DKorablin/ByteCodeReader: JVM ClassFile format reader · GitHub

A managed .NET library for parsing Java Virtual Machine (JVM) class (*.class) files. It provides strongly typed access to the full constant pool, fields, methods, interfaces and all attribute blocks defined in the Java Class File Format (JVM Spec §4).

• .NET Framework 2.0
• .NET Standard 2.0 (compatible with .NET 8 and earlier)

• Zero external native dependencies
• Lazy constant pool parsing (constants are materialized as needed)
• Full fidelity access to raw attribute data
• Works with very old and modern class file versions

Package manager:

Install-Package AlphaOmega.ByteCodeReader

dotnet CLI:

dotnet add package AlphaOmega.ByteCodeReader

using System;
using AlphaOmega.Debug;

using(ClassFile cls = new ClassFile(StreamLoader.FromFile(@"C:\Test\MyClass.class")))
{
    if(!cls.IsValid)
        throw new InvalidOperationException("Invalid JVM class file");

    Console.WriteLine($"Class: {cls.ThisClass.Bytes}");
    Console.WriteLine($"Super: {cls.SuperClass?.Bytes ?? "<Object>"}");
    Console.WriteLine($"Version: {cls.MajorVersion}.{cls.MinorVersion}");

    // Fields
    foreach(var f in cls.Fields)
        Console.WriteLine($"Field: {f.Name.Bytes} {f.Descriptor.Bytes}");

    // Methods
    foreach(var m in cls.Methods)
        Console.WriteLine($"Method: {m.Name.Bytes}{m.Descriptor.Bytes}");

    // Constant Pool (example: list all Utf8 entries)
    foreach(var utf8 in cls.ConstantPool.Utf8)
        Console.WriteLine($"Utf8[{utf8.Index}]: {utf8.Bytes}");
}

• ClassFile: Root parser for one JVM class / interface file.
• ConstantTables: Provides typed access to each constant pool table (Class, Fieldref, Methodref, InterfaceMethodref, String, Integer, Float, Long, Double, NameAndType, Utf8, MethodHandle, MethodType, InvokeDynamic).
• AttributeTables / AttributePool: Unified access to attributes from class, fields, and methods.
• Field_Info / MethodInfo: High-level wrappers exposing metadata and per-member attributes.

The reader currently handles:

• Header (magic, versioning)
• ConstantPool (all spec tags including invokedynamic)
• ThisClass / SuperClass
• Interfaces
• Fields (field_info)
• Methods (method_info)
• Attributes (attribute_info) at class, field, and method level
• Aggregated AttributePool (all attributes from all structures)

Common attribute types recognized (JVM Spec §4.7):

• InnerClasses
• EnclosingMethod
• Synthetic
• Signature
• SourceFile
• SourceDebugExtension
• Deprecated
• RuntimeVisibleAnnotations
• RuntimeInvisibleAnnotations
• BootstrapMethods

Unrecognized attributes are preserved (spec requires silent ignore) so tooling can still inspect raw data.

You can access raw constant pool bytes or attribute blocks for advanced analysis or re-emission.

Byte[] rawConstants = cls.ConstantPool.GetData();

• Invalid magic number throws ArgumentException.
• Use ClassFile.IsValid for a quick check after constructing.

• Constant pool rows are wrapped in lightweight objects; repeated access reuses instances.
• Long / Double constants advance index per JVM spec (extra unusable slot handled internally).

• Does not modify or write class files (read-only).
• StackMapTable, Module, Record attributes (newer JVM versions) may need extension.
• Bytecode instruction stream decoding (Code attribute body) can be added.

Planned:

• Detailed Code attribute parser (instructions, exception table, line numbers)
• CLI tool for quick inspection
• Optional rewrite / emit API

Class file version exposed via ClassFile.Version, MajorVersion, MinorVersion (e.g. 61.0 for Java 17).

Only parses local byte arrays / streams. No code execution or dynamic loading.

• JVM Class File Format (Java SE 8+): https://docs.oracle.com/javase/specs/jvms/se8/html/jvms-4.html

Feedback and improvements welcome.