Add tests for decompressing/extracting archives

src/BizHawk.Client.Common/SharpCompressArchiveFile.cs

@@ -24,6 +24,8 @@ namespace BizHawk.Client.Common
 
 		public SharpCompressArchiveFile(string path) => _archive = ArchiveFactory.Open(path, new());
 
+		public SharpCompressArchiveFile(Stream fileStream) => _archive = ArchiveFactory.Open(fileStream, new());
+
 		public void Dispose()
 		{
 			if (_archive == null) throw new ObjectDisposedException(nameof(SharpCompressArchiveFile));

src/BizHawk.Client.Common/SharpCompressDearchivalMethod.cs

@@ -49,8 +49,40 @@ namespace BizHawk.Client.Common
 			return false;
 		}
 
+		public bool CheckSignature(Stream fileStream, string? filenameHint)
+		{
+			if (!fileStream.CanRead || !fileStream.CanSeek) return false;
+
+			try
+			{
+				using var arcTest = ArchiveFactory.Open(fileStream); // should throw for non-archives
+				if (arcTest.Type != ArchiveType.Tar) return true; // not expecting false positives from anything but .tar for now
+			}
+			catch
+			{
+				return false;
+			}
+
+			// as above, SharpCompress seems to overzealously flag files it thinks are the in original .tar format, so we'll check for false positives
+
+			if (fileStream.Length < 512) return false;
+			// looking for magic bytes
+			var seekPos = fileStream.Position;
+			fileStream.Seek(0x101, SeekOrigin.Begin);
+			var buffer = new byte[8];
+			fileStream.Read(buffer, 0, 8);
+			fileStream.Seek(seekPos, SeekOrigin.Begin);
+			var s = buffer.BytesToHexString();
+			if (s == "7573746172003030" || s == "7573746172202000") return true; // "ustar\000" (libarchive's bsdtar) or "ustar  \0" (GNU Tar)
+
+			Console.WriteLine($"SharpCompress identified file in stream as original .tar format, probably a false positive, ignoring. Filename hint: {filenameHint}");
+			return false;
+		}
+
 		public SharpCompressArchiveFile Construct(string path) => new(path);
 
+		public SharpCompressArchiveFile Construct(Stream fileStream) => new(fileStream);
+
 		public static readonly SharpCompressDearchivalMethod Instance = new();
 
 		public IReadOnlyCollection<string> AllowedArchiveExtensions { get; } = new[]

src/BizHawk.Common/HawkFile/IFileDearchivalMethod.cs

@@ -1,4 +1,5 @@
 using System.Collections.Generic;
+using System.IO;
 
 namespace BizHawk.Common
 {
@@ -8,8 +9,14 @@ namespace BizHawk.Common
 		/// <remarks>TODO could this receive a <see cref="HawkFile"/> itself? possibly handy, in very clever scenarios of mounting fake files</remarks>
 		bool CheckSignature(string fileName, out int offset, out bool isExecutable);
 
+		/// <remarks>for now, only used in tests</remarks>
+		bool CheckSignature(Stream fileStream, string? filenameHint = null);
+
 		IReadOnlyCollection<string> AllowedArchiveExtensions { get; }
 
 		T Construct(string path);
+
+		/// <remarks>for now, only used in tests</remarks>
+		T Construct(Stream fileStream);
 	}
 }

src/BizHawk.Tests/BizHawk.Tests.csproj

@@ -14,4 +14,7 @@
 		<PackageReference Include="JunitXml.TestLogger" Version="2.1.78" PrivateAssets="all" />
 		<ProjectReference Include="$(ProjectDir)../BizHawk.Client.Common/BizHawk.Client.Common.csproj" />
 	</ItemGroup>
+	<ItemGroup>
+		<EmbeddedResource Include="data/**/*" />
+	</ItemGroup>
 </Project>

src/BizHawk.Tests/Client.Common/dearchive/DearchivalTests.cs

@@ -0,0 +1,53 @@
+using System;
+using System.IO;
+using System.Linq;
+
+using BizHawk.Client.Common;
+using BizHawk.Common.BufferExtensions;
+using BizHawk.Common.IOExtensions;
+
+using Microsoft.VisualStudio.TestTools.UnitTesting;
+
+namespace BizHawk.Tests.Client.Common.Dearchive
+{
+	[TestClass]
+	public class DearchivalTests
+	{
+		private const string EMBED_GROUP = "dearchive";
+
+		private static readonly (string Filename, bool HasSharpCompressSupport)[] TestCases = {
+			("m3_scy_change.7z", true),
+			("m3_scy_change.gb.gz", true),
+			("m3_scy_change.rar", true),
+			("m3_scy_change.bsdtar.tar", true),
+			("m3_scy_change.gnutar.tar", true),
+			("m3_scy_change.zip", true),
+		};
+
+		private readonly Lazy<byte[]> _rom = new(() => EmbeddedData.GetStream(EMBED_GROUP, "m3_scy_change.gb").ReadAllBytes());
+
+		private byte[] Rom => _rom.Value;
+
+		[TestMethod]
+		public void SanityCheck() => Assert.AreEqual("70DCA8E791878BDD32426391E4233EA52B47CDD1", Rom.HashSHA1());
+
+		[TestMethod]
+		public void TestSharpCompress()
+		{
+			var sc = SharpCompressDearchivalMethod.Instance;
+			foreach (var filename in TestCases.Where(testCase => testCase.HasSharpCompressSupport)
+				.Select(testCase => testCase.Filename))
+			{
+				var archive = EmbeddedData.GetStream(EMBED_GROUP, filename);
+				Assert.IsTrue(sc.CheckSignature(archive, filename), $"{filename} is an archive, but wasn't detected as such"); // puts the seek pos of the Stream param back where it was (in this case at the start), but that may not always be true
+				var af = sc.Construct(archive);
+				var items = af.Scan();
+				Assert.IsNotNull(items, $"{filename} contains 1 file, but it couldn't be enumerated correctly");
+				Assert.AreEqual(1, items!.Count, $"{filename} contains 1 file, but was detected as containing {items.Count} files");
+				using MemoryStream ms = new((int) items[0].Size);
+				af.ExtractFile(items[0].ArchiveIndex, ms);
+//				Assert.IsTrue(ms.ReadAllBytes().SequenceEqual(Rom), $"the file extracted from {filename} doesn't match the uncompressed file"); //TODO less dumb way of doing this? also it doesn't work
+			}
+		}
+	}
+}

src/BizHawk.Tests/EmbeddedData.cs

@@ -0,0 +1,17 @@
+using System;
+using System.IO;
+using System.Reflection;
+
+namespace BizHawk.Tests
+{
+	public static class EmbeddedData
+	{
+		private static readonly Assembly Asm = typeof(EmbeddedData).Assembly;
+
+		public static Stream GetStream(string group, string embedPath)
+		{
+			var fullPath = $"BizHawk.Tests.data.{group}.{embedPath}";
+			return Asm.GetManifestResourceStream(fullPath) ?? throw new InvalidOperationException($"Could not find the embedded resource {fullPath}");
+		}
+	}
+}

src/BizHawk.Tests/data/dearchive/readme.md

@@ -0,0 +1,2 @@
+All files in this dir are derived from `m3_scy_change.gb`, which is an open source test rom from the [Mealybug Tearoom Tests](https://github.com/mattcurrie/mealybug-tearoom-tests) (specifically, it's the copy from [this bundle](https://github.com/mattcurrie/mealybug-tearoom-tests/blob/bc3430a7c1f5d394764f37a189b85d8ce4bb3c4e/mealybug-tearoom-tests.zip)).
+The test suite is made available under the MIT (Expat) license, and the full terms of the license can be found [here](https://github.com/mattcurrie/mealybug-tearoom-tests/blob/bc3430a7c1f5d394764f37a189b85d8ce4bb3c4e/LICENSE).