Scrawl - A Tiny Programming Language

Scrawl is a programming language designed to be very small, yet sufficient to express useful programs (such as implementations of more complex languages). A parser, bytecode compiler, runtime, and standard library have been implemented in 498 lines of C code (as counted by sloccount).

Scrawl provides the following features:

• Control flow: call/return, branches based on comparisons, and unconditional jumps.
• Integer arithmetic: addition, subtraction, and bitwise operations.
• Input/output: opening, creating, reading, and writing files.
• Command line parameters.

This is enough for many useful programs.

To keep the size of the implementation small, the language itself is intentionally very limited. For example, Scrawl offers no named functions or variables, identifying them by number instead. Expressions cannot be nested, and there are no string literals or even comments. Many of these could be implemented by a preprocessor that runs before the Scrawl compiler itself. Indeed, the Scrawl distribution includes a few such preprocessors (written in Scrawl):

string

translates string literals (enclosed by double quotes) to byte sequences.

comment

removes comments (introduced by a ‘;’ and running until the end of the line). Semicolons enclosed by double quotes are ignored.

const

allows the definition of named constants that can be used instead of regular Scrawl code. This can be used, for example, to give names to locations.

As an example, here is a Hello program in Scrawl:

        write $1 $401 @$400
        exit $0

$400:   word $6
$401:   byte $48 $65 $6c $6c $6f  $0a

Here is the same program, with comments, string literals, and named constants added to make it more readable:

;;; Scrawl program that prints Hello.
def msg $401
def msglen $400
def stdout $1

        write stdout msg @msglen
        exit $0

msglen: word $6
msg:    byte "Hello" $0a