WIP: full gemini support

- DomainNotFound needs to be handled. Probably makes sense to not generate certs for unconfigured domains, to prevent DOS. - Probably needs more testing.
Fix self-signed cert generation, not_before is required apparently
2023-07-31 20:49:00 +02:00 · 2023-07-31 20:48:00 +02:00 · 2023-07-31 20:47:14 +02:00
10 changed files with 216 additions and 36 deletions
--- a/Cargo.lock
+++ b/Cargo.lock
@ -111,6 +111,12 @@ version = "1.6.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bddcadddf5e9015d310179a59bb28c4d4b9920ad0f11e8e14dbadf654890c9a6"

+[[package]]
+name = "arrayvec"
+version = "0.5.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "23b62fc65de8e4e7f52534fb52b0f3ed04746ae267519eef2a83941e8085068b"
+
 [[package]]
 name = "arrayvec"
 version = "0.7.2"
@ -158,6 +164,18 @@ version = "2.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "24a6904aef64d73cf10ab17ebace7befb918b82164785cb89907993be7f83813"

+[[package]]
+name = "bitvec"
+version = "0.19.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "55f93d0ef3363c364d5976646a38f04cf67cfe1d4c8d160cdea02cab2c116b33"
+dependencies = [
+ "funty",
+ "radium",
+ "tap",
+ "wyz",
+]
+
 [[package]]
 name = "block-buffer"
 version = "0.10.4"
@ -458,9 +476,11 @@ name = "domani"
 version = "0.1.0"
 dependencies = [
 "acme2",
+ "bytes",
 "clap",
 "env_logger",
 "futures",
+ "gemini",
 "gix",
 "handlebars",
 "hex",
@ -660,6 +680,12 @@ version = "0.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a06f77d526c1a601b7c4cdd98f54b5eaabffc14d5f2f0296febdc7f357c6d3ba"

+[[package]]
+name = "funty"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fed34cd105917e91daa4da6b3728c47b068749d6a62c59811f06ed2ac71d9da7"
+
 [[package]]
 name = "futures"
 version = "0.3.28"
@ -749,6 +775,18 @@ dependencies = [
 "slab",
 ]

+[[package]]
+name = "gemini"
+version = "0.0.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "12a820f5a9ac6f433b34944dc8d17b759d5009275c8fe12f73b873153dbcd4e0"
+dependencies = [
+ "nom 6.1.2",
+ "paste",
+ "thiserror",
+ "url",
+]
+
 [[package]]
 name = "generic-array"
 version = "0.14.7"
@ -828,7 +866,7 @@ dependencies = [
 "btoi",
 "gix-date",
 "itoa",
- "nom",
+ "nom 7.1.3",
 "thiserror",
 ]

@ -891,7 +929,7 @@ dependencies = [
 "gix-sec",
 "log",
 "memchr",
- "nom",
+ "nom 7.1.3",
 "once_cell",
 "smallvec",
 "thiserror",
@ -1100,7 +1138,7 @@ dependencies = [
 "gix-validate",
 "hex",
 "itoa",
- "nom",
+ "nom 7.1.3",
 "smallvec",
 "thiserror",
 ]
@ -1195,7 +1233,7 @@ dependencies = [
 "gix-hash",
 "gix-transport",
 "maybe-async",
- "nom",
+ "nom 7.1.3",
 "thiserror",
 ]

@ -1226,7 +1264,7 @@ dependencies = [
 "gix-tempfile",
 "gix-validate",
 "memmap2",
- "nom",
+ "nom 7.1.3",
 "thiserror",
 ]

@ -1767,6 +1805,19 @@ version = "1.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"

+[[package]]
+name = "lexical-core"
+version = "0.7.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6607c62aa161d23d17a9072cc5da0be67cdfc89d3afb1e8d9c842bebc2525ffe"
+dependencies = [
+ "arrayvec 0.5.2",
+ "bitflags 1.3.2",
+ "cfg-if",
+ "ryu",
+ "static_assertions",
+]
+
 [[package]]
 name = "libc"
 version = "0.2.142"
@ -1954,6 +2005,19 @@ dependencies = [
 "smallvec",
 ]

+[[package]]
+name = "nom"
+version = "6.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e7413f999671bd4745a7b624bd370a569fb6bc574b23c83a3c5ed2e453f3d5e2"
+dependencies = [
+ "bitvec",
+ "funty",
+ "lexical-core",
+ "memchr",
+ "version_check",
+]
+
 [[package]]
 name = "nom"
 version = "7.1.3"
@ -2081,6 +2145,12 @@ dependencies = [
 "windows-sys 0.45.0",
 ]

+[[package]]
+name = "paste"
+version = "1.0.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "de3145af08024dea9fa9914f381a17b8fc6034dfb00f3a84013f7ff43f29ed4c"
+
 [[package]]
 name = "pem"
 version = "2.0.1"
@ -2249,6 +2319,12 @@ dependencies = [
 "proc-macro2",
 ]

+[[package]]
+name = "radium"
+version = "0.5.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "941ba9d78d8e2f7ce474c015eea4d9c6d25b6a3327f9832ee29a4de27f91bbb8"
+
 [[package]]
 name = "radix_trie"
 version = "0.2.1"
@ -2805,6 +2881,12 @@ dependencies = [
 "unicode-ident",
 ]

+[[package]]
+name = "tap"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "55937e1799185b12863d447f42597ed69d9928686b8d88a1df17376a097d8369"
+
 [[package]]
 name = "tempdir"
 version = "0.3.7"
@ -3121,7 +3203,7 @@ version = "3.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "794a32261a1f5eb6a4462c81b59cec87b5c27d5deea7dd1ac8fc781c41d226db"
 dependencies = [
- "arrayvec",
+ "arrayvec 0.7.2",
 ]

 [[package]]
@ -3508,3 +3590,9 @@ checksum = "80d0f4e272c85def139476380b12f9ac60926689dd2e01d4923222f40580869d"
 dependencies = [
 "winapi",
 ]
+
+[[package]]
+name = "wyz"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "85e60b0d1b5f99db2556934e21937020776a5d31520bf169e851ac44e6420214"
--- a/Cargo.toml
+++ b/Cargo.toml
@ -31,7 +31,7 @@ mime_guess = "2.0.4"
 hyper = { version = "0.14.26", features = [ "server", "stream" ]}
 http = "0.2.9"
 serde_urlencoded = "0.7.1"
-tokio-util = "0.7.8"
+tokio-util = { version = "0.7.8", features = [ "io" ]}
 acme2 = "0.5.1"
 openssl = "0.10.52"
 rustls = "0.21.1"
@ -45,6 +45,8 @@ serde_yaml = "0.9.22"
 rand = "0.8.5"
 reqwest = "0.11.18"
 hyper-reverse-proxy = "0.5.1"
+gemini = "0.0.5"
+bytes = "1.4.0"

 [patch.crates-io]
 tokio-rustls = { git = "https://code.betamike.com/micropelago/tokio-rustls.git", branch = "start-handshake-into-inner" }
--- a/src/domain/tls/certificate.rs
+++ b/src/domain/tls/certificate.rs
@ -35,6 +35,12 @@ impl Certificate {
            .set_subject_name(&name)
            .or_unexpected_while("setting subject name")?;

+        // 1970/01/01
+        let not_before = Asn1Time::from_unix(0).expect("initializing not_before");
+        builder
+            .set_not_before(not_before.as_ref())
+            .or_unexpected_while("setting not_before")?;
+
        // 9999/07/23
        let not_after = Asn1Time::from_unix(253388296800).expect("initializing not_after");
        builder
--- a/src/main.rs
+++ b/src/main.rs
@ -151,6 +151,7 @@ async fn main() {

    let _ = domani::service::gemini::Service::new(
        &mut task_stack,
+        domain_manager.clone(),
        sync::Arc::new(domain_gemini_store),
        config.service,
    );
--- a/src/origin/git.rs
+++ b/src/origin/git.rs
@ -332,7 +332,7 @@ impl super::Store for FSStore {

        // TODO this is very not ideal, the whole file is first read totally into memory, and then
        // that is cloned.
-        let data = file_object.data.clone();
+        let data = bytes::Bytes::copy_from_slice(file_object.data.as_slice());
        Ok(Box::pin(stream::once(async move { Ok(data) })))
    }
 }
--- a/src/service.rs
+++ b/src/service.rs
@ -3,3 +3,32 @@ pub mod gemini;
 pub mod http;

 pub use config::*;
+
+use std::borrow;
+
+fn append_index_to_path<'path, 'index>(
+    path: &'path str,
+    index: &'index str,
+) -> borrow::Cow<'path, str> {
+    if path.len() == 0 {
+        let mut path = String::with_capacity(1 + index.len());
+        path.push('/');
+        path.push_str(index);
+        return borrow::Cow::Owned(path);
+    }
+
+    if path.ends_with('/') {
+        let mut indexed_path = String::with_capacity(path.len() + index.len());
+        indexed_path.push_str(path.as_ref());
+        indexed_path.push_str(index);
+        return borrow::Cow::Owned(indexed_path);
+    }
+
+    borrow::Cow::Borrowed(path)
+}
+
+fn guess_mime(path: &str) -> String {
+    mime_guess::from_path(path)
+        .first_or_octet_stream()
+        .to_string()
+}
--- a/src/service/gemini.rs
+++ b/src/service/gemini.rs
@ -4,12 +4,13 @@ mod proxy;
 pub use config::*;

 use crate::error::unexpected::{self, Mappable};
-use crate::{domain, service, task_stack};
+use crate::{domain, service, task_stack, util};

 use std::sync;
 use tokio_util::sync::CancellationToken;

 pub struct Service {
+    domain_manager: sync::Arc<dyn domain::manager::Manager>,
    cert_resolver: sync::Arc<dyn rustls::server::ResolvesServerCert>,
    config: service::Config,
 }
@ -26,10 +27,12 @@ enum HandleConnError {
 impl Service {
    pub fn new(
        task_stack: &mut task_stack::TaskStack<unexpected::Error>,
+        domain_manager: sync::Arc<dyn domain::manager::Manager>,
        cert_resolver: sync::Arc<dyn rustls::server::ResolvesServerCert>,
        config: service::Config,
    ) -> sync::Arc<Service> {
        let service = sync::Arc::new(Service {
+            domain_manager,
            cert_resolver,
            config,
        });
@ -37,6 +40,72 @@ impl Service {
        service
    }

+    async fn respond_conn<W>(
+        &self,
+        w: W,
+        code: &str,
+        meta: &str,
+        body: Option<util::BoxByteStream>,
+    ) -> unexpected::Result<()>
+    where
+        W: tokio::io::AsyncWrite + Unpin,
+    {
+        use tokio::io::{copy, AsyncWriteExt, BufWriter};
+        let mut w = BufWriter::new(w);
+
+        w.write_all(code.as_bytes()).await.or_unexpected()?;
+        w.write_all(" ".as_bytes()).await.or_unexpected()?;
+        w.write_all(meta.as_bytes()).await.or_unexpected()?;
+        w.write_all("\r\n".as_bytes()).await.or_unexpected()?;
+
+        if let Some(body) = body {
+            let mut body = tokio_util::io::StreamReader::new(body);
+            copy(&mut body, &mut w).await.or_unexpected()?;
+        }
+
+        w.flush().await.or_unexpected()?;
+
+        Ok(())
+    }
+
+    async fn serve_conn<IO>(&self, domain: &domain::Name, conn: IO) -> Result<(), HandleConnError>
+    where
+        IO: tokio::io::AsyncRead + tokio::io::AsyncWrite + Unpin,
+    {
+        use tokio::io::*;
+
+        let (r, w) = split(conn);
+        let mut r = BufReader::new(r);
+
+        let mut req = String::with_capacity(64);
+        r.read_line(&mut req)
+            .await
+            .map_err(|e| HandleConnError::ClientError(format!("failed to read request: {e}")))?;
+
+        let req = gemini::request::parse::request(req.as_bytes())
+            .map(|(_, req)| req)
+            .map_err(|e| HandleConnError::ClientError(format!("failed to parse request: {e}")))?
+            .into_gemini_request()
+            .map_err(|e| HandleConnError::ClientError(format!("failed to parse request: {e}")))?;
+
+        let path = service::append_index_to_path(req.path(), "index.gmi");
+
+        let f = match self.domain_manager.get_file(domain, &path) {
+            Ok(f) => f,
+            Err(domain::manager::GetFileError::DomainNotFound) => panic!("TODO"),
+            Err(domain::manager::GetFileError::FileNotFound) => {
+                return Ok(self.respond_conn(w, "51", "File not found", None).await?)
+            }
+            Err(domain::manager::GetFileError::Unexpected(e)) => return Err(e.into()),
+        };
+
+        let content_type = service::guess_mime(&path);
+
+        Ok(self
+            .respond_conn(w, "20", content_type.as_str(), Some(f))
+            .await?)
+    }
+
    async fn proxy_conn<IO>(
        &self,
        proxied_domain: &ConfigProxiedDomain,
@ -59,7 +128,7 @@ impl Service {
    async fn handle_conn(
        &self,
        conn: tokio::net::TcpStream,
-        _tls_config: sync::Arc<rustls::ServerConfig>,
+        tls_config: sync::Arc<rustls::ServerConfig>,
    ) -> Result<(), HandleConnError> {
        let teed_conn = {
            let (r, w) = tokio::io::split(conn);
@ -92,9 +161,8 @@ impl Service {
                    return Ok(());
                }

-                return Err(HandleConnError::ClientError(format!(
-                    "unknown domain {domain}"
-                )));
+                let conn = start.into_stream(tls_config).await.or_unexpected()?;
+                self.serve_conn(&domain, conn).await
            }
            Err(err) => {
                return Err(unexpected::Error::from(
@ -113,7 +181,7 @@ async fn listen(
    let tls_config = sync::Arc::new(
        rustls::server::ServerConfig::builder()
            .with_safe_defaults()
-            .with_no_client_auth() // TODO maybe this isn't right?
+            .with_no_client_auth()
            .with_cert_resolver(service.cert_resolver.clone()),
    );

--- a/src/service/http.rs
+++ b/src/service/http.rs
@ -79,13 +79,9 @@ struct DomainSyncArgs {

 impl<'svc> Service {
    fn serve(&self, status_code: u16, path: &str, body: Body) -> Response<Body> {
-        let content_type = mime_guess::from_path(path)
-            .first_or_octet_stream()
-            .to_string();
-
        match Response::builder()
            .status(status_code)
-            .header("Content-Type", content_type)
+            .header("Content-Type", service::guess_mime(path))
            .body(body)
        {
            Ok(res) => res,
@ -161,20 +157,10 @@ impl<'svc> Service {
    }

    async fn serve_origin(&self, domain: domain::Name, req: Request<Body>) -> Response<Body> {
-        let mut path_owned;
-        let path = req.uri().path();
+        let path = service::append_index_to_path(req.uri().path(), "index.html");

-        let path = match path.ends_with('/') {
-            true => {
-                path_owned = String::from(path);
-                path_owned.push_str("index.html");
-                path_owned.as_str()
-            }
-            false => path,
-        };
-
-        match self.domain_manager.get_file(&domain, path) {
-            Ok(f) => self.serve(200, path, Body::wrap_stream(f)),
+        match self.domain_manager.get_file(&domain, &path) {
+            Ok(f) => self.serve(200, &path, Body::wrap_stream(f)),
            Err(domain::manager::GetFileError::DomainNotFound) => {
                return self.render_error_page(404, "Domain not found")
            }
--- a/src/util.rs
+++ b/src/util.rs
@ -10,6 +10,6 @@ pub fn open_file(path: &path::Path) -> io::Result<Option<fs::File>> {
    }
 }

-pub type BoxByteStream = futures::stream::BoxStream<'static, io::Result<Vec<u8>>>;
+pub type BoxByteStream = futures::stream::BoxStream<'static, io::Result<bytes::Bytes>>;

 pub type BoxFuture<'a, O> = pin::Pin<Box<dyn futures::Future<Output = O> + Send + 'a>>;
--- a/static/index.gmi
+++ b/static/index.gmi
@ -2,8 +2,8 @@

 This is the gemini index.

-=> foo /foo.gmi
+=> /foo.gmi foo

-=> bar /subdir/bar.gmi
+=> /subdir/bar.gmi bar

-=> penguin /penguin.jpg
+=> /penguin.jpg penguin
Author	SHA1	Message	Date
Brian Picciano	b1717b6ded	WIP: full gemini support - DomainNotFound needs to be handled. Probably makes sense to not generate certs for unconfigured domains, to prevent DOS. - Probably needs more testing.	2023-07-31 20:49:00 +02:00
Brian Picciano	d429b51cf8	Fix self-signed cert generation, not_before is required apparently	2023-07-31 20:48:00 +02:00
Brian Picciano	8a530d4f7d	Fix static file index.gmi	2023-07-31 20:47:14 +02:00