Jonathan de Jong

We can have it be derived from the likes of defining paths with `"/:type[/:key]"` and having that be validated.

I think, at the moment, as we're only working with one trailing path segment, I'll do the following; - Add `trailing_segment: bool` to `Metadata`, to easily define this - Add...

After some talk in #ruma-dev, i've come to the conclusion to scrap the above idea, and go with this instead; Use a specific enum to represent both a non-truncated path...

We could possibly also generate some neat documentation from this; ```md Endpoint History: - Unstable ([MSC1234]), using path "`/_matrix/client/unstable/org.matrix.msc1234`". - Matrix Version [1.2], stablized, using path "`/_matrix/client/v3/stuff`". - Matrix Version...

Additionally, on the `Metadata` struct itself, this could refactor `added` and friends to be methods, with just one `history` slice field which contains all this information.

After some discussion we decided to split the work, the new plan should reflect this. https://github.com/ruma/ruma/pull/1298 was a previous attempt to merge this in a larger chunk, we decided to...

I forgot about https://github.com/ruma/ruma/issues/1067, which I kind-of want to have fixed here. https://github.com/ruma/ruma/issues/901 also was part of #1298, and had a fix that worked nicely there.

Made a issue to track for that on here: https://github.com/poljar/matrix-nio/issues/229

Additionally; "Ignore" can be expanded to "ignore for this match" and "ignore for any match", the former will notify if a new Notify match appears in a different list, but...

Sounds like some spec/schema changes need to occur then to signal that it's a manually-approved delegated ban, sourcing from a different list, instead of being from your own list.