feat: add Gitea/Forgejo integration for pull request management

crates/git-host/Cargo.toml

@@ -9,6 +9,7 @@ backon = "1.5.1"
 chrono = { version = "0.4", features = ["serde"] }
 db = { path = "../db" }
 enum_dispatch = "0.3.13"
+reqwest = { workspace = true }
 serde = { workspace = true }
 serde_json = { workspace = true }
 tempfile = "3.21"

crates/git-host/src/detection.rs

@@ -8,6 +8,8 @@ use crate::types::ProviderKind;
 /// - GitHub.com: `https://github.com/owner/repo` or `git@github.com:owner/repo.git`
 /// - GitHub Enterprise: URLs containing `github.` (e.g., `https://github.qkg1.toppany.com/owner/repo`)
 /// - Azure DevOps: `https://dev.azure.com/org/project/_git/repo` or legacy `https://org.visualstudio.com/...`
+/// - Gitea/Forgejo: instances registered via `GITEA_URL` env var, or URLs containing
+///   `/pulls/` (Gitea PR URL pattern), or `gitea.` / `forgejo.` in the hostname
 pub(crate) fn detect_provider_from_url(url: &str) -> ProviderKind {
     let url_lower = url.to_lowercase();
 
@@ -33,15 +35,81 @@ pub(crate) fn detect_provider_from_url(url: &str) -> ProviderKind {
         return ProviderKind::GitHub;
     }
 
+    // Gitea/Forgejo: explicit GITEA_URL match
+    if let Ok(gitea_url) = std::env::var("GITEA_URL") {
+        let gitea_host = gitea_url
+            .to_lowercase()
+            .trim_start_matches("https://")
+            .trim_start_matches("http://")
+            .trim_end_matches('/')
+            .to_string();
+        if !gitea_host.is_empty() && url_lower.contains(&gitea_host) {
+            return ProviderKind::Gitea;
+        }
+    }
+
+    // Gitea PR URL pattern: /pulls/ in path (GitHub uses /pull/, Azure uses /pullrequest/)
+    if url_lower.contains("/pulls/") {
+        return ProviderKind::Gitea;
+    }
+
+    // Well-known Gitea/Forgejo hostnames
+    if url_lower.contains("gitea.") || url_lower.contains("forgejo.") {
+        return ProviderKind::Gitea;
+    }
+
+    // Codeberg is a large Forgejo instance
+    if url_lower.contains("codeberg.org") {
+        return ProviderKind::Gitea;
+    }
+
     ProviderKind::Unknown
 }
 
+/// Extract the base URL for a Gitea instance from a remote or PR URL.
+///
+/// Uses `GITEA_URL` env var only when the URL matches the configured instance.
+/// Otherwise derives the base URL from the URL itself.
+pub(crate) fn gitea_base_url(url: &str) -> String {
+    // Use env var only if the URL actually matches the configured instance
+    if let Ok(gitea_url) = std::env::var("GITEA_URL") {
+        let gitea_host = gitea_url
+            .to_lowercase()
+            .trim_start_matches("https://")
+            .trim_start_matches("http://")
+            .trim_end_matches('/')
+            .to_string();
+        if !gitea_host.is_empty() && url.to_lowercase().contains(&gitea_host) {
+            return gitea_url.trim_end_matches('/').to_string();
+        }
+    }
+
+    // Derive from URL
+    if let Ok(parsed) = url::Url::parse(url) {
+        let mut base = format!("{}://{}", parsed.scheme(), parsed.host_str().unwrap_or(""));
+        if let Some(port) = parsed.port() {
+            base.push_str(&format!(":{port}"));
+        }
+        return base;
+    }
+
+    // SSH-style: git@host:owner/repo.git → https://host
+    if let Some(host_part) = url.strip_prefix("git@")
+        && let Some(host) = host_part.split(':').next()
+    {
+        return format!("https://{host}");
+    }
+
+    url.to_string()
+}
+
 /// Detect the git hosting provider from a PR URL.
 ///
 /// Supports:
 /// - GitHub: `https://github.com/owner/repo/pull/123`
 /// - GitHub Enterprise: `https://github.qkg1.toppany.com/owner/repo/pull/123`
 /// - Azure DevOps: `https://dev.azure.com/org/project/_git/repo/pullrequest/123`
+/// - Gitea/Forgejo: `https://gitea.example.com/owner/repo/pulls/123`
 #[cfg(test)]
 fn detect_provider_from_pr_url(pr_url: &str) -> ProviderKind {
     let url_lower = pr_url.to_lowercase();
@@ -59,7 +127,7 @@ fn detect_provider_from_pr_url(pr_url: &str) -> ProviderKind {
         return ProviderKind::AzureDevOps;
     }
 
-    // Fall back to general URL detection
+    // Fall back to general URL detection (handles Gitea /pulls/ pattern too)
     detect_provider_from_url(pr_url)
 }
 
@@ -136,6 +204,38 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_gitea_well_known_hostname() {
+        assert_eq!(
+            detect_provider_from_url("https://gitea.company.com/owner/repo"),
+            ProviderKind::Gitea
+        );
+        assert_eq!(
+            detect_provider_from_url("https://forgejo.example.org/owner/repo"),
+            ProviderKind::Gitea
+        );
+    }
+
+    #[test]
+    fn test_gitea_codeberg() {
+        assert_eq!(
+            detect_provider_from_url("https://codeberg.org/owner/repo"),
+            ProviderKind::Gitea
+        );
+        assert_eq!(
+            detect_provider_from_url("git@codeberg.org:owner/repo.git"),
+            ProviderKind::Gitea
+        );
+    }
+
+    #[test]
+    fn test_gitea_pr_url_pattern() {
+        assert_eq!(
+            detect_provider_from_url("https://git.example.com/owner/repo/pulls/42"),
+            ProviderKind::Gitea
+        );
+    }
+
     #[test]
     fn test_unknown_provider() {
         assert_eq!(
@@ -175,4 +275,90 @@ mod tests {
             ProviderKind::AzureDevOps
         );
     }
+
+    #[test]
+    fn test_pr_url_gitea() {
+        assert_eq!(
+            detect_provider_from_pr_url("https://gitea.example.com/owner/repo/pulls/42"),
+            ProviderKind::Gitea
+        );
+        assert_eq!(
+            detect_provider_from_pr_url("https://codeberg.org/owner/repo/pulls/7"),
+            ProviderKind::Gitea
+        );
+    }
+
+    #[test]
+    fn test_gitea_base_url_from_https() {
+        let base = super::gitea_base_url("https://gitea.example.com/owner/repo.git");
+        assert_eq!(base, "https://gitea.example.com");
+    }
+
+    #[test]
+    fn test_gitea_base_url_with_port() {
+        let base = super::gitea_base_url("http://localhost:3000/owner/repo");
+        assert_eq!(base, "http://localhost:3000");
+    }
+
+    #[test]
+    fn test_gitea_base_url_from_ssh() {
+        let base = super::gitea_base_url("git@gitea.example.com:owner/repo.git");
+        assert_eq!(base, "https://gitea.example.com");
+    }
+
+    // Edge-case tests for GITEA_URL handling (Bugbot findings)
+    //
+    // SAFETY: These tests manipulate env vars which is unsafe in Rust 2024.
+    // They must run single-threaded (--test-threads=1) to avoid races.
+
+    unsafe fn set_gitea_url(val: &str) {
+        std::env::set_var("GITEA_URL", val);
+    }
+
+    unsafe fn remove_gitea_url() {
+        std::env::remove_var("GITEA_URL");
+    }
+
+    #[test]
+    fn test_empty_gitea_url_does_not_match_all() {
+        // str::contains("") is always true in Rust — ensure we guard against that
+        unsafe { set_gitea_url("") };
+        assert_eq!(
+            detect_provider_from_url("https://gitlab.com/owner/repo"),
+            ProviderKind::Unknown,
+        );
+        assert_eq!(
+            detect_provider_from_url("https://bitbucket.org/owner/repo"),
+            ProviderKind::Unknown,
+        );
+        unsafe { remove_gitea_url() };
+    }
+
+    #[test]
+    fn test_scheme_only_gitea_url_does_not_match_all() {
+        unsafe { set_gitea_url("https://") };
+        assert_eq!(
+            detect_provider_from_url("https://gitlab.com/owner/repo"),
+            ProviderKind::Unknown,
+        );
+        unsafe { remove_gitea_url() };
+    }
+
+    #[test]
+    fn test_gitea_base_url_derives_from_url_when_env_differs() {
+        // GITEA_URL points to one instance, but URL is for Codeberg —
+        // should derive base URL from the URL, not the env var
+        unsafe { set_gitea_url("https://gitea.company.com") };
+        let base = super::gitea_base_url("https://codeberg.org/owner/repo.git");
+        assert_eq!(base, "https://codeberg.org");
+        unsafe { remove_gitea_url() };
+    }
+
+    #[test]
+    fn test_gitea_base_url_uses_env_when_matching() {
+        unsafe { set_gitea_url("https://gitea.company.com") };
+        let base = super::gitea_base_url("https://gitea.company.com/owner/repo.git");
+        assert_eq!(base, "https://gitea.company.com");
+        unsafe { remove_gitea_url() };
+    }
 }