Can't destroy through association.

[yonahforst]

class Configuration
  include Her::Model
  has_many :publish_sets
end

class PublishSet
  include Her::Model
  belongs_to :configuration
end

This association works as expected for get requests:

Configuration.find(1234).publish_sets
GET .../configurations/1234/publish_sets

but for delete requests:

Configuration.find(1234).publish_sets.first.destroy
DELETE .../publish_sets/5678

I need to make it so that delete requests use the full nested url DELETE .../configurations/1234/publish_sets/5678

I tried setting collection_path("/configurations/:configuration_id/publish_sets") but the publish_set objects i get back from the api don't have the configuration_id param, only a configuration object.

Is there anyway to do this kind of delete through the association instead of a delete_raw (how i'm doing it now)

[remi]

No, unfortunately setting collection_path("/configurations/:configuration_id/publish_sets") is the right way to do it.

I’ll check if there’s a way we could add support for intersting path parameters back into the fetched resource’s attributes when her initializes it.

[yonahforst]

thanks, that would be amazing.

I'm not sure how this API has their models set up, maybe has_many :through, or polymorphic. In any case, they don't return the relation id, rather the entire object. would it make sense to be able to do something like:

 collection_path("/configurations/#{configuration.id}/publish_sets")

Also, I noticed that every object has its own url embedded as an attribute.

#<Skytap::Configuration(configurations/799324) id="799324" url="https://cloud.skytap.com/configurations/799324" name="234921" error="" runstate="stopped" suspend_on_idle=1800 routable=false .....>

I don't know if this is standard practice with REST APIs. If it is, would it make sense to allow the objects path to be inferred from the object itself (url attribute)?

[billc]

I am having a similar issue with attempting to destroy_existing on a nested resource. Has any progress been made to address this issue?

[billc]

@joshblour Do you have a gist of how you used delete_raw as a work around?