Chance
Chance
It wasn't horrible to write the iterator below, but I think it could be made a lot cleaner. If `Component` were something like ```rust enum Component Self { Self {...
Ah, that is interesting. I can't quite get my head around how it'd feel interacting with it though. It may be too complex for some use cases while incredibly useful...
One thing that would have cut down the complexity of the iterator a bit is a `resolve_token` method on `Resolve` / `Pointer`. It would have eliminated this: ```rust let resolvable...
I think we'd need some sort of constructed pointer type if the idea is to use the cursor to traverse a value. `BTreeMap` has the advantage of owning both key...
I guess the `Cursor` could have an internal `PointerBuf` that it manipulates and then hands out `&'cursor Pointer` **edit**: nah, this would be a bad idea.
The niche comment likely answers a question I was going to present, which is whether providing a walk impl like the above would make any sense. I actually have it...
@asmello I was mistaken on the `split_back`. I had to revert it. I submitted a draft PR #83 that has the `WalkTo` impl. The codebase is a complete mess right...
@asmello I was on mobile earlier and hadn't read all of your reply - I just wanted to push up the source I had mentioned before responding to the rest....
Ah, sorry! Yea, it's a mess over in that repo at the moment. I'm in the middle of rewriting it to use the iterator but I have to swap things...
The more I think about this, the more convinced I am that the walk iterators don't belong in this crate, at least not in their current form. In fact, I...