Daniel Mewes

The paths are not expanded through a shell, so they are currently just taken as they are. I'm not sure if this is important enough for us to support within...

@shinzui That depends on your particular use case. The easiest option is often to just implement the recursive part in your application code, and to fetch the data you need...

@fixpunkt One option for doing arbitrary recursions with `fold` is emulating a call stack in the accumulator. That's not necessary for a BFS traversal though, since all you need for...

Looks like `fold` is not currently implemented for grouped streams. I wonder if there's a reason for this or if we just didn't consider that case. @nighelles do you remember...

This is by design. Keeping open as a ReQL proposal to add an option to change the behavior. Pretty low priority though since there's a simple work-around.

https://github.com/rethinkdb/horizon/issues/704 has a list of the corresponding issues for the missing pieces in the main Horizon tracker.

@chipotle `hz schema apply` does not convert the schema file itself. If you want to get a new schema file that uses the current format, you have to apply it...

One more thing, which I think we forgot about: There are new rdb_timeout, rdb_user and rdb_password options for the `config.toml` file, the CLI commands, and the Horizon server constructor. See...

@chipotle The new format for specifying an index in the `schema.toml` file for `hz schema apply`: https://github.com/rethinkdb/horizon/issues/706#issuecomment-238413049

@intellix Thanks for the example. Your example is essentially using the old schema.toml syntax for defining indexes (where each index is simply described by a string), though with the new...