Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 2 years ago.
Is there a nice place for learning the JVM bytecode instruction set. The specification perhaps and maybe some tutorials?
I ask because I would like to design a toy language and a compiler for it that generates JVM bytecode.
Thanks for your knowledge and perhaps googling.
A good reference for Java bytecode specification is the The Java Virtual Machine Specification.
See Chapter 4. The class File Format and Chapter 6. The Java Virtual Machine Instruction Set.
A little more "graphic" explanation, IBM developer works: Understanding bytecode makes you a better programmer.
Also useful are the javap disassembler and bytecode manipulation frameworks like ASM and BCEL, even if all you want to do is verify your classes.
The book Programming for the Java Virtual Machine explains the JVM instruction set and how to write code for it. It also introduces a bytecode assembler called Oolong, which I have not been able to download. You can, however, use Jasmin, the predecessor of Oolong. Essentially, you write a text file with instructions and Jasmin will spit out a .class file. The book was published in 1999, but it is still a good and gentle introduction to the VM.
The Java Virtual Machine Specification is a good place to start.
See Chapter 4 The class File Format. The updates cover new attributes added since the 2nd edition was made.
To start with, I suggest generating Java code from your language.
This will make reading and debugging much simpler.
A new version of the specification has been released:
The Java Virtual Machine Specification, Java SE 7 Edition
This new addition does exclude the chapter on Java Programming Language Concepts which folks may or may not find useful. See here for that chapter in the Second Edition of the spec.
Perhaps check out Preon's example on how to parse a Java class file. It has a fairly complete representation of the bytecode in a Java object model.
This is a bit more specialized, but here is an on-line presentation on how to optimize generated bytecode for running on the JVM. It was presented at the recent JVM Languages Summit conferences. InfoQ has a collection of presentations from that conference which might be of help to someone wanting to bring up a language on the JVM (or to see what's already been done).
