Compiler “brainc-brainduck”
BrainDuck compiler is used as a part of BrainDuck computer, which acts as a translator of brainfuck “human-readable” language into binary form, used by BrainDuck CPU. Those instructions and their binary codes have no relation with brainfuck itself, therefore the computer is not called brainfuck computer, because it is not brainfuck. But it does not mean you cannot write and run brainfuck programs in it :)
At first, each compiler, including BrainDuck compiler, provides a lexical analyzer for help with tokenizing the source code, used in syntax highlighting. Secondly, the compiler compiles the source code into other (usually binary) form which is then understood by CPU.
Compilation takes part by user request (clicking on ‘compile’ icon in the main window). After compilation is successful, the compiler usually loads the translated program into operating memory and saves the translation into a file. So it is with BrainDuck compiler. Files have .hex extension (format is called Intel HEX).
Language Syntax
The language of BrainDuck compiler is almost identical to the original brainfuck. However, brainfuck interpreter is not specified well-enough, so there are open questions on how to treat with some special situations, which are described below.
Generally, the language knows eight instructions. They are best described when they are compared with C language equivalent. Brainfuck uses only a single data pointer called P, pointing to bounded memory. The boundary is specified in brainduck-mem plugin.
NOTE: BrainDuck architecture conforms to the true von-Neumann model, instead of classic Harvard-style interpreters. It means that program memory and data memory are not separated. The data pointer is therefore not initialized to 0 as programmers might expect and potentially there can be written brainfuck programs with self-modifications.
| Brainfuck instruction | C language equivalent |
|---|---|
> | P++ |
< | P-- |
+ | ++*P |
- | --*P |
, | *P = getchar() |
. | putchar(*P); |
[ | while (*P) { |
] | } |
The compiler is supplied with many example programs written in brainfuck.
Additional details
Specification of brainfuck language or interpreter implementation is not complete. There are left some details which might be solved differently in different implementations. In this version of BrainDuck implementation in emuStudio, the details are hardcoded, as described below.
Comments
The compiler considers a comment being everything that is not a brainfuck instruction. From the first occurrence of the unknown character, everything to the end of the line is treated as a comment. Exceptions are whitespaces, tabulators, and newlines. This practically means that it is impossible to write brainfuck program with syntax errors.
In the following example, everything starting with # is treated as a comment, up to the end of the line.
++++[-] # Useless program in brainfuck. [-] clears the content of the memory cell.
Cell size
A memory cell has 8-bits (cells are bytes).
Memory size
Memory size is defined in byte-mem plugin. In this version of emuStudio, it is 65536 bytes.
End-of-line code
EOL is defined in vt100-terminal plugin. In the current version of emuStudio, it is a Newline character with ASCII code 10.
End-of-file behavior
EOF is defined in brainduck-cpu and vt100-terminal plugins. In the current version of emuStudio, the current cell (where P is pointing at) is changed to value 0. This is not how original brainfuck behaves, which does not change the cell on EOF.