This commit refactors the node database handling within `MeshService` to improve performance and reliability during initial configuration synchronization.

Previously, each `NodeInfo` received from the device was individually written to the database, leading to multiple transactions. This could be slow and inefficient, especially on startup.

The key changes are:
- Introduced a batch update mechanism for the initial node list. Instead of writing each node to the database one by one, all nodes are now collected and saved in a single transaction when the `NodeInfoComplete` signal is received.
- The `haveNodeDB` flag has been split into two more descriptive flags:
    - `isNodeDbReady`: Signals that the initial node database is synchronized and ready for use.
    - `allowNodeDbWrites`: Controls when individual node updates can be written to the persistent database, which is enabled only after the initial batch-write is complete.
- `updateNodeInfo()` now includes a `withBroadcast` parameter to prevent unnecessary broadcasts and database writes during the initial batch processing.
- The `myNodeInfo` object is now updated more reliably from the `nodeRepository` flow at service start.

Signed-off-by: James Rich <2199651+jamesarich@users.noreply.github.com>