Dylan Anthony
Dylan Anthony
Yep, pretty sure this one was a bug in composition from 2.7.5 through 2.7.7, 2.7.8 should work as expected.
Most of the install script is about checking for dependencies and architecture. As a workaround, you could instead download the binary from your internal repo and run it directly (or...
> +1 for `get_line_column` and `get_line_column_range`. The latter could even return `std::ops::Range` (or `Option` of) I like having a struct that encodes the meaning, rather than a tuple (which is...
I've reworked this PR to a completely new proposal that uses _new_ structs for representing both line/col, decoupled from the GraphQL error representation. However, given that we're going to need...
On closer inspection, `TextRange` is [already using exclusive ends](https://docs.rs/text-size/latest/text_size/struct.TextRange.html#method.contains). So I updated to `Range` and made it 0-indexed. I then switch to 1-indexing only when converting to a `GraphQLError`. I...
Also, a quick prototype of these changes in `router`: https://github.com/apollographql/router/pull/5263
We're working on a more general interface for iterating over mapping inputs/outputs. We'll use that for validating bodies instead of the AST directly
At a glance, the issue appears to be that the environment variable was not defined. We need to turn this into a readable error instead of a crash, though.
Something I didn't consider, that we should test, is how this works with unstable router features. We have new config for Router 2.0, I wonder if that shows up as...
@jerelmiller can we keep/merge the supergraph schema one (which pretty much never changes and should be safe) and just take out the Router config stuff for now? I don't think...