When using the relationship link with a block, calling "meta" without "href", e.g.
has_one :bio do
link :related do
meta id: 1
end
end
results in in a nil "href", e.g.
{ links: { posts: { related: { href: nil, meta: { id: 1 } } } } }.
According to JSONAPI, we should be able to use meta without href
(http://jsonapi.org/format/#document-links).
PR #1454 was merged with some missing fixes. All these fixes are
addressed by this commit:
- Rename ActiveModel::Serializer::Adapter::JsonApi::Association to
ActiveModel::Serializer::Adapter::JsonApi::Relationship
- Move ActiveModel::Serializer::Adapter:: JsonApi::Relationship and
ActiveModel::Serializer::Adapter::JsonApi::ResourceIdentifier to
ActiveModel::Serializer::Adapter::JsonApi::ApiObjects module
- Add unit test for
ActiveModel::Serializer::Adapter::JsonApi::Relationship
- Add unit test for
ActiveModel::Serializer::Adapter::JsonApi::ResourceIdentifier
Breaking change:
- Adapters now inherit Adapter::Base
- 'Adapter' is now a module, no longer a class
Why?
- using a class as a namespace that you also inherit from is complicated and circular at time i.e.
buggy (see https://github.com/rails-api/active_model_serializers/pull/1177)
- The class methods on Adapter aren't necessarily related to the instance methods, they're more
Adapter functions
- named `Base` because it's a Rails-ism
- It helps to isolate and highlight what the Adapter interface actually is
I was seeing transient failures where adapters may not be registered.
e.g. https://travis-ci.org/rails-api/active_model_serializers/builds/77735382
Since we're using the Adapter, JsonApi, and Json classes
as namespaces, some of the conventions we use for modules don't apply.
Basically, we don't want to define the class anywhere besides itself.
Otherwise, the inherited hooks may not run, and some adapters may not
be registered.
For example:
If we have a class Api `class Api; end`
And Api is also used as a namespace for `Api::Product`
And the classes are defined in different files.
In one file:
```ruby
class Api
autoload :Product
def self.inherited(subclass)
puts
p [:inherited, subclass.name]
puts
end
end
```
And in another:
```ruby
class Api
class Product < Api
def sell_sell_sell!
# TODO: sell
end
end
end
```
If we load the Api class file first, the inherited hook will be defined on the class
so that when we load the Api::Product class, we'll see the output:
```plain
[ :inherited, Api::Product]
```
However, if we load the Api::Product class first, since it defines the `Api` class
and then inherited from it, the Api file was never loaded, the hook never defined,
and thus never run.
By defining the class as `class Api::Product < Api` We ensure the the Api class
MUST be defined, and thus, the hook will be defined and run and so sunshine and unicorns.
Appendix:
The below would work, but triggers a circular reference warning.
It's also not recommended to mix require with autoload.
```ruby
require 'api'
class Api
class Product < Api
def sell_sell_sell!
# TODO: sell
end
end
end
```
This failure scenario was introduced by removing the circular reference warnings in
https://github.com/rails-api/active_model_serializers/pull/1067
Style note:
To make diffs on the adapters smalleer and easier to read, I've maintained the same
identention that was in the original file. I've decided to prefer ease of reading
the diff over style, esp. since we may later return to the preferred class declaration style.
with '#' will be ignored, and an empty message aborts the commit.
Pagination links will be included in your response automatically as long
as the resource is paginated using Kaminari or WillPaginate
and if you are using a JSON-API adapter. The others adapters does not have this feature.
It's an upgrade based on the new Cache implementation #693.
It allows to use the Rails conventions to cache
specific attributes or associations.
It's based on the Cache Composition implementation.
