From 20418240f78ec7105550ab79f16f13c2bb5acda2 Mon Sep 17 00:00:00 2001 From: Erich Blume Date: Thu, 4 Jun 2026 13:56:27 -0700 Subject: [PATCH] C2(hephd-self-update): impl correct spawn_self_update_loop doc The poller now installs + restarts (not just logs); fix the stale doc and point at service-env-forge-access as the deployment step that makes the apply path operational. --- crates/hephd/src/server.rs | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/crates/hephd/src/server.rs b/crates/hephd/src/server.rs index 16e7494..bf245ba 100644 --- a/crates/hephd/src/server.rs +++ b/crates/hephd/src/server.rs @@ -124,8 +124,13 @@ impl Daemon { } /// If self-update is enabled, spawn its background poller: every - /// `cfg.interval` it checks the forge for a newer release and (for now) logs - /// when one is available. No-op when the mode is off. + /// `cfg.interval` it checks the forge for a newer release and, when one is + /// found, installs it and restarts onto the new binary. No-op when off. + /// + /// Note: the *apply* path runs `cargo install` and exits, which only works + /// when the daemon's service environment can reach the forge over SSH and + /// find cargo — see the `service-env-forge-access` card (the deployment step + /// that makes this operational). pub fn spawn_self_update_loop(&self) { let Some(cfg) = self.ctx.self_update.clone() else { return;