feat: JOB_STAT_TRANSIENT, job_should_retain, worker_should_retain, dead_since (P0-5a, P0-8, P0-13)

Add terminal-state constants and GC retention predicates for the
checkpoint xzarray_thaw() GC pass:

- types.h: add JOB_STAT_TRANSIENT (0x400), XCE_WORKER_HIDDEN (0x00200000),
  QUEUE_STAT_HIDDEN (0x0100)
- persist.c: implement job_is_terminal(), job_should_retain(),
  worker_should_retain() — placed here so the unit test binary
  (which links only persist.c) can access them without RTTI deps
- persist.h: declare the three GC predicates; include xce/types.h
- worker.c: set dead_since = time(NULL) in mark_worker_unavail()
  when dead_since is not already set
- cluster.c: add TODO P0-8b marker on zarray_del endpoint cleanup
- meson.build: register test_gc_retention.c; raise test timeout 60→120s
- testsuite/test_gc_retention.c: 19 tests covering job_is_terminal,
  job_should_retain (TRANSIENT TTL, terminal TTL, non-terminal always
  retained), and worker_should_retain (HIDDEN/UNAVAIL/healthy/zero
  dead_since); all pass (63 total, 0 failures)

src/controller/persist.c

@@ -3,8 +3,10 @@
 #include "config.h"
 #include "hook.h"
 #include "xce/cluster_def.h"
+#include "xce/types.h"
 
 #include <assert.h>
+#include <stdbool.h>
 #include <dirent.h>
 #include <errno.h>
 #include <fcntl.h>
@@ -873,3 +875,56 @@ int persist_worker_hook_fun(int type, int situation, int detail,
     uint32_t payload = (uint32_t)detail;
     return persist_wal_async(p, wal_type, &payload, sizeof(payload));
 }
+
+/* ============================================================
+ * GC retention predicates (P0-5a / P0-8 / P0-13)
+ *
+ * Placed in persist.c so the unit test binary (which links only
+ * persist.c from the controller sources) can access them without
+ * pulling in RTTI-dependent job.c / worker.c.
+ * ============================================================ */
+
+bool job_is_terminal(const job_t* j)
+{
+    return (j->job.state & (JOB_STAT_EXIT | JOB_STAT_DONE | JOB_STAT_PERR)) != 0;
+}
+
+bool job_should_retain(const job_t* j, const persist_config_t* cfg)
+{
+    time_t now = time(NULL);
+
+    /* TRANSIENT jobs: auto-terminate when TTL elapses after failover.
+     * transient_since == 0 means the timestamp was never set — do not GC. */
+    if (j->job.state & JOB_STAT_TRANSIENT) {
+        if (j->job.transient_since > 0 &&
+            now > (time_t)j->job.transient_since + (time_t)cfg->transient_job_timeout_s) {
+            return false;
+        }
+        return true;
+    }
+
+    /* Non-terminal jobs are always retained */
+    if (!job_is_terminal(j)) return true;
+
+    /* Terminal jobs: retain until finished_retain_s expires */
+    return (j->job.finish_time + (time_t)cfg->finished_retain_s >= now);
+}
+
+bool worker_should_retain(const xce_worker_t* w, const persist_config_t* cfg)
+{
+    time_t now = time(NULL);
+
+    /* HIDDEN workers: admin explicitly removed — clean up at next checkpoint */
+    if (w->status & XCE_WORKER_HIDDEN) return false;
+
+    /* UNAVAILABLE workers: retain until worker_retain_s expires.
+     * dead_since == 0 means not yet set — do not GC. */
+    if (w->status & XCE_WORKER_UNAVAIL) {
+        if (w->dead_since > 0 &&
+            now > w->dead_since + (time_t)cfg->worker_retain_s) {
+            return false;
+        }
+    }
+
+    return true;
+}