diag: log merge-reap values to debug disappearance bug

Temporary diagnostic added to reap_stale_merge_jobs to surface the t,
current_boot, and decoded values being compared on every reap pass.
Will revert once the disappearance bug is understood.

server/src/agents/pool/pipeline/merge/runner.rs

@@ -39,10 +39,23 @@ impl AgentPool {
                 if job.status != "running" {
                     continue;
                 }
-                let stale = match decode_server_start_time(job.error.as_deref()) {
+                let decoded = decode_server_start_time(job.error.as_deref());
+                let stale = match decoded {
                     Some(t) => t < current_boot,
                     None => true, // Legacy (pid-encoded) or malformed: stale
                 };
+                // DIAGNOSTIC: log the values being compared so we can see why
+                // fresh jobs are being marked stale. Remove once the disappearance
+                // bug is understood.
+                slog!(
+                    "[merge-reap] job='{}' status='{}' error={:?} decoded={:?} current_boot={} stale={}",
+                    job.story_id,
+                    job.status,
+                    job.error,
+                    decoded,
+                    current_boot,
+                    stale,
+                );
                 if stale {
                     slog!(
                         "[merge] Cleared stale Running merge job for '{}' (server restarted) — re-dispatching",