GZipStream Class

Is it just me or does this just take your .tar.gz file and leave you with a .tar file, not very helpful.

Please tell me if I'm missing something.

If you just want to zip a file and don't need all the bells, whistles and complexities of third party code. This works fine in .net 2.0:

Imports System.IO
Imports System.IO.Compression

    Public Sub Compress(ByVal FilePath As String)
        Dim UncompressedData As Byte() = System.IO.File.ReadAllBytes(FilePath)
        Dim CompressedData As New MemoryStream()
        Dim GZipper As New GZipStream(CompressedData, CompressionMode.Compress, True)
        GZipper.Write(UncompressedData, 0, UncompressedData.Length)
        GZipper.Dispose()
        System.IO.File.WriteAllBytes(IO.Path.GetPathRoot(FilePath) + "\" + IO.Path.GetFileNameWithoutExtension(FilePath) + ".zip", CompressedData.ToArray)
        CompressedData.Dispose()
    End Sub

AND IT IS VB.NET NATIVE!!! NO C# to go bald over trying to read and understand.

Here was my solution:

FileStream sourceStream = File.OpenRead(file.FullName);
FileInfo destFile =
    new FileInfo(ArchiveLocation + nsyskey.ToString() + ".gz");
FileStream destStream = File.Create(destFile.FullName); 
GZipStream compressedzipStream = 
    new GZipStream(destStream, CompressionMode.Compress, true); 
bool isError = false; 
byte[] buffer = new byte[0]; 
FileStream infile;
try
{
    const int MAX_BUFFER_SIZE = 100000; 
    // Read the file to ensure it is readable. 
    int start = 0; 
    int last = System.Convert.ToInt32(infile.Length); 
    int bufferSize = MAX_BUFFER_SIZE; 
    while (start < last) 
    {
        if (start + MAX_BUFFER_SIZE > last) 
        {
            bufferSize = last - start;
        }
        buffer = newbyte[bufferSize];
        int count = infile.Read(buffer, 0, bufferSize); 
        if (count != bufferSize) 
        {
            infile.Close();
            compressedzipStream.Close();
            Console.WriteLine("Test Failed: Unable to read data from file"); 
            return; 
        }
        compressedzipStream.Write(buffer, 0, bufferSize);
        start += bufferSize;
    }
}
finally 
{
    infile.Close();
    compressedzipStream.Close();
}

Decompression was very similar expect you convert the last 4 bytes to be your last and read from the gzipstream and write to the filestream

If you just want to read a gzip-compressed text file, you can use stream composition:

      using (var fs = new System.IO.FileStream(path, FileMode.Open, FileAccess.Read))
        using (var gzs = new GZipStream(fs, CompressionMode.Decompress))
        {
            using (var reader = new System.IO.StreamReader(gzs))
            {
                string line;
                while ((line = reader.ReadLine()) != null)
                {
                    System.Console.WriteLine(line);
                }
            }
        }

GZipStream doesn't handle ZIP files, but DotNetZip does. You can use DotNetZip (http://dotnetzip.codeplex.com ), a free 3rd-party library, to create and read zip files from within any .NET application.

This code in C#, zips all the files in a specified directory.

using (ZipFile zip = new ZipFile())
{
   zip.AddDirectory(@"MyDocuments\ProjectX", "ProjectX");
   zip.Comment = "This zip was created at " + System.DateTime.Now.ToString("G") ; 
   zip.Save(zipFileToCreate);
}

This code in C#, unzips a zipfile:

      string unpackDirectory = "ExtractedFiles";
      using (ZipFile zip1 = ZipFile.Read(zipToUnpack))
      {
          // here, we extract every entry, but we could extract conditionally
          // based on entry name, size, date, checkbox status, etc.  
          foreach (ZipEntry e in zip1)
          {
            e.Extract(unpackDirectory, ExtractExistingFileAction.OverwriteSilently);
          }
       }

DotNetZip is free.

Notice DeflateStream cannot read/write a .zip file directly (neither GZipStream)
ZipStorer library provides support for Zip files for .net and .net compact frameworks in a simple and monolithic class: http://zipstorer.codeplex.com
Also non-compressed storage support for Silverlight.

The example from the help really confuses things. I had a simple requirement to use XElement.Load to load an XMLdocument from a compressed file rather than from a regular file. Against a regular XML file this is easy:

Dim doc as XElement
doc = XElement.Load("testdata.xml")

Dim infile as FileStream' infile is the testdata.xml.gz file
Dim Decompressed as GZipStream' The output byte stream of uncompressed data
Dim charsDecompressed as StreamReader' The TextReader character input required by XElement.Load()
' A TextReader cannot be used as it is an abstract base class

Try
infile = New FileStream("testdata.xml.gz", FileMode.Open)
        Decompressed = New GZipStream(infile, CompressionMode.Decompress)
        charsDecompressed = New StreamReader(Decompressed)
        SRD = XElement.Load(charsDecompressed)
        charsDecompressed.Dispose()
        infile.Close()
Catch ex As XmlException
        ' The XML itself was not properly formed
        MessageBox.Show("The file appears to be corrupt.", "Fatal error", MessageBoxButtons.OK, MessageBoxIcon.Error)
        Return False

Catch ex As InvalidDataException
        ' The decompression process failed somehow
        MessageBox.Show("Extraction error while loading data file.", "Fatal error", MessageBoxButtons.OK, MessageBoxIcon.Error)
        Return False

Catch ex As FileNotFoundException
        ' File could not be found
        MessageBox.Show("The testdata.xml.gz file could not be found, "Fatal error", MessageBoxButtons.OK, MessageBoxIcon.Error)
        Return False

Catch ex As Exception
        ' Catch everything else
        MessageBox.Show("An unexpected error occurred: " & ex.Message, "Unknown error", MessageBoxButtons.OK, MessageBoxIcon.Error)
        Return False

End Try

Using GZipStream for Compression in .NET [Brian Grunkemeyer] http://blogs.msdn.com/bclteam/archive/2005/06/15/429542.aspx

Gives the nice clean code for Compress and Decompress methods

Using a MemoryStream with GZipStream [Lakshan Fernando] http://blogs.msdn.com/bclteam/archive/2006/05/10/592551.aspx

To explain the very unintuitive aspect of this stream

>I would leave optimization to the compiler, myself, and not bother with buffering the ReadAllBytes method

Doing it that way without a buffer will be much much slower.
I timed both the code you provide and original posters code - and without a buffer of at least 1024bytes - the code ran 3 times slower for a sample 25mb file.

One more thing - if you are looking to handle Zlib Compression (IETF RFC 1950), that is in DotNetZip, too. There is a ZlibStream class to handle it nicely. Works very similarly to the GZipStream class.

There are 2 problems with this class.

1. It does not handle ZIP files.
2. It is dysfunctional, can actually inflate data in "compression". There's something wrong with the logic. It's a known problem but as yet unfixed.

On the first item, the Deflate compression algorithm is what is used in .zip files. A GZIP stream is just a DEFLATEd stream with a header and trailer surrounding it. A zip file is a set of DEFLATEd streams with metadata surrounding all of them. This class helps you read or write a GZIP stream, but it will not help you read or write a zip file. A GZIP file (sometimes with the extension .gz) is not the same thing as a ZIP file. Similar ideas but not the same thing.

The java.util.zip classes do handle .zip files. These classes are available through the VJ# runtime. But, there are a few problems with them: (1) the java classes are unwieldy to use in .NET. There are no progress events, for example. The enumeration is all Java-esque. etc.; (2) there are a bunch of bugs in those classes that have not been and will not be fixed in the VJ# runtime; (3) the VJ# runtime is huge; (4) The VJ# runtime is no longer supported! (not shipped with VS2008).

There is a good 3rd party library, though, that solves all these problems: DotNetZip. It provides a GZipStream that NEVER inflates your data, and actually deflates it substantially better than the built-in class. Also, DotNetZip can read and write .zip files.

Find it at:
http://www.codeplex.com/DotNetZip

It's free to use. It is written in C#, but you can use it from any language. It works in Winforms apps, console apps, ASP.NET apps, Powershell, anything you write in .NET. Fast, good compression (better than the built-in GZipStream). Easy to use. You'll find lots of examples on that site.

Any suggestions on how to zip an entire directory using GZipStream?
I'm currently using java.util.zip classes in the vjslib.dll. It works well but I'd rather avoid using an outside library if I can avoid it.

Sweet!

I would leave optimization to the compiler, myself, and not bother with buffering the ReadAllBytes method...laziness. (and create a ToArray() or ToMemoryStream() extension method on Stream, itself. Extension methods really help plug some huge holes!

const int EndOfStream = -1;
private static byte[] ReadAllBytes(Stream stream)
{
using (var ms = new MemoryStream())
{
for (int b = stream.ReadByte(); b != EndOfStream; b = stream.ReadByte())
ms.WriteByte((byte) b);
return ms.ToArray();
}
}