to force the user to...
Traversal is in a weird spot; it was really a sort of proof-of-concept I wanted to play around with, and then ended up using it to implement Debug. It is...
Yea, so part of my thinking with traversal is that it would give us a common mechanism that could then be easily reused for `Debug`, `PartialEq`, `Hash`, etc; otherwise we...
Let me see what that looks like and we can go from there.
wonderful, happy to have someone more comfortable than I am with the soundness rules looking at this stuff. For the overlapping refs issue: I think the problem here is that...