Graeme Coupar

bors r+

I assume the fix for this is to be smarter about [this set here](https://github.com/obmarg/cynic/blob/master/cynic-querygen/src/query_parsing/inputs.rs#L95) - it could potentially be a HashMap or similar, with a lookup to avoid the recursion.

Hi @jxnu-liguobin - what two categories were you thinking? If I remember right the main motivation behind this issue is when we're facing a scalar decoding error - custom scalars...

Decoding is now done with serde, so this should now be easily do-able with the https://github.com/dtolnay/path-to-error crate. Not sure about how it should be exposed to users. Perhaps the http...

Ah yeah, was wondering when GraphQLs three state nullables would crop up and bite me. Guess today's the day :) So I can see two solutions off the top of...

Ok, so after some discussion on discord there are a few steps to resolve this: 1. [x] Allow users to omit optional fields from their `InputObject`s without complaining. (Implemented in...

Hi @idmyn - are these arguments nested inside an `InputObject` or are they specifically arguments on a field? If the former, you can do ``` #[cynic(skip_serializing_if="Option::is_none")] whatever_your_field_is_called: Option ``` the...

> On second thought, the arguments will be nested inside an `InputObject` in this case (so we'll try that solution you suggest - thanks!) Cool, let me know if you...

Hey @tim-maddison-rnf - that does sound quite annoying, sorry for the hassle. Would be happy to accept a PR along these lines. I had kind of hoped that users could...

#132 adds the `cynic::Errors` type that wraps up several `syn::Error`s into one, and supports emitting all of them. That seems a better solution than the one proposed in this issue,...