German Velasco

Hi @yordis, I believe `params_for` intentionally removes any associations. From the [params_for docs](https://hexdocs.pm/ex_machina/ExMachina.Ecto.html#c:params_for/2) > Will return a map with the fields and virtual fields, but without the Ecto metadata, the...

@jeremy-hanna thanks so much for opening such a detailed and helpful issue! 🙏 This is still an issue that I'm not sure how to resolve, so I'm very open to...

Thanks @robacarp and @jeremy-hanna for commenting on this. I just wanted to loop back and say that I still haven't figured out how to solve this, but it is something...

Hi @speeddragon, you're totally right that it doesn't set the foreign key when using `build/2`. `build/2` simply grabs the struct defined in the factory and merges any attributes you pass...

@s3cur3 thanks for the suggestion! I think that's possible for `assert_has/3`, but I wonder if it'll lead to two other points of confusion: 1. Will people try to do use...

@RobertoSchneiders thanks so much for opening this issue! Could you confirm what version of PhoenixTest you're using? I'm wondering if this is related to #203. If you point your app...

@RobertoSchneiders sorry for the confusion. We had briefly fixed the issue, but it introduced a regression in forms that didn't have a `phx-change` at all (so I didn't ship it...

@sevenseacat thanks so much for adding a demo app that reproduces the error! 🙏 Can't tell you how helpful that is. I've spent a few hours looking at this, and...

Yeah, I don't know that Ash is doing anything out of the ordinary there. I think it's just a limitation of how much PhoenixTest knows about the changing UI, and...

@sevenseacat I think I have a solution for your problem. Just got merged in #208. Can you test out `main` whenever you have a chance and see if that fixes...