From c5c9908a2e2f468a69421bb9129a1caeeb623d71 Mon Sep 17 00:00:00 2001 From: JF Date: Tue, 20 Oct 2026 11:29:33 -0400 Subject: [PATCH] ci(canary): give the doctor backend-source step a 120 s probe timeout (#968) The step asserts where CodeLLDB resolves from; a rust row that timed out carries no backend.source to assert on. doctor is the first thing to run each toolchain on a fresh runner, and on windows-latest the same published 0.25.1 took 5-6 s for this step on 2026-09-29, 11-15 s on 09-30, 12-27 s on 10-13 and 36-37 s on 10-20, when rust's probe passed 30 s. The doctor fix in #968's other PR reaches this job only with the next release. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01Y3ULnoVSoCZ5t2Y8d98Ctv --- .github/workflows/canary.yml | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/.github/workflows/canary.yml b/.github/workflows/canary.yml index 6dc26237..f04d8d40 100644 --- a/.github/workflows/canary.yml +++ b/.github/workflows/canary.yml @@ -174,8 +174,14 @@ jobs: else EXPECT="vendored" fi + # The timeout is generous on purpose (#968): this step asserts where + # CodeLLDB resolves from, and a timed-out rust row carries no + # backend.source to assert on. doctor is the first thing to run each + # toolchain on a fresh runner, and on Windows the same published + # 0.25.1 took 5 s here on 2026-09-29 and 37 s on 2026-10-20, when + # rust timed out at 30 s. set +e - node "$CLI" doctor rust --json --timeout 30000 \ + node "$CLI" doctor rust --json --timeout 120000 \ > "$TRANSCRIPTS/doctor.json" 2> "$TRANSCRIPTS/doctor.err" DOCTOR_EXIT=$? set -e