feat: Implement baseline profile generation and startup benchmarking

- Added baseline profile generator for Rosetta Messenger to optimize startup performance.
- Created startup benchmark tests to measure cold start times under different compilation modes.
- Introduced a new Gradle module for baseline profile and benchmark tests.
- Updated ChatsListViewModel to show loading skeleton while data is being fetched.
- Improved keyboard handling in MessageInputBar by using SHOW_IMPLICIT instead of SHOW_FORCED.
- Minor code cleanups and optimizations across various components.

app/src/main/java/com/rosetta/messenger/database/MessageEntities.kt

@@ -367,10 +367,24 @@ interface DialogDao {
           AND i_have_sent = 1
           AND last_message_timestamp > 0
         ORDER BY last_message_timestamp DESC
+        LIMIT 30
     """
     )
     fun getDialogsFlow(account: String): Flow<List<DialogEntity>>
 
+    /** Получить все диалоги с пагинацией */
+    @Query(
+            """
+        SELECT * FROM dialogs
+        WHERE account = :account
+          AND i_have_sent = 1
+          AND last_message_timestamp > 0
+        ORDER BY last_message_timestamp DESC
+        LIMIT :limit OFFSET :offset
+    """
+    )
+    suspend fun getDialogsPaged(account: String, limit: Int, offset: Int): List<DialogEntity>
+
     /**
      * Получить requests - диалоги где нам писали, но мы не отвечали Исключает пустые диалоги (без
      * сообщений)
@@ -382,6 +396,7 @@ interface DialogDao {
           AND i_have_sent = 0
           AND last_message_timestamp > 0
         ORDER BY last_message_timestamp DESC
+        LIMIT 30
     """
     )
     fun getRequestsFlow(account: String): Flow<List<DialogEntity>>