CNeyn
is a fast Http library with C interface. You can checkout C++ and Python interfaces too. Some of the features:
- Fast
- Very Easy to Use
- No External Dependencies
Since the project is new there are some limitations:
Windows
platform isn't supported for now.- Partially implements HTTP/1.1 for now.
- Uses some new features of
Linux
kernel so version 4.5 and above kernel is supported for now.
You have two options:
- Building and installing the library system-wide.
- Adding to your
CMake
project as a subdirectory.
You can download the latest release and extract (or you can clone the repository but the latest release is more stable).
You can do these in the cneyn
directory:
mkdir build && cd build
cmake -DCNEYN_BUILD_TESTS=OFF -DCNEYN_INSTALL_LIB=ON -DCMAKE_BUILD_TYPE=Release ..
cmake --build .
sudo cmake --install .
sudo
might be needed or not depending on the install destination. You can use CMAKE_INSTALL_PREFIX
CMake variable to control the install destination and BUILD_SHARED_LIBS
to control the type of library(shared or static).
Then you can use it with various build systems. Here is an example of CMake
:
find_package(cneyn REQUIRED)
add_executable(myexec main.cpp)
target_link_libraries(myexec cneyn::cneyn)
You can put the cneyn
directory inside your project's directory and add it as a subdirectory. For example:
add_subdirectory(cneyn)
add_executable(myexec main.c)
target_link_libraries(myexec ${CNEYN_LIBRARIES})
target_include_directories(myexec PUBLIC ${CNEYN_INCLUDE_DIRS})
You can include the library like this:
#include <cneyn/cneyn.h>
Here are the options you can set:
- Port: port number of the server.
- IP Version: it can be
neyn_address_ipv4
orneyn_address_ipv6
. - Timeout: server in milliseconds. set 0 for no timeout.
- Limit: request size limit in bytes. set 0 for no limit.
- Threads: number of threads.
- Address: address string of the server.
Example:
struct neyn_config config;
config.port = 8081;
config.ipvn = neyn_address_ipv4;
config.timeout = 0;
config.limit = 0;
config.threads = 1;
config.address = "0.0.0.0";
Handling of input requests is done by passing a function to the server. This function takes neyn_request
and neyn_response
and a user-defined data. Request struct has port, address, major, minor, method, path, body, header
fields. Response struct has status, header, body
fields.
Example:
void handler(const struct neyn_request *request, struct neyn_response *response, void *data)
{
response->body.len = 5;
response->body.ptr = "Hello";
neyn_response_write(request, response);
}
You need to create a server object and pass the created configuration and handler function to it. Here is how:
struct neyn_server server;
neyn_server_init(&server);
server.handler = handler;
server.config = config;
You must call neyn_server_run
function and pass the server object to it. If you want the function to be non-blocking you can pass 0 as the last arguement and 1 otherwise. You can stop a non-blocking server by calling neyn_server_kill
on it.
Example:
enum neyn_error error = neyn_server_run(&server, 1);
printf("%i\n", error);
You can report bugs, ask questions and request features on issues page. Pull requests are not accepted right now.
This library is licensed under BSD 3-Clause permissive license. You can read it here.