Share the target filter, fix cd-status branch, record deploy status

- enabled_targets/names_for: one implementation of 'a host acts only on its
  own targets'; --host s5 and --host s5.fisher.hu now select the same targets
- shell_value: JSON booleans reach shell consumers as true/false
- cd-status: capture before eval, so the unconfigured-host branch is reachable
- cd-deploy writes a .status record; cd-status reads it instead of grepping
  the log's prose
- cd-render-hooks --list-urls replaces two inline JSON readers

bin/cd-deploy

@@ -107,6 +107,7 @@ eval "$resolved"
 
 mkdir -p "$STATE_DIR"
 readonly LOG_FILE="${STATE_DIR}/${CD_NAME}.log"
+readonly STATUS_FILE="${STATE_DIR}/${CD_NAME}.status"
 exec > >(stdbuf -oL tee -a "$LOG_FILE") 2>&1
 
 log "=============================================================="
@@ -129,6 +130,15 @@ notify() {
         fail) icon="❌" ;;
         *)    icon="ℹ️" ;;
     esac
+
+    # Every terminal outcome notifies, so this is also where the machine-readable
+    # record of that outcome belongs. cd-status reads this file rather than
+    # grepping the log's prose, which quietly reported "no completed deploy"
+    # for a healthy target whenever a log line was reworded.
+    printf '%s\t%s\t%s\t%s\n' \
+        "$(date -u '+%Y-%m-%dT%H:%M:%SZ')" "$status" "${SHA:0:12}" "$message" \
+        > "$STATUS_FILE"
+
     topic="${REPO//\//-}"
     curl -fsS --max-time 10 \
         -H "Title: ${CD_NAME} deploy" \
@@ -299,7 +309,7 @@ fi
 
 # ---------------------------------------------------------------- rollback ---
 
-if [[ "$CD_ROLLBACK_ON_FAILURE" != "True" && "$CD_ROLLBACK_ON_FAILURE" != "true" ]]; then
+if [[ "$CD_ROLLBACK_ON_FAILURE" != "true" ]]; then
     notify fail "health check failed for ${SHA:0:12}; rollback disabled, stack left on the new image"
     die "health check failed and rollback is disabled for this target"
 fi