test(server-test): add CLI e2e tests against real PostgreSQL database

[pyramation]

Summary

Adds 5 end-to-end tests that exercise the full CLI pipeline: codegen → transpile → execute as child process against a running PostgreSQL database with a real GraphQL server.

Approach (Approach A):

1. getConnections() spins up a real Postgres DB + GraphQL HTTP server
2. generateCli() and generateOrm() produce TypeScript source files
3. ts.transpileModule strips types without resolving imports (avoids needing all type packages in a temp dir)
4. Appstash context is configured in a temp $HOME pointing at the test server endpoint
5. A child process (spawn, not execFileSync) runs the compiled CLI against the live server
6. stdout is parsed and assertions verify the results

Why spawn instead of execFileSync: The GraphQL server runs in the same Node.js process. Synchronous child execution blocks the event loop, preventing the server from responding — causing a deadlock/timeout.

5 focused corner-case tests:

1. Paginated list with --where (dot-notation) + --fields projection
2. Cursor-based forward pagination (--limit + --after)
3. find-first with --where.name.equalTo
4. Combined --where + --orderBy + --fields
5. Empty result set handling

Package.json changes: Added @0no-co/graphql.web, gql-ast, appstash, inquirerer, and nested-obj as devDependencies in server-test — these are ORM/CLI runtime deps needed by the child process.

Review & Testing Checklist for Human

• Hardcoded buildAnimalsTable() object (lines 100–205): This manually mirrors the simple-seed animals schema. Verify the field names, types, inflection keys (especially orderByType: 'AnimalOrderBy'), and query names match what introspection actually produces. If the codegen's expected Table shape evolves, this will silently generate wrong code.
• Output structure parsing (raw.data?.animals ?? raw): Tests assume the CLI prints the raw ORM execute() result as JSON. Verify this matches the actual handleList/handleFindFirst output format. The fallback chain could mask shape mismatches.
• Hardcoded UUID in Test 3: Asserts node.id === 'a0000001-0000-0000-0000-000000000001' — tied to simple-seed/test-data.sql. Confirm this ID exists and won't change.
• resolveNodePaths() pnpm discovery: Walks require.resolve() paths to find all intermediate node_modules dirs for pnpm strict isolation. Verify this works in CI (which may have a different pnpm version or hoisting config).

Recommended test plan: Run cd graphql/server-test && pnpm test -- --testPathPattern cli-e2e --verbose locally and in CI to confirm all 5 tests pass. Also verify the existing test suite still passes (pnpm test).

Notes

• The pnpm-lock.yaml diff is large but is purely formatting changes (single-line vs multi-line resolution objects) from a lockfile refresh — no dependency version changes.
• The graphile-cache ERROR log during teardown (PostGraphile instance has been released) is pre-existing in other server-test suites and is harmless.

Link to Devin session: https://app.devin.ai/sessions/c92c3a11450342f8875625a60fa1be28
Requested by: @pyramation

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring