The build machine takes one ask at a time, and says so while it builds

With the worker consumer's default of many deliveries in flight, every ask behind the one being
built was delivered at once, left unacknowledged for the length of the build, redelivered after the
ack wait and dropped after the fifth time: on 2026-10-01 twenty-six of forty-three builds asked in two
minutes were never built and the queue read as empty (hq issue 186). The holder's worker now has one
in flight, and a running build tells the bus it is still working, as the controller's long handlers
do, so a build longer than the ack wait is neither redelivered nor counted out.

internal/link/builds_nats.go

@@ -173,7 +173,13 @@ func (m *natsMachine) Take(ctx context.Context, do func(context.Context, Build))
 				_ = msg.Term()
 				continue
 			}
+			// A build outlives the acknowledgement window many times over; said while it runs,
+			// as the controller says it for its own long handlers, so the server neither hands
+			// the ask to a second machine nor counts the wait against its deliveries.
+			working := make(chan struct{})
+			go stillWorking(msg, working)
 			do(ctx, &natsBuild{request: request, msg: msg, on: m.on, js: m.js})
+			close(working)
 		}
 	}
 }