From bd40ca6d9df68cfa46c21c6b3900d1733bf8fbd0 Mon Sep 17 00:00:00 2001
From: Matt Nadareski <mnadareski@outlook.com>
Date: Thu, 23 Jun 2022 13:58:48 -0700
Subject: [PATCH] Use OpenMcdf for MSI

---
 BurnOutSharp/BurnOutSharp.csproj |   4 +-
 BurnOutSharp/FileType/MSI.cs     | 100 ++++++++++++++++++++++++++++++-
 README.md                        |   3 +-
 3 files changed, 100 insertions(+), 7 deletions(-)

diff --git a/BurnOutSharp/BurnOutSharp.csproj b/BurnOutSharp/BurnOutSharp.csproj
index 88d59127..11fb4a17 100644
--- a/BurnOutSharp/BurnOutSharp.csproj
+++ b/BurnOutSharp/BurnOutSharp.csproj
@@ -22,6 +22,7 @@
   </PropertyGroup>
 
   <ItemGroup>
+    <PackageReference Include="OpenMcdf" Version="2.2.1.9" />
     <PackageReference Include="SharpCompress" Version="0.32.1" />
     <PackageReference Include="System.Text.Encoding.CodePages" Version="6.0.0" />
     <PackageReference Include="Teronis.MSBuild.Packaging.ProjectBuildInPackage" Version="1.0.0">
@@ -58,9 +59,6 @@
   <ItemGroup>
     <ProjectReference Include="..\HLLibSharp\HLLibSharp\HLLibSharp.csproj" />
     <ProjectReference Include="..\LibMSPackSharp\LibMSPackSharp\LibMSPackSharp.csproj" />
-    <ProjectReference Include="..\WixToolset\src\WixToolset.Dtf.WindowsInstaller\WixToolset.Dtf.WindowsInstaller.csproj">
-      <PrivateAssets>all</PrivateAssets>
-    </ProjectReference>
   </ItemGroup>
 
 </Project>
diff --git a/BurnOutSharp/FileType/MSI.cs b/BurnOutSharp/FileType/MSI.cs
index e49475f2..2307ff0f 100644
--- a/BurnOutSharp/FileType/MSI.cs
+++ b/BurnOutSharp/FileType/MSI.cs
@@ -1,9 +1,10 @@
 using System;
 using System.Collections.Concurrent;
 using System.IO;
+using System.Text;
 using BurnOutSharp.Interfaces;
 using BurnOutSharp.Tools;
-using WixToolset.Dtf.WindowsInstaller;
+using OpenMcdf;
 
 namespace BurnOutSharp.FileType
 {
@@ -40,9 +41,39 @@ namespace BurnOutSharp.FileType
                 string tempPath = Path.Combine(Path.GetTempPath(), Guid.NewGuid().ToString());
                 Directory.CreateDirectory(tempPath);
 
-                using (Database msidb = new Database(file, DatabaseOpenMode.ReadOnly))
+                using (CompoundFile msi = new CompoundFile(stream, CFSUpdateMode.ReadOnly, CFSConfiguration.Default))
                 {
-                    msidb.ExportAll(tempPath);
+                    msi.RootStorage.VisitEntries((e) =>
+                    {
+                        if (!e.IsStream)
+                            return;
+
+                        var str = msi.RootStorage.GetStream(e.Name);
+                        if (str == null)
+                            return;
+
+                        byte[] strData = str.GetData();
+                        if (strData == null)
+                            return;
+
+                        string decoded = DecodeStreamName(e.Name).TrimEnd('\0');
+                        byte[] nameBytes = Encoding.UTF8.GetBytes(e.Name);
+
+                        // UTF-8 encoding of 0x4840.
+                        if (nameBytes[0] == 0xe4 && nameBytes[1] == 0xa1 && nameBytes[2] == 0x80)
+                            decoded = decoded.Substring(3);
+
+                        foreach (char c in Path.GetInvalidFileNameChars())
+                        {
+                            decoded = decoded.Replace(c, '_');
+                        }
+
+                        string filename = Path.Combine(tempPath, decoded);
+                        using (Stream fs = File.OpenWrite(filename))
+                        {
+                            fs.Write(strData, 0, strData.Length);
+                        }
+                    }, recursive: true);
                 }
 
                 // Collect and format all found protections
@@ -70,5 +101,68 @@ namespace BurnOutSharp.FileType
 
             return null;
         }
+
+        /// <remarks>Adapted from LibMSI</remarks>
+        private static string DecodeStreamName(string input)
+        {
+            if (input == null)
+                return null;
+
+            int count = 0;
+            byte[] inputBytes = Encoding.UTF8.GetBytes(input);
+            int p = 0; // inputBytes[0]
+
+            byte[] output = new byte[inputBytes.Length + 1];
+            int q = 0; // output[0]
+            while (p < inputBytes.Length && inputBytes[p] != 0)
+            {
+                int ch = inputBytes[p];
+                if ((ch == 0xe3 && inputBytes[p + 1] >= 0xa0) || (ch == 0xe4 && inputBytes[p + 1] < 0xa0))
+                {
+                    // UTF-8 encoding of 0x3800..0x47ff. 
+                    output[q++] = (byte)Mime2Utf(inputBytes[p + 2] & 0x7f);
+                    output[q++] = (byte)Mime2Utf(inputBytes[p + 1] ^ 0xa0);
+                    p += 3;
+                    count += 2;
+                    continue;
+                }
+
+                if (ch == 0xe4 && inputBytes[p + 1] == 0xa0)
+                {
+                    // UTF-8 encoding of 0x4800..0x483f.
+                    output[q++] = (byte)Mime2Utf(inputBytes[p + 2] & 0x7f);
+                    p += 3;
+                    count++;
+                    continue;
+                }
+
+                output[q++] = inputBytes[p++];
+                if (ch >= 0xc1)
+                    output[q++] = inputBytes[p++];
+                if (ch >= 0xe0)
+                    output[q++] = inputBytes[p++];
+                if (ch >= 0xf0)
+                    output[q++] = inputBytes[p++];
+
+                count++;
+            }
+
+            output[q] = 0;
+            return Encoding.ASCII.GetString(output);
+        }
+
+        /// <remarks>Adapted from LibMSI</remarks>
+        private static int Mime2Utf(int x)
+        {
+            if (x < 10)
+                return x + '0';
+            if (x < (10 + 26))
+                return x - 10 + 'A';
+            if (x < (10 + 26 + 26))
+                return x - 10 - 26 + 'a';
+            if (x == (10 + 26 + 26))
+                return '.';
+            return '_';
+        }
     }
 }
diff --git a/README.md b/README.md
index 5f058050..09d2cb11 100644
--- a/README.md
+++ b/README.md
@@ -7,12 +7,13 @@ C# port of the protection scanning ability of [BurnOut](http://burnout.sourcefor
 In addition to the original BurnOut code, the following libraries (or ports thereof) are used for file handling:
 
 - [HLLibSharp](https://github.com/mnadareski/HLLibSharp) - Various Valve archive format extraction
+- [LibMSPackSharp](https://github.com/mnadareski/LibMSPackSharp) - Microsoft CAB extraction
+- [openmcdf](https://github.com/ironfede/openmcdf) - MSI extraction
 - [psxt001z](https://github.com/Dremora/psxt001z) - PS1 LibCrypt detection
 - [SharpCompress](https://github.com/adamhathcock/sharpcompress) - 7zip/GZip/RAR/PKZIP extraction
 - [StormLibSharp](https://github.com/robpaveza/stormlibsharp) - MPQ extraction
 - [UnshieldSharp](https://github.com/mnadareski/UnshieldSharp) - InstallShield CAB extraction
 - [WiseUnpacker](https://github.com/mnadareski/WiseUnpacker) - Wise Installer extraction
-- [WixToolset.Dtf](https://github.com/wixtoolset/Dtf) - MSI and Microsoft CAB extraction
 
 Please note that due to current library limitations, the functionality of StormLibSharp and WixToolset.Dtf are locked to Windows only.