lzasm is a lightweight runtime assembler written in C++.
Its main purpose is to assemble the decompression code in shrinkler-gba.
Currently, only ARMv4T Thumb instructions are supported. Moreover, only little endian CPUs are supported.
The name lzasm stands for Laufzeitassembler - runtime assembler in German. lzasm has nothing to do with LZASM (Lazy Assembler, some sort of TASM clone/replacement).
lzasm is licensed under the MIT license.
The production code of lzasm uses C++20 and depends on the C++ Standard Library only. The unit tests are written using Boost.Test, but lzasm can be used without Boost.
lzasm has been tested with the following compilers, x86 and x86-64 each:
- gcc 10.3.0
- clang 11.0 and 12.0
- Visual Studio 2019
lzasm is a header-only library, so you just need to point your compiler at its include directory.
That said, the build process of lzasm uses CMake and provides a basic install target,
which installs the lzasm headers and makes lzasm available through CMake's find_package command.
For more information refer to INSTALL.md.
Below is a short usage example. For more information refer to USAGE.md.
#include <iostream>
#include <lzasm/arm/arm32/divided_thumb_assembler.hpp>
int main()
{
using namespace lzasm::arm::arm32;
// Create assembler
divided_thumb_assembler a;
// Assemble code
a.add(r0, r1, r2);
a.mul(r3, r4);
a.add(r5, 42); // Add immediate: add r5, #42
// Finally call link(), specifying the load address.
// link() may throw std::runtime_error, for instance
// if there are undefined symbols.
bytevector program = a.link(0x1000);
// Do whatever you like with the program
std::cout << "Program size: " << program.size() << " bytes" << std::endl;
}