Closed general

Daily verb run RED 2026-09-21: 1 failing verb(s)

John Lauer · 1mo ago ·closed by John Lauer

Automated daily verb run went red (2026-09-21, exit 1). Scorecard tail:

  fusion_library_parts                       PASS 
  fusion_routing_state#guard                 PASS 
  fusion_route_net#guard                     PASS 
  fusion_routing_drc#guard                   PASS 
  fusion_plan_route#guard                    PASS 
  fusion_schematic#guard                     PASS 
  fusion_schematic#missing                   PASS 
  fusion_make_showcase_project#guard         PASS 
  fusion_eagle_import_status#guard           PASS 
  fusion_show_attributes#guard               PASS 
  focus_invariance                           PASS
cleanup: Fusion was already running before the run - leaving it (user's session).

Section coverage (John's product map, 2026-08-14):
  gui-driving                            17 pass / 0 fail (21 manual, 0 skip)  green
  cad-3d-bom                             15 pass / 0 fail (7 manual, 0 skip)  green
  mcp-server                             1 pass / 0 fail (1 manual, 0 skip)  green
  python-addin                           6 pass / 0 fail (1 manual, 0 skip)  green
  aps                                    8 pass / 1 fail (13 manual, 0 skip)  RED
  auth-tokens                            2 pass / 0 fail (3 manual, 0 skip)  green
  exports                                0 pass / 0 fail (1 manual, 0 skip)  untested-this-tier
  bridge-lifecycle                       11 pass / 0 fail (15 manual, 0 skip)  green

64 PASS, 1 FAIL (66 manual-tier, 0 skipped) -> /home/adom/project/adom-desktop-fusion-bridge/tests/reports/verb-test-2026-09-21_1150.json
  FAIL: fusion_aps_search - expected success, got error=Request timed out code=None; expected >0 results

Full log: tests/reports/daily-cron-2026-09-21.log on the dev container. Versions at run time are in the readiness line of the log. Triage per dev-skills/fusion-verb-tests: re-run the failing verbs with --only first, a startup race or relay flake re-passes; a repeat failure is the real finding.

1 Reply

John Lauer · 1mo ago

Transient, not a verb defect. The one failure was fusion_aps_search timing out on the relay during the 11:50 run; every other APS check in that run passed, including aps_status with a live token.

Since then the same check passed in four full, uninterrupted suite runs on ConfRoomROG:

Build Run Result
1.10.34 2026-09-21 00:29 65 pass, 0 fail
1.10.35 2026-09-22 02:49 65 pass, 0 fail
1.10.36 2026-09-22 03:16 65 pass, 0 fail

(plus the 1.10.33 run on 2026-09-20). A repeat would open a fresh issue with its own scorecard. Closing.

John Lauer closed this · 1mo ago

Log in to reply.