Parsing Error for missing empty block

[cornerman]

We are getting a Parsing Error when generating client code for a graphql schema with the sbt plugin.

Our graphql schema contains the following line:

input foobar

And we get the following error:

Parsing Error: Position 5143:1, found ""
caliban.parsing.Parser$.$anonfun$parseQuery$7(Parser.scala:23)
zio.ZIO$.$anonfun$fail$1(ZIO.scala:2898)
zio.internal.FiberContext.evaluateNow(FiberContext.scala:413)
zio.Runtime.unsafeRunWith(Runtime.scala:221)
zio.Runtime.unsafeRunSync(Runtime.scala:84)
zio.Runtime.unsafeRunSync$(Runtime.scala:81)
zio.Runtime$$anon$3.unsafeRunSync(Runtime.scala:288)
zio.Runtime.unsafeRun(Runtime.scala:59)
zio.Runtime.unsafeRun$(Runtime.scala:58)
zio.Runtime$$anon$3.unsafeRun(Runtime.scala:288)
caliban.codegen.CalibanSourceGenerator$.generateSources$1(CalibanSourceGenerator.scala:122)
caliban.codegen.CalibanSourceGenerator$.$anonfun$apply$35(CalibanSourceGenerator.scala:131)
caliban.codegen.CalibanSourceGenerator$.$anonfun$apply$35$adapted(CalibanSourceGenerator.scala:130)
sbt.util.Tracked$.$anonfun$outputChangedW$1(Tracked.scala:128)
caliban.codegen.CalibanSourceGenerator$.apply(CalibanSourceGenerator.scala:144)
caliban.codegen.CalibanPlugin$.$anonfun$calibanScopedSettings$3(CalibanPlugin.scala:40)
scala.Function1.$anonfun$compose$1(Function1.scala:49)
sbt.internal.util.$tilde$greater.$anonfun$$u2219$1(TypeFunctions.scala:62)
sbt.std.Transform$$anon$4.work(Transform.scala:68)
sbt.Execute.$anonfun$submit$2(Execute.scala:282)
sbt.internal.util.ErrorHandling$.wideConvert(ErrorHandling.scala:23)
sbt.Execute.work(Execute.scala:291)
sbt.Execute.$anonfun$submit$1(Execute.scala:282)
sbt.ConcurrentRestrictions$$anon$4.$anonfun$submitValid$1(ConcurrentRestrictions.scala:265)
sbt.CompletionService$$anon$2.call(CompletionService.scala:64)
java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1130)
java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:630)
java.base/java.lang.Thread.run(Thread.java:831)

We can fix it by rewriting our graphql schema to:

input foobar {}

This also happens for enum foobar without braces.

Our graphql schema was generated from hasura with the following command:

npx --package=graphqurl gq http://localhost:8080/v1/graphql --introspect

[ghostdogpr]

That's not valid graphql, is it?

In the spec:

An Input Object type must define one or more input fields.

An Enum type must define one or more unique enum values.

[ghostdogpr]

Related code in the parser:

• Scala 2:

  caliban/core/src/main/scala-2/caliban/parsing/parsers/Parsers.scala

  Line 79 in 55dab61

  def inputObjectTypeDefinition(implicit ev: P[Any]): P[InputObjectTypeDefinition] =

  and

  caliban/core/src/main/scala-2/caliban/parsing/parsers/Parsers.scala

  Line 92 in 55dab61

  def enumTypeDefinition(implicit ev: P[Any]): P[EnumTypeDefinition] =

• Scala 3:

  caliban/core/src/main/scala-3/caliban/parsing/Parser.scala

  Line 361 in 55dab61

  private def inputObjectTypeDefinition(description: Option[String]): P[InputObjectTypeDefinition] =

  and

  caliban/core/src/main/scala-3/caliban/parsing/Parser.scala

  Line 377 in 55dab61

  private def enumTypeDefinition(description: Option[String]): P[EnumTypeDefinition] =

[cornerman]

Good point, I do not know. But that is what we get from hasura :)

Mind you, there are normally two schema with hasura. One for authorized users and one unauthorized users. As of now, we get multiple empty inputs and empty enums in the unauthorized case.

I just opened another issue regarding enums here: #1253.

[ghostdogpr]

According to the spec, enum foo and input foo should be supported during parsing. It's not a valid schema so code generation should fail, but it shouldn't fail during parsing.