Repository navigation
ci: harden integration job — fix PGDATABASE, step order, negative control - #199
Closed
amar-python wants to merge 1 commit into
Closed
amar-python wants to merge 1 commit into
amar-python wants to merge 1 commit into
Conversation
… control Three latent problems survive on main even though CI is green: 1. Job-level PGDATABASE was "postgres" — psql commands without an explicit -d flag connected to the wrong database. Changed to "te_mgmt_dev". 2. The eval step ran before deploy_all.sh, relying on Tier E's internal _deploy_env() to paper over the ordering. Moved deploy before evals so the full test suite also sees the deployed schemas. 3. API_ALLOW_DESTRUCTIVE was missing at job level, so DELETE-endpoint tests were silently skipped. Added it. Additionally, a Tier E negative control (02_empty_schema_is_detected) is introduced: it creates a temporary empty schema, asserts _column_signature returns 0 tables, and drops the schema — preventing the false green that masked the original CI failure. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VBrxqChRJtxdvSpFhiUWUy
Owner
Author
|
Closing — CI on main is already green. The changes here (job-level Generated by Claude Code |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
PGDATABASE: changed frompostgrestote_mgmt_devso psql commands without an explicit-dflag connect to the correct databaseDeploy all environmentsbeforeEvals — Tiers P, I, S, Eso the full test suite sees deployed schemas (previously relied on Tier E's internal_deploy_env())API_ALLOW_DESTRUCTIVEat job level so DELETE-endpoint tests are no longer silently skipped02_empty_schema_is_detected): creates a temporary empty schema, asserts_column_signaturereturns 0 tables, drops it — prevents false greens from empty/missing schemasWhy these are latent
CI on main is currently green because Tier E deploys internally and the test suite has a per-step
PGDATABASEoverride. But the bugs remain: any new step that relies on the job-levelPGDATABASEor assumes deploy ran first will fail silently.Files changed
.github/workflows/quality-gate.ymlevals/runner.py02_empty_schema_is_detectedhandlerevals/expected/tier_e/02_empty_schema_is_detected.jsonevals/datasets/tier_e/02_empty_schema_is_detected/NOTES.txtTest plan
02_empty_schema_is_detectedruns and passes01_envs_have_identical_structurestill passesGenerated by Claude Code