Prevent 'Too many open files' crash on store source switch

- Raise the process soft file-descriptor limit to the hard limit at startup
  (GLib aborts the process with 'Creating pipes for GWakeup: Too many open
  files' when cover/metadata loading exhausts a 1024-soft-limit environment).
- Reduce per-loader network concurrency 8 -> 6 threads to lower socket
  pressure against the store server.

app/gui.py

@@ -136,7 +136,7 @@ class CoverLoader(QObject):
     def run(self):
         from concurrent.futures import ThreadPoolExecutor, as_completed
 
-        pool = ThreadPoolExecutor(max_workers=8)
+        pool = ThreadPoolExecutor(max_workers=6)
         futures = {
             pool.submit(load_cover, it): i
             for i, it in enumerate(self.items)
@@ -1634,7 +1634,7 @@ class StoreCoverLoader(QObject):
     def run(self):
         from concurrent.futures import ThreadPoolExecutor, as_completed
 
-        pool = ThreadPoolExecutor(max_workers=8)
+        pool = ThreadPoolExecutor(max_workers=6)
         futures = {
             pool.submit(load_cover, it, 8, self.quick): row
             for row, it in self.tasks
@@ -5954,8 +5954,24 @@ class MainWindow(QMainWindow):
         event.accept()
 
 
+def _raise_nofile_limit():
+    """Raise the soft file-descriptor limit to the hard limit.
+
+    Cover/metadata loading opens many short-lived sockets, and each worker
+    QThread's GLib event dispatcher creates a wakeup pipe. On systems with a
+    1024 soft limit this aborts the process with GLib's "Too many open files".
+    """
+    try:
+        import resource
+        _soft, hard = resource.getrlimit(resource.RLIMIT_NOFILE)
+        resource.setrlimit(resource.RLIMIT_NOFILE, (hard, hard))
+    except Exception:  # noqa: BLE001 - best effort, never block startup
+        pass
+
+
 def main():
     logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)-7s %(name)-14s %(message)s")
+    _raise_nofile_limit()
     init_db()  # migrate legacy on-share DB/covers to local storage, then schema
     # Auto-ingest any missing DATs so scanning matches against the catalog
     # even on a fresh/reset database (no manual `python -m app.ingest` step).