Skip to content

GFQL deletion: remove obsolete M1 differential scaffold legacy-baseline tests #1472

Description

@lmeyerov

Summary

graphistry/tests/compute/gfql/cypher/test_m1_differential_scaffold.py still carries an M1-era legacy-vs-candidate differential scaffold. The strict binder/runtime rollout has since landed, so this test harness should be audited for deletion or replacement by direct conformance tests.

Current evidence

On current origin/master:

  • test_m1_differential_scaffold.py defines _run_legacy() and compares binder-prepass/candidate behavior against legacy behavior.
  • Tests include test_diff_corpus_legacy_baseline and test_diff_corpus_legacy_vs_candidate.
  • Other strict binder and runtime parity tests now pin the active paths directly.

Goal

Delete obsolete legacy-baseline scaffolding after confirming its coverage is duplicated by current binder/runtime tests, or replace any unique cases with direct assertions and then delete the scaffold.

Scope

  • Inventory each _DiffCase and map it to current direct tests, or promote missing cases into focused tests.
  • Remove _run_legacy() and candidate-vs-legacy comparison scaffolding if no longer needed.
  • Keep only direct behavior tests that assert current intended semantics.
  • Record test LOC removed.

Out of scope

  • Changing runtime/compiler semantics.
  • Reintroducing broad differential framework work.
  • Public API deprecation removal.

Acceptance

  • Obsolete M1 differential scaffold is deleted or materially reduced.
  • Any unique cases are preserved as direct tests.
  • Focused Cypher binder/lowering tests pass.
  • Full relevant GFQL CI passes.

Related

Activity

  1. lmeyerov commented on May 17, 2026

    @lmeyerov
    ContributorAuthor

    #1472 implementation status:

    • PR: Delete obsolete M1 differential scaffold tests #1491
    • Deleted obsolete M1 legacy-vs-candidate differential scaffold test file.
    • Preserved the unique independent OPTIONAL MATCH row/null behavior as direct lowering coverage across both optional-clause orders.
    • No runtime/compiler behavior changed.
    • LOC delta: CHANGELOG.md +1/-0, test_lowering.py +41/-0, test_m1_differential_scaffold.py +0/-282; net +42/-282, overall -240 LOC.
    • Local validation: focused preservation tests passed; binder/lowering non-cuDF slice passed (1022 passed); ./bin/ruff.sh, ./bin/typecheck.sh, and git diff --check passed.
    • Review skill converged after two no-finding waves.
    • CI is green on current PR head. GPU/RAPIDS was not run because this is a test-only deletion/preservation slice with no cuDF/GPU runtime changes; local full cuDF collection is blocked on the non-GPU host by cudaErrorNoDevice.
  2. lmeyerov commented on May 17, 2026

    @lmeyerov
    ContributorAuthor

    #1472 merged.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions