aboutsummaryrefslogtreecommitdiffhomepage
path: root/crates/mozart-registry/src/packagist.rs
diff options
context:
space:
mode:
authornsfisis <nsfisis@gmail.com>2026-02-22 00:37:54 +0900
committernsfisis <nsfisis@gmail.com>2026-02-22 00:37:54 +0900
commit0a8e5935e6305819bb02d8c69e2f046ff397913a (patch)
treee5a288e679477b1603d7989e986ca22bbe590aa4 /crates/mozart-registry/src/packagist.rs
parentb5af594fec7da72b15c9a202c641af0494db6355 (diff)
downloadphp-mozart-0a8e5935e6305819bb02d8c69e2f046ff397913a.tar.gz
php-mozart-0a8e5935e6305819bb02d8c69e2f046ff397913a.tar.zst
php-mozart-0a8e5935e6305819bb02d8c69e2f046ff397913a.zip
refactor(workspace): split monolithic crate into 6 workspace crates
Extract modules from the single `mozart` crate into 5 focused library crates to improve compilation parallelism and architectural clarity: - mozart-constraint: version constraint parser (independent) - mozart-core: base types, console, validation, platform utilities - mozart-archiver: archive creation (tar, zip, bzip2) - mozart-registry: Packagist API, cache, resolver, downloader, lockfile - mozart-autoload: autoloader generation and PHP scanner Refactor Console::from_cli and build_cache_config to accept primitive args instead of &Cli to break circular dependencies. Introduce [workspace.dependencies] for centralized version management. Remove 9 unused direct dependencies from the CLI crate. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Diffstat (limited to 'crates/mozart-registry/src/packagist.rs')
-rw-r--r--crates/mozart-registry/src/packagist.rs629
1 files changed, 629 insertions, 0 deletions
diff --git a/crates/mozart-registry/src/packagist.rs b/crates/mozart-registry/src/packagist.rs
new file mode 100644
index 0000000..ba80e7e
--- /dev/null
+++ b/crates/mozart-registry/src/packagist.rs
@@ -0,0 +1,629 @@
+use crate::cache::Cache;
+use serde::{Deserialize, Serialize};
+use std::collections::BTreeMap;
+
+#[derive(Debug, Clone, Deserialize)]
+pub struct PackagistDist {
+ #[serde(rename = "type")]
+ pub dist_type: String,
+ pub url: String,
+ pub reference: Option<String>,
+ pub shasum: Option<String>,
+}
+
+#[derive(Debug, Clone, Deserialize)]
+pub struct PackagistSource {
+ #[serde(rename = "type")]
+ pub source_type: String,
+ pub url: String,
+ pub reference: Option<String>,
+}
+
+#[derive(Debug, Clone, Deserialize)]
+pub struct PackagistVersion {
+ pub version: String,
+ pub version_normalized: String,
+ #[serde(default)]
+ pub require: BTreeMap<String, String>,
+ #[serde(default)]
+ pub replace: BTreeMap<String, String>,
+ #[serde(default)]
+ pub provide: BTreeMap<String, String>,
+ #[serde(default)]
+ pub conflict: BTreeMap<String, String>,
+ pub dist: Option<PackagistDist>,
+ pub source: Option<PackagistSource>,
+
+ #[serde(rename = "require-dev", default)]
+ pub require_dev: BTreeMap<String, String>,
+
+ #[serde(default)]
+ pub suggest: Option<BTreeMap<String, String>>,
+
+ #[serde(rename = "type")]
+ pub package_type: Option<String>,
+
+ pub autoload: Option<serde_json::Value>,
+
+ #[serde(rename = "autoload-dev")]
+ pub autoload_dev: Option<serde_json::Value>,
+
+ pub license: Option<Vec<String>>,
+
+ pub description: Option<String>,
+
+ pub homepage: Option<String>,
+
+ pub keywords: Option<Vec<String>>,
+
+ pub authors: Option<Vec<serde_json::Value>>,
+
+ pub support: Option<serde_json::Value>,
+
+ pub funding: Option<Vec<serde_json::Value>>,
+
+ pub time: Option<String>,
+
+ pub extra: Option<serde_json::Value>,
+
+ #[serde(rename = "notification-url")]
+ pub notification_url: Option<String>,
+}
+
+impl PackagistVersion {
+ /// Extract the `extra.branch-alias` map from this version's metadata.
+ ///
+ /// Composer packages can declare branch aliases in `extra.branch-alias`:
+ /// ```json
+ /// {
+ /// "extra": {
+ /// "branch-alias": {
+ /// "dev-master": "2.x-dev"
+ /// }
+ /// }
+ /// }
+ /// ```
+ ///
+ /// Returns a map from branch name (e.g. `"dev-master"`) to alias target
+ /// (e.g. `"2.x-dev"`). Returns an empty map when no aliases are declared.
+ pub fn branch_aliases(&self) -> BTreeMap<String, String> {
+ let Some(extra) = &self.extra else {
+ return BTreeMap::new();
+ };
+
+ let Some(branch_alias) = extra.get("branch-alias") else {
+ return BTreeMap::new();
+ };
+
+ let Some(map) = branch_alias.as_object() else {
+ return BTreeMap::new();
+ };
+
+ map.iter()
+ .filter_map(|(k, v)| v.as_str().map(|s| (k.clone(), s.to_string())))
+ .collect()
+ }
+}
+
+/// Parse a Packagist p2 API JSON response.
+///
+/// The response format is: `{"packages": {"vendor/package": [...]}}`.
+pub fn parse_p2_response(json: &str, package_name: &str) -> anyhow::Result<Vec<PackagistVersion>> {
+ #[derive(Deserialize)]
+ struct P2Response {
+ packages: BTreeMap<String, Vec<PackagistVersion>>,
+ }
+
+ let response: P2Response = serde_json::from_str(json)?;
+ response
+ .packages
+ .into_iter()
+ .find(|(key, _)| key == package_name)
+ .map(|(_, versions)| versions)
+ .ok_or_else(|| anyhow::anyhow!("Package \"{package_name}\" not found in response"))
+}
+
+/// Fetch package version metadata from the Packagist p2 API.
+///
+/// If `repo_cache` is provided, the JSON response is cached on disk under the
+/// key `"provider-{vendor}~{package}.json"`. Subsequent calls for the same
+/// package are served from cache without a network request.
+pub fn fetch_package_versions(
+ package_name: &str,
+ repo_cache: Option<&Cache>,
+) -> anyhow::Result<Vec<PackagistVersion>> {
+ // Build cache key: replace `/` with `~` per cache key convention
+ let cache_key = format!("provider-{}.json", package_name.replace('/', "~"));
+
+ // Check cache first
+ if let Some(cache) = repo_cache
+ && let Some(cached) = cache.read(&cache_key)
+ {
+ return parse_p2_response(&cached, package_name);
+ }
+
+ // Cache miss — fetch from Packagist
+ let url = format!("https://repo.packagist.org/p2/{package_name}.json");
+ let response = reqwest::blocking::get(&url)?;
+
+ if !response.status().is_success() {
+ anyhow::bail!(
+ "Failed to fetch package \"{package_name}\" from Packagist (HTTP {})",
+ response.status()
+ );
+ }
+
+ let body = response.text()?;
+
+ // Write to cache
+ if let Some(cache) = repo_cache {
+ let _ = cache.write(&cache_key, &body);
+ }
+
+ parse_p2_response(&body, package_name)
+}
+
+// ─────────────────────────────────────────────────────────────────────────────
+// Packagist search API
+// ─────────────────────────────────────────────────────────────────────────────
+
+/// A single search result from the Packagist search API.
+#[derive(Debug, Deserialize, Serialize, Clone)]
+pub struct SearchResult {
+ pub name: String,
+ pub description: String,
+ pub url: String,
+ pub repository: Option<String>,
+ pub downloads: u64,
+ pub favers: u64,
+}
+
+#[derive(Debug, Deserialize)]
+pub struct SearchResponse {
+ pub results: Vec<SearchResult>,
+ pub total: u64,
+ pub next: Option<String>,
+}
+
+/// Maximum number of pages to fetch from the Packagist search API.
+const SEARCH_MAX_PAGES: usize = 20;
+
+/// Percent-encode a string for use in a URL query parameter value.
+fn url_encode(s: &str) -> String {
+ let mut encoded = String::with_capacity(s.len());
+ for byte in s.bytes() {
+ match byte {
+ b'A'..=b'Z' | b'a'..=b'z' | b'0'..=b'9' | b'-' | b'_' | b'.' | b'~' => {
+ encoded.push(byte as char);
+ }
+ b' ' => encoded.push_str("%20"),
+ other => {
+ encoded.push_str(&format!("%{other:02X}"));
+ }
+ }
+ }
+ encoded
+}
+
+/// Search Packagist for packages matching `query`.
+///
+/// Fetches up to `SEARCH_MAX_PAGES` pages of results and returns the full list.
+/// An optional `package_type` filter can narrow results (e.g. `"library"`).
+pub fn search_packages(
+ query: &str,
+ package_type: Option<&str>,
+) -> anyhow::Result<(Vec<SearchResult>, u64)> {
+ let client = reqwest::blocking::Client::builder()
+ .user_agent("mozart/0.1.0")
+ .build()?;
+
+ let mut all_results: Vec<SearchResult> = Vec::new();
+ let mut page = 1usize;
+ let mut next_url: Option<String> = None;
+ let mut total: u64 = 0;
+
+ loop {
+ let response: SearchResponse = if let Some(ref url) = next_url {
+ let resp = client.get(url).send()?;
+ if !resp.status().is_success() {
+ anyhow::bail!("Packagist search request failed (HTTP {})", resp.status());
+ }
+ resp.json()?
+ } else {
+ let encoded_query = url_encode(query);
+ let mut url = format!("https://packagist.org/search.json?q={encoded_query}");
+ if let Some(t) = package_type {
+ url.push_str("&type=");
+ url.push_str(&url_encode(t));
+ }
+
+ let resp = client.get(&url).send()?;
+ if !resp.status().is_success() {
+ anyhow::bail!("Packagist search request failed (HTTP {})", resp.status());
+ }
+ resp.json()?
+ };
+
+ if page == 1 {
+ total = response.total;
+ }
+
+ all_results.extend(response.results);
+ next_url = response.next;
+ page += 1;
+
+ if next_url.is_none() || page > SEARCH_MAX_PAGES {
+ break;
+ }
+ }
+
+ Ok((all_results, total))
+}
+
+// ─────────────────────────────────────────────────────────────────────────────
+// Security Advisories API
+// ─────────────────────────────────────────────────────────────────────────────
+
+/// A single security advisory from the Packagist API.
+#[derive(Debug, Clone, Deserialize, Serialize)]
+pub struct SecurityAdvisory {
+ #[serde(rename = "advisoryId")]
+ pub advisory_id: String,
+
+ #[serde(rename = "packageName")]
+ pub package_name: String,
+
+ #[serde(rename = "remoteId")]
+ pub remote_id: String,
+
+ pub title: String,
+
+ pub link: Option<String>,
+
+ pub cve: Option<String>,
+
+ /// Composer version constraint string, e.g. ">=1.0,<1.5.1|>=2.0,<2.3"
+ #[serde(rename = "affectedVersions")]
+ pub affected_versions: String,
+
+ pub source: String,
+
+ #[serde(rename = "reportedAt")]
+ pub reported_at: String,
+
+ #[serde(rename = "composerRepository")]
+ pub composer_repository: Option<String>,
+
+ pub severity: Option<String>,
+
+ #[serde(default)]
+ pub sources: Vec<AdvisorySource>,
+}
+
+/// A source entry within a security advisory.
+#[derive(Debug, Clone, Deserialize, Serialize)]
+pub struct AdvisorySource {
+ pub name: String,
+ #[serde(rename = "remoteId")]
+ pub remote_id: String,
+}
+
+/// Response from POST `https://packagist.org/api/security-advisories/`.
+#[derive(Debug, Deserialize)]
+pub struct SecurityAdvisoriesResponse {
+ pub advisories: BTreeMap<String, Vec<SecurityAdvisory>>,
+}
+
+/// Fetch security advisories for the given package names from the Packagist API.
+///
+/// Sends a POST request to `https://packagist.org/api/security-advisories/`
+/// with form-encoded package names. Returns advisories grouped by package name.
+///
+/// If the package list is very large (500+), requests are batched in chunks of
+/// 500 names per request and the results are merged.
+pub fn fetch_security_advisories(
+ package_names: &[&str],
+) -> anyhow::Result<BTreeMap<String, Vec<SecurityAdvisory>>> {
+ let client = reqwest::blocking::Client::builder()
+ .user_agent("mozart/0.1.0")
+ .build()?;
+
+ let mut all_advisories: BTreeMap<String, Vec<SecurityAdvisory>> = BTreeMap::new();
+
+ for chunk in package_names.chunks(500) {
+ // Build an application/x-www-form-urlencoded body manually.
+ // Each package is encoded as `packages[]=<name>` and joined with `&`.
+ let body: String = chunk
+ .iter()
+ .map(|name| format!("packages[]={}", url_encode(name)))
+ .collect::<Vec<_>>()
+ .join("&");
+
+ let response = client
+ .post("https://packagist.org/api/security-advisories/")
+ .header("Content-Type", "application/x-www-form-urlencoded")
+ .body(body)
+ .send()?;
+
+ if !response.status().is_success() {
+ anyhow::bail!(
+ "Packagist security advisories request failed (HTTP {})",
+ response.status()
+ );
+ }
+
+ let parsed: SecurityAdvisoriesResponse = response.json()?;
+
+ for (pkg_name, advisories) in parsed.advisories {
+ if !advisories.is_empty() {
+ all_advisories
+ .entry(pkg_name)
+ .or_default()
+ .extend(advisories);
+ }
+ }
+ }
+
+ Ok(all_advisories)
+}
+
+#[cfg(test)]
+mod tests {
+ use super::*;
+
+ #[test]
+ fn parse_p2_response_basic() {
+ let json = r#"{
+ "packages": {
+ "monolog/monolog": [
+ {
+ "version": "3.8.0",
+ "version_normalized": "3.8.0.0",
+ "require": {"php": ">=8.1"},
+ "dist": {
+ "type": "zip",
+ "url": "https://example.com/monolog-3.8.0.zip",
+ "reference": "abc123",
+ "shasum": ""
+ },
+ "source": {
+ "type": "git",
+ "url": "https://github.com/Seldaek/monolog.git",
+ "reference": "abc123"
+ }
+ },
+ {
+ "version": "3.7.0",
+ "version_normalized": "3.7.0.0",
+ "require": {"php": ">=8.1"}
+ }
+ ]
+ }
+ }"#;
+
+ let versions = parse_p2_response(json, "monolog/monolog").unwrap();
+ assert_eq!(versions.len(), 2);
+ assert_eq!(versions[0].version, "3.8.0");
+ assert_eq!(versions[0].version_normalized, "3.8.0.0");
+ assert_eq!(versions[0].require.get("php").unwrap(), ">=8.1");
+ assert!(versions[0].dist.is_some());
+ assert!(versions[0].source.is_some());
+ assert_eq!(versions[1].version, "3.7.0");
+ assert!(versions[1].dist.is_none());
+ }
+
+ #[test]
+ fn parse_p2_response_not_found() {
+ let json = r#"{"packages": {"other/pkg": []}}"#;
+ let result = parse_p2_response(json, "monolog/monolog");
+ assert!(result.is_err());
+ }
+
+ #[test]
+ fn parse_p2_response_with_dev_version() {
+ let json = r#"{
+ "packages": {
+ "test/pkg": [
+ {
+ "version": "dev-master",
+ "version_normalized": "dev-master",
+ "require": {}
+ },
+ {
+ "version": "1.0.0",
+ "version_normalized": "1.0.0.0",
+ "require": {}
+ }
+ ]
+ }
+ }"#;
+
+ let versions = parse_p2_response(json, "test/pkg").unwrap();
+ assert_eq!(versions.len(), 2);
+ assert_eq!(versions[0].version, "dev-master");
+ assert_eq!(versions[1].version, "1.0.0");
+ }
+
+ // ──────────── branch_aliases() tests ────────────
+
+ #[test]
+ fn test_branch_aliases_present() {
+ let json = r#"{
+ "packages": {
+ "test/pkg": [
+ {
+ "version": "dev-master",
+ "version_normalized": "dev-master",
+ "require": {},
+ "extra": {
+ "branch-alias": {
+ "dev-master": "2.x-dev"
+ }
+ }
+ }
+ ]
+ }
+ }"#;
+
+ let versions = parse_p2_response(json, "test/pkg").unwrap();
+ let aliases = versions[0].branch_aliases();
+ assert_eq!(aliases.len(), 1);
+ assert_eq!(aliases.get("dev-master").unwrap(), "2.x-dev");
+ }
+
+ #[test]
+ fn test_branch_aliases_multiple() {
+ let json = r#"{
+ "packages": {
+ "test/pkg": [
+ {
+ "version": "dev-master",
+ "version_normalized": "dev-master",
+ "require": {},
+ "extra": {
+ "branch-alias": {
+ "dev-master": "2.x-dev",
+ "dev-1.x": "1.5.x-dev"
+ }
+ }
+ }
+ ]
+ }
+ }"#;
+
+ let versions = parse_p2_response(json, "test/pkg").unwrap();
+ let aliases = versions[0].branch_aliases();
+ assert_eq!(aliases.len(), 2);
+ assert_eq!(aliases.get("dev-master").unwrap(), "2.x-dev");
+ assert_eq!(aliases.get("dev-1.x").unwrap(), "1.5.x-dev");
+ }
+
+ #[test]
+ fn test_branch_aliases_no_extra() {
+ let json = r#"{
+ "packages": {
+ "test/pkg": [
+ {
+ "version": "dev-master",
+ "version_normalized": "dev-master",
+ "require": {}
+ }
+ ]
+ }
+ }"#;
+
+ let versions = parse_p2_response(json, "test/pkg").unwrap();
+ let aliases = versions[0].branch_aliases();
+ assert!(aliases.is_empty());
+ }
+
+ #[test]
+ fn test_branch_aliases_extra_without_branch_alias_key() {
+ let json = r#"{
+ "packages": {
+ "test/pkg": [
+ {
+ "version": "dev-master",
+ "version_normalized": "dev-master",
+ "require": {},
+ "extra": {
+ "installer-name": "my-plugin"
+ }
+ }
+ ]
+ }
+ }"#;
+
+ let versions = parse_p2_response(json, "test/pkg").unwrap();
+ let aliases = versions[0].branch_aliases();
+ assert!(aliases.is_empty());
+ }
+
+ // ──────────── SecurityAdvisory parsing tests ─────────────────────────────
+
+ #[test]
+ fn test_parse_security_advisories_response() {
+ let json = r#"{
+ "advisories": {
+ "monolog/monolog": [
+ {
+ "advisoryId": "PKSA-b2m0-qqf7-qck4",
+ "packageName": "monolog/monolog",
+ "remoteId": "monolog/monolog/2017-11-13-1.yaml",
+ "title": "Header injection in NativeMailerHandler",
+ "link": "https://github.com/Seldaek/monolog/pull/683",
+ "cve": null,
+ "affectedVersions": ">=1.8.0,<1.12.0",
+ "source": "FriendsOfPHP/security-advisories",
+ "reportedAt": "2017-11-13T00:00:00+00:00",
+ "composerRepository": "https://packagist.org",
+ "severity": "low",
+ "sources": [
+ {
+ "name": "FriendsOfPHP/security-advisories",
+ "remoteId": "monolog/monolog/2017-11-13-1.yaml"
+ }
+ ]
+ }
+ ]
+ }
+ }"#;
+
+ let response: SecurityAdvisoriesResponse = serde_json::from_str(json).unwrap();
+ assert_eq!(response.advisories.len(), 1);
+ let advisories = response.advisories.get("monolog/monolog").unwrap();
+ assert_eq!(advisories.len(), 1);
+ let adv = &advisories[0];
+ assert_eq!(adv.advisory_id, "PKSA-b2m0-qqf7-qck4");
+ assert_eq!(adv.package_name, "monolog/monolog");
+ assert_eq!(adv.title, "Header injection in NativeMailerHandler");
+ assert_eq!(adv.affected_versions, ">=1.8.0,<1.12.0");
+ assert_eq!(adv.severity.as_deref(), Some("low"));
+ assert!(adv.cve.is_none());
+ assert_eq!(adv.sources.len(), 1);
+ assert_eq!(adv.sources[0].name, "FriendsOfPHP/security-advisories");
+ }
+
+ #[test]
+ fn test_parse_security_advisories_empty() {
+ let json = r#"{"advisories": {"other/package": []}}"#;
+ let response: SecurityAdvisoriesResponse = serde_json::from_str(json).unwrap();
+ assert_eq!(response.advisories.len(), 1);
+ let advisories = response.advisories.get("other/package").unwrap();
+ assert!(advisories.is_empty());
+ }
+
+ #[test]
+ fn test_parse_security_advisories_null_fields() {
+ let json = r#"{
+ "advisories": {
+ "vendor/pkg": [
+ {
+ "advisoryId": "PKSA-0000-0000-0000",
+ "packageName": "vendor/pkg",
+ "remoteId": "vendor/pkg/2024-01-01.yaml",
+ "title": "Some vulnerability",
+ "link": null,
+ "cve": null,
+ "affectedVersions": ">=1.0,<2.0",
+ "source": "FriendsOfPHP/security-advisories",
+ "reportedAt": "2024-01-01T00:00:00+00:00",
+ "composerRepository": null,
+ "severity": null,
+ "sources": []
+ }
+ ]
+ }
+ }"#;
+
+ let response: SecurityAdvisoriesResponse = serde_json::from_str(json).unwrap();
+ let advisories = response.advisories.get("vendor/pkg").unwrap();
+ assert_eq!(advisories.len(), 1);
+ let adv = &advisories[0];
+ assert!(adv.link.is_none());
+ assert!(adv.cve.is_none());
+ assert!(adv.severity.is_none());
+ assert!(adv.composer_repository.is_none());
+ assert!(adv.sources.is_empty());
+ }
+}