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
refactor(storage): update blob sweeping mechanism
Steven Enamakel2c453e1bcd46
diff --git a/crates/openhuman-core/src/storage/blob_sweep.rs b/crates/openhuman-core/src/storage/blob_sweep.rsindex fcfe87e762..3e3a2f141a 100644--- a/crates/openhuman-core/src/storage/blob_sweep.rs+++ b/crates/openhuman-core/src/storage/blob_sweep.rs@@ -2,8 +2,8 @@//!//! A blob written with a lifetime reads as absent once it lapses, but its//! bytes stay in the backend until something sweeps them-//! (`StorageBackend::sweep_expired_blobs`, the trait-level hook the-//! tinystoragedrivers drivers implement). The host owns the schedule: this+//! (`BlobStore::sweep_expired`, the primitive hook the+//! tinystoragedrivers blob engines implement). The host owns the schedule: this//! module visits every storage scope background work covers//! ([`super::agents::for_each_scope`]) on a fixed tick and asks the installed//! backend to reclaim that scope's lapsed blobs.@@ -42,7 +42,18 @@ pub(crate) async fn sweep_scope_within(scope: &Scope,limit: Duration,) -> u64 {- let outcome = match tokio::time::timeout(limit, backend.sweep_expired_blobs(scope)).await {+ let outcome = match backend.for_scope(scope) {+ Ok(storage) => tokio::time::timeout(limit, storage.blobs().sweep_expired()).await,+ Err(error) => {+ tracing::warn!(+ scope = scope.as_str(),+ %error,+ "[storage::blob_sweep] could not open scope; retrying on the next tick"+ );+ return 0;+ }+ };+ let outcome = match outcome {Ok(outcome) => outcome,Err(_) => {tracing::warn!(