Roll back a Space Missions entry whose mission index fails to load

The launches layer swallowed every fetch error in its update, so the
manager treated a 429 or timeout on /api/launches as a successful start.
Context then committed the isolating mode with no missions: the other
layers stayed off, unrelated enables were refused, the camera stayed at
the full-globe view and the roster kept reading "loading".

Until the first index loads, the update now reports the failure. The
manager fails the enable closed and both entry routes roll back to the
pre-entry layers. The camera pose captured at entry is restored when the
mode is left before any index loaded, and the roster shows a failed
state instead of loading. A valid empty index still enters, and a
refresh failure after a loaded index keeps its records.

scripts/format-scope.json

@@ -474,6 +474,7 @@
   "src/layers/satellites/stale.test.mjs",
   "src/layers/launches/source.test.mjs",
   "src/layers/launches/ingestion.test.mjs",
+  "src/layers/launches/failedEntry.test.mjs",
   "src/layers/installations/lifecycle.js",
   "src/layers/installations/viewport.js",
   "src/layers/installations/controls.js",