fix(ci): stabilize regtest Postgres tests and regtest workflows

[b-l-u-e]

10‑minute job timeout – Jobs were cancelled mid‑pytest (operation canceled), so often never saw a real pass/fail. so i raise timeout-minutes so the suite can finish.
After the timeout was raised – Many runs completed and showed ~3–4 minutes for tests/wallet / tests/mint in some matrix cells saw tests. So duration was under the old 10‑minute cap in those cases; the cap was still wrong for slower/heavier rows or when setup + tests add up.
then noticed one matrix run hit DeadlockDetectedError on DROP SCHEMA public CASCADE in the ledger fixture

Improves reliability of mint tests against PostgreSQL in CI, deduplicates regtest setup via a shared composite action, and hardens Postgres orchestration (service containers, digest-pinned image, consistent cleanup). Raises job timeouts so slow Lightning/regtest runs are not killed mid-suite, and fixes test assertions and DB cleanup so Postgres matrix jobs pass consistently.

Tests (tests/conftest.py)

• Replace per-test DROP SCHEMA public CASCADE with a TRUNCATE-based cleanup that keeps the schema and avoids heavy locking/deadlock issues.
• Build table names with quote_ident (no % in SQL strings) so Postgres queries stay valid after Connection.rewrite_query doubles % for SQLAlchemy.
• Do not truncate dbversions, so migration version rows stay aligned with the existing schema; avoids ADD COLUMN … already exists when migrations would otherwise re-run after truncate.
• Remove stray print("teardown") from the ledger fixture.

Tests (tests/mint/test_mint_melt.py)

• Extend assert_err with optional contains=True for substring checks; default remains exact match.
• Fix test_mint_pay_with_duplicate_checking_id: await assert_err(...) and use contains=True so wrapped wallet errors (e.g. mint error prefix) still match.

CI

• Composite action .github/actions/regtest-run: clone/start regtest; for Postgres, create a unique DB per run/backend (MINT_TEST_DATABASE), run make test-mint / test-wallet, pg_terminate_backend + DROP DATABASE cleanup with retries, dump Postgres logs on failure.
• Postgres in CI: regtest-mint.yml and regtest-wallet.yml use GitHub service containers (health-checked) instead of ad-hoc docker run, with use-github-postgres-service so the app talks to localhost:5432 and admin ops use psql + shared PGADMINURL. The docker run path remains for other callers that keep the default.
• Reproducibility: workflow-level REGTEST_POSTGRES_IMAGE is digest-pinned; pass postgres-image into the composite so services.postgres.image, fallback docker run, and log capture all reference the same image ref.
• regtest-mint.yml / regtest-wallet.yml: timeout-minutes: 120, upload artifacts on failure, Codecov with if: always() where applicable.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 75.05%. Comparing base (a1ff72c) to head (aae9635).
⚠️ Report is 2 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #966      +/-   ##
==========================================
+ Coverage   74.90%   75.05%   +0.14%     
==========================================
  Files          99       99              
  Lines       11685    11714      +29     
==========================================
+ Hits         8753     8792      +39     
+ Misses       2932     2922      -10     

Flag	Coverage Δ
mint-CLNRestWallet	55.53% <ø> (?)
mint-CoreLightningRestWallet	55.18% <ø> (?)
mint-LNbitsWallet	55.39% <ø> (?)
mint-LndRPCWallet	55.06% <ø> (?)
mint-LndRestWallet	55.45% <ø> (?)
wallet-CLNRestWallet	49.11% <ø> (?)
wallet-CoreLightningRestWallet	48.78% <ø> (?)
wallet-LNbitsWallet	48.63% <ø> (?)
wallet-LndRPCWallet	49.14% <ø> (?)
wallet-LndRestWallet	49.25% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.