Skip to content

v1.0.24 - YIELD pipeline fixes

Choose a tag to compare

View all tags
@orneryd orneryd released this 19 Mar 18:11
· 603 commits to main since this release
v1.0.24
96972e0

[v1.0.24] - 2026-03-19

Changed

  • CALL/YIELD pipeline execution:
    • generalized post-YIELD clause handling so MATCH, WITH, RETURN, ORDER BY, SKIP, and LIMIT pipelines execute consistently after procedure calls
    • removed brittle fixed-shape assumptions and aligned handling with broader valid clause combinations.
  • Server test fixture reuse:
    • refactored high-frequency server test paths to share fixtures through grouped subtests where isolation was not required
    • reduced repeated full server/bootstrap setup in hot test paths.
  • Test startup behavior:
    • disabled external embedding initialization in shared server test setup to avoid unnecessary async retry work in generic tests.

Fixed

  • CALL clause boundary parsing:
    • fixed YIELD parsing to respect CALL-clause boundaries instead of matching by raw string position
    • resolved cases where downstream clauses were incorrectly parsed or skipped in multi-clause statements.
  • N-column YIELD projection correctness:
    • fixed projection behavior so YIELD supports variable-width output columns without shape-specific fallbacks
    • resolved incorrect/empty result sets in valid procedure-followed query pipelines.
  • Concurrent metadata snapshot race in multi-database management:
    • made database metadata snapshots lock-safe when listing and fetching database info
    • prevents race conditions between storage-size cache initialization and metadata reads under concurrent access.

Tests

  • Added regression coverage for:
    • procedure-call pipelines with trailing clause permutations
    • boundary-aware YIELD parsing in multi-clause statements
    • multi-column YIELD projection with downstream MATCH/WITH/RETURN flows.
  • Added a dedicated race regression test for concurrent storage-size initialization vs database metadata listing.
  • Consolidated selected multi-database and server branch tests into shared-fixture suites while preserving isolated setup for lifecycle-sensitive cases.

Technical Details

  • Range covered: v1.0.23..HEAD
  • Commits in range: 3 (non-merge)
  • Repository delta: 5 files changed, +806 / -108 lines
  • Primary focus areas: procedure pipeline correctness, parser robustness, and generalized clause compatibility for Cypher execution.
Assets 6
Loading