Lee Byron

Another thing to add here - seems like we agree that our goal is to find a way to move to a better method of representing `defaultValue` while maintaining support...

**EDIT: This graphic is no longer totally accurate:** #3074 (in the stack referenced at the original comment) implements "New Introspection Path for SDL defined Schema" - it's non breaking and...

Some things to note in those flow diagrams above: - The flow you're defining as `toLiteral`, or the a method for converting "Internal input value" to "External input value as...

I totally agree. This is currently broken, or at least poorly named. There's still a lot of work left to clean up this PR, but the next change coming will...

An update on strategy now that this stack is complete: The initial attempt at this PR tried to preserve existing behavior by automatically "uncoercing" default values at schema construction time....

> This is interesting. So graphql-js is moving to specifying default values as "on the wire" / "external" values, not as "application" / "internal" values? That's exactly right. Defining a...

The primary upside that this original RFC required was that a default value should have identical behavior should that same value be provided at runtime. For default values of input...

Added code review responses from @benjie

@graphql/implementers please take one last look at this stack of changes. I'm planning to clean this up and land soon, I'm pretty confident this is the right change. One last...

@benjie thanks for the detailed thoughts, responding to your comment further above: > in particular trying to figure out the practical difference between "complete normally" and "cancel" The practical difference...