absinthe_error_payload
absinthe_error_payload copied to clipboard
mirego
Bridges the gap between Ecto and Absinthe for mutation payload
Thanks for the great library, I've got a point of confusion I'd like help understanding. > The primary philosophy is that messages generated by invalid and/or unexpected user input are...
## 📖 Description and reason Since we initialize the struct with an empty list, we should type the GraphQL field accordingly. ## 🦀 Dispatch `#dispatch/elixir`
## 📖 Description and reason Casting a `nil` value for an association adds an error of type `{:array, :map}`. We did not support this type of key on message builder....
## 📖 Description and reason The messages list included in the payload object was nullable, and each message in the list was also nullable. This makes the responses more ambiguous...
Previously, a GraphQL document containing nested fields could not be tested if one of the values was null / `nil`. The library stringifies the expected response, coercing `nil` to `""`....
https://github.com/mirego/absinthe_error_payload/blob/691194ea3360e6e38c8278d5c51f9f9546cd0e5b/lib/absinthe_error_payload/payload.ex#L125 A better schema seems should be like this: ``` field(:messages, list_of(non_null:validation_message)) ``` Any reason allowing nullable validation messages in the list?
Add `t:AbsintheErrorPayload.ValidationMessage.t/0` for typespec referencing
Hi guys, I ran into an issue with elixirs `Range` not being handled properly. This snippet from my changeset ```elixir |> validate_inclusion(:value, 6..20) ``` produces this error message when the...
