From c35f259a4a48d13b48e0da284686d7545ee4c542 Mon Sep 17 00:00:00 2001 From: Andre van Zuydam Date: Sat, 26 Sep 2026 19:58:23 +0200 Subject: [PATCH] ci(test): trial un-ignore of the http-hardening module (tina4-python#170) The built-in-server deadlock only ever manifested on the 2-vCPU GitHub runner; GitHub has since moved ubuntu-latest to 4 vCPU, and 400+ repro attempts (lab 1-core+load, real runner isolated, real runner 2-core recreated) never wedged. Remove the --ignore so the full suite runs the hardening file again, behind the --timeout=180 signal backstop (a wedge fails loudly in minutes, never a silent hang). If this stays green across several runs, the mitigation is retired and #170 closes as an environmental resolution. Signed-off-by: Andre van Zuydam Co-Authored-By: Claude Opus 4.8 Co-Authored-By: Tina4 <82961293+tina4stack@users.noreply.github.com> --- .github/workflows/test.yml | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml index caebc9f4..3daf3e25 100644 --- a/.github/workflows/test.yml +++ b/.github/workflows/test.yml @@ -406,12 +406,15 @@ jobs: # tina4stack/tina4-python#170. The rest of the suite runs as 4 SEQUENTIAL # pytest-split groups (fresh process each) so a separate CI-only resource # leak cannot accumulate to a wedge. --timeout is a loud backstop. + # TRIAL (tina4-python#170): the http-hardening file is NO LONGER --ignored -- + # runners moved 2->4 vCPU so the original deadlock trigger is gone; if this + # stays green across several runs the mitigation is retired. run: | uv pip install pytest-timeout pytest-split overall=0 for g in 1 2 3 4; do echo "::group::pytest split group $g/4" - uv run pytest tests/ --ignore=tests/test_http_hardening_contract.py \ + uv run pytest tests/ \ --splits 4 --group "$g" -v --timeout=180 --timeout-method=signal -rfE || overall=1 echo "::endgroup::" done