From 86e737d7bad09d3c4e21a0deb360c8158537f443 Mon Sep 17 00:00:00 2001
From: Sebastian Baunsgaard <baunsgaard@apache.org>
Date: Wed, 6 Sep 2023 10:33:53 +0200
Subject: [PATCH] [MINOR] FileSystemConfig refinement

This commit change the loading of the filesystem to return a future
to guarantee that we do not launch multiple filesystem requests at once.

Also included is a cleanup in pom.xml that fixes some space vs tabs
introduced.
---
 pom.xml                                       | 28 +++++++++++--------
 .../sysds/conf/ConfigurationManager.java      | 19 +++++++++----
 2 files changed, 31 insertions(+), 16 deletions(-)

diff --git a/pom.xml b/pom.xml
index 69942a2fbc0..4fc1dd2ccdc 100644
--- a/pom.xml
+++ b/pom.xml
@@ -235,9 +235,9 @@
 									<classpathPrefix>lib/</classpathPrefix>
 									<mainClass>org.apache.sysds.performance.Main</mainClass>
 								</manifest>
-								 <manifestEntries>
-									 <Class-Path>SystemDS.jar ${project.build.directory}/${project.artifactId}-${project.version}-tests.jar</Class-Path>
-								 </manifestEntries>
+								<manifestEntries>
+									<Class-Path>SystemDS.jar ${project.build.directory}/${project.artifactId}-${project.version}-tests.jar</Class-Path>
+								</manifestEntries>
 							</archive>
 						</configuration>
 					</execution>
@@ -309,7 +309,7 @@
 				<artifactId>maven-compiler-plugin</artifactId>
 				<version>${maven-compiler-plugin.version}</version>
 				<configuration>
-			        <compilerArgument>-Xlint:unchecked</compilerArgument>
+					<compilerArgument>-Xlint:unchecked</compilerArgument>
 					<source>${java.level}</source>
 					<target>${java.level}</target>
 					<release>${java.level}</release>
@@ -408,12 +408,12 @@
 					<output>file</output>
 					<destFile>${project.build.directory}/jacoco.exec</destFile>
 					<fileSets>
-					  <fileSet>
-					    <directory>${project.build.directory}</directory>
-					    <includes>
-						    <include>*/jacoco.exec</include>
-					    </includes>
-					  </fileSet>
+						<fileSet>
+							<directory>${project.build.directory}</directory>
+							<includes>
+								<include>*/jacoco.exec</include>
+							</includes>
+						</fileSet>
 					</fileSets>
 				</configuration>
 				<executions>
@@ -655,6 +655,7 @@
 								<exclude>src/test/scripts/functions/jmlc/tfmtd_example/dummycoded.column.names</exclude>
 								<exclude>src/test/scripts/functions/jmlc/tfmtd_example2/column.names</exclude>
 								<exclude>src/test/scripts/functions/jmlc/tfmtd_frame_example/tfmtd_frame</exclude>
+								<exclude>src/test/scripts/applications/entity_resolution/connected_components/expected/**</exclude>
 								<!-- csv test input not captured by *.csv -->
 								<exclude>src/test/scripts/functions/io/csv/in/**</exclude>
 								<!-- IOGEN input examples -->
@@ -1488,5 +1489,10 @@
 			<artifactId>log4j-core</artifactId>
 			<version>${log4j.version}</version>
 		</dependency>
+		<dependency>
+			<groupId>ch.randelshofer</groupId>
+			<artifactId>fastdoubleparser</artifactId>
+			<version>0.9.0</version>
+		</dependency>
 	</dependencies>
-</project>
\ No newline at end of file
+</project>
diff --git a/src/main/java/org/apache/sysds/conf/ConfigurationManager.java b/src/main/java/org/apache/sysds/conf/ConfigurationManager.java
index 088545b8ed8..a4b5c0ffecd 100644
--- a/src/main/java/org/apache/sysds/conf/ConfigurationManager.java
+++ b/src/main/java/org/apache/sysds/conf/ConfigurationManager.java
@@ -20,6 +20,7 @@
 package org.apache.sysds.conf;
 
 import java.util.concurrent.ExecutorService;
+import java.util.concurrent.Future;
 
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
@@ -56,17 +57,20 @@ public class ConfigurationManager{
 	/** Local compiler configuration for thead-local config updates */
 	private static ThreadLocalCompilerConfig _lcconf = new ThreadLocalCompilerConfig();
 
+	/** Indicate if the filesystem is loaded or not */
+	private static Future<?> loaded;
+
 	//global static initialization
 	static {
 		_rJob = new JobConf();
 		
 		//initialization after job conf in order to prevent cyclic initialization issues 
-		//ConfigManager -> OptimizerUtils -> InfrastructureAnalyer -> ConfigManager 
+		//ConfigManager -> OptimizerUtils -> InfrastructureAnalyzer -> ConfigManager 
  		_dmlconf = new DMLConfig();
 		_cconf = new CompilerConfig();
 
 		final ExecutorService pool = CommonThreadPool.get();
-		pool.submit(() ->{
+		loaded = pool.submit(() ->{
 			try{
 				IOUtilFunctions.getFileSystem(_rJob);
 			}
@@ -74,19 +78,24 @@ public class ConfigurationManager{
 				LOG.warn(e.getMessage());
 			}
 		});
-		// pool.shutdown();
 	}
 	
 	
 	/**
 	 * Returns a cached JobConf object, intended for global use by all operations
 	 * with read-only access to job conf. This prevents to read the hadoop conf files
-	 * over and over again from classpath. However,
+	 * over and over again from classpath.
 	 *
 	 * @return the cached JobConf
 	 */
 	public static JobConf getCachedJobConf() {
-		return _rJob;
+		try{
+			loaded.get();
+			return _rJob;
+		}
+		catch(Exception e){
+			throw new RuntimeException(e);
+		}
 	}
 	
 	public static void setCachedJobConf(JobConf job) {