Move some GraphQL analysis earlier in the pipeline

[SimonSapin]

The following is now done in QueryAnalysisLayer instead of in the BridgeQueryPlanner service:

• Finding the relevant operation in a GraphQL document based on operationName. Now ParsedDocument in context contains a reference to the parsed apollo_compiler::executable::Operation.
• Scan that operation’s root fields to find which are present: root __typename, schema intropsection, and "normal" non-meta fields. This is a preliminary step to moving introspection out of BridgeQueryPlanner.

As a side-effect observable to clients, the exact response for some error cases like an empty document changes slightly as the error no longer comes from JavaScript code.

Also optional drive-by refactor in the second commit: change spec::Query to contain a single Operation instead of a Vec<Operation>. This removes conditionals and the need to pass operationName around in a lot of places, notably in execution code.

---

Checklist

Complete the checklist (and note appropriate exceptions) before the PR is marked ready-for-review.

• Changes are compatible¹
• Documentation² completed
• Performance impact assessed and acceptable
• Tests added and passing³
  • Unit Tests
  • Integration Tests
  • Manual Tests

Exceptions

Note any exceptions here

Notes

Footnotes

1. It may be appropriate to bring upcoming changes to the attention of other (impacted) groups. Please endeavour to do this before seeking PR approval. The mechanism for doing this will vary considerably, so use your judgement as to how and when to do this. ↩

2. Configuration is an important part of many changes. Where applicable please try to document configuration examples. ↩

3. Tick whichever testing boxes are applicable. If you are adding Manual Tests, please document the manual testing (extensively) in the Exceptions. ↩

[svc-apollo-docs]

✅ Docs Preview Ready

No new or changed pages found.

[github-actions]

@SimonSapin, please consider creating a changeset entry in /.changesets/. These instructions describe the process and tooling.

[goto-bus-stop]

this could/should also use operation.root_fields(), i guess, but doesn't need to be addressed here

[SimonSapin]

root_fields() is used to compute doc.has_root_typename and implicitly descends into inline and named fragments regardless of any directive they might have. This specifically only looks at field, in order to not deal with potential @skip or @include that fragments might have.

… but now I realize that fields can also have @skip or @include, so I’ll add this restriction to the condition here.

[SimonSapin]

I’ll simplify this further in the next PR, having a fast path only for literally {__typename} and letting introspection code deal with any other combination of fragments, aliases, directives, etc.

[goto-bus-stop]

I figured that if we have a fast path for root typename, we might as well have a fast path for something like this?

fragment tyname on Query { __typename }
{
  ... { ...tyname }
}

Though I wonder how useful a __typename fast path is in general with Rust introspection.

[goto-bus-stop]

LGTM, removing Vec is very nice 🥳

[router-perf]

CI performance tests

• connectors-const - Connectors stress test that runs with a constant number of users
• const - Basic stress test that runs with a constant number of users
• demand-control-instrumented - A copy of the step test, but with demand control monitoring and metrics enabled
• demand-control-uninstrumented - A copy of the step test, but with demand control monitoring enabled
• enhanced-signature - Enhanced signature enabled
• events - Stress test for events with a lot of users and deduplication ENABLED
• events_big_cap_high_rate - Stress test for events with a lot of users, deduplication enabled and high rate event with a big queue capacity
• events_big_cap_high_rate_callback - Stress test for events with a lot of users, deduplication enabled and high rate event with a big queue capacity using callback mode
• events_callback - Stress test for events with a lot of users and deduplication ENABLED in callback mode
• events_without_dedup - Stress test for events with a lot of users and deduplication DISABLED
• events_without_dedup_callback - Stress test for events with a lot of users and deduplication DISABLED using callback mode
• extended-reference-mode - Extended reference mode enabled
• large-request - Stress test with a 1 MB request payload
• no-tracing - Basic stress test, no tracing
• reload - Reload test over a long period of time at a constant rate of users
• step-jemalloc-tuning - Clone of the basic stress test for jemalloc tuning
• step-local-metrics - Field stats that are generated from the router rather than FTV1
• step-with-prometheus - A copy of the step test with the Prometheus metrics exporter enabled
• step - Basic stress test that steps up the number of users over time
• xlarge-request - Stress test with 10 MB request payload
• xxlarge-request - Stress test with 100 MB request payload