Simon Binder

The line you've referenced would actually be executed both for batches and for regular inserts, so I don't see how that should make a difference. I have slightly different line...

The API still looks good to me, but I think there should be a way to be more explicit about the impact of joining a reverse reference. Things that suddenly...

> I've spent a considerable amount of time on this, but I'll have to agree, in practice this is handing a loaded gun to a child. Yeah, unfortunately this looks...

It's not an easy API to design for sure. My "list all joins beforehand" idea also suffers from poor discoverability. I'm happy to help if as well if you have...

Pagination would definitely be awesome to have, but I think it doesn't really solve the stream invalidation problem (every write to either table would still invalidate every page in every...

Since you have to await the query either way I'm not sure it makes much of a difference in the end, most methods using the API will be `async` either...

> Can I just do `todosTable.category as RowId` or should I be doing `RowId(todosTable.category)`? If we want to allow this, we should be applying the type converter (`categoriesTable.id.converter.fromSql(todosTable.category)` probably).

Returning pre-populated managers that can then be fetched afterwards is brilliant! I definitely agree with that approach (and also with potentially adding prefetching/caching under the same interface later where it...

Alright, I'll wait :+1: > Static Shock or MkDocs sound good? I agree that the content could certainly use a structured rewrite and the landing pages need some love, but...

That sounds good to me, thanks for taking the time to make the changes easier to review :+1: As for the breaking changes, I'd like to coordinate them with other...