Enable storage retention for cargo

docs/configuration.md

@@ -89,7 +89,7 @@ A package rule wins over an ecosystem rule, which wins over `default`. Package r
 | Ecosystem key | Retention supported | Package rule form |
 |---------------|---------------------|-------------------|
 | `alpine` | yes | `pkg:apk/curl` or `pkg:apk/alpine/curl`; case doesn't matter. A rule covers the package in every configured APK repository. Alpine mirrors keep only the current build of each package per release, so an evicted older one can't be fetched again. |
-| `cargo` | no | - |
+| `cargo` | yes | `pkg:cargo/serde`, the crate name as published; case, `-` and `_` have to match, so `pkg:cargo/inflector` doesn't cover `Inflector`. |
 | `composer` | no | - |
 | `conan` | no | - |
 | `conda` | no | - |

internal/handler/cargo.go

@@ -9,6 +9,8 @@ import (
 	"net/url"
 	"strings"
 	"time"
+
+	"github.com/git-pkgs/proxy/internal/retention"
 )
 
 const (
@@ -20,6 +22,14 @@ const (
 	cargoIndexLen3 = 3
 )
 
+// Crates are stored under pkg:cargo/{name}, with the name cargo puts in the
+// download URL, which is the name as published. purl keeps cargo names as
+// they are, so package rules use the default key mapping and have to match
+// that name exactly.
+func init() {
+	retention.Default.Register(retention.Spec{Key: "cargo", CanonicalPackage: retention.DefaultCanonical("cargo")})
+}
+
 // CargoHandler handles cargo registry protocol requests.
 type CargoHandler struct {
 	proxy       *Proxy

internal/server/retention_cargo_test.go

@@ -0,0 +1,239 @@
+package server
+
+import (
+	"context"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"strings"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/git-pkgs/proxy/internal/config"
+	"github.com/git-pkgs/proxy/internal/database"
+	"github.com/git-pkgs/proxy/internal/handler"
+	"github.com/git-pkgs/proxy/internal/mirror"
+	"github.com/git-pkgs/proxy/internal/retention"
+	"github.com/git-pkgs/registries/fetch"
+)
+
+// cargoRetentionUpstream serves crate files the way static.crates.io does,
+// counting downloads.
+func cargoRetentionUpstream(t *testing.T) (*httptest.Server, *atomic.Int64) {
+	t.Helper()
+	var downloads atomic.Int64
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		switch r.URL.Path {
+		case "/crates/serde/serde-1.0.0.crate", "/crates/Inflector/Inflector-0.11.4.crate":
+			downloads.Add(1)
+			_, _ = io.WriteString(w, "crate bytes")
+		default:
+			http.NotFound(w, r)
+		}
+	}))
+	t.Cleanup(srv.Close)
+	return srv, &downloads
+}
+
+func newCargoRetentionProxy(t *testing.T, upstream *httptest.Server) (*httptest.Server, *database.DB) {
+	t.Helper()
+	db, store := setupEvictionTest(t)
+	// The upstream is a plain-HTTP loopback server; the default fetcher
+	// refuses loopback, a stock client reaches it.
+	fetcher := fetch.NewFetcher(fetch.WithHTTPClient(upstream.Client()), fetch.WithMaxRetries(0))
+	t.Cleanup(func() { _ = fetcher.Close() })
+	proxy := handler.NewProxy(db, store, fetcher, fetch.NewResolver(), discardLogger())
+	proxy.HTTPClient = upstream.Client()
+	h := handler.NewCargoHandler(proxy, "http://proxy.test", upstream.URL, upstream.URL+"/crates")
+	srv := httptest.NewServer(h.Routes())
+	t.Cleanup(srv.Close)
+	return srv, db
+}
+
+func cargoGet(t *testing.T, url string) {
+	t.Helper()
+	resp, err := http.Get(url)
+	if err != nil {
+		t.Fatalf("GET %s: %v", url, err)
+	}
+	body, _ := io.ReadAll(resp.Body)
+	_ = resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		t.Fatalf("GET %s = %d %s, want 200", url, resp.StatusCode, body)
+	}
+}
+
+// cargoArtifact returns the record of the crate file named filename.
+func cargoArtifact(t *testing.T, db *database.DB, filename string) database.Artifact {
+	t.Helper()
+	var art database.Artifact
+	if err := db.Get(&art, db.Rebind(`SELECT * FROM artifacts WHERE filename = ?`), filename); err != nil {
+		t.Fatalf("reading artifact %s: %v", filename, err)
+	}
+	return art
+}
+
+func cargoBackdate(t *testing.T, db *database.DB, filename string, fetched, accessed time.Time) {
+	t.Helper()
+	query := db.Rebind(`UPDATE artifacts SET fetched_at = ?, last_accessed_at = ? WHERE filename = ?`)
+	if _, err := db.Exec(query, fetched, accessed, filename); err != nil {
+		t.Fatalf("backdating %s: %v", filename, err)
+	}
+}
+
+func TestCargoRegistersForRetention(t *testing.T) {
+	if _, ok := retention.Default.Lookup("cargo"); !ok {
+		t.Fatal("cargo is not registered for retention")
+	}
+	if key, ok := retention.Default.KeyForDBEcosystem("cargo"); !ok || key != "cargo" {
+		t.Errorf("KeyForDBEcosystem(cargo) = %q, %v, want cargo, true", key, ok)
+	}
+}
+
+// A crate served from the cache stays while it is used, is evicted once its
+// rule runs out, comes back from upstream on the next download, and stays
+// after that even with the evicted copy's old access time on the record.
+func TestCargoRetentionEvictsAndRefetches(t *testing.T) {
+	upstream, downloads := cargoRetentionUpstream(t)
+	srv, db := newCargoRetentionProxy(t, upstream)
+	const crate = "serde-1.0.0.crate"
+
+	cargoGet(t, srv.URL+"/crates/serde/1.0.0/download")
+	cargoGet(t, srv.URL+"/crates/serde/1.0.0/download")
+	if got := downloads.Load(); got != 1 {
+		t.Fatalf("upstream downloads = %d, want 1", got)
+	}
+	if !cargoArtifact(t, db, crate).LastAccessedAt.Valid {
+		t.Fatal("cache hit did not record an access")
+	}
+
+	rules := retention.Rules{Ecosystems: map[string]time.Duration{"cargo": retentionDay}}
+	now := time.Now()
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, now)
+	if !cargoArtifact(t, db, crate).StoragePath.Valid {
+		t.Fatal("crate used today was evicted")
+	}
+
+	old := now.Add(-2 * retentionDay)
+	cargoBackdate(t, db, crate, old, old)
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, now)
+	if cargoArtifact(t, db, crate).StoragePath.Valid {
+		t.Fatal("crate unused for two days kept under a one-day rule")
+	}
+
+	cargoGet(t, srv.URL+"/crates/serde/1.0.0/download")
+	if got := downloads.Load(); got != 2 {
+		t.Errorf("upstream downloads = %d, want 2 after the eviction", got)
+	}
+	art := cargoArtifact(t, db, crate)
+	cargoBackdate(t, db, crate, art.FetchedAt.Time, now.Add(-30*retentionDay))
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, time.Now())
+	if !cargoArtifact(t, db, crate).StoragePath.Valid {
+		t.Error("refetched crate evicted again by the next sweep")
+	}
+}
+
+// A key with a namespace parses as a cargo PURL but can never match a
+// stored crate, so it is refused at validation. A plain key is accepted.
+func TestCargoRetentionRefusesNamespacedKeys(t *testing.T) {
+	cfg := config.Default()
+	cfg.Storage.Retention.Packages = map[string]string{"pkg:cargo/serde": "0"}
+	if _, err := cfg.RetentionRules(retention.Default); err != nil {
+		t.Fatalf("RetentionRules(pkg:cargo/serde): %v", err)
+	}
+	cfg.Storage.Retention.Packages = map[string]string{"pkg:cargo/rust-lang/serde": "0"}
+	_, err := cfg.RetentionRules(retention.Default)
+	if err == nil || !strings.Contains(err.Error(), "does not match how cargo packages are stored") {
+		t.Errorf("RetentionRules(pkg:cargo/rust-lang/serde) error = %v, want the key refused", err)
+	}
+}
+
+// A package rule reaches a crate under its published name, case included;
+// a rule with the name in another case covers nothing.
+func TestCargoRetentionPackageRules(t *testing.T) {
+	upstream, _ := cargoRetentionUpstream(t)
+	srv, db := newCargoRetentionProxy(t, upstream)
+	const inflector, serde = "Inflector-0.11.4.crate", "serde-1.0.0.crate"
+
+	tests := []struct {
+		key           string
+		keepInflector bool
+	}{
+		{"pkg:cargo/Inflector", true},
+		{"pkg:cargo/inflector", false},
+	}
+	for _, tt := range tests {
+		cargoGet(t, srv.URL+"/crates/Inflector/0.11.4/download")
+		cargoGet(t, srv.URL+"/crates/serde/1.0.0/download")
+		old := time.Now().Add(-30 * retentionDay)
+		cargoBackdate(t, db, inflector, old, old)
+		cargoBackdate(t, db, serde, old, old)
+
+		cfg := config.Default()
+		cfg.Storage.Retention.Ecosystems = map[string]string{"cargo": "1d"}
+		cfg.Storage.Retention.Packages = map[string]string{tt.key: "0"}
+		rules, err := cfg.RetentionRules(retention.Default)
+		if err != nil {
+			t.Fatalf("RetentionRules(%s): %v", tt.key, err)
+		}
+		sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, time.Now())
+
+		if got := cargoArtifact(t, db, inflector).StoragePath.Valid; got != tt.keepInflector {
+			t.Errorf("with %s: Inflector cached = %v, want %v", tt.key, got, tt.keepInflector)
+		}
+		if cargoArtifact(t, db, serde).StoragePath.Valid {
+			t.Errorf("with %s: serde kept past the cargo rule", tt.key)
+		}
+	}
+}
+
+// cargoRewriteTransport sends every request to the test upstream, so the
+// mirror command's fixed static.crates.io URLs reach it.
+type cargoRewriteTransport struct {
+	target *url.URL
+	base   http.RoundTripper
+}
+
+func (r cargoRewriteTransport) RoundTrip(req *http.Request) (*http.Response, error) {
+	req = req.Clone(req.Context())
+	req.URL.Scheme = r.target.Scheme
+	req.URL.Host = r.target.Host
+	req.Host = r.target.Host
+	return r.base.RoundTrip(req)
+}
+
+// A crate preloaded with the mirror command falls under the cargo rule.
+func TestCargoRetentionCoversMirroredCrates(t *testing.T) {
+	upstream, downloads := cargoRetentionUpstream(t)
+	target, err := url.Parse(upstream.URL)
+	if err != nil {
+		t.Fatal(err)
+	}
+	db, store := setupEvictionTest(t)
+	client := &http.Client{Transport: cargoRewriteTransport{target: target, base: upstream.Client().Transport}}
+	fetcher := fetch.NewFetcher(fetch.WithHTTPClient(client), fetch.WithMaxRetries(0))
+	t.Cleanup(func() { _ = fetcher.Close() })
+	proxy := handler.NewProxy(db, store, fetcher, fetch.NewResolver(), discardLogger())
+
+	m := mirror.New(proxy, db, store, discardLogger(), 1)
+	progress, err := m.Run(context.Background(), &mirror.PURLSource{PURLs: []string{"pkg:cargo/serde@1.0.0"}})
+	if err != nil || progress.Failed != 0 || downloads.Load() != 1 {
+		t.Fatalf("mirror run = %+v, %v, downloads %d; want one successful download", progress, err, downloads.Load())
+	}
+
+	const crate = "serde-1.0.0.crate"
+	if cargoArtifact(t, db, crate).LastAccessedAt.Valid {
+		t.Fatal("mirrored crate has an access recorded before anyone downloaded it")
+	}
+	query := db.Rebind(`UPDATE artifacts SET fetched_at = ? WHERE filename = ?`)
+	if _, err := db.Exec(query, time.Now().Add(-30*retentionDay), crate); err != nil {
+		t.Fatalf("backdating %s: %v", crate, err)
+	}
+	rules := retention.Rules{Ecosystems: map[string]time.Duration{"cargo": retentionDay}}
+	sweepRetention(context.Background(), db, discardLogger(), retention.Default, rules, time.Now())
+	if cargoArtifact(t, db, crate).StoragePath.Valid {
+		t.Error("never-served mirrored crate kept past the cargo rule")
+	}
+}