'Maximum call stack size exceeded' on large schema

[wimklerkx]

My schema.graphql is rather large, about 2500 lines without comments

Until version 2.0.1 there were no problems with this schema,
but after updating to v2.1.1 the following error occurs:
RangeError: Maximum call stack size exceeded

Reducing the schema to about 250 lines removes the error.

Increasing the node stack size does not help
When increasing the stack size to 8500
--stack-size=8500
another error takes precedence:
Segmentation fault: 11

Is there a limit to how large a schema can be?
Or is there a way to circumvent this problem when handling large schema's?

[jexp]

Is that in neo4j-graphql-js ?

Or the server plugin?

[wimklerkx]

The error occurs at the makeAugmentedSchema({typeDefs}) function,
which is neo4j-graphql-js

[johnymontana]

@wimklerkx Would it be possible for you to share your schema SDL so we can reproduce the error?

[rdimicheleb]

I had the same problem, solved disabling auto-generating querys and mutations:

module.exports = augmentSchema(executableSchema, {
    query: false, 
    mutation: false
  });

[wimklerkx]

@rdimicheleb That would defeat the purpose of the makeAugmentedSchema function
That said, indeed, I'm now headed to writing my own query and mutation resolvers.

[rdimicheleb]

@wimklerkx you are right, you need to write your own mutations, but I can still query the graph without write every single query in the resolver. I just need to define in the resolver the principals querys using the neo4jgraphql function, example:

resolver:

.
.
.
Query: {
        company: (object, params, context, resolveInfo) => {
            return neo4jgraphql(object, params, context, resolveInfo, process.env.NODE_ENV === 'development')
        },
.
.
.

schema:

type Company{
    name: String
    employees: [Person] @relation(name: "EMPLOYEE_OF", direction:"IN")
}
type Query {
    company (id: ID): Company
}

And I don't need to write any custom resolver for query employees or persons.

I know this issue should be fixed, but the above allowed me to continue working without crash the server.

I Hope you find a solution soon.

[wimklerkx]

@rdimicheleb Thanks for the elaboration, will do :)

[wimklerkx]

@johnymontana Here is the large schema causing the problem
schema.txt

[michaeldgraham]

Hey everyone, this problem was caused when I introduced the use of lodash/cloneDeep while refactoring some of the augmentation code a while back. It was having problems when cloning a type in a large schema. Sorry I didn't catch it before.

Thanks for sharing your schema, @wimklerkx. As of #188, if you're using a larger schema, you should again be able to run the augmentation process on it. If you run into any further problems, feel free to ping me here.

[johnymontana]

The fix for this should be available now in v2.3.0 - please test and let us know if it works.

[wimklerkx]

Yes, the call stack size error is now gone
Thanks :)