36 lines
1.4 KiB
Markdown
36 lines
1.4 KiB
Markdown
# A C++ Gemini server
|
|
## Introduction
|
|
As I saw other implementations in Rust but no C++ to compete, I felt that I had to do something ;-).
|
|
It is also a chance for me and a friend (Brumaire) to practice C++ and to try to get as much performance as possible (my speciality is HPC so… Yeah).
|
|
|
|
## Design
|
|
I use PlantUML to do the class diagram for the design. It's a work in progress.
|
|
|
|
## Goal
|
|
What we want is a high performance server using asynchrony at first, and then maybe multi-threading to have the maximum possible performance.
|
|
|
|
## Configuration
|
|
**WIP**
|
|
The configuration of the server is made through a JSON file `config.json`. The JSON parser is [simdjson](https://github.com/simdjson/simdjson) directly integrated
|
|
in our source code.
|
|
|
|
### Structure of the file
|
|
- content
|
|
- path : string, folder to the content of your site.
|
|
- cache
|
|
- enable_cache: bool, to enable cache of files or not
|
|
- cache_size : unsigned int, the maximum size in bytes to put in cache
|
|
- ssl
|
|
- pem_path : string
|
|
- cert_path : string
|
|
|
|
## Dependencies
|
|
It uses a bundled version of `spdlog` as a git submodule, but it does not use the `fmt` lib that is inside `spdlog`.
|
|
|
|
## Documentation
|
|
The doc of the code is made with [doxygen](https://www.doxygen.nl/), the `Doxyfile` is at the root of the repo.
|
|
|
|
## TODO/features
|
|
* Possibility of an applicative gateway (like CGI for example)
|
|
* Content in RAM cache
|
|
* Asynchronous (boost ASIO deals with it) |