UrlResolver PR https://github.com/CodexCoder21Organization/UrlResolver/pull/770 remote buildtest

Reported (UTC): 2026-07-14 00:30

UrlResolver PR https://github.com/CodexCoder21Organization/UrlResolver/pull/770 remote buildtest replacement run 7395d3a9 produced six false test failures because dependency resolution timed out before test logic ran. The affected tests were testE2eObservableInResultOverP2P, testHostOriginatedInterfaceNestedInContainersIsProjected, testIdleUrlProtocolDrainsInternalThreadsToZero, testLiveProjectionPushFailureReapsSubscription, testMapReturnWithStringValue, and testNatNodeGetsCircuitRelayAddressesAfterRelayRegistration. Four failed resolving workspace-local artifact foundation.url:resolver:0.0.632 and two failed resolving the unified test classpath. Each exhausted all three built-in attempts, with every attempt hitting kompile.coursier.fetchAttemptTimeoutMillis=60000ms and total failure around 180.3s. Concurrent evidence showed buildtest frontend /api/test-results and /api/runs returning HTTP 503 Read timed out, while host files continued advancing; 1090 other tests completed successfully and the PR's GitHub Actions suite passed. Impact: a required remote CI check goes red despite no test assertion executing in the six cases, forcing an expensive full-suite rerun after an earlier run had already been lost to restart recovery. Workaround: classify via host test-events.jsonl/xml_results, wait for terminal collection, then warm the webhook receiver and rerequest only the failed remote check suite. Suggested durable fix: pre-resolve/cache the workspace artifact and unified classpath once per shard/run before scheduling tests, avoid many concurrent per-test Coursier requests against kotlin.directory, and classify exhausted repository fetches as an infrastructure-level retry/rerun rather than individual test failures.

Production verification — 2026-07-17

Status: STILL EXISTS. Live GitHub verification found 1 referenced tracker(s) still open: https://github.com/CodexCoder21Organization/UrlResolver/pull/770. A live ProductionHealth connection also emitted repeated NothingToCompleteException gossip failures, while the stopped HardwareControlFabric daemon log ends with Netty ByteBuf leak reports.

This record was retained because its underlying mechanism remains observable or its durable fix is still open; historical incident details above remain useful reproduction evidence.