You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If I use the JSON-AD parser, I know it will throw because Commits don't have an @id by default. I guess it might be a good idea to give commits identifiers.
I've used the signature of commits + some base URL in identifiers in the past, e.g. https://example.com/commits/YTIfURstCKlYfibyuYUfFyu. If I require this, the one creating the commit needs to know which URL is acceptable - and that means that a client making a change needs to know which route is used for commits. That seems like it could bite me in the future.
I can also just use the signature as the identifier, but that means that they cannot be resolved, because there is no schema. In a way, this is true - a commit doesn't live anywhere yet (until it is validated).
I can also add some conditional to the JSON_AD parser, or create a new function for parsing commits which uses the same logic. That way, Commits still won't have an ID.
I think that feels like the most sensible thing to do. So no new IDs for commits, yet.
Currently uses serde, and has nonsensical behavior (only accepts strings, no support for nested resources or arrays).
The text was updated successfully, but these errors were encountered: