From 1187d029eae7ebbc7a4a99bfa89f72512411d7e2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ralph=20K=C3=BCpper?= Date: Mon, 31 Aug 2026 14:52:25 +0200 Subject: [PATCH] test(parity): drop the stale test_stress_promises known-failure entry parity (10) fails the stale-entry check: test_stress_promises is listed as a known Linux failure but now passes there. Per #7582 a triage entry that outlives its bug is a permanent suppression of that test's next regression, so the gate requires it removed. Recorded uncertainty: #8841 is still open and no targeted fix for this SIGSEGV appears in main's history, so this may be a bug that is currently masked rather than fixed. The entry's own note claims it reproduced 5/5 on the revisions it was filed against. That cannot be re-verified from macOS, and the gate's verdict is the only Linux evidence available. If it returns, the right response is a fresh entry against a fresh issue with current evidence -- not restoring this one, whose reasoning is two weeks and many merges stale. --- changelog.d/9286-stale-known-failure.md | 4 ++++ test-parity/known_failures.json | 9 --------- 2 files changed, 4 insertions(+), 9 deletions(-) create mode 100644 changelog.d/9286-stale-known-failure.md diff --git a/changelog.d/9286-stale-known-failure.md b/changelog.d/9286-stale-known-failure.md new file mode 100644 index 0000000000..1dba8d04bf --- /dev/null +++ b/changelog.d/9286-stale-known-failure.md @@ -0,0 +1,4 @@ +The parity triage list no longer suppresses `test_stress_promises`, which now +passes on Linux. A `known_failures.json` entry that outlives its bug silently +absorbs that test's next regression, so the gate requires stale entries to be +removed. diff --git a/test-parity/known_failures.json b/test-parity/known_failures.json index 202e64d354..dea95fa392 100644 --- a/test-parity/known_failures.json +++ b/test-parity/known_failures.json @@ -922,15 +922,6 @@ "linux" ] }, - "test_stress_promises": { - "issue": "8841", - "added": "2026-08-25", - "category": "bug-open", - "reason": "Reproduces 5/5 on both pre-#8835 r13 and r14: Perry dies with SIGSEGV (exit 139) after six output lines while Node exits 0. Tracked in #8841.", - "platforms": [ - "linux" - ] - }, "test_take_screenshot": { "issue": "8271", "added": "2026-08-17",