Support GitHub and GitLab URLs

lib/gofer/Cargo.toml

@@ -23,15 +23,15 @@ std = [
     "percent-encoding?/std",
     "thiserror/std",
 ]
-all = ["data", "file", "ftp", "http", "https", "ipfs", "stdin"]
-unstable = ["ftps", "git", "scp"]
+all = ["data", "file", "ftp", "git", "http", "https", "ipfs", "stdin"]
+unstable = ["ftps", "scp"]
 
 # Protocols:
 data = ["dep:data-url"]
 file = ["std"]
 ftp = ["dep:percent-encoding", "dep:suppaftp"]
 ftps = ["ftp", "suppaftp?/rustls"]
-git = ["dep:gix-protocol"]
+git = ["dep:reqwest", "reqwest?/blocking"]
 http = ["dep:reqwest", "reqwest?/blocking"]
 https = ["http", "reqwest?/default", "reqwest?/rustls-tls-native-roots"]
 ipfs = ["dep:reqwest", "reqwest?/blocking"]
@@ -56,9 +56,6 @@ dogma = { version = "0.1.5", default-features = false, features = [
     "iri",
     "uri",
 ] }
-gix-protocol = { version = "0.49", default-features = false, features = [
-    "blocking-client",
-], optional = true }
 miette = { version = "7.5", default-features = false, features = [
     "derive",
 ], optional = true }

lib/gofer/examples/git.rs

@@ -0,0 +1,11 @@
+// This is free and unencumbered software released into the public domain.
+
+use std::{boxed::Box, error::Error, io::stdout, result::Result};
+
+pub fn main() -> Result<(), Box<dyn Error>> {
+    let mut output = stdout().lock();
+    let mut input = gofer::open("git://github.com/dryrust/gofer.rs/master/VERSION")?;
+    std::io::copy(&mut input, &mut output)?;
+
+    Ok(())
+}

lib/gofer/src/error.rs

@@ -113,6 +113,18 @@ pub enum Error {
     )]
     FailedHttpRequest(#[from] reqwest::Error),
 
+    #[cfg(feature = "git")]
+    #[error("invalid Git URL: {0}")]
+    #[cfg_attr(
+        feature = "miette",
+        diagnostic(
+            code(gofer::invalid_git_url),
+            help("it seems that the URL is malformed in some way"),
+            url(docsrs),
+        )
+    )]
+    InvalidGitUrl(String),
+
     #[cfg(feature = "ipfs")]
     #[error("invalid IPFS URL: {0}")]
     #[cfg_attr(
@@ -155,6 +167,9 @@ impl From<Error> for std::io::Error {
             #[cfg(any(feature = "http", feature = "https"))]
             Error::FailedHttpRequest(_e) => std::io::Error::from(ErrorKind::Other), // FIXME
 
+            #[cfg(feature = "git")]
+            Error::InvalidGitUrl(u) => std::io::Error::new(ErrorKind::InvalidInput, u.as_str()),
+
             #[cfg(feature = "ipfs")]
             Error::InvalidIpfsUrl(u) => std::io::Error::new(ErrorKind::InvalidInput, u.as_str()),
         }

lib/gofer/src/schemes/git.rs

@@ -1,9 +1,103 @@
 // This is free and unencumbered software released into the public domain.
 
-use crate::{Read, Result, Url};
+use crate::{Error, Read, Result, Url};
+use reqwest::{blocking::ClientBuilder, redirect};
 
-/// See: https://en.wikipedia.org/wiki/Git
+static USER_AGENT: &str = concat!(env!("CARGO_PKG_NAME"), "/", env!("CARGO_PKG_VERSION"));
+
+/// Downloads a file from a git repository.
+///
+/// Supports:
+/// - GitHub: git://github.com/owner/repo/branch/...path
+/// - GitLab: git://gitlab.com/owner/repo/branch/...path
+///
 /// See: https://git-scm.com/docs/protocol-v2
-pub fn open<'a, 'b>(_url: &'a Url<'b>) -> Result<Box<dyn Read>> {
-    todo!() // TODO
+/// See: https://git-scm.com/docs/gitweb
+/// See: https://docs.github.com/en/repositories/working-with-files/using-files/getting-permanent-links-to-files
+pub fn open<'a, 'b>(url: &'a Url<'b>) -> Result<Box<dyn Read>> {
+    // See: https://docs.rs/reqwest/latest/reqwest/blocking/struct.ClientBuilder.html
+    let client = ClientBuilder::new()
+        .user_agent(USER_AGENT)
+        .redirect(redirect::Policy::default())
+        .https_only(true);
+
+    let url = map_git_url_to_raw_url(url.as_str())?;
+
+    // See: https://docs.rs/reqwest/latest/reqwest/blocking/struct.Client.html#method.get
+    // See: https://docs.rs/reqwest/latest/reqwest/blocking/struct.RequestBuilder.html
+    let response = client.build()?.get(url).send()?;
+
+    Ok(Box::new(response))
+}
+
+/// Maps a git URL to a raw content URL for supported git providers.
+/// - GitHub: git://github.com/owner/repo/branch/...path -> https://raw.githubusercontent.com/owner/repo/refs/heads/branch/...path
+/// - GitLab: git://gitlab.com/owner/repo/branch/...path -> https://gitlab.com/owner/repo/-/raw/branch/...path
+fn map_git_url_to_raw_url(url_str: &str) -> Result<String> {
+    let Some(path) = url_str.strip_prefix("git://") else {
+        return Err(Error::InvalidGitUrl(format!(
+            "Invalid git URL format, expected `git://`: {}",
+            url_str
+        )));
+    };
+
+    let components: Vec<&str> = path.split('/').collect();
+
+    let host = components
+        .first()
+        .ok_or_else(|| Error::InvalidGitUrl(format!("Invalid git URL format: {}", url_str)))?;
+
+    if components.len() < 5 {
+        return Err(Error::InvalidGitUrl(format!(
+            "Invalid GitHub git URL format (need at least 5 components): {}",
+            url_str
+        )));
+    }
+
+    let owner = components[1];
+    let repo = components[2];
+    let version = components[3];
+    let file_path = components[4..].join("/");
+
+    match *host {
+        "github.com" => Ok(format!(
+            "https://raw.githubusercontent.com/{}/{}/{}/{}",
+            owner, repo, version, file_path
+        )),
+        "gitlab.com" => Ok(format!(
+            "https://gitlab.com/{}/{}/-/raw/{}/{}",
+            owner, repo, version, file_path
+        )),
+        _ => Err(Error::InvalidGitUrl(format!(
+            "Unsupported git host: {}",
+            host
+        ))),
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+
+    #[test]
+    fn url_mapping() {
+        assert_eq!(
+            map_git_url_to_raw_url(
+                "git://github.com/dryrust/gofer.rs/master/lib/gofer/src/schemes/git.rs"
+            )
+            .unwrap(),
+            "https://raw.githubusercontent.com/dryrust/gofer.rs/master/lib/gofer/src/schemes/git.rs"
+        );
+        assert_eq!(
+            map_git_url_to_raw_url(
+                "git://github.com/dryrust/gofer.rs/f4ea4a585c009aefd570cefcb6062dc5d579c6ab/VERSION"
+            )
+            .unwrap(),
+            "https://raw.githubusercontent.com/dryrust/gofer.rs/f4ea4a585c009aefd570cefcb6062dc5d579c6ab/VERSION"
+        );
+        assert_eq!(
+            map_git_url_to_raw_url("git://gitlab.com/rust-lang/rust/master/src/README.md").unwrap(),
+            "https://gitlab.com/rust-lang/rust/-/raw/master/src/README.md"
+        )
+    }
 }