If you have general knowledge about how HTTP works, the documentation and the well-documented examples are good resources to get you started.
Licensed under either of
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you shall be dual licensed as above, without any additional terms or conditions.
Async I/O, green threads, coroutines, etc. in Rust are still very immature.
The rouille library just ignores this optimization and focuses on providing an easy-to-use synchronous API instead, where each request is handled in its own dedicated thread.
Even if rouille itself was asynchronous, you would need asynchronous database clients and asynchronous file loading in order to take advantage of it. There are currently no such libraries in the Rust ecosystem.
Once async I/O has been figured out, rouille will be (hopefully transparently) updated to take it into account.
On the author's old Linux machine, some basic benchmarking with wrk -t 4 -c 4
shows the following results:
http.createServer
) yields ~14k requests/sec.While not the fastest, rouille has reasonable performances. Amongst all these examples, rouille is the only one to use synchronous I/O.
It should be trivial to integrate a database or templates to your web server written with rouille. Moreover plugins need maintenance and tend to create a dependency hell. In the author's opinion it is generally better not to use plugins.
Instead of doing this: (pseudo-code)
server.add_middleware(function() { // middleware 1 }); server.add_middleware(function() { // middleware 2 }); server.add_middleware(function() { // middleware 3 });
In rouille you just handle each request entirely manually:
// initialize everything here rouille::start_server(..., move |request| { // middleware 1 // middleware 2 // middleware 3 });