added support for expanding tar.gz archives

Author: ayousuf23

src/ArchiveFactory.cs

@@ -34,6 +34,7 @@ internal static bool TryGetArchiveFormatFromExtension(string path, out ArchiveFo
                 ".zip" => ArchiveFormat.Zip,
                 ".tar" => ArchiveFormat.Tar,
                 ".gz" => path.EndsWith(".tar.gz") ? ArchiveFormat.Tgz : null,
+                ".tgz" => ArchiveFormat.Tgz,
                 _ => null
             };
             return archiveFormat is not null;

src/Cmdlets/ExpandArchiveCommand.cs

@@ -324,23 +324,14 @@ private string DetermineDestinationPath(IArchive archive)
             var workingDirectory = SessionState.Path.CurrentFileSystemLocation.ProviderPath;
             string? destinationDirectory = null;
 
-            // If the archive has a single top-level directory only, the destination will be: "working directory"
-            // This makes it easier for the cmdlet to expand the directory without needing addition checks
-            if (archive.HasTopLevelDirectory())
+            var filename = System.IO.Path.GetFileName(archive.Path);
+            // If filename does have an exension, remove the extension and set the filename minus extension as destinationDirectory
+            if (System.IO.Path.GetExtension(filename) != string.Empty)
             {
-                destinationDirectory = workingDirectory;
-            }
-            // Otherwise, the destination path will be: "working directory/archive file name"
-            else
-            {
-                var filename = System.IO.Path.GetFileName(archive.Path);
-                // If filename does have an exension, remove the extension and set the filename minus extension as destinationDirectory
-                if (System.IO.Path.GetExtension(filename) != string.Empty)
-                {
-                    int indexOfLastPeriod = filename.LastIndexOf('.');
-                    destinationDirectory = filename.Substring(0, indexOfLastPeriod);
-                }
+                int indexOfLastPeriod = filename.LastIndexOf('.');
+                destinationDirectory = filename.Substring(0, indexOfLastPeriod);
             }
+              
 
             if (destinationDirectory is null)
             {

src/Formats/GzipArchive.cs

@@ -12,19 +12,19 @@ namespace Microsoft.PowerShell.Archive
 {
     internal class GzipArchive : IArchive
     {
-        protected bool _disposedValue;
+        private bool _disposedValue;
 
-        protected readonly ArchiveMode _mode;
+        private readonly ArchiveMode _mode;
 
-        protected readonly string _path;
+        private readonly string _path;
 
-        protected readonly FileStream _fileStream;
+        private readonly FileStream _fileStream;
 
-        protected readonly CompressionLevel _compressionLevel;
+        private readonly CompressionLevel _compressionLevel;
 
         private bool _addedFile;
 
-        protected bool _didCallGetNextEntry;
+        private bool _didCallGetNextEntry;
 
         ArchiveMode IArchive.Mode => _mode;
 
@@ -38,7 +38,7 @@ public GzipArchive(string path, ArchiveMode mode, FileStream fileStream, Compres
             _compressionLevel = compressionLevel;
         }
 
-        public virtual void AddFileSystemEntry(ArchiveAddition entry)
+        public void AddFileSystemEntry(ArchiveAddition entry)
         {
             if (_mode == ArchiveMode.Extract)
             {
@@ -61,7 +61,7 @@ public virtual void AddFileSystemEntry(ArchiveAddition entry)
             _addedFile = true;
         }
 
-        IEntry? IArchive.GetNextEntry()
+        public IEntry? GetNextEntry()
         {
             // Gzip has no concept of entries
             if (!_didCallGetNextEntry) {
@@ -71,7 +71,7 @@ public virtual void AddFileSystemEntry(ArchiveAddition entry)
             return null;
         }
 
-        protected virtual void Dispose(bool disposing)
+        private void Dispose(bool disposing)
         {
             if (!_disposedValue)
             {
@@ -94,11 +94,6 @@ public void Dispose()
             GC.SuppressFinalize(this);
         }
 
-        bool IArchive.HasTopLevelDirectory()
-        {
-            throw new NotSupportedException();
-        }
-
         internal class GzipArchiveEntry : IEntry {
 
             private GzipArchive _gzipArchive;
@@ -116,7 +111,7 @@ public GzipArchiveEntry(GzipArchive gzipArchive)
 
             void IEntry.ExpandTo(string destinationPath)
             {
-                using var destinationFileStream = new FileStream(destinationPath, FileMode.CreateNew, FileAccess.Write, FileShare.None);
+                using var destinationFileStream = new FileStream(destinationPath, FileMode.Create, FileAccess.Write, FileShare.None);
                 using var gzipDecompressor = new GZipStream(_gzipArchive._fileStream, CompressionMode.Decompress);
                 gzipDecompressor.CopyTo(destinationFileStream);
             }

src/Formats/TarArchive.cs

@@ -2,7 +2,6 @@
 // Licensed under the MIT License.
 
 using System;
-using System.Collections.Generic;
 using System.Formats.Tar;
 using System.IO;
 using System.Diagnostics;
@@ -66,7 +65,7 @@ public void AddFileSystemEntry(ArchiveAddition entry)
             _tarWriter.WriteEntry(fileName: entry.FileSystemInfo.FullName, entryName: entryName); 
         }
 
-        IEntry? IArchive.GetNextEntry()
+        public IEntry? GetNextEntry()
         {
             // If _tarReader is null, create it
             if (_tarReader is null)
@@ -147,36 +146,6 @@ public void Dispose()
             GC.SuppressFinalize(this);
         }
 
-        bool IArchive.HasTopLevelDirectory()
-        {
-            // Go through each entry and see if it is a top-level entry
-            _tarReader = new TarReader(_fileStream, leaveOpen: true);
-
-            int topLevelDirectoriesCount = 0;
-            var entry = _tarReader.GetNextEntry();
-            while (entry is not null) {
-                
-                if (entry.EntryType == TarEntryType.Directory)
-                {
-                    topLevelDirectoriesCount++;
-                    if (topLevelDirectoriesCount > 1)
-                    {
-                        break;
-                    }
-                } else
-                {
-                    _tarReader.Dispose();
-                    _tarReader = null;
-                    return false;
-                }
-                entry = _tarReader.GetNextEntry();
-            }
-
-            _tarReader.Dispose();
-            _tarReader = null;
-            return topLevelDirectoriesCount == 1;
-        }
-
         internal class TarArchiveEntry : IEntry {
 
             // Underlying object is System.Formats.Tar.TarEntry

src/Formats/TarGzArchive.cs

@@ -2,69 +2,129 @@
 // Licensed under the MIT License.
 
 using System;
-using System.Collections.Generic;
-using System.Formats.Tar;
 using System.IO;
 using System.IO.Compression;
 using System.Diagnostics;
 
 namespace Microsoft.PowerShell.Archive
 {
-    internal class TarGzArchive : GzipArchive
+    internal class TarGzArchive : IArchive
     {
 
         // Use a tar archive because .tar.gz file is a compressed tar file
         private TarArchive? _tarArchive;
 
+        private FileStream? _tarFileStream;
+
         private string? _tarFilePath;
 
-        private FileStream? _tarFileStream;
+        private string _path;
+
+        private bool _disposedValue;
+
+        private bool _didCallGetNextEntry;
+
+        private readonly ArchiveMode _mode;
+
+        private readonly FileStream _fileStream;
+
+        private readonly CompressionLevel _compressionLevel;
+
+        public string Path => _path;
 
-        public TarGzArchive(string path, ArchiveMode mode, FileStream fileStream, CompressionLevel compressionLevel) : base(path, mode, fileStream, compressionLevel)
+        ArchiveMode IArchive.Mode => _mode;
+
+        string IArchive.Path => _path;
+
+        public TarGzArchive(string path, ArchiveMode mode, FileStream fileStream, CompressionLevel compressionLevel)
         {
+            _path = path;
+            _mode = mode;
+            _fileStream = fileStream;
+            _compressionLevel = compressionLevel;
         }
 
-        public override void AddFileSystemEntry(ArchiveAddition entry)
+        public void AddFileSystemEntry(ArchiveAddition entry)
         {
-            if (_mode == ArchiveMode.Extract || _mode == ArchiveMode.Update) {
+            if (_mode == ArchiveMode.Extract || _mode == ArchiveMode.Update)
+            {
                 throw new ArgumentException("Adding entries to the archive is not supported in extract or update mode");
             }
 
             if (_tarArchive is null)
             {
-                var outputDirectory = Path.GetDirectoryName(_path);
-                var tarFilename = Path.GetRandomFileName();
-                _tarFilePath = Path.Combine(outputDirectory, tarFilename);
-                _tarFileStream = new FileStream(_tarFilePath, FileMode.CreateNew, FileAccess.ReadWrite, FileShare.None);
+                // This will create a temp file and return the path
+                _tarFilePath = System.IO.Path.GetTempFileName();
+                // When creating the stream, the file already exists
+                _tarFileStream = new FileStream(_tarFilePath, FileMode.Open, FileAccess.ReadWrite, FileShare.None);
                 _tarArchive = new TarArchive(_tarFilePath, ArchiveMode.Create, _tarFileStream);
 
             }
             _tarArchive.AddFileSystemEntry(entry);
         }
 
-        protected override void Dispose(bool disposing)
+        public IEntry? GetNextEntry()
+        {
+            if (_mode == ArchiveMode.Create)
+            {
+                throw new ArgumentException("Getting next entry is not supported when the archive is in Create mode");
+            }
+
+            if (_tarArchive is null)
+            {
+                // Create a Gzip archive
+                using var gzipArchive = new GzipArchive(_path, _mode, _fileStream, _compressionLevel);
+                // Where to put the tar file when expanding the tar.gz archive
+                _tarFilePath = System.IO.Path.GetTempFileName();
+                // Expand the gzip portion
+                var entry = gzipArchive.GetNextEntry();
+                Debug.Assert(entry is not null);
+                entry.ExpandTo(_tarFilePath);
+                // Create a TarArchive pointing to the newly expanded out tar file from the tar.gz file
+                FileStream tarFileStream = new FileStream(_tarFilePath, FileMode.Open, FileAccess.ReadWrite, FileShare.None);
+                _tarArchive = new TarArchive(_tarFilePath, ArchiveMode.Extract, tarFileStream);
+            }
+            return _tarArchive?.GetNextEntry();
+        }
+
+        public void Dispose()
+        {
+            // Do not change this code. Put cleanup code in 'Dispose(bool disposing)' method
+            Dispose(disposing: true);
+            GC.SuppressFinalize(this);
+        }
+
+        // Performs gzip compression on _path
+        private void CompressArchive() {
+            Debug.Assert(_tarFilePath is not null);
+            _tarFileStream = new FileStream(_tarFilePath, FileMode.Open, FileAccess.Read);
+            using var gzipCompressor = new GZipStream(_fileStream, _compressionLevel, true);
+            _tarFileStream.CopyTo(gzipCompressor);
+            _tarFileStream.Dispose();
+        }
+
+        private void Dispose(bool disposing)
         {
             if (!_disposedValue)
             {
                 if (disposing)
                 {
-                    // TODO: dispose managed state (managed objects)
+                    // Do this before compression because disposing a tar archive will add necessary EOF markers
+                    _tarArchive?.Dispose();
+                    if (_mode == ArchiveMode.Create) {
+                        CompressArchive();
+                    }
                     _fileStream.Dispose();
-                    CompressArchive();
+                    if (_tarFilePath is not null) {
+                        // Delete the tar file created in the process of created the tar.gz file
+                        File.Delete(_tarFilePath);
+                    }
                 }
 
                 // TODO: free unmanaged resources (unmanaged objects) and override finalizer
                 // TODO: set large fields to null
                 _disposedValue = true;
             }
         }
-
-        // Performs gzip compression on _path
-        private void CompressArchive() {
-            Debug.Assert(_tarFileStream is not null);
-            _tarFileStream.Position = 0;
-            using var gzipCompressor = new GZipStream(_fileStream, _compressionLevel, true);
-            _tarFileStream.CopyTo(gzipCompressor);
-        }
     }
 }

src/Formats/ZipArchive.cs

@@ -151,28 +151,6 @@ public void Dispose()
             GC.SuppressFinalize(this);
         }
 
-        bool IArchive.HasTopLevelDirectory()
-        {
-            int topLevelDirectoriesCount = 0;
-            foreach (var entry in _zipArchive.Entries)
-            {
-                if (entry.FullName.EndsWith(ZipArchiveDirectoryPathTerminator) &&
-                    entry.FullName.LastIndexOf(ZipArchiveDirectoryPathTerminator, entry.FullName.Length - 2) == -1)
-                {
-                    topLevelDirectoriesCount++;
-                    if (topLevelDirectoriesCount > 1)
-                    {
-                        break;
-                    }
-                } else
-                {
-                    return false;
-                }
-            }
-
-            return topLevelDirectoriesCount == 1;
-        }
-
         internal class ZipArchiveEntry : IEntry
         {
             // Underlying object is System.IO.Compression.ZipArchiveEntry

src/IArchive.cs

@@ -21,8 +21,5 @@ interface IArchive: IDisposable
         public void AddFileSystemEntry(ArchiveAddition entry);
 
         public IEntry? GetNextEntry();
-
-        // Does the archive have only a top-level directory?
-        public bool HasTopLevelDirectory();
     }
 }