Skip to content
This repository has been archived by the owner on Nov 23, 2020. It is now read-only.

H2CO3/Sparkling

Repository files navigation

What is Sparkling?

Sparkling is a little C-style scripting language I've started as a pet project back in late 2012. It has evolved into an active enough project since then, so I'm open-sourcing it in the hopes that

  1. it will be useful for the community, and
  2. others seeing the potential in it will help me make it better.

On the one hand, the name "Sparkling" comes from my intent to make the language nice, fast and lightweight, properties I associate with sparks in my mind. On the other hand, my nickname (H2CO3) has a lot to do with carbonated water and bubbles.

Sparkling is influenced by other programming languages, namely:

  • C, above all. C is a wonderful programming language, it's well-balanced between a high-level and a low-level language. It has enough but not too much abstraction. Its syntax is clear, concise and well-structured. Quite a lot of languages inherit some of C's design decisions, especially its syntax.
  • Lua. Being small, compact, fast and embeddable was a primary goal while I have been designing Sparkling. In addition, having a separate operator for concatenation is just inevitable.
  • Python. Python has dynamic and strict typing. That's good because it's convenient and safe at the same time. Following this pattern, Sparkling has a strict strong type system, but variables don't have types, only values do, so any value can be stored in a variable, and runtime checks enforce the correctness of operations.
  • Design errors of JavaScript. JavaScript is a horrible language in my opinion. (Sorry, but I feel like that, nobody has to agree.) It has some fundamental flaws that I wanted to eliminate in order to create a sane language. Overloading the '+' operator for concatenation, equality operators that don't work the way one would expect, block statements without scope, implicitly global undeclared variables, semi-reserved (unused and unusable) keywords, object and function names hard-coded into the language -- argh. And at the end, people even call JavaScript a "small and lightweight" language, even though when we look at any decent JavaScript engine (just think of V8, Nitro, SquirrelFish, SpiderMonkey, etc.), all of them consist of dozens or even hundreds of megabytes of sophisticated, overly complex code. There isn't any need for that in an embedded scripting language.
  • Other "default" mistakes that almost every scripting language has had so far. For example, the exclusive use of floating-point numbers for arithmetic operations, even for integers. It's a well-known fact that floating-point computations aren't always exact, they behave in a quite counter-intuitive manner and they can even be significantly slower than integer operations. Another thing is garbage collection. It seems to be the silver bullet of automatic memory management when it comes to scripting languages, but it has some definitive, serious downsides. Non-deterministic behavior is one, speed is two, memory overhead is three, the difficulty of a decent implementation is four. Reference counting is a simpler approach that is more lightweight, easier to implement, completely deterministic and it has no memory allocation overhead (a program uses only as much memory as needed). Refcounting has its disadvantages too (e. g. memory leaks caused by cyclic references), but I believe that its good properties outweigh them.

Why Sparkling?

  • Small, embeddable: compiled size around 150 kB; tiny grammar, simple rules
  • Portable: written in pure ANSI C89 - runs on Windows, Linux, OS X, iOS, ... it can also be used from within a C++ program (apart from the requirement that the library itself should be compiled as C and not as C++, the public parts of the API are entirely C++-compatible.)
  • Well-designed (at least it is supposed to be well-designed):
    • strict yet dynamic typing and consistent semantics: less programmer errors
    • uses integers to perform integer operations (faster, exact, allows bit ops)
    • easy-to-grasp, readable C-style syntax
    • no global variables
    • no implicit conversions or type coercion
  • Fast: speed comparable to that of Lua
  • Friendly: automatic memory management
  • Extensible: simple and flexible C API
  • Free: Sparkling is free software, the reference implementation is licensed under the 2-clause BSD License (see LICENSE.txt for details).

How do I use it?

To learn Sparkling, look at the tutorial/reference manual in doc/, then have a look at the examples in the examples/ directory. Don't worry, there will be more and more documentation over time, but for now that's all I've got.

Using the Sparkling engine is fairly easy. As in the case of practically any modern scripting language, running a program involves three simple steps:

  1. Parse the source text;
  2. Compile it into bytecode;
  3. Execute the bytecode.

The Sparkling C API provides functions for these tasks. For usage information, have a look at implementation of the stand-alone interpreter in spn.c.

Building the library and the REPL

To obtain a debug build (runs slowly, easy to debug):

make
sudo make install

To make a release build (runs fast, hard to debug):

make BUILD=release
sudo make install

To build the JavaScript language bindings:

make -f Makefile.emscripten

To run the unit tests:

make test

To run the unit tests and examine the interpreter using Valgrind:

make test-valgrind

How do I hack on it?

If you have fixed a bug, improved an algorithm or otherwise contributed to the library and you feel like sharing your work with the community, please send me a pull request on GitHub with a concise description of the changes. I only ask you to please respect and follow my coding style (placement of brackets and asterisks, indentation and alignment, whitespace, comments, etc.)

The Sparkling API also has some very basic debugging facilities: namely, it is possible to dump the abstract syntax tree of a parsed program (in order to examine the behavior of the parser) and one can disassemble compiled bytecode as well (so as to debug the compiler and the virtual machine).

What else?

If you have any questions or suggestions, you have used Sparkling in your project or you want to share anything else with me, feel free to drop me an e-mail or a tweet (I run by the name @H2CO3_iOS on Twitter). You may also find me on irc.freenode.net by the same nick, on the channel #sparkling. I appreciate any feedback, including constructive (and polite) criticism, improvement suggestions, questions about usage (if the documentation is unclear), and even donations :P

I've created an entry/wiki page for Sparkling on Rosetta Code, feel free to browse, edit and/or suggest modifications to it. Also check out the list of not implemented tasks and implement some of them at your will (please let me know if/when you implement one, so that I can check it).

The official Sparkling website is h2co3.github.io.

This is an alpha release, so don't expect the engine to be perfect (actually, not even feature-complete). Although I always try to do my best and test my code as much as possible, there may still be bugs - let me know if you find one and I'll fix it as soon as possible. The syntax and semantics of the language are subject to change, too (at least until it leaves alpha), so in the early days, code that ran yesterday can break today. But this is done only in order to let the community decide what kind of features, syntactic and semantic rules would be the best, and when I will have gathered enough suggestions, I'll freeze the language specification. (This is also good for me since now I can procrastinate writing the full specs until the beta release...)

In the meantime, please experiment with the library, write extensions, try to break the code (I appreciate bug reports), play around with the engine in various situations, on different platforms. The more people use Sparkling, the better it will become. Check out the Makefile (with special regards to the BUILD variable) as well and tailor it to your needs.

A word about text editors

If you are using Emacs, then you will for sure appreciate the Sparkling major mode (tools/sparkling-mode.el) and the Flycheck syntax checking plug-in (tools/sparkling-flycheck.el).

To use the major mode, put sparkling-mode.el. into your load-path, then add the following line to your Emacs init file (init.el or .emacs): (require 'sparkling-mode) (You might want to adjust the default tab width in the major mode file if the default - 8 spaces - does not suit you.)

Similarly, for using the Flycheck plug-in, place sparkling-flycheck.el inside your load-path, copy the tools/spnlint script in $PATH, then add (require 'sparkling-flycheck) to the init file after the line that says (require 'flycheck).

If you are using Gedit for coding, install the tools/sparkling.lang file in the appropriate location to have Gedit recognize the syntax of Sparkling and apply syntax highlighting on your code.

Notepad++ users can import tools/sparkling-npp.xml via Language->Define your language...->[ Import... ]*

Atom users can run the script tools/sparkling-atom-install.sh.

Portability note:

The code is portable and cross-platform (at least that is my aim), but the Makefile isn't. I can only test this on Linux, OS X and iOS. There are a couple of variables you can change in the Makefile if it doesn't work out of the box on your platform. A non-exhaustive list of common problems and their possible solution, respectively:

  • The compiler and/or the linker may need an explicit (or different) development sysroot to be specified, perhaps using the -isysroot flag.
  • The use of the editline library can be turned off if it isn't installed on your platform (just make LIBEDIT=0). This does not affect the behavior of the library (since the library itself doesn't depend on any 3rd-party libraries), only the usage of the REPL will be less convenient.
  • In order to create a shared library, position-independent code must be generated. This is done using the -fpic compiler flag by default, which usually generates faster and/or smaller code, but it doesn't always work. If it doesn't (the linker will tell you that), try -fPIC instead.
  • explicit linkage against some components of the C standard library (maths, time, I/O, etc.) may be necessary using different linker flags (e. g. -lm)
  • You may not be using GCC or clang, in which case I'm sorry but you're on your own. (a properly configured IDE should accept and import the code as-is; most notably, if you are on Windows, then you are probably not using GCC or clang but an IDE and a compiler of which the name I don't even dare to mention; in this case, drag 'n dropping the src/ folder into your project should still work.)

Happy programming!

-- H2CO3