Arno Dirlam

Sorry for re-opening an old issue, but this is of critical importance for us as well. Numerous components use sub-processes, most recently puma in clustered mode (for zero-downtime deployment) and...

Sorry for reviving and old (but open) issue. Here's how I "solved" the issue: ``` # Override `BunnyMock::Queue#subscribe` to mimic the default exchange behavior # of RabbitMQ. Also add the...

Tried with - Dataloader 1.0.8 and master - Ecto 3.6.2 and master

Turns out it just expects an `:ok` tuple, which we did not return. So this works: ```elixir Ecto.Multi.new() |> Ecto.Multi.run(:whatever, fn _, _ -> loader = Dataloader.new() |> Dataloader.add_source(:repo, Dataloader.Ecto.new(MyApp.Repo))...

We ran into issues again, now with a correct return value for `Multi.run/3`. Please see my updated [PR branch](https://github.com/absinthe-graphql/dataloader/pull/130). ``` [error] Task #PID started from #PID terminating ** (DBConnection.ConnectionError) could...

@benwilson512 not sure if you've seen I re-opened this. I tried to fix it, but didn't get very far in my attempts. Can you confirm that it's an issue in...

@benwilson512 Yes, the tests in the linked PR should fail: #130

Thanks a lot for confirming the issue and for coming up with a PR, @giddie 🙏 I've done more research, and found out the following: ### Why does dataloader use...

Also @seddy who worked on the async parts a lot 👌

So this wasn't you? @seddy https://github.com/absinthe-graphql/dataloader/commit/ce9ac3f3e4649dd15a3094ad50ba841511455d31 🤔 EDIT: Yea, we're talking ages ago 😄