We now request only property paths from the repository query
that are requested through a GraphQL query if the result type
is not a projection. That reduces the amount of data being
retrieved from the underlying data store.
See gh-168
Prior to this commit, we could only test Spring GraphQL applications
with a complete application - all application and infrastructure
components were involved. While using `@SpringBootTest` is often useful
for complete integration tests (with or without a live running server),
we often want to write lean integration tests and test slices of our
application.
Just like `@WebMvcTest` or `@WebFluxTest`, this commit introduces the
support for `@GraphQlTest`. This annotation helps us to test a
particular slice of our application: a hand-picked selection of
`@Controller`, plus `RuntimeWiringConfigurer` and `WebInterceptor`
beans.
Other `@Component` must be imported or mocked for those tests.
This commit also refactors the existing auto-configuration to enable
this use case. The `WebGraphQlHandlerAutoConfiguration` now holds the
required components for `@GraphQlTest`, while other web-related
auto-configurations bring the web framework and transport
infrastructures.
Closes gh-75
This commit ensures that Java beans using primary constructors are
properly instantiated when they're nested and that the instantiation
algorithm is called recursively.
Fixes gh-155
Show map-based batch loading first which is easier to implement without
having to order results.
Use JUnit named arguments in BatchMappingInvocationTests.
See gh-130
Also separate RequestStrategy from WebRequestStrategy since those are
delegated to internally and never used as a hierarchy. The generics
required to make them extend from each other bring no benefit.
This commit configures a `FormattingConversionService` into the argument
resolution infrastructure, using the expected conversion service for
both MVC and WebFlux.
Closes gh-148
Rename AnnotatedDataFetcherConfigurer -> AnnotatedControllerConfigurer,
since handler methods are now not only data fetchers but also
batch loaders.
Move DataFetcherHandlerMethod and BatchLoaderHandlerMethod into the
support package next to AnnotatedControllerConfigurer where they're used.
See gh-130
Given that BatchLoaderRegistry is declared as a bean and injected into
ExecutionGraphQlService as a DataLoaderRegistrar, it makes sense to
have be a DataLoaderRegistrar vs declaring the bean as the
implementation type. A BatchLoaderRegistry is supposed to result in
DataLoader registrations so this makes sense in any case.
Prior to this commit, the `@Argument` annotation would support two
attributes: `required` and `defaultValue`, specifying whether arguments
should be considered as required and default values to use if they're
not present.
This can be problematic as this can already be specified at the schema
level.
To avoid duplication and invalid setups betweens schema and Controllers,
this commit removes those attributes in favor of the schema definition.
Closes gh-150
Prior to this commit, the `GraphQlArgumentInstantiator` would not
instantiate and bind properly types with a list attribute that has
elements which need to be instantiated with a primary constructor.
This commit ensures that every list that is encountered is
recursively passed to the instantiator, instead of trying to
convert the list using the databinder.
This also fixes gh-145, by directly accessing the arguments from
the arguments map, instead of the mapped propertyvalues used for
the DataBinder.
See gh-147
This commit adds a new `"graphql.request.datafetch.count"` distribution
summary metric for counting the number of data fetching calls per
request.
Closes gh-146
Prior to this commit, a missing, non-required input argument would throw
an NullPointerException instead of returning `null` or
`Optional.empty()`.
Fixes gh-144
