* Make the executor and validation APIs public to enable splitting parsing and execution into two stages
Based on https://github.com/graphql-rust/juniper/pull/773#issuecomment-703783048 and https://github.com/graphql-rust/juniper/pull/773#issuecomment-704009918
* Fix fmt warning for visit_all_rules
* Add `Definition` to the public API so it's available for the result of compilation
* Make OperationType public so that user land code can tell the difference between query/mutation and subscription
* Add integrations tests for execute_validated_query_async, visit_all_rules, get_operation, and resolve_validated_subscription
Co-authored-by: Christian Legnitto <LegNeato@users.noreply.github.com>
* Mark `Arc` with `IsInputType` & `IsOutputType`.
We've just updated a service we work with to the latest release of
juniper (it's been on a fairly old master commit for some time). In
this service we have some fields that are contained within `Arc`s, which
I added support for in #479.
Since then it seems the new marker traits of `IsInputType` &
`IsOutputType` have been added, but they do not support `Arc`, leading
to:
```
error[E0277]: the trait bound `Arc<menu::Menu>: IsOutputType<__S>` is not satisfied
--> src/graphql.rs:36:1
|
36 | #[juniper::graphql_object(Context=Context)]
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the trait `IsOutputType<__S>` is not implemented for `Arc<menu::Menu>`
|
= note: required by `juniper::marker::IsOutputType::mark`
= note: this error originates in an attribute macro (in Nightly builds, run with -Z macro-backtrace for more info)
```
This commit adds support for those, mirroring the support for `Box`.
I'm not sure if there's a good place to add a test for this
functionality so that regressions can be avoided in the future?
* Add Arc tests & FromInputValue for Arc<T>
- add associated type to IntoResolvable and IntoFieldResult traits allowing to name the GraphQLType being resolved
- relax Sized requirement on some IsInputType and IsOutputType impls
- add object safe GraphQLValue, GraphQLValueAsync and GraphQLSubscriptionValue traits containing methods for value resolution
- refactor GraphQLType, GraphQLTypeAsync and GraphQLSubscriptionType traits to register and provide GraphQL type information only
- make GraphQLUnion trait generic over ScalarValue
- generate generic over ScalarValue impls of GraphQLUnion trait
- add codegen tests with a custom ScalarValue for union macros
- reimplement #[derive(GraphQLUnion)] macro to support:
- both structs and enums
- generics in type definition
- multiple #[graphql] attributes
- external resolver functions
- remove From trait impls generation for enum variants
- reimplement #[graphql_union] macro to support:
- traits
- generics in trait definition
- multiple attributes
- external resolver functions
- GraphQLType implemetation for a raw trait object
- GraphQLTypeAsync implemetation (#549)
- add marker::GraphQLUnion trait
- rewrite "2.5 Unions" section in Book (Juniper user documentation)
- rewrite `codegen` and `codegen_fail` integration tests for GraphQL unions
Additionally:
- re-export `futures` crate in `juniper` for convenient reuse in the generated code without requiring library user to provide `futures` crate by himself (#663)
- use unit type () as default context for EmptyMutation and EmptySubscriptions
- relax Sized trait bound on some GraphQLType and GraphQLTypeAsync definitions, implementations and usages
* Implemented most test cases from the specification
* Unified error handling for all generators
- Removed proc-macro-error -> not required -> use syn::Error
- Everything below lib.rs uses proc_macro2::TokenStream
instead of proc_macro::TokenStream
- Replaced error handling in attribute parsers
* WIP better error messages for *all* macros
* Refactored GraphQLInputObject and minor tweaks
- removed support for Scalar within a string ("DefaultScalarValue")
- removed unraw function and replaced it with the built-in one
- added error messages and return types for all functions within utils
- added more constraints to fulfill the GraphQL spec
* Fixed test-cases which are not compliant with the specification
* Removed unused function
* Added constrains, updated error messages, added marker
* Added argument rename within impl_graphql and fixed `__` tests
* Formatted and cleanup
* Added GraphQLTypeAsync for input object
* Moved codegen tests to separate module
Nightly and stable produce different outputs, thus only test nightly.
* Added IsInputType/IsOutputType traits for type checking
Co-authored-by: Christian Legnitto <LegNeato@users.noreply.github.com>
Fixes#571
Signed-off-by: andrisak <andrisak@gmail.com>
* Replaced the old macro with the new proc macro.
Updated documentation.
Signed-off-by: andrisak <andrisak@gmail.com>
Co-authored-by: Christian Legnitto <LegNeato@users.noreply.github.com>
* Updated implementation of derive enum
- allows context specification
- allows scalar specification
- shares code with derive object
* Added this feature to CHANGELOG
* Added matrix with supported macro attributes for enums
* Added case which checks for custom context
* GraphQLUnion now can use a different context per variant
* Moved context switch test for union into right folder
* Sync resolve expression has the same form as the other impls
* Disabled custom scalar on GraphQLEnum
* Fixed CHANGELOG
* Fixed support matrix of GraphQLEnum in the book
- scalar not supported!
- skip not supported!
* Added test case for "noasync" derive attribute
* Disallowed generics and lifetimes on GraphQLEnum
* Added error message for duplicate naming
* Added error message for empty variant
* Implemented device macro for GraphQLUnion's
* Updated PR link in CHNAGELOG
* Disabled documentation on enumeration fields
* Disabled skip on fields
* Changed implementation for std::convert::Into since skip is not possible
* Added documentation for GraphQLUnion
* Added tests for GraphQLUnion
* Fixed typos in error messages (as suggested by review)
* Fixed failing documentation example
* Utilized `resolver_code` in `util::GraphQLTypeDefinitionField`.
Simplifies code and provides the idea of using
`util::GraphQLTypeDefinitionField` for different types than objects.
* Removed wrong statement about skip annotation in docs.
Co-authored-by: Christian Legnitto <LegNeato@users.noreply.github.com>
there is still some weirdness going on. Running async and
non-async tests in `integration_tests/*` works, but running it
from `integration_tests` does not.
* feat: Raw identifier support in object macro
This commit implements raw identifier (`r#name`) support
for field names (methods) and arguments in the `object` proc macro.
Eg:
```rust
impl T {
fn r#type(r#trait: String) -> bool {}
}
```
* Rebase onto master
* Fix merge [skip ci]
This commit implements raw identifier (`r#name`) support
for field names (methods) and arguments in the `object` proc macro.
Eg:
```rust
impl T {
fn r#type(r#trait: String) -> bool {}
}
```
style: Enable rustfmt merge_imports and format
This commit enables the rustfmt merge_imports setting
and formats the whole code base accordingly.
Note that the setting is not stable yet, but will be with Rust 1.38.
In the meantime, running fmt on stable will just leave the
changes alone so no problems should occur.
