interpreter VS jmurmel

Here you go: https://github.com/bamless/interpreter. Look at the interpret/memenv package. The implementation really couldn't be more simple: it is just a queue of frame objects. pushStackFrame is called when calling a function, and popStackFrame when hitting a return or a function end. The MemoryEnvironment is instantiated by the interpreter visitor so all parts of it have access to the memory. The variables themselves are kept in an hashmap inside each Frame. Not the most efficient way but fast enough for a tree walking interpreter

https://github.com/bamless/interpreter/tree/master.

See also format.lisp for what I have so far. This will also run with sbcl and/ or abcl. If you've made it this far I'd also appreciate feedback on whether my chosen subset (see the comment at the top of the file) of Common-Lisp's format is somewhat useful and/ or which features you would miss the most.

JMurmel also has commandline flags to turn off language features for experimentation purposes, see e.g. implementing cons, car and cdr in Lambda Calculus.

To my defense, I have started my Lisp compiler/ interpreter mostly for recreational purposes to do the exact opposite of what the checkstyle nazis at my $job demand.

Murmel:

I may be a little late to the party but here's my Murmel solution:

Code is on Github, the latest release with a precompiled jar is at Release V 1.3.

Re: automated tests; there are JUnit tests in https://github.com/mayerrobert/jmurmel/tree/master/lambda/src/test/java, and the files in https://github.com/mayerrobert/jmurmel/tree/master/lambda/src/test/lisp are run automatically, too, and their output and result is checked. Maybe I should add a file HACKING.md or something that gives an intro of the project structure and build system?