mark-sweep VS c-examples

Seconding the Make-A-Lisp and Crafting Interpreters recommendations. This post is also really great, it contains a hard written parser that you may learn from.

You mentioned you want to write your lisp in javascript so you'll be getting quite a lot of functionality for free: memory allocation and garbage collection. Just in case anyone reading is going to write in C, here's an excellent introduction to the topic:

https://journal.stuffwithstuff.com/2013/12/08/babys-first-ga...

Don't forget to spill the registers.

Never forget:

https://journal.stuffwithstuff.com/2013/12/08/babys-first-ga...

> Let me stress here that while this collector is simple, it isn’t a toy.

> There are a ton of optimizations you can build on top of this—in GCs and programming languages, optimization is 90% of the effort—but the core code here is a legitimate real GC.

> It’s very similar to the collectors that were in Ruby and Lua until recently.

> You can ship production code that uses something exactly like this.

> Now go build something awesome!

Bob Nystrom also has a blog, and his articles are really well written (see his post on Pratt parsers / garbage collectors). I'd also recommend going through the source code for Wren, it shares a lot of code with Lox. Despite the deceptive simplicity of the implementation, it (like Lox) is incredibly fast - it's a great way to learn how to build production grade compilers in general.

Bob Nystrom (of Game Programming Patterns, Crafting Interpreters, and dartfmt fame) also wrote a tutorial[1], of a precise as opposed to a conservative garbage collector.

Regarding register scanning, Andreas Kling has made (or at least quoted) an amusing observation[2] that your C runtime already has a primitive to dump all callee-save registers onto the stack: setjmp(). So all you have to do to scan registers is to put a jmp_buf onto the stack, setjmp() to it, then scan the stack normally starting from its address.

[1] https://journal.stuffwithstuff.com/2013/12/08/babys-first-ga...

[2] https://youtu.be/IzB6iTeo8kk

- http://journal.stuffwithstuff.com/2013/12/08/babys-first-gar...

Might not be that hard: https://journal.stuffwithstuff.com/2013/12/08/babys-first-garbage-collector/

So I actually tested your code: https://gist.github.com/bjourne/c2d0db48b2e50aaadf884e4450c6...

On my machine single-threaded OpenBLAS multiplies two single precision 4096x4096 matrices in 0.95 seconds. Your code takes over 30 seconds. For comparison, my own matrix multiplication code (https://github.com/bjourne/c-examples/blob/master/libraries/...) run in single-threaded mode takes 0.89 seconds. Which actually beats OpenBLAS, but OpenBLAS retakes the lead for larger arrays when multi-threading is added.