IACore (Independent Architecture Core)

The Battery-Included Foundation for High-Performance C++ Applications.

📖 Description

IACore is a high-performance, battery-included C++20 foundation library designed to eliminate "dependency hell." It bundles essential systems—IPC, Logging, Networking, Compression, and Async Scheduling—into a single, coherent API.

Originally developed as the internal core for IASoft (PVT) LTD., it is now open-source to provide a standardized bedrock for C++ applications where performance matters.

✨ Features

• 🚀 High-Performance IPC: Shared-Memory Ring Buffers with wait-free SPSC synchronization.
• 🌐 Networking: Integrated HTTP/HTTPS client (wrapper around cpp-httplib with automatic Zlib/Gzip handling).
• 🧵 Async Scheduler: A job system with high/normal priority queues.
• 💾 File I/O: Memory-mapped file operations and optimized binary stream readers/writers.
• 📦 Compression: Unified API for Zlib, Gzip, and Zstd.
• 📜 Logging: Thread-safe, colored console and disk logging.
• ⚡ Modern C++: Heavily utilizes modern C++20 concepts, std::span, and tl::expected for error handling.

🛠️ Integration

IACore is built with CMake. You can include it in your project via FetchContent or by adding it as a subdirectory.

Note: On Windows, you must have VCPKG installed and the VCPKG_ROOT environment variable set, for OpenSSL support.

CMake Example

add_subdirectory(IACore)

add_executable(MyApp Main.cpp)
target_link_libraries(MyApp PRIVATE IACore)

📦 Dependencies

IACore manages its own dependencies via CMake's FetchContent. You do not need to install these manually:

• Networking: cpp-httplib
• Compression: zlib-ng & zstd
• Utilities: tl-expected & unordered_dense
• JSON: glaze

Note: Following dependencies are not directly used by IACore, but bundles them (+ helper wrappers) for user convenience: nlohmann_json, simdjson, pugixml

💡 Usage Examples

1. IPC (Manager & Node)

IACore provides a manager/node architecture using shared memory.

Manager:

#include <IACore/IPC.hpp>

// Spawns a child process and connects via Shared Memory
auto nodeID = manager.SpawnNode("MyChildNodeExe");
manager.WaitTillNodeIsOnline(*nodeID);

// Send data with zero-copy overhead
String msg = "Hello Node";
manager.SendPacket(*nodeID, 100, {(PCUINT8)msg.data(), msg.size()});

Node:

#include <IACore/IPC.hpp>

class Node : public IACore::IPC_Node {
public:
    void OnSignal(uint8_t signal) override {
        // Handle signals
    }

    void OnPacket(uint16_t packetID, std::span<const uint8_t> payload) override {
        // Handle packets
    }
};

int main(int argc, char* argv[]) {
    // The connection string is passed as the first argument by the Manager
    if (argc < 2) return -1;
    
    Node node;
    // Connect back to the manager via Shared Memory
    if (!node.Connect(argv[1])) return -1;
    
    while(true) {
        node.Update();
    }
    return 0;
}

2. Async Jobs

#include <IACore/AsyncOps.hpp>

// Initialize worker threads (hardware_concurrency - 2)
IACore::AsyncOps::InitializeScheduler();

// Schedule a task
IACore::AsyncOps::Schedule *mySchedule = new IACore::AsyncOps::Schedule();

IACore::AsyncOps::ScheduleTask([](auto workerID) {
    printf("Running on worker %d\n", workerID);
}, 0, mySchedule);

// Wait for completion
IACore::AsyncOps::WaitForScheduleCompletion(mySchedule);

3. HTTP Request

#include <IACore/HttpClient.hpp>

IACore::HttpClient client("https://api.example.com");
auto res = client.JsonGet<MyResponseStruct>("/data", {});

if (res) {
    std::cout << "Data: " << res->value << "\n";
} else {
    std::cerr << "Error: " << res.error() << "\n";
}

🤝 Contributing

We welcome contributions from the community!

What we accept immediately:

• 📚 Documentation: Improvements to comments, the README, or external docs.
• 🧪 Tests: New unit tests (in Tests/) to improve coverage or reproduce bugs.
• 💡 Examples: New usage examples or sample projects.
• 🐛 Bug Reports: detailed issues describing reproduction steps are highly valued.

Core Library Policy (Src/ Directory)

Currently, we are not accepting Pull Requests that modify the core source code (Src/).

If you find a critical bug in Src/, please open an Issue rather than a PR, and the core team will implement a fix ASAP.

⚖️ License

This project is licensed under the Apache License Version 2.0.

🤖 Use of Generative AI at IASoft

While we never let Generative AI to make architecural/design decisions, and we hand-write almost all of the implementations, we do use Generative AI (Google Gemini) for the following and (ONLY following) tasks:

1. Controlled Repititive Boilerplate Generation: Each and every single line of AI generated code is manually reviewed one-by-one for hallucinations and logic errors. Trust, but verify.

2. Concept Research and Development: Design pattern comparisions, cost-benefit analysis and as a second pair of eyes to evalute and critique our design decisions.

3. Documentation: Repititive method doc strings (parameter, return value descriptions) are done mostly using Generative AI.

4. Code Review: Automated logic checking and static analysis on top of deterministic tools.