Java java.io.IOException: Not in GZIP format

Question

I searched for an example of how to compress a string in Java.

I have a function to compress then uncompress. The compress seems to work fine:

   public static String encStage1(String str)
   {
      String format1 = "ISO-8859-1";
      String format2 = "UTF-8";
      if (str == null || str.length() == 0)
      {
         return str;
      }
      System.out.println("String length : " + str.length());
      ByteArrayOutputStream out = new ByteArrayOutputStream();
      String outStr = null;
      try
      {
         GZIPOutputStream gzip = new GZIPOutputStream(out);
         gzip.write(str.getBytes());
         gzip.close();
         outStr = out.toString(format2);
         System.out.println("Output String lenght : " + outStr.length());
      } catch (Exception e)
      {
                  e.printStackTrace();

      }
      return outStr;
   }

But the reverse is complaining about the string not being in GZIP format, even when I pass the return from encStage1 straight back into the decStage3:

   public static String decStage3(String str)
   {
      if (str == null || str.length() == 0)
      {
         return str;
      }
      System.out.println("Input String length : " + str.length());
      String outStr = "";
      try
      {
         String format1 = "ISO-8859-1";
         String format2 = "UTF-8";
         GZIPInputStream gis = new GZIPInputStream(new ByteArrayInputStream(str.getBytes(format2)));
         BufferedReader bf = new BufferedReader(new InputStreamReader(gis, format2));
         String line;
         while ((line = bf.readLine()) != null)
         {
            outStr += line;
         }
         System.out.println("Output String lenght : " + outStr.length());
      } catch (Exception e)
      {
         e.printStackTrace();
      }
      return outStr;
   }

I get this error when I call with a string return from encStage1:

   public String encIDData(String idData)
   {
      String tst = "A simple test string";
      System.out.println("Enc 0: " + tst);
      String stg1 = encStage1(tst);
      System.out.println("Enc 1: " + toHex(stg1));
      String dec1 = decStage3(stg1);
      System.out.println("unzip: " + toHex(dec1));
   }

Output/Error:

Enc 0: A simple test string
String length : 20
Output String lenght : 40
Enc 1: 1fefbfbd0800000000000000735428efbfbdefbfbd2defbfbd495528492d2e51282e29efbfbdefbfbd4b07005aefbfbd21efbfbd14000000
Input String length : 40
java.io.IOException: Not in GZIP format
    at java.util.zip.GZIPInputStream.readHeader(GZIPInputStream.java:137)
    at java.util.zip.GZIPInputStream.<init>(GZIPInputStream.java:58)
    at java.util.zip.GZIPInputStream.<init>(GZIPInputStream.java:68)

Answer 1

A small error is:

     gzip.write(str.getBytes());

takes the default platform encoding, which on Windows will never be ISO-8859-1. Better:

     gzip.write(str.getBytes(format1));

You could consider taking "Cp1252", Windows Latin-1 (for some European languages), instead of "ISO-8859-1", Latin-1. That adds comma like quotes and such.

The major error is converting the compressed bytes to a String. Java separates binary data (byte[], InputStream, OutputStream) from text (String, char, Reader, Writer) which internally is always kept in Unicode. A byte sequence does not need to be valid UTF-8. You might get away by converting the bytes as a single byte encoding (ISO-8859-1 for instance).

The best way would be

     gzip.write(str.getBytes(StandardCharsets.UTF_8));

So you have full Unicode, every script may be combined.

And uncompressing to a ByteArrayOutputStream and new String(baos.toByteArray(), StandardCharsets.UTF_8). Using BufferedReader on an InputStreamReader with UTF-8 is okay too, but a readLine throws away the newline characters

outStr += line + "\r\n"; // Or so.

---

Clean answer:

public static byte[] encStage1(String str) throws IOException
{
   try (ByteArrayOutputStream out = new ByteArrayOutputStream())
   {
       try (GZIPOutputStream gzip = new GZIPOutputStream(out))
       {
           gzip.write(str.getBytes(StandardCharsets.UTF_8));
       }
       return out.toByteArray();
       //return out.toString(StandardCharsets.ISO_8859_1);
       // Some single byte encoding
   }
}

public static String decStage3(byte[] str) throws IOException
{
   ByteArrayOutputStream baos = new ByteArrayOutputStream();
   try (GZIPInputStream gis = new GZIPInputStream(new ByteArrayInputStream(str)))
   {
       int b;
       while ((b = gis.read()) != -1) {
           baos.write((byte) b);
       }
   }
   return new String(baos.toByteArray(), StandardCharset.UTF_8);
}

Answer 2

usage of toString/getBytes for encoding/decoding is a wrong way. try to use something like BASE64 encoding for this purpose (java.util.Base64 in jdk 1.8)

as a proof try this simple test:

import org.testng.annotations.Test;
import java.io.ByteArrayOutputStream;
import static org.testng.Assert.assertEquals;

public class SimpleTest {

    @Test
    public void test() throws Exception {

        final String CS = "utf-8";

        byte[] b0 = {(byte) 0xff};
        ByteArrayOutputStream out = new ByteArrayOutputStream();
        out.write(b0);
        out.close();

        byte[] b1 = out.toString(CS).getBytes(CS);

        assertEquals(b0, b1);
    }
}