Stephen Haberman

Huh, I'm surprised that is not already getting pruned; seems like it should be.

@chr1sjf0x I tried to reproduce this in `EntityManager.queries.test.ts` but couldn't, seems like it's working? ```ts it("can prune subtype only fields", async () => { const where = { publisherLargePublisher: {...

Added a passing test in https://github.com/joist-orm/joist-orm/pull/1250

I was able to reproduce this, and pushed up a branch with a test showing that the STI "where type_id = ..." condition is not dropped: https://github.com/joist-orm/joist-orm/pull/1465

:tada: This issue has been resolved in version 1.242.1 :tada: The release is available on: - `v1.242.1` - [GitHub release](https://github.com/joist-orm/joist-orm/releases/tag/v1.242.1) Your **[semantic-release](https://github.com/semantic-release/semantic-release)** bot :package::rocket:

@chr1sjf0x per "could return a TaskNew or TaskOld subclass", doesn't the presence of `specialNewField` basically dictate that only `TaskNew`s would be returned from this query, so the code should be...

Hi @coreyzzp ; apologies for the late response, definitely sounds like a bug around here: https://github.com/stephenh/ts-proto/blob/b4bf37d0bc7c68abc7929d82be09e378a28f762b/src/main.ts#L1894 If you could work on a PR, that'd be great! Thanks!

Currently our preloading looks like: ```sql -- select authors & preload a.books and a.comments select b._ as b, c._ as c, a.* from authors a -- join into books >...

We should try and do this instead: ```sql select json_build_object('a', row_to_json(a), 'c', c.c, 'b', b.b) from authors a cross join lateral ( select json_agg(json_build_object('br', br.br, 'b', row_to_json(b))) as b from...

We've got this `entitiesDirectory` that is supposed to drive most things: ``` entitiesDirectory: z.string().default("./src/entities"), ``` But looks like maybe the `graphql-codegen` side of things never bothered using that? Is this...