enhance(stitch): introduce SubschemaSetConfig

[yaacovCR]

to improve ergonomics of sharing endpoint details between subschemas

[theguild-bot]

The latest changes of this PR are available as alpha in npm: 6.2.1-alpha-6d92adab.0

Quickly update your package.json by running:

npx match-version @graphql-tools 6.2.1-alpha-6d92adab.0

[gmac]

I see. So this structure is what computed field isolations would/should produce? Looks good, I like that there’s still just a single subschema.

[yaacovCR]

Yes, but now we don't even need a new type you can just return an array of SubschemaConfig objects.

#2001

[yaacovCR]

To share an endpoint and pool batching, just use identical object with endpoint details in the endpoint property of the members of the array instead of putting the endpoint details within the parent SubschemaConfig object.

May make sense in v7 to make the endpoint details always be required to be within the endpoint property instead of parent object just for consistency, but I think small potatoes.

[gmac]

I’m not familiar with how endpoint works yet... is it a new expectation that executor goes into endpoint to facilitate batching, and one needs to update their subschema config with this property to enable batching?

[yaacovCR]

Batching will work if executor in SubschemaConfig, but to enable batching across SubschemaConfigs, it must be a property of an endpoint object with the same object then used in multiple SubschemaConfig objects. See

graphql-tools/packages/delegate/tests/batchExecution.test.ts

Lines 107 to 135 in 6a2e5ab

	const endpoint: Endpoint = {
	batch: true,
	executor: ((params: ExecutionParams): ExecutionResult => {
	executions++;
	return execute(innerSchemaA, params.document, undefined, params.context, params.variables) as ExecutionResult;
	}) as SyncExecutor
	};
	const innerSubschemaConfigA: Array<SubschemaConfig> = [{
	schema: innerSchemaA,
	transforms: [new FilterObjectFields((typeName, fieldName) => typeName !== 'Object' || fieldName !== 'field2')],
	merge: {
	Object: {
	fieldName: 'objectA',
	args: () => ({}),
	},
	},
	endpoint,
	}, {
	schema: innerSchemaA,
	transforms: [new FilterObjectFields((typeName, fieldName) => typeName !== 'Object' || fieldName !== 'field1')],
	merge: {
	Object: {
	fieldName: 'objectA',
	args: () => ({}),
	},
	},
	endpoint,
	}];

[gmac]

Got it. Thanks!

…

On Sat, Sep 5, 2020 at 8:05 PM Yaacov Rydzinski ***@***.***> wrote: Batching will work if executor in SubschemaConfig, but to enable batching across SubschemaConfigs, it must be a property of an endpoint object with the same object then used in multiple SubschemaConfig objects. See https://github.com/ardatan/graphql-tools/blob/6a2e5ab8a1efd4b7524437ee0be5a03d7f312b1e/packages/delegate/tests/batchExecution.test.ts#L107-L135 — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1994 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFRROCWTNPSKBQB7ME44SLSELG4DANCNFSM4QW7ZDTA> .