Jesse Rosenberger

+1 for Rust!

Is it worth enumerating what I believe to be a third case: * graphql that contains **just the schema** ? In my experience, GraphQL that contains the schema and an...

> ```rust > is there a world where you'd want to start a compiler > with just the executable definition?? > ``` Perhaps in a case where you write an...

Will `compiler.executable_definition` present any problems for the case where you want to build up a context with a _set_ of operations where, for example, each operation itself is an anonymous/unnamed...

Cool, I think that could work! I'm interested in any other thoughts, too. And yes, you've caught onto my important nuance that I could have been even more clear about:...

Thanks for opening this! It definitely seems reasonable that your schema would not need to be needlessly augmented with `@cacheControl` directives if `cacheControl` was explicitly set to `false`. Though we'll...

We've done a meta-analysis of the schemas we track at Apollo, being careful to respect the privacy of schema owners, and we've found that: - 6.5% graphs contain object fields...

I also had to rebuild the cache manually for a newly-created file to show up.

I can confirm that our remote API doesn't **yet** support contract variants! Once that lands, this will just _start working_ right? With no changes on Rover?

I think we should carefully define the desired behavior for how subgraph schema files are merged. Is merely concatenation acceptable? If so, then there is a short-term, already supported answer...