4 files changed, 333 insertions, 0 deletions
diff --git a/src/web/Cargo.toml b/src/web/Cargo.toml
new file mode 100644
index 00000000..751b9ace
--- /dev/null
+++ b/src/web/Cargo.toml
@@ -0,0 +1,28 @@
+[package]
+name = "garage_web"
+version = "0.1.0"
+authors = ["Alex Auvolat <alex@adnab.me>", "Quentin Dufour <quentin@dufour.io>"]
+edition = "2018"
+license = "GPL-3.0"
+description = "S3-like website endpoint crate for the Garage object store"
+repository = "https://git.deuxfleurs.fr/Deuxfleurs/garage"
+
+[lib]
+path = "lib.rs"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+garage_util = { version = "0.1", path = "../util" }
+garage_table = { version = "0.1.1", path = "../table" }
+garage_model = { version = "0.1.1", path = "../model" }
+garage_api = { version = "0.1.1", path = "../api" }
+
+err-derive = "0.2.3"
+log = "0.4"
+futures = "0.3"
+http = "0.2"
+hyper = "0.13"
+percent-encoding = "2.1.0"
+roxmltree = "0.11"
+idna = "0.2"
diff --git a/src/web/error.rs b/src/web/error.rs
new file mode 100644
index 00000000..14bc3b75
--- /dev/null
+++ b/src/web/error.rs
@@ -0,0 +1,39 @@
+use err_derive::Error;
+use hyper::StatusCode;
+
+use garage_util::error::Error as GarageError;
+
+#[derive(Debug, Error)]
+pub enum Error {
+	#[error(display = "API error: {}", _0)]
+	ApiError(#[error(source)] garage_api::error::Error),
+
+	// Category: internal error
+	#[error(display = "Internal error: {}", _0)]
+	InternalError(#[error(source)] GarageError),
+
+	#[error(display = "Not found")]
+	NotFound,
+
+	// Category: bad request
+	#[error(display = "Invalid UTF-8: {}", _0)]
+	InvalidUTF8(#[error(source)] std::str::Utf8Error),
+
+	#[error(display = "Invalid header value: {}", _0)]
+	InvalidHeader(#[error(source)] hyper::header::ToStrError),
+
+	#[error(display = "Bad request: {}", _0)]
+	BadRequest(String),
+}
+
+impl Error {
+	pub fn http_status_code(&self) -> StatusCode {
+		match self {
+			Error::NotFound => StatusCode::NOT_FOUND,
+			Error::ApiError(e) => e.http_status_code(),
+			Error::InternalError(GarageError::RPC(_)) => StatusCode::SERVICE_UNAVAILABLE,
+			Error::InternalError(_) => StatusCode::INTERNAL_SERVER_ERROR,
+			_ => StatusCode::BAD_REQUEST,
+		}
+	}
+}
diff --git a/src/web/lib.rs b/src/web/lib.rs
new file mode 100644
index 00000000..f28937b9
--- /dev/null
+++ b/src/web/lib.rs
@@ -0,0 +1,6 @@
+#[macro_use]
+extern crate log;
+
+pub mod error;
+
+pub mod web_server;
diff --git a/src/web/web_server.rs b/src/web/web_server.rs
new file mode 100644
index 00000000..24d111a9
--- /dev/null
+++ b/src/web/web_server.rs
@@ -0,0 +1,260 @@
+use std::{borrow::Cow, convert::Infallible, net::SocketAddr, sync::Arc};
+
+use futures::future::Future;
+
+use hyper::{
+	header::HOST,
+	server::conn::AddrStream,
+	service::{make_service_fn, service_fn},
+	Body, Method, Request, Response, Server,
+};
+
+use idna::domain_to_unicode;
+
+use crate::error::*;
+use garage_api::s3_get::{handle_get, handle_head};
+use garage_model::bucket_table::*;
+use garage_model::garage::Garage;
+use garage_table::*;
+use garage_util::error::Error as GarageError;
+
+pub async fn run_web_server(
+	garage: Arc<Garage>,
+	shutdown_signal: impl Future<Output = ()>,
+) -> Result<(), GarageError> {
+	let addr = &garage.config.s3_web.bind_addr;
+
+	let service = make_service_fn(|conn: &AddrStream| {
+		let garage = garage.clone();
+		let client_addr = conn.remote_addr();
+		async move {
+			Ok::<_, Error>(service_fn(move |req: Request<Body>| {
+				let garage = garage.clone();
+				handle_request(garage, req, client_addr)
+			}))
+		}
+	});
+
+	let server = Server::bind(&addr).serve(service);
+	let graceful = server.with_graceful_shutdown(shutdown_signal);
+	info!("Web server listening on http://{}", addr);
+
+	graceful.await?;
+	Ok(())
+}
+
+async fn handle_request(
+	garage: Arc<Garage>,
+	req: Request<Body>,
+	addr: SocketAddr,
+) -> Result<Response<Body>, Infallible> {
+	info!("{} {} {}", addr, req.method(), req.uri());
+	let res = serve_file(garage, req).await;
+	match &res {
+		Ok(r) => debug!("{} {:?}", r.status(), r.headers()),
+		Err(e) => warn!("Response: error {}, {}", e.http_status_code(), e),
+	}
+
+	Ok(res.unwrap_or_else(error_to_res))
+}
+
+fn error_to_res(e: Error) -> Response<Body> {
+	let body: Body = Body::from(format!("{}\n", e));
+	let mut http_error = Response::new(body);
+	*http_error.status_mut() = e.http_status_code();
+	http_error
+}
+
+async fn serve_file(garage: Arc<Garage>, req: Request<Body>) -> Result<Response<Body>, Error> {
+	// Get http authority string (eg. [::1]:3902 or garage.tld:80)
+	let authority = req
+		.headers()
+		.get(HOST)
+		.ok_or(Error::BadRequest(format!("HOST header required")))?
+		.to_str()?;
+
+	// Get bucket
+	let (host, _) = domain_to_unicode(authority_to_host(authority)?);
+	let root = &garage.config.s3_web.root_domain;
+	let bucket = host_to_bucket(&host, root);
+
+	// Check bucket is exposed as a website
+	let bucket_desc = garage
+		.bucket_table
+		.get(&EmptyKey, &bucket.to_string())
+		.await?
+		.filter(|b| !b.is_deleted())
+		.ok_or(Error::NotFound)?;
+
+	match bucket_desc.state.get() {
+		BucketState::Present(params) if *params.website.get() => Ok(()),
+		_ => Err(Error::NotFound),
+	}?;
+
+	// Get path
+	let path = req.uri().path().to_string();
+	let index = &garage.config.s3_web.index;
+	let key = path_to_key(&path, &index)?;
+
+	info!("Selected bucket: \"{}\", selected key: \"{}\"", bucket, key);
+
+	let res = match req.method() {
+		&Method::HEAD => handle_head(garage, &bucket, &key).await?,
+		&Method::GET => handle_get(garage, &req, bucket, &key).await?,
+		_ => return Err(Error::BadRequest(format!("HTTP method not supported"))),
+	};
+
+	Ok(res)
+}
+
+/// Extract host from the authority section given by the HTTP host header
+///
+/// The HTTP host contains both a host and a port.
+/// Extracting the port is more complex than just finding the colon (:) symbol due to IPv6
+/// We do not use the collect pattern as there is no way in std rust to collect over a stack allocated value
+/// check here: https://docs.rs/collect_slice/1.2.0/collect_slice/
+fn authority_to_host(authority: &str) -> Result<&str, Error> {
+	let mut iter = authority.chars().enumerate();
+	let (_, first_char) = iter
+		.next()
+		.ok_or(Error::BadRequest(format!("Authority is empty")))?;
+
+	let split = match first_char {
+		'[' => {
+			let mut iter = iter.skip_while(|(_, c)| c != &']');
+			match iter.next() {
+				Some((_, ']')) => iter.next(),
+				_ => {
+					return Err(Error::BadRequest(format!(
+						"Authority {} has an illegal format",
+						authority
+					)))
+				}
+			}
+		}
+		_ => iter.skip_while(|(_, c)| c != &':').next(),
+	};
+
+	match split {
+		Some((i, ':')) => Ok(&authority[..i]),
+		None => Ok(authority),
+		Some((_, _)) => Err(Error::BadRequest(format!(
+			"Authority {} has an illegal format",
+			authority
+		))),
+	}
+}
+
+/// Host to bucket
+///
+/// Convert a host, like "bucket.garage-site.tld" or "john.doe.com"
+/// to the corresponding bucket, resp. "bucket" and "john.doe.com"
+/// considering that ".garage-site.tld" is the "root domain".
+/// This behavior has been chosen to follow AWS S3 semantic.
+fn host_to_bucket<'a>(host: &'a str, root: &str) -> &'a str {
+	if root.len() >= host.len() || !host.ends_with(root) {
+		return host;
+	}
+
+	let len_diff = host.len() - root.len();
+	let missing_starting_dot = root.chars().next() != Some('.');
+	let cursor = if missing_starting_dot {
+		len_diff - 1
+	} else {
+		len_diff
+	};
+	&host[..cursor]
+}
+
+/// Path to key
+///
+/// Convert the provided path to the internal key
+/// When a path ends with "/", we append the index name to match traditional web server behavior
+/// which is also AWS S3 behavior.
+fn path_to_key<'a>(path: &'a str, index: &str) -> Result<Cow<'a, str>, Error> {
+	let path_utf8 = percent_encoding::percent_decode_str(&path).decode_utf8()?;
+
+	if path_utf8.chars().next() != Some('/') {
+		return Err(Error::BadRequest(format!(
+			"Path must start with a / (slash)"
+		)));
+	}
+
+	match path_utf8.chars().last() {
+		None => unreachable!(),
+		Some('/') => {
+			let mut key = String::with_capacity(path_utf8.len() + index.len());
+			key.push_str(&path_utf8[1..]);
+			key.push_str(index);
+			Ok(key.into())
+		}
+		Some(_) => match path_utf8 {
+			Cow::Borrowed(pu8) => Ok((&pu8[1..]).into()),
+			Cow::Owned(pu8) => Ok((&pu8[1..]).to_string().into()),
+		},
+	}
+}
+
+#[cfg(test)]
+mod tests {
+	use super::*;
+
+	#[test]
+	fn authority_to_host_with_port() -> Result<(), Error> {
+		let domain = authority_to_host("[::1]:3902")?;
+		assert_eq!(domain, "[::1]");
+		let domain2 = authority_to_host("garage.tld:65200")?;
+		assert_eq!(domain2, "garage.tld");
+		let domain3 = authority_to_host("127.0.0.1:80")?;
+		assert_eq!(domain3, "127.0.0.1");
+		Ok(())
+	}
+
+	#[test]
+	fn authority_to_host_without_port() -> Result<(), Error> {
+		let domain = authority_to_host("[::1]")?;
+		assert_eq!(domain, "[::1]");
+		let domain2 = authority_to_host("garage.tld")?;
+		assert_eq!(domain2, "garage.tld");
+		let domain3 = authority_to_host("127.0.0.1")?;
+		assert_eq!(domain3, "127.0.0.1");
+		assert!(authority_to_host("[").is_err());
+		assert!(authority_to_host("[hello").is_err());
+		Ok(())
+	}
+
+	#[test]
+	fn host_to_bucket_test() {
+		assert_eq!(
+			host_to_bucket("john.doe.garage.tld", ".garage.tld"),
+			"john.doe"
+		);
+
+		assert_eq!(
+			host_to_bucket("john.doe.garage.tld", "garage.tld"),
+			"john.doe"
+		);
+
+		assert_eq!(host_to_bucket("john.doe.com", "garage.tld"), "john.doe.com");
+
+		assert_eq!(
+			host_to_bucket("john.doe.com", ".garage.tld"),
+			"john.doe.com"
+		);
+
+		assert_eq!(host_to_bucket("garage.tld", "garage.tld"), "garage.tld");
+
+		assert_eq!(host_to_bucket("garage.tld", ".garage.tld"), "garage.tld");
+	}
+
+	#[test]
+	fn path_to_key_test() -> Result<(), Error> {
+		assert_eq!(path_to_key("/file%20.jpg", "index.html")?, "file .jpg");
+		assert_eq!(path_to_key("/%20t/", "index.html")?, " t/index.html");
+		assert_eq!(path_to_key("/", "index.html")?, "index.html");
+		assert_eq!(path_to_key("/hello", "index.html")?, "hello");
+		assert!(path_to_key("", "index.html").is_err());
+		assert!(path_to_key("i/am/relative", "index.html").is_err());
+		Ok(())
+	}
+}