From 422ee477d94637f54435ef79eb2914e879b2cc68 Mon Sep 17 00:00:00 2001
From: James McMullan <jpmcmu@gmail.com>
Date: Wed, 5 Jun 2024 09:55:34 -0400
Subject: [PATCH] HPCC4J-605 Connection: Improve Invalid URL Error Message
 (#712)

Signed-off-by: James McMullan James.McMullan@lexisnexis.com
---
 .../ws/client/utils/Connection.java           | 27 ++++++++++++++++++-
 .../ws/client/utils/ConnectionTest.java       | 24 +++++++++++++++++
 2 files changed, 50 insertions(+), 1 deletion(-)

diff --git a/wsclient/src/main/java/org/hpccsystems/ws/client/utils/Connection.java b/wsclient/src/main/java/org/hpccsystems/ws/client/utils/Connection.java
index a28da183d..b7e4ec50c 100644
--- a/wsclient/src/main/java/org/hpccsystems/ws/client/utils/Connection.java
+++ b/wsclient/src/main/java/org/hpccsystems/ws/client/utils/Connection.java
@@ -12,6 +12,8 @@
 import java.util.Base64;
 import java.util.Base64.Decoder;
 import java.util.Base64.Encoder;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
 
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
@@ -205,6 +207,9 @@ public int hashCode()
     private StringBuffer        baseUrl;
     private StringBuffer        uriAndParams;
 
+    // Note: this pattern is very basic and is only meant to extract hostnames from URLs
+    public final static Pattern URL_HOSTNAME_PATTERN = Pattern.compile("((https?|ftp|file):\\/\\/)?(?<hostname>([\\da-z\\.\\-_]+)(\\.[a-z\\.]{2,6})?)(:\\d{2,6})?.*");
+
     /** Constant <code>CONNECT_TIMEOUT_PARAM="connecttimeoutmillis"</code> */
     final static public String  CONNECT_TIMEOUT_PARAM         = "connecttimeoutmillis";
     /** Constant <code>READ_TIMEOUT_PARAM="readtimeoutmillis"</code> */
@@ -287,7 +292,27 @@ public static boolean isSslProtocol(String protocol)
      */
     public Connection(String connectionstring) throws MalformedURLException
     {
-        URL theurl = new URL(connectionstring);
+        URL theurl = null;
+        try
+        {
+            theurl = new URL(connectionstring);
+        }
+        catch (MalformedURLException e)
+        {
+            Matcher matcher = URL_HOSTNAME_PATTERN.matcher(connectionstring);
+            if (matcher.matches())
+            {
+                String hostName = matcher.group("hostname");
+                if (hostName.contains("_"))
+                {
+                    throw new MalformedURLException("Invalid URL: Hostname contains invalid underscores: '" + connectionstring + "': " + e.getMessage());
+                }
+            }
+            else
+            {
+                throw e;
+            }
+        }
 
         setProtocol(theurl.getProtocol());
 
diff --git a/wsclient/src/test/java/org/hpccsystems/ws/client/utils/ConnectionTest.java b/wsclient/src/test/java/org/hpccsystems/ws/client/utils/ConnectionTest.java
index edfc9d640..278b40540 100644
--- a/wsclient/src/test/java/org/hpccsystems/ws/client/utils/ConnectionTest.java
+++ b/wsclient/src/test/java/org/hpccsystems/ws/client/utils/ConnectionTest.java
@@ -19,6 +19,7 @@
 import static org.junit.Assert.*;
 
 import java.net.MalformedURLException;
+import java.util.regex.Matcher;
 
 import org.junit.Test;
 
@@ -107,4 +108,27 @@ public void testInvalidProtHostPort() throws MalformedURLException
         assertFalse(con.getIsHttps());
         assertEquals(con.getProtocol(), http);
     }
+
+    @Test
+    public void hostNamePatternTest() throws MalformedURLException
+    {
+        // Note: we want to test improved error messaging with underscores, but not all versions
+        // of Java throw an exception for underscores in hostnames.
+        // So we are testing the pattern instead
+        String[] urls = {
+            "https://invalid_host_name.test:8010?params",
+            "https://invalid_host_name.test:8010",
+            "http://invalid_host_name.test:8010",
+            "invalid_host_name.test:8010",
+            "invalid_host_name.test"
+        };
+
+        String hostName = "invalid_host_name.test";
+        for (String url : urls)
+        {
+            Matcher matcher = Connection.URL_HOSTNAME_PATTERN.matcher(url);
+            assertTrue(matcher.matches());
+            assertEquals(matcher.group("hostname"), hostName);
+        }
+    }
 }