Introduction

The panther language is a language made for exploitations.

It's main focus is minimal code size through code size optimizations and compiling directly to a shellcode (position independent code).

Flow

Lexer: Parse text file into tokens
Parser: Parse tokens into AST
Analyzer: Validate that the AST is correct panther syntax
Bytecode: Parse AST into bytecode
LLVM: Parse bytecode into LLVM IR
Optimizations + Backend: llvm

What can it do currently?

Currently, the panther binary can compile a single file using the llvm backend, with minimal type inference and type checking.

When running the panther compiler on main.pan

# Compile the panther compiler
nimble build -d:release
# Panther compiles an object file, in this example I compile to a windows COFF
./panther c -i main.pan -o output.o -t x86_64-pe-windows-coff
# Link the object file to an executable, needed to get rid of relocations
ld -nostartfiles output.o -o output.exe
# Extract the .shell section created by the compiler
objcopy -j .shell -O binary output.exe shellcode.bin

When main.pan looks like:

proc _start();  // First function in the shellcode
proc fib(a: s32) -> s32 {
    if a < 2 {
        return a;
    } else {
        return fib(a - 1) + fib(a - 2);
    }
}
// The printf symbol predetermined by an address.
// In the future I would like to add a feature of searching for a symbol at runtime.
let printf = 0x402ba0 as proc(fmt: string, num1: s32, num2: s32);
proc _start() {
    let i = 1;
    while i <= 15 {
        printf("%d: %d\n", i, fib(i));
        i = i + 1;
    }
}

Results in a shellcode (shellcode.bin) that prints the the first 15 fibonnaci numbers.