Let fetch.Resolver take base URLs and ecosystem aliases

README.md

@@ -486,6 +486,15 @@ info, _ := resolver.Resolve(ctx, "npm", "lodash", "4.17.21")
 // info.URL = "https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz"
 ```
 
+`WithBaseURLs` points those patterns at a mirror instead of the public host. Ecosystems left out of the map keep their defaults:
+
+```go
+resolver := fetch.NewResolver(fetch.WithBaseURLs(map[string]string{
+	"npm":   "https://npm.mycompany.com",
+	"cargo": "https://crates.mycompany.com/crates",
+}))
+```
+
 ## Private Registries
 
 PURLs with a `repository_url` qualifier automatically use that URL:

fetch/resolver.go

@@ -9,6 +9,7 @@ import (
 	"strings"
 	"unicode"
 
+	"github.com/git-pkgs/purl"
 	"github.com/git-pkgs/registries"
 	"github.com/git-pkgs/registries/client"
 )
@@ -28,21 +29,60 @@ type Registry interface {
 	URLs() client.URLBuilder
 }
 
+// defaultBaseURLs are the public registry hosts used when no registry is
+// registered for an ecosystem and WithBaseURLs does not override them.
+var defaultBaseURLs = map[string]string{
+	"npm":    "https://registry.npmjs.org",
+	"cargo":  "https://static.crates.io/crates",
+	"gem":    "https://rubygems.org",
+	"golang": "https://proxy.golang.org",
+	"hex":    "https://repo.hex.pm",
+	"pub":    "https://pub.dev",
+	"maven":  "https://repo1.maven.org/maven2",
+	"nuget":  "https://api.nuget.org",
+}
+
 // Resolver determines download URLs for package artifacts.
 type Resolver struct {
 	registries map[string]Registry
+	baseURLs   map[string]string
+}
+
+// ResolverOption configures a Resolver.
+type ResolverOption func(*Resolver)
+
+// WithBaseURLs overrides the base URL used for ecosystems resolved without a
+// registered registry, keyed by ecosystem. Ecosystems not in the map keep
+// the public registry host.
+func WithBaseURLs(baseURLs map[string]string) ResolverOption {
+	return func(r *Resolver) {
+		for ecosystem, baseURL := range baseURLs {
+			if baseURL != "" {
+				r.baseURLs[purl.EcosystemToPURLType(ecosystem)] = strings.TrimRight(baseURL, "/")
+			}
+		}
+	}
 }
 
 // NewResolver creates a new URL resolver.
-func NewResolver() *Resolver {
-	return &Resolver{
+func NewResolver(opts ...ResolverOption) *Resolver {
+	r := &Resolver{
 		registries: make(map[string]Registry),
+		baseURLs:   make(map[string]string, len(defaultBaseURLs)),
+	}
+	for ecosystem, baseURL := range defaultBaseURLs {
+		r.baseURLs[ecosystem] = baseURL
+	}
+	for _, opt := range opts {
+		opt(r)
 	}
+	return r
 }
 
-// RegisterRegistry adds a registry for URL resolution.
+// RegisterRegistry adds a registry for URL resolution. Ecosystem names are
+// matched by PURL type, so "rubygems" and "gem" are the same ecosystem.
 func (r *Resolver) RegisterRegistry(reg Registry) {
-	r.registries[reg.Ecosystem()] = reg
+	r.registries[purl.EcosystemToPURLType(reg.Ecosystem())] = reg
 }
 
 // ArtifactInfo contains information about a downloadable artifact.
@@ -73,6 +113,7 @@ func (r *Resolver) ResolveWithOptions(
 	version string,
 	options ResolveOptions,
 ) (*ArtifactInfo, error) {
+	ecosystem = purl.EcosystemToPURLType(ecosystem)
 	reg, ok := r.registries[ecosystem]
 	if !ok {
 		info, err := r.resolveWithoutRegistry(ecosystem, name, version)
@@ -96,35 +137,36 @@ func (r *Resolver) ResolveWithOptions(
 // when no registry client is configured.
 func (r *Resolver) resolveWithoutRegistry(ecosystem, name, version string) (*ArtifactInfo, error) {
 	var url, filename string
+	base := r.baseURLs[ecosystem]
 
 	switch ecosystem {
 	case "npm":
 		shortName := name
 		if idx := strings.LastIndex(name, "/"); idx >= 0 {
 			shortName = name[idx+1:]
 		}
-		url = fmt.Sprintf("https://registry.npmjs.org/%s/-/%s-%s.tgz", name, shortName, version)
+		url = fmt.Sprintf("%s/%s/-/%s-%s.tgz", base, name, shortName, version)
 		filename = fmt.Sprintf("%s-%s.tgz", shortName, version)
 
 	case "cargo":
-		url = fmt.Sprintf("https://static.crates.io/crates/%s/%s-%s.crate", name, name, version)
+		url = fmt.Sprintf("%s/%s/%s-%s.crate", base, name, name, version)
 		filename = fmt.Sprintf("%s-%s.crate", name, version)
 
 	case "gem":
-		url = fmt.Sprintf("https://rubygems.org/downloads/%s-%s.gem", name, version)
+		url = fmt.Sprintf("%s/downloads/%s-%s.gem", base, name, version)
 		filename = fmt.Sprintf("%s-%s.gem", name, version)
 
 	case "golang":
 		encoded := encodeGoModule(name)
-		url = fmt.Sprintf("https://proxy.golang.org/%s/@v/%s.zip", encoded, version)
+		url = fmt.Sprintf("%s/%s/@v/%s.zip", base, encoded, version)
 		filename = fmt.Sprintf("%s@%s.zip", lastPathComponent(name), version)
 
 	case "hex":
-		url = fmt.Sprintf("https://repo.hex.pm/tarballs/%s-%s.tar", name, version)
+		url = fmt.Sprintf("%s/tarballs/%s-%s.tar", base, name, version)
 		filename = fmt.Sprintf("%s-%s.tar", name, version)
 
 	case "pub":
-		url = fmt.Sprintf("https://pub.dev/packages/%s/versions/%s.tar.gz", name, version)
+		url = fmt.Sprintf("%s/packages/%s/versions/%s.tar.gz", base, name, version)
 		filename = fmt.Sprintf("%s-%s.tar.gz", name, version)
 
 	case "maven":
@@ -134,13 +176,13 @@ func (r *Resolver) resolveWithoutRegistry(ecosystem, name, version string) (*Art
 			return nil, fmt.Errorf("invalid maven name format, expected group:artifact")
 		}
 		group = strings.ReplaceAll(group, ".", "/")
-		url = fmt.Sprintf("https://repo1.maven.org/maven2/%s/%s/%s/%s-%s.jar", group, artifact, version, artifact, version)
+		url = fmt.Sprintf("%s/%s/%s/%s/%s-%s.jar", base, group, artifact, version, artifact, version)
 		filename = fmt.Sprintf("%s-%s.jar", artifact, version)
 
 	case "nuget":
 		// NuGet package IDs are case-insensitive, use lowercase
 		lowername := strings.ToLower(name)
-		url = fmt.Sprintf("https://api.nuget.org/v3-flatcontainer/%s/%s/%s.%s.nupkg", lowername, version, lowername, version)
+		url = fmt.Sprintf("%s/v3-flatcontainer/%s/%s/%s.%s.nupkg", base, lowername, version, lowername, version)
 		filename = fmt.Sprintf("%s.%s.nupkg", lowername, version)
 
 	default:

fetch/resolver_test.go

@@ -94,6 +94,59 @@ func TestResolveWithoutRegistry(t *testing.T) {
 	}
 }
 
+func TestResolveWithBaseURLs(t *testing.T) {
+	r := NewResolver(WithBaseURLs(map[string]string{
+		"npm":    "https://npm.internal/",
+		"cargo":  "https://crates.internal/crates",
+		"gem":    "https://gems.internal",
+		"golang": "https://goproxy.internal",
+		"hex":    "https://hex.internal",
+		"pub":    "https://pub.internal",
+		"maven":  "https://maven.internal/releases",
+		"nuget":  "https://nuget.internal",
+	}))
+
+	tests := []struct {
+		ecosystem string
+		name      string
+		version   string
+		wantURL   string
+	}{
+		{"npm", "@babel/core", "7.23.0", "https://npm.internal/@babel/core/-/core-7.23.0.tgz"},
+		{"cargo", "serde", "1.0.193", "https://crates.internal/crates/serde/serde-1.0.193.crate"},
+		{"gem", "rails", "7.1.2", "https://gems.internal/downloads/rails-7.1.2.gem"},
+		{"golang", "github.com/Azure/azure-sdk-for-go", "v68.0.0", "https://goproxy.internal/github.com/!azure/azure-sdk-for-go/@v/v68.0.0.zip"},
+		{"hex", "phoenix", "1.7.10", "https://hex.internal/tarballs/phoenix-1.7.10.tar"},
+		{"pub", "flutter", "3.16.0", "https://pub.internal/packages/flutter/versions/3.16.0.tar.gz"},
+		{"maven", "com.google.guava:guava", "32.1.3-jre", "https://maven.internal/releases/com/google/guava/guava/32.1.3-jre/guava-32.1.3-jre.jar"},
+		{"nuget", "Newtonsoft.Json", "13.0.3", "https://nuget.internal/v3-flatcontainer/newtonsoft.json/13.0.3/newtonsoft.json.13.0.3.nupkg"},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.ecosystem, func(t *testing.T) {
+			info, err := r.Resolve(context.Background(), tt.ecosystem, tt.name, tt.version)
+			if err != nil {
+				t.Fatalf("Resolve failed: %v", err)
+			}
+			if info.URL != tt.wantURL {
+				t.Errorf("URL = %q, want %q", info.URL, tt.wantURL)
+			}
+		})
+	}
+}
+
+func TestResolveWithBaseURLsKeepsDefaultsForOtherEcosystems(t *testing.T) {
+	r := NewResolver(WithBaseURLs(map[string]string{"npm": "https://npm.internal", "cargo": ""}))
+
+	info, err := r.Resolve(context.Background(), "cargo", "serde", "1.0.193")
+	if err != nil {
+		t.Fatalf("Resolve failed: %v", err)
+	}
+	if want := "https://static.crates.io/crates/serde/serde-1.0.193.crate"; info.URL != want {
+		t.Errorf("URL = %q, want %q", info.URL, want)
+	}
+}
+
 func TestResolveUnsupportedEcosystem(t *testing.T) {
 	r := NewResolver()
 
@@ -301,3 +354,64 @@ func TestFilenameFromURL(t *testing.T) {
 		}
 	}
 }
+
+type namedRegistry struct {
+	fakeRegistry
+	ecosystem string
+}
+
+func (n *namedRegistry) Ecosystem() string { return n.ecosystem }
+
+func TestResolveEcosystemAliases(t *testing.T) {
+	for _, ecosystem := range []string{"gem", "rubygems", "RubyGems"} {
+		t.Run(ecosystem, func(t *testing.T) {
+			info, err := NewResolver().Resolve(context.Background(), ecosystem, "rake", "13.2.1")
+			if err != nil {
+				t.Fatalf("Resolve failed: %v", err)
+			}
+			if want := "https://rubygems.org/downloads/rake-13.2.1.gem"; info.URL != want {
+				t.Errorf("URL = %q, want %q", info.URL, want)
+			}
+		})
+	}
+}
+
+func TestResolveRegisteredEcosystemAliases(t *testing.T) {
+	want := "https://repo.packagist.org/dist/monolog/monolog-3.5.0.zip"
+	for _, tt := range []struct{ registered, requested string }{
+		{"composer", "packagist"},
+		{"packagist", "composer"},
+		{"composer", "composer"},
+	} {
+		t.Run(tt.registered+"/"+tt.requested, func(t *testing.T) {
+			r := NewResolver()
+			r.RegisterRegistry(&namedRegistry{
+				ecosystem: tt.registered,
+				fakeRegistry: fakeRegistry{versions: []registries.Version{{
+					Number:   "3.5.0",
+					Metadata: map[string]any{"download_url": want},
+				}}},
+			})
+
+			info, err := r.Resolve(context.Background(), tt.requested, "monolog/monolog", "3.5.0")
+			if err != nil {
+				t.Fatalf("Resolve failed: %v", err)
+			}
+			if info.URL != want {
+				t.Errorf("URL = %q, want %q", info.URL, want)
+			}
+		})
+	}
+}
+
+func TestWithBaseURLsAcceptsEcosystemAliases(t *testing.T) {
+	r := NewResolver(WithBaseURLs(map[string]string{"rubygems": "https://gems.internal"}))
+
+	info, err := r.Resolve(context.Background(), "gem", "rake", "13.2.1")
+	if err != nil {
+		t.Fatalf("Resolve failed: %v", err)
+	}
+	if want := "https://gems.internal/downloads/rake-13.2.1.gem"; info.URL != want {
+		t.Errorf("URL = %q, want %q", info.URL, want)
+	}
+}