Rename classes to avoid name clashes

[xirc]

This PR solves the issue #543.

#543 (comment)

How about renaming those internal classes to __???Value directly?

As suggested by @ghostdogpr, I've changed the internal class names.

• NullValue to __NullValue
• NumberValue to __NumberValue
• EnumValue to __EnumValue
• StringValue to __StringValue
• BooleanValue to __BooleanValue
• ListValue to __ListValue
• ObjectValue to __ObjectValue

[ghostdogpr]

Nice! There's one more thing to do: modifying the code generation to use those new names. The relevant code is there: https://github.com/ghostdogpr/caliban/blob/master/tools/src/main/scala/caliban/tools/ClientWriter.scala

Let me know if you need more explanations.

[xirc]

@ghostdogpr Thanks for the review! I will try it tomorrow morning (JST+9:00).
Because the CI is success, I guess that the code generation is not tested yet, right?
I think it's better to test the code generation.
So, I will also try to test it if there is no technical problem.

[ghostdogpr]

It is tested (here: https://github.com/ghostdogpr/caliban/blob/master/tools/src/test/scala/caliban/tools/ClientWriterSpec.scala) but the output is a string (generated code), and we just verify that this is the expected code, we don't actually verify that it compiles. That's why the tests passed. If you have a better idea for testing this, feel free to add it 👍

[xirc]

Thanks! I got it.
In this PR, I will focus on changing the code generation and its existing test because I think writing a test verifying that generated code compiles is out of topic in this PR and makes the PR huge.

[xirc]

Is it better to rename Value to __Value?
I think the type may have the potential to collide with a user type name.

[ghostdogpr]

Yeah, I agree.

[ghostdogpr]

Perfect! Thanks for the contribution!