From 637cf89e180a42fe85817738707c629bb5e56a4c Mon Sep 17 00:00:00 2001
From: Dhruba Borthakur <dhruba@apache.org>
Date: Tue, 16 Mar 2010 08:03:38 +0000
Subject: [PATCH] HADOOP-6591. HarFileSystem can handle paths with the
 whitespace characters. (Rodrigo Schmidt via dhruba)

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@923619 13f79535-47bb-0310-9956-ffa450edef68
---
 CHANGES.txt                                   |  3 +++
 .../org/apache/hadoop/fs/HarFileSystem.java   | 25 ++++++++++++++-----
 2 files changed, 22 insertions(+), 6 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 178b3ddf21..b9d07341fa 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -190,6 +190,9 @@ Trunk (unreleased changes)
     HADOOP-6486. fix common classes to work with Avro 1.3 reflection.
     (cutting via tomwhite)
 
+    HADOOP-6591. HarFileSystem can handle paths with the whitespace characters.
+    (Rodrigo Schmidt via dhruba)
+
   OPTIMIZATIONS
 
     HADOOP-6467. Improve the performance on HarFileSystem.listStatus(..).
diff --git a/src/java/org/apache/hadoop/fs/HarFileSystem.java b/src/java/org/apache/hadoop/fs/HarFileSystem.java
index 051c4433e2..584b653927 100644
--- a/src/java/org/apache/hadoop/fs/HarFileSystem.java
+++ b/src/java/org/apache/hadoop/fs/HarFileSystem.java
@@ -19,8 +19,10 @@
 
 import java.io.FileNotFoundException;
 import java.io.IOException;
+import java.io.UnsupportedEncodingException;
 import java.net.URI;
 import java.net.URISyntaxException;
+import java.net.URLDecoder;
 import java.util.ArrayList;
 import java.util.EnumSet;
 import java.util.List;
@@ -45,7 +47,7 @@
  */
 
 public class HarFileSystem extends FilterFileSystem {
-  public static final int VERSION = 1;
+  public static final int VERSION = 2;
   // uri representation of this Har filesystem
   private URI uri;
   // the version of this har filesystem
@@ -121,7 +123,8 @@ public void initialize(URI name, Configuration conf) throws IOException {
       throw new IOException("Unable to " +
           "read the version of the Har file system: " + this.archivePath);
     }
-    if (this.version != HarFileSystem.VERSION) {
+    // make it always backwards-compatible
+    if (this.version > HarFileSystem.VERSION) {
       throw new IOException("Invalid version " + 
           this.version + " expected " + HarFileSystem.VERSION);
     }
@@ -213,6 +216,15 @@ private URI decodeHarURI(URI rawURI, Configuration conf) throws IOException {
     return tmp;
   }
   
+  private String decodeFileName(String fname) 
+    throws UnsupportedEncodingException {
+    
+    if (version == 2){
+      return URLDecoder.decode(fname, "UTF-8");
+    }
+    return fname;
+  }
+  
   /**
    * return the top level archive.
    */
@@ -479,6 +491,7 @@ private String fileStatusInIndex(Path harPath) throws IOException {
         read += tmp;
         String lineFeed = line.toString();
         String[] parsed = lineFeed.split(" ");
+        parsed[0] = decodeFileName(parsed[0]);
         if (harPath.compareTo(new Path(parsed[0])) == 0) {
           // bingo!
           retStr = lineFeed;
@@ -502,16 +515,16 @@ private String fileStatusInIndex(Path harPath) throws IOException {
   // the format is of the form 
   // filename "dir"/"file" partFileName startIndex length 
   // <space seperated children>
-  private static class HarStatus {
+  private class HarStatus {
     boolean isDir;
     String name;
     List<String> children;
     String partName;
     long startIndex;
     long length;
-    public HarStatus(String harString) {
+    public HarStatus(String harString) throws UnsupportedEncodingException {
       String[] splits = harString.split(" ");
-      this.name = splits[0];
+      this.name = decodeFileName(splits[0]);
       this.isDir = "dir".equals(splits[1]) ? true: false;
       // this is equal to "none" if its a directory
       this.partName = splits[2];
@@ -520,7 +533,7 @@ public HarStatus(String harString) {
       if (isDir) {
         children = new ArrayList<String>();
         for (int i = 5; i < splits.length; i++) {
-          children.add(splits[i]);
+          children.add(decodeFileName(splits[i]));
         }
       }
     }