Add extraction for installshield executables. (#27)

* Add extractor for installshield executables.

* Add comment

* Fix intendation.

* Ensure PRs recursively pull

* Rebase

* Fix formatting for parsex

* newline before obj.name

* Changed while loop check to a position vs length check

* Specify type for chunkSize

* declare chunksize as a const int outside of the loop

* Open file writer with using

* Call fs.flush after every write.

* Improved chunk writing loop.

* Remove now-unecessary variable

* Properly handle extracting files of size greater than int32.

* Use long for overlay address

* create deserializer outside of loop.

* Move null check.

* Don't cache position.

---------

Co-authored-by: Matt Nadareski <mnadareski@outlook.com>

SabreTools.Serialization/Models/InstallShieldExecutable/ExtractableFile.cs

@@ -0,0 +1,25 @@
+namespace SabreTools.Data.Models.InstallShieldExecutable
+{
+    public class ExtractableFile
+    {
+        /// <summary>
+        /// Name of the file, only ASCII characters(?)
+        /// </summary>
+        public string? Name { get; set; }
+
+        /// <summary>
+        /// Path of the file, only ASCII characters(?), seems to usually use \ filepaths
+        /// </summary>
+        public string? Path  { get; set; }
+
+        /// <summary>
+        /// Version of the file
+        /// </summary>
+        public string? Version { get; set; }
+
+        /// <summary>
+        /// Length of the file. Stored in the installshield executable as a string.
+        /// </summary>
+        public ulong Length { get; set; }
+    }
+}

SabreTools.Serialization/Readers/InstallShieldExecutableFile.cs

@@ -0,0 +1,66 @@
+using System.IO;
+using SabreTools.Data.Models.InstallShieldExecutable;
+using SabreTools.IO.Extensions;
+
+namespace SabreTools.Serialization.Readers
+{
+    public class InstallShieldExecutableFile : BaseBinaryReader<ExtractableFile>
+    {
+        public override ExtractableFile? Deserialize(Stream? data)
+        {
+            // If the data is invalid
+            if (data == null || !data.CanRead)
+                return null;
+
+            try
+            {
+                // Cache the initial offset
+                long initialOffset = data.Position;
+
+                // Try to parse the header
+                var header = ParseExtractableFileHeader(data);
+                if (header == null)
+                    return null;
+
+                return header;
+            }
+            catch
+            {
+                // Ignore the actual error
+                return null;
+            }        
+        }
+        
+        /// <summary>
+        /// Parse a Stream into an InstallShield Executable file header
+        /// </summary>
+        /// <param name="data">Stream to parse</param>
+        /// <returns>Filled InstallShield Executable file header on success, null on error</returns>
+        public static ExtractableFile? ParseExtractableFileHeader(Stream? data)
+        {
+            var obj = new ExtractableFile();
+            
+            obj.Name = data.ReadNullTerminatedAnsiString();
+            if (obj.Name == null)
+                return null;
+        
+            obj.Path = data.ReadNullTerminatedAnsiString();
+            if (obj.Path == null)
+                return null;
+        
+            obj.Version = data.ReadNullTerminatedAnsiString();
+            if (obj.Version == null)
+                return null;
+        
+            var versionString = data.ReadNullTerminatedAnsiString();
+            if (versionString == null || !ulong.TryParse(versionString, out var foundVersion))
+                return null;
+        
+            obj.Length = foundVersion;
+            if (obj.Name == null)
+                return null;
+        
+            return obj;
+        }
+    }
+}

SabreTools.Serialization/Wrappers/PortableExecutable.Extraction.cs

@@ -18,6 +18,7 @@ namespace SabreTools.Serialization.Wrappers
         /// - SFX archives
         ///     + 7z
         ///     + Advanced Installer
+        ///     + InstallShield Executables
         ///     + MS-CAB
         ///     + PKZIP
         ///     + RAR
@@ -28,6 +29,7 @@ namespace SabreTools.Serialization.Wrappers
         {
             bool cai = ExtractAdvancedInstaller(outputDirectory, includeDebug);
             bool cexe = ExtractCExe(outputDirectory, includeDebug);
+            bool issexe = ExtractInstallShieldExecutable(outputDirectory, includeDebug);
             bool matroschka = ExtractMatroschka(outputDirectory, includeDebug);
             bool resources = ExtractFromResources(outputDirectory, includeDebug);
             bool spoon = ExtractSpoonInstaller(outputDirectory, includeDebug);
@@ -37,10 +39,10 @@ namespace SabreTools.Serialization.Wrappers
             bool wiseSection = wiseOverlay || ExtractWiseSection(outputDirectory, includeDebug);
 
             // Overlay can be skipped in some situations
-            bool overlay = cai || spoon || wiseOverlay
+            bool overlay = cai || issexe || spoon || wiseOverlay
                 || ExtractFromOverlay(outputDirectory, includeDebug);
 
-            return cai || cexe || matroschka || overlay || resources || spoon
+            return cai || cexe || issexe || matroschka || overlay || resources || spoon
                 || wiseOverlay || wiseSection;
         }
 
@@ -158,6 +160,85 @@ namespace SabreTools.Serialization.Wrappers
                 return false;
             }
         }
+        
+        /// <summary>
+        /// Extract data from an InstallShield Executable
+        /// </summary>
+        /// <param name="outputDirectory">Output directory to write to</param>
+        /// <param name="includeDebug">True to include debug data, false otherwise</param>
+        /// <returns>True if extraction succeeded, false otherwise</returns>
+        public bool ExtractInstallShieldExecutable(string outputDirectory, bool includeDebug)
+        {
+            try
+            {
+                long overlayAddress = OverlayAddress;
+                
+                // Return if overlay doesn't exist.
+                if (overlayAddress == -1) 
+                    return false;
+                
+                // Ensure the stream is starting at the overlay address
+                _dataSource.Seek(overlayAddress, SeekOrigin.Begin);
+
+                var streamLength = _dataSource.Length;
+                const int chunkSize = 65536;
+                var deserializer = new Readers.InstallShieldExecutableFile();
+                
+                while (_dataSource.Position < streamLength)
+                {
+                    lock (_dataSourceLock)
+                    {
+                        // Try to deserialize the source data
+                        
+                        var entry = deserializer.Deserialize(_dataSource);
+                        if (entry?.Path == null)
+                            return false;
+                    
+                        // Get the length, and make sure it won't EOF
+                        var length = (long)entry.Length;
+                        if (length > streamLength - _dataSource.Position)
+                            break;
+
+                        // Ensure directory separators are consistent
+                        // Path is used instead of Name because Path contains the filename anyways.
+                    
+                        var filename = entry.Path.TrimEnd('\0');
+                        if (Path.DirectorySeparatorChar == '\\')
+                            filename = filename.Replace('/', '\\');
+                        else if (Path.DirectorySeparatorChar == '/')
+                            filename = filename.Replace('\\', '/');
+
+                        // Ensure the full output directory exists
+                        filename = Path.Combine(outputDirectory, filename);
+                        var directoryName = Path.GetDirectoryName(filename);
+                        if (directoryName != null && !Directory.Exists(directoryName))
+                            Directory.CreateDirectory(directoryName);
+
+                        // Write the output file
+                        using var fs = File.Open(filename, FileMode.Create, FileAccess.Write, FileShare.ReadWrite);
+                        Console.WriteLine($"Attempting to extract {entry.Name} from potential InstallShield Executable");
+
+                        // Read from file in chunks in order to save memory, since some extracted files will be large
+                        // Chunk size is purely arbitrary and can be adjusted as needed.
+                        // Read file from InstallShield Executable and write it as an output file.
+                        while (length > 0)
+                        {
+                            var bytesToRead = (int)Math.Min(length, chunkSize);
+                            var buffer = _dataSource.ReadBytes(bytesToRead);
+                            fs.Write(buffer, 0, bytesToRead);
+                            fs.Flush();
+                            length -= bytesToRead;
+                        }
+                    }
+                }
+                return true;
+            }
+            catch (Exception ex)
+            {
+                if (includeDebug) Console.Error.WriteLine(ex);
+                return false;
+            }
+        }
 
         /// <summary>
         /// Extract data from the overlay