mesh-controller/internal/builder/mirror_shared_test.go

package builder

import (
	"context"
	"net/http"
	"net/http/httptest"
	"strings"
	"sync"
	"testing"
)

// One copy per upstream image, whichever modules stand on it (novox/hq ADR 0257).

func TestAnUpstreamImageHasOneRepositoryNamedForIt(t *testing.T) {
	for from, want := range map[string]string{
		"golang@sha256:abc":                       "upstream/docker.io/library/golang",
		"n8nio/n8n@sha256:abc":                    "upstream/docker.io/n8nio/n8n",
		"quay.io/minio/mc@sha256:abc":             "upstream/quay.io/minio/mc",
		"ghcr.io/Mailu/Admin@sha256:abc":          "upstream/ghcr.io/mailu/admin",
		"localhost:5000/x/y@sha256:abc":           "upstream/localhost-5000/x/y",
		"docker.io/library/alpine:3.20@sha256:ab": "upstream/docker.io/library/alpine",
	} {
		got, err := MirrorRepository(from)
		if err != nil {
			t.Fatalf("%s: %v", from, err)
		}
		if got != want {
			t.Errorf("%s: got %q want %q", from, got, want)
		}
	}
	if got := FormerMirrorRepository("route-proxy", "GO_BASE"); got != "route-proxy/on-go_base" {
		t.Fatalf("the former repository is %q", got)
	}
}

// aRegistryOfRepositories is a registry the way the real one is: blobs stored once, and each
// repository linking the blobs and manifests it holds. It mounts a blob across repositories.
type aRegistryOfRepositories struct {
	mu        sync.Mutex
	blobs     map[string][]byte          // digest → bytes, stored once
	links     map[string]map[string]bool // repository → blob digests it links
	manifests map[string][]byte          // repository@digest → document
	uploads   int
	mounts    int
	gets      int
}

func newRegistryOfRepositories() *aRegistryOfRepositories {
	return &aRegistryOfRepositories{blobs: map[string][]byte{}, links: map[string]map[string]bool{},
		manifests: map[string][]byte{}}
}

func (m *aRegistryOfRepositories) link(repository, digest string) {
	if m.links[repository] == nil {
		m.links[repository] = map[string]bool{}
	}
	m.links[repository][digest] = true
}

// split reads /v2/<repository>/<kind>/<rest> with a repository of any depth.
func splitPath(path string) (repository, kind, rest string) {
	path = strings.TrimPrefix(path, "/v2/")
	for _, k := range []string{"/manifests/", "/blobs/uploads/", "/blobs/"} {
		if i := strings.Index(path, k); i >= 0 {
			return path[:i], strings.Trim(k, "/"), path[i+len(k):]
		}
	}
	return "", "", ""
}

func (m *aRegistryOfRepositories) handler() http.Handler {
	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		m.mu.Lock()
		defer m.mu.Unlock()
		repository, kind, rest := splitPath(r.URL.Path)
		switch {
		case kind == "manifests" && (r.Method == http.MethodHead || r.Method == http.MethodGet):
			body, ok := m.manifests[repository+"@"+rest]
			if !ok {
				w.WriteHeader(http.StatusNotFound)
				return
			}
			if r.Method == http.MethodGet {
				m.gets++
				w.Header().Set("Content-Type", mediaTypeOf(body))
				_, _ = w.Write(body)
				return
			}
			w.WriteHeader(http.StatusOK)
		case kind == "manifests" && r.Method == http.MethodPut:
			body, _ := readAll(r)
			m.manifests[repository+"@"+digestOf(body)] = body
			w.WriteHeader(http.StatusCreated)
		case kind == "blobs" && (r.Method == http.MethodHead || r.Method == http.MethodGet):
			if !m.links[repository][rest] {
				w.WriteHeader(http.StatusNotFound)
				return
			}
			if r.Method == http.MethodGet {
				m.gets++
				_, _ = w.Write(m.blobs[rest])
				return
			}
			w.WriteHeader(http.StatusOK)
		case kind == "blobs/uploads" && r.Method == http.MethodPost:
			if digest, from := r.URL.Query().Get("mount"), r.URL.Query().Get("from"); digest != "" && m.links[from][digest] {
				m.link(repository, digest)
				m.mounts++
				w.WriteHeader(http.StatusCreated)
				return
			}
			w.Header().Set("Location", "/v2/"+repository+"/blobs/uploads/one")
			w.WriteHeader(http.StatusAccepted)
		case kind == "blobs/uploads" && r.Method == http.MethodPut:
			body, _ := readAll(r)
			digest := r.URL.Query().Get("digest")
			if digestOf(body) != digest {
				w.WriteHeader(http.StatusBadRequest)
				return
			}
			m.blobs[digest] = body
			m.link(repository, digest)
			m.uploads++
			w.WriteHeader(http.StatusCreated)
		default:
			w.WriteHeader(http.StatusNotFound)
		}
	})
}

func mediaTypeOf(body []byte) string {
	switch {
	case strings.Contains(string(body), mediaIndexOCI):
		return mediaIndexOCI
	default:
		return mediaManifestOCI
	}
}

// A base a module's own repository already holds is copied into the image's repository from there:
// every blob mounted, nothing asked of upstream — which may be gone, or rationing anonymous pulls.
func TestABaseHeldUnderTheModulesFormerRepositoryIsMountedNotFetchedAgain(t *testing.T) {
	src, indexDigest, srcBlobs := anUpstreamRegistry(t)
	dst := newRegistryOfRepositories()
	dstServer := httptest.NewServer(dst.handler())
	defer dstServer.Close()
	address := strings.TrimPrefix(dstServer.URL, "http://")
	r := Registry{Address: address, HTTP: src.Client()}
	host := strings.TrimPrefix(src.URL, "http://")

	// Before: copied the old way, under the module's repository.
	if _, err := r.MirrorImage(context.Background(), host+"/library/thing:latest", "hello-web/on-thing_base"); err != nil {
		t.Fatal(err)
	}
	uploaded := dst.uploads
	if uploaded != len(srcBlobs) {
		t.Fatalf("the first copy uploaded %d of %d blobs", uploaded, len(srcBlobs))
	}
	src.Close()

	from := host + "/library/thing@" + indexDigest
	repository, err := MirrorRepository(from)
	if err != nil {
		t.Fatal(err)
	}
	reference, err := r.MirrorBase(context.Background(), from, repository, "hello-web/on-thing_base")
	if err != nil {
		t.Fatalf("a base the registry holds was asked of an upstream that is gone: %v", err)
	}
	if reference != address+"/"+repository+"@"+indexDigest {
		t.Fatalf("pinned as %q", reference)
	}
	if dst.uploads != uploaded {
		t.Fatalf("the move to one repository uploaded %d blobs again", dst.uploads-uploaded)
	}
	if dst.mounts != len(srcBlobs) {
		t.Fatalf("%d of %d blobs mounted", dst.mounts, len(srcBlobs))
	}
	// The index and both platform manifests are in the image's repository, and every blob is linked.
	for key := range dst.manifests {
		if strings.HasPrefix(key, "hello-web/") {
			continue
		}
		if !strings.HasPrefix(key, repository+"@") {
			t.Fatalf("a manifest went to %s", key)
		}
	}
	if n := countPrefix(dst.manifests, repository+"@"); n != 3 {
		t.Fatalf("%d manifests in %s, want the index and two platforms", n, repository)
	}
	for digest := range srcBlobs {
		if !dst.links[repository][digest] {
			t.Fatalf("blob %s is not linked into %s", digest, repository)
		}
	}

	// A second module standing on the same image: already held, nothing copied, the same reference.
	again, err := r.MirrorBase(context.Background(), from, repository, "other-module/on-thing_base")
	if err != nil || again != reference {
		t.Fatalf("a second module's copy: %q %v", again, err)
	}
}

func countPrefix(m map[string][]byte, prefix string) int {
	n := 0
	for k := range m {
		if strings.HasPrefix(k, prefix) {
			n++
		}
	}
	return n
}

// What a build copied is said whether it worked or not: the copy is in the store either way, and a
// build that failed after copying is the one record that it is there.
func TestABuildSaysWhatItMirroredEvenWhenItFails(t *testing.T) {
	// A recipe that reaches for an image nobody declared is refused — after the base was copied.
	r, workspace := aRepository(t, anImage, map[string]string{
		"modules/bus/Dockerfile": "ARG BASE\nFROM ${BASE}\nCOPY --from=vendor/other:1 /x /x"})
	r.contents["modules/bus/module.json"] = anImage
	r.tree = "aaaa"
	m := &mirroring{recorded: r, at: "registry-a:5000"}
	got, err := Build(context.Background(), r.run, m, "https://forge.invalid/catalogue.git", "modules/bus", "", workspace,
		nil, Npmrc{}, GitCredential{}, nil)
	if err == nil {
		t.Fatal("a recipe fetching an undeclared image was built")
	}
	want := "registry-a:5000/upstream/docker.io/vendor/server@sha256:" + strings.Repeat("1", 64)
	if len(got.Mirrored) != 1 || got.Mirrored[0] != want {
		t.Fatalf("a failed build said it mirrored %v, want [%s]", got.Mirrored, want)
	}

	// And a build that works says it too.
	r2, workspace2 := aRepository(t, anImage, map[string]string{"modules/bus/Dockerfile": "ARG BASE\nFROM ${BASE}"})
	r2.contents["modules/bus/module.json"] = anImage
	r2.tree = "aaaa"
	ok, err := Build(context.Background(), r2.run, &mirroring{recorded: r2, at: "registry-a:5000"},
		"https://forge.invalid/catalogue.git", "modules/bus", "", workspace2, nil, Npmrc{}, GitCredential{}, nil)
	if err != nil {
		t.Fatal(err)
	}
	if len(ok.Mirrored) != 1 || ok.Mirrored[0] != want {
		t.Fatalf("a build said it mirrored %v, want [%s]", ok.Mirrored, want)
	}
}