The fastest, cheapest, most efficient open-source agent harness. Run more than 500 agents on a $10 VPS.
Rust agent-orchestrationai-agentsai-assistantdesktopllmlocal-firstmcppersonal-aiprivacyrustsecond-braintauri
test(flows): increase concurrency for resume checkpoint tests
Steven Enamakel237cc8ab14a7
diff --git a/crates/openhuman-core/src/flows/ops_resume_checkpoint_tests.rs b/crates/openhuman-core/src/flows/ops_resume_checkpoint_tests.rsindex 4b55c11512..158f53c40d 100644--- a/crates/openhuman-core/src/flows/ops_resume_checkpoint_tests.rs+++ b/crates/openhuman-core/src/flows/ops_resume_checkpoint_tests.rs@@ -301,7 +301,15 @@ async fn resume_uses_the_captured_profile_scope_for_identical_run_ids() {/// upgrading mid-park can never strand an otherwise-valid in-flight approval/// — even if the flow's graph was *also* edited in the meantime, since there/// is nothing recorded to compare it against.-#[tokio::test]+///+/// Multi-thread on purpose: this test calls the synchronous store bridge+/// (`update_parked_run_document_for_test`, `store::update_flow_graph`) between+/// async calls. The bridge blocks the calling thread while the TinyFlows+/// module calls back into the host's document port, and a current-thread+/// runtime has no other thread left to make progress on the work that call+/// depends on (the resumed run's poller task lives on that runtime). The+/// production runtime is multi-thread, where the bridge uses `block_in_place`.+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]async fn flows_resume_allows_a_legacy_row_with_null_graph_hash() {let tmp = TempDir::new().unwrap();let config = test_config(&tmp);