Merge pull request #460 from itn3000/tar-fix-garbled2

fix filename garbling in tar(#414)

src/SharpCompress/Common/Tar/Headers/TarHeader.cs

@@ -39,16 +39,17 @@ namespace SharpCompress.Common.Tar.Headers
             WriteOctalBytes(0, buffer, 116, 8); // group ID
 
             //ArchiveEncoding.UTF8.GetBytes("magic").CopyTo(buffer, 257);
-            if (Name.Length > 100)
+            var nameByteCount = ArchiveEncoding.GetEncoding().GetByteCount(Name);
+            if (nameByteCount > 100)
             {
                 // Set mock filename and filetype to indicate the next block is the actual name of the file
                 WriteStringBytes("././@LongLink", buffer, 0, 100);
                 buffer[156] = (byte)EntryType.LongName;
-                WriteOctalBytes(Name.Length + 1, buffer, 124, 12);
+                WriteOctalBytes(nameByteCount + 1, buffer, 124, 12);
             }
             else
             {
-                WriteStringBytes(Name, buffer, 0, 100);
+                WriteStringBytes(ArchiveEncoding.Encode(Name), buffer, 0, 100);
                 WriteOctalBytes(Size, buffer, 124, 12);
                 var time = (long)(LastModifiedTime.ToUniversalTime() - EPOCH).TotalSeconds;
                 WriteOctalBytes(time, buffer, 136, 12);
@@ -69,10 +70,17 @@ namespace SharpCompress.Common.Tar.Headers
 
             output.Write(buffer, 0, buffer.Length);
 
-            if (Name.Length > 100)
+            if (nameByteCount > 100)
             {
                 WriteLongFilenameHeader(output);
-                Name = Name.Substring(0, 100);
+                // update to short name lower than 100 - [max bytes of one character].
+                // subtracting bytes is needed because preventing infinite loop(example code is here).
+                //
+                // var bytes = Encoding.UTF8.GetBytes(new string(0x3042, 100));
+                // var truncated = Encoding.UTF8.GetBytes(Encoding.UTF8.GetString(bytes, 0, 100));
+                //
+                // and then infinite recursion is occured in WriteLongFilenameHeader because truncated.Length is 102.
+                Name = ArchiveEncoding.Decode(ArchiveEncoding.Encode(Name), 0, 100 - ArchiveEncoding.GetEncoding().GetMaxByteCount(1));
                 Write(output);
             }
         }
@@ -184,6 +192,17 @@ namespace SharpCompress.Common.Tar.Headers
             return buffer;
         }
 
+        private static void WriteStringBytes(byte[] name, byte[] buffer, int offset, int length)
+        {
+            int i = Math.Min(length, name.Length);
+            Buffer.BlockCopy(name, 0, buffer, offset, i);
+            // if Span<byte>.Fill can be used, it is more efficient
+            for (; i < length; ++i)
+            {
+                buffer[offset + i] = 0;
+            }
+        }
+
         private static void WriteStringBytes(string name, byte[] buffer, int offset, int length)
         {
             int i;

tests/SharpCompress.Test/Tar/TarArchiveTests.cs

@@ -5,6 +5,10 @@ using SharpCompress.Archives.Tar;
 using SharpCompress.Common;
 using SharpCompress.Writers;
 using Xunit;
+using System.Text;
+using SharpCompress.Readers;
+using SharpCompress.Writers.Tar;
+using SharpCompress.Readers.Tar;
 
 namespace SharpCompress.Test.Tar
 {
@@ -26,7 +30,7 @@ namespace SharpCompress.Test.Tar
         {
             ArchiveFileRead("Tar.tar");
         }
-        
+
         [Fact]
         public void Tar_FileName_Exactly_100_Characters()
         {
@@ -135,10 +139,17 @@ namespace SharpCompress.Test.Tar
             string scratchPath = Path.Combine(SCRATCH_FILES_PATH, "Tar.tar");
             string unmodified = Path.Combine(TEST_ARCHIVES_PATH, "Tar.noEmptyDirs.tar");
 
+            // var aropt = new Ar
+
             using (var archive = TarArchive.Create())
             {
                 archive.AddAllFromDirectory(ORIGINAL_FILES_PATH);
-                archive.SaveTo(scratchPath, CompressionType.None);
+                var twopt = new TarWriterOptions(CompressionType.None, true);
+                twopt.ArchiveEncoding = new ArchiveEncoding()
+                {
+                    Default = Encoding.GetEncoding(866)
+                };
+                archive.SaveTo(scratchPath, twopt);
             }
             CompareArchivesByPath(unmodified, scratchPath);
         }
@@ -195,5 +206,40 @@ namespace SharpCompress.Test.Tar
                 Assert.True(archive.Type == ArchiveType.Tar);
             }
         }
+        [Theory]
+        [InlineData(10)]
+        [InlineData(128)]
+        public void Tar_Japanese_Name(int length)
+        {
+            using (var mstm = new MemoryStream())
+            {
+                var enc = new ArchiveEncoding()
+                {
+                    Default = Encoding.UTF8
+                };
+                var twopt = new TarWriterOptions(CompressionType.None, true);
+                twopt.ArchiveEncoding = enc;
+                var fname = new string((char)0x3042, length);
+                using (var tw = new TarWriter(mstm, twopt))
+                using (var input = new MemoryStream(new byte[32]))
+                {
+                    tw.Write(fname, input, null);
+                }
+                using (var inputMemory = new MemoryStream(mstm.ToArray()))
+                {
+                    var tropt = new ReaderOptions()
+                    {
+                        ArchiveEncoding = enc
+                    };
+                    using (var tr = TarReader.Open(inputMemory, tropt))
+                    {
+                        while (tr.MoveToNextEntry())
+                        {
+                            Assert.Equal(fname, tr.Entry.Key);
+                        }
+                    }
+                }
+            }
+        }
     }
 }