I have created several hand written compilers for very simple languages but now I want to try my hand at developing a dynamic language, similar to a simplified Python or Ruby. However, it was easy for me to wrap my head around how compilers work. Primitive compilers just translate. But I can't do this if the language is dynamic. I have to write an interpreter or VM that keeps track of information at runtime and puts a lot more work on me.
In short, are there any resources I should check out considering I know how compilers work but want to migrate to creating an interpreter? There are a few VMs out there for dynamic languages, but I have no problem with rolling my own. This is all just for my personal experience.
I am seeking information on how to go from a compiler to an interpreter. If I have already made a compiler for language X but now what to write an interpreter, what needs to be done and are there any resources that go over the process?
I do not want broad or abstract resources that go over how compilers or virtual machines work. I have plenty of textbooks on the subject. All of the resources I found online either assume you have 0 experience and thus start you off with lexical or syntactic analysis or they are extremely abstract. I have a working compiler, but I now want to turn this into an interpreter and add dynamic features to the language.
I could not find resources on this process, it may be too limited in scope, or resources on the "back end" of an interpreter without being too theoretical which is why I posted here.
Best Answer
First learn about implementing interpreters. I recommend PLAI (Programming Languages: Application and Interpretation). It gets to the meat of interpretation quickly without dwelling overlong on syntax.
For your language, you'll be able to reuse the compiler's front-end (parser, mostly) and run-time library (GC, data structures, primitive operations, etc).
Of course, you can also implement a dynamic language with a compiler that produces code that manipulates (some of) the same data structures that you would use in an interpreter. For example, in an interpreter you might implement global variables as a string-indexed hash table. In a compiler, you would compile global variable references into the code that does the lookup using the same table. In contrast, you could compile lexical variables into a more efficient representation ("native" arguments and closure structure references).