BrainFuck is an esoteric programming language that is surprisingly easy to implement. It is almost on the same level as “Hello, world!”, but for compilers and interpreters. In this post, ill share my new little BrainFuck compiler I built with a bash script.
The BrainFuck instruction set
BrainFuck has 8 simple instructions:
| Instruction | Operation |
|---|---|
> |
increment data pointer |
< |
decrement data pointer |
+ |
increment the byte at the data pointer |
- |
decrement the byte at the data pointer |
. |
print the current byte to stdout |
, |
read one byte from stdin to the current byte |
[ |
jump to the matching ] if the current byte is 0 |
] |
jump to the matching [ if the current byte is nonzero |
The C equivalent
BrainFuck works on a “tape”. This is essentially a massive array, with a pointer that moves around. Luckily, this can be implemented with a tiny bit of C. (Thanks wikipedia)
| BF | C code |
|---|---|
> |
++ptr; |
< |
--ptr; |
+ |
++*ptr; |
- |
--*ptr; |
. |
putchar(*ptr); |
, |
*ptr=getchar(); |
[ |
while (*ptr) { |
] |
} |
Implementation
Due to the fact BF has a direct conversion to C, I figured: “Why not just use sed to make a BF compiler?”. And so I did.
The script is available at git.io/JvIHm, and works as follows:
- Create a file containing a “header” that contains some C code that imports
stdio.hand creates a char array - Use SED to replace all BF instructions with the matching C code
- Append a file footer with code to return the current value at the program pointer
- Compile this c file with GCC