ZipStorer64 3.5.1

ZipStorer64

Fast managed library for working with ZIP archives. A fork of ZipStorer (https://github.com/jaime-olivares/zipstorer)

Introduction

There are many techniques to produce Zip files in a .NET environment, like the following:

• Using the System.IO.Compression.ZipArchive class
• Using the java.util.zip namespace
• Invoking Shell API features
• Using a third-party .NET library
• Wrapping and marshalling a non-.NET library
• Invoking a compression tool at command-line

ZipStorer is a minimalistic class to create Zip files and store/retrieve files to/from it, by using the Deflate algorithm. No other compression methods supported.

Using the code

The ZipStorer class is the unique one needed to create the zip file. It contains a nested structure (ZipFileEntry) for collecting each directory entry. The class has been declared inside the System.IO namespace.

There is no default constructor. There are two ways to construct a new ZipStorer instance, depending on specific needs: use either the Create() or the Open() static method. To create a new Zip file, use the Create() method like this:

ZipStorer zip = ZipStorer.Create(filename, comment);  // file-oriented version
ZipStorer zip = ZipStorer.Create(stream, comment);  // stream-oriented version

It is required to specify the full path for the new zip file, or pass a valid stream, and optionally add a comment. For opening an existing zip file for appending, the Open() method is required, like the following:

ZipStorer zip = ZipStorer.Open(filename, fileaccess);  // file-oriented version
ZipStorer zip = ZipStorer.Open(stream, fileaccess);  // stream-oriented version

Where fileaccess should be of type System.IO.FileAccess enumeration type. Also, as now ZipStorer is derived from IDisposable interface, the using keyword can be used to ensure proper disposing of the storage resource:

using (ZipStorer zip = ZipStorer.Create(filename, comment))
{
    // some operations with zip object
    //
}   // automatic close operation here

For adding files into an opened zip storage, there are two available methods:

public void AddFile(ZipStorer.Compress _method, string _pathname, string _filenameInZip, string _comment);
public void AddStream(ZipStorer.Compress _method, string _filenameInZip, Stream _source, DateTime _modTime, string _comment);

The first method allows adding an existing file to the storage. The first argument receives the compression method; it can be Store or Deflate enum values. The second argument admits the physical path name, the third one allows to change the path or file name to be stored in the Zip, and the last argument inserts a comment in the storage. Notice that the folder path in the _pathname argument is not saved in the Zip file. Use the _filenameInZip argument instead to specify the folder path and filename. It can be expressed with both slashes or backslashes.

The second method allows adding data from any kind of stream object derived from the System.IO.Stream class. Internally, the first method opens a FileStream and calls the second method.

Finally, it is required to close the storage with the Close() method. This will save the central directory information too. Alternatively, the Dispose() method can be used.

Extracting stored files

For extracting a file the ExtractFile() method method should be used, like in the following minimal sample code:

using (var zip = ZipStorer.Open(@"c:\data\sample.zip", FileAccess.Read))
    if (zip.Files.TryGetValue("sample.jpg", out var zfe))
        zip.ExtractFile(zfe, @"C:\data\sample.jpg");

File and stream usage

The current release of ZipStorer supports both files and streams for creating and opening a zip storage. Several methods are overloaded for this dual support. The advantage of file-oriented methods is simplicity, since those methods will open or create files internally. On the other hand, stream-oriented methods are more flexible by allowing to manage zip storages in streams different than files. File-oriented methods will invoke internally to equivalent stream-oriented methods. Notice that not all streams will apply, because the library requires the streams to be randomly accessed (CanSeek = true). The RemoveEntries method will work only if the zip storage is a file.

    // File-oriented methods:
    public static ZipStorer Create(string _filename, string _comment)
    public static ZipStorer Open(string _filename, FileAccess _access)
    public void AddFile(Compression _method, string _pathname, string _filenameInZip, string _comment)
    public bool ExtractFile(ZipFileEntry _zfe, string _filename)

    // Stream-oriented methods:
    public static ZipStorer Create(Stream _stream, string _comment, bool _leaveOpen)
    public static ZipStorer Open(Stream _stream, FileAccess _access, bool _leaveOpen)
    public void AddStream(Compression _method, string _filenameInZip, Stream _source, DateTime _modTime, string _comment)
    public bool ExtractFile(ZipFileEntry _zfe, Stream _stream)
    public Stream LoadFile(ZipFileEntry _zfe)

    // Async method (experimental):
    public async Task<bool> ExtractFileAsync(ZipFileEntry _zfe, Stream _stream)    

The new _leaveOpen argument will prevent the stream to be closed after completing the generation of the zip package.

Filename encoding

Traditionally, the ZIP format supported DOS encoding system (a.k.a. IBM Code Page 437) for filenames in header records, which is a serious limitation for using non-occidental and even some occidental characters. Since 2007, the ZIP format specification was improved to support Unicode's UTF-8 encoding system.

ZipStorer class detects UTF-8 encoding by reading the proper flag in each file's header information. For enforcing filenames to be encoded with UTF-8 system, set the EncodeUTF8 member of ZipStorer class to true. All new filenames added will be encoded with UTF8. Notice this doesn't affect stored file contents at all. Also be aware that Windows Explorer's embedded Zip format facility does not recognize well the UTF-8 encoding system, as WinZip or WinRAR do.

.Net Standard support

Now ZipStorer supports .Net Standard 2.0 and hence a broad range of platforms.

Advantages and usage

ZipStorer has the following advantages:

• Unlike System.IO.Compression.ZipArchive, ZipStorer doesn't read the whole ZIP file into memory and write it back to disk when closing the archive. It writes all added files to the disk immediately, and keeps only the Central Directory records in memory. For example, ZipArchive requires 5 GB of memory and several minutes of time to add a 100 KB file to a 5 GB archive. ZipStorer performs the same task in 350 MB of memory and 10 seconds of time on the same computer.
• It is a short and monolithic C# class that can be embedded as source code in any project (1 source file of 37K, 800+ lines)
• Async and stream methods for extracting files (experimental)
• ZIP-64 support for working with archives larger than 4 GB or containing more than 65535 files
• UTF8 Encoding support
• Available as a nuget package