From ea9c7ee5d6635405b31b4a1d462cca746478b040 Mon Sep 17 00:00:00 2001
From: Sebastian Nagel <snagel@apache.org>
Date: Thu, 11 Apr 2024 13:28:37 +0200
Subject: [PATCH] NUTCH-3039 Failure to handle ftp:// URLs

Pass ftp:// URLs to the standard JVM URLStreamHandler
---
 .../org/apache/nutch/plugin/URLStreamHandlerFactory.java    | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/src/java/org/apache/nutch/plugin/URLStreamHandlerFactory.java b/src/java/org/apache/nutch/plugin/URLStreamHandlerFactory.java
index bd7e377d0f..0916f4c9db 100644
--- a/src/java/org/apache/nutch/plugin/URLStreamHandlerFactory.java
+++ b/src/java/org/apache/nutch/plugin/URLStreamHandlerFactory.java
@@ -72,9 +72,13 @@ public class URLStreamHandlerFactory
    * Protocols covered by standard JVM URL handlers. These protocols must not be
    * handled by Nutch plugins, in order to avoid that basic actions (eg. loading
    * of classes and configuration files) break.
+   * 
+   * Also the "ftp" protocol is included: it's usually supported by the standard
+   * JVM URL handler and Nutch does not yet provide a dedicated URL stream
+   * handler.
    */
   public static final String[] SYSTEM_PROTOCOLS = { //
-      "http", "https", "file", "jar" };
+      "http", "https", "file", "jar", "ftp" };
 
   static {
     instance = new URLStreamHandlerFactory();