feat(ir): RETURN * projects all in-scope variables (#598)

[DecisionNerd]

Advances #598 (read-path). The last engine piece for the With1 cluster.

Summary

MATCH (a)-->(b) RETURN * now returns one column per in-scope named variable instead of erroring (variable * used before it was introduced). * parses as a Var named "*"; the binder now expands it to the current named vars from scope (sorted → deterministic plan), each lowered as the normal per-var projection — so node vars materialize as whole node values (#785).

Combined with WITH whole-node forwarding (#814 slice 2) + MATCH-after-WITH, the With1 shape MATCH (a) WITH a MATCH (a)-->(b) RETURN * now works end-to-end.

Scope

• Anonymous pattern elements (no user name) aren't in scope → not returned.
• Named relationship/path vars would expand here but still fail downstream (rel/path values unsupported) — not silently dropped.

Validation

• New e2e: return_star_projects_all_in_scope_vars (columns a+b, node values), return_star_after_with_forwarded_node (the With1 MATCH-after-WITH shape) — 87 pass.
• IR/rel goldens unchanged; cargo clippy --workspace -- -D warnings + fmt clean.

With this on main, #889 slice 2 (render + un-skip) finally has a runnable node-returning corpus cluster to land against.

🤖 Generated with Claude Code

Note

Expand RETURN * to project all in-scope named variables in the IR binder

• Adds expand_return_wildcard in binder.rs to detect a Var named "*" in return items and replace it with one item per in-scope named variable from BinderState.vars, sorted for deterministic ordering.
• Binder.lower_return now calls this expansion before aggregate detection and lowering, so wildcard projections resolve correctly instead of propagating a raw "*" variable.
• Anonymous elements are excluded from expansion; named relationship and path variables are included but subject to downstream support limitations.
• Adds two e2e tests covering basic RETURN * after MATCH and RETURN * after WITH node forwarding.

^{Macroscope summarized 39a7a3c.}

Summary by CodeRabbit

• Tests

  • Added end-to-end baseline tests for RETURN * behavior, validating proper projection of all in-scope variables as individual columns and correct handling in WITH clause scenarios with variable re-matching.

• Bug Fixes

  • Fixed RETURN * wildcard expansion to systematically include all named variables currently in scope in query results, sorted alphabetically.

[coderabbitai]

Walkthrough

Implements RETURN * wildcard support in the binder: a new expand_return_wildcard helper replaces wildcard ReturnItems with one item per in-scope variable from BinderState.vars, sorted by name. lower_return calls this helper before aggregate detection and Project lowering. Two e2e baseline tests validate the behavior for plain MATCH and WITH + re-MATCH pipelines.

Changes

RETURN * Wildcard Expansion

Layer / File(s)	Summary
Wildcard expansion in lower_return crates/gf-ir/src/binder.rs	expand_return_wildcard replaces RETURN * AST items with one ReturnItem per in-scope variable from BinderState.vars, emitted in sorted order. lower_return calls this helper first, then runs aggregate detection and Project lowering on the expanded list.
E2E baseline tests for RETURN * crates/gf-api/tests/e2e_baseline.rs	Two new tests: one verifies RETURN * after MATCH binding a and b produces all in-scope variable columns; the second verifies RETURN * after a WITH + re-MATCH pipeline produces both the forwarded and re-matched variable columns with correct row counts.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• DecisionNerd/graphforge#890: Modifies Binder::lower_return in the same file to handle whole-node materialization via _node_struct, directly adjacent to the return-lowering code path extended here.
• DecisionNerd/graphforge#894: Reworks WITH clause scoping in binder.rs, which determines the in-scope variables that expand_return_wildcard reads from BinderState.vars after a WITH + re-MATCH pipeline.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately describes the main feature: RETURN * now projects all in-scope variables instead of erroring.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description check	✅ Passed	The pull request description is comprehensive, well-structured, and includes all key information: a clear summary of the change, the problem being solved, scope limitations, validation details, and context about blocking/unblocking other work.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/598-return-star

---

_{Comment @coderabbitai help to get the list of available commands.}