Peter Jaszkowiak

I think we should heavily discourage abbreviating the debug output of map or list. I'm not sure if making that easier is the right course of action.

Personally, I want my debug output to include all the information available. Hiding information just because there's a lot of it seems antithetical to the whole idea of debug. Like...

I'm curious if this RFC supports the creation of a "resizeable" `Box`. This would support when-necessary reallocation when the internal unsized type is grown. It would hold extra capacity to...

> I personally wouldn't call the old range types "legacy" ranges because, if the new ranges impl `IntoIterator`, then it makes sense for their `IntoIter` to be the old range...

@ahicks92 thanks for bringing this up, I hadn't considered the indexing case. I think many common cases will be covered by `SliceIndex`, but there will be collections that will need...

I will just mention that `(a..b).into()` probably won't work unless the only index type is `Range` - otherwise it will probably fail with a type inference error. We may want...

> I can't find a crate right now but ironically the example that came to mind is ndarray (which sadly uses a different type for slicing and thus isn't applicable)...

My thoughts: - We need actual data about this. Right now we don't know if this will be an actual issue in practice. The evaluation of `std` was only an...

> I don't think we can. One either picks 2024 and converts for 2021 crates, or picks 2021 and converts for 2024 crates. Right? Not really. For the indexing issue,...

> I doubt anyone is going to cover both. That's not a big ask but you then have to teach newbies to the language that due to legacy reasons they'd...