This project is part of the csBigInteger macro project, a C# BigInteger implementation on C++ (portable format desired)
This project is an ongoing port of csBigInteger.js project: https://github.com/neoresearch/csBigInteger.js
Existing frameworks use high-level languages that may not be suitable for very lightweight architectures, such as microcontrollers with very limited computing capabilities.
C/C++ is interoperable with nearly all existing languages, so the idea is to provide modules that can be reused on other projects (on other languages too).
The project is very flexible, but still very simple to use.
See examples in demo/ folder. GNU make with build three versions:
- demo_hand: using HandBigInt.hpp (slow and the simplest to use)
- demo_cshand: using BigInteger.h + BigIntegerHand.cpp (slow)
- demo_csgmp: using BigInteger.h + BigIntegerGMP.cpp (fast)
Example from ./demo_cshand (print -1, 128 and 2^128):
demonstration for csBigInteger (Engine: 'HandBigInt')
usage: <operation> <enter> <number> <enter> <number <enter>
avaliable operations are: + - / * % ^ > < 0x 0b (0x and 0b are conversions to hex big endian and binary)
please type the operation: 0x
expects 1 number(s)
please type first number (in decimal, or unsigned hex with prefix 0x, or binary prefix 0b):-1
number is (decimal) '-1'
conversion: 0xff <----- Negative values have MSb set to one
please type the operation: 0x
expects 1 number(s)
please type first number (in decimal, or unsigned hex with prefix 0x, or binary prefix 0b):128
number is (decimal) '128'
conversion: 0x0080 <----- This zero on the left is correct and necessary for csBigInteger
please type the operation: ^
expects 2 number(s)
please type first number (in decimal, or unsigned hex with prefix 0x, or binary prefix 0b):2
number is (decimal) '2'
please type second number (in decimal):128
number is '128'
result: 340282366920938463463374607431768211456
Just include the Single Header solution src/HandBigInt.hpp.
#include<iostream>
#include "HandBigInt.hpp"
int main() {
HandBigInt big(10);
big = big + 1;
std::cout << big.toString() << std::endl;
// output is 11
return 0;
}
This is VERY innefficient, but it is tested against all BigInteger tests (except Online_Pack test, which takes so long...). The BigInteger layer can be implemented with HandBigInt, GMP or Mono.
If you have a C++ project and want to use csBigInteger, just:
#include "BigInteger.h"
using namespace csbiginteger;
// ...
BigInteger big{"123456789123456789"};
big.ToString(); // prints '123456789123456789' in base 10
//
BigInteger big1{1};
BigInteger big2{2};
BigInteger big3 = big1 + big2;
big3.ToString(); // prints '3' in base 10To compile this using GNU MP library (install its libs -lgmp -lgmpxx), just include flag GMP_CSBIG (or link together with BigIntegerGMP.cpp). Example with GCC: g++ -DGMP_CSBIG yourfile.cpp -o output -lgmp -lgmpxx.
Other options is to use MONO_CSBIG (or link against BigIntegerMono.cpp).
With Mono you may have an "equivalent" version of "original" C#.
It may be more efficient, but it's harder to build (requires dotnet and mono dependencies).
To use this library on other languages, you need to build a static (.a) or shared library (.dll,.so). You can find some experimental versions on build folder, but it's safer to build by yourself.
The process is similar, using C header csBigIntegerLib.h and alsocsBigIntegerLib.cpp (see makefile for an example).
It may look crazy, but csBigIntegerLib.h is also useful to "get external implementations" from other languages (like javascript, for example). In this case, your C++ code should #include "csBigIntegerLibClass.hpp" and using namespace csbigintegerlib.
The implementation may be provided by C++ itself (via csBigIntegerLib.cpp) or any other language you like.
There are currently two implementations for BigInteger.h: BigIntegerGMP.cpp or BigIntegerMono.cpp.
Getting submodules: git submodule update --init --recursive and git pull --recurse-submodules
On debian-based systems (or ubuntu), just type make vendor (it will install libgmp-dev package).
Mono implementation will generate a csbiginteger_dotnet.dll and load it through csbiginteger_mono.so. It is much more complex process, but it guarantees that original BigInteger C# library is being used.
Right now, tests are passing for both gmp and mono implementations, so gmp C++ native implementation is much preferred (and more lightweight) on practice.
If you want to go this way, you will need mono:
sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys 3FA7E0328081BFF6A14DA29AA6A19B38D3D831EF
sudo apt install apt-transport-https ca-certificates
echo "deb https://download.mono-project.com/repo/ubuntu stable-xenial main" | sudo tee /etc/apt/sources.list.d/mono-official-stable.list
sudo apt update
sudo apt install mono-complete
It will also configure test library (as long as you cloned this project with --submodules too).
To test, just run make test.
Currently, C++11 is adopted, in order to keep the best compatibility between conversors and compilers. However, it is recommended to migrate to C++17 as soon as possible, if this does not break compatibility with any existing modules and tools.
Let's please follow the CppCoreGuidelines.
If using vscode IDE, it is recommended to install the following extensions:
- C/C++ (currently 0.23.0-insiders2)
- C++ Intellisense (currently 0.2.2)
- GoogleTest Adapter (currently 1.8.3)
The currently adopted style for C++ is Mozilla, with indentation level set to 3.
Recommended configuration for vscode:
{
"[cpp]": {
"editor.tabSize" : 3,
"editor.detectIndentation": false
},
"C_Cpp.clang_format_fallbackStyle": "{ BasedOnStyle : Mozilla , ColumnLimit : 0, IndentWidth: 3, AccessModifierOffset: -3}"
}The naming style for variables and methods is based on C# language.
So, CamelCase is used and public variables start with upper-case letters, while private and local variables start with lower-case.
The idea is to preseve maximum compatibility with reference project (which is on C#).
Anything that is beyond the scope of the reference project can use classic CamelCase C++ naming (for example, starting variables with lower-case).
Code follows MIT License.
Implementation BigIntegerGMP.cpp (class implementation of standard BigInteger.h) is LGPLv3. The reason is that this implementation depends on GNU MP Bignum Library (licensed LGPLv3 since version 6), what means that all modifications of BigIntegerGMP.cpp, or usage of its code (even partially) on other projects should also adopt LGPLv3 (not MIT License).
Implementation BigIntegerMono.cpp depends on Mono license, which is also MIT License.
The binaries generated by this project (csbiginteger_gmp.so or csbiginteger_mono.so) can be freely used on other projects, regardless of license.