Elvis Pranskevichus

Agree with @vpetrovykh re scoping of tools under `e.group`. Option 1 seems fine to me, as the loss of immediate shape is easily alleviated by composing a `e.select`.

How do other database drivers handle this?

On the `edgedb migrate` path the CLI isn't doing anything special other than sending the content of `*.edgeql` files from `dbschema/migrations` (which is where `edgedb create-migration` puts them). The only...

`apply` isn't exactly right, because we don't "apply" `and` to each argument, it's more like "join".

> I'm against sugar (`e.and()`) on top of sugar I'm also a bit wary of adding `e.and()` sugar right away. IMO we should stick to the "generic first, sugar later"...

> Oh, wait... maybe "join" is a bad and confusing keyword in the database context. Exactly.

That's right. `edgedb project` and `edgedb instance` are intended to be ran on a development machine and the instances they manage listen on localhost by default. Here's a [draft guide](https://github.com/edgedb/edgedb/blob/cb086f3032d67210d66d2e3e288c1549a318316f/docs/guides/deployment/gcp.rst)...

The instructions are [here] (https://github.com/edgedb/imdbench/blob/master/DEVELOP.rst), though I don't think you'll be able to run this in Codespaces.

There is also precedent for a top-level function in the [Python standard library](https://docs.python.org/3/library/dataclasses.html#dataclasses.asdict).

> unctions like `dataclasses.asdict` and `namedtuple._asdict` are not recursive `dataclasses.asdict` _is_ recursive: > Each dataclass is converted to a tuple of its field values. dataclasses, dicts, lists, and tuples are...