Clarification on cache_key options about graphql-ruby-fragment_cache HOT 6 CLOSED

Hey! The difference is that when you pass only argument (second snippet) it will be used both for building the cache key and as a value for caching, and when you pass an argument and a block — it will build the key using object but return the value.

For instance, imagine that post somehow depends on the user. In this case you might want to cache the post for this particular user, not for everyone:

def post(id:)
  post = Post.find(id)
  cache_fragment([post, current_user]) { post }
end

In case of cache_fragment(Post.find(id)) you do not have this granularity of control: it will grab post, build a key from it and return it back.

from graphql-ruby-fragment_cache.

Sorry for being obtuse here, but its still not clear from the documentation what the difference between :value and :object is in terms of cache_key:.

from graphql-ruby-fragment_cache.

@gsdean but the explanation above is more clear, right?

from graphql-ruby-fragment_cache.

Sadly no, I think what would be helpful is to clarify the difference in terms of just the attribute without overriding the getter. In other words, what is the difference between:

field :post, PostType, null: true, cache_fragment: {cache_key: :value}

and

field :post, PostType, null: true, cache_fragment: {cache_key: :object}

?

from graphql-ruby-fragment_cache.

Also not to hijack the thread, but have you seen this #72 (comment)

There is a pretty serious race condition that we should probably address.

from graphql-ruby-fragment_cache.

When cache_key is equals :value then the value from the resolver method will be used as a key, when you use object — current object will be used as a key.

For instance impagine a post that has the author, and for some reason we want author cache to be refreshed only when post is updated. In this case we can do this:

class Post
  field :cached_author, Types::User, null: true, cache_fragment: {cache_key: :object}

  def cached_author
    object.author
  end
end

class Query < Base
  field :post, Post, null: true do
    argument :id, GraphQL::Types::ID, required: true
  end

  def post(id:)
    ::Post.find(id)
  end
end

If you keep this code as is than Post instance will be a key for this cached author, otherwise — author itself.

The same thing but in code https://github.com/DmitryTsepelev/graphql-ruby-fragment_cache/blob/master/spec/graphql/fragment_cache/field_extensions_spec.rb#L171.

from graphql-ruby-fragment_cache.