increase chunk download timeout

Author: void-memories

steady_fetch/src/main/java/dev/namn/steady_fetch/impl/di/DependencyContainer.kt

@@ -11,19 +11,29 @@ import java.util.concurrent.TimeUnit
 import kotlinx.coroutines.CoroutineScope
 import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.SupervisorJob
+import okhttp3.ConnectionPool
 import okhttp3.OkHttpClient
 
 internal class DependencyContainer private constructor(
     private val application: Application
 ) {
     private val okHttpClient: OkHttpClient by lazy {
+        // Connection pool: max 50 idle connections, keep alive for 5 minutes
+        // This allows many concurrent chunk downloads across multiple files
+        val connectionPool = ConnectionPool(50, 5, TimeUnit.MINUTES)
+        
         OkHttpClient.Builder()
             .connectTimeout(Constants.DEFAULT_CONNECT_TIMEOUT_SECONDS, TimeUnit.SECONDS)
-            .readTimeout(Constants.DEFAULT_READ_TIMEOUT_SECONDS, TimeUnit.SECONDS)
-            .callTimeout(
-                Constants.DEFAULT_READ_TIMEOUT_SECONDS * 2,
-                TimeUnit.SECONDS
-            )
+            // Very generous read timeout for large chunks and slow connections
+            // For 8MB chunk: at 0.1MB/s = 80s, at 0.05MB/s = 160s
+            // 10 minutes provides ample margin for very slow connections and network hiccups
+            .readTimeout(10, TimeUnit.MINUTES)
+            // Write timeout for slow uploads (though we're downloading)
+            .writeTimeout(5, TimeUnit.MINUTES)
+            // Call timeout should be much longer than read timeout to allow for retries
+            // 30 minutes allows for multiple retries even on very slow connections
+            .callTimeout(30, TimeUnit.MINUTES)
+            .connectionPool(connectionPool)
             .retryOnConnectionFailure(true)
             .build()
     }