Skip to content

perf(rules): pin SQLClean 0.4.3 for cheaper unused-CTE facts - #1296

Merged
kvlonge merged 3 commits into
mainfrom
perf/pin-sqlclean-0-4-3
Nov 2, 2026
Merged

kvlonge merged 3 commits into
mainfrom
perf/pin-sqlclean-0-4-3

Conversation

@kvlonge

@kvlonge kvlonge commented Nov 2, 2026

Copy link
Copy Markdown
Contributor

Why

SQLClean 0.4.3 makes SQLCLEAN005 (unused CTE) facts cheaper: it no longer deep-clones every query with a WITH clause. This is a small part of the default-Rules cold-compile cost and does not recover the regression tracked in #1289.

Changes

  • Pin sqlclean to the v0.4.3 release commit cf79656. The lockfile changes only the sqlclean entry; its polyglot revision is unchanged.

Verification

  • SQLClean 0.4.3 passed its full Verify and published to PyPI; the installed sqlclean 0.4.3 reports the expected SQLCLEAN005 findings.
  • Replaying the 2,249-request lint batch from the 3,000-model inspection project gave byte-identical lint output before and after the change.
  • Full SQLBuild Verify on this pull request.

@kvlonge
kvlonge merged commit 1ba0a3d into main Nov 2, 2026
44 checks passed
@kvlonge
kvlonge deleted the perf/pin-sqlclean-0-4-3 branch November 2, 2026 08:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant