Stabilize protocol schema fixture generation (#13886)

## What changed
- TypeScript schema fixture generation now goes through in-memory tree
helpers rather than a heavier on-disk generation path.
- The comparison logic normalizes generated banner and path differences
that are not semantically relevant to the exported schema.
- TypeScript and JSON fixture coverage are split into separate tests,
and the expensive schema-export tests are serialized in `nextest`.

## Why this fixes the flake
- The original fixture coverage mixed several heavy codegen paths into
one monolithic test and then compared generated output that included
incidental banner/path differences.
- On Windows CI, that combination created both runtime pressure and
output variance unrelated to the schema shapes we actually care about.
- Splitting the coverage isolates failures by format, in-memory
generation reduces filesystem churn, normalization strips generator
noise, and serializing the heavy tests removes parallel resource
contention.

## Scope
- Production helper change plus test changes.

codex-rs/app-server-protocol/src/protocol/common.rs

@@ -171,6 +171,12 @@ macro_rules! client_request_definitions {
             Ok(())
         }
 
+        pub(crate) fn visit_client_response_types(v: &mut impl ::ts_rs::TypeVisitor) {
+            $(
+                v.visit::<$response>();
+            )*
+        }
+
         #[allow(clippy::vec_init_then_push)]
         pub fn export_client_response_schemas(
             out_dir: &::std::path::Path,
@@ -549,6 +555,12 @@ macro_rules! server_request_definitions {
             Ok(())
         }
 
+        pub(crate) fn visit_server_response_types(v: &mut impl ::ts_rs::TypeVisitor) {
+            $(
+                v.visit::<$response>();
+            )*
+        }
+
         #[allow(clippy::vec_init_then_push)]
         pub fn export_server_response_schemas(
             out_dir: &Path,