27 Commits

Author	SHA1	Message	Date
Dmitrii Skurikhin	8503c0f288	enhance query errors	2022-02-16 08:58:16 +01:00
Dmitrii Skurikhin	1f7bd92d11	fix index position in error path	2022-01-07 08:31:47 +01:00
Eugen Wissner	116aa1f6bb	Put JSON support behind a flag	2021-12-24 13:35:18 +01:00
Eugen Wissner	eedab9e742	Don't append a trailing newline in gql	2021-09-22 08:50:20 +02:00
Eugen Wissner	60d1167839	Test nullability on value completion	2021-09-17 10:01:14 +02:00
Eugen Wissner	b96d75f447	Replace the old executor	2021-09-03 22:47:49 +02:00
Eugen Wissner	7b4c7e2b8c	Handle argument locations	2021-09-02 08:45:23 +02:00
Eugen Wissner	233a58094d	Adjust value completion tests	2021-09-01 09:27:12 +02:00
Eugen Wissner	b580d1a988	Attach the field location to resolver exceptions	2021-06-27 13:42:58 +02:00
Eugen Wissner	96bb061666	Fail with a location for result coercion ... The intermediate representation was further modified so that the operation definitions contain location information. Probably I should introduce a data type that generalizes fields and operations, so it contains object type, location and the selection set, so the functions don't accept so many arguments.	2021-06-24 09:29:24 +02:00
Eugen Wissner	812f6967d4	Provide locations for argument errors ... The executor still doesn't give an error per argument, but a single error per field with locations for all arguments. If a non-null argument isn't specified, only the error location of the field is given. If some arguments cannot be coerced, only the locations of these arguments are given, non-null arguments are ignored. This should still be improved, so the executor returns all errors at once. The transformation tree is changed, so that argument map contains locations of the arguments (but not the locations of the argument values yet).	2021-06-22 09:13:27 +02:00
Eugen Wissner	6fe9eb72e4	Fix merging fields with arguments ... executeField shouldn't assume that a selection has only one field with a given name, but it should take the first field. The underlying cause is a wrong pattern, which (because of the laziness) is executed only if the field has arguments.	2021-06-18 06:51:14 +02:00
Eugen Wissner	2ce2be5d91	Provide location information for interface errors	2021-06-17 08:15:27 +02:00
Eugen Wissner	c311cb0070	Add constructor with additional schema types	2021-05-13 17:40:38 +02:00
Eugen Wissner	1b7cd85216	Add location information to the intermediate tree	2021-05-12 06:51:59 +02:00
Eugen Wissner	0d23df3da2	Provide an internal function to add errors ... The old function, addErrMsg, takes only a string with an error description, but more information is required for the execution errors: locations and path. addErrMsg should be deprecated after the switching to the new addError.	2021-05-09 12:42:02 +02:00
Eugen Wissner	7c0b0ace4d	Collect types once the schema is created	2020-10-07 05:24:51 +02:00
Eugen Wissner	4602eb1df3	Validate arguments are defined	2020-09-28 07:06:15 +02:00
Eugen Wissner	08998dbd93	Validate fragments don't form cycles	2020-09-11 08:03:49 +02:00
Eugen Wissner	09135c581a	Constrain base monad to MonadCatch ... Let's try MonadThrow/MonadCatch. It looks nice at a first glance. The monad transformer stack contains only the ReaderT, less lifts are required. Exception subtyping is easier, the user can (and should) define custom error types and throw them. And it is still possible to use pure error handling, if someone doesn't like runtime exceptions or need to run a query in a pure environment. Fixes #42.	2020-07-17 07:05:03 +02:00
Eugen Wissner	e24386402b	Respect subscriptions in the executor ... After the last commit there were a few places needed to be adjusted to support subscriptions. This is done and a test case is added. It is important to implement subscriptions now, because they require changes to the library API, and they are a big missing part to finish the executor. When the executor is finished, we can start to provide more stable API without breaking everything every release. Validation and introspection shouldn't require much changes to the API; AST would require some changes to report good errors after the validation - this is one thing I can think of. Fixes #5.	2020-07-15 19:15:31 +02:00
Eugen Wissner	ae2210f659	Support subscriptions ... This is experimental support. The implementation is based on conduit and is boring. There is a new resolver data constructor that should create a source event stream. The executor receives the events, pipes them through the normal execution and puts them into the response stream which is returned to the user. - Tests are missing. - The executor should check field value resolver on subscription types. - The graphql function should probably return (Either ResponseEventStream Response), but I'm not sure about this. It will make the usage more complicated if no subscriptions are involved, but with the current API implementing subscriptions is more difficult than it should be.	2020-07-14 19:37:56 +02:00
Eugen Wissner	a6f9cec413	Handle errors using custom types ... Fixes #32.	2020-07-05 14:36:00 +02:00
Eugen Wissner	705e506c13	Combine Resolver and ActionT in ResolverT	2020-06-29 13:14:23 +02:00
Eugen Wissner	882276a845	Coerce result ... Fixes #45.	2020-06-13 07:20:19 +02:00
Eugen Wissner	e8c54810f8	Merge selection sets	2020-06-12 07:58:08 +02:00
Eugen Wissner	c37b9c88b1	Skip unknown fields	2020-06-10 11:42:00 +02:00