Pocketlang is a lightweight (~3000 semicolons) and fast object oriented, embeddable scripting language written in C. It has a ruby flavoured python syntax, that can be learned within 15 minutes. Including the compiler, bytecode VM and runtime, it's a standalone executable with zero external dependencies just as it's self descriptive name. The pocketlang VM can be embedded in another hosting program very easily.

Wren Language and their wonderful book Crafting Interpreters were used as a reference to write this language.

# Python like import statement.
from time import clock as now

# A recursive fibonacci function.
def fib(n)
  if n < 2 then return n end
  return fib(n-1) + fib(n-2)
end

# Prints all fibonacci from 0 to 10 exclusive.
for i in 0..10
  print("fib($i) = ${fib(i)}")
end

You can try pocketlang on your browser. It's a WebAssembly build of the VM compiled using emscripten. Note that in the webassembly version of the language, some features (input, file handling, relative import, etc.) have disabled, has limited memory allocations, and the stdout calls might be slower.

The pocketlang documentation is hosted at https://thakeenathees.github.io/pocketlang/ which is built from the docs branch generated by a little python script at docs/generate.py. Note that the documentations are WIP and might not be up to date.

Pocketlang uses NaN-Boxing which is a memory efficient way to represent dynamic types and dealing with them are much faster. It supports tail call optimization. When a function returns a call, the callee can re-use the caller's stack frame, this will optimize memory from O(n) to O(1) and for tail recursive it'll completely prevent stackoverflows and yet it's faster.

All benchmarks below were executed on: Windows10 (64bit), ASUS N552VX, Intel Core i7-6700HQ 2.6GHz with 12GB SODIMM Ram. And the language versions are: pocketlang (pre-alpha), wren v0.3.0, python v3.7.4, ruby v2.7.2.

The source files used to run benchmarks can be found at test/benchmarks/ directory. They were executed using a small python script in the test directory.

It can be build from source easily without any dependencies, or additional requirements except for a c99 compatible compiler. It can be compiled with the following command.

gcc -o pocket cli/*.c src/core/*.c src/libs/*.c -Isrc/include -lm -ldl

cl /Fepocket cli/*.c src/core/*.c src/libs/*.c /Isrc/include && rm *.obj

make

To run the make file on windows with mingw, you require the GNU make tool which you can get from msys2 or cygwin.

scripts\build.bat

You don't have to run the script from a Visual Studio .NET developer command prompt, It'll search for the MSVS installation path and setup the build environment.

1. Create an empty project file / makefile.
2. Add all C files in the src/core/ directory.
3. Add all C files in the src/libs/ directory.
4. Add all C files in the cli/ directory.
5. Add src/include to include path.
6. If *nix link m, dl
7. Compile.

Visual studio project files can be generated with the premake, see scripts/README for more information. If you weren't able to compile it, please report us by opening an issue.

• Bob Nystrom.(2021) craftinginterpreters [online] Available at www.craftinginterpreters.com/ (Accessed January 2021)

• Mark W. Bailey, Nathan C. Weston (June 2001) Technical report. Performance Benefits of Tail Recursion Removal in Procedural Languages [online] Available at http://cs.hamilton.edu/~mbailey/pubs/techreps/TR-2001-2.pdf

• Roberto Ierusalimschy, Luiz Henrique de Figueiredo, Waldemar Celes Closures in Lua [pdf] Available at https://www.cs.tufts.edu/~nr/cs257/archive/roberto-ierusalimschy/closures-draft.pdf (Accessed March 2022)

• Leonard schütz.(2020) Dynamic Typing and NaN Boxing [online] Available at https://leonardschuetz.ch/blog/nan-boxing/ (Accessed December 2020)

• Bob Nystrom.(2011) Pratt Parsers: Expression Parsing Made Easy [online] Avaliable at http://journal.stuffwithstuff.com/2011/03/19/pratt-parsers-expression-parsing-made-easy/ (Accessed December 2020)