Merge PR #240, Add support for POSIX Extended Headers

Fixes #121
Only "path" keyword supported as it's used for non-GNU long file names.

Author: piksel

src/ICSharpCode.SharpZipLib/Tar/TarExtendedHeaderReader.cs

@@ -0,0 +1,89 @@
+using System;
+using System.Collections.Generic;
+using System.IO;
+using System.Text;
+
+namespace ICSharpCode.SharpZipLib.Tar
+{
+	public class TarExtendedHeaderReader
+	{
+		const byte LENGTH = 0;
+		const byte KEY = 1;
+		const byte VALUE = 2;
+		const byte END = 3;
+
+		private readonly Dictionary<string, string> headers = new Dictionary<string, string>();
+
+		private string[] headerParts = new string[3];
+
+		int bbIndex;
+		private byte[] byteBuffer;
+		private char[] charBuffer;
+
+		private readonly StringBuilder sb = new StringBuilder();
+		private readonly Decoder decoder = Encoding.UTF8.GetDecoder();
+
+		private int state = LENGTH;
+
+		private static readonly byte[] StateNext = new[] { (byte)' ', (byte)'=', (byte)'\n' };
+
+		public TarExtendedHeaderReader()
+		{
+			ResetBuffers();
+		}
+
+		public void Read(byte[] buffer, int length)
+		{
+			for (int i = 0; i < length; i++)
+			{
+				byte next = buffer[i];
+
+				if (next == StateNext[state])
+				{
+					Flush();
+					headerParts[state] = sb.ToString();
+					sb.Clear();
+
+					if (++state == END)
+					{
+						headers.Add(headerParts[KEY], headerParts[VALUE]);
+						headerParts = new string[3];
+						state = LENGTH;
+					}
+				}
+				else
+				{
+					byteBuffer[bbIndex++] = next;
+					if (bbIndex == 4)
+						Flush();
+				}
+			}
+		}
+
+		private void Flush()
+		{
+			decoder.Convert(byteBuffer, 0, bbIndex, charBuffer, 0, 4, false, out int bytesUsed, out int charsUsed, out bool completed);
+
+			sb.Append(charBuffer, 0, charsUsed);
+			ResetBuffers();
+		}
+
+		private void ResetBuffers()
+		{
+			charBuffer = new char[4];
+			byteBuffer = new byte[4];
+			bbIndex = 0;
+		}
+
+
+		public Dictionary<string, string> Headers
+		{
+			get
+			{
+				// TODO: Check for invalid state? -NM 2018-07-01
+				return headers;
+			}
+		}
+
+	}
+}

src/ICSharpCode.SharpZipLib/Tar/TarInputStream.cs

@@ -434,7 +434,30 @@ public TarEntry GetNextEntry()
 						SkipToNextEntry();
 						headerBuf = this.tarBuffer.ReadBlock();
 					} else if (header.TypeFlag == TarHeader.LF_XHDR) {  // POSIX extended header
-																		// Ignore things we dont understand completely for now
+
+						byte[] nameBuffer = new byte[TarBuffer.BlockSize];
+						long numToRead = this.entrySize;
+
+						var xhr = new TarExtendedHeaderReader();
+
+						while (numToRead > 0)
+						{
+							int numRead = this.Read(nameBuffer, 0, (numToRead > nameBuffer.Length ? nameBuffer.Length : (int)numToRead));
+
+							if (numRead == -1)
+							{
+								throw new InvalidHeaderException("Failed to read long name entry");
+							}
+
+							xhr.Read(nameBuffer, numRead);
+							numToRead -= numRead;
+						}
+
+						if (xhr.Headers.TryGetValue("path", out string name))
+						{
+							longName = new StringBuilder(name);
+						}
+
 						SkipToNextEntry();
 						headerBuf = this.tarBuffer.ReadBlock();
 					} else if (header.TypeFlag == TarHeader.LF_GNU_VOLHDR) {

test/ICSharpCode.SharpZipLib.Tests/Tar/TarTests.cs

@@ -231,6 +231,58 @@ public void LongNames()
 			}
 		}
 
+		[Test]
+		[Category("Tar")]
+		public void ExtendedHeaderLongName()
+		{
+			string expectedName = "lftest-0000000000111111111122222222223333333333444444444455555555556666666666777777777788888888889999999999";
+
+			var input64 = @"Li9QYXhIZWFkZXJzLjExOTY5L2xmdGVzdC0wMDAwMDAwMDAwMTExMTExMTExMTIyMjIyMjIyMjIz
+							MzMzMzMzMzMzNDQ0NDQ0NDQ0NDU1NTU1NTU1NTU2NjY2NjY2NjY2Nzc3NzAwMDA2NDQAMDAwMDAw
+							MAAwMDAwMDAwADAwMDAwMDAwMzE3ADEzMzE2MTYyMzMzADAyMTYwNgAgeAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAB1c3RhcgAwMAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAx
+							MTcgcGF0aD1sZnRlc3QtMDAwMDAwMDAwMDExMTExMTExMTEyMjIyMjIyMjIyMzMzMzMzMzMzMzQ0
+							NDQ0NDQ0NDQ1NTU1NTU1NTU1NjY2NjY2NjY2Njc3Nzc3Nzc3Nzc4ODg4ODg4ODg4OTk5OTk5OTk5
+							OQozMCBtdGltZT0xNTMwNDU1MjU5LjcwNjU0ODg4OAozMCBhdGltZT0xNTMwNDU1MjU5LjcwNjU0
+							ODg4OAozMCBjdGltZT0xNTMwNDU1MjU5LjcwNjU0ODg4OAoAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGxm
+							dGVzdC0wMDAwMDAwMDAwMTExMTExMTExMTIyMjIyMjIyMjIzMzMzMzMzMzMzNDQ0NDQ0NDQ0NDU1
+							NTU1NTU1NTU2NjY2NjY2NjY2Nzc3Nzc3Nzc3Nzg4ODg4ODg4ODg5OTkwMDAwNjY0ADAwMDE3NTAA
+							MDAwMTc1MAAwMDAwMDAwMDAwMAAxMzMxNjE2MjMzMwAwMjM3MjcAIDAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAdXN0YXIAMDBuaWxzAAAAAAAAAAAAAAAAAAAAAAAA
+							AAAAAAAAAAAAAG5pbHMAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAMDAwMDAwMAAwMDAwMDAw";
+
+			var buffer = new byte[2560];
+			var truncated = Convert.FromBase64String(input64);
+			Array.Copy(truncated, buffer, truncated.Length);
+			truncated = null;
+
+			using (var ms = new MemoryStream(buffer))
+			using (var tis = new TarInputStream(ms))
+			{
+				var entry = tis.GetNextEntry();
+				Assert.IsNotNull(entry, "Entry is null");
+
+				Assert.IsNotNull(entry.Name, "Entry name is null");
+
+				Assert.AreEqual(expectedName.Length, entry.Name.Length, $"Entry name is truncated to {entry.Name.Length} bytes.");
+
+				Assert.AreEqual(expectedName, entry.Name, "Entry name does not match expected value");
+			}
+
+		}
+
+
 		/// <summary>
 		/// Test equals function for tar headers.
 		/// </summary>