various: simplify code; work on security and performance

Signed-off-by: NotAShelf <raf@notashelf.dev>
Change-Id: I9a5114addcab5fbff430ab2b919b83466a6a6964

crates/pinakes-core/src/plugin/loader.rs

@@ -0,0 +1,407 @@
+//! Plugin loader for discovering and loading plugins from the filesystem
+
+use anyhow::{Result, anyhow};
+use pinakes_plugin_api::PluginManifest;
+use std::path::{Path, PathBuf};
+use tracing::{debug, info, warn};
+use walkdir::WalkDir;
+
+/// Plugin loader handles discovery and loading of plugins from directories
+pub struct PluginLoader {
+    /// Directories to search for plugins
+    plugin_dirs: Vec<PathBuf>,
+}
+
+impl PluginLoader {
+    /// Create a new plugin loader
+    pub fn new(plugin_dirs: Vec<PathBuf>) -> Self {
+        Self { plugin_dirs }
+    }
+
+    /// Discover all plugins in configured directories
+    pub async fn discover_plugins(&self) -> Result<Vec<PluginManifest>> {
+        let mut manifests = Vec::new();
+
+        for dir in &self.plugin_dirs {
+            if !dir.exists() {
+                warn!("Plugin directory does not exist: {:?}", dir);
+                continue;
+            }
+
+            info!("Discovering plugins in: {:?}", dir);
+
+            match self.discover_in_directory(dir).await {
+                Ok(found) => {
+                    info!("Found {} plugins in {:?}", found.len(), dir);
+                    manifests.extend(found);
+                }
+                Err(e) => {
+                    warn!("Error discovering plugins in {:?}: {}", dir, e);
+                }
+            }
+        }
+
+        Ok(manifests)
+    }
+
+    /// Discover plugins in a specific directory
+    async fn discover_in_directory(&self, dir: &Path) -> Result<Vec<PluginManifest>> {
+        let mut manifests = Vec::new();
+
+        // Walk the directory looking for plugin.toml files
+        for entry in WalkDir::new(dir)
+            .max_depth(3) // Don't go too deep
+            .follow_links(false)
+        {
+            let entry = match entry {
+                Ok(e) => e,
+                Err(e) => {
+                    warn!("Error reading directory entry: {}", e);
+                    continue;
+                }
+            };
+
+            let path = entry.path();
+
+            // Look for plugin.toml files
+            if path.file_name() == Some(std::ffi::OsStr::new("plugin.toml")) {
+                debug!("Found plugin manifest: {:?}", path);
+
+                match PluginManifest::from_file(path) {
+                    Ok(manifest) => {
+                        info!("Loaded manifest for plugin: {}", manifest.plugin.name);
+                        manifests.push(manifest);
+                    }
+                    Err(e) => {
+                        warn!("Failed to load manifest from {:?}: {}", path, e);
+                    }
+                }
+            }
+        }
+
+        Ok(manifests)
+    }
+
+    /// Resolve the WASM binary path from a manifest
+    pub fn resolve_wasm_path(&self, manifest: &PluginManifest) -> Result<PathBuf> {
+        // The WASM path in the manifest is relative to the manifest file
+        // We need to search for it in the plugin directories
+
+        for dir in &self.plugin_dirs {
+            // Look for a directory matching the plugin name
+            let plugin_dir = dir.join(&manifest.plugin.name);
+            if !plugin_dir.exists() {
+                continue;
+            }
+
+            // Check for plugin.toml in this directory
+            let manifest_path = plugin_dir.join("plugin.toml");
+            if !manifest_path.exists() {
+                continue;
+            }
+
+            // Resolve WASM path relative to this directory
+            let wasm_path = plugin_dir.join(&manifest.plugin.binary.wasm);
+            if wasm_path.exists() {
+                // Verify the resolved path is within the plugin directory (prevent path traversal)
+                let canonical_wasm = wasm_path
+                    .canonicalize()
+                    .map_err(|e| anyhow!("Failed to canonicalize WASM path: {}", e))?;
+                let canonical_plugin_dir = plugin_dir
+                    .canonicalize()
+                    .map_err(|e| anyhow!("Failed to canonicalize plugin dir: {}", e))?;
+                if !canonical_wasm.starts_with(&canonical_plugin_dir) {
+                    return Err(anyhow!(
+                        "WASM binary path escapes plugin directory: {:?}",
+                        wasm_path
+                    ));
+                }
+                return Ok(canonical_wasm);
+            }
+        }
+
+        Err(anyhow!(
+            "WASM binary not found for plugin: {}",
+            manifest.plugin.name
+        ))
+    }
+
+    /// Download a plugin from a URL
+    pub async fn download_plugin(&self, url: &str) -> Result<PathBuf> {
+        // Only allow HTTPS downloads
+        if !url.starts_with("https://") {
+            return Err(anyhow!(
+                "Only HTTPS URLs are allowed for plugin downloads: {}",
+                url
+            ));
+        }
+
+        let dest_dir = self
+            .plugin_dirs
+            .first()
+            .ok_or_else(|| anyhow!("No plugin directories configured"))?;
+
+        std::fs::create_dir_all(dest_dir)?;
+
+        // Download the archive with timeout and size limits
+        let client = reqwest::Client::builder()
+            .timeout(std::time::Duration::from_secs(300))
+            .build()
+            .map_err(|e| anyhow!("Failed to build HTTP client: {}", e))?;
+
+        let response = client
+            .get(url)
+            .send()
+            .await
+            .map_err(|e| anyhow!("Failed to download plugin: {}", e))?;
+
+        if !response.status().is_success() {
+            return Err(anyhow!(
+                "Plugin download failed with status: {}",
+                response.status()
+            ));
+        }
+
+        // Check content-length header before downloading
+        const MAX_PLUGIN_SIZE: u64 = 100 * 1024 * 1024; // 100 MB
+        if let Some(content_length) = response.content_length()
+            && content_length > MAX_PLUGIN_SIZE {
+                return Err(anyhow!(
+                    "Plugin archive too large: {} bytes (max {} bytes)",
+                    content_length,
+                    MAX_PLUGIN_SIZE
+                ));
+            }
+
+        let bytes = response
+            .bytes()
+            .await
+            .map_err(|e| anyhow!("Failed to read plugin response: {}", e))?;
+
+        // Check actual size after download
+        if bytes.len() as u64 > MAX_PLUGIN_SIZE {
+            return Err(anyhow!(
+                "Plugin archive too large: {} bytes (max {} bytes)",
+                bytes.len(),
+                MAX_PLUGIN_SIZE
+            ));
+        }
+
+        // Write archive to a unique temp file
+        let temp_archive = dest_dir.join(format!(".download-{}.tar.gz", uuid::Uuid::now_v7()));
+        std::fs::write(&temp_archive, &bytes)?;
+
+        // Extract using tar with -C to target directory
+        let canonical_dest = dest_dir
+            .canonicalize()
+            .map_err(|e| anyhow!("Failed to canonicalize dest dir: {}", e))?;
+        let output = std::process::Command::new("tar")
+            .args([
+                "xzf",
+                &temp_archive.to_string_lossy(),
+                "-C",
+                &canonical_dest.to_string_lossy(),
+            ])
+            .output()
+            .map_err(|e| anyhow!("Failed to extract plugin archive: {}", e))?;
+
+        // Clean up the archive
+        let _ = std::fs::remove_file(&temp_archive);
+
+        if !output.status.success() {
+            return Err(anyhow!(
+                "Failed to extract plugin archive: {}",
+                String::from_utf8_lossy(&output.stderr)
+            ));
+        }
+
+        // Validate that all extracted files are within dest_dir
+        for entry in WalkDir::new(&canonical_dest).follow_links(false) {
+            let entry = entry?;
+            let entry_canonical = entry.path().canonicalize()?;
+            if !entry_canonical.starts_with(&canonical_dest) {
+                return Err(anyhow!(
+                    "Extracted file escapes destination directory: {:?}",
+                    entry.path()
+                ));
+            }
+        }
+
+        // Find the extracted plugin directory by looking for plugin.toml
+        for entry in WalkDir::new(dest_dir).max_depth(2).follow_links(false) {
+            let entry = entry?;
+            if entry.file_name() == "plugin.toml" {
+                let plugin_dir = entry
+                    .path()
+                    .parent()
+                    .ok_or_else(|| anyhow!("Invalid plugin.toml location"))?;
+
+                // Validate the manifest
+                let manifest = PluginManifest::from_file(entry.path())?;
+                info!("Downloaded and extracted plugin: {}", manifest.plugin.name);
+
+                return Ok(plugin_dir.to_path_buf());
+            }
+        }
+
+        Err(anyhow!(
+            "No plugin.toml found after extracting archive from: {}",
+            url
+        ))
+    }
+
+    /// Validate a plugin package
+    pub fn validate_plugin_package(&self, path: &Path) -> Result<()> {
+        // Check that the path exists
+        if !path.exists() {
+            return Err(anyhow!("Plugin path does not exist: {:?}", path));
+        }
+
+        // Check for plugin.toml
+        let manifest_path = path.join("plugin.toml");
+        if !manifest_path.exists() {
+            return Err(anyhow!("Missing plugin.toml in {:?}", path));
+        }
+
+        // Parse and validate manifest
+        let manifest = PluginManifest::from_file(&manifest_path)?;
+
+        // Check that WASM binary exists
+        let wasm_path = path.join(&manifest.plugin.binary.wasm);
+        if !wasm_path.exists() {
+            return Err(anyhow!(
+                "WASM binary not found: {}",
+                manifest.plugin.binary.wasm
+            ));
+        }
+
+        // Verify the WASM path is within the plugin directory (prevent path traversal)
+        let canonical_wasm = wasm_path.canonicalize()?;
+        let canonical_path = path.canonicalize()?;
+        if !canonical_wasm.starts_with(&canonical_path) {
+            return Err(anyhow!(
+                "WASM binary path escapes plugin directory: {:?}",
+                wasm_path
+            ));
+        }
+
+        // Validate WASM file
+        let wasm_bytes = std::fs::read(&wasm_path)?;
+        if wasm_bytes.len() < 4 || &wasm_bytes[0..4] != b"\0asm" {
+            return Err(anyhow!("Invalid WASM file: {:?}", wasm_path));
+        }
+
+        Ok(())
+    }
+
+    /// Get plugin directory path for a given plugin name
+    pub fn get_plugin_dir(&self, plugin_name: &str) -> Option<PathBuf> {
+        for dir in &self.plugin_dirs {
+            let plugin_dir = dir.join(plugin_name);
+            if plugin_dir.exists() {
+                return Some(plugin_dir);
+            }
+        }
+        None
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use tempfile::TempDir;
+
+    #[tokio::test]
+    async fn test_discover_plugins_empty() {
+        let temp_dir = TempDir::new().unwrap();
+        let loader = PluginLoader::new(vec![temp_dir.path().to_path_buf()]);
+
+        let manifests = loader.discover_plugins().await.unwrap();
+        assert_eq!(manifests.len(), 0);
+    }
+
+    #[tokio::test]
+    async fn test_discover_plugins_with_manifest() {
+        let temp_dir = TempDir::new().unwrap();
+        let plugin_dir = temp_dir.path().join("test-plugin");
+        std::fs::create_dir(&plugin_dir).unwrap();
+
+        // Create a valid manifest
+        let manifest_content = r#"
+[plugin]
+name = "test-plugin"
+version = "1.0.0"
+api_version = "1.0"
+kind = ["media_type"]
+
+[plugin.binary]
+wasm = "plugin.wasm"
+"#;
+        std::fs::write(plugin_dir.join("plugin.toml"), manifest_content).unwrap();
+
+        // Create dummy WASM file
+        std::fs::write(plugin_dir.join("plugin.wasm"), b"\0asm\x01\x00\x00\x00").unwrap();
+
+        let loader = PluginLoader::new(vec![temp_dir.path().to_path_buf()]);
+        let manifests = loader.discover_plugins().await.unwrap();
+
+        assert_eq!(manifests.len(), 1);
+        assert_eq!(manifests[0].plugin.name, "test-plugin");
+    }
+
+    #[test]
+    fn test_validate_plugin_package() {
+        let temp_dir = TempDir::new().unwrap();
+        let plugin_dir = temp_dir.path().join("test-plugin");
+        std::fs::create_dir(&plugin_dir).unwrap();
+
+        // Create a valid manifest
+        let manifest_content = r#"
+[plugin]
+name = "test-plugin"
+version = "1.0.0"
+api_version = "1.0"
+kind = ["media_type"]
+
+[plugin.binary]
+wasm = "plugin.wasm"
+"#;
+        std::fs::write(plugin_dir.join("plugin.toml"), manifest_content).unwrap();
+
+        let loader = PluginLoader::new(vec![]);
+
+        // Should fail without WASM file
+        assert!(loader.validate_plugin_package(&plugin_dir).is_err());
+
+        // Create valid WASM file (magic number only)
+        std::fs::write(plugin_dir.join("plugin.wasm"), b"\0asm\x01\x00\x00\x00").unwrap();
+
+        // Should succeed now
+        assert!(loader.validate_plugin_package(&plugin_dir).is_ok());
+    }
+
+    #[test]
+    fn test_validate_invalid_wasm() {
+        let temp_dir = TempDir::new().unwrap();
+        let plugin_dir = temp_dir.path().join("test-plugin");
+        std::fs::create_dir(&plugin_dir).unwrap();
+
+        let manifest_content = r#"
+[plugin]
+name = "test-plugin"
+version = "1.0.0"
+api_version = "1.0"
+kind = ["media_type"]
+
+[plugin.binary]
+wasm = "plugin.wasm"
+"#;
+        std::fs::write(plugin_dir.join("plugin.toml"), manifest_content).unwrap();
+
+        // Create invalid WASM file
+        std::fs::write(plugin_dir.join("plugin.wasm"), b"not wasm").unwrap();
+
+        let loader = PluginLoader::new(vec![]);
+        assert!(loader.validate_plugin_package(&plugin_dir).is_err());
+    }
+}